haritzpuerto/the_pile_00_PubMed_Central · Datasets at Hugging Face

text stringlengths 4 2.52M	meta dict
1.. Introduction ================ Cirrhotic cardiomyopathy (CCM) was first described by Kowalski and Abelmann in 1953. The estimated incidence rate is about 56% of patients waiting for orthotopic liver transplantation (OLT) without a previous history of cardiac disease \[[@b1-ijms-15-08037]--[@b4-ijms-15-08037]\]. The CCM mechanism has been discussed relative to hyperdynamic circulation, but even after liver transplantation 7%--15% of deaths are related to cardiac-related dysfunction \[[@b5-ijms-15-08037]\]. Humoral factor abnormalities are suggested to play important roles in cirrhosis, finally causing blunted cardiac response or cardiac dysfunction \[[@b6-ijms-15-08037]\]. In liver fibrosis the transforming growth factor (TGF-β) is required and TGF-β signaling blunting can reduce fibrogenesis \[[@b7-ijms-15-08037]--[@b9-ijms-15-08037]\]. TGF-β production occurs in non-parenchymal liver cells when the liver is damaged, especially by Kupffer cells and hematopoietic stem cells (HSC) rather than fully differentiated epithelial cells \[[@b10-ijms-15-08037]\]. In cardiac remodeling TGF-β leads in modulating fibroblast phenotype and gene expression \[[@b11-ijms-15-08037]\]. TGF-β also promotes extracellular matrix deposition by upregulating collagen and fibronectin synthesis in the infarct \[[@b12-ijms-15-08037]\]. In addition, dilative ventricular remodeling by inducing interstitial fibrosis is also mediated through the TGF-β signaling pathway \[[@b12-ijms-15-08037]\]. Up to now, there has been no standard treatment because CCM does not have a solidly established diagnosis and is based on high clinical suspicion \[[@b13-ijms-15-08037]\]. In the therapeutic strategy for fibrogenesis prevention, TGF-β signaling should also be reduced using any treatment for CCM patients \[[@b14-ijms-15-08037]--[@b16-ijms-15-08037]\]. Here, silymarin and a traditional Chinese herbal medicine formula (San Huang Shel Shin Tang, SHSST) and its water-soluble beta-cyclodextrin (β-CD) complex modification compound (SHSSTc) were used in treating CCM rats, which were induced from carbon tetrachloride (CCl~4~) intraperitoneal (IP) injection induced cirrhosis model \[[@b17-ijms-15-08037]--[@b19-ijms-15-08037]\]. Silymarin is a well known drug against cirrhosis and is a cocktail-like herbal liver-protective drug with four flavonolignan isomers, silybin, silychristin, silydianin and isosilybin \[[@b20-ijms-15-08037],[@b21-ijms-15-08037]\]. SHSST is also a cocktail-like traditional herbal decoction used for liver and heart protection in China. SHSST is composed of 50% Rheum officinale Baill, 25% Scutellaria baicalnsis Geprgi and 25% Coptis chinensis Franch in weight \[[@b22-ijms-15-08037],[@b23-ijms-15-08037]\]. Rheum was reported to have a liver protection effect that can protect the liver in CCl~4~-induced injury treatment in rats. Scutellaria and Coptis were also reported to have similar liver protection effects in acute hepatotoxicity. The liver protection effects between Rheum, Scutellaria and Coptis chinensis are due to the same bioactive compounds, baicalein and other flavonoids \[[@b24-ijms-15-08037]--[@b28-ijms-15-08037]\]. Both SHSST and silymarin are potential liver protection drugs, but both present poor water solubility and poor bioavailability \[[@b17-ijms-15-08037]\]. A formulation approach is necessary to increase the solubility of these liver protection drugs. β-CD modification can increase the solubility and spectral properties of guest molecules, especially hydrophobic drugs, without changing their intrinsic property to permeate the cell membranes. Thus, SHSST was modified into SHSST-β-CD-complex (SHSSTc) and evaluated for its therapeutic effects in a CCM animal model. 2.. Results and Discussion ========================== The heart phenomena changes in each group were measured and presented in [Table 1](#t1-ijms-15-08037){ref-type="table"}. The average heart weight in CCl~4~-induced CCM groups was higher than the control. The average heart weights in silymarin, baicalein, SHSST, and SHSSTc low and high dose treatments were reduced and similar to the control group average heart weight. The ratio between left ventricular weight (LVW) and tibia length (TL) is a particularly accurate indicator for cardiomegaly. The average LVW/TL of CCl~4~-induced CCM groups are higher, and silymarin and baicalein can slightly reduce cardiac hypertrophy. Interestingly, the SHSST, SHSSTc low and high dose treatments can significantly reverse cardiac hypertrophy in CCl~4~-induced CCM rat hearts. In CCl~4~-induced CCM heart proteins analysis, the high level expression of brain natriuretic peptide (BNP) presents a stress load on the beating heart. ([Figure 1](#f1-ijms-15-08037){ref-type="fig"}) After 4 weeks of silymarin, baicalein and SHSST treatments, BNP levels were slightly reduced and lower than that in the CCM only group. Low and high dose SHSSTc treatments efficiently inhibited the BNP secretion in a dose dependent manner. The heart biopsy showed a large area of collagen accumulation in the CCl~4~-induced CCM rat heart. The hematoxylin and eosin (H&E) staining assay also showed cardiomyocytes presenting a disordered arrangement with more inter-space between the CCM heart cells ([Figure 2](#f2-ijms-15-08037){ref-type="fig"}). After 4 weeks of silymarin, baicalein and SHSST treatments, the cell arrangement became neat and close. Low and high dose SHSSTc treatments caused the cell arrangement to present just like the control and the collagen accumulation areas disappeared. The protein analysis results show that CCl~4~-induced CCM efficiently increased the TGF-β, phoshphorylated mothers against decapentaplegic homolog 3 (p-Smad3) and CTGF protein levels. ([Figure 3](#f3-ijms-15-08037){ref-type="fig"}) The silymarin, baicalein and SHSST treatments can only slightly reduce the TGF-β signaling pathway. However, low and high SHSSTc dose treatments significantly reduced TGF-β pro-form (monomer) and mature-form (dimer) expression in a dose dependent manner. Cirrhosis usually causes a hyper-dynamic state and this syndrome leads to increased cardiac output and decreased systemic vascular resistance \[[@b29-ijms-15-08037]\]. In some CCM cases the cardiac dysfunction in cirrhosis is not associated with the liver disease severity \[[@b30-ijms-15-08037]\]. On the contrary, one in five CCM patients suffers death after liver transplantation \[[@b31-ijms-15-08037]\]. The death risk might be caused by the maladaptive cardiac re-circulation system with a new liver. The traceable biomarker is BNP, which is directly secreted from the cardiac response to output stress \[[@b32-ijms-15-08037],[@b33-ijms-15-08037]\]. BNP expressions were indeed increased in the CCl~4~-induced CCM animal model in this research. ([Figure 1](#f1-ijms-15-08037){ref-type="fig"}) Cardiac hypertrophy was also induced in this animal model ([Table 1](#t1-ijms-15-08037){ref-type="table"}). In the results of this study cardiac fibrosis was serious in the CCM rat heart. ([Figure 2](#f2-ijms-15-08037){ref-type="fig"}) The protein analysis results also show a high mature-form of TGF-β expression in CCM. ([Figure 3](#f3-ijms-15-08037){ref-type="fig"}) Such results suggest the high expression might cause cardiac fibrosis or remodeling \[[@b11-ijms-15-08037],[@b12-ijms-15-08037]\]. Thus, in addition to support therapy for OLT patient treatment for TGF-β expression reduction might work and needs further clinical research to confirm. It is interesting that BNP expression levels after low and high dose SHSSTc treatments are comparable in the heart, while the fibrosis parameters (TGF-β pathway activation) are different (less activated in the high dose group). This might suggest that high dose SHSSTc treatment could provide a stronger effect on the TGF-β pathway. Another possible mechanism might exist in BNP expression reduction in CCM rat hearts after both low and high dose SHSSTc treatments. Silymarin and baicalein (one of the bioactive compounds in SHSST) were reported to reduce TGF-β expression in CCl~4~-induced acute liver injury in a rat model \[[@b34-ijms-15-08037],[@b35-ijms-15-08037]\]. This study also found a similar therapeutic effect can be used in CCM. SHSST treatment in CCM rats also inhibited TGF-β expression and its downstream proteins p-Smad3 and CTGF, similar to silymarin and baicalein. Although, SHSST and silymarin are effective, they are still limited by their hydrophobic characteristics. Low drug water solubility often presents poor bioavailability, preventing cirrhosis patients from accepting the normal use dose. This study used β-CD modified SHSSTc, which requires only minimal dose treatment to suppress TGF-β expression in the CCM rat heart. SHSST was reported to have an anti-atherogenic effect (0.2 mg/mL) in human aortic smooth muscle cells and suggested to present cytokine production inhibition effects \[[@b36-ijms-15-08037]\]. SHSST and SHSSTc were used for the first time and their cardiac protect effects against CCM evaluated in this research. The experimental results from this research show that low dose (30 mg/kg/day) SHSSTc can provide a protective effect in the CCM animal model through reducing TGF-β. SHSSTc does not just inhibit TGF-β dominated fibrosis in CCM rat hearts, but also slows CCM rat cardiac hypertrophy in a dose dependent manner. Thus, TGF-β might play an important role in heart remodeling in CCl~4~-induced CCM rat hearts. Silymarin, SHSST and baicalein treatments can reduce TGF-β expression and the downstream pathway, but SHSSTc treatment leads to a significantly improved result. 3.. Experimental Section ======================== 3.1.. Preparation of SHSST-β-CD (San Huang Shel Shin Tang-β-cyclodextrin) Complex and Drugs ------------------------------------------------------------------------------------------- SHSST-β-CD complex was prepared by co-precipitation. β-CD (70.0 g) was dissolved in distilled water (85 mL) at 70 °C in a water bath for 1 h. SHSST (10.0 g) in ethanol (15 mL) was slowly added to the β-CD solution with continuous agitation for 6 h. Forty milliliter of ethanol was then added drop wise to regulate the hydrophobic solute solubility in β-CD solution. Afterwards the solution was refrigerated overnight at 4 °C. The precipitated SHSSTc (SHSST:β-CD = 1:9 in weight) was recovered by filtration and washed with ethanol to remove unencapsulated SHSST. This residue was dried in a vacuum oven at −20 °C for 48 h. The final powder was stored at 4 °C until used. The silymarin and baicalein were purchased from Sigma (St. Louis, MO, USA). The silymarin, baicalein, SHSST and SHSSTc stock solutions for treatments were prepared by dissolving in distilled deionized water at 100 mg/mL each. CCl~4~ was dissolved in olive oil at the concentration of 4% v/v. 3.2.. Animal Model ------------------ The animal use protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of China Medical University (No. 100-3-B, date: 1 September 2010). There were 42 Sprague-Dawley (SD) rats purchased from BioLASCO Taiwan Co., Ltd. (Taipei, Taiwan) and divided into 7 groups (n = 6 each). CCl~4~ intraperitoneal (IP) injection to rats for cirrhosis induction was applied twice with 0.2 mL/kg at the first and fourteenth days. After 4 weeks from the final CCl~4~ IP injection treatments, all CCl~4~-induced CCM rat hearts were confirmed by echocardiography with the ejection fraction (EF) \<50% and fractional shortening (FS) \<85% parameters (control group EF \> 50% and FS \> 85%). After CCM induction, 4 more weeks of drug treatments were applied through gavage assay to each rat group. Groups were designated control, CCl~4~-induced CCM, silymarin (100 mg/kg/day) oral treatment, baicalein (30 mg/kg/day) oral treatment, SHSST (30 mg/kg/day) oral treatment, SHSSTc (30 mg/kg/day) oral treatment, SHSSTc (300 mg/kg/day) oral treatment. 3.3.. Cardiac Echocardiography ------------------------------ M-mode echocardiographic examination was performed using a 6--15 MHz linear transducer (15--6 L) via a parasternal long axis approach. Left ventricular (LV) M-mode measurements at the papillary muscles level included left ventricular internal end-diastolic dimensions (LVIDd), left ventricular internal end-systolic dimensions (LVIDs). Fractional shortening (FS%) was calculated according to the following equation: FS% = \[(LVIDd − LVIDs)/LVIDd\] × 100%. Ejection fraction (EF) is defined as the ratio between the volume of blood pumped out of the LV and total volume of blood in LV. 3.4.. Hemotoxylin and Eosin Staining ------------------------------------ The rat hearts from each group were soaked in 10% formalin, dehydrated through graded alcohols and embedded in paraffin wax. The 0.2 μm-thick paraffin sections were then cut into slices from these paraffin-embedded tissue blocks. The tissue sections were deparaffinized by immersing in xylene and rehydrated. All slices were dyed with hematoxylin and eosin (H&E) and then rinsed with water. Each slide was dehydrated through graded alcohols. Heart sections were finally soaked in xylene twice. Photomicrographs were obtained using Zeiss Axiophot microscopes (Taiwan Instrument Co., Taipei, Taiwan). 3.5.. Masson's Trichrome Staining --------------------------------- The rat hearts from each group were soaked in 10% formalin, dehydrated through graded alcohols and embedded in paraffin wax. The 0.2 μm-thick paraffin sections were then cut into slices from these paraffin-embedded tissue blocks. The tissue sections were deparaffinized by immersing in xylene and rehydrated. Samples were then stained with Masson's trichrome staining to investigate liver histo-logical and fibrotic changes. Photomicrographs were obtained using Zeiss Axiophot microscopes. 3.6.. Tissue Protein Extraction ------------------------------- Heart tissue extracts from 6 rats in each group were obtained by homogenizing in a lysis buffer (0.05 M Tris--HCl, pH 7.4, 0.15 M NaCl, 0.25% deoxycholic acid, 1% nonyl phenoxypolyethoxylethanol, 1 mM EDTA) at a ratio of 100 mg tissue/1 mL buffer. The homogenates were placed on ice and then centrifuged at 13,000 rpm for 40 min. The supernatants were collected and stored at −80 °C for further experiments. 3.7.. Western Blot Assay ------------------------ Heart tissue protein concentration extracts were determined using the Lowry protein assay. Protein samples were separated in a 12% SDS polyacrylamide gel electrophoresis (SDS-PAGE) with a constant voltage of 75 V for 120 min. Proteins were then transferred to Hybond-C membranes (GE healthcare UK Ltd., Little Chalfont, Buckinghamshire, UK) using 50 V for 3 h. Polyvinylidene difluoride (PVDF) membranes were incubated in 3% bovine serum albumin (BSA) in tricine buffer solution. Primary antibodies including brain natriuretic peptide (BNP, SC-18818, Santa Cruz Biotechnology, Dallas, TX, USA), TGF-β (SC-31609, Santa Cruz Biotechnology), Smad-3 (SC-8332, Santa Cruz Biotechnology), α-tubulin (SC-5286, Santa Cruz Biotechnology), CTGF (SC-14939, Santa Cruz Biotechnology), were added into the membranes for recognizing the fitted proteins. Horseradish peroxidase-labeled antibodies were finally used and pictures were then taken with Fujifilm LAS-4000 (GE healthcare UK Ltd.). 3.8.. Statistical Analysis -------------------------- The results shown are the means ± SD of three independent experiments. Statistical analysis was performed using one-way analysis of variants. The Student's t test was applied for paired samples. 4.. Conclusions =============== The experimental evidence reported here suggests that SHSSTc treatment can improve the present clinical treatment for non-cirrhotic heart failure. Present clinical treatment includes bed rest, oxygen and diuretics. TGF-β expression elimination in CCM through SHSSTc treatment can improve heart function in CCM and may also improve the survival rate for OLT patients after clinical liver transplantation surgery. This study is supported in part by Taiwan Ministry of Health and Welfare Clinical Trial and Research Center of Excellence (DOH102-TD-B-111-004). Author Contributions The authors' contributions were as follows: F.-J.T, Y.T and C.-Y.H. designed the experiments; C.-H.Y., W.-J.T., D.-T.J. and C.-H.D. acquired and analyzed the results; C.-H.Y., W.-J.T., D.-T.J., Y.-L.Y. and C.-Y.H. interpreted the results; L.-C.C., C.-H.T., F.-J.T., Y.T. and C.-Y.H. prepared and edited the manuscript. All authors read and approved the final manuscript. The authors declare no conflict of interest. ![Phoshphorylated GATA binding protein 4 (p-GATA4) and brain natriuretic peptide (BNP) expressions in cirrhotic cardiomyopathy (CCM) hearts. (A) The p-GATA4 and BNP expressions (n = 6 in each group) were increased in CCM and reduced by silymarin, baicalein, SHSST (San Huang Shel Shin Tang), and SHSSTc low dose and high dose treatments; (B) The normalized protein expression folds of p-GATA4 with GATA4; (C) The normalized protein expression folds of BNP with α-tubulin (the scale bars is also presented as the normalized folds with α-tubulin); (D) Echocardiography analysis images (n = 6 in each group) of the heart function is compared by the left ventricular systolic and diastolic distance (cm). \\ p \< 0.01, \\\* p \< 0.001 compared with control group; \# p \< 0.05, \#\# p \< 0.01, \#\#\# p \< 0.001 compared with CCM group.](ijms-15-08037f1){#f1-ijms-15-08037} ![Morphology assessments by hematoxylin and eosin (H&E) stain, Massion's trichrome stain (MS) assay in cirrhotic cardiomyopathy (CCM) rat hearts. In H&E stain slides, cell nuclei are stained with blue color, other intracellular or extracellular protein are stained with pink color. The heart fibrosis can be assessed using the collagen accumulation (indicated by blue color). Normal cells are indicated by pink color in the MS assay. All heart sections were obtained from the ventricular septal of each rat.](ijms-15-08037f2){#f2-ijms-15-08037} ###### The transforming growth factor (TGF)-β pathway protein expression level analysis. (A) TGF-β/phoshphorylated mothers against decapentaplegic homolog 3 (Smad-3)/connective tissue growth factor (CTGF) expression levels were increased in CCM and reduced by silymarin, baicalein, SHSST, and SHSSTc low and high dose treatments; (B) The normalized TGF-β protein expression folds with α-tubulin; (C) The normalized protein expression folds of p-Erk with Erk; (D) The normalized protein expression folds of p-Smad3 with Smad3; (E) The normalized protein expression folds of CTGF with α-tubulin. n = 6; \* p \< 0.05, \\\* p \< 0.001 compared with control group; \#\#\# p \< 0.001 compared with the CCM group. ![](ijms-15-08037f3a) ![](ijms-15-08037f3b) ###### Cardio vascular structure physiological characteristics assessment. Groups Control CCM CCM silymarin CCM Baicalein CCM SHSST CCM SHSSTc (low dose) CCM SHSSTc (high dose) ---------------- -------------- ----------------------------------------------------------------- ----------------------------------------------------------------- ----------------------------------------------------------------- ----------------------------------------------------------------- ----------------------------------------------------------------- ----------------------------------------------------------------- BW (g) 323.0 ± 8.9 352.3 ± 14.0 [\](#tfn2-ijms-15-08037){ref-type="table-fn"} 343.7 ± 6.1 [\](#tfn2-ijms-15-08037){ref-type="table-fn"} 325.7 ± 11.6 366.0 ± 22.3 [\](#tfn2-ijms-15-08037){ref-type="table-fn"} 349.7 ± 25.8 316.7 ± 9.1 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} TL (cm) 3.80 ± 0.15 4.10 ± 0.00 3.97 ± 0.06 4.10 ± 0.00 4.10 ± 0.10 4.03 ± 0.06 4.00 ± 0.10 HW (mg) 936.7 ± 53.3 1146.0 ± 21.9 [\\](#tfn3-ijms-15-08037){ref-type="table-fn"} 927.3 ± 49.8 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 921.7 ± 17.8 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 940.3 ± 25.1 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 944.0 ± 20.1 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 965.7 ± 11.2 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} LVW (mg) 681.7 ± 15.0 922.7 ± 11.9 [\\\](#tfn4-ijms-15-08037){ref-type="table-fn"} 711.3 ± 62.5 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 720.0 ± 62.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 683.7 ± 52.5 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 687.3 ± 54.7 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 683.0 ± 22.7 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} HW/BW (mg/g) 2.90 ± 0.10 3.26 ± 0.15 [\](#tfn2-ijms-15-08037){ref-type="table-fn"} 2.70 ± 0.19 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 2.83 ± 0.13 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 2.85 ± 0.22 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 2.71 ± 0.25 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 3.15 ± 0.12 HW/TL (mg/cm) 238.6 ± 20.5 279.5 ± 5.3 [\](#tfn2-ijms-15-08037){ref-type="table-fn"} 233.9 ± 16.0 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 224.8 ± 4.3 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 22.9 ± 11.7 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 234.1 ± 8.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 241.6 ± 8.3 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} LVW/BW (mg/g) 2.1 ± 0.1 2.6 ± 0.1 [\\](#tfn3-ijms-15-08037){ref-type="table-fn"} 2.1 ± 0.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 2.2 ± 0.3 1.9 ± 0.3 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 2.0 ± 0.3 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 2.2 ± 0.0 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} LVW/HW (mg/mg) 0.73 ± 0.03 0.81 ± 0.01 [\](#tfn2-ijms-15-08037){ref-type="table-fn"} 0.77 ± 0.03 0.78 ± 0.05 0.73 ± 0.04 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 0.73 ± 0.04 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 0.71 ± 0.03 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} LVW/TL (mg/cm) 173.5 ± 8.1 225.0 ± 2.9 [\\\](#tfn4-ijms-15-08037){ref-type="table-fn"} 179.5 ± 15.1 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 175.6 ± 15.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 167.0 ± 16.9 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 170.6 ± 15.8 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 170.7 ± 2.4 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} EF (%) 80 ± 2.6 68.5 ± 1.0 [\\\](#tfn4-ijms-15-08037){ref-type="table-fn"} 72.8 ± 3.3 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 72.3 ± 1.5 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 74 ± 1.0 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 76.7 ± 1.5 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 77 ± 0.3 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} FS (%) 43.7 ± 2.1 33.7 ± 1.4 [\\\](#tfn4-ijms-15-08037){ref-type="table-fn"} 38.8 ± 6.1 36.5 ± 1.0 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 38.3 ± 0.6 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 40.7 ± 1.5 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 41.0 ± 0.7 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} LVIDd (mm) 8.5 ± 0.1 8.0 ± 0.2 [\\\](#tfn4-ijms-15-08037){ref-type="table-fn"} 8.2 ± 0.4 8.1 ± 0.3 8.1 ± 0.3 8.3 ± 0.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 8.4 ± 0.1 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} BW, body weight; TL, tibia length; HW, whole heart weight; LVW, left vestibular weight; EF, ejection fraction; FS, fractional shortening; LVIDd, left ventricular interior dimention; p* \< 0.05, p \< 0.01, p \< 0.001 compared with control group; p \< 0.05, p \< 0.01, p \< 0.001 compared with cirrhotic cardiomyopathy (CCM) group.	{ "pile_set_name": "PubMed Central" }
1. Introduction {#sec1-jcm-09-01661} =============== Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection \[[@B1-jcm-09-01661]\] that may culminate in organ failure and death. Despite continuous efforts to understand and improve outcomes of sepsis, it remains a commonly fatal disease \[[@B2-jcm-09-01661],[@B3-jcm-09-01661],[@B4-jcm-09-01661]\]. In addition, the current definition of sepsis identifies a heterogeneous population of individuals with diverse patterns of immune response, organ dysfunction, and clinical outcomes \[[@B1-jcm-09-01661]\]. Therefore, early diagnosis, precise stratification of severity, and accurate outcome prediction are critical in managing patients with sepsis. Various biomarkers, including C-reactive protein and procalcitonin, have been investigated to date \[[@B5-jcm-09-01661],[@B6-jcm-09-01661]\]; however, no marker has demonstrated sufficient discriminatory power \[[@B7-jcm-09-01661]\]. Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) is a cytokine and member of the TNF superfamily. TRAIL initiates apoptosis of transformed cells or tumor cells by binding to death receptors (DR) 4 or 5 \[[@B8-jcm-09-01661]\] and functions as an immune response regulator in sepsis \[[@B8-jcm-09-01661],[@B9-jcm-09-01661]\]. More recently, however, it was discovered that TRAIL also functions as a trigger for necroptosis, a specific form of programmed cell necrosis characterized by its caspase-independent activation and release of damage-associated molecular patterns (DAMPs) which convey highly proinflammatory properties \[[@B10-jcm-09-01661],[@B11-jcm-09-01661]\]. In a recent study, plasma TRAIL was related to poor outcomes in patients with sepsis \[[@B12-jcm-09-01661]\]. A subsequent multicenter study suggested a relationship between the plasma level of TRAIL with necroptosis in sepsis; however, its association with mortality demonstrated a contradictory result \[[@B13-jcm-09-01661]\]. Therefore, further studies are necessary to clarify the relationship between plasma TRAIL and predicting outcomes of patients with sepsis. In this study, we analyzed the association between the plasma level of TRAIL and severity and outcomes of sepsis. We also compared the plasma level of TRAIL to that of receptor-interacting protein kinase-3 (RIPK3), a well-known necroptosis mediator \[[@B14-jcm-09-01661]\], to identify the correlation between TRAIL and necroptosis. 2. Materials and Methods {#sec2-jcm-09-01661} ======================== 2.1. Study Design and Registry {#sec2dot1-jcm-09-01661} ------------------------------ This was a prospective observational study of the Samsung Medical Center Registry of Critical Illness (SMC RoCI), which is an on-going single-center prospective registry of the Samsung Medical Center (1989-bed, university affiliated, tertiary referral hospital in Seoul, South Korea) initiated in April 2014 for the purpose of establishing a human sample repository and developing new biological markers for critical illness \[[@B15-jcm-09-01661]\]. The study was approved by the institutional review board of Samsung Medical Center. Written informed consent was obtained from patients or their legally authorized representative prior to enrollment. 2.2. Study Patients {#sec2dot2-jcm-09-01661} ------------------- Critically ill adult (≥19 years old) patients admitted to the medical intensive care unit (ICU) of Samsung Medical Center were considered eligible for inclusion in the registry. Exclusion criteria were as follows: (1) cognitive impairment, (2) inability to provide informed consent, (3) ICU admission for a simple procedure or postsurgical care, (4) transfer from other hospitals, (5) end-of-life decision or admission to facilitate comfort care, (6) hemoglobin \< 8 g/dL upon admission or persistent bleeding, and (7) discharge within 24 h of admission to ICU. Screening and enrollment were completed within 24 h of ICU admission. Patients registered between April 2014 and December 2016 were included in the analysis. Some clinical data and RIPK3 levels from patients enrolled until August 2016 were reported in a previous study \[[@B15-jcm-09-01661]\]. 2.3. Data Collection {#sec2dot3-jcm-09-01661} -------------------- A trained study coordinator used hospital records for each patient to prepare a standardized case report form. Clinical data consisting of patient demographics, reason for ICU admission, severity of illness scoring, and laboratory data were obtained at the time of enrollment. Illness severity was assessed by the Acute Physiology and Chronic Health Evaluation II (APACH II) \[[@B16-jcm-09-01661]\], Simplified Acute Physiology Score 3 (SAPS 3) \[[@B17-jcm-09-01661]\], and Sequential Organ Failure Assessment (SOFA) scores \[[@B18-jcm-09-01661]\]. The Revised Trauma Score was used to determine the severity of trauma patients \[[@B19-jcm-09-01661]\]. The primary outcome was 28-day mortality. Secondary outcomes were in-hospital and 90-day mortality. Sepsis was defined according to the third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) \[[@B1-jcm-09-01661]\]. Since enrollment for the registry began in April 2014, patients enrolled before release of the new definition were reclassified. Patients in the registry who did not meet the definition of either sepsis or septic shock were defined as controls. 2.4. Measurement of Plasma TRAIL and RIPK3 {#sec2dot4-jcm-09-01661} ------------------------------------------ Along with clinical data, 19 mL of whole blood was drawn from each patient within 48 h of study enrollment. When possible, additional blood samples were collected at day 3 (15 mL) and day 7 (15 mL). Blood samples were centrifuged within 4 h of collection. Plasma was separated and stored at −80 °C until further analysis. Plasma TRAIL level was measured from stored aliquots using commercially available TRAIL Human ELISA kits according to the manufacturer's recommendations (R & D systems, Minneapolis, MN, USA). Plasma RIPK3 level was measured using a commercially available ELISA kit as per the manufacturer's recommendations (CUSABIO, Houston, TX, USA) \[[@B15-jcm-09-01661]\]. 2.5. Statistical Analysis {#sec2dot5-jcm-09-01661} ------------------------- Data are presented as numbers (percentages) for categorical variables, and as the median and interquartile range (IQR, 25th--75th percentiles) for continuous variables. Categorical variables were compared using the Chi-square test or Fisher's exact test, while continuous variables were compared using the Mann--Whitney U test. Differences in plasma TRAIL level across control, sepsis, and septic shock groups were assessed with the Kruskal--Wallis test. Baseline characteristics, clinical status at ICU admission, illness severity, and mortality were compared between patients with low and high TRAIL level divided by the median level of plasma TRAIL. To evaluate the association between plasma TRAIL level and patient outcomes, a Kaplan--Meier curve was used to determine the 90-day survival curves according to plasma TRAIL. These were then compared using the log-rank test. Linear regression was applied to assess the association between plasma levels of TRAIL and RIPK3. The Friedman test was performed to determine differences among serial levels of TRAIL. The Wilcoxon signed rank test with Bonferroni correction was used to compare plasma TRAIL levels between days 0 and 3 and between days 3 and 7. All tests were two-sided, and a p value \< 0.05 was considered significant. Data were analyzed using IBM SPSS Statistics 20.0 (IBM, Chicago, IL, USA). 3. Results {#sec3-jcm-09-01661} ========== Over the study period, 1419 patients were admitted to the ICU. After excluding 1224 patients who met the exclusion criteria and one patient who missed screening, 194 patients were enrolled in the registry. Since four patients withdrew their consent, 190 patients were finally included in the analysis ([Figure 1](#jcm-09-01661-f001){ref-type="fig"}). The baseline characteristics of 190 patients at the time of ICU admission are summarized in [Table 1](#jcm-09-01661-t001){ref-type="table"}. The median age was 64 (IQR, 54--73) years, and 121 (63.7%) patients were male. Of 62 (32.6%) patients with cancer, 32 patients had hematologic malignancy. The reason for ICU admission was sepsis in 59 (31.1%) patients and septic shock in 84 (44.2%) patients. For 47 patients of the control group, the most common cause of ICU admission was pulmonary edema in nine patients, followed by airway stenosis in six, interstitial lung disease in five, multiple trauma in five, and acute exacerbation of chronic obstructive pulmonary disease or asthma in four patients. The median revised trauma score for five patients with multiple trauma was 13.26 (IQR, 7.19--22.15). No patients in the control group were with infection. Ninety-five (50.0%) patients were on mechanical ventilator support, and 116 (61.1%) patients required vasopressor support upon ICU admission. The median plasma TRAIL level was 34.48 pg/mL in the study patients. While patients on mechanical ventilatory support were more common in the control group, patients requiring vasopressor support were more prevalent in the sepsis group. Furthermore, severity scores of SAPS 3 and SOFA, as well as the C-reactive protein level were higher in the sepsis group. However, no difference in 28-day ICU mortality or in-hospital mortality was observed. The plasma TRAIL level for 143 patients with either sepsis or septic shock was 31.55 pg/mL, which was lower compared to that of 52.00 pg/mL in the control group. There was a statistically significant trend of decreased plasma TRAIL level across control, sepsis, and septic shock groups (52.00 \[35.66--74.41\] pg/mL vs. 35.83 \[28.07--60.82\] pg/mL vs. 26.08 \[15.29--40.97\] pg/mL; p \< 0.001) ([Figure 2](#jcm-09-01661-f002){ref-type="fig"}). When correlation between TRAIL and RIPK3 was assessed, the plasma level of TRAIL was inversely related to the plasma level of RIPK3 in patients with sepsis and septic shock (r = −0.172, r^2^ = 0.02973, p = 0.039) ([Figure 3](#jcm-09-01661-f003){ref-type="fig"}). To evaluate the association between characteristics of patients with sepsis and plasma TRAIL level, 143 patients who were diagnosed with either sepsis or septic shock were divided into two groups---low and high plasma TRAIL level---according to the median TRAIL level ([Table 2](#jcm-09-01661-t002){ref-type="table"}). Septic shock and patients requiring vasopressor support were more prevalent in the group with low plasma TRAIL. Furthermore, lactic acid, the SAPS 3 score, APACHE II score, and SOFA score at initial ICU admission were significantly higher in patients with low plasma TRAIL. However, statistical differences in 28-day ICU mortality, in-hospital mortality, and 90-day mortality were not observed between the two groups. In addition, Kaplan--Meier survival estimation did not demonstrate a difference in 90-day survival between patients with high and low plasma TRAIL level (p = 0.419) ([Figure 4](#jcm-09-01661-f004){ref-type="fig"}). Finally, serial levels of plasma TRAIL were analyzed in 143 patients with sepsis. Blood samples at days 3 and 7 were available for TRAIL measurement in 59 and 32 patients, respectively. Plasma TRAIL levels increased over time, with median levels of 31.55 pg/mL (19.29--47.56) at day 0, 40.00 pg/mL (27.00--56.00) at day 3, and 40.50 pg/mL (25.00--53.75) at day 7 (p \< 0.001) ([Figure 5](#jcm-09-01661-f005){ref-type="fig"}A). Although the TRAIL level on day 3 was significantly higher compared to that of day 0 (p = 0.036), there was no difference in levels between days 3 and 7. Plasma TRAIL levels were also assessed according to 28-day ICU mortality ([Figure 5](#jcm-09-01661-f005){ref-type="fig"}B). While serial levels of TRAIL significantly increased over time in survivors (p = 0.005), TRAIL levels in non-survivors did not demonstrate any differences (p = 0.102). The TRAIL level in survivors was higher on day 3 compared to day 0 (p = 0.039); however, TRAIL levels of day 3 and day 7 did not differ. Conversely, there was no difference in levels of TRAIL between day 0 and day 3, or between day 3 and day 7 in non-survivors. Finally, when plasma TRAIL levels on days 3 and 7 were compared between survivors and non-survivors, there was a tendency for a higher TRAIL level only on day 3 in survivors compared to non-survivors \[42.00 (32.00--57.00) vs. 26.00 (17.50--53.50), p = 0.0502\]. Nonetheless, the day 7 levels demonstrated no difference between survivors and non-survivors \[40.00 (25.00--52.50) vs. 53.00 (23.00--75.00), p = 0.801\]. 4. Discussion {#sec4-jcm-09-01661} ============= The aim of this prospective observational study was to investigate the association between plasma TRAIL level and sepsis severity, as well as to determine the role of TRAIL in predicting mortality. In addition, we planned to analyze the correlation between plasma levels of TRAIL and RIPK3 to identify the potential relationship of TRAIL with necroptosis. Our study of 190 consecutively enrolled critically ill patients demonstrated that the plasma TRAIL level was inversely associated with sepsis severity and plasma level of RIPK3; however, it was not predictive of mortality. Until now, only two studies have investigated the significance of plasma TRAIL in sepsis. Tian et al. reported that the level of soluble TRAIL was lower in patients with septic shock (defined by an old definition) compared to healthy controls or patients with sepsis; soluble TRAIL was also associated with mortality \[[@B12-jcm-09-01661]\]. Very recently, a multicenter study demonstrated an inverse correlation between the level of circulating TRAIL and both organ dysfunction and RIPK3; \[[@B13-jcm-09-01661]\] however, the association of TRAIL with mortality was portrayed in only two of three cohorts. The inverse relationship between plasma TRAIL level and sepsis severity observed in our study concurs with the results of two previous studies, strengthening the growing evidence of the potential role of TRAIL in sepsis. The reasons for the inverse relationship between TRAIL and sepsis severity and the role of TRAIL in sepsis pathogenesis are not fully understood. Nonetheless, there are several suggested mechanisms to explain this phenomenon. One possible explanation is based on the anti-inflammatory effect of TRAIL. TRAIL is thought to facilitate and accelerate apoptosis of leukocytes, contributing to resolution of inflammation \[[@B20-jcm-09-01661]\]. Animal studies have shown that TRAIL is associated with improved survival in murine models of infection and autoimmune diseases \[[@B21-jcm-09-01661],[@B22-jcm-09-01661],[@B23-jcm-09-01661]\]. However, this protective anti-inflammatory effect remains controversial. In a study by Unsinger et al. on murine cecal ligation and puncture models, TRAIL-deficient mice were better able to control bacterial infection due to TRAIL-dependent immune suppression and immune unresponsiveness \[[@B24-jcm-09-01661],[@B25-jcm-09-01661]\]. In addition, plasma TRAIL level was measured at the early phase of sepsis in our studies, as well as previous studies. Since TRAIL-dependent apoptosis affects immune suppression and immune unresponsiveness, which plays a role in the relatively later course of sepsis \[[@B25-jcm-09-01661]\], changes in plasma TRAIL level may not be reflective of immune status. Considering those facts, the anti-inflammatory effect of TRAIL may not fully explain its association with the pathogenesis of sepsis. Another possible mechanism is infection clearance by the TRAIL-dependent pathway. In vitro studies have shown that TRAIL contributes to infection control by mediating elimination of infected cells or restriction of viral replication in influenza or encephalomyocarditis viral infection models \[[@B26-jcm-09-01661],[@B27-jcm-09-01661]\]. However, this observation is also conflicting. Cardoso Alves et al. demonstrated in their study that TRAIL depletion resulted in reduced NK-cell mediated antiviral CD8^+^ T cell killing, thereby leading to faster pathogen clearance in a mouse model of viral infection \[[@B28-jcm-09-01661]\]. Moreover, since pathogen clearance by T cell mediated cell death is crucial in viral infection but not in bacterial infection, it may not fully explain the results of our finding. Finally, the capability of TRAIL to initiate necroptosis, a programmed cell death mechanism, may be another possible theory \[[@B10-jcm-09-01661]\]. Although the precise function and effect of necroptosis in sepsis has yet to be elucidated, the detrimental contribution of necroptosis is thought to lie in cell death, leading to local tissue injury and excess release of pro-inflammatory DAMPs. DAMPs stimulate the innate immune response and trigger an inflammatory cascade, resulting in organ damage \[[@B15-jcm-09-01661],[@B29-jcm-09-01661],[@B30-jcm-09-01661],[@B31-jcm-09-01661]\]. The relationship between sepsis and necroptosis has been observed in murine models, as well as in critically ill patients. Recent studies examining the necroptosis regulators of RIPK1, RIPK3, and MLKL in patients with sepsis reported that their levels were associated with disease severity and/or mortality \[[@B32-jcm-09-01661],[@B33-jcm-09-01661]\]. In certain pathologic environments, such as acidic states, TRAIL triggers necroptosis via RIPK1 and RIPK3 to hinder caspase-dependent apoptosis \[[@B10-jcm-09-01661],[@B34-jcm-09-01661]\]. Our study, as well as the study by Schenck et al. \[[@B13-jcm-09-01661]\], illustrated a significant correlation between the plasma levels of TRAIL and RIPK3, proposing the possibility that TRAIL may have an impact on sepsis through the necroptosis pathway. Nonetheless, it should be noted that this correlation between the TRAIL and RIPK3 proteins may not provide direct evidence that TRAIL reflects the process of necroptosis of sepsis. Although we have identified in our preliminary analysis that the plasma level of TRAIL also correlates with that of mitochondrial DNA (a well-known DAMP) in sepsis ([Figures S1--S3](#app1-jcm-09-01661){ref-type="app"}), supporting the hypothesis that TRAIL is associated with necroptosis, further research to investigate the pathogenesis of TRAIL in sepsis would be reasonable. Contrary to previous studies, the plasma TRAIL level was not associated with either 28-day ICU mortality or in-hospital mortality in our study. Interestingly, in the study by Schenck et al. \[[@B13-jcm-09-01661]\], the association between the plasma level of TRAIL with mortality was depicted only in two American cohorts of the United States, but not in a Korean cohort. The authors speculated that different baseline characteristics or processes of care might have contributed to this discrepancy. However, definite causes could not be determined. No relationship between TRAIL and mortality was observed in our cohort, in addition to another Korean cohort studied by Schenck et al. \[[@B13-jcm-09-01661]\]. This raises an ethnic difference as one of the many possibilities. Since the mortality of patients with sepsis may be attributable to various factors, further studies are warranted to confirm the relationship between mortality and TRAIL level and to further investigate the reason for the current discrepancy. One noteworthy finding of this study is that serial change in plasma TRAIL levels was available for analysis. To our knowledge, this is the first study to provide trends of the plasma TRAIL in sepsis. In particular, two discoveries require attention. First, TRAIL levels increased over time; however, the levels between days 3 and 7 did not differ. Second, when survivors and non-survivors were divided, the plasma TRAIL level on day 3 compared to day 0 was higher in survivors, but not in non-survivors. That there was no difference between levels on days 3 and 7 suggests that TRAIL, or the mechanism underlying the decrease in TRAIL, stabilizes or recovers within the initial days of sepsis. This suggests that TRAIL may be important in the relatively early phase of sepsis. Furthermore, because the TRAIL level on day 3 increased in survivors but not in non-survivors, perhaps it is not only the initial TRAIL level, but also the early serial changes of TRAIL that are crucial in understanding the patient's course and predicting mortality. However, since the day 3 and 7 samples were available only in some of the 143 patients with sepsis, we could not draw a definitive conclusion. As we have limited understanding of the role of TRAIL in sepsis, changes in TRAIL levels may be a consequence of sepsis or organ failure, rather than demonstration of a key process underlying sepsis. Comprehension of the pathogenesis and kinetics of TRAIL in sepsis is required to determine the significance of serial changes in TRAIL. To fully appreciate the results of our study, several limitations should be acknowledged. First, our study was conducted at a single referral center, which may limit the generalizability of our data. Second, follow-up data on TRAIL were not available in some patients due to death, recovery, and refusal to collect additional samples. Third, based on our study protocol, enrollment of a study patient and sampling of blood were done within 24 h of ICU admission and 48 h of enrollment, respectively. Therefore, patients with high severity and early mortality may not have been included in the study. Interpretation of results requires caution, since there may be a selection bias. Fourth, the correlation between TRAIL and RIPK3 and its association with severity of sepsis does not provide sufficient evidence that TRAIL functions as an initiator of necroptosis in sepsis or necroptosis is activated in sepsis. Further studies regarding gene expression and quantification are necessary to confirm this relationship. 5. Conclusions {#sec5-jcm-09-01661} ============== In conclusion, plasma TRAIL level was inversely associated with sepsis severity and plasma level of RIPK3. Plasma TRAIL levels increased following the course of recovery. However, they were not predictive of mortality. The following are available online at <https://www.mdpi.com/2077-0383/9/6/1661/s1>, Association of plasma level of TNF-related apoptosis inducing ligand with mitochondrial DNA in sepsis. Figure S1: Correlation between plasma level of TRAIL and mtDNA in patients with 49 patients with sepsis, Figure S2: Correlation between plasma level of RIPK3 and mtDNA in patients with 49 patients with sepsis, Figure S3: Plasma levels of mtDNA in sepsis and septic shock. ###### Click here for additional data file. H.Y. collected and analyzed the data and drafted this manuscript. J.Y.L. conducted experiments and analyzed the data. J.P. collected and analyzed the data. J.H.Y. and G.Y.S. analyzed the data and revised the manuscript. K.J. conceived and designed this study, analyzed the data, and wrote the final manuscript. All authors have read and agreed to the published version of the manuscript. This work was supported by a Samsung Medical Center grant (SMO1190131). The authors disclose that they do not have any conflict of interest. APACHE II: Acute physiology, age, chronic health evaluation II; DAMPs: Damage-associated molecular patterns; ICU: Intensive care unit; RIPK3: Receptor-interacting protein kinase-3; SAPS3: Simplified acute physiology score 3; SOFA: Sequential organ failure assessment; TRAIL, Tumor necrosis factor related apoptosis-inducing ligand. ![Study flow diagram.](jcm-09-01661-g001){#jcm-09-01661-f001} ![Plasma levels of Tumor necrosis factor TNF-related apoptosis-inducing ligand (TRAIL) in control, sepsis, and septic shock (p for trend \< 0.001).](jcm-09-01661-g002){#jcm-09-01661-f002} ![Correlation between plasma level of TRAIL and RIPK3 in patients with sepsis (n = 143). Slope: −8.713 (95% CI: −0.426--−17.000), r^2^: 0.02973, Pearson's r: −0.172 (p = 0.039).](jcm-09-01661-g003){#jcm-09-01661-f003} ![Kaplan--Meier survival analysis comparing patients with sepsis or septic shock with high and low plasma level of TRAIL (n = 143). The 30-day and 90-day survival estimates are 76.6% and 60.5%, respectively for patients with high TRAIL (solid line), while it is 80.2% and 66.9% for patients with low TRAIL (dotted line) (p = 0.419, log-rank test).](jcm-09-01661-g004){#jcm-09-01661-f004} ![(A) Serial plasma levels of TRAIL on days 0, 3, and 7 of 143 patients with sepsis. (B) Comparison of serial plasma TRAIL levels in 28-day intensive care unit survivors (solid line) and non-survivors (dotted line). Upper and lower bars represent the 75th and 25th interquartile ranges, respectively.](jcm-09-01661-g005){#jcm-09-01661-f005} jcm-09-01661-t001_Table 1 ###### Baseline characteristics of study patients (n = 190). Characteristics Total (N = 190) Control (n = 47) Sepsis (n = 143) p Value ---------------------------------------- ------------------------- ------------------------- ------------------------- ----------- Age, years 64 (54--73) 60 (54--68) 67 (53--74) 0.086 Gender, male 121 (63.7) 25 (46.8) 96 (67.1) 0.085 Co-morbidities Solid tumor 62 (32.6) 11 (23.4) 51 (35.7) 0.120 Hematologic malignancy 32 (16.8) 7 (14.9) 25 (17.5) 0.681 Diabetes 55 (28.9 12 (25.5) 43 (30.1) 0.552 Chronic obstructive pulmonary disease 18 (9.5) 7 (14.9) 11 (7.7) 0.157 Chronic kidney disease 15 (7.9) 4 (8.5) 11 (7.7) 1.000 Myocardial infarction 11 (5.8) 4 (8.5) 7 (4.9) 0.470 Congestive heart failure 9 (4.7) 4 (8.5) 5 (3.5) 0.228 Cerebrovascular disease 6 (3.2) 0 6 (4.2) 0.339 Chronic liver disease 1 (0.5) 0 1 (0.7) 1.000 Charlson Comorbidity Index 2 (1--3) 2 (1--3) 2 (1--3) 0.554 Clinical status on ICU admission Need for MV 95 (50.0) 32 (68.1) 66 (46.2) 0.009 Need for vasopressor support 116 (61.1) 15 (31.9) 101 (70.6) \<0.001 Laboratory findings PaO~2~/FiO~2~ ratio 139.3 (125.2--298.1) 255.1 (125.4--362.1) 181.3 (124.7--280.5) 0.112 CRP, mg/dL 10.0 (3.0--19.0) 5.4 (0.8--9.1) 12.5 (4.1--22.5) \<0.001 Lactic acid, mg/dL 3.0 (2.0--4.0) 2.4 (1.2--4.3) 2.7 (1.8--4.0) 0.230 Severity of illness SAPS 3, points 52 (44--59) 48 (35--56) 53 (46--59) 0.008 APACHE II score 23 (18--28) 22 (15--27) 23 (19--29) 0.184 SOFA score, initial 8 (5--11) 5 (3--9) 6 (6--11) \<0.001 Plasma TRAIL, pg/mL 34.482 (23.446--56.946) 52.000 (35.657--74.410) 31.545 (19.289--47.564) \<0.001 Outcome 28-day ICU mortality 39 (20.5) 10 (21.3) 29 (20.3) 0.883 In-hospital mortality 48 (25.3) 12 (25.5) 36 (25.2) 0.961 Note: APACHE II, acute physiology and chronic health evaluation II; CRP, C-reactive protein; ICU, intensive care unit; MV, mechanical ventilation; SAPS 3, simplified acute physiology score 3; SOFA, sequential organ failure assessment; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand. jcm-09-01661-t002_Table 2 ###### Characteristics of patients with sepsis stratified according to the median level of plasma TRAIL (n = 143). Characteristics Low TRAIL (n = 72) High TRAIL (n = 71) p Value ---------------------------------- ---------------------- ---------------------- ----------- Age, years 67 (55--73) 62 (51--74) 0.418 Gender, male 40 (69.4) 46 (64.8) 0.553 Co-morbidities Solid tumor 10 (13.9) 16 (22.5) 0.448 Hematologic malignancy 14 (19.4) 11 (15.5) 0.534 Diabetes 23 (31.9) 20 (28.2) 0.623 Chronic kidney disease 11 (11.6) 7 (7.4) 0.322 Myocardial infarction 8 (11.1) 4 (5.6) 0.217 Congestive heart failure 3 (4.2) 3 (4.2) 1.000 Charlson Comorbidity Index 2 (1--3) 2 (1--3) 0.668 Septic shock 52 (72.2%) 32 (45.1%) 0.001 Clinical status on ICU admission Need for MV 36 (50.0) 30 (42.3) 0.353 Need for vasopressor support 61 (84.7) 40 (56.3) \<0.001 Laboratory findings PaO~2~/FiO~2~ ratio 191.7 (120.2--282.6) 180.6 (130.5--277.5) 0.436 CRP, mg/dL 13.7 (4.2--13.7) 12.0 (4.0--22.5) 0.305 Lactic acid, mg/dL 3.1 (2.1--4.5) 2.3 (1.7--3.3) \<0.001 Severity of illness SAPS 3, points 56 (50--68) 50 (42--56) \<0.001 APACHE II score 24 (19--30) 22 (19--27) 0.007 SOFA score, initial 9 (7--11) 7 (5--11) \<0.001 Outcome 28-day ICU mortality 12 (16.7) 17 (23.9) 0.279 In-hospital mortality 19 (26.4) 17 (23.9) 0.736 90-day mortality 23 (31.9) 26 (36.6) 0.556 Note: APACHE II, acute physiology and chronic health evaluation II; CRP, C-reactive protein; ICU, intensive care unit; MV, mechanical ventilation; SAPS 3, simplified acute physiology score 3; SOFA, sequential organ failure assessment; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand.	{ "pile_set_name": "PubMed Central" }
![](indmedgaz72376-0058){#sp1 .638} ![](indmedgaz72376-0059){#sp2 .639} ![](indmedgaz72376-0060){#sp3 .640} ![](indmedgaz72376-0061){#sp4 .641} ![](indmedgaz72376-0062){#sp5 .642} ![](indmedgaz72376-0063){#sp6 .643} ![](indmedgaz72376-0064){#sp7 .644} ![](indmedgaz72376-0065){#sp8 .645} ![](indmedgaz72376-0066){#sp9 .646}	{ "pile_set_name": "PubMed Central" }
INTRODUCTION ============ Prompt detection of stimuli whose novelty or familiarity is unexpected in the current environment (i.e., contextual novelty) provides an evolutionary advantage, enabling adaptive behavior and flexible learning of the new information (Kakade and Dayan, [@b24]). The ability to discriminate potentially salient familiar and novel information, in a context abundant with old and new stimuli, becomes crucial for ensuring effective contextual learning. Efficient discrimination may trigger an orienting response toward the new unexpected information and may therefore enhance memory formation (Tulving and Kroll, [@b53]). Indeed, the brain\'s dopaminergic system, including striatal and midbrain structures, has been shown to respond to reward anticipation and to prediction errors as well as having an effect on memory formation (Adcock et al., [@b1]; Wittmann et al., [@b56]). It has also been proposed (Lisman and Grace, [@b29]; O\'Carroll et al., [@b33]; Shohamy and Adcock, [@b47]) that interaction between the midbrain dopaminergic regions and the hippocampus is crucial for the construction of new memories, especially when novel information is involved (Bunzeck and Duzel, [@b3]; Bunzeck et al., [@b4]). Drawing on the measures of eye tracking, subsequent memory, and functional magnetic resonance imaging (fMRI), we further explored this proposal by manipulating the contextual novelty of novel and familiar visual stimuli in three separate experiments. The response of the dopaminergic midbrain and striatum to unexpected information may be associated with increased exploration of such a stimulus (Düzel et al., [@b9]), perhaps to establish its absolute novelty or familiarity status. Dopamine release, from the midbrain to the evaluation centers in the ventral striatum and from there to the hippocampus, may therefore be manifested behaviorally through an enhanced visual exploration of the presented information, leading to enhanced memory formation for this information. Eye movements have been found to be the markers of the increased visual exploration that leads to enhanced encoding into long-term memory (Kafkas and Montaldi, [@b19]), whereas pupil responses are argued to reflect resource allocation, cognitive processing (Granholm et al., [@b13]; Kahneman and Beatty, [@b23]; for a recent review, see Laeng et al., [@b28]), and more recently also encoding and retrieval processes (Kafkas and Montaldi, [@b19],[@b20]; Papesh et al., [@b36]). The current research explored eye movement and pupil response patterns as well as striatal/midbrain and medial temporal lobe (MTL) blood oxygenation level-dependent (BOLD) responses while unexpected novel and familiar stimuli were presented. Critically, in contrast with the previous studies that have focused on the encoding of rewarding information, the current experiments focused on retrieval when familiar and novel, expected and unexpected stimuli were (re-)encoded. The aim of this design was to further characterize the role of the dopaminergic system in memory at retrieval, for both familiar and novel stimuli, when these are either expected or unexpected, as determined by the experimental context. Retrieval provides a setting in which to vary the probability of old and new stimuli, so as to explore how memory is updated depending on both the novelty and the familiarity status of a stimulus, and the interaction of that status with contextual expectation. This investigation promises to provide a more comprehensive, but so far, lacking (Scimeca and Badre, [@b45]), understanding of the role of the dopaminergic system in memory. Specifically, the purpose of this research was to investigate the degree to which the detection of contextual novelty (for both novel and familiar stimuli) would trigger an increased response in the striatal and midbrain dopaminergic system and whether this would, in turn, modulate the eye behavior accompanying visual processing, and in so doing, enhance further encoding of contextual information. We hypothesized that if, indeed, striatal and midbrain activity accompanies the processing of contextual novelty this would drive exploratory behavior, reflected in eye movements and pupil responses. Furthermore, we hypothesized that functional connectivity of the striatal/midbrain structures with the MTL, in response to the processing of unexpected (compared to expected) information, would explain the enhanced learning. These hypotheses were explored in two separate, but matched, experiments drawing on fMRI BOLD responses and eye measures (pupil response and fixation patterns). Finally, subsequent memory for unexpected familiar and novel stimuli was explored in a third experiment using a surprise follow-up memory test. MATERIALS AND METHODS ===================== Participants ------------ In the fMRI study 17 (10 females) right-handed volunteers with a mean age of 22.6 years (SD = 4.2) gave informed consent and participated. Remuneration was given at a rate of £20 per session. Another 37 (27 females) undergraduate psychology students, with a mean age of 20.3 years (SD = 2.0) took part in the eye-tracking experiment in return for course credits. In both experiments, volunteers with normal or corrected (with contact lenses) to normal vision were recruited. No history of psychiatric or neurological disorder and no systematic use of psychotropic medicines or drugs were reported by participants in either experiment. Finally, 27 participants (24 females) with a mean age of 20 years (SD = 1.97) participated (receiving £5) in Experiment 3. The experiments were approved by the Research Ethics Committee of the University of Manchester (eye tracking and behavioral) and the National Research Ethics Service (fMRI study). Stimulus Materials ------------------ The pictures of manmade and natural objects, normalized to a mean gray level, were used in all three experiments. In the fMRI study, 318 of these stimuli were used (18 for practice), whereas the eye-tracking experiment used 220 stimuli (20 stimuli for practice), each subtending a visual angle of 18.69° horizontally and 14.03° vertically at presentation. In the behavioral subsequent memory study (Experiment 3), 280 stimuli (220 from the eye-tracking experiment plus 60 new foils) were presented. Owing to the sensitivity of pupillary responses to light intensity (Cheng et al., [@b5]), as well as to chromatic changes (Tsujimura et al., [@b52]), stimulus properties such as ambient light, brightness, contrast, and the color of the presented visual items were controlled (for further details regarding the development and standardization of these stimuli, see Kafkas and Montaldi, [@b19]). Experimental Design and Task ---------------------------- In the fMRI and eye-tracking experiments, participants completed two phases: an encoding phase and a retrieval phase, with the fMRI and eye-tracking data recorded during the retrieval phase. At encoding, a series of visual stimuli were encoded using a perceptual-matching-to-sample task with each stimulus appearing for 4 s. In each trial, participants were presented with triplet images of the same object and were asked to decide which of the two bottom images was identical to the target image presented on top ([Fig. 1](#fig01){ref-type="fig"}; for this task, see Montaldi et al., [@b32]; Kafkas and Montaldi, [@b20],[@b21]). After a short filled interval (∼10 min), participants were presented with two recognition conditions, a novelty task (NT) emphasizing the detection of novel stimuli, and a familiarity task (FT) emphasizing the detection of familiar stimuli. Critically, 70% of the items presented in the two tasks were the "target" stimuli (i.e., novel in the NT condition and familiar in the FT condition), whereas 30% of the items in each task were foils (familiar stimuli in the NT condition and novel stimuli in the FT condition; [Fig.1](#fig01){ref-type="fig"}). These foil stimuli, which were presented with considerably lower probability, are classified here as unexpected items, compared to the target familiar and novel stimuli, which are classified as expected. For each target stimulus, participants provided a rating of how strongly familiar (FT) or novel (NT) they felt they were. In response to the unexpected stimuli (foils), on the other hand, participants were only required to report them as familiar or novel. The comparison of the BOLD response (Experiment 1) and the eye-tracking measures (Experiment 2) produced by unexpected versus expected novel and familiar stimuli is therefore the main focus of this study. ![Experimental design of the fMRI (Experiment 1) and eye-tracking (Experiment 2) studies.](hipo0025-1262-f1){#fig01} All stimuli were presented for a period of 3 s during which participants responded. In the fMRI experiment, 6 alternating blocks of familiarity (FT) and novelty (NT) conditions were presented, whereas in the eye-tracking experiment the two tasks were presented sequentially in an order that was counterbalanced across participants. In the fMRI study, 60 null events (providing an implicit baseline) were intermixed with the 300 visual stimuli, pseudorandomly allocated across the six blocks with the only restriction being that null events were never presented consecutively. These null events involve a fixation cross presented centrally and constitute therefore slightly extended presentations of the interstimulus fixation cross. This ensures that the null events are implicitly processed as they are not distinguished from standard interstimulus presentations by participants. A debriefing procedure after the end of the fMRI session indicated that no participant perceived the null events as unexpected in any way. The design of the behavioral follow-up study (Experiment 3) was identical to the eye-tracking experiment with the only difference being the addition of a surprise retrieval task at the end in which memory for all the stimuli presented in the two main recognition conditions (FT and NT) was tested. Due to the addition of the surprise recognition procedure, the number of stimuli presented in the two main recognition conditions (FT, NT) was reduced maintaining, however, the same probability (30%) for new and old foils in FT and NT, respectively. The completion of the familiarity and novelty tasks was followed by a 10-min distracter task, after which participants were presented with the surprise recognition memory test in which 120 stimuli from the FT and NT conditions were presented together with 60 new foils. Participants were instructed to provide familiarity (F) or recollection (R) responses to each stimulus if they believed they had seen it at some point in the experiment, or to report it as new, if they felt they were encountering the stimulus for the first time (remember/know procedure; for the instructions used see Migo et al., [@b31]). fMRI and Eye-Tracking Data Acquisition -------------------------------------- MR scanning was carried out on a 3T MRI scanner (Philips, Achieva) using a gradient echo-planar imaging (EPI) sequence to obtain T2\* images, employing the BOLD contrast. Forty slices positioned parallel to the AC--PC line and covering the whole brain (matrix size, 96 × 96; voxel size, 2.5 × 2.5 × 3.5 mm) were acquired for each volume (TR = 2.5 s; TE = 35 ms). In total, 726 volumes were recorded per participant across two sessions (363 per session). High-resolution T1 anatomical images (size, 256 × 256; 180 slices; size: 1 mm isotropic) were acquired for each participant before the two functional sessions. In the eye-tracking study, pupillary responses and eye movements were recorded using an ASL infrared eye-tracking system (Applied Science Laboratories, Bedford, MA; Model Eye-Trac 6000; sampling rate 60 Hz). A nine-point standard calibration procedure was carried out separately for each participant prior to data collection. For each trial, the eye-tracking data were recorded and analyzed from stimulus onset until the point of behavioral response. Blinks and other losses in the raw eye signal were identified by the eye-tracking software and were discarded from further analyses. Pupil recordings departing more than 3 SDs from the mean of each trial were discarded, as such abrupt pupil dilations or constrictions are considered noise in the pupil signal (Granholm et al., [@b13]; Beatty and Lucero-Wagoner, [@b2]). Trials containing \<60% of valid recordings were excluded (\<5% of total). The peak pupil size for each trial was calculated as the average of three pupil recordings preceding and three recordings following the maximum pupil value of the corresponding trial. The resulting peak pupil responses for each trial of the experiment were expressed as the deviation from the baseline pupil size, recorded during a period of 1,000 ms preceding each trial (Beatty and Lucero-Wagoner, [@b2]), consisting of a gray background (RGB level, 130). Finally, from the raw eye movement signal, a measure of spatial fixation (i.e., the number of fixations) was extracted. Data Analyses ------------- ### fMRI preprocessing The fMRI data were preprocessed using SPM8 (Statistical Parametric Mapping, Wellcome Department of Cognitive Neurology, University College London, London, United Kingdom; <http://www.fil.ion.ucl.ac.uk/spm/>). The EPI data were realigned to the first image and resliced using a six-parameter rigid body transformation and sinc interpolation in space. Slice-timing correction was also applied to the resliced images, which were then spatially normalized to the Montreal Neurological Institute (MNI) EPI template. Smoothing was performed using an isotropic (FWHM, 6 mm) Gaussian kernel. A canonical hemodynamic response function (Friston et al., [@b12]) was used to model the event-related responses for each participant. Correctly recognized expected and unexpected familiar and novel stimuli were modeled as separate conditions, whereas regressors of no interest (e.g., cue screens preceding each block, trials with no response, etc.) were also included in each individual model. The six movement parameters for each of the two sessions of the experiment were also modeled as regressors to capture residual movement-specific variance in the time series. ### fMRI data analyses Contrasts of the t-statistic for the four conditions of interest (expected novelty, unexpected novelty, expected familiarity, and unexpected familiarity) were produced for each participant in the first-level analysis, and carried forward to a second-level group analysis treating participants as a random (mixed) effect. A two-way ANOVA was conducted in SPM with the stimulus type (familiar or novel) and expectation (expected or unexpected) as the within-subjects factors. Two direct t contrasts (unexpected novelty \> expected novelty and unexpected familiarity \> expected familiarity) were used to further explore the functional data. Activations surviving a cluster-wise P \< 0.05 significance level (cluster FWE-corrected) are reported unless otherwise stated. MNI coordinates are reported for the activation data. ### Psychophysiological interaction analyses Effective connectivity was explored using psychophysiological interaction (PPI) analysis in SPM8 (Friston et al., [@b11]). PPI models were used to explore the hypothesis that the detection of unexpected familiar or novel stimuli would be associated with increased connectivity between the midbrain/striatal areas and the MTL. A PPI analysis explores the effect of activity in one brain region, on the activity in another region under the influence of a psychological contrast (i.e., unexpected vs. expected events). A significant PPI between two brain regions indicates that the slope of the regression characterizing their activity changes owing to the effect of the psychological context (Friston et al., [@b11]). Two areas, within the globus pallidus (GP) and the substantia nigra/ventral tegmental area (SN/VTA), showing selective responses to unexpected familiarity and novelty respectively (Results section) were chosen as the two functional seeds. Spheres of 6 mm radius within the GP (centered at x = 17, y = 1, z = −5) and SN/VTA (centered at x = 7, y = −20, z = −8) were created for each participant and the deconvolved BOLD activation data were extracted from these volumes of interests. Two separate PPI analyses were set up for the two seed areas, one for unexpected familiarity (contrast: unexpected familiar \> expected familiar stimuli; seed area: GP) and another for unexpected novelty (contrast: unexpected novel \> expected novel stimuli; seed area: SN/VTA). At the first-level analysis, three conditions were entered for each participant: (a) the BOLD signal time series of the seed region (physiological variable), (b) the contrast of interest (unexpected \> expected familiarity or novelty; psychological variable), and (c) the interaction term expressed as the multiplication of the psychological and physiological variables (PPI). A general linear model, with these three regressors, was estimated and t-contrasts of the interaction term were created for each participant and analyzed at the group level using a random-effects model (for further details, see Friston et al., [@b11]). The same statistical inference criteria as those applied to the main fMRI analyses were used for identifying significant activations. The MTL regions are reported as significant when the activation survived a FWE small-volume correction for the volume of the MTL or when eight or more contiguous voxels are active at P \< 0.001. This level was established as being equivalent to a corrected probability level of P \< 0.05 for the volume of the MTL using a Monte Carlo simulation (implemented in the AlphaSim tool in AFNI). ### Eye-tracking data analysis The fixation and pupil data were analyzed using a 2 × 2 repeated measures ANOVA with stimulus type (familiar, novel) and expectation (expected, unexpected) as the within-subject factors. Direct pairwise comparisons are also reported, where relevant, adopting a significance level of P \< 0.05 (Bonferroni corrected). ### Follow-up surprise recognition memory experiment analysis Data from the follow-up recognition memory experiment (Experiment 3) were analyzed to determine whether the unexpected novel and familiar stimuli were later remembered better than the expected novel and familiar stimuli. Using signal detection, d′ scores were calculated for familiarity and recollection responses provided at later test for the originally expected and unexpected novel and familiar stimuli. Two ANOVAs were run separately for the novel and familiar stimuli, with response in the follow-up task (familiar or recollected) and expectation status in the original task (expected or unexpected) as the within-subjects factors. Paired t-tests were also employed to further explore the significant interactions using a significance level of P \< 0.05 (Bonferroni corrected). RESULTS ======= Experiment 1: fMRI Results -------------------------- Behavioral data acquired during the fMRI study showed that the participants were able to accurately categorize novel and familiar stimuli in both tasks with well above chance accuracy (0.76 ± 0.08 and 0.77 ± 0.06). The analysis of the response times (RTs) for novel and familiar expected and unexpected stimuli showed a significant main effect of expectation (F~1,16~ = 5.50, P = 0.03). Expected stimuli (familiar~expected~: M = 1,749 ms, SD = 236 ms, and novel~expected~: M = 1,839, SD = 198 ms) had longer RTs than the unexpected ones (familiar~unexpected~: M = 1,679 ms, SD = 193 ms, and novel~unexpected~: M = 1,714 ms, SD = 220 ms). The fMRI analyses revealed striatal, midbrain, and neocortical activations for the main effect of expectation as well as for the expectation by stimulus-type interaction ([Fig. 2](#fig02){ref-type="fig"}). Four sets of neural response emerged. First, one region within the left caudate nucleus (caudate body, [Fig. 2a](#fig02){ref-type="fig"}) responded selectively to expected familiar and novel stimuli compared to unexpected stimuli. A second set included areas within the midbrain/brainstem ([Fig. 2c](#fig02){ref-type="fig"}) and the inferior temporal gyrus (BA 37; x = −53, y = −52, z = −19; 37 voxels, P \< 0.001 uncorrected), which responded to both familiar and novel unexpected stimuli relative to expected ones. A third set of regions, within the right GP ([Fig. 2b](#fig02){ref-type="fig"}), the left inferior parietal lobe (extending into the occipito-parietal junction; [Fig. 2d](#fig02){ref-type="fig"}), and the left precuneus ([Fig. 2e](#fig02){ref-type="fig"}) showed selective involvement in the processing of unexpected familiarity when compared to expected familiarity, but no modulation for novelty. Finally, the interaction effect revealed a set of regions in the right caudate (caudate head, [Fig. 2f](#fig02){ref-type="fig"}), the left middle frontal gyrus ([Fig. 2g](#fig02){ref-type="fig"}), the right inferior parietal lobe ([Fig. 2h](#fig02){ref-type="fig"}), and the right precuneus ([Fig. 2i](#fig02){ref-type="fig"}) that respond to unexpected familiarity (relative to expected familiarity) and to expected novelty relative to unexpected novelty. Critically, in these latter regions unexpected familiarity responses were significantly higher than unexpected novelty responses, whereas there were no significant differences between expected familiar and expected novel stimuli. This pattern of results suggests that the regions isolated in the interaction analysis have a predominant role in responding to unexpected familiarity. ![fMRI results from Experiment 1. A main effect of expectation status (expected/unexpected) was observed within (a) the left caudate body x = −21, y = 6, z = 20 (b) the right GP x = 15, y = −7, z = −1, (c) the left midbrain x = −3, y = −27, z = −26, (d) the left occipito-parietal junction x = −36, y = −70, z = 31; BA39 and BA19 and (e) the left precuneus x = −11, y = −75, z = 45; BA7. A significant interaction between expectation status (expected/unexpected) and stimulus type (familiar/novel) was observed in (f) the right caudate head x = 5, y = 11, z = 6, (g) the left middle frontal gyrus x = −31, y = 16, z = 34; BA8/9, (h) the right inferior parietal lobe x = 42, y = −72, z = 48; BA39 and (i) the right precuneus x = 12, y = -57, z = 41; BA7. Parameter estimates are plotted for each region separately. Error bars show one standard error of the mean. Asterisks denote statistically significant effects; \*P \< 0.05; \\P \< 0.01; \\\*P \< 0.001. \[Color figure can be viewed in the online issue, which is available at [wileyonlinelibrary.com](http://wileyonlinelibrary.com).\]](hipo0025-1262-f2){#fig02} To further constrain these effects, direct contrasts between unexpected and expected familiar and novel stimuli were run separately. The unexpected novelty versus expected novelty revealed significant activity in the right SN (x = 7, y = −20, z = −8, P \< 0.001 uncorrected, 21 voxels, Z = 4.17; [Fig. 3a](#fig03){ref-type="fig"}). The same contrast (unexpected vs. expected) for familiar stimuli revealed significant activity in the bilateral lentiform nucleus (including GP and putamen; [Fig. 4a](#fig04){ref-type="fig"}; Right: x = 17, y = 1, z = −5, cluster FWE-corrected; Left: x = −13, y = −2, z = −1, P \< 0.001 uncorrected, 13 voxels) extending into the thalamus (x = 10, y = −10, z = −1) and the caudate nucleus (x = 7, y = 6, z = −5). In the whole-brain analysis, consistent with the ANOVA effects, the bilateral angular gyri (BA 39; left: x = −38, y = −72, z = 38; right: x = 42, y = −70, z = 41), the bilateral middle occipital gyri (BA 19; left: x = −36, y = −67, z = 24; right: x = 37, y = −77, z = 27), and the bilateral precuneus (BA 7; left: x = −6, y = 72, z = 41; right: x = 12, y = −57, z = 34) also showed significant activity (P \< 0.05 cluster FWE-corrected) for the unexpected versus expected familiar stimuli. ![(a) Activity and parameter estimates within SN/VTA (x = 7, y = −20, z = −8) for unexpected novelty and (b) PPI connectivity with the hippocampus (x = −36, y = −32, z = −12 and x = −38, y = −12, z = −26). \[Color figure can be viewed in the online issue, which is available at [wileyonlinelibrary.com](http://wileyonlinelibrary.com).\]](hipo0025-1262-f3){#fig03} ![(a) Activity and parameter estimates within the bilateral ventral striatum/GP (x = 17, y = 1, z = −5 and x = −13, y = −2, z = −1) for unexpected familiarity and (b) PPI connectivity with the perirhinal cortex (x = -31, y = -5, z = -33) and the hippocampus (x = 35, y = −35, z = −5). Parameter estimates are plotted for the right GP. \[Color figure can be viewed in the online issue, which is available at [wileyonlinelibrary.com](http://wileyonlinelibrary.com).\]](hipo0025-1262-f4){#fig04} PPI: Effective Connectivity Analysis ------------------------------------ A PPI analysis with the right SN cluster as the seed area (centered at x = 7, y = −20, z = −8), for unexpected versus expected novelty, showed significantly increased connectivity with the left hippocampus (x = −36, y = −32, z = −12; P \< 0.001, 8 voxels and x = −38, y = −12, z = −26; P \< 0.001, 16 voxels; [Fig. 3b](#fig03){ref-type="fig"}). The connectivity analysis of the right ventral striatum/GP (centered at x = 17, y = 1, z = −5) showed increased connectivity with the right posterior hippocampus (x = 35, y = −35, z = −5; P \< 0.001, 8 voxels) and the left PRC (x = −31, y = −5, z = −33; P \< 0.001, 40 voxels; [Fig. 4b](#fig04){ref-type="fig"}) for unexpected versus expected familiarity. Both PPI analyses revealed increased connectivity between the two seed areas and the MTL structures, but no other cortical or subcortical effect was found. Experiment 2: Eye-Tracking Results ---------------------------------- To further investigate the behavior that accompanies the processing of unexpected familiar and novel stimuli, we analyzed the pupil responses and fixation patterns produced while participants processed these stimuli. In the eye-tracking experiment (Experiment 2), familiar and novel stimuli in the two tasks were categorized with a mean accuracy of 0.81 (SD = 0.07) and 0.82 (SD = 0.10). Consistent with the fMRI experiment, the detection of expected stimuli was accompanied by longer RTs than unexpected ones as indicated by the main effect of expectation (F~1,36~ = 28.87, P \< 0.001; familiar~expected~: M = 1,562 ms, SD = 229 ms; familiar~unexpected~: M = 1,434 ms, SD = 315 ms; novel~expected~: M = 1,591 ms, SD = 261 ms; novel~unexpected~: M = 1,394 ms, SD = 228 ms). The stimulus by expectation ANOVA of the peak pupil dilation produced a significant main effect of stimulus (F~1,36~ = 4.99, P = 0.03), showing that the detection of familiar stimuli produced more elevated pupil dilations than did the detection of novel stimuli. The main effect of expectation was not significant (F \< 1); however, a significant interaction between stimulus and expectation (F~1,36~ = 11.16, P = 0.002) showed that unexpected familiar stimuli produced larger pupil dilations than expected familiar stimuli (t~36~ = 2.21, P = 0.03). Similarly, unexpected novel produced larger pupil dilations than expected novel stimuli (t~36~ = −2.56, P = 0.015; [Fig. 5a](#fig05){ref-type="fig"}). The analysis of the number of fixations did not yield any significant main effect (stimulus: F~1,36~ = 1.12, P = 0.30; expectation: F \< 1), but produced a significant interaction between the two factors (F~1,36~ = 6.67, P = 0.014; [Fig. 5b](#fig05){ref-type="fig"}), revealing that unexpected novel stimuli were accompanied by more fixations than expected novel stimuli (t~36~ = 2.80, P = 0.008). The same pattern was also found for unexpected versus expected familiar stimuli, but this difference did not reach significance (t~36~ = 1.27, P = 0.21). ![Pupil dilation (a) and fixation patterns (b) for expected and unexpected familiar and novel stimuli in Experiment 2. Error bars show one standard error of the mean; \*P \< 0.05; \\P \< 0.01.](hipo0025-1262-f5){#fig05} Experiment 3: Subsequent Memory for Unexpected Stimuli ------------------------------------------------------ Finally, we investigated the extent to which the patterns of midbrain/striatal--MTL connectivity (from Experiment 1) and the eye-tracking patterns that accompanied unexpected stimuli (from Experiment 2) are associated with any memory advantages for unexpected stimuli versus expected ones. In a separate experiment (Experiment 3), participants completed the same experimental procedures as in the previous experiments. Critically, their memory for the originally expected and unexpected familiar and novel stimuli was tested in a surprise recognition memory task at the end of the session (Materials and Methods section). In this task, participants recognized old and new stimuli (from across the original FT and NT conditions) with a mean accuracy of 0.88 (SD = 0.05) and a false alarm (FA) rate of 0.22 (SD = 0.15), giving an overall memory performance (hit rate − FA rate) of 0.66 (SD = 0.19), which was significantly above chance (t~26~ = 18.31, P \< 0.001). Familiarity d′ scores and recollection d′ scores for originally expected and unexpected stimuli were entered into a 2 × 2 ANOVA separately for originally familiar and novel stimuli. Both ANOVAs showed a significant expectation by response interaction for novel (F~1,26~ = 21.32, P \< 0.001) and familiar stimuli (F~1,26~ = 12.32, P = 0.002). As summarized in Table[1](#tbl1){ref-type="table"}, processing unexpected familiar stimuli produced subsequently higher levels of recollection accuracy than processing expected familiar stimuli (t~26~ = −4.15, P \< 0.001), whereas there was no difference in the familiarity (d′) accuracy for unexpected versus expected familiar stimuli (t~26~ \< 1). Similarly, processing unexpected novel stimuli produced higher levels of recollection accuracy than expected novel stimuli (t~26~ = −4.03, P \< 0.001), whereas familiarity d′ scores were higher for expected than unexpected novel stimuli (t~26~ = 3.16, P = 0.004). ###### Subsequent Familiarity and Recollection Performance (d′) for Unexpected and Expected Novel and Familiar Stimuli Response Unexpected Expected ------------------ -------------- -------------------------------------------------- ------------- Novel stimuli Recollection 0.53 (0.08)[\\\](#tf1-1){ref-type="table-fn"} 0.11 (0.10) Familiarity 1.17 (0.08) 1.30 (0.08)[\\](#tf1-2){ref-type="table-fn"} Familiar stimuli Recollection 0.72 (0.11)[\\\](#tf1-1){ref-type="table-fn"} 0.29 (0.16) Familiarity 1.14 (0.12) 1.11 (0.13) Note: The categorization of stimuli as novel and familiar refers to their status in the main recognition task. Familiarity and recollection responses are obtained from the surprise subsequent memory test. Numbers in the parentheses are standard errors of the mean. P\<0.001. P\<0.01. DISCUSSION ========== The findings from this research expand our understanding of the nature of the behavioral conditions that might trigger increased connectivity between striatal/midbrain and MTL regions during the detection of novelty and the effect of this on later memory. In addition, the research illustrates the role of visual exploration (eye movements) in the facilitation of encoding during the detection of contextual novelty. The main findings can be summarized as follows. First, we confirmed that striatal and midbrain structures (along with some cortical areas, [Fig. 2](#fig02){ref-type="fig"}) responded to contextual novelty for both familiar and novel stimuli at retrieval. Critically, the detection of unexpected novel versus expected novel stimuli elicited increased activity in the SN/VTA, whereas the detection of unexpected familiar versus expected familiar stimuli was associated with the activity in the striatum (predominantly within the GP and putamen). In addition, these regions displayed increased effective connectivity during the detection of unexpected versus expected stimuli, with important memory formation regions within the MTL; the hippocampus (connecting with both SN/VTA and GP) and the perirhinal cortex (connecting with GP). Finally, a small set of cortical regions (involving frontal and parietal areas) was found to be sensitive to unexpected versus expected familiarity, but no such cortical effect was found for novelty. In Experiment 2, we showed that the detection of unexpected stimuli (familiar or novel) was associated with greater pupillary dilation and a greater number of fixations, denoting enhanced visual exploration and processing of the unexpected information. Both the fMRI connectivity and the eye-tracking findings suggest that enhanced learning during the detection of unexpected familiar and novel stimuli might be taking place. This hypothesis was confirmed in the final experiment (Experiment 3), which revealed that both unexpected familiar and unexpected novel stimuli were later remembered with higher recollection accuracy than expected stimuli. Connectivity Between Midbrain/Striatum and the Hippocampus is Modulated by Contextual Novelty --------------------------------------------------------------------------------------------- According to an influential model proposed by Lisman and Grace ([@b29]), a functional loop between the hippocampus and the SN/VTA underlies long-term memory encoding of new information. According to this view, a novelty signal generated in the hippocampus travels via subcortical connections to the midbrain where it stimulates the VTA neurons. The dopaminergic cells in SN/VTA release dopamine, which converges in the hippocampus resulting in the enhancement of long-term potentiation and leads to the construction of new memories (for an extension of this model, see Lisman et al., [@b30]). Dopaminergic neurons within the SN/VTA and the ventral/dorsal striatal structures have been linked to positive reward prediction errors (D\'Ardenne et al., [@b7]), reward learning (Schultz, [@b44]), the processing of unexpected information (Joshua et al., [@b18]), as well as to the facilitation of the learning of new information (Packard and Knowlton, [@b35]; Lisman and Grace, [@b29]; Lisman et al., [@b30]). Furthermore, SN/VTA has been found to respond to novel stimuli independent of their reward value (Schott et al., [@b43]), whereas coactivation with the hippocampus has been reported as facilitating long-term memory formation for rewarding or highly motivational stimuli (Wittmann et al., [@b54],[@b55]; Adcock et al., [@b1]; Bunzeck and Duzel, [@b3]; Shohamy and Wagner, [@b48]). The findings presented here further extend these observations by showing that the activation of SN/VTA and the striatum is not driven at all by the absolute novelty of a stimulus, but by the contextual novelty of both novel and familiar stimuli. Contextual expectation or prediction of novel stimuli has been reported before as modulating the amplitude and the topography of the P3a novelty ERP component (Cycowicz and Friedman, [@b6]) and as playing a pivotal role in determining the magnitude of novelty-related activity in the MTL (Dudukovic and Wagner, [@b8]; Bunzeck et al., [@b4]). However, this is the first time that such a mechanism is revealed in the SN/VTA. In addition, this study illustrates that contextual expectation modulates the connectivity between the midbrain or the striatum, and the hippocampus. Specifically, the effective connectivity analysis identified a close coupling between midbrain/striatum and MTL structures when unexpected information is detected. Intrinsic connectivity between the SN/VTA and the hippocampus has recently been found in a resting-state fMRI study (Kahn and Shohamy, [@b22]), denoting a functional coupling between the two structures. However, an important question, arising from the current models of midbrain--hippocampal crosstalk (Lisman and Grace, [@b29]; Lisman et al., [@b30]), asks under what conditions such a coupling is utilized to boost memory formation for novel or even familiar information. The results of the current PPI connectivity analysis suggest that the close coupling between the dopaminergic midbrain and the hippocampus is modulated by stimulus expectancy and is especially responsive to contextual novelty. This means that the suggested hippocampal--SN/VTA loop (Lisman and Grace, [@b29]) becomes strengthened under conditions when unexpected information is detected and less so when predictable or expected information is detected (whether novel or familiar). Indeed, expected novel stimuli did not yield midbrain activity when contrasted with expected familiar stimuli. Furthermore, a contrast between unexpected familiar stimuli and expected novel stimuli resulted in significant activity in the SN/VTA (x = 2, y = −27, z = −22, cluster FWE-corrected). This indicates that activity in the SN/VTA as a response to the detection of contextual novelty even for familiar stimuli exceeds any SN/VTA response to the absolute novelty of a stimulus alone. It should be noted here that while a few previous studies (Schott et al., [@b43]; Adcock et al., [@b1]; Wittmann et al., [@b55]; Shohamy and Wagner, [@b48]) have reported concurrent activation of the hippocampus and the midbrain/striatal system for novel stimuli, especially when these are contrasted with familiar stimuli, coactivation indicates activity synchronization between regions, it does not denote functional coupling (or indeed effective connectivity) between the two structures. In contrast, a PPI (as illustrated here) indicates that the slope of the BOLD response observed in two regions (i.e., SN/VTA and the hippocampus) is modulated by the psychological context (in this case, unexpected vs. expected novelty or familiarity; Materials and Methods section). Although, not precisely identified by the fMRI data, it can be assumed that the character of this midbrain/striatum--hippocampal connectivity is dopaminergic (Eckart and Bunzeck, [@b10]). Therefore, the present findings provide especially strong support for the proposed functional memory circuit linking the dopaminergic midbrain and the hippocampus (Lisman and Grace, [@b29]; O\'Carroll et al., [@b33]; Shohamy and Adcock, [@b47]). Furthermore, this circuit appears to exclusively engage striatal/midbrain and MTL regions without the effect of any other intervening brain structures as revealed by the lack of any significant PPI connectivity of SN/VTA and GP with any other cortical or subcortical region. This makes it highly unlikely that the increased connectivity between the midbrain/striatum and the hippocampus arises from the common influence of a third region. It remains to be established whether there are any functional differences in the contribution made by the ventral striatum (GP) and the SN/VTA to the detection of unexpected incoming information and, critically, whether the nature of their connectivity with the hippocampus is modulated by stimulus type (i.e., nominally novel or familiar). As noted above, SN/VTA and midbrain activity (see also [Fig. 2b](#fig02){ref-type="fig"}) appears to respond to both unexpected novel and unexpected familiar stimuli. In contrast, GP activity selectively signals the detection of unexpected familiar stimuli although this activity most likely also originates from the dopaminergic midbrain. Indeed, the ventral pallidum receives major afferent inputs from the dopaminergic midbrain neurons (Haber et al., [@b14]) and it is very likely that the activity of the dopaminergic neurons in the midbrain results in the modulation of BOLD activity in regions to which these neurons project (Yacubian et al., [@b57]), such as the GP. Therefore, the connectivity patterns between these two regions (GP and SN/VTA) and the hippocampus, during the detection of contextually unexpected stimuli, may suggest two distinct routes within the striatal/VTA--hippocampus loop. More specifically, and based on the connectivity findings from this study, a direct projection between the SN/VTA and the hippocampus appears to be employed during the detection of unexpected novel stimuli. This pathway may be considered a reciprocal one, with the hippocampus signaling the detection of contextual novelty to the VTA and the dopaminergic neurons of the VTA projecting their response back to the hippocampus to trigger encoding. This is consistent with the documented role of the hippocampus in novelty detection (Knight, [@b25]; Köhler et al., [@b26]; Stoppel et al., [@b50]; Kafkas and Montaldi, [@b21]). In contrast, midbrain activity for unexpected familiar stimuli seems to originate from other key memory-related brain regions such as the perirhinal cortex, which showed increased connectivity with the GP for unexpected familiar stimuli, and this signal might project to the hippocampus through the ventral striatum (GP). The selectivity of the deployment of the GP for the processing of unexpected familiar stimuli may reflect the onset of an evaluation process that computes the significance of the incoming familiar information in terms of the need for further encoding by the hippocampus. Therefore, although the hippocampus is not involved in the detection of familiar stimuli (Yonelinas et al., [@b58]; Montaldi et al., [@b32]; Kafkas and Montaldi, [@b20]), it responds to the novelty of the unexpected familiar information through its connectivity with the ventral striatum. Indeed, neurons in the monkey GP have been reported to respond not only to the learning, but critically to the evaluation of the reward value of an action (Pasquereau et al., [@b37]) and are responsible for initiating the transmission of this signal to dopaminergic centers like the SN/VTA (Hong and Hikosaka, [@b16]; see also Tecuapetla et al., [@b51]). Greater Visual Exploration and Learning Facilitation for Unexpected Stimuli --------------------------------------------------------------------------- What is the functional significance of a mechanism whereby midbrain/striatum activation and its connectivity with the hippocampus is selectively strengthened for unexpected stimuli and how is it manifested behaviorally? We hypothesized that midbrain and striatal activations may trigger enhanced visual exploration and processing of unexpected stimuli (see also Düzel et al., [@b9]). It has been shown earlier that less frequent or unanticipated events attract greater pupil dilation even when the cognitive demands of the task are kept minimal (Richer and Beatty, [@b42]; Reinhard and Lachnit, [@b39]). This effect vanishes only when the anticipation of an event plays a less crucial role or when other distinctive features of a stimulus become more salient (Reinhard and Lachnit, [@b40]; Reinhard et al., [@b41]). Indeed, in Experiment 2 we showed that the unexpected familiar and novel stimuli were accompanied by increased pupil responses and that unexpected novel stimuli were characterized additionally by increased numbers of fixations compared to those triggered by expected novel stimuli ([Fig. 2](#fig02){ref-type="fig"}). These findings suggest that the contextually novel stimuli, which are accompanied by striatal/midbrain responses and enhanced connectivity with the hippocampus, are also characterized by enhanced visual exploration and processing. Based on these findings, we hypothesized that the SN/VTA triggered hippocampal activity and the richer visual exploration of the unexpected stimuli would lead to enhanced subsequent memory for those items. Experiment 3 confirmed that unexpected stimuli (whether familiar or novel) were recollected later with higher accuracy than expected familiar and novel stimuli. These findings are consistent with the hypothesis that the role of the enhanced midbrain/striatal--hippocampal connectivity is to facilitate learning for unexpected stimuli (see also Lisman et al., [@b30]) and that this is supported by an enhanced visual exploration of these stimuli (as indicated in Experiment 2). Importantly, as it might be expected from the enhanced connectivity between the midbrain/striatum and the hippocampus, a key region for encoding subsequently recollected events (Ranganath et al., [@b38]; Staresina and Davachi, [@b49]), the memory enhancement for unexpected stimuli was selective to recollection memory with no enhancement of familiarity responses. It should be noted here that the enhanced pupil response and number of fixations characterizing the unexpected stimuli and the ensuing memory benefit cannot be explained by a RT confound. It might have been argued, for example, that the detection of unexpected information is more effortful and thus slower, leading to greater memory for these stimuli owing to the extra time allocated to processing these stimuli. However, in both experiments, the detection of unexpected familiar and novel stimuli was faster than the detection of expected familiar and novel stimuli. Therefore, neither increased effort nor extra processing time can account for the findings in the experiments reported here. The faster RT found for unexpected stimuli relative to expected ones may relate to the differential emphasis put on familiar and novel stimuli within the two retrieval conditions (FT and NT, respectively), or it may result from the engagement in a strength-rating task in the case of target (expected) but not the unexpected stimuli. One limitation of the presented studies, however, is that the fMRI and the eye-tracking data are derived from different experiments. Nevertheless, absolutely matched experimental designs were adopted in both studies with the only difference being that the fMRI data were acquired while participants lay in the MRI scanner, whereas the eye-tracking data were recorded while they were seated in a testing room. We therefore do not believe that this difference could undermine the potential relationship between the results and an intriguing possibility is that visual exploration is the direct product of the enhanced midbrain/striatal activation and connectivity with the hippocampus. Indeed, the activity of SN neurons in the monkey has been found to relate to saccadic eye movements especially when these are directed toward rewarded locations (Handel and Glimcher, [@b15]). More recently, neurons in monkey GP demonstrated synchronized activity with saccadic generation (Shin and Sommer, [@b46]). Therefore, it can perhaps be concluded that the memory facilitation for unexpected stimuli, driven by the hippocampal--midbrain system (Lisman and Grace, [@b29]), is mediated by, or at the very least correlates with, the modulation of eye movement behavior illustrated in Experiment 2. Unexpected Familiarity: a Role for Parietal Cortex -------------------------------------------------- Apart from the striatal/midbrain activations and connectivity patterns with the hippocampus, the analyses also identified areas within the inferior parietal lobe, the occipito-parietal junction, the precuneus, and the middle frontal gyrus ([Fig. 2](#fig02){ref-type="fig"}) that showed a selective response to unexpected versus expected familiar stimuli. The parietal regions, in particular, have been previously found to play a role in identifying unexpected stimuli (O\'Connor et al., [@b34]; Jaeger et al., [@b17]) or stimulus incongruency within a sequence of events (Krebs et al., [@b27]). For example, in a recent study, Jaeger et al. ([@b17]) found that distinct areas within the inferior parietal lobe monitor unexpected novelty and unexpected familiarity and proposed that their function is to orient current attention to the unexpected information. Our findings are consistent with this model although we observed only inferior parietal activations in response to unexpected familiarity, not to unexpected novelty. In addition, our connectivity data did not show any interactions between the parietal regions and the midbrain/striatal--MTL circuitry. Further studies are needed to explore any communication between the midbrain/striatal--hippocampal circuit and the inferior parietal regions. CONCLUSIONS =========== Taken together, these studies show that unexpected familiar and novel stimuli selectively activate striatal and midbrain regions and are also associated with increased visual exploration and enhanced subsequent memory. Consistent with the theoretical model proposed by Lisman and Grace ([@b29]), this memory facilitation is accompanied by increased connectivity between striatal/midbrain and the hippocampus, and the present findings show that this effect is selective to contextual novelty (for both familiar and novel stimuli) not related to absolute stimulus novelty. The findings also discriminate between two potential pathways within this system; one for the detection and processing of unexpected familiar stimuli and the other for the detection and processing of unexpected novel stimuli. Overall, the findings support the hypothesis that striatal/midbrain activity and its connectivity with the hippocampus and other medial temporal lobe structures drives and supports exploratory behavior, subsequently leading to the increased recollection of contextually novel events.	{ "pile_set_name": "PubMed Central" }
1. Introduction {#sec1-plants-09-00675} =============== Seed germination is a crucial process, characterized by a series of steps, normally occurring before radicle emergence from the seed coat \[[@B1-plants-09-00675],[@B2-plants-09-00675]\]. Radicle emergence percentage can deeply affect the production and the crop quality since the germination process is a sensitive step for plant growth and it is strictly linked to the seed quality, being the radicle emergence percentage related to seed germination potential \[[@B3-plants-09-00675],[@B4-plants-09-00675],[@B5-plants-09-00675]\]. In fact, during the germination of seeds, different enzymes such as amylases, proteases, and lipases lead to the hydrolysis of reserve substances, producing compounds that are transported to the growing seedlings for their development \[[@B6-plants-09-00675]\]. Sugar beet (Beta vulgaris subsp. vulgaris) is mainly grown in temperate climates and it is an important industrial biennial root crop, providing about 20% of the world's annual sugar production and representing a source material for bioethanol and animal feed production \[[@B7-plants-09-00675]\]. Sugar beet seed health and quality are crucial for adequate plant growth and they are strictly associated with the productive yield both in terms of quantity and quality of the crop \[[@B8-plants-09-00675]\]. Unfortunately, the productivity of sugar beet is often limited by heterogeneous germination in the field, probably due to the presence of inhibitory substances in the pericarp of the seed as well as pathogen attacks \[[@B9-plants-09-00675],[@B10-plants-09-00675]\]. Moreover, about 35--40% of the sugar beet seeds need to be discarded before sowing in the field because they are defective \[[@B11-plants-09-00675]\]. Finally, after germination, sugar beet seedlings often have to confront different biotic and abiotic stresses, therefore, rapid germination of sugar beet seeds is crucial for plant development and overall yield \[[@B12-plants-09-00675],[@B13-plants-09-00675]\]. For this purpose, different techniques have been adopted in order to increase the physiological potential of sugar beet seeds and their treatment may become an indispensable procedure aiming at the enhancement of the seed vigor and the reduction of the variability in the germination process. Among these, seed priming with water, salicylic acid, or gibberellic acid has been shown to promote sugar beet seed germination as well as enhance sugar beet seedling growth \[[@B10-plants-09-00675]\]. Szajsner et al. \[[@B14-plants-09-00675]\] showed that the seed vigor, germination speed, and the germination time significantly improved by using a magnetic field or a semiconductor laser radiation as pre-sowing treatments of sugar beet seeds. An aqueous extract from microalga Acutodesmus dimorphus was successfully used as a seed primer, improving the germination energy of tomato seeds (cv. Roma) and increasing lateral root development and extract concentrations \[[@B15-plants-09-00675]\]. Conversely, other authors found that higher seaweed extract concentrations have caused inhibition of seed germination \[[@B16-plants-09-00675],[@B17-plants-09-00675],[@B18-plants-09-00675]\]. Recently, Barone et al. \[[@B19-plants-09-00675]\] have shown that extracts from microalgae Chlorella vulgaris or Scenedesmus quadricauda may act as biostimulant in the early stages of sugar beet cultivation, when added after 5 days to the Hoagland growth solution, by improving root and plant growth and modulating gene expression related to the nutrient acquisition in sugar beet. Moreover, these extracts were successfully applied at the root level in the cultivation substrates, showing to exert a biostimulant effect on tomato and lettuce seedlings \[[@B20-plants-09-00675],[@B21-plants-09-00675]\]. Considering that C. vulgaris and S. quadricauda extract have been successfully applied as a biostimulant to several crops, including sugar beet, the aim of this work was to evaluate whether these microalgae extracts may also be able to positively affect the germination process as a priming treatment for sugar beet seed. Moreover, the concentration-dependent effects of microalgal extracts from C. vulgaris or S. quadricauda were also evaluated. Therefore, seed germination of B. vulgaris was monitored with the aim to calculate several physiological indices, useful to evaluate the effect of pre-soaking seeds with the microalgal extracts. In addition, root morphological traits were also evaluated by using WinRHIZO software in order to evaluate the effectiveness of these priming treatments. 2. Results and Discussions {#sec2-plants-09-00675} ========================== Several germination indices of sugar beet seeds were calculated and monitored (as detailed in the Materials and Methods section) in order to evaluate the effect of the treatments and their concentration dependence, using five amounts of C. vulgaris (CVextr) or S. quadricauda extracts (SQextr). In all data analysis, a significant effect (p \< 0.01) of microalgal extract concentration was observed on all germination variables of B. vulgaris seedlings. Similarly, all concentration × microalgal extract interactions were significant (p \< 0.01) for all germination parametric indices. Therefore, the experiments were always presented for each type of microalgal extract (C. vulgaris or S. quadricauda). The percentage of sugar beet seed germination (GP) is reported in [Figure 1](#plants-09-00675-f001){ref-type="fig"}. All the C. vulgaris extract-based treatments significantly affected the GP values in comparison to the untreated seeds (C0), 3 days after priming (DAP) ([Figure 1](#plants-09-00675-f001){ref-type="fig"}A). In particular, C2 and C3 C. vulgaris extract concentrations increased GP values, showing at 3 DAP a significant increase with respect to the control of 2.8 and 3.8 folds, respectively. At the remaining monitoring times, only C2 and C3 C. vulgaris extract concentrations significantly affected GP values in comparison to the untreated seeds, reaching the highest values at 4 DAP, when GP values calculated for C2 and C3 were 1.8 and 2 times higher than those calculated in the control seeds, respectively. The final GP values, calculated 7 DAP, were 1.4 (C2) and 1.5 (C3) times higher than those calculated in the untreated seeds. Conversely, among the different S. quadricauda extract concentrations, the effectiveness of the treatment was observed only after 4 DAP ([Figure 1](#plants-09-00675-f001){ref-type="fig"}B). From 4 to 7 DAP, C2 S. quadricauda extract concentration positively affected seed germination, by increasing the GP values of around 1.3 times with respect to the untreated seeds. Noteworthy also that C3 concentration significantly increased GP value with respect to the untreated seeds (1.2 folds), limited to 5 DAP ([Figure 1](#plants-09-00675-f001){ref-type="fig"}B). Results reported in [Figure 2](#plants-09-00675-f002){ref-type="fig"} show that at all the monitoring times, mean daily germination (MDG) values calculated for seeds treated with C2 and C3 C. vulgaris extract concentrations, were always greater than those calculated for untreated seeds, reaching the highest increase only after 3 DAP (2.8 and 3.8 times higher than those calculated in control seeds for C2 and C3, respectively) ([Figure 2](#plants-09-00675-f002){ref-type="fig"}A). Among S. quadricauda extract concentrations, at all monitoring times except for 3 DAP, the highest MDG values, if compared to that calculated in untreated seeds, were recorded in seeds treated with C2, showing always an increase of 1.3 times ([Figure 2](#plants-09-00675-f002){ref-type="fig"}B). These results suggest that treatments exerted different effects on GP and MDG values, strictly related to the microalgae species as well as the extract concentrations. Interestingly, the positive effect on GP values, mainly observed by using C2 and C3 concentrations of CVextr and C2 concentration of SQextr, were higher than those obtained by Szajsner et al. \[[@B14-plants-09-00675]\], who found GP values around 1 fold higher than controls by treating sugar beet seeds with a magnetic field or laser radiation. Therefore, these results indicate that these priming treatments, in particular C2 and C3 CVextr, may be a promising alternative practice aiming to enhance seed germination performance and increase GP and MDG indices, strictly related to the variability of this important physiological process. Interestingly, the lower and higher extract concentrations (C1, C4, and C5) do not significantly affect the GP and MDG germination indices, as these values were always similar to the respective controls. These data are in accordance with that reported by Santos et al. \[[@B22-plants-09-00675]\], who found that the application of algae Ascophyllum nodosum extract-based biostimulants in ornamental sunflower requires an optimal concentration to increase germination indices. Conversely, an aqueous extract from microalga Acutodesmus dimorphus behaved as an effective primer on tomato seeds, at increasing extract concentrations \[[@B15-plants-09-00675]\]. Indeed, natural biostimulants may contain various biologically active compounds which may determine concentration-dependent effects, making crucial the testing of a broad range of concentrations \[[@B23-plants-09-00675]\]. Moreover, depending on the extract type and concentration applied, natural biostimulants may elicit different responses in treated plants, being sometimes also potentially phytotoxic \[[@B24-plants-09-00675]\]. Our data suggest that the microalgal extracts did not negatively affect the germination process at lower and higher concentrations, showing no phytotoxic effect, as supported by Ronga et al. \[[@B25-plants-09-00675]\], who found no phytotoxicity in an aqueous C. vulgaris microalgal extract on cress. Interestingly, the two extracts showed an evident divergent effect on germination indices probably due to the difference in extract compositions \[[@B19-plants-09-00675]\]. Indeed, the two extracts showed a different composition in terms of organic carbon distribution ([Supplementary Table S1](#app1-plants-09-00675){ref-type="app"}) and element composition ([Supplementary Table S2](#app1-plants-09-00675){ref-type="app"}), as reported in Barone et al. \[[@B19-plants-09-00675]\]. In particular, the degree of hydrophobicity for humic substances determined according to Baglieri et al. \[[@B26-plants-09-00675]\], resulted in being much higher for CVextr (6.1) than that calculated for SQextr (3.8), showing CVextr to be more apolar than SQextr ([Supplementary Table S1](#app1-plants-09-00675){ref-type="app"}). These data suggest that the different extract compositions in organic carbon distribution as well as polarity, may be strictly related to the different effect on GP and MDG germination indices, as confirmed by Piccolo et al. \[[@B27-plants-09-00675]\], who found a relationship between the structure of the formulate and its bioactivity in humic substances. Data reported in [Figure 3](#plants-09-00675-f003){ref-type="fig"} show that the two concentrations C3 and C2 of C. vulgaris extract significantly increased the germination indices (GI) (1.7 and 1.6 times, respectively), germination energy (GE) (3.8 and 2.8 times, respectively), speed of emergence (SE) (2.4 and 2 times, respectively) and coefficient of the rate of germination (CRG) (around 1.1 times for both concentrations) ([Figure 3](#plants-09-00675-f003){ref-type="fig"}A, B, C and D), if compared to the control; on the other hand, mean germination time (MGT) and T~50~ ([Figure 3](#plants-09-00675-f003){ref-type="fig"}E and F) were significantly reduced (around 1.1 times for both indices and concentrations). As regards to the T~50~ index, all the C. vulgaris extract concentrations significantly reduced the time required for 50% germination ([Figure 3](#plants-09-00675-f003){ref-type="fig"}F). These findings support the hypothesis that the treatment with C. vulgaris extract at all the concentrations tested may be a very useful priming treatment in order to improve seed germination performance, in terms of reduction of the time required to obtain 50% seed germination. Indeed, it is well known that the higher the GI, GE, SE, and CRG values, the higher the positive effect on seed germination \[[@B28-plants-09-00675]\]. In contrast, the lower the T~50~ and MGT values with respect to the control, the lower the inhibition on seed germination \[[@B28-plants-09-00675]\]. As regards the S. quadricauda extract, the concentration effects on germination indices were less evident and differences among concentrations were not always significant ([Figure 3](#plants-09-00675-f003){ref-type="fig"}). Nevertheless, the concentration C2 positively affected GI index (1.4 times) with respect to the control, and the CRG index values were significantly higher in seeds treated with C3 and C5 concentrations than those calculated for untreated seeds. Conversely, MGT values were significantly reduced by the treatments with C3 and C5 concentrations (around 1.1 for both amounts of extract). All other germination indices were not significantly affected by S. quadricauda extract at all the tested concentrations ([Figure 3](#plants-09-00675-f003){ref-type="fig"}). These data show that S. quadricauda extract also positively affected seed germination, although to a lesser extent than C. vulgaris extract. In detail, the calculated GI, GE, and SE values were almost always lower than those relative to the corresponding concentrations of C. vulgaris, moreover, T~50~ values were not positively affected by S. quadricauda extract, showing values always similar to those calculated for the control ([Figure 3](#plants-09-00675-f003){ref-type="fig"}F). It is noteworthy to underline that these differences between the two algal extracts were significant only for the C2, C3, and C4 concentrations with regard to GI, for the C2 and C3 concentrations with regard to GE and SE indices, for C2 concentration with regard to CRG and MGT, and for C3 concentration with regard to the T~50~ index ([Figure 3](#plants-09-00675-f003){ref-type="fig"}). These results show that CVextr seems to be more effective than SQextr, with C2 and C3 the optimal concentrations, although a higher amount of CVextr did not negatively affect seed germination. These findings are supported by Ronga et al \[[@B25-plants-09-00675]\], who found that a two-fold concentration (around 25 mgC~org~/L) of an aqueous C. vulgaris microalgal extract did not show phytotoxicity effect, by measuring the GI index, on a sensitive species to phytotoxic compounds such as cress. Interestingly, from a physiological point of view, several authors \[[@B29-plants-09-00675],[@B30-plants-09-00675],[@B31-plants-09-00675],[@B32-plants-09-00675]\] have reported that the faster emergence in sugar beet showed an enormous influence on plant characteristics, showing a higher dry matter weight compared to the plants emerging later. Among these authors, Podlaski et al. \[[@B32-plants-09-00675]\] demonstrated that the time of emergence was the strongest factor influencing plant weight in sugar beet during harvest season. Analysis of morphological data provided always a significant effect (p \< 0.01) of microalgal extract concentrations on all morpho-biometric parameters (length, surface area, mean root diameter, root volume, tips, root 0.000 \< L \< 0.500 and root 0.500 \< L \< 1.000) of B. vulgaris seedlings. Since concentration × microalgal extract interactions were significant for all parametric variables, the data were presented for each microalgal extract ([Table 1](#plants-09-00675-t001){ref-type="table"} and [Table 2](#plants-09-00675-t002){ref-type="table"}). The positive effect of C2 and C3 C. vulgaris extract concentrations were also confirmed by morphological parameters ([Table 1](#plants-09-00675-t001){ref-type="table"}). In particular, all C. vulgaris extract concentrations significantly increased all root morphological parameters, both after 5 and 7 DAP, being the C2 and C3 CVextr the highest values, in comparison to the control ([Table 1](#plants-09-00675-t001){ref-type="table"}). The positive effect of C2 and C3 C. vulgaris extract concentrations were also observed at the root volume level, whereas mean root diameter values were unaffected or reduced by the treatments ([Table 1](#plants-09-00675-t001){ref-type="table"}). Similarly, all root morphological traits were positively affected by S. quadricauda extracts too ([Table 2](#plants-09-00675-t002){ref-type="table"}), although to a lesser extent than C. vulgaris. Noteworthy, the highest values of morpho-biometric parameters were reached using C2 and C3 S. quadricauda extract concentrations, whereas each amount of the SQextr did not positively affect the mean root diameter ([Table 2](#plants-09-00675-t002){ref-type="table"}), as it was already observed using C. vulgaris extract ([Table 1](#plants-09-00675-t001){ref-type="table"}). These results are in accordance with Barone et al. \[[@B19-plants-09-00675]\], who found that by applying to the hydroponic solution 1 and 2 mg Corg/L of the two microalgal extracts (C. vulgaris or S. quadricauda), root apparatus of sugar beet seedlings was positively affected by increasing total root length, root surface area, and the number of root tips, whereas the average diameter and the volume of roots were not affected by the treatments. The effect on root morphology may be related to the degree of hydrophobicity of extracts ([Supplementary Table S1](#app1-plants-09-00675){ref-type="app"}), the latter being closely related to the increase of root growth \[[@B27-plants-09-00675]\]. C. vulgaris extract showed a degree of hydrophobicity value (6.1) greater than those observed for humic substances of a different origin, ranging between 0.61 and 4.75 \[[@B33-plants-09-00675]\], whereas S. quadricauda extract showed an intermediate value (3.8). In particular, methoxilic groups, aryl groups, and carboxylic acids seem to be involved in the bioactivity of natural biostimulant substances, and often related to hormone-like compounds \[[@B27-plants-09-00675]\]. Therefore, both extracts, on the basis of their characterization, seem to be perfectly compatible with the effect observed on the morphological traits in sugar beet roots \[[@B34-plants-09-00675]\]. Moreover, the performances obtained in the present study seem to be higher than those obtained by Szajsner et al. \[[@B14-plants-09-00675]\], who treated sugar beet seeds with a magnetic field or laser radiation. In particular, Szajsner et al. \[[@B14-plants-09-00675]\] achieved an increase of the seedling length of 1.5 times with respect to the control after 4 days from the pre-sowing treatments, whereas the treatment with C3 C. vulgaris extract, induced an increase of the seedling length of around 6 times with respect to the untreated seeds at 5 days after priming treatment ([Table 1](#plants-09-00675-t001){ref-type="table"}). Finally, SVI values were also calculated both at 5 and 7 days after priming treatments ([Figure 4](#plants-09-00675-f004){ref-type="fig"}). All CVextr treatments positively affected SVI values, in accordance with other calculated germination indices. In particular, the C3 C. vulgaris extract resulted in being the most performant treatment, determining an increase of seedling vigor index (SVI) of around 9 and 3 times higher than the controls after 5 and 7 days, respectively. As regards to S. quadricauda extract, C2 induced the greatest increase of 8 and 3 times higher than the controls after 5 and 7 days, respectively ([Figure 4](#plants-09-00675-f004){ref-type="fig"}). These results suggest that the use of these microalgae extract as priming treatment, may be a good alternative to other priming methods adopted for sugar beet, according to Islam et al. \[[@B28-plants-09-00675]\], who observed that the higher the seedling vigor index (SVI) value, the higher the positive effect on seed germination. Moreover, in accordance with Ugena et al. \[[@B35-plants-09-00675]\], pre-sowing treatment with different biostimulant compounds, aiming to increase the vigor of seedlings, represents an innovative alternative to cope with different kinds of stresses. Comprehensively, all these results taken together suggest that best results, in terms of the germination process as well as root morphological traits, were reached by using the concentrations C2 and C3 of C. vulgaris extract, showing that this microalgal extract, besides exerting a biostimulant effect when added to the growth medium of sugar beet seedlings \[[@B19-plants-09-00675]\], may also be used as a priming method positively affecting the sugar beet seed germination. 3. Materials and Methods {#sec3-plants-09-00675} ======================== 3.1. Microalgae Cultivation and Extract Preparation {#sec3dot1-plants-09-00675} --------------------------------------------------- The microalgae used in this study were C. vulgaris (Beijerinck, CCAP 211/11C) and S. quadricauda (isolated from an algal company raceway pond, located in Borculo, Gelderland, the Netherlands in 2011). They were obtained by and maintained in the algal collection of the Department of Agriculture, Food and Environment (Di3A) (University of Catania, Italy) as described in Baglieri et al. \[[@B36-plants-09-00675]\]. Microalgal growth was conducted in 250 mL flask containing 150 mL of sterile BG11 culture medium \[[@B37-plants-09-00675]\] at pH 8.4, incubated on a mechanical shaker (100 rpm) at 25--30 °C, illuminated by a 3500-lx, average photon flux (PPF) 100 μmol photons m^−2^ s^−1^ light source (PHILIPS SON-T AGRO 400) with a 12 h photoperiod for 30 days in a growth chamber and aerated by pumps with 20 L h^--1^ 1.5% CO~2~ \[[@B38-plants-09-00675]\]. Microalgal biomasses were harvested by centrifugation (at 5000 rpm for 15 min), washed with distilled water (up conductivity \< 200 μS cm^−1^), and freeze-dried as described in Puglisi et al. \[[@B39-plants-09-00675]\]. Microalgal extract stock solutions (referred to as CVextr and SQextr) were prepared as described in Barone et al. \[[@B19-plants-09-00675]\]. Briefly, microalgae cells were collected and centrifuged at 5000 rpm for 15 min and the final pellets obtained from each microalgal biomass were added to methanol to lyse the cell wall in order to obtain the intracellular extracts. After centrifugation and evaporation of the organic solvent, the extracts were freeze-dried and collected with distilled water to obtain the microalgal extract stock solution. The complete characterization of extracts was reported in Barone et al. \[[@B19-plants-09-00675]\]. 3.2. Plant Material and Experimental Conditions {#sec3dot2-plants-09-00675} ----------------------------------------------- The sugar beet variety used in this study is the hybrid "Shannon" provided by the Department of Agronomy, Food, Natural Resources, Animals and Environment (DAFNAE) of the University of Padova (Italy) \[[@B40-plants-09-00675]\]. Seeds were soaked in 76% ethanol for 5 min, rinsed with sterilized water, and placed on distilled water moistened filter paper. The treatments were performed by diluting different amounts of microalgal extracts in distilled water, used to moisten the filter paper. For each microalgal extract, five different concentrations were tested and they were calculated on the basis of the extract organic carbon (C~org~) content: C1 = 0,1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Control samples using untreated seeds (C0 = 0 mg C~org~/L) were routinely performed. Germination was carried out in a growth chamber in the dark at 25 °C. Sugar beet seeds were considered germinated when a radicle of at least 2 mm emerged. The germinated seeds were counted and monitored daily for 7 days, after this time no germination, even of those not yet germinated, was detected. The experimental procedure was repeated twice in a complete randomized block design and for each treatment, four replicates consisting of 100 seeds were tested according to the methods of the International Rules for Seed Testing \[[@B41-plants-09-00675],[@B42-plants-09-00675]\]. 3.3. Germination Indices {#sec3dot3-plants-09-00675} ------------------------ In order to evaluate the effect of microalgal extract treatments, several germination indices were calculated as detailed below. The germination percentage (GP) was calculated for each treatment as a percentage of total germinated seeds after 3, 4, 5, and 7 days after priming (DAP):$${GP} = \left( {{{number}~{of}~{germinated}~{seeds}}/{{number}~{of}~{total}~{seeds}~{for}~{bioassay}}} \right) \times 100$$ The mean daily germination (MDG), representing the mean number of seeds germinated per day, was calculated at 3, 4, 5, and 7 DAP \[[@B43-plants-09-00675]\]:$${MDG} = {GP}/t$$ where GP is the germination percentage, and t is the number of DAP. The germination index (GI), also known as mean germination rate or rate of Maguire \[[@B44-plants-09-00675]\], is a measure assigning the maximum arithmetic weight to seeds that germinate at the first days of count and less weight to those germinating later. GI was calculated as follows:$${GI} = \left\lbrack {{{number}~{of}~{germinated}~{seeds}}/{{days}~{of}~{first}~{count}}} \right\rbrack + \ldots + \left\lbrack {{{number}~{of}~{germinated}~{seeds}}/{{days}~{of}~{final}~{count}}} \right\rbrack$$ The mean germination time (MGT) was calculated according to Soltani et al. \[[@B45-plants-09-00675]\] as follows:$${MGT} = \sum\left( {n,t} \right)/\sum n$$ where n is the number of newly germinated seeds at time t. The germination energy (GE) was calculated according to Ruan et al. \[[@B46-plants-09-00675]\] as follows: The speed of emergence (SE), was calculated according to Islam et al. \[[@B28-plants-09-00675]\] as follows: The coefficient of the rate of germination (CRG) was calculated according to Chiapusio et al. \[[@B47-plants-09-00675]\] as follows:$${CRG} = \left\lbrack {\left\lbrack {N_{1} + N_{2} + \ldots + N_{n}} \right\rbrack/\left\lbrack {N_{1} \times T_{1}} \right\rbrack + \left\lbrack {N_{2} \times T_{2}} \right\rbrack + \ldots + \left\lbrack {N_{n} \times T_{n}} \right\rbrack} \right\rbrack \times 100$$ where N~1~, N~2~, ..., N~n~ are the number of germinated seeds on time T~1~, T~2~, ..., T~n~. The time required for 50% germination (T~50~), was calculated according to Coolbear et al. \[[@B48-plants-09-00675]\] as follows:$$T_{50} = t_{i} + \left\lbrack {\left\lbrack {\left\lbrack {N/2} \right\rbrack - n_{i}} \right\rbrack \times \left\lbrack {t_{i} - t_{j}} \right\rbrack} \right\rbrack/(n_{i} - n_{j})$$ where N is the final number of germinated seeds, n~i~ and n~j~ the cumulative numbers of seeds germinated by adjacent counts at times t~i~ and t~j~. Finally, the seedling vigor index (SVI) was calculated at 5 and 7 DAP, according to Noorhossein et al. \[[@B43-plants-09-00675]\] as follows: 3.4. Root Morphological Analysis {#sec3dot4-plants-09-00675} -------------------------------- Root morphological analysis was performed by using WinRHIZO software (Regent Instruments) and total root length, root surface area, mean root diameter, root volume, total number of root tips, lateral root (0.000 \< L \< 0.500 and 0.500 \< L \< 1.000) were determined by computerized scanning (STD 1600, Regent Instruments, Canada) at 5 and 7 days from sowing \[[@B49-plants-09-00675]\]. 3.5. Statistical Analysis {#sec3dot5-plants-09-00675} ------------------------- Data on performances of microalgal extracts from the repeated experiment were analyzed by using the Statistica package software (version 10; Statsoft Inc., Tulsa, OK, USA). The arithmetic means of parameters were calculated, by averaging the values determined for the single replicates of each treatment. Percentage data were transformed into the arcsine (sin^−1^ square root x) prior to the analysis of variance (ANOVA). Initial analyses were performed by calculating associated F and P values to evaluate whether the effects of a single factor (concentration) and microalga × concentration interactions were significant. In the post-hoc analysis, the corresponding mean values of all parameters were subsequently separated by Fisher's least significant difference test (p = 0.05). Untransformed arithmetic means of parameters are presented in the figures and tables. 4. Conclusions {#sec4-plants-09-00675} ============== This study leads us to employ microalgae extracts as a priming agent in order to improve the germination process of an important industrial crop such as sugar beet. The novelty of this work consists of the successful application of C. vulgaris and S. quadricauda as pre-sowing treatments, resulting in the first work in which microalgae extracts are employed as a priming method for sugar beet seed germination. Microalgae extracts, especially C. vulgaris, was found to improve all calculated germination indices, the root traits as well as the seedling vigor of sugar beet, putatively promoting the further nutrient acquisition and plant growth. Among the tested extract concentrations, 1 mg C~org~/L and 2 mg C~org~/L of C. vulgaris extract revealed to be the best priming treatments, exerting a positive effect both on the germination process and morphological traits of sugar beet seedlings. Although further investigations should be performed, based on our findings, C. vulgaris extract may represent a promising practice to increase the physiological potential of sugar beet seeds. The following are available online at <https://www.mdpi.com/2223-7747/9/6/675/s1>, Table S1: Distribution of C intensity of 13C NMR (%) of Chlorella vulgaris (CVextr) and Scenedesmus quadricauda (SQextr) extracts; Table S2: Element composition (%) of Chlorella vulgaris (CVextr) and Scenedesmus quadricauda (SQextr) extracts. ###### Click here for additional data file. Conceptualization, A.B.; methodology, A.B. and P.S.; validation, A.B. and I.P.; formal analysis, A.V.; investigation, V.B. and F.F.; resources, A.B.; data curation, I.P. and A.V.; writing---original draft preparation, I.P.; writing---review and editing, A.B., I.P. and A.V.; visualization, I.P. and A.L.; supervision, A.B.; project administration, A.B.; funding acquisition, A.B. All authors have read and agreed to the published version of the manuscript. Authors thank the grants which supported this work: 'Programma ricerca di Ateneo UNICT 2020-22 linea 2′---University of Catania (Italy); "STARTING GRANT 2020"---University of Catania (Italy); Fondi di Ateneo 2020--2022 - University of Catania (Italy), linea Open Access. The authors declare no conflicts of interest. ![Germination percentage (GP) of sugar beet seeds treated with different concentrations of C. vulgaris (A) or S. quadricauda (B) extracts, for each day after priming (DAP). C0 = 0 mg C~org~/L; C1= 0.1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Data (± standard error bar) are the means of three replicates (each formed by 100 seeds). Columns within each sampling point followed by the same letters are not significantly different according to Fisher's least significant difference test (α = 0.05). The absence of letters above the columns shows the lack of significant differences.](plants-09-00675-g001){#plants-09-00675-f001} ![Mean daily germination (MDG) of sugar beet seeds treated with different concentrations of C. vulgaris (A) and S. quadricauda (B) extracts, for each day after priming (DAP). C0 = 0 mg C~org~/L; C1 = 0.1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Data (± standard error bar) are the means of three replicates (each formed by 100 seeds). Columns within each sampling point followed by the same letters are not significantly different according to Fisher's least significant difference test (α = 0.05). The absence of letters above the columns shows the lack of significant differences.](plants-09-00675-g002){#plants-09-00675-f002} ![Germination index (GI) (A), germination energy (GE) (B), speed of emergence (SE) (C), coefficient of the rate of germination (CRG) (D), mean germination time (MGT) (E) and time required for 50% germination (T~50~) (F) of sugar beet seeds treated with different concentrations of C. vulgaris (CV) and S. quadricauda (SQ) extracts. C0 = 0 mg C~org~/L; C1 = 0.1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Data (± standard error bar) are the means of three replicates. The same colored columns representing 5 concentrations of each algal extract followed by the same letters are not significantly different according to Fisher's least significant difference test (α = 0.05). The absence of letters above the columns shows the lack of significant differences. The presence of an asterisk (\) within each concentration show a significant difference between the two algal extracts.](plants-09-00675-g003){#plants-09-00675-f003} ![Seedling vigor index (SVI) after 5 and 7 days of sugar beet treated with different concentrations of C. vulgaris* (A) and S. quadricauda (B) extracts. C0 = 0 mg C~org~/L; C1 = 0.1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Data (± standard error bar) are the means of three replicates. The same colored columns representing 5 concentrations of each algal extract followed by the same letters are not significantly different according to Fisher's least significant difference test (α = 0.05). The presence of an asterisk (\) within each concentration indicates a significant difference between the two algal extracts.](plants-09-00675-g004){#plants-09-00675-f004} plants-09-00675-t001_Table 1 ###### Performance of different dosages of Chlorella vulgaris* extract on different morpho-biometric parameters on Beta vulgaris seedlings. Data derived from two repeated experiments. Standard error of the means = SEM, n = 4. Means followed by different letters within the column are significantly different according to Fisher's least significant difference test (α = 0.05). Length (cm) Surface Area (cm^2^) Mean Root Diameter (mm) Root Volume (cm^3^) Tips Number Root Number (0.000 \< L \< 0.500) Root Number (0.500 \< L \< 1.000) -------- -------------- ---------------------- ------------------------- --------------------- --------------- ----------------------------------- ----------------------------------- ----------------- --------------- -------------- -------------- -------------- -------------- ----------------- C0 0.8 ± 0.1 d 3.0 ± 0.2 c 0.2 ± 0.01 e 0.3 ± 0.01 c 0.6 ± 0.01 ab 0.6 ± 0.02 a 0.003 ± 0.0 d 0.01 ± 0.0 ab 1.5 ± 0.07 e 2.9 ± 0.1 e 0.2 ± 0.04 b 1.4 ± 0.1 d 0.6 ± 0.02 c 1.4 ± 0.1 a C1 2.8 ± 0.2 c 4.5 ± 0.2 b 0.5 ± 0.05 d 0.6 ± 0.05 b 0.6 ± 0.02 b 0.4 ± 0.04 bc 0.01 ± 0.0 c 0.01 ± 0.0 bc 3.0 ± 0.14 d 5.04 ± 0.2 c 3.0 ± 1.1 a 3.2 ± 0.2 bc 1.2 ± 0.2 b 1.2 ± 0.2 a--c C2 4.7 ± 0.1 a 6.7 ± 0.1 a 0.8 ± 0.01 b 0.9 ± 0.03 a 0.6 ± 0.01 b 0.4 ± 0.02 c 0.01 ± 0.0 ab 0.01 ± 0.0 a 3.7 ± 0.22 c 8.0 ± 0.2 b 2.3 ± 0.6 a 5.3 ± 0.04 a 2.1 ± 0.2 a 1.3 ± 0.06 a--c C3 5.1 ± 0.1 a 6.4 ± 0.2 a 0.9 ± 0.02 a 0.9 ± 0.01 a 0.6 ± 0.02 b 0.4 ± 0.01 c 0.01 ± 0.0 a 0.01 ± 0.0 a--c 4.6 ± 0.18 a 11.2 ± 0.2 a 2.7 ± 0.4 a 5.3 ± 0.2 a 2.3 ± 0.2 a 1.05 ± 0.1 bc C4 3.2 ± 0.2 bc 4.2 ± 0.1 b 0.6 ± 0.02 c 0.6 ± 0.01 b 0.6 ± 0.02 a 0.5 ± 0.01 b 0.01 ± 0.0 b 0.01 ± 0.0 a--c 4.3 ± 0.14 ab 4.4 ± 0.1 d 1.7 ± 0.3 ab 2.7 ± 0.2 c 1.2 ± 0.1 b 1.3 ± 0.1 ab C5 3.5 ± 0.1 b 4.5 ± 0.2 b 0.6 ± 0.02 c 0.6 ± 0.01 b 0.6 ± 0.02 ab 0.4 ± 0.01 c 0.01 ± 0.0 b 0.01 ± 0.0 c 4.1 ± 0.11 bc 5.2 ± 0.2 c 2.5 ± 0.7 a 3.4 ± 0.2 b 1.2 ± 0.1 b 1.0 ± 0.1 c plants-09-00675-t002_Table 2 ###### Performance of different dosages of Scenedesmus quadricauda extract on different morpho-biometric parameters on Beta vulgaris seedlings. Data derived from two repeated experiments. Standard error of the means = SEM, n = 4. Means followed by different letters within the column are significantly different according to Fisher's least significant difference test (α = 0.05). Length (cm) Surface Area (cm^2^) Mean Root Diameter (mm) Root Volume (cm^3^) Tips Number Root Number (0.000 \< L \< 0.500) Root Number (0.500 \< L \< 1.000) -------- -------------- ---------------------- ------------------------- --------------------- --------------- ----------------------------------- ----------------------------------- --------------- -------------- ------------- -------------- -------------- ------------- ----------------- C0 0.8 ± 0.1 d 3.0 ± 0.2 d 0.2 ± 0.01 c 0.3 ± 0.01 d 0.6 ± 0.01 b 0.6 ± 0.02 a 0.003 ± 0.0 c 0.01 ± 0.0 ab 1.5 ± 0.07 d 2.9 ± 0.1 b 0.2 ± 0.04 c 1.4 ± 0.1 e 0.6 ± 0.02c 1.4 ± 0.1 bc C1 2.7 ± 0.1 c 3.9 ± 0.1 c 0.5 ± 0.01 b 0.6 ± 0.03 c 0.6 ± 0.01 b 0.5 ± 0.01 c 0.01 ± 0.0 ab 0.01 ± 0.0 b 2.5 ± 0.11 c 3.1 ± 0.1 b 1.1 ± 0.1 a 2.8 ± 0.1 bc 1.9 ± 0.1a 1.3 ± 0.1 c C2 5.2 ± 0.3 a 6.2 ± 0.2 a 0.7 ± 0.03 a 0.9 ± 0.03 a 0.7 ± 0.01 ab 0.5 ± 0.01 c 0.01 ± 0.0 bc 0.01 ± 0.0 a 2.8 ± 0.13 c 5.3 ± 0.2 a 0.5 ± 0.2 bc 4.1 ± 0.2 a 1.2 ± 0.4b 1.9 ± 0.1 a C3 3.5 ± 0.2 b 5.4 ± 0.3 b 0.7 ± 0.04 a 0.9 ± 0.01 a 0.6 ± 0.04 b 0.5 ± 0.02 c 0.01 ± 0.0 a 0.01 ± 0.0 a 3.4 ± 0.14 b 5.0 ± 0.2 a 1.2 ± 0.3 a 3.3 ± 0.2 b 2.1 ± 0.1a 1.8 ± 0.1 a C4 2.9 ± 0.04 c 3.6 ± 0.2 c 0.5 ± 0.02 b 0.7 ± 0.02 b 0.7 ± 0.02 b 0.6 ± 0.04 ab 0.01 ± 0.0 a 0.01 ± 0.0 ab 4.0 ± 0.12 a 4.7 ± 0.2 a 1.0 ± 0.2 ab 1.9 ± 0.2 de 2.1 ± 0.1a 1.6 ± 0.04 a--c C5 2.6 ± 0.1 c 3.9 ± 0.1 c 0.5 ± 0.01 b 0.7 ± 0.01 b 0.8 ± 0.06 a 0.5 ± 0.3 bc 0.01 ± 0.0 a 0.01 ± 0.0 ab 2.6 ± 0.3 c 5.1 ± 0.4 a 0.6 ± 0.1 bc 2.5 ± 0.2 cd 1.8 ± 0.1 a 1.8 ± 0.3 ab	{ "pile_set_name": "PubMed Central" }
Dear Editor-in-Chief ==================== Every year in the world there are 600,000 new cases of patients diagnosed with head and neck cancer ([@B1],[@B2]). In Poland head and neck cancer accounts for 4.5% of all cancer cases which in 2008 reached 6046 ([@B3]). The aim of study was to assess the quality of life in patients' treatment due to head and/or neck cancer. Study was conducted of 48 patients (52.08% men), treatment at the Oncology Center, Maria Skłodowska-Curie Institute in Warsaw, Poland, in 2015. The standardized questionnaire EORTC QLQ-C30 and QLQH&N35 module for these patients. The study was approved by the local Ethics Committee and the management of the hospital. All subjects gave their written informed consent to participation in the study. The mean age was 56.29±6.94 yr (range: 43--67 yr). An equal number of respondents had location of tumour: larynx and laryngopharynx and oral cavity (43.75% each) and oropharynx (12.50%). Patients was treatment: radiotherapy (54.17%); chemotherapy (4.17%); surgical treatment (2.08%); radiotherapy and chemotherapy (33.33%); surgery, radiotherapy and chemotherapy (4.17%); surgery and radiotherapy (2.08%). The mean value of subjective assessment of health condition reached 4.04±1.09 and quality of life 3.79±1.17. The statistical analysis did not reveal correlations between subjective assessment of health condition and level of quality of life and sex, age, location of tumour, stage, and type of treatment. Each of aspects of physical function (difficulties in performing wearisome activities, fatigue during long and short walks, help required in the performance of everyday activities, limitations in the performance of everyday activities, limitations in pursuing hobbies, need for rest during the day and the sense of fatigue) was correlated with sex, age, location of tumour, stage and type of treatment. Physical function was affected solely by the stage and type of treatment. Thus, sex, age, location of tumour did not determine the physical ability of patients. The sense of pain (dyspnoea, weakness and pain that made the performance of everyday activities more difficult) was not determined by variables: sex, age, location of tumour, stage and type of treatment. 16.7% of patients\` health condition did not affect their family life, with 29.2% claimed there was no connection between their health condition and social life. Senior patients experienced significant or very significant difficulties (P=0.01). The results of QLQ-N&H 35 questionnaire are illustrated in [Table 1](#T1){ref-type="table"}. ###### The results of QLQ-H&N35 questionnaire *QLQ-H&N35* *N (%)* -------------------------------------------------- ------------- ----------- ----------- ----------- Pain in the mouth 10 (20.8) 19 (39.6) 11 (22.9) 8 (16.7) Pain in the maxilla 13 (27.1) 19 (39.6) 10 (20.8) 6 (12.5) Oral sensitivity 4 (8.4) 22 (45.8) 12 (25.0) 10 (20.8) Sore throat 4 (8.4) 18 (37.5) 16 (33.3) 10 (20.8) Difficulties in swallowing of liquids 7 (14.6) 17 (35.4) 18 (37.5) 6 (12.5) Difficulties in swallowing of mashed foods 8 (16.7) 20 (41.7) 11 (22.9) 9 (18.7) Difficulties in swallowing of solid foods 3 (6.3) 16 (33.3) 18 (37.5) 11 (22.9) Choking when swallowing 13 (27.1) 12 (25.0) 13 (27.1) 10 (20.8) Teeth problems 19 (39.6) 12(25.0) 12 (25.0) 5 (10.4) Difficulties in mouth opening 10 (20.8) 18 (37.5) 17 (35.4) 3 (6.3) Dry mouth 1 (2.1) 19 (39.6) 22 (45.8) 6 (12.5) Gluey saliva 4 (8.3) 14 (29.2) 23 (47.9) 7 (14.6) Impaired smell 7 (14.6) 22 (45.8) 12 (25.0) 7 (14.6) Impaired taste 2 (4.2) 15 (31.3) 18 (37.5) 13 (27.1) Cough 10 (20.8) 18 (37.5) 11 (22.9) 9 (18.8) Hoarse voice 10 (20.8) 17 (35.4) 12 (25.0) 9 (18.8) Feeling sick 6 (12.5) 17 (35.4) 17 (35.4) 8 (16.7) Preoccupied with appearance 11 (22.9) 13 (27.1) 13 (27.1) 11 (22.9) Difficulties in eating 7 (14.6) 17 (35.4) 13 (27.1) 11 (22.9) Difficulties in eating with the family 17 (35.4) 10 (20.8) 10 (20.8) 11 (22.9) Difficulties in eating with other people 17 (35.4) 12 (25.0) 13 (27.1) 6 (12.5) Difficulties in deriving pleasure from eating 6 (12.5) 19 (39.6) 14 (29.2) 9 (18.8) Difficulties in speaking with other people 12 (25.0) 17 (35.4) 11 (22.9) 8 (16.7) Difficulties in phone conversations 8 (16.7) 20 (41.7) 10 (20.8) 10 (20.8) Difficulties in family communication 19 (39.6) 12 (25.0) 12 (25.0) 5 (10.4) Difficulties in social relations 16 (33.3) 16 (33.3) 11 (22.9) 5 (10.4) Difficulties when leaving someone else's house 17 (35.4) 14 (29.2) 9 (18.8) 8 (16.7) Difficulties in contacts with family and friends 15 (31.3) 17 (35.4) 12 (25.0) 4 (8.3) Decreased interest in sex 16 (33.3) 16 (33.3) 13 (27.1) 6 (12.5) Decreased satisfaction with sex 15 (31.3) 14 (29.2) 12 (27.1) 6 (12.5) Oral sensitivity was more common among patients completed treatment compared against those who still received treatment (P=0.04). Women reported primarily difficulties in swallowing of mashed foods (P=0.03) and impaired smell (P=0.02). A statistically significant relation was found between the sense of a dry mouth and the stage of treatment (P=0.00). The difficulties associated with phone conversations were dependent on the health condition of patients (P=0.04). The difficulties in family communication significantly impaired the quality of life of patients (P=0.01). No such relation was found in case of social relations. There was a correlation between satisfaction with sex and health condition of patients (P=0.03). The study was not sponsored (own resources). Conflict of Interests The authors declare that there is no conflict of interest.	{ "pile_set_name": "PubMed Central" }
Polarized centriole positioning is important for properly oriented cell division, cilia positioning and cell migration. In mammals, planar polarized centriole postioning, as a part of the basal body of cilia, is necessary for proper directional beating of cilia at the apical surface within the node to establish left--right asymmetry, or in ependymal cells to promote cerebrospinal fluid circulation, among many other vital functions[@b1][@b2][@b3][@b4][@b5]. The coordination of cilia/centriole positioning from cell to cell across a tissue has been shown to be dependent on Frizzled/planar cell polarity (Fz/PCP) signalling in vertebrates[@b6][@b7][@b8]. Epithelial cells in Drosophila, on the other hand, do not have cilia and so the effect on cilia positioning versus general centriole positioning can be separated. PCP refers to polarization within the epithelial plane and is regulated by two distinct and conserved protein pathways, the Fat/Dachsous pathway (Ft/Ds-PCP) and the core Fz pathway (Fz/PCP)[@b9][@b10]. The conserved mechanism(s) of Fz/PCP signalling are mediated by Wnts and two protein complexes, Fz/Dishevelled/Diego/Flamingo (Fz/Dsh/Dgo/Fmi (a.k.a. Starry night, Stan)) and Van Gogh/Prickle/Flamingo (Vang (a.k.a. Strabismus, Stbm)/Pk/Fmi), which localize to opposite junctional domains within each epithelial cell[@b11][@b12][@b13][@b14][@b15][@b16][@b17], for example, to distal and proximal sides, respectively, in Drosophila wing cells. PCP signalling generally coordinates cell polarity across tissues, including ciliary positioning, the latter being reflected in the growing number of human diseases linked to aberrant Wnt-Fz/PCP pathways[@b18]. Ciliopathies, including Bardett--Biedl, Joubert and Meckel--Gruber syndromes, as well as neural tube closure defects in the embryo[@b19][@b20], are linked to vital roles of PCP in cilia positioning and orientation, and directed cell movements during gastrulation[@b21]. There is growing evidence linking core components of the Fz/PCP pathway to ciliary positioning in vertebrates, that is, in the developing mouse embryo the basal body of node cilia shifts from the centre towards the posterior side of the node cells in a PCP-dependent manner[@b22]. In fact, Inversin, a vertebrate homologue of Dgo, localizes to the basal body and axoneme and is part of the NPHP (nephronophthisis) disease module, and its loss-of-function (LOF) alleles affect cilia morphogenesis, convergent extension, and left--right determination[@b23]. Vangl2 can also localize to the basal body and axoneme in ciliated cells[@b24], and it affects the position and tilting of cilia[@b8]. Moreover, a Dishevelled triple knockout (mDvl1, 2, 3^−/−^) in multi-ciliated ependymal cells causes hydrocephalus and mis-positioning of cilia[@b25], a phenotype similar to LOF of the mammalian Fmi homologues, Celsr2 and Celsr3 (refs [@b26], [@b27]). Fz/PCP signalling controls ciliary positioning in all vertebrates examined, including mice, zebrafish and Xenopus[@b6][@b7][@b8][@b28], and it controls intracellular positioning of centrioles during zebrafish gastrulation, biasing it towards the posterior cell region[@b29]. In other contexts, including non-ciliated epithelial cells in Drosophila, the effect of Fz/PCP signalling on ciliary components, including acetylated tubulin or centrioles, remains unknown. It is thus an important evolutionary question whether Fz/PCP acts on centriole positioning in general, including in non-ciliated tissues where PCP signalling is 'active\'. We thus decided to explore centriole distribution and positioning in Drosophila wing epithelia, where the effects of PCP are well established, but which is a non-ciliated epithelium, as are all Drosophila imaginal disc epithelia. The Drosophila wing is one of the best-established tissues in which to study PCP pathways[@b11][@b12][@b13][@b14][@b15][@b16][@b17]. Adult wings manifest PCP with a single distally pointing actin-based hair in each cell (a trichome)[@b30]. At the pupal stage, when the wing is formed by two monolayers of non-ciliated epithelial cells juxtaposed at their basal membranes, at around 30--32 h APF, the trichomes start to appear, as actin polymerization becomes activated and focused at the distal apical vertex of each cell. This process depends on Rho family GTPases, which are recruited and activated by Fz-Dsh/PCP complexes[@b31][@b32][@b33]. Microtubules (MTs) also change in arrangement, from a radial to parallel distribution, projecting towards the distal apical portion of the cell with a distal bias of MT plus ends[@b34][@b35][@b36]. Although Fz/PCP signalling induces changes to the cytoskeleton, many unanswered questions remain how PCP regulates cytoskeletal elements, and it is for example unknown what type of MTs are involved in actin-hair formation in pupal wings and if these are linked to actin polymerization or centriole positioning among other options. In this study, we demonstrate that in non-ciliated cells of the Drosophila wing the positioning of centrioles is polarized towards the Fz/Dsh side of each cell and, importantly, under the control of the core Fz/PCP system. Our in vivo data in Drosophila wings argue for and provide evidence that centriole positioning is a conserved PCP readout, likely shared in all epithelial cells. Results ======= Centriole polarization in pupal wing cells ------------------------------------------ Using non-ciliated cells in imaginal discs, we asked whether centriole positioning is linked to Fz/PCP signalling as an evolutionarily conserved readout of Fz/PCP establishment (or if the presence of cilia is a pre-requisite for a Fz/PCP signalling-centriole connection). We first established a quantitative method to assess centriole positioning during pupal wing development, at the time when cytoskeletal rearrangements are being established downstream of Fz/PCP signalling, establishing a distally oriented trichome/actin hair. Two core centriolar components, Sas4 and Asterless (Asl), serve as excellent markers for centrioles; Sas4 and Asl, which is a centriolar scaffold required for centriolar assembly[@b37]. We analysed the localization of centrioles, via Sas4 and Asl staining, in pupal wing epithelial cells relative to other cellular markers, leading to two initial general observations on centriolar positioning: (1) centrioles are localized apically in cells at the level of the adherens junctions ([Fig. 1](#f1){ref-type="fig"}); and (2) Centriole positioning became progressively more polarized and localized to the distal vertex of each cell ([Fig. 1](#f1){ref-type="fig"} and [Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). Centrioles were detected at the adherens junction level, which were labelled with Fmi, and were never detected more basally (for example, at the level of Dlg/Discs large, a marker for baso-lateral membrane[@b38]; [Fig. 1](#f1){ref-type="fig"} and [Supplementary Fig. 1](#S1){ref-type="supplementary-material"}). This apical localization is very similar to that in vertebrate polarized epithelial cells. As the cells matured and started to display polarized, distal actin polymerization (phalloidin staining; [Fig. 1](#f1){ref-type="fig"}), centrioles became also localized to the distal vertex of each cell ([Fig. 1](#f1){ref-type="fig"} and [Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). To characterize the timing of centriole re-positioning in pupal wing epithelial cells, we determined centriole positions before and during hair formation relative to the centre of the cell, measuring distance and angles between the centroid of the cell and the centroid of the centriole in individual cells ([Fig. 1f--w](#f1){ref-type="fig"}; see Methods and [Supplementary Fig. 2](#S1){ref-type="supplementary-material"} for technical details). Before the actin-based hair is formed (28--30 APF), centrioles are positioned near randomly but close to the centroid of the cell ([Fig. 1f--i](#f1){ref-type="fig"}; see quantification in [Fig. 1j--k](#f1){ref-type="fig"}), and can be found at any angle (rosette diagram in [Fig. 1j](#f1){ref-type="fig"}; also [Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). At 31 APF when actin starts to be enriched at distal vertex, centrioles appear biased towards the distal cellular vertex, the distal quadrant between −45° and +45° ([Fig. 1l--q](#f1){ref-type="fig"}, compared with non-polarized localization at 29--30 APF, [Figure 1f--k](#f1){ref-type="fig"}). Once hairs are detected in all cells (32--33 APF), centriole localization is fully polarized to the distal vertex of cells (relative to the wing margin) ([Fig. 1r--w](#f1){ref-type="fig"}). This subcellular distribution of centrioles is very similar to that of the actin-based hair centroids ([Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). Taken together, we conclude that centriole positioning becomes planar polarized in non-ciliated epithelial cells, following largely the same distribution as the trichomes, the actin-based hairs and the best defined 'cellular effect\' of Fz/PCP signalling in Drosophila wings. PCP signalling instructs centriole positioning ---------------------------------------------- As PCP signalling regulates wing hair (trichome) formation in Drosophila and the position of basal bodies (cilia) in vertebrates, we next investigated whether Fz/PCP regulates the position of centrioles during pupal wing development. We characterized the positioning of the centrioles in loss or gain-of-function core Fz/PCP pathway backgrounds, with the same quantitative approach described above, using en-Gal4 driven fmi-IR (dsRNA knockdown; see Methods) and fz null allele (fz^P21^) wings ([Fig. 2](#f2){ref-type="fig"}, and [Supplementary Fig. 3](#S1){ref-type="supplementary-material"}) and dpp-driven Fz overexpression (Fz-OE; GOF) background. In both genetic scenarios, centriole positioning was altered ([Fig. 2](#f2){ref-type="fig"} and [Supplementary Fig. 3](#S1){ref-type="supplementary-material"}). Although apical localization was maintained, centriole positioning remained unpolarized and more centered within mutant cells in both cases (cf. heat maps in [Fig. 2f,l,r,t](#f2){ref-type="fig"}), consistent with randomized PCP in each case. Moreover, angle distribution was spread over a much wider range of angles with WT regions of same wings remaining polarized and serving as controls (compare rosette diagrams in [Fig. 2e,k,q,s](#f2){ref-type="fig"}, respectively; see also [Supplementary Fig. 3](#S1){ref-type="supplementary-material"}). These data are consistent with the notion that in epithelial cells centriole positioning is generally connected to Fz/PCP signalling. Cytoskeleton and centriole localization are linked -------------------------------------------------- To determine how centriole positioning in pupal wing cells relates to the cytoskeleton, we next analysed both MTs and actin in pupal wing cells. MT localization and actin have been previously stained during PCP establishment in pupal wings, using confocal and electron microscopy[@b32][@b34][@b39], but the nature of the MTs remained unexplored. We tested whether they were acetylated, a marker of stable MTs, since acetylated MTs are generally associated with cilia and the basal body in ciliated cells. In pupal wings before hair formation, actin was enriched in the apical plane of the cell and acetylated MTs also formed a web-like structure in the apical plane of each cell ([Fig. 3a--c](#f3){ref-type="fig"}). Once actin-based hairs started to form and actin polymerization was focused at the distal vertex of the apical membrane, acetylated MTs became enriched at the base of the hairs ([Fig. 3d--f](#f3){ref-type="fig"}). Moreover, during hair formation, acetylated MTs started to 'invade\' the hair itself, forming what looks like a scaffold-structure for the trichome ([Fig. 3d--f](#f3){ref-type="fig"}). Of note, the juxtaposition between actin and acetylated tubulin was maintained into deeper areas in the cell, not just in the most apical planes from which the hair projects ([Supplementary Fig. 4](#S1){ref-type="supplementary-material"}). When centrioles were co-labelled with actin in the presence of the hair structure/trichome, they were localized adjacent to the base of the hair (co-labelled with actin; [Supplementary Fig. 5](#S1){ref-type="supplementary-material"}). The close co-localization of centrioles, acetylated tubulin and actin raises the question of whether they remain connected or become disconnected in PCP LOF or GOF backgrounds. Knockdown of fmi in the posterior wing compartment causes PCP defects, reflected in aberrant hair orientation and some multiple cellular hair (mch) defects ([Fig. 2g](#f2){ref-type="fig"} and [Supplementary Fig. 3i](#S1){ref-type="supplementary-material"}). Generally in mutant PCP backgrounds, the actin hair in each cell appears unpolarized (within the apical membrane plane), often near the centre of a cell. When acetylated tubulin was examined in fmi-IR conditions, it remained associated with the growing actin hair, independently of where the hair growth was positioned within the cell. Furthermore, this link between acetylated tubulin, centrioles and actin remained in GOF scenarios, for example, when Fz was over-expressed (for example in the dpp stripe of 8--14 cells close to vein L3; [Supplementary Fig. 6](#S1){ref-type="supplementary-material"}). These results suggested that acetylated tubulin/ MTs are a structural component of the actin-based hair and that the link between actin and acetylated tubulin is independent (or downstream) of Fz/PCP establishment, even though the position of the structure within the cell is regulated by Fz/PCP activity. Next we asked whether the number of actin hairs or hair positioning and thus actin polymerization downstream of Fz/PCP could affect centriole localization. The null allele of multiple wing hair (mwh^1^) shows mch formation. Mwh, becomes localized in response to Fz/PCP factor interactions and acts downstream of the Fz/PCP effectors Inturned and Fuzzy, and it is thought to directly affect actin polymerization. Moreover, it does not affect localization of the core Fz/PCP factors (for example, Fmi or Fz (refs [@b40], [@b41])). Centriole localization was less polarized in mwh mutants ([Fig. 3g--j](#f3){ref-type="fig"}; see heatmap in [Fig. 3e](#f3){ref-type="fig"}, although angular distribution was less affected when compared with fz or fmi LOF and GOF backgrounds; for example, compare [Figs 3k](#f3){ref-type="fig"} and [2k,s](#f2){ref-type="fig"}). These data suggest that centriole positioning is a downstream event of localized actin polymerization, which in turn is regulated by Fz/PCP. Over-expression of the Sple isoform of pk in developing wings has recently been reported to reverse PCP orientation, resulting in actin hair formation being moved to the proximal cellular vertex without affecting Vang or Dsh localization[@b36][@b42]. Importantly, Sple-OE (see Methods) caused not only a reversal of actin hair polymerization but also a reversal of centriole positioning, with the majority of centrioles located to the proximal vertex ([Fig. 3m--r](#f3){ref-type="fig"}). Taken together, these data are consistent with the model that centriole positioning is a downstream readout of core Fz/PCP signalling, similarly to properly localized actin polymerization, and, more importantly (based on the mwh LOF effects) that centriole positioning is a downstream effect of localized actin polymerization as regulated by the core Fz/PCP pathway. Loss or gain of centrioles does not affect PCP ---------------------------------------------- We next asked if loss or gain of centrioles could affect core PCP factor localization. Importantly, loss of centrioles, via Sas4 knockdown or the sas4^S2214^ null allele, did not result in detectable PCP phenotypes in pupal wing cells when Fmi staining/polarization was assessed ([Fig. 4a--e](#f4){ref-type="fig"} and [Supplementary Fig. 8](#S1){ref-type="supplementary-material"}), although growth defects were observed as recently reported ([Supplementary Fig. 8](#S1){ref-type="supplementary-material"}; surprisingly, flies lacking Asl or Sas4, and thus centrioles, survive to adulthood due to compensatory cell proliferation; their adult wings exhibit blisters, overgrowth, and vein mis-patterning[@b43][@b44][@b45]). In contrast, Asl and Sak/Plk4 (Polo-like kinase 4) overexpression caused an increase in centriole number[@b37][@b46], for example under en-Gal4 conditions (en\>Asl; [Fig. 4f,m](#f4){ref-type="fig"}) or ubiquitously (Sak/Plk4, expressed via ubiquitin promoter; not shown). In such conditions, we did not detect defects in Fmi localization at the pupal stage ([Fig. 4f--l](#f4){ref-type="fig"}), with respect to either core PCP generated nematic order, actin hair orientation or centriole positioning (see also [Supplementary Fig. 7](#S1){ref-type="supplementary-material"}). Strikingly, however, in the thus generated multi-centriolar cells, when more than two centrioles were present (Asl normally only labels one centriole per cell, the mother centriole; [Supplementary Fig. 1](#S1){ref-type="supplementary-material"}), all centrioles were positioned near the distal vertex ([Fig. 4k--m](#f4){ref-type="fig"}), and moreover were always located near the base of the hair ([Supplementary Fig. 7](#S1){ref-type="supplementary-material"}). These results indicate that the number of centrioles does not impact Fz/PCP signalling and further confirm the notion that centriole positioning is a downstream effect of core Fz/PCP signalling via its effects on actin polymerization. Discussion ========== Taken together with observations that Fz/PCP signalling regulates basal body and cilia positioning in vertebrates[@b6][@b7][@b8][@b28], our data on centriole positioning as a Fz/PCP readout in non-ciliated Drosophila wing cells indicate that centriole/MTOC (MT organizing centre)/basal body positioning is an evolutionarily conserved downstream effect of Fz/PCP signalling. Its link with actin polymerization (hair formation in Drosophila wing cells) suggests that actin polymerization effectors also affect cilia positioning, possibly through docking of the basal bodies to the apical membranes. Inturned, Fuzzy and Rho GTPases regulate apical actin assembly necessary for the docking of basal bodies to the apical membrane[@b47][@b48] and this apical actin membrane accumulation is lost in Dvl1-3-depleted cells[@b47][@b48][@b49][@b50]. In left--right asymmetry establishment of the Drosophila hindgut, which is not a Fz/PCP-dependent process[@b51], asymmetric centriole positioning is observed. During this so-called planar cell shape chirality process, which affects gut-looping and thus embryonic left/right asymmetry, centriole positioning is however still dependent upon actin polymerization downstream of Rho GTPases (Rac and Rho), via MyoD and DE-cadherin control[@b51][@b52]. As Rho GTPases (Rac, Cdc42 and Rho) are downstream effectors of Fz-Dsh/PCP complexes, and their mutants cause PCP-like phenotypes including mchs or loss of hairs in wing cells[@b31][@b32][@b53]. It is thus tempting to infer that both processes, planar cell shape chirality and Fz/PCP, regulate centriole positioning through a common Rho GTPase-mediated actin polymerization pathway, initiated by an upstream cellular communication system, although this assumption will require experimental confirmation. In the mouse, Fz/PCP signalling regulates cilia movement/positioning in cochlear sensory cells via Rho GTPase-mediated processes[@b54], suggesting a similar mechanism in a representative mammalian PCP model system ([Fig. 4n](#f4){ref-type="fig"}). In conclusion, the positioning of centrioles appears to be a key and an evolutionary conserved downstream readout of Fz/PCP signalling, ranging from flies to mammals in both ciliated and non-ciliated cells. Methods ======= Fly strains ----------- Flies were raised on standard medium and maintained at 25 °C, unless otherwise indicated. GAL4/UAS system[@b55] was used for gene expression and RNAi studies. The Gal4 expression drivers were en-GAL4, dpp-GAL4 and nub-GAL4. In addition the following lines were used: fmi RNAi (Mlodzik lab stock, ML117 (2)), Sas4 RNAi (KK106051 from VDRC and BL35049 from Bloomington Stock Centre), Sple-OE (gift from Masakazu Yamazaki; fly ID: TID29239 ref [@b42]), GFP-Asl-OE, Sas4-GFP, sas4^S2214^ (gift from Jordan Raff), fz^p21^, mwh^1^ (described in Flybase), UAS-Fz for Fz-OE experiments[@b56]. The different Gal4 lines applied in this study were used to direct expression of the UAS-constructs to distinct wing compartments, linked to the localized expression of developmental genes such engrailed (en), restricted to the posterior compartment of the wing, decapentaplegic (dpp), expressed in a stripe between L3 and L4, and nubbin (nub) expressed in the whole wing. Immunohistochemistry and immunocytochemistry -------------------------------------------- White pupae (prepupae) were collected and staged at 25 °C for indicated time points. Wings were dissected in PBS with 0.1% Triton X-100 (PBT) and fixed with 4% formaldehyde for 45 min at room temperature. Pupae were then washed twice in PBT and blocked in PBT with 2% bovine serum albumin for 30--45 min. Samples were incubated with primary antibodies overnight at 4° in PBT-0.2% bovine serum albumin. Samples were washed five times in PBT and incubated for 1 h in fluorescent secondary antibodies diluted in PBT and fluorescent phalloidin when indicated. Five additional washes in PBT were performed before pupal wings were detached from the pupal cage and mounted on slides with Vectashield (Vector Laboratories). Pupal wing images were acquired using a confocal microscope (× 40∼oil immersion, 1.4 NA; SP5 DM; Leica) with LAS AF (Leica) software. The following antibodies were used: Anti-Asl (gift from Jordan Raff); anti-Fmi (from DSHB); anti-Dlg (gift from Kuyng-Ok Cho); anti-Cnn (gift from Jordan Raff); and acetylated-tubulin (from Sigma). Secondary antibodies used at 1:200 were from Invitrogen with different Alexa fluorophores (Alexa 568 and Alexa 647). FITC-phalloidin, rhodamine-phalloidin and Alexa 647-phalloidin were used at 1:500--1,000 (Invitrogen). Quantitative analyses of centriole positioning ---------------------------------------------- A novel function in MATLAB was developed to assess the relative position of centrioles within cells using immunohistochemical images. As input, this function uses (i) an immunofluorescent image of centrioles marked by Sas4 or Asl and (ii) the same image, marked by Fmi or Dlg and processed with the software 'packing_analyzer_V2\' to obtain the sketches of the cell borders. First, each cell is automatically detected and its centroid (R) is calculated as the center of the mass of the cell, whose Cartesian coordinates are where (r~ix~, r~iy~) are the Cartesian coordinates of the pixel r~i~, and n represents the sum of pixels in the cell. Then, the centrioles present within the cell are recognized and, again, the centroid (C )representing the center of mass of each one is obtained as where (p~jx~, p~jy~) are the coordinates of the pixel p~j~, and m represents the sum of pixels in the centriole. The angle (α) formed between the centroid of the cell and the centroid of each centriole is calculated as with a correction in the orientation for angle detection in the 0\<=α\<=2π radians range. The final results from a full set of cells were plotted as a rosette diagram representing these angles, and χ^2^-tests were used to detect significant differences between angle distributions (see figure legends). Another representation of the relative centriole positions inside the cells was performed as follows: A general cell model for a specific genotype was generated as a 51 × 51 matrix called M. Then, each pixel p of each centriole was assigned to an element E in the matrix following the equation: rounding each element to the nearest integer. eqRad represents the equivalent radius of the cell (the radius of a circle with the same area as the cell), and calculated as The final cell model is obtained as the density histogram of M using the whole set of pixels of each centriole, and represented as a colour heatmap. This function is available on the website <http://www.neural-circuits.org/>other-software, and can be applied to the relative location of other subcellular structures, like the actin hair. Quantitative analyses of polarized Fmi localization --------------------------------------------------- Polarity as determined from anti-Fmi stained cells was calculated with the software 'packing_analyzer_V2\' as described in ref. [@b57]. The software calculated both angles and strength of polarization (nematic order)[@b57]. Rosette figures were generated to represent the 360° orientation of the nematic order/polarity vector population, with 0° always being oriented as pointing distally using MATLAB. Statistical tests were used to assess differences between cellular orientation distributions (χ^2^-test) or polarity strength/nematic order (t-test). Additional information ====================== How to cite this article: Carvajal-Gonzalez, J. M. et al. Positioning of centrioles is a conserved readout of Frizzled planar cell polarity signalling. Nat. Commun. 7:11135 doi: 10.1038/ncomms11135 (2016). Supplementary Material {#S1} ====================== ###### Supplementary Information Supplementary Figures 1-8 We are most grateful to Jordan Raff for many reagents (flies and antibodies), and all members of the Mlodzik and Carlo Iomini labs for helpful suggestions, comments and discussions during the development and execution of the project. We thank the Bloomington and VDRC Stock Centers for fly strains, and the DSHB for antibodies. We are grateful to Giovanna Collu for helping with the manuscript. J.M.C.-G. is a recipient of an Atracción y Retención de talento contract from the GOBEX (Extremadura government). This work was supported by NIGMS and NEI grants from the National Institutes of Health to M.M., and BFU2014-54699-P grant from the Ministry of economy to J.M.C.-G. Confocal microscopy was performed at the Tisch Cancer Institute Microscopy Core, supported by grant P30 CA196521 from the NCI of the National Institutes of Health. Author contributions J.M.C.-G. and M.M. designed the experiments, analysed the data and wrote the manuscript. J.M.C.-G. performed all of the experiments. A.-C.R. designed and performed the data analysis. ![Centriole localization and positioning during PCP establishment in pupal wings.\ (a) Illustration of pupal wing and its orientation (b--d) Sas4-labelled centrioles (cyan in b and monochrome d) are distributed in the junctional planes (X--Z plane---upper panels) (marked by Fmi staining, red in b, monochrome in c). Top panels are x--z sections of respective x--y views shown below. Scale bar, 10 mm. (e) Schematic representation of a pupal wing epithelial cell and the parameters used to study centriole positioning. (f--w) During pupal wing development, centriole localization changes. (f,l,r) Sas4 (green), Fmi stained in red, and actin (phalloidin) in blue, and the respective monochromes. (f--k) Before hair formation (29 h APF), centrioles are unpolarized in a central position in the apical portion of each cell (quantified in f,k). (l--q) At the onset of hair formation (31 h APF), centrioles begin to localize to the distal portion of each cell (quantified in p,q). (r--w) Subsequently, when hairs are fully present in all wing cells (32--34 h APF), centrioles appear to be polarized mostly to the distal sector of each cell (quantified in v,w). Scale bar, 10 mm. Red sectors in f,p,v---% within distal quadrant. Statistical analyses: rosette diagram distributions panel g versus p: P\<0.0001; p versus v: P\<0.0001 (χ^2^-test).](ncomms11135-f1){#f1} ![Centriole position is affected by PCP signalling.\ (a--l) Fmi LOF, using en-driven fmi-IR knockdown affects centriole localization; (a,g) Dlg: red, cell outline; Sas4: green, centriole; actin (phalloidin): blue; and respective monochromes in a--d,h--j. Centrioles within cells in the en\>fmi-IR area (g) are less polarized and are distributed more centrally, quantified in e--f and k--l, respectively. (m--t) Fz gain-of-function (GOF; see Methods), causes Fmi depolarization (see polarity vectors in p), and defects in centriole distribution (Fmi in red, monochrome in n; Sas4 in green, monochrome in o). Fz overexpressing cells have central distribution of centrioles (quantified in s--t; see also ROI: Fz-OE in p), as compared with WT areas of dpp\>Fz-OE wings (quantified in h,i; see ROI: WT in p). Scale bar, 10 μm. Statistical analyses: centriole rosette diagrams, e versus k: P\<0.0001; q versus s: P\<0.0001 (χ^2^-test).](ncomms11135-f2){#f2} ![Centriole localization relative to trichome position.\ (a--c) XZ and two XY optical sections of same wing area, showing polymerized actin (phalloidin, cyan) and acetylated tubulin staining (red, monochrome in b). At 29 h APF before actin-based hairs are formed, both actin and acetylated tubulin appear enriched at the apical portion of the pupal wing cells; two planes, apical (XY1) and subapical (XY2) are shown in the XY-axis (indicated in the XZ sections as black lines 1 and 2). (d--f) After hair formation (32 h APF), acetylated tubulin is enriched at the base of each hair and within the polymerized actin structure. XZ and two XY planes, apical (XY1) and subapical (XY2) are shown in the XY-axis from the same; stainings as in a. (g--l) Centriole localization in the mwh^1^ mutant is less polarized than wt (in 32 h pupal wing cells; compared with [Fig. 1v,w](#f1){ref-type="fig"}). Asl (green), Fmi (red) and phalloidin (blue) stainings and the respective monochromes are shown. k and l show centriole distribution quantifications in mwh^1^ mutant, note shift to less polarized, central distribution most evident in heat map (l). (m--q) Overexpression of Sple (Sple-OE) causes reversal of cellular polarity and hair position, and accordingly also centriole positioning is inverted; quantifications in q,r. Asl staining position was used for quantifications in rosettes (k,q) and heat maps (l,r). Scale bar, 10 μm. Statistical analysis: centriole rosette diagrams, k versus [Fig. 1w](#f1){ref-type="fig"} (wt control): P\<0.001; q versus wt control: P\<0.0001 (χ^2^-test).](ncomms11135-f3){#f3} ![Loss or gain of centrioles does not affect core PCP factor localization.\ (a--e) Imaging and quantifications (sed Methods) of Fmi staining in Sas4 RNAi-mediated knockdown (in cells marked with GFP) in the posterior compartment in 32--34 h APF pupal wings. en\>Sas4^IR^ RNAi-mediated KD (a--b) did not disrupt Fmi localization as compared with the WT anterior compartment from same wings. (c,d) Quantifications of Fmi polarization using polarity vector angle orientation revealed no significant differences between en\>Sas4^IR^ (n=1,872 cells from five independent wings) and WT cells (n=1,513 cells, five independent wings). NS: non-significant (P=0.655) (χ^2^-test)). (e) Polarity vector lengths (relative to Fmi fluorescence polarization) did not show significant differences between in Sas4^IR^ cells and adjacent WT cells (five independent wings). NS: non-significant. (f--j) Gain of centrioles (\>2 per cell) through Asl overexpression (n=1,099 cells, three independent wings) in the posterior compartment (en\>Asl), did not affect Fmi localization as compared with wild type from the same wings (n=1,096 cells, three independent wings). (h,i) Quantification of angle distribution of polarity vectors, and (j) polarity vector length. Note no significant changes between wt and en\>Asl cells (P=0.783) (χ^2^-test). (k--m) Centriole positioning follows PCP core factor localization even when higher numbers of Asl-positive centrioles are present per cell. Quantifications are depicted in rosette (k) and heatmap (l) diagrams. (m,n) Specific examples of centriole localization in multi-centriolar cells, note several examples with 3--4 centrioles per cell (m, top), and two Asl-positive centrioles per cell (m, bottom); in WT only 1 centriole is stained by Asl for comparison (Asl in green; cf. to [Fig. 1a,b](#f1){ref-type="fig"}). In all cases, centrioles remained close to the distal side of each cell (marked with Fmi; magenta). (n) Schematic representation of signalling pathways involved in centriole polarization related to planar cell polarity in vertebrates versus Drosophila. Scale bar, 10 μm.](ncomms11135-f4){#f4} [^1]: Present address: Departamento de Bioquímica y Biología Molecular, Facultad de Ciencias, Universidad de Extremadura, Badajoz 06071, Spain.	{ "pile_set_name": "PubMed Central" }
![](yjbm00632-0073.tif "scanned-page"){.329} ![](yjbm00632-0074.tif "scanned-page"){.330}	{ "pile_set_name": "PubMed Central" }
Background ========== The notion \'hard to reach\' is a contested and ambiguous term \[[@B1]\] that is commonly used within the spheres of social care and health, especially in discourse around health and social inequalities. Reducing inequalities in health and health provision is a key theme underpinning the NHS Plan and National Service Framework\' \[[@B2]\]. This same report acknowledged that certain groups are marginalized from services and therefore \'harder to reach\' for health services whose goal is to provide appropriate and equitable health care for all populations. The Framework stated that primary care has a significant role to play in providing services that reach those in the greatest need but a Home Office report in 2004 \[[@B3]\] described the \'hard-to-reach\' as \'inaccessible to most traditional and conventional methods for any reason.\'. On the other hand, voluntary and community organisations (VCOs) are considered \'uniquely placed to reach marginalised groups\' \[[@B4]\] and the \'hardest to reach\' \[[@B5]\], and play an important role in the provision of health and social care services for these groups \[[@B6]\]. This paper describes the first stage of a planned study aimed at improving understanding of the \'hard to reach\' and improving access to primary health services for \'hard to reach\' groups. Acknowledgement of the established role of the Voluntary and Community Sector in providing services for these groups informed a decision to begin the study by exploring the perceptions and experiences of organisations from this sector. This first stage of the study had two research objectives - ◦ To describe how service providers from the voluntary and community sector (VCS) conceive of the notion \'hard to reach\'. ◦ To explore perceptions of the barriers and facilitators to accessing services for \'hard to reach\' groups. Using the search terms: \'hard to reach populations\' \'hidden populations\', \'marginalized populations\', a literature search was conducted using the following databases: Books @ Ovid, Journal @ Ovid Full text, British Nursing Index (BNI), CAB Abstracts (1973-2006), CINAHL, EMBASE, Ovid Medline, PSYCINFO, Sociological Abstracts and Web of Science. Searches were restricted to English-language papers and limited from 1997 to present day. A search of the grey literature was also conducted using the Home Office and Department of Health websites. A Google search was undertaken limited to last 12 months using the term \'hard to reach Groups.\' Despite the apparent familiarity of the term and its use in social research and public policy, there would appear to be a lack of consensus about the meaning of the term. Within the literature, \'hard to reach\' is often synonymised with other terms and the sheer multiplicity of alternatives reflects the divergence in the discourse as well as the difficulty in arriving at a definitive description of its meaning. \'Vulnerable\', transient\' \[[@B7]\] \'marginalised\' \'refusers\' \[[@B8]\], \'hidden\' \[[@B9]\], \'forgotten populations\' \[[@B10]\], \'underserved\' \[[@B11]\], \'special populations\' \[[@B12]\] \'disadvantaged populations\' \[[@B13]\] are terms that have all been utilised in the literature pertaining to addressing issues faced by the \'hard to reach\', or indeed those who are trying to reach them. There was an overarching sense that trying to engage the \'hard to reach\' is problematic, as described almost two decades ago by Freimuth & Mettger \[[@B14]\](1990) and this may be due to the inherent ambiguity and lack of clarity of definition. Assignation of the term appears largely dependent upon the context of the organisation doing the reaching. The \'hard to reach\' groups most commonly identified in the literature are sex workers, drug users, people living with HIV and people from lesbian, gay, bisexual, transgender and intersex communities \[[@B15]-[@B17]\], but there are a number of other groups to which the description applies including asylum seekers, refugees, black and minority ethnic communities (BME), children and young people, disabled people, elderly people \[[@B18]\] and traveller families \[[@B19]\]. The \'hard to reach\' may be people who are stigmatised due to the fact they are perceived as being somehow different. If we think in terms of stigma and social exclusion in a healthcare context, the \'hard to reach\' may include people with a variety of conditions or life limiting illnesses. This may include people with congenital abnormalities \[[@B20]\] or genetic conditions \[[@B21]\]; people with hearing loss \[[@B22]\]; people with mental health problems \[[@B23],[@B24]\]. It is important to address also the notion that there are populations who are deemed \'hard to reach\' but who are \'non-associative\'; that is they do not generally associate with other members \[[@B25]\]. These populations are comparatively under-researched and it is acknowledged that trying to gain access to disparate and often isolated individuals, although labour intensive, is an important area of consideration for researchers and indeed service providers working with the \'hard to reach.\' \'Hard to reach\' audiences have been defined as \'inaccessible to most traditional and conventional methods for any reason\' \[[@B3]\], highlighting the difficulty facing statutory health care provision. The voluntary and community sector has arguably been more successful in penetrating some of the barriers for the \'hard to reach\' and has an important role to play in the understanding of service delivery provision. There are around half a million voluntary and community organisations (VCOs) in the UK and there is a long tradition of voluntary action and community service in this country \[[@B4]\]. They have been used increasingly to deliver public services, especially over the last ten years \[[@B26]-[@B28]\]. The VCS is a term used to describe self-governing organisations working in areas of public benefit and they are \'value-driven\' in that they \'exist for the good of the community\' \[[@B29]\]. They can range in size from small community based groups or projects to large national and international organisations. As the term suggests, these services often rely upon volunteers for service provision and delivery and are characterised by their independence from the \'formal structures of government and the profit sector\'. However, in practice it appears that the sector \'defies easy definition\' \[[@B30]\] and it is suggested that the inherent ambiguity and variation in terms of how various organisations define and describe themselves is the very reason they are able to provide such diversity and flexibility. Research Objectives ------------------- The aim of the study was to explore \'hard to reach\'-from the experiences and perspectives of the VCS. The findings would contribute to a larger study considering how these maybe translated to provision of services in primary health care. The specific objectives of the work described in this paper were: ◦ To describe how service providers from the voluntary and community sector (VCS) conceive of the notion \'hard to reach\'. ◦ To explore perceptions of the barriers and facilitators to accessing services for \'hard to reach\' groups. Methods ======= This study utilised a qualitative design, since qualitative methods are well placed to explore sensitive and relatively poorly understood aspects of social life \[[@B31]\]. Semi-structured interviews were conducted with purposefully sampled team leaders or practitioners from VCS organisations across Birmingham. Purposive sampling or criterion based sampling \[[@B32],[@B33]\] was used to enable recruitment of participants who have particular experiences and roles that are pertinent to the study objectives. The Birmingham Index of Voluntary Organisations was employed to access details of organisations working with \'hard to reach groups\'. Letters of invitation were sent to 30 organisations and these were followed up by a telephone call, introducing the researcher and offering further details about the project. The organisations contacted included services working with people with mental health problems; homeless and poorly housed people; problem drug users, people identifying as gay, lesbian, bi-sexual and transgender; parents of young children in disadvantaged areas; people from BME communities; people working as carers; older adults; people with learning difficulties. These organisations were contacted as the groups they worked with are described in the literature pertaining to stigmatised, marginalised or hard to reach groups \[[@B23],[@B34],[@B15],[@B17],[@B34]-[@B38]\]. The initial plan was to interview service leads or managers as it was felt that they may be in the best position to give both an overview of their services, as well as providing accounts of front-line experience. However, it was not always possible to interview team leads, and participation was delegated by the service manager/team leads to other practitioners from within the projects, who in all but one case held a management role of some kind within the project. Eight services agreed to participate and the organisations worked with the following groups - young people, members of the Irish community, homeless people, people from BME communities living with HIV and people from the lesbian, gay and transgender communities, families and carers of drug users, young homeless people and people with mental health problems. The organisations that declined to participate cited time restrictions as their reason for non-participation. The literature review on \'hard to reach groups\' informed the development of the topic guide for the semi structured interviews. The topic guide covered the following areas: • remit of your organisation and the client group you work. • Experience of working with this group/in this field. • Understanding of the term hard to reach. • Who do you regard as \'hard to reach\'? • Why is it difficult to reach these groups? Why do you think they don\'t come? • What measures have been put in place to engage the \'hard to reach\'? • Examples of successes in engaging the \'hard to reach\'. Ethical approval was sought and obtained from the university ethics committee. IRAS approval was not required as no NHS patients or employees were interviewed. All interviews took place at the participants\' place of work. All Participants agreed for the interviews to be audio-recorded and signed a consent form to reflect this. Recorded interviews were transcribed by SF and each transcript was checked for any errors or omissions. Each manuscript was read and the emerging themes and key points were highlighted. Transcripts were analysed by both authors to ensure reliability of interpretation of the results. Labels and categories are used to \'organise and analyse the data\' \[[@B32]\] by assigning key themes and emerging concepts to the raw data. This thematic approach to analysis provides a focus on the identifiable themes and patterns that emerge from the raw data. Three broad areas emerged which include the following ◦ How the term \'hard to reach\' is defined ◦ What barriers the \'hard to reach\' face in accessing services. ◦ What can be done to engage \'the hard to reach\' Each of these broad themes comprised a number of categories which are discussed below. Quotations have been used throughout to add power to participants\' accounts, and to encapsulate the views and experiences of individual participants \[[@B39]\]. Results ======= How the term \'hard to reach\' is defined ----------------------------------------- Two respondents alluded to the apparent familiarity of the term \'hard to reach\', for example, \"one of those terms that people bandy about\" while also indicating their unease with it. As one respondent explained (R5): \"The problem I have with the term is that I don\'t really feel as if groups are actually hard to reach.\" before going on to suggest that it would be better to \"find some other terminology to explain \[it\].\" The term was attributed to describe the service (non) user and attributed to \'services\' themselves. The unengaged - non users of services ------------------------------------- There was an overall sense here that one conception of the \'hard to reach\' refers to the people who are not seen in services. The \"unengaged\" was used to by one respondent to refer to those \"who don\'t come to see us\" (R1). Two other respondents described them as; \"groups and communities who aren\'t accessing a particular service for whatever reason\" (R5) and those who \"traditionally wouldn\'t seek support from the usual avenues\" (R6). It is suggested that there are certain characteristics pertaining to how we may commonly define the notion of the \'hard to reach\': People that would generally have difficulty accessing services\... younger and more vulnerable people\... you are talking about people who are more deprived, who are less well educated\... may have restrictions placed upon them by society generally or their more narrow social group. (R1) This respondent further intimated that this does not tell the whole story and acknowledged in the context of providing a service for young people that the \"white middle classes\" may lead \"controlled lives\" that may inhibit them accessing particular services. This alternative explanation of the meaning of the term; the sense that the term \'hard to reach\' can refer to those inhabiting a more privileged section of society \[[@B40],[@B41]\] is an issue that is not raised in the remaining interviews but is worthy of note. More commonly we found that the responses support the sense that the \'hard to reach\' are in some ways more needy; more marginalized; more likely to have experienced poverty, whether that be economic poverty or poverty of opportunity \[[@B7],[@B19],[@B9]\]. Lack of choice is another feature that was used to explain why a group maybe \'hard to reach\' - You don\'t have those choices\... either you can\'t afford them because you\'ve got limited income and you haven\'t got the information either to be able to make an informed choice(R7) The transgender community is also particularly stigmatised \[[@B42]\]. The Diagnostic and Statistical Manual of Mental Disorders (DSM) includes \'Gender Identity Disorder\' as a category and this \'continues to raise questions of consistency, validity, and fairness\' \[[@B43]\]. To help overcome this, the project employs a transgender doctor to help engage with members of the transgender community. \'Hard to reach\' understood from a service and societal stance --------------------------------------------------------------- It is difficult to define the \'hard to reach\' without thinking of the wider societal issues that contribute to or create such conditions. Service restrictions and limitations may mean that it is the services themselves that are \'hard to reach.\' One respondent alluded to the sense that it may be the way that an organisation is configured that creates this notion of the \'hard to reach\' and she stated that the term can mean \"people to whom we put up some barriers\" - it makes it sound like the fault of the non service-user - \"you are hard to reach\" - like they are sat up on a shelf and we have got to lure them down with biscuits, or something\.... actually if you just got a ladder and sat next to them that would be fine. (R1) Other ways of describing this were given in terms of \'hard to reach\' individuals having \"fallen through the net\" (R7) or being \"left out of the loop.\" This suggests a sense that certain groups or individuals are being excluded from services and this is the fault of the services themselves - organisations in the mainstream don\'t know \... how to reach them, how to engage them in activities, don\'t know where to find certain people. (R2) These people have in some way been let down by the system: \"for whatever circumstances \[they\] have been through the system and have either failed the system of the system has failed them.\" (R3) One respondent made explicit the sense that the term is inappropriately attributed to groups and communities \[[@B18]\]. Instead, he stated that \"it is about looking at our own services and looking at why individuals aren\'t engaging. R6 suggested his organisation aims to engage the hardest to reach people who traditionally wouldn\'t seek support from the usual avenues and traditionally hadn\'t succeeded very well in either society or within the education system. Two key points are raised here; firstly of somebody being on the margins of what is considered to be the mainstream of society and secondly, a sense that the reason for this maybe due to a failure in the system. This sentiment is developed and made more explicit later when the respondent intimates that the problem may be systemic - It\'s just about pushing them in that direction and perhaps that\'s where our school education doesn\'t quite serve everybody as well as it should. This respondent also stated that \"we are very good at stereotyping and boxing people in\" and he uses a quotation from a service user to demonstrate this notion: I am not \"hard to reach\", generally people don\'t know how to reach me. Barriers and Facilitators to Services ------------------------------------- In order to understand how appropriate services can be offered to the \'hard to reach\', the experiences of VCS services providers in providing services to these groups were explored. Service providers were asked to describe the kinds of barriers they face in offering services to \'hard to reach\' clients, as well as the factors that helped to facilitate the engagement of these groups. Barriers -------- ### Previous experiences of accessing services A number of participants made explicit reference to the fact that many potential service users may not engage because of their previous experiences of accessing services. In particular, it would seem that statutory services were conceived as being particularly impenetrable \[[@B44]\], thus discouraging individuals to access help. The sense that such experiences may begin in the formative years is indicated by one respondent who suggested that the \"experience of the education system as being quite negative\" (R6) is likely to have an impact upon future expectations. ### Physical Factors There are logistical factors to be considered that may serve as barriers to accessing services, including factors such as location and opening times. One respondent indicated that \"geography and transport\" were barriers to services, but also the fact that young people may be conscious of gang boundaries: quite a lot of young people aren\'t sure if it is ok for them to go to certain places. (R1) A similar point was made by R7 who explained that their particular project moved location from the middle of an estate because: a lot of young people wouldn\'t use it, because it is all territorial. Another respondent (R7) suggested that geography may be a barrier to accessing services as potential service users \"don\'t know we are here.\" However, once people become aware of the location and are engaged in the service, limitations can arise in terms of how many people can be accommodated. By becoming victims of their own success, organisations may inadvertently exclude some more needy but less vocal members of their target audience. This idea is encapsulated by this quotation from R1; the worry is that when you are very busy\... it\'s the more vulnerable that you squeeze out. ### Funding and Partnership working A common issue raised by a number of respondents concerns the way that VCS organisations are funded \[[@B45]\]. These broad areas of concern were based around the fact the organisations have to vie for different pots of money to provide specific services for the particular remit of client group. A similar issue was raised by one respondent who suggested that there may be a fear of competition amongst partner organisations culminating in the concern that - they are going to steal my clients. What service am I going to be left to run?(R4) There was also a concern that excessive attention is given to those things that are easier to measure at the expense of those things that were harder to quantify. Furthermore, some \'causes\' are perceived as more worthy resulting in larger funding pools for certain projects. One respondent describes the inequity in accessing project funding for lesbians: If you work in BME communities, or people with HIV, it can be a lot easier to get funding.(R4) Despite a good deal of praise for the practice of partnership working (see below) one respondent drew attention to its potentially problematic nature, stating that co-working or partnership working may be perceived as synonymous with \"extra work\" and therefore maybe avoided. ### Attitudes and constraints of the statutory services towards the VCS Three of the respondents were explicit in their concerns around the relationship between themselves and statutory sector organisations. One expressed concern about the attitude held by the statutory sector towards the VCS. He suggested that the use of the term \'third sector\' with reference to the VCS reinforces a view that this sector is somehow of less value: the image the statutory sector has of the voluntary sector is \'there, there nice little organisations. You do your specialist thing and leave us to do the professional work. (R3) One respondent made explicit their frustrations with the statutory sector suggesting that social services are not equipped to deal with the complexity of problems faced by some young people: if that young person doesn\'t co-operate, for whatever reason, then they\'re \[social services\] not interested and the young person is left high and dry. (R7) Problems of access are also faced by the VCS themselves: as an agency, it\'s also difficult to access these services on behalf of young people, especially when the average user will have more than one, like sort of, special need. (R7) ### Expectations/limitations of services Expectations or preconceptions of the VCS services may also become a barrier in terms of accessing \'hard to reach\' groups. This is encapsulated by one respondent, who states: it is about ourselves, the expectations that you must come to us.(R5) It is suggested that for a service that is dominated by clients from a more narrow demographic this can become a barrier for potential service users who may feel like \"outsiders\" (R3). Even within services for the \'hard to reach\' there is a hard core of individuals who do not access services. In some homeless day centres, for example, where the client group are predominantly white men, then women and people from other ethnic groups may feel more reluctant to utilise the service. Furthermore, the views held by some clients about other service users can contribute to the difficulties faced by service providers in serving diverse groups: we see individuals coming in with a very blinkered view about some of the service users that we see, maybe refugees - \"why are they over here stealing our jobs?(R4) Facilitators ------------ As described, there are a number of perceived barriers in terms of providing services to the \'hard to reach.\' Conversely, a number of facilitators in providing accessible services to the difficult to engage were cited. ### Treatment of Clients - trust, respect A perception that service users have been poorly treated by other services, and in particular by statutory service has already been raised. The most commonly cited facilitator for engaging clients was the way VCS services communicate with and treat clients. This key theme is illustrated by the following quotations from R1: we don\'t interrogate clients when they come here. We ask what questions we need to give them the service they want. (R1) Reference is also made to the attention paid to the attitude of the staff members: I think the important thing is to spend an awful lot of time and effort on making sure that staff are genuinely welcoming and non-judgmental about absolutely everyone that walks through the door. (R1) The quality of the relationship that develops between staff and service user is a major factor in engaging clients; building trust and respect; being non-judgmental and being able to relate to and empower people. Ok, you can trust us, but it is very much about them being able to recognise that they can trust us, or build up that relationship which can take weeks. (R3) ### Flexibility Offering flexible services that respond to the needs of the service users including running outreach services, listening to feedback, offering flexible opening hours and providing service users with the kinds of services they want were all key facilitators cited by the respondents. One respondent made the point that although they are not as \"flexible as we would like,\" many of the people using this service are very much on the peripheries of normal society\... a lot of people on benefits \...so they can come as and when. (R4) Furthermore, by offering an outreach service to encourage potential service users to engage, they are able to access people who may not ordinarily feel comfortable to attending a HIV service. Not all the organisations interviewed conducted outreach and this was largely due to resource limitations. There appeared to be a tension between offering a service to all the target audience, including those \'harder to reach\' clients and being able to provide an adequate service to all. Offering services that people want is a key facilitator for engagement. One respondent described how changing their opening times as well as offering service users a hot meal had resulted in a fifty per cent increase in demand for the service (R3). ### Partnership Working As described above there are a number of drawbacks to partnership working that can be perceived as barriers to service participation. However, there are also a number of positive aspects to working in partnership and it seemed to be a priority for a number of the organisations. One respondent makes the point that working in partnership is actively encouraged by the funding bodies who support organisations. Partnerships can be useful in assisting in: getting information out there, and ensuring that people know who we are. (R4) Acknowledgement of the limitations of a service was also important: we don\'t claim to know everything. We don\'t want to do everything. It isn\'t possible to do everything really, really well, but what you can do is build very, very strong partnerships with other organisations that do specialise in particular areas. (R5) ### User involvement For some of the respondents, emphasis was placed upon the importance of user participation. Service participation and interview panels were conducted at one organisation throughout the year. \"Member helpers\" or trainee volunteers form part of the team with the aim to build the esteem and skills of the service users. Using service users as part of the team encourages other clients to participate and feel more settled: It\'s about having a different relationship. A service user can better describe it \[how it feels to be a service user\]. They can share that with each other, use that as a bridging gap and a way of opening doors. (R3) The kinds of relationships the VCS is able to develop with their clients is of the kind that is often found wanting in statutory services. The emphasis they place on user involvement and use of volunteers as demonstrated above can act as an enabler for those they seek to serve. Discussion ========== The VCS a model for the NHS? ---------------------------- As we have seen, tackling health inequalities is a top priority for the Government and part of this includes engaging those \'seldom seen, seldom heard\' in services. Recent Department of Health initiatives have highlighted the need for improved service provision and configuration. The NHS Next Stage Review Interim Report \[[@B46]\] reported that despite \'sustained investment and improvement in the NHS over the past 10 years, access to primary medical care and services and quality of those services continues to vary significantly across the country.\' (pp 3) As noted by a number of the respondents, the importance of listening to the voices of users has been taken onboard by the NHS as a key resource in providing appropriate services \[[@B47]\]. Already, changes in how health services, and in particular primary care services, are delivered seem to bear the hallmarks of VCS practice, but there is a need to go further. Some of the changes in provision of GP services aimed at improving patient access and experience include the introduction of patient feedback and quality assessment (surveying patients about their local GP practice); the extension of opening hours to accommodate the varying lifestyles of patient populations and thus engaging with the \'seldom seen\'. As we have seen the VCS ability to offer flexibility in terms of opening hours and responsive services for service users are cited as facilitators for the more difficult to engage. The Department of Health acknowledges the need to \'establish effective links with frontline services, utilising the potential of VCF sector agencies as valuable catalysts for dialogue, mutual understanding and empowerment \[[@B48]\] (pp4). Partnership working is becoming increasingly important to the health service. From the perspective of primary care provision it is acknowledged that as commissioners of services, Primary Care Trusts (PCTs) working alongside local authorities should \'identify specific communities with particularly poor health, such as travellers, migrant workers, people with learning disabilities, those living in disadvantaged areas or demographic groups\' \[[@B49]\]. It seems that the NHS is moving in the right direction in terms of offering choice and a voice to patients. Before equity in health services can be fully realised that work must continue in engaging the most excluded and \'hard to reach\' in society. Study Limitations ----------------- The study is clearly limited by the small sample size. Most qualitative studies will aim to interview between twenty and thirty people although the aim is to continue data collection until data saturation is met. This is the point were no further themes are being generated. Unfortunately, only eight out of the thirty organisations felt they had the time to participate. It could be postulated that organisations providing services for more marginal groups are resource and time limited and therefore find it more difficult to put time aside, so the views of practitioners from these organisations may have been lost. The participant organisations operate services in a large, diversely populated urban area, and the kind of experiences and views found here are likely to differ from those from organisations operating in rural communities, with a widely dispersed population. There are also likely to regional variations that are not reflected in this study. However as a pilot study, one of the objectives is to provide background to inform a larger scale study proposal which aims to involve the views of a larger number of participants including service users and primary care practitioners. Conclusion ========== The VCS have a long history of working with marginalised, disadvantaged and \'hard to reach\' groups and as discussed their experiences may provide potential lessons for statutory service providers, particularly health services. The VCS is able to provide added value in terms of the resources they offer; the procedures they employ and how they are organised \[[@B30]\]. From the interviews conducted there are four main areas of importance that have arisen that relate to how best to engage \'hard to reach groups\': attitude of staff; service flexibility; working in partnership with other organisations and empowering users involvement. In order to fully engage with the \'hard to reach\' and provide an equitable health service for all, the NHS must embrace some of the philosophies that appear to underpin the VCS. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= SF conducted the interviews and drafted the original manuscript. BH participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1472-6963/10/92/prepub> Acknowledgements ================ This study was funded by The University of Birmingham Research Support Facility 2 year fellowship (May 2006 - September 2008) awarded to SF. The RSF is an academic support unit based in the Department of Primary Care and General Practice. It is funded by the National Co-ordinating Centre for Research Capacity Development (NCCRCD) to contribute to research capacity building by supporting individuals with the potential to develop a career in primary care research.	{ "pile_set_name": "PubMed Central" }
Introduction {#s1} ============ Adaptive phenotypic plasticity across a range of biological traits, from individual cell form and function to advanced learning, enables individuals to cope with fluctuating environments ([@bib64]; [@bib59]). This flexibility is well-recognised in the diploid life stage, where a complex multicellular organism can generate adaptive plasticity in behaviour, morphology and physiology. However, far less is known about what plastic adaptive responses are possible at the seemingly simple unicellular gamete stage, when environmental variation can be profound. We therefore assessed whether males and females can adaptively vary sperm and egg function through gametogenesis in anticipation of impending functional environment. Although sperm and eggs have universal primary roles that are vital for reproductive success, how they achieve these and the environments in which they must succeed can fluctuate considerably, both biotically and abiotically. Sperm almost always operate after ejaculation and release in a non-self and demanding environment, either within the female reproductive tract or through external fertilisation, and many factors that directly influence sperm function can vary profoundly across these environments ([@bib65]; [@bib71]). Ova face a similarly challenging set of biotic and abiotic variables, especially in the many species that are oviparous where zygotes and embryos develop in eggs to hatch in the natural environment outside the mother ([@bib43]; [@bib37]). These intrinsically variable environments for gamete function could lead to selection for adaptive plasticity, allowing males and females to improve their reproductive fitness by matching sperm and egg phenotypes through development in anticipation of different fertilisation and embryogenesis environments. One of the most important abiotic environmental variables is temperature ([@bib12]; [@bib4]), especially in the context of climate change when thermal environments are expected to both warm and become much variable and extreme ([@bib62]; [@bib68]). Thermal variation has profound impacts on living systems, and numerous examples of adaptive plastic responses to temperature variability have been described, from acclimated mitochondrial function ([@bib63]) and sex determination ([@bib95]), up to complex shifts in behaviour ([@bib60]) and phenology ([@bib94]). Temperature also influences gamete function across a number of levels, with sperm production and function being especially sensitive to warming ([@bib82]; [@bib75]) and egg development being directly influenced by thermal regime ([@bib38]; [@bib39]). Here, we therefore use a combination of experimental approaches (overview on [Figure 1](#fig1){ref-type="fig"}) to examine whether males and females adaptively vary sperm and egg biology through gametogenesis at different temperatures, in anticipation of varying thermal environments for fertilisation and reproduction. We test for this potential in the flour beetle Tribolium castaneum which, like most sexually reproducing animals, is both ectothermic and oviparous ([@bib84]), and in which spermatozoal sensitivity to temperature is known ([@bib75]). After exposing adult males and females and their gametes to different temperatures, we compare sperm and egg development and reproductive function within thermal regimes that mimic the increasingly variable conditions faced by warmer tropical regions ([@bib62]). We find that thermal regime regulates gamete size development: in the short-term, as the developmental environment warms, eggs get larger and sperm become smaller. Having established these opposing male-versus-female responses, we use reciprocal transplant experiments to measure whether adaptive gamete plasticity exists. We find that warmer reproductive environments present challenges for reproduction. However, gamete plasticity enabled significant improvements in reproductive performance. Our findings reveal a potentially important route by which ectothermic populations can buffer their reproductive output against increasingly stressful and unpredictable temperature fluctuations under climate change. ![Experimental design.\ Overview of experiments investigating thermal adaptation and plasticity in sperm and egg biology and its adaptive significance. (a) Set up of long-term thermal selection lines (TSLs) maintained at 30°C or 38°C for 50+ non-overlapping generations with eight replicate populations per regime. (b) Experimental design to investigate gamete size divergence in 30°C and 38°C TSLs at their long-term evolving environments, short-term plasticity in gamete size measured after a single-generation of novel temperature exposures, and short-term plasticity in individuals from the ancestral KSS (Krakow Super Strain) population. Symbol fill colour represents long-term background (30°C TSLs in blue, 38°C TSLs in red and KSS in yellow) while outline colour represents short-term exposure temperature (30°C blue, 38°C red). bi) and bii) illustrate experiments on temporal patterns of short-term plasticity in egg size (see main text). (c) Experiments investigating the adaptive significance of sperm and egg morphological plasticity in KSS adults. Gametic divergence was achieved by having adults produce gametes at either 30°C or 38°C, whose performance was then tested at 30°C, 38°C or 40°C.](elife-49452-fig1){#fig1} Results {#s2} ======= Sperm morphological divergence and plasticity {#s2-1} --------------------------------------------- Following 54 generations of experimental evolution under increased temperature, we found that at both development temperatures sperm length differed by an average of \~4% between long-term selection regimes, with males from lines evolved at 38°C producing significantly longer sperm than males from lines evolved at 30°C (effect size β = 3.4, t~1,14~ = 3.4, p=0.004; eight lines per regime, five males per line, and five sperm per male, [Figure 2](#fig2){ref-type="fig"} and [Table 1](#table1){ref-type="table"}). In contrast, a within-generation increase in temperature during pupation and post-eclosion maturation reduced sperm length development irrespective of evolutionary background, with sperm produced at 38°C being \~7% shorter, and indicating developmental plasticity in sperm size (β = −6.3, t~1,24~ = -11.2, p\<0.001; [Figure 2](#fig2){ref-type="fig"}). The interaction between selection regime and the development environment was not significant (β = 0.5, t~1,24~ = 0.4, p=0.672). Male body sizes of 30°C and 38°C thermal selection lines (TSLs) did not differ (F~1,78~ = 0.38, p=0.53), and similarly we found no evidence for body size divergence between 30°C or 38°C thermal environments from the pupal stage in the ancestral stock population (the Krakow Super Strain, KSS) (F~1,46~ = 0.90, p=0.35). ![Long-term divergence and short-term plasticity in sperm size.\ Symbol fill colour represents long-term background (30°C TSLs in blue, 38°C TSLs in red and KSS in yellow) while outline colour represents short-term exposure temperature (30°C blue, 38°C red; [Figure 1b](#fig1){ref-type="fig"}). (a) Sperm length in µm ± SEM of the experimentally evolved TSLs, measured either following production from pupation in their long-term thermal environment (highlighted in shaded boxes) or at the reciprocal temperature. (b) Sperm length of mature KSS males exposed to either 30°C or 38°C from pupation through eclosion.](elife-49452-fig2){#fig2} 10.7554/eLife.49452.004 ###### Sperm and egg length in relation to long-term selection and short-term exposure. Line Temperature Sperm length N Egg length N ---------- ------------- -------------- ---- -------------- ----- 30°C TSL 30°C 88.4 ± 5.2 40 615.4 ± 51.5 240 30°C TSL 38°C 81.9 ± 7.9 40 685.5 ± 42.2 240 38°C TSL 30°C 91.5 ± 4.9 40 638.4 ± 48.3 240 38°C TSL 38°C 85.5 ± 4.3 40 682.7 ± 42.7 240 KSS 30°C 89.3 ± 4.1 26 662.6 ± 39.4 180 KSS 38°C 84.3 ± 4.8 26 697.9 ± 48.4 180 Shown are mean, standard deviation and sample size (sperm: number of males; eggs: number of eggs) for sperm length and egg length measured in individuals from temperature selection lines (TSL) and the ancestral population (KSS), exposed to different temperatures from pupation onward (see main text and [Figures 1](#fig1){ref-type="fig"}--[3](#fig3){ref-type="fig"}). Egg morphological divergence and plasticity {#s2-2} ------------------------------------------- Egg size also showed divergence and plasticity in relation to thermal regime. In contrast to sperm, egg size showed an increase in response to a hotter proximate temperature (β = 70.0, t~1,14~ = 13.9, p\<0.001; [Figure 3a](#fig3){ref-type="fig"} and [Table 1](#table1){ref-type="table"}). However, there was a significant interaction between long-term selection regime and short-term temperature exposure (β = −25.7, t~1,14~ = -3.6, p=0.003; [Figure 3a](#fig3){ref-type="fig"}). Following 58 generations of experimental evolution (logistic contraints prevented simultaneous measurement at generation 54 when sperm lengths were assayed), eggs of females from TSLs evolved at 38°C were larger than those of 30°C TSL females when produced at 30°C (β = 23.0, t~1,14~ = 3.2, p=0.006), but were very similar in size when produced at 38°C (β = −2.8, t~1,14~ = -0.4, p=0.677). ![Long-term divergence and short-term plasticity in egg size.\ Symbol fill colour represents long-term background (30°C TSLs in blue, 38°C TSLs in red and KSS in yellow) while outline colour represents short-term exposure temperature (30°C blue, 38°C red; [Figure 1b](#fig1){ref-type="fig"}). (a) Egg length in µm ± SEM of the experimentally evolved TSLs, measured either following production in their long-term thermal environment (highlighted in shaded boxes), or at the opposite temperature. (b) Egg length produced by mature KSS females exposed to either 30°C or 38°C for mating and oviposition, measured across three experimental replicates (grey lines). (c) Egg length of KSS females mated to a standard KSS male ovipositing individually either at 30°C or 38°C for 10 days in two-day intervals. The two bold-face lines indicate the averages across all females within a given treatment, while thin lines show average values for individual females. (d) Egg length of groups of 50 females, mated to standard KSS males, and ovipositing alternately at 30°C and 38°C for 12 days in two-day intervals. Egg width did not differ between thermal regimes, demonstrating that oviposition temperature increased egg volumes ([Figure 3---figure supplements 1](#fig3s1){ref-type="fig"} and [2](#fig3s2){ref-type="fig"}).](elife-49452-fig3){#fig3} Egg size was also thermally plastic in standard ancestral stock Krakow Super Strain (KSS) females, and showed significant divergence according to adult rearing and oviposition temperature when assessed in three experimental repeats. KSS females produced larger eggs when ovipositing at 38°C compared to 30°C (β = 35.3, t~1,356~ = 8.2, p\<0.001; 50 females per group and 60 eggs measured per group at either rearing temperature, [Figure 3b](#fig3){ref-type="fig"}). In addition, this plasticity was shown by individual females ovipositing alone at either 30°C or 38°C (β = 81.5, t~1,42~ = 11.6, p\<0.001; [Figure 3c](#fig3){ref-type="fig"}), and was reversible when females were alternated between 30°C and 38°C thermal environments (β = 58.9, t~1,4~ = 7.3, p=0.002; [Figure 3d](#fig3){ref-type="fig"}). Egg width was not different between the oviposition temperatures (F~1,198~ = 0.1, p=0.686, [Figure 3---figure supplements 1](#fig3s1){ref-type="fig"} and [2](#fig3s2){ref-type="fig"}) and the interaction between oviposition temperature and egg length was not statistically significant (t = −0.77, p=0.44), demonstrating that oviposition temperature increased egg volumes (see [Figure 3---figure supplement 2](#fig3s2){ref-type="fig"} for volume calculations). Adaptive sperm plasticity {#s2-3} ------------------------- As for the selection lines, sperm size was thermally plastic in ancestral stock KSS males, and showed significant divergence according to rearing temperatures (two experimental repeats), with KSS males producing significantly shorter sperm when reared at 38°C compared to 30°C (β = −5.1, t~1,49~ = -7.0, p\<0.001; [Figure 2b](#fig2){ref-type="fig"} and [Table 1](#table1){ref-type="table"}). To test for the adaptive value of any functional plasticity, the performance of sperm from males of the same ancestral KSS population reared from pupation at either 30°C or 38°C was tracked by comparing the total number of offspring sired across 100 days of oviposition by KSS females at either 30°C or 38°C ([Figure 4a](#fig4){ref-type="fig"} and [Table 2](#table2){ref-type="table"}), by which time all females in the experiment had exhausted their viable sperm stores ([@bib58]) and ceased to produce offspring ([Figure 4b and c](#fig4){ref-type="fig"}). Reproduction was generally sensitive to the proximal thermal environment, with on average 299 (± 20.8, mean ± s.e.m.) offspring eclosing at 30°C, compared with 135 (± 10.3, mean ± s.e.m.) at 38°C (β = −5.8, F~1,104~ = 69.8, p\<0.001). Despite these overall differences, it was clear that the thermal regime in which sperm production and insemination took place also gave individual males a reproductive advantage when matched to the same thermal environment for fertilisation and offspring development ([Figure 4a](#fig4){ref-type="fig"}). Indeed, the interaction between spermatogenesis temperature (sperm production) and offspring production temperature was highly significant (β = 7.5, F~1,104~ = 29.6, p\<0.001), while the main effect of male treatment was not significant (β = 0.7, F~1,104~ = 1.1, p=0.290). Sperm from males exposed as pupae and maturing adults to 30°C achieved greater reproductive success in the 30°C reproductive environment than sperm from males that completed development at 38°C. By contrast, in the 38°C reproductive environment sperm produced by males in the 38°C treatment achieved double the reproductive success compared with sperm from males developed through the 30°C treatment. Across 100 days of oviposition in the 30°C reproductive environment, males reared at 30°C sired an average of 349 (± 29.1, mean ± s.e.m.) offspring, while males reared at 38°C sired 249 (± 27.1, mean ± s.e.m.) offspring. Using the same protocols in the 38°C reproductive environment, males reared at 38°C sired 180 (± 12, mean ± s.e.m. offspring), while males reared at 30°C sired only 91 (± 12, mean ± s.e.m.) offspring ([Figure 4a](#fig4){ref-type="fig"} and [Table 2](#table2){ref-type="table"}). Exploration of the temporal patterns of these effects showed a clear decline in reproduction across consecutive 10-day blocks, and that sperm matched to the fertilisation and development temperature consistently outperformed sperm that were thermally mismatched ([Figure 4b and c](#fig4){ref-type="fig"}). Our model comparison based on AIC values confirmed that the interaction between male rearing temperature and fertilisation temperature was important: the best model included the main effects of male temperature, offspring temperature and time, and the interaction between male and offspring temperature for both parts of the model, and additionally the interaction between rearing temperature and time for the zero-inflation model ([Supplementary files 1](#supp1){ref-type="supplementary-material"} and [2](#supp2){ref-type="supplementary-material"}). ![Adaptive thermal plasticity in sperm.\ Reproductive output of ancestral KSS males following a 24 hr mating bout with a single female. Symbol outline colour represents male short-term exposure temperature (30°C blue, 38°C red) while background colour indicates fertilisation and offspring development temperature (30°C blue, 38°C red; see [Figure 1c](#fig1){ref-type="fig"}). (a) Total offspring produced over a 100d period across ten 10 day blocks from sperm produced in either 30°C or 38°C conditions when functioning in either 30°C or 38°C reproductive environments. Temporal patterns in 30°C and 38°C environments are illustrated in (b and c), respectively (note the log-scale of the Y-axis). Analyses of individual male reproductive performance and average sperm length across a range of thermal regimes indicate a longer-sperm advantage in this system ([Figure 4---figure supplement 1](#fig4s1){ref-type="fig"}).](elife-49452-fig4){#fig4} 10.7554/eLife.49452.010 ###### Adaptive thermal plasticity in sperm and eggs improves gamete performance. Sperm Eggs ----- ------ ------ --------------- ------ --------- --- KSS 30°C 30°C 349.5 ± 151.5 27 90 ± 2% 8 KSS 38°C 30°C 248.8 ± 140.9 27 80 ± 5% 8 KSS 30°C 38°C 91.0 ± 61.8 27 78 ± 8% 8 KSS 38°C 38°C 179.6 ± 61.8 27 80 ± 3% 8 KSS 30°C 40°C 27 ± 5% 8 KSS 38°C 40°C 36 ± 5% 8 Sperm performance was measured by mating focal males to tester females and counting all offspring produced over a 100d period. Egg performance was measured as the proportion of eggs that developed into adult offspring, tested in groups of 50 eggs (see main text and [Figures 3](#fig3){ref-type="fig"} and [4](#fig4){ref-type="fig"}). Adaptive egg plasticity {#s2-4} ----------------------- The performance of eggs produced by ancestral stock KSS females (mated to 30°C-reared KSS males) ovipositing at either 30°C or 38°C was compared by measuring egg-to-adult offspring viability of pre-counted groups of 50 eggs when incubated and reared at either 30°C, 38°C or 40°C ([Table 2](#table2){ref-type="table"}). There was a significant interaction between oviposition thermal regime and the environmental temperature treatments at which the eggs were incubated, and offspring hatched and developed (z = −4.1, p\<0.001; [Figure 5](#fig5){ref-type="fig"}). Thus, there was similar evidence of adaptive egg plasticity as for sperm, although this was only evident in the 30°C and 40°C environmental treatments ([Figure 5](#fig5){ref-type="fig"} and [Table 2](#table2){ref-type="table"}). Post hoc testing showed that at 30°C eggs oviposited at 30°C achieved significantly greater egg-to-adult offspring viability rates than eggs oviposited at 38°C (z = 4.6, p\<0.0001), whereas the opposite was true at 40°C (z = 3.01, p=0.01) where 30°C-oviposited eggs resulted in 25% fewer offspring than 38°C-oviposited eggs. In the 38°C environment, there was no evidence for any adaptive plasticity in egg biology, with the same relative number of eggs from the 30°C and 38°C regime females hatching and producing adult offspring ([Figure 5](#fig5){ref-type="fig"} and [Table 2](#table2){ref-type="table"}). ![Adaptive thermal plasticity in eggs.\ Symbol fill colour represents production and oviposition temperature (30°C blue, 38°C red) for ancestral KSS females while outline and background colour indicate egg incubation, offspring hatching and development temperature (30°C blue, 38°C red, 40°C dark red; see [Figure 1c](#fig1){ref-type="fig"}). Egg-to-adult viability was measured in a fodder medium with 0% yeast, with eight replicate groups of 50 eggs per treatment combination. Point surface area is proportional to the number of observations with identical outcomes.](elife-49452-fig5){#fig5} Discussion {#s3} ========== Our experiments revealed that: (1) gamete function and reproduction is highly sensitive to the local thermal environment; (2) developmental plasticity exists within sperm and eggs in response to temperature; (3) plastic responses in gamete size proceed in different directions in either sex; and (4) gamete plasticity is adaptive, enabling males and females to significantly improve their reproductive success via mechanisms that match sperm and egg development to the imminent thermal environment in which they must function. Sperm production and function is known to be affected by many environmental variables ([@bib65]; [@bib71]), with particular sensitivities to temperature ([@bib82]; [@bib17]), which may be especially important in ectotherms where environmental temperature varies through the reproductive window ([@bib71]; [@bib93]). Such variation will directly influence important elements within the sperm storage and fertilisation environment ([@bib65]; [@bib71]), including biophysical properties of fluids and membranes ([@bib53]), mitochondrial metabolic sensitivity ([@bib85]), flagellar function ([@bib48]) and haplotype integrity ([@bib61]). Our transplant experiments reveal obvious challenges for gamete performance in warmer environments, with an overall halving of reproductive output when sperm and eggs are challenged to function at 38°C versus 30°C ([Figures 4](#fig4){ref-type="fig"} and [5](#fig5){ref-type="fig"}). Interestingly, between-line variance in sperm length also increased when males were forced to develop sperm at 38°C ([Figure 2a and b](#fig2){ref-type="fig"}), possibly indicating a response to thermal stress at the higher temperature. Although we found evidence that laboratory selection across 54 generations at 38°C caused the overall evolution of \~4% longer sperm compared with the thermal line selection at 30°C, our within-generation experiments revealed consistent \~7% reductions in sperm length for all populations when produced through the warmer 38°C regime compared with 30°C ([Figure 2](#fig2){ref-type="fig"}). The different sperm length responses to experimental evolution versus within-generation plasticity under different temperatures is puzzling, and may have arisen through indirect selection acting on the thermal lines for \~five years, given that the 38°C thermal selection lines showed generally longer sperm at both 30°C and 38°C rearing temperatures ([Figure 2a](#fig2){ref-type="fig"}). Recent work examining sperm length evolution in T. castaneum revealed that a history of heightened sexual selection led to the evolution of longer sperm ([@bib41]), so one possibility is that increased metabolic and developmental rates in the 38°C regime had elevated mating activity, and therefore promoted male-male competition and selection for increased sperm length. However, our current results ([Figure 4](#fig4){ref-type="fig"}) and past findings ([@bib75]) indicate that warmer conditions generally reduce male reproductive fitness, and we find no evidence that operational sex ratios have deviated between our 30°C or 38°C selection line regimes, with both showing an average adult sex ratio of 50% (54 adults randomly sampled from six lines per regime: 30°C regime male ratios = 0.52, 0.46, 0.50, 0.52, 0.48, 0.52; 38°C regime male ratios = 0.50, 0.59, 0.48, 0.48, 0.52, 0.43); offspring sex ratios are also not changed following male exposure to 42°C heatwaves ([@bib76]). Another possible explanation is that increased developmental rate at 38°C has hastened sperm ageing, leading to correlated changes in length. However, T. castaneum is a relatively long-lived insect, with adult males showing no change in fertility even after one year of lifespan ([@bib40]), so reproductive ageing differences over the few days of the sperm length assays conducted here are unlikely to explain the overall sperm length increases in the thermal selection lines at 38°C. Another developmental possibility is that elevated spermatogenic rates at 38°C leads to the production of smaller cells, and longer-term selection has compensated for this reduction in sperm cell size within warmer regimes to evolve longer sperm. There is some evidence for longer-sperm advantage in T. castaneum ([@bib41]), and our results for individual male fertilisation success across a range of thermal regimes suggest that males producing longer sperm have improved reproductive fitness ([Figure 4---figure supplement 1](#fig4s1){ref-type="fig"}). Additional possibilities for indirect effects acting on sperm length in the 38°C selection regime include genetic bottlenecking at the start of selection if fertility and offspring production had been compromised, or heightened metabolic rate throughout the entire life cycle allowing improved access to key nutrients which may enable the development of longer sperm ([@bib41]). Whatever the cause behind long-term experimental evolution of sperm length within our thermal lines, it is clear that short-term within-generation impacts of temperature have strong and direct effects on sperm development, with an experimental switch to warmer regimes for either the thermal selection lines or the ancestral stock population resulting in consistently clear reductions in sperm length ([Figure 2a and b](#fig2){ref-type="fig"}). Temperature variation during gametogenesis is known to affect sperm size in ectotherms ([@bib7]; [@bib72]), and cool and warm thermal extremes can reduce fertile function in tandem with reduced sperm length development ([@bib90]). However, no previous study has revealed that the thermal regime during spermatogenesis can shape sperm function to be optimal for the forthcoming thermal reproductive environment. We demonstrate clear evidence for adaptive plasticity in sperm function in anticipation of thermal regime, enabling males (and their mates) to increase their reproductive success by 40% to 100% when males are developmentally 'matched' to the temperature of the subsequent reproductive environment ([Figure 4](#fig4){ref-type="fig"}). Adaptive plasticity in the production of sperm numbers is known in relation to environmental risks of male-male competition ([@bib96]), with the capability for spermatogenesis to increase in response to elevated risks of sperm competition ([@bib70]; [@bib36]). This male plasticity can also change individual sperm cell form and function: domestic cockerels (Gallus gallus domesticus) rapidly changed their sperm mobility in relation to their own competitive status ([@bib66]), and male Gouldian finches (Erythrura gouldiae) adjusted sperm morphometry in relation to social factors ([@bib49]). Similar changes occured in sperm velocity and density in Arctic charr (Salvelinus alpinus) when dominance switched ([@bib74]), and broadcast-spawning ascidians (Styela plicata) produced larger, more motile sperm when adults were kept at densities with greater risks of sperm competition ([@bib15]). Far fewer studies have explored adaptive sperm plasticity in relation to abiotic variation, which is a gap because physico-chemical factors can greatly influence sperm function and also vary across reproductive environments ([@bib71]). Acclimation and in vitro fertilisation experiments with the broadcast spawning tubeworm Hydroides diramphus revealed adaptive plasticity in sperm and egg function in relation to salinity, with gametes performing best at salinities experienced by their parents prior to spawning ([@bib51]). Likewise, sticklebacks (Gasterosteus aculeatus) showed adaptive plasticity in spermatozoal sensitivity to the salinity and osmolarity signal for initiating flagellar motility ([@bib88]). To our knowledge, only two studies have investigated sperm plasticity in relation to temperature: male mosquitofish (Gambusia holbrooki) acclimated to cool (18°C) and warm (30°C) regimes for five weeks showed no signs of spermatozoal acclimation or change in thermal limits ([@bib2]). Likewise, although warm conditions reduced sperm motility in brown trout (Salmo trutta), warm-acclimated males did not produce sperm with improved relative motility ([@bib26]). Our transplant experiments revealed clear evidence for adaptive developmental plasticity in sperm function according to thermal regime, with males exposed to warmer 38°C temperatures producing sperm that enabled a doubling of reproductive success in warmer 38°C fertilisation and reproductive environments compared with sperm produced at 30°C ([Figure 4a and c](#fig4){ref-type="fig"}). The opposite also applied, with sperm produced in cooler 30°C regimes gaining \~40% greater reproductive success at 30°C compared with sperm produced at 38°C ([Figure 4a and b](#fig4){ref-type="fig"}). This plasticity will confer direct fitness advantages if the thermal regime through development before mating anticipates the temperature for sperm function and reproduction following insemination. Sperm manufacture proceeds rapidly in T. castaneum, with production of mature, functional gametes taking around four days at 30°C ([@bib31]). As in many insects, fertilisation and oviposition proceeds within hours of mating in T. castaneum ([@bib25]), so sperm production temperature will usually predict insemination, storage and fertilisation temperature. In contrast to males, female T. castaneum produced larger gametes in the warmer environment ([Figure 3](#fig3){ref-type="fig"}), which is unusual for arthropods where smaller eggs are usually produced as temperatures increase ([@bib33]); but see [@bib81] and [@bib87]. This egg size plasticity occurred within two days of exposure to the novel thermal regime, and was reversible ([Figure 3d](#fig3){ref-type="fig"}). Egg plasticity was also adaptive when comparing performance between the more extreme 30°C versus 40°C environments ([Figure 5](#fig5){ref-type="fig"}): 36% of eggs produced at 38°C generated viable offspring following development at 40°C compared with only 27% of eggs produced at 30°C. Conversely, 90% of eggs produced at 30°C generated offspring when developed at 30°C, versus 80% of eggs developed at 38°C ([Figure 5](#fig5){ref-type="fig"}). Adaptive plasticity in egg biology is known in relation to a number of environmental factors, but we believe this is the first study to demonstrate it under environmental warming. Adaptive egg plasticity in relation to biotic variation is shown by seed beetles (Stator limbatus), where females vary egg size in relation to anticipated host plant quality ([@bib32]). Experiments show that females respond to a switch in host plant in a manner that maximises reproductive fitness by matching egg size (and number) to the host type ([@bib32]; [@bib77]). Likewise, female cowpea weevils (Callosobruchus maculatus) are sensitive to levels of larval competition predicted by increased adult density, laying larger eggs to improve larval fitness ([@bib52]). Broadcast-spawning ascidians (S. plicata) produce smaller eggs in high density populations, but their embryo-yielding ovicells are larger than eggs from low-density adults ([@bib15]). Similarly, female Nasonia vitripennis wasps adjust the sex ratio of their broods depending on whether they are first or second to parasitise a host, improving offspring fitness according to anticipated variation in local mate competition ([@bib97]). Adaptive egg plasticity in relation to some abiotic variables is also recognised. Female stink bugs (Podisus maculiventris) detect reflectance at the site where they oviposit, and invest more protective pigment into eggs that will be exposed to stronger ultraviolet solar radiation ([@bib1]). A number of studies have found variation in egg phenotypes according to temperature during development and oviposition (e.g. [@bib33]; [@bib24]; [@bib6]), but few have identified that the changes are adaptive, with thermal variation creating physiological constraints or stress during egg production ([@bib33]). A notable exception is in Bicyclus anynana butterflies, where females lay larger (and fewer) eggs when they are exposed to lower temperatures during oviposition ([@bib27]); reciprocal transfer experiments, as we employ here, show that this behaviour is adaptive ([@bib28]). We identify adaptive plasticity in response to the upper ranges of reproductive tolerance by T. castaneum females, with adaptive matching through warmer egg development temperatures enabling a 33% improvement in offspring production in the hottest reproductive environment ([Table 2](#table2){ref-type="table"}, [Figure 5](#fig5){ref-type="fig"}). Mechanisms for optimising gamete function in different thermal environments are to be uncovered, but four broad and related possibilities exist through: 1) optimising size, 2) physiological matching, 3) haploid selection, and/or 4) epigenetic modifications for offspring development. Although we observed opposing responses by sperm and egg sizes under temperature variation, it seems unlikely that this phenotypic variation is solely responsible for enabling improvements in reproductive success. Sperm length decreased as temperatures increased, but additional correlative analyses gave no indication that reduced sperm size per se improved reproductive performance in hotter thermal regimes, and the reverse was more evident ([Figure 4---figure supplement 1](#fig4s1){ref-type="fig"}), consistent with previous work showing that sperm elongation is costly and positively selected by competition ([@bib41]). It therefore seems more likely that changes in sperm physiology, rather than morphometry, allow plasticity in sperm function to match thermal reproductive environments. Thermal adaptations influencing cell physiology and biochemistry are known to exist within mitochondrial metabolism ([@bib23]; [@bib42]), essential for sperm function ([@bib69]), and membrane properties influencing cell structure and physiology ([@bib13]) for sperm and flagellum function ([@bib9]). In addition, Heat Shock Proteins, expressed throughout spermatogenesis in the testis, play key roles in sperm development ([@bib22]). These adaptations for different thermal and hydrodynamic environments could be adaptively varied through spermiogenesis so that sperm function is matched to challenges facing sperm migration, storage and fertilisation in warmer environments. Within-ejaculate haplotype selection provides another mechanism for improving offspring performance ([@bib3]). Adaptive thermal plasticity may be possible via changes in egg size, because larger-volume eggs produced at higher temperatures will contain more fluids, and possibly a greater nutrient load, both of which could improve offspring viability under heat stress and desiccation ([@bib29]). However, within the 30°C environment, larger eggs produced at 38°C were outperformed by smaller eggs produced at 30°C, suggesting additional mechanisms beyond size benefits. As for sperm, thermally plastic traits that are essential for egg fertilisation and subsequent function could include differential mitochondrial activity ([@bib21]) and egg plasma membrane properties ([@bib86]). Heat Shock Proteins could also play vital roles in protecting egg development from thermal stress; recent work demonstrates that increased loading of HSP23 into D. melanogaster eggs improves embryo thermal tolerance ([@bib55]). In addition to the potential for adaptive plasticity within gamete function, reproductive fitness could be improved if short-term epigenetic modification through gamete development can pass adaptive information to the zygote, embryo and offspring ([@bib35]). The potential for adaptive transgenerational plasticity via the matriline and through transcription factors within eggs is increasingly recognised ([@bib44]), potentially enabling rapid responses to climate change via adaptive plasticity ([@bib18]). There is growing evidence that sperm also have the potential to be transcriptionally labile ([@bib50]), passing environmentally-driven epigenetic information to offspring through histone or protamine modifications, haplotype DNA methylation remodelling, and/or small RNAs ([@bib20]). Sperm carry complex payloads of coding and non-coding small RNAs which can be transcribed into the oocyte and embryo ([@bib16]; [@bib46]) with conserved functions across mammalian and insect models ([@bib30]); the potential importance of these RNA transcripts remains largely unexplored ([@bib11]; [@bib10]; [@bib83]). However, the possibility for transgenerational information transmission via sperm or egg epigenomes in response to environmental variation during gametogenesis is an obvious mechanism to enable adaptive thermal plasticity for populations facing the challenge of reproducing under climate change where increases in both thermal averages, maxima and variation are expected ([@bib62]; [@bib68]). Materials and methods {#s4} ===================== -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Reagent\ Designation Source or reference Identifiers Additional\ type (species)\ information or resource --------------------------------------------------- ---------------------------------------------- ----------------------------------------------------- ---------------------------------------------- ------------------------------------------------------------------ Strain, strain background (Tribolium castaneum) Krakow\ KSS: [@bib19] [@bib76]\ KSS\ Live beetles Super Strain ancestral stock\ TSL: this paper and [@bib19] [@bib76] &\ and Thermal Selection Lines at 30°C and 38°C TSL30 or TSL38 Biological sample (Tribolium castaneum) Spermatozoa and ova This paper Sperm and eggs Sperm recovered from sacrificed live males, eggs recovered from\ oviposition food medium Software, algorithm R Studio R Studio ([@bib73]) in R ([@bib67]) (version 3.4.1) R Studio version 1.1.463 and R version 3.4.1 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Model system, selection lines and thermal exposure {#s4-1} -------------------------------------------------- Tribolium castaneum flour beetles were maintained as previously described (e.g. [@bib56]; [@bib41]) in a 16 hr: 8 hr light: dark photoperiod at 60 ± 10% relative humidity in small stock populations in ad libitum fodder comprising 90% organic white flour, 10% brewer's yeast and a sprinkling of rolled oats to aid traction. Approval for the research was granted by UEA\'s Animal Welfare and Ethical Review Board. To identify males from females in mating pairs, individuals were marked with a small dot on the dorsal thorax with correction fluid (Tippex, France) ([@bib92]). We used beetles from the Krakow Super Strain (KSS) in this study, a wild-type outbred lab strain created in 2008 by breeding together eleven different strains, and maintained since at 30°C (see [@bib19]). The thermal selection lines were established in 2010 and maintained since as eight independent replicates for either regime with synchronous generations at either 30 ± 1°C or 38 ± 1°C. Each line was reproduced through every adult generation using 100 haphazardly-selected, sexually mature adults (\>ten days since pupal emergence). The mixed adults were free to mate and oviposit in 150 ml of fodder in 300 ml jars with mesh lids for one week, after which they were removed and the resulting eggs and larvae allowed to develop to the next generation at their respective temperatures. After 50+ generations of experimental evolution, gametes from males and females were assayed from eight independent lines in either thermal regime. To measure the impacts of thermal regime on gamete biology, beetles were exposed to either 30°C, 38°C or 40°C temperatures in a controlled environment facility. Thermal exposure was regulated within a controlled environment facility held at 30°C, or in A.B. Newlife 75 Mk4 forced air egg incubators (A.B. Incubators, Suffolk, UK) at 38°C within the facility. 30°C is our standard rearing temperature, and the optimum laboratory temperature for population productivity in T. castaneum is 35°C ([@bib84]; [@bib76]). 38°C approaches the upper thermal limit across more sensitive juvenile stages ([@bib19]; [@bib76]), and this temperature has been recorded in 150 countries ([@bib57]). Our experimental exposure to this thermal regime will have relevance to more extreme thermal conditions and heatwaves, predicted to increase in frequency, severity and duration under climate change (e.g. [@bib62]), and which we know specifically constrain reproduction and sperm function in this system ([@bib75]). Gamete measurements {#s4-2} ------------------- Measures of mature sperm were performed following microdissection (see [@bib41]) and eggs were sieved from fodder following oviposition ([Figure 6](#fig6){ref-type="fig"}). Flour beetle eggs are oblong in shape, and size was quantified by measuring length across the long axis of rotational symmetry of the egg. Initial measures of length and width revealed a positive correlation between the two measures across 100 individual eggs, so a single length measure was used to quantify egg size. Eggs were sieved from the fodder using mechanical sieves (pore size: 300 µm, Endecotts Ltd, London, UK), and placed on a dark tile using a fine paintbrush. Eggs from the fodder were coated with a single layer of flour, and were measured at x30 magnification using a Zeiss Discovery V12 binocular microscope, AxioCam MRc5 camera and AxioVision V5.1 imaging software. Mature sperm were recovered from the base of the testicular follicles dissected out of males frozen at --6°C. Follicles were isolated on microscope slides in 30 µl drops of insect saline (0.9% NaCl), and then ruptured using fine forceps under an Olympus SZX9 binocular microscope. Once ruptured, sperm were dispersed by spreading out the area of the saline drop across the microscope slide using fine forceps, and the smears left to dry at room temperature so that the sperm cells lay in flat two-dimensional planes adhering to the glass. Images of intact spermatozoa were captured at 600X magnification under dark-field phase contrast using an Olympus BX41 microscope connected to a GT Vision GX CAM digital camera and GXCapture 8.2 software (GT Vision Ltd, UK). Sperm length was measured by creating a segmented line that traced the entire length of the cell using the 'ImageJ' analysis package and segmented line tool ([@bib80]). Previous work shows this approach has high repeatability ([@bib41]). ![Gamete measurements were performed on mature sperm dissected from males and eggs following oviposition.\ Sperm length was measured at 600X magnification in Image J by drawing segmented lines along the backbone of the cell (a and b). Egg lengths were measured along the long axis of each ovoid at 30X magnification (c and d). Further details in [@bib41].](elife-49452-fig6){#fig6} Long-term and short-term gamete divergence ([Figure 1a and b](#fig1){ref-type="fig"}) {#s4-3} ------------------------------------------------------------------------------------- We assessed the impact of temperature on development of gamete sizes following both long-term and short-term variation in thermal regime. Responses to long-term variation were measured following five years of selection (50+ generations) within replicate lines maintained in either 'warm' (30 ± 1°C) or 'hot' (38 ± 1°C) conditions. To measure impacts of short-term thermal variation we created duplicates of each of the eight replicate 30°C and 38°C lines at generation 54 to measure sperm effects and 58 for egg effects, and reared these at both 30°C and 38°C ([Figure 1b](#fig1){ref-type="fig"}) using a balanced, factorial design. Logistic contraints prevented simultaneous measurement at generation 54. To measure effects on sperm, males were exposed to either temperature from the pupal stage in 6 cm Petri dishes containing 15 ml of standard fodder, with adults allowed to emerge in groups of 20 per dish at their treatment temperature ([Figure 1b](#fig1){ref-type="fig"}). Ten days later, when reproductively mature, they were frozen for dissection and sperm measurement. We also measured full body size of males from both regimes (N = 80; five males from each of eight lines across two regimes). Sperm length variation was measured in five males per replicate line, recovering sperm from frozen males and measuring the total length of five sperm per male (=200 sperm measures from 40 males across eight lines in each of four thermal x selection treatment combinations). To measure temperature effects on eggs, mature adult females from eight replicate 30°C and 38°C lines at generation 58 were allocated to oviposit for two days at the two temperatures in a fully factorial design. Two days after the introduction of adults, we sieved oviposited eggs from the line's fodder, isolating them for measurement. Egg length measurement followed a balanced design across temperature regimes and lines, with 30 eggs measured from each replicate line at either temperature regime (=960 egg measures across eight replicate lines of either selection regime and two short-term temperature exposures). We also measured gamete size plasticity in the ancestral KSS stock population maintained at 30 ± 1°C, following methods for testing gamete size plasticity as above. Males were developed from pupae at either 30°C or 38°C and, ten days following emergence, five sperm and body length were measured from each of 26 males per treatment. Eggs were measured in three experimental blocks, within which two groups of 50 KSS females (eclosed and mated at 30°C with standard KSS males also developed at 30°C) oviposited at either 30°C or 38°C for two days in 100 ml jars with perforated lids and 80 ml of standard fodder. We measured 60 eggs per oviposition group (total N = 3600 eggs). Two additional tests measured short-term thermal impacts on egg size: the first measured egg size plasticity within individual females (N = 20 at 30°C and N = 20 at 38°C), with 40 KSS females mated to individual 30°C-developed KSS males for 24 hr at 30°C, after which 20 each were randomly allocated to oviposit at either 30°C or 38°C in 4 ml vials containing 0.5 g of standard fodder, transferring females to new vials every two days for a total of ten days, and eggs measured (length and width, µm) from vials immediately after females had been transferred on ([Figure 1bi](#fig1){ref-type="fig"}). The second test examined short-term reversibility of egg size in a group of 50 KSS mated females which were alternated for oviposition between 30°C and 38°C thermal regimes every two days for a total of 12 days, starting at 30°C ([Figure 1bii](#fig1){ref-type="fig"}). Females had been mated to standard 30°C-developed KSS males, and kept in 100 ml jars containing 80 ml of standard fodder, and 30 eggs per jar were measured immediately after transfers (total N = 180 eggs). Assessing the adaptive significance of gamete plasticity ([Figure 1c](#fig1){ref-type="fig"}) {#s4-4} --------------------------------------------------------------------------------------------- To measure adaptive plasticity in sperm, we reared KSS male pupae through either 30 ± 1°C or 38 ± 1°C temperatures as described above, and then tested the relative performance of sperm from eclosed males within KSS females and ova that were maintained in either 30 ± 1°C or 38 ± 1°C fertilisation and oviposition regimes for 100 days. Male pupae were isolated from the KSS stock population and completed development to eclosion in groups of 20 in 6 cm plastic Petri dishes with ad libitum fodder at either 30 ± 1°C or 38 ± 1°C. Three of these male groups (N = 60 pupae) were reared and maintained at 30°C, and three (N = 60 pupae) at 38°C. When sexually mature at 10 days post eclosion, individual males from either the 30°C or 38°C eclosion regime were paired with similar-aged, marked virgin females from the KSS stock population (reared at 30°C) in 7 ml vials containing 0.5 g of fodder. Pairs were allowed to mate at the male's eclosion temperature. After 24 hr, pairs were separated, and females isolated in individual 6 cm Petri dishes containing 10 g of fodder. Half the mated females within either male thermal treatment group were allowed to oviposit at 30 ± 1°C, and the other half at 38 ± 1°C, with N = 27 females in each of the four oviposition groups. Thus, we executed a fully factorial and balanced experiment in which females developed, fertilised and oviposited eggs at either 30 or 38°C, using sperm that had been produced at either 30°C or 38°C. Every ten days, females were transferred to new Petri dishes containing fresh fodder, for a maximum of 100 days (ten x 10 day blocks), by which time females had ceased to produce fertile eggs (following a single mating period, female T. castaneum typically use up all viable sperm within 100 days, after which a new mating allows resumption of fertility and offspring production \[[@bib58]\]). The number of adult offspring emerging from each Petri dish across up to 100 days of oviposition at either 30°C or 38°C quantified the reproductive success of each pair, comparing performance of sperm developed at either 30°C or 38°C when challenged with functioning at either 30°C or 38°C. To measure the adaptive significance of egg plasticity, we isolated eggs that had been developed and laid at either 30 ± 1°C or 38 ± 1°C temperatures from KSS adults, and then tested their egg-to-adult viability through either 30 ± 1°C, 38 ± 1°C or 40 ± 1°C thermal regimes. To generate phenotypic divergence in egg biology, groups of 300 females previously mated to KSS males at 30°C oviposited for 2 days at either 30 ± 1°C or 38 ± 1°C (two groups at either temperature) in 1200 ml tubs containing 600 ml of standard fodder. 600 eggs per group were sieved and isolated from the flour, counted, and transferred in clutches of 50 to develop in 100 ml jars containing yeast-free fodder (applying stronger environmental selection on offspring development). Egg clutches produced at either 30°C or 38°C were transferred to hatch and develop at either 30°C, 38°C or 40°C, with eight groups assayed in each of these three temperature treatments (=2400 eggs assayed across a total of 48 treatment groups). Statistical analysis {#s4-5} -------------------- All analyses were carried out using R Studio ([@bib73]) (v 1.1.463) in R ([@bib67]) (version 3.4.1) with Rmisc ([@bib45]), multcomp ([@bib47]), car ([@bib34]), MASS ([@bib91]), glmmTMB ([@bib8]) and lmerTest ([@bib54]) packages for data exploration and analysis. Graphical figures were plotted using ggplot2 ([@bib98]). Unless otherwise specified, all data were analysed using Linear Mixed Models (LMM) and Generalised Linear Mixed models (GLMM) in lme4 ([@bib5]), with the specific approach for each set of results described below. All data generated from the experiments described above were included for analysis, and all replication is biological. To determine the appropriate error distributions the relationship between the variance and the mean of the response variable and the assumptions for data distribution were checked ([@bib14]). Models were fitted using Restricted Maximum Likelihood (REML) methods to enable refinement and validation ([@bib89]). Residuals from linear models were checked for violations of the assumptions of normality and homoscedasticity. Significance of fixed effects in LMMs were obtained using t-tests with Satterthwaite\'s approximation for degrees of freedom implemented in lmerTest ([@bib54]). To facilitate the interpretation of main effects in the presence of interactions, we centred the contrasts between factors with two levels by coding them as minus 0.5 and 0.5, respectively ([@bib78]). Total sperm length divergence following long-term evolution and short-term temperature exposure in the thermal selection lines (TSL) was analysed using an LMM, with the selection regime (30°C or 38°C), the exposure temperature and their interaction entered as fixed factors, and the replicate male (1--5) as a random factor, nested within each of the eight replicate lines. To account for our fully factorial design, we additionally included random slopes for our replicate lines. Divergence in KSS sperm length was analysed using an LMM with thermal environment (30°C or 38°C) as a fixed effect, and male (five sperm from each of 26 males per thermal environment) nested within the two experimental runs as random effects. Adaptiveness of sperm length plasticity was assessed using a General Linear Model (LM) on the total number of offspring per mating pair (after square root transformation), with male temperature treatment (30°C or 38°C), fertilisation temperature (regime; 30°C or 38°C) and their interaction as explanatory variables. To additionally explore temporal variation in offspring production, we additionally ran zero-inflated models with gaussian distribution on offspring counts, implemented in glmmTMB ([@bib8]). Almost all females ceased to produce offspring in the last few 10-day blocks which has previously been shown to be due to sperm limitation ([@bib58]). We ran a selection of models, and selected the model with the lowest AIC value as our best model ([Supplementary files 1](#supp1){ref-type="supplementary-material"} and [2](#supp2){ref-type="supplementary-material"}). In our conditional full model, we included KSS male treatment, fertilization temperature, time (using blocks as a continuous variable), and two- and three-way interactions as fixed effects. We included random intercepts for 10-day blocks and random slopes for individual pairs to account for repeated measures across time ([@bib79]). Our zero-inflated full model included male treatment, fertilization temperature (regime), time, and two- and three-way interactions as fixed effects. Egg length divergence in groups of TSL females was analysed analogous to sperm length, using an LMM with selection regime (30°C or 38°C), short-term exposure temperature and their interaction as fixed effects, and replicate line (1--8) as a random factor, including random intercepts and slopes. Egg length plasticity in the KSS stock following oviposition at either 30°C or 38°C was analysed in an LMM with thermal regime as a fixed effect and experimental block included as a random effect. Plasticity in egg morphology in individual females was analysed in an LMM with thermal regime as a fixed effect and random intercepts for female ID as well as for two-day blocks. We additionally modelled temporal trends by including two-day blocks as a continuous fixed effect, and random slopes for individual KSS females, but found no evidence for temporal trends on egg morphology (p=0.9). To analyse reversibility of egg size plasticity we used an LMM with thermal regime as a fixed effect and two-day blocks as a random effect. Additionally, on a subset of the eggs, we measured egg width to quantify subsequent changes in morphology and overall volume with oviposition temperature (N = 20 KSS females per thermal exposure and five eggs per female measured). To assess the correlation between egg length and egg width (egg morphology) at the two oviposition temperatures, a simple LM was fitted to the data with egg width as a response variable, and egg length, oviposition temperature and their interaction as predictor variables. Adaptive developmental plasticity in egg function was assessed by testing the reproductive performance of replicate groups of 50 eggs produced by females at either 30°C or 38°C when hatching and developing in 30°C, 38°C or 40°C thermal environments. Egg performance was analysed using a Generalised Linear Model (GLM) with binomial error structure (logit link) in which egg production temperature (30°C or 38°C), developmental thermal environment (30°C, 38°C or 40°C), and their interaction, were entered as fixed effects. We included the number of successes (developed offspring) and failures (eggs that failed to hatch/develop) using the cbind function, and confirmed that our model was not overdispersed. Finally, temperature impacts on male body size were analysed using a simple LM. Female body sizes were not assessed because egg size plasticity tests under thermal variation were conducted on already-emerged mature adult females. An overview of sample sizes is given in [Tables 1](#table1){ref-type="table"} and [2](#table2){ref-type="table"}. Box plots indicate the median and interquartile ranges (IQR), with whiskers indicating data within 1.5 IQR. A central filled marker indicates the mean of the sample. Data accessibility statement {#s4-6} ---------------------------- All data generated and analysed in this study, together with R codes, are openly provided as an associated source file in our Dryad Data Repository with the identifier: <https://doi.org/10.5061/dryad.83bg17q>. Funding Information =================== This paper was supported by the following grants: - http://dx.doi.org/10.13039/501100000270Natural Environment Research Council NE/K013041/1 to Matthew Gage. - http://dx.doi.org/10.13039/501100000867Commonwealth Scholarship Commission Commonwealth Rutherford Fellowship to Ramakrishnan Vasudeva. - http://dx.doi.org/10.13039/501100001711Swiss National Science Foundation P300PA_177906 to Andreas Sutter. This work was supported by NERC grant NE/K013041/1, a Commonwealth Rutherford Fellowship (to RV) by the Commonwealth Scholarship Commission, UK, and a Swiss National Science Foundation fellowship (to AS; grant number P300PA_177906). We thank Maya Krishnan Kumar for help with ggplot2, and the eLIFE editorial team and three expert referees for reviewing the work in detail and substantially improving the manuscript. Additional information {#s5} ====================== No competing interests declared. Conceptualization, Data curation, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing---original draft, Project administration, Writing---review and editing. Conceptualization, Data curation, Formal analysis, Supervision, Validation, Visualization, Writing---original draft, Project administration, Writing---review and editing. Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing---review and editing. Data curation, Investigation, Methodology, Writing---review and editing. Data curation, Supervision, Investigation, Methodology, Project administration, Writing---review and editing. Conceptualization, Resources, Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing---original draft, Project administration, Writing---review and editing. Animal experimentation: This study was approved by, and followed strict guidelines to, the University of East Anglia\'s Animal Welfare and Ethical Review Board. Additional files {#s6} ================ 10.7554/eLife.49452.013 ###### Table model summary of the best glmmTMB model (lowest AIC value; see [Supplementary file 1](#supp1){ref-type="supplementary-material"} Table 2) for reproductive output of males reared at 30°C or 38°C (Treatment) with offspring developing at 30°C or 38°C (Regime). 10.7554/eLife.49452.014 ###### Table overview of models for adaptive significance of sperm plasticity. Models were run using glmmTMB and were sorted along ascending AIC values. All conditional models additionally included random intercepts for ten-day blocks as well as random intercepts and random slopes for individual males. 10.7554/eLife.49452.015 Data availability {#s7} ================= Data accessibility statement: All data generated and analysed in this study, together with R codes, are openly provided as an associated source file in our Dryad Data Repository with the identifier: <https://doi.org/10.5061/dryad.83bg17q>. The following dataset was generated: RamakrishnanVasudevaAndreasSutterKrisSalesMatthewJ. G. Gage2019Data from: Adaptive thermal plasticity enhances sperm and egg performance in a model insectDryad10.5061/dryad.83bg17q 10.7554/eLife.49452.019 Decision letter Donoso David Reviewing Editor Escuela Politécnica Nacional Ecuador Gemmell Neil J Reviewer University of Otago New Zealand Pizzari Tom Reviewer University of Oxford United Kingdom In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included. Thank you for submitting your article \"Adaptive thermal plasticity enhances sperm and egg performance in a model insect\" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Patricia Wittkopp as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Neil J Gemmell (Reviewer \#2); Tom Pizzari (Reviewer \#3). The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission. Summary: The authors make use of the Tribolium flour beetle model to assess the potential for thermal adaptation in sperm and egg function. Exposing flour beetles to increased temperature under both long-term experimental evolution and short term acclimation, the authors found that thermal regime regulates gamete size development in both males and females. They find that long-term selection regimes at high temperatures (38°C) led to males producing sperm 4% longer than those evolved at lower temperatures (30°C). However, short term changes in temperature generated the opposite pattern: sperm production in males subjected to 38°C within a single generation produced sperm 7% shorter than those exposed to 30°C. Females also showed a response to both the long and short term selection regimes, producing longer eggs at higher temperatures. Using a series of well-conducted thermal translocation experiments, the authors show that this gametic phenotypic plasticity was adaptive. Both sperm and eggs produced in warmer conditions had superior reproductive performance in warmer environments, and vice versa for colder conditions and reproductive environments. Further, in warmer environments, the gamete plasticity observed enabled males to double their reproductive success, and females could increase offspring production by one-third. Collectively the authors\' work identifies a new dimension in gametic plasticity important for understanding how organisms may respond to the increasing thermal variation in the natural environment. This is an excellent study, presenting a uniquely exhaustive experimental investigation into the consequences of environmental temperature for gametic phenotype and related fitness consequences in an ectotherm invertebrate model species. While the biological ramifications of climatic change, including environmental temperature, have been the focus of intense and increasing research interest in recent years, surprisingly little remains known about the impact on gametic performance and associated implications for reproductive fitness. The study therefore makes an influential contribution to a topical and unresolved question. Essential revisions: Perhaps the most intriguing result is the strong contrast in phenotypic response in sperm length under long-term and short-term regimes. Under long term selection to higher temperatures sperm get longer, while under the short term regime, they got shorter. I find this a puzzling result, which I cannot easily reconcile. The authors have put forward some plausible explanations, which I think are sensible, but I did wonder if other factors might contribute to or explain this result. First, was there potential for the sex ratios in the selection lines to have altered as a consequence of thermal exposure, which may have affected the level of sperm competition across some treatments? I wondered if this might be a particular factor in explaining the difference in sperm length response between short-term and long-term selection experiments? I know the authors are expert in this realm, so assume they controlled for this possibility, but some details on whether sex ratios differed and how these were handled might be useful to include. Similarly, I wondered if there were any data on sperm volumes or concentrations. While sperm length clearly responded to the selection regime, I wondered if the authors might also know how sperm number, ejaculate volume and motility altered as there are often strong interactions among these traits? Probably these traits are not easily measured in flour beetles, but if there are data, they might further strengthen this already impressive piece of work. Last, I wondered if there might be an ageing effect that was driving some of the findings observed. In particular, I was curious as to whether males age faster and/or sperm mature faster at higher temperatures. Might part of the effect observed be driven by male age and ejaculate age? Perhaps there is a way to factor in degree days to control for any variation that might exist among experiments? The authors need to justify the experimental temperatures the beetles were exposed to. In special, 38 C is not a yearly average temperature that exists anywhere in the world. Even the lower temperature (30 C) is not a common average temperature anywhere in the world. And very unlikely it will exist, even under global warming scenarios. So, the paper can be better framed in the context of seasonal temperature variability (growing seasons, summer, etc.) and not necessarily in a global warming context. Comparing only two temperatures 30 vs. 38 C is not ideal, besides being temperatures that are very far apart from each other, it is impossible to draw a general trend when you only have one comparison. It would have been great had the manuscript had an intermediate temperature treatment, at 34 C (or so). 10.7554/eLife.49452.020 Author response > Essential revisions: > > Perhaps the most intriguing result is the strong contrast in phenotypic response in sperm length under long-term and short-term regimes. Under long term selection to higher temperatures sperm get longer, while under the short term regime, they got shorter. I find this a puzzling result, which I cannot easily reconcile. The authors have put forward some plausible explanations, which I think are sensible, but I did wonder if other factors might contribute to or explain this result. First, was there potential for the sex ratios in the selection lines to have altered as a consequence of thermal exposure, which may have affected the level of sperm competition across some treatments? I wondered if this might be a particular factor in explaining the difference in sperm length response between short-term and long-term selection experiments? I know the authors are expert in this realm, so assume they controlled for this possibility, but some details on whether sex ratios differed and how these were handled might be useful to include. Similarly, I wondered if there were any data on sperm volumes or concentrations. While sperm length clearly responded to the selection regime, I wondered if the authors might also know how sperm number, ejaculate volume and motility altered as there are often strong interactions among these traits? Probably these traits are not easily measured in flour beetles, but if there are data, they might further strengthen this already impressive piece of work. Last, I wondered if there might be an ageing effect that was driving some of the findings observed. In particular, I was curious as to whether males age faster and/or sperm mature faster at higher temperatures. Might part of the effect observed be driven by male age and ejaculate age? Perhaps there is a way to factor in degree days to control for any variation that might exist among experiments? We have added a major section to the Discussion where we offer additional explanations for the general evolution of longer sperm in the warmer regime, despite short-term responses leading to shorter sperm. We include the reviewer suggestions for factors arising from sexual selection and/or developmental ageing. We also suggest nutrient access through the whole life cycle for spermatogenesis as a potential factor, since nutrient limited males develop shorter sperm in T. castaneum (Godwin et al., 2017). Finally, we consider the possibility that longer-term evolution in a warmer 38^o^C environment, where sperm are pushed by short-term effects (perhaps via faster development rate) to be shorter, has led to compensatory selection for sperm elongation. We do not have data on sperm number/volume/concentration or motility in our thermal lines, because achieving meaningful measures of these pose problems in this system. Individual male T. castaneum can mate with multiple females throughout their long life, successfully inseminating an average of 50 females within 7 days if given the opportunity (Lumley et al., 2015), so individual ejaculate sperm number/volume/concentration will not reveal the capacity for overall sperm number production. (If the reviewer refers to sperm cell volume for consideration, T. castaneum sperm cells are also small and extremely thin, and may vary in diameter along their long axis, so volume will also be a challenge to measure accurately.) Finally, and as in most insects, motility is also a problem to measure meaningfully in T. castaneum because sperm evidently function through the female tract by moving within narrow tubules or within dense sperm masses, so the motility of individual isolated sperm (even if the tract biochemistry could be replicated) on a microscope slide will likely not reflect function in the natural fertilisation environment. We therefore now suggest the following additional interpretations for why the 38^o^C thermal selection lines have evolved slightly longer sperm, whereas short-term temperature increases through pupal and adult development cause a reduction in sperm length: "Although we found evidence that laboratory selection across 54 generations at 38°C caused the overall evolution of \~4% longer sperm compared with the thermal line selection at 30°C, our within-generation experiments revealed consistent \~7% reductions in sperm length for all populations when produced through the warmer 38°C regime compared with 30°C (Figure 2). \[...\] Whatever the cause behind long-term experimental evolution of sperm length within our thermal lines, it is clear that short-term within-generation impacts of temperature have strong and direct effects on sperm development, with an experimental switch to warmer regimes for either the thermal selection lines or the ancestral stock population resulting in consistently clear reductions in sperm length (Figure 2). Temperature variation during gametogenesis is known to affect sperm size in ectotherms..." > The authors need to justify the experimental temperatures the beetles were exposed to. In special, 38 C is not a yearly average temperature that exists anywhere in the world. Even the lower temperature (30 C) is not a common average temperature anywhere in the world. And very unlikely it will exist, even under global warming scenarios. So, the paper can be better framed in the context of seasonal temperature variability (growing seasons, summer, etc.) and not necessarily in a global warming context. Comparing only two temperatures 30 vs. 38 C is not ideal, besides being temperatures that are very far apart from each other, it is impossible to draw a general trend when you only have one comparison. It would have been great had the manuscript had an intermediate temperature treatment, at 34 C (or so). Indeed, we agree that our experimental thermal exposures are not relevant to annual average temperature effects, but seasonal variation, which is becoming more extreme as climate change drives up atmospheric volatility. We now frame the work more in this context, and make it clearer throughout the manuscript that our study is relevant to thermal variability, and not annual averages (including three new references to this). Although we cite climate change as being relevant to our study, we believe that our findings have primary impact in revealing a novel example of basic biological function for dealing with inherent thermal variation in the natural environment. Temperatures exceeding 38°C have been recorded in 150 countries (Mherrera, 2019), so are very frequently encountered in tropical latitudes. It is also relevant to note that the laboratory 'optimum' for population productivity in our T. castaneum model is a warm 35°C (stated in the manuscript Materials and methods), so 38°C is not too far apart from that. We also agree that additional thermal regimes in the study would have been great, but the existing workload to complete this work through a number of years at control and extreme (but relevant) experimental temperatures was itself more than a stretch. Changes to framing the work: "One of the most important abiotic environmental variables is temperature (Cossins and Bowler 1987;Angilletta Jr., 2009), especially in the context of climate change when thermal environments are expected to both warm and become much variable and extreme (Perkins et al., 2012, Raftery et al., 2017)." "After exposing adult males and females and their gametes to different temperatures, we compare sperm and egg development and reproductive function within thermal regimes that mimic the increasingly variable conditions faced by warmer tropical regions (Perkins et al., 2012)." "...an obvious mechanism to enable adaptive thermal plasticity for populations facing the challenge of reproducing under climate change where increases in both thermal averages, maxima and variation are expected (Perkins et al., 2012, Raftery et al., 2017)." "30°C is our standard rearing temperature, and the optimum laboratory temperature for population productivity in T. castaneum is 35°C (Sokoloff, 1974, Sales, 2019). \[...\] Our experimental exposure to this thermal regime will have relevance to more extreme thermal conditions and heatwaves, predicted to increase in frequency, severity and duration under climate change (e.g. Perkins et al., 2012), and which we know specifically constrain reproduction and sperm function in this system (Sales et al., 2018)."	{ "pile_set_name": "PubMed Central" }
Introduction ============ Breast cancer is one of the most common female malignant tumors with high morbidity and mortality rate caused by its strong metastatic ability ([@b1-ol-0-0-6571],[@b2-ol-0-0-6571]). There are no consensus biomarkers for early diagnosis and prognosis assessment of breast cancer with applications in clinical practice. Therefore, the development of breast cancer biomarkers has attracted increased attention recently. Special AT-rich sequence-binding protein 1 (SATB1) binds to T-rich sequences in chromosomes to regulate the expression of downstream genes ([@b3-ol-0-0-6571],[@b4-ol-0-0-6571]). The expression level of SATB1 is low in normal tissue, but is elevated in a variety of tumors ([@b5-ol-0-0-6571]--[@b7-ol-0-0-6571]). Toll-like receptor 4 (TLR4) is mainly expressed in immune cells, but can also be expressed in tumor cells ([@b8-ol-0-0-6571]--[@b10-ol-0-0-6571]). In this study, we detected the expression of SATB1 and RLR4 in 120 cases of cancer and 53 cases of adjacent non-cancerous tissue by immunohistochemistry. The correlation between the expression of these two proteins and the clinical characteristics of patients were analyzed. Patients and methods ==================== ### Patient information A total of 120 patients diagnosed with breast cancer in Yuhuangding Hospital of Yantai from October 2014 to October 2016 was enrolled in the study. Cancer tissue was collected after surgical resection. At the same time, adjacent non-cancerous tissue was collected from 53 patients. All the patients were females, and their ages ranged from 28 to 65 years, with a mean age of 46.5±11.7 years. No patient had been treated with chemotherapy before the study. Specimens were collected from necrotic cancer tissue and the adjacent non-cancerous tissue within 3 cm, fixed, and embedded in paraffin. Cancerous samples were diagnosed as breast cancer by pathological examination. The study was approved by the Ethics Committee of Yuhuangding Hospital of Yantai. All the patients signed an informed consent before being enrolled in the study. ### Reagents and methods Anti-human SATB1 monoclonal antibody and rabbit anti-human TLR4 monoclonal antibody were purchased from Abcam (Cambridge, UK). The DAB kit and hematoxylin were purchased from ZSBG-Bio (Beijing, China). Horseradish peroxidase-conjugated secondary antibody was purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Paraffin-embedded breast cancer samples were cut into 4 µm sections and transferred onto glass slides. After baking for 2 h at 90°C, tissue sections were dewaxed and rehydrated. After that, antigen retrieval was performed by incubating with 0.01 M sodium citrate buffer for 15 min. Endogenous peroxidase blocker was then added and incubated at 37°C for 10 min. After blocking with goat serum at room temperature for 20 min, the primary antibodies of SATB1 (1:300) and TLR4 (1:300) were incubated with the slides at 4°C overnight. After washing, secondary antibody was added and incubated at 37°C for 1 h. DAB staining was then performed and tissue sections were examined under the microscope to observe the staining. After hematoxylin staining, the slides were dehydrated, cleared, and sealed. All the operations were performed in accordance with the manufacturer\'s instructions. ### Determination of experimental results The brown or yellow granules on the slides showed the positive expression of SATB1 and TLR4. SATB1 mainly accumulated in the nucleus and TLR4 mainly accumulated in the cytoplasm. Using ×400 magnification in a bright field microscope (Leica, Wetzlar, Germany), 10 distinct visual fields were selected to count the positive cells and record the degree of staining. We also calculated the percentage of positive cells. Scoring was performed according to the degree of staining: no staining, 0 points; light yellow, 1 point; yellowish-brown, 2 points; chocolate brown, 3 points. Scoring was also performed according to the percentage of positive cells: 0--25%, 1 point; 25--65%, 2 points; 65--100%, 3 points. The product of the 2 scores greater than 3 was taken as positive expression; values below 3 were considered negative expression ([@b11-ol-0-0-6571]). ### Statistical analysis SPSS 19.0 statistical software (IBM SPSS, Armonk, NY, USA) was used to analyze the data. The count data were analyzed by Chi-square test. Correlation analysis was performed by Spearman\'s rank correlation analysis. P\<0.05 was considered to be statistically significant. Results ======= ### Expression of SATB1 and TLR4 in breast cancer and adjacent non-cancerous tissue SATB1 expression was observed in the nucleus under a microscope. SATB1 was positively expressed in 89 cases of breast cancer, and the positive expression rate of SATB1 was 74.1% ([Fig. 1](#f1-ol-0-0-6571){ref-type="fig"}). Positive expression of SATB1 was only detected in 7 cases of adjacent non-cancerous tissue, and the positive expression rate of SATB1 was 13.21%. A statistically significant difference in the expression of SATB1 was found between breast cancer and adjacent non-cancerous tissue (P\<0.05; [Table I](#tI-ol-0-0-6571){ref-type="table"}). TLR4 expression was detected in the cytoplasm. TLR4 was positively expressed in 70 cases of breast cancer, and the positive rate was 58.33% ([Fig. 1](#f1-ol-0-0-6571){ref-type="fig"}). Positive expression of SATB1 was detected in 48 cases of adjacent non-cancerous tissues, and the positive rate was 90.57%. A statistically significant difference in the expression of TLR4 was found between breast cancer and adjacent non-cancerous tissues (P\<0.05; [Table I](#tI-ol-0-0-6571){ref-type="table"}). ### Correlation between expression of SATB1 and TLR4 Following the immunohistochemistry results, we next analyzed the correlation between SATB1 and TLR4. As shown in [Table II](#tII-ol-0-0-6571){ref-type="table"}, the expression of SATB1 was negatively correlated with the expression of TLR4 (r=−0.624, P\<0.05). ### Correlation of SATB1 and TLR4 with the clinical and pathological features of patients We last examined the correlation of the expression of SATB1 and TLR4, and the clinical and pathological features of patients. We found that the expression levels of SATB1 and TLR4 were not significantly correlated with the age, menopause, PR protein, and HER-2 protein expression (P\>0.05). However, the expression levels of SATB1 and TLR4 were significantly correlated with tumor size, local lymphatic metastasis, histopathological grade, tumor stage, and the expression of ER protein (P\<0.05; [Table III](#tIII-ol-0-0-6571){ref-type="table"}). Discussion ========== SATB1 is a nuclear matrix attachment-binding protein with tissue-specific expression. The SATB1 gene is located on chromosome 3p23 and encodes for a 763-amino acid protein ([@b12-ol-0-0-6571]). SATB1 is highly expressed in thymus where it regulates the development and maturation of T cells ([@b13-ol-0-0-6571],[@b14-ol-0-0-6571]). Previous studies showed that SATB1 gene knockout in mice can inhibit the production of CD4^+^ and CD8^+^ double positive T cells, leading to disorders of thymus cell maturation ([@b15-ol-0-0-6571]). SATB1 plays a role as a 'gene organizer' in the genome. SATB1 can interact with more than 1,000 proteins to specifically regulate the expression of its target genes by chromatin remodeling and protein modification ([@b16-ol-0-0-6571]). SATB1 can bind the BUR region of target genes and anchor the BUR region on the nuclear matrix to alter the higher-order structure of the chromatin, and regulate gene expression ([@b17-ol-0-0-6571]). SATB1 can also regulate DNA binding capacity and the subcellular localization of proteins through phosphorylation, ubiquitination, and acetylation ([@b18-ol-0-0-6571]). Liu et al found that high expression of SATB1 in breast cancer cells significantly increased cell invasion ability ([@b19-ol-0-0-6571]). Clinical data from 1,318 breast cancer patients showed that the expression level of SATB1 was negatively correlated with survival time ([@b20-ol-0-0-6571]). Our study shows that SATB1 is strongly expressed in breast cancer and weakly expressed in adjacent non-cancerous tissue. The expression of SATB1 was not significantly correlated with age, menopause, and the expression of the PR and HER-2 proteins, but was significantly correlated with tumor size, local lymphatic metastasis, histopathological grade, tumor stage, and ER protein expression. Our findings are consistent with previous studies ([@b21-ol-0-0-6571],[@b22-ol-0-0-6571]). Toll-like receptors were first found in Drosophila. In 1997, TLR4 homologue was identified in humans, and so far, there are 12 members of the TLRs family ([@b23-ol-0-0-6571]). The TLR4 gene is located on chromosome 9q32-q33, and encodes for a 224-amino acid protein. TLR4 is widely distributed on the cell surface to sense pathogens ([@b24-ol-0-0-6571]). TLR4 is widely distributed in human monocytes ([@b25-ol-0-0-6571]), neutrophils ([@b26-ol-0-0-6571]), and epithelial cells ([@b27-ol-0-0-6571]). TLR4 can recognize a variety of pathogen-associated molecular patterns (e.g., LPS of Gram-negative bacteria) to induce different immune responses ([@b28-ol-0-0-6571]). Through binding to the corresponding ligands and mediating intracellular signal transduction, TLR4 plays a role as a transcription factor to activate the expression of a variety of cell growth and apoptosis-related factors ([@b29-ol-0-0-6571],[@b30-ol-0-0-6571]). TLR4 can mediate MyD88-dependent pathways through the interaction with a series of cytokines to promote tumorigenesis ([@b31-ol-0-0-6571]). Clinical studies also found that TLR4 was correlated with the growth and metastasis of gastric cancer ([@b32-ol-0-0-6571]), ovarian cancer ([@b33-ol-0-0-6571],[@b34-ol-0-0-6571]), cervical cancer ([@b35-ol-0-0-6571]), and other types of tumor cells. In our study, we found that TLR4 was positively expressed in 58.33% cases of breast cancer tissues and in 90.57% cases of adjacent non-cancerous tissue. The positive expression rate of TLR4 in this study is consistent with previous studies ([@b35-ol-0-0-6571],[@b36-ol-0-0-6571]). We also found that TLR4 expression was not significantly correlated with age, menopause, or the expression of the PR and HER-2 proteins. However, TLR4 was significantly correlated with tumor size, local lymphatic metastasis, histopathological grade, tumor stage, and ER protein expression. In addition, correlation analysis indicated that the expression level of SATB1 was negatively correlated with the expression level of TLR4. In conclusion, SATB1 and TLR4 are involved in the development of breast cancer, which is of great significance for the identification of potential therapeutic targets and prognosis of breast cancer. ![Expression of SATB1 and TLR4 in breast cancer and adjacent non-cancerous tissue. (A) Representative result of SATB1 expression in adjacent non-cancerous tissue (×400). (B) Representative result of SATB1 expression in breast cancer (×400). (C) Representative result of TLR4 expression in adjacent non-cancerous tissue (×400). (D) Representative result of TLR4 expression in breast cancer (×400). The positive expression rate of SATB1 in breast cancer tissues was significantly higher than that in adjacent non-cancerous tissue. The positive expression rate of TLR4 in adjacent non-cancerous tissues was significantly higher than that in breast cancer. SATB1, special AT-rich sequence-binding protein 1; TLR4, toll-like receptor 4.](ol-14-03-3611-g00){#f1-ol-0-0-6571} ###### Expression of SATB1 and TLR4 in breast cancer and adjacent non-cancerous tissue. SATB1 TLR4 ------------------------------- ------ --------- ------ ----------- ----- ---- ------- Breast cancer 120 31 89 74.17 50 70 58.33 Adjacent non-cancerous tissue 53 46 7 13.21 5 48 90.57 χ^2^ value 37.413 26.481 P-value 0.006 0.011 SATB1, special AT-rich sequence-binding protein 1; TLR4, toll-like receptor 4. ###### Correlation between the expression of SATB4 and TLR4 in breast cancer tissues. SATB1 ---- ------- ----- -------- ------- \+ 43 27 −0.624 0.003 − 46 4 SATB1, special AT-rich sequence-binding protein 1; TLR4, toll-like receptor 4. ###### Correlation of the expression of SATB1 and TLR4 with the clinical and pathological features of patients. SATB1 TLR4 ------------------------- ---- ------- ------- ------- ---- ------ ------- Age (years) \<50 46 35 0.25 0.416 28 0.74 0.352 ≥50 74 54 42 Menopause Yes 52 38 1.36 0.129 30 0.62 0.391 No 68 51 40 Tumor size \<2 cm 44 29 5.24 0.017 23 8.49 0.013 ≥2 cm 76 60 47 Lymph node metastasis Yes 56 31 9.15 0.013 26 6.37 0.015 No 64 57 44 Histopathological grade I 38 19 11.48 0.007 18 7.14 0.015 II 53 44 36 III 29 26 16 Tumor stage I 32 18 12.53 0.007 16 9.61 0.011 II 57 43 33 III, IV 31 28 21 PR (−) 48 35 0.94 0.172 28 1.74 0.114 (+) 72 54 42 ER (−) 53 29 6.74 0.015 24 7.93 0.013 (+) 67 60 46 HER-2 (−) 52 38 0.87 0.181 30 0.96 0.172 (+) 68 51 40 SATB1, special AT-rich sequence-binding protein 1; TLR4, toll-like receptor 4. [^1]: Contributed equally	{ "pile_set_name": "PubMed Central" }
The prevalence of diabetes continues to escalate around the world, with an estimated 24 million people affected in the U.S. ([@B1]). Its prevalence has more than doubled in Tennessee the last decade from 5.0% in 1997 to 11.0% in 2007 ([@B2]), with estimated diabetes related mortality of 31 per 100,000 population ([@B2]). Rising prevalence and shortage of physicians, especially in rural areas, contribute to poor outcome in diabetic patients ([@B3]). Telemedicine, the transfer of electronic medical data to a remote location utilizing telecommunications technology, may be beneficial in improving access, care, and outcome in diabetic patients. Therefore, we investigated the impact of telemedicine on the quality of care in diabetic patients in five health professional shortage areas with diabetes related mortality rates of 41.5--84.7 per 100,000. ADT was a prospective interventional study, which recruited diabetic patients aged ≥18 years, with A1C ≥8.0%. Diabetes self-management education (DSME) was delivered by a certified diabetes educator via videoconference every 3 months. The DSME classes addressed pathogenesis of diabetes, nutritional education, physical activity, self blood glucose monitoring, effects of insulin and other diabetes medications, sick day management, and complications of diabetes. DSME classes were delivered from the telemedicine studio of the University of Tennessee Health Science Center in Memphis, which was connected to the remote locations by videoconference using Polycom VSX 7000 video cameras (Pleasanton, CA), television monitors, and the Internet. Vital signs and anthropometry were obtained at each visit. Laboratory data such as A1C and lipid levels were obtained from the patients' primary physician. Statistical analysis was done using Student t test or ANOVA and χ^2^ test. Thirty-six patients aged 55.6 ± 10.0 years, with mean duration of diabetes of 12.0 ± 10.0 years completed the study; 55% of the subjects had concomitant dyslipidemia and hypertension. Subjects showed significant improvement in A1C after 3 months (9.8 vs. 8.4%, P \< 0.001). Reduction in A1C at 3 months was sustained over 12 months of the study (8.4 vs. 8.3%, P \> 0.5). The proportion of subjects achieving target A1C of \<7% increased to 40% after intervention (P \< 0.001). There was also a significant rise in the proportion of patients who achieved target blood pressure (61 vs. 78%), HDL (11 vs. 22%), and triglyceride (3 vs. 30%) (P \< 0.001); and 97% of the participants were satisfied with the program, and 90% acknowledged it saved time and money. Given the average duration of uncontrolled diabetes of 12 years, this degree of improvement in 3 months is quite encouraging. The improvement in A1C, which was maintained throughout the study, may suggest that a 3-month program may be more cost-effective than a longer one. This prospective study is a useful addition to the evidence for the efficacy of a well-designed telemedicine program in diabetic patients ([@B4],[@B5]). However, it is limited by the small number of subjects. A prospective randomized study would be required to verify the observations in this study. In conclusion, a telemedicine-based management program that delivers diabetes education and consultative service through videoconference was effective in improving diabetes outcome measures. It proved to be an acceptable means of providing care in diabetic patients at high risk for cardiovascular disease. ADT was supported by a grant (ED-08-22875-00) from the Tennessee Department of Health. No potential conflicts of interest relevant to this article were reported. E.A.N. contributed to data research, discussion, and manuscript writing and review. S.A. contributed to data research and manuscript review. J.T. contributed to data research and manuscript review. S.S.N. contributed to data research and manuscript review. A.E.K. contributed to discussion and manuscript writing and review.	{ "pile_set_name": "PubMed Central" }
Background {#Sec1} ========== Natural enemies are important in controlling the population of insect pests in agricultural ecosystems \[[@CR1]--[@CR6]\]. The presence of natural enemies is detected through chemical cues emitted by prey and host plants, alone or in association, and received by the antennal sensory system in the multi-trophic environment to result in various behavioral choices including prey forage, localization, oviposition and escape \[[@CR7]--[@CR17]\]. Recently the study of insect peripheral sensory system and chemosignal transduction has experienced considerable progress due to the development of bioinformatics-based approaches and protein function prediction methods \[[@CR18]\]. In particular, a large amount of information is provided by antennal transcriptome projects from various insect taxa \[[@CR19]--[@CR28]\]. In general, the process of chemoreception, including olfaction and taste, involves several families of genes, including odorant receptors (ORs), ionotropic receptors (IRs), and gustatory receptors (GRs) \[[@CR20], [@CR29]--[@CR31]\]. In addition, odorant binding proteins (OBPs), chemosensory proteins (CSPs) and sensory neuron membrane proteins (SNMPs) also play crucial roles in chemoreception \[[@CR32]--[@CR36]\]. The insect chemoreceptor superfamily including ORs and GRs was first identified in the Drosophila melanogaster genome \[[@CR18], [@CR21]\]. Insect odorant and gustatory receptors were once thought to be G-protein-coupled receptors just like ORs in worms and vertebrates, but subsequent studies have shown a lack of homology to vertebrate ORs \[[@CR37]\]. One such superfamily encoding ORs is highly divergent across insect taxa with sequences and frequencies varying to a large extent \[[@CR18], [@CR38], [@CR39]\]. ORs are broadly tuned to alcohols, ketones, and esters generally present in the environment \[[@CR40], [@CR41]\]. Another family encoding GRs, or receptors for taste or contact stimuli, is also very divergent across insect taxa \[[@CR31]\]. On the contrary, one example of exceptionally conserved GRs are GR21a and GR63a, which work together as a CO~2~ receptor in Drosophila \[[@CR42]\]. Such chemoreceptors play an important role in host seeking behaviors in many insects but, especially seen in mosquitoes \[[@CR43], [@CR44]\]. A new insect chemosensory family was identified recently and given the name ionotropic receptors (IRs). These IRs belong to the ionotropic glutamate receptor superfamily (iGluRs) and were identified in both the olfactory and gustatory systems \[[@CR30], [@CR45]\]. IRs are more greatly conserved than ORs and GRs but considerable variations can be observed in ligand-binding domains. They are mainly tuned to acids, amines and other odorants that are not detected by ORs \[[@CR30], [@CR45], [@CR46]\]. Since chemosensory gene families were characterized in two important model species, D. melanogaster and Anopheles gambiae \[[@CR47], [@CR48]\], a growing number of chemosensory genes have been identified from many Dipteran species, such as Musca domestica \[[@CR49]\], Bactrocera dorsalis (genome: assembly ASM78921v2), Calliphora stygia \[[@CR50]\], Glossina morsitans morsitans \[[@CR51]\], and Mayetiola destructor Say \[[@CR20]\]. Protein prediction methods have been the first step for functional identification of chemosensory genes. All information regarding insect chemosensory was obtained bioinformatically and has been beneficial in understanding insect processing of diverse volatile compounds and cross-species differences in chemical communication. The syrphids belong within the Diptera order and their larvae are aphid-specific natural enemies \[[@CR1], [@CR2], [@CR4], [@CR5], [@CR8]\]. Due to the larvae's agricultural importance via potential applications, several reports have been published on the chemical ecology of these insects. In Episyrphus balteatus DeGeer, larvae may use a sesquiterpene as a kairomone \[[@CR17], [@CR52]\] and other potential semiochemicals to locate their prey \[[@CR8], [@CR53]\]. In Sphaerophoria rueppellii, adult females are strongly attracted to odors from aphid colonies showing that specific volatile compounds are important to detect their prey \[[@CR1]\]. Some studies on the relationships between aphid or host plant volatile emissions and aphid localization and foraging behavior have shown strong associations with syrphid recognition. One striking finding has shown that volatiles from plants attacked by aphids produce strong electrophysiological responses from the antennae of syrphids \[[@CR7], [@CR8]\]. These studies indicate that detecting prey-derived volatiles ((E)-β-farnesene), herbivore-induced plant volatiles (monoterpenes and sesquiterpenes), or naturally occurring general leaf volatiles (GLVs; alcohols, aldehydes and esters) help natural enemies to select oviposition sites and locate their prey \[[@CR8], [@CR12], [@CR13], [@CR17], [@CR52]--[@CR55]\]. Despite these reports on chemosensory behavior, little is known on the molecular basis of syrphid olfaction. Therefore, the identification of predatory syrphid chemosensory gene families will help reveal how syrphids forage on their prey and choose oviposition sites. In this study we selected two syrphid species, E. balteatus and Eupeodes corollae Fabricius, active [in](http://cn.bing.com/dict/search?q=in&FORM=BDVSP6&mkt=zh-cn) northern China cotton fields to perform antennal transcriptome sequencing in order to explore and compare chemosensory genes in the two species. A total of 154 and 134 chemosensory candidate genes were identified in E. balteatus and E. corollae transcriptomes, respectively, including ORs, IRs, GRs, OBPs, CSPs and SNMPs. Furthermore, we report the expression profile of the OR families found in each insect transcriptome. A comparison between these two syrphids and other insect species revealed candidate chemosensory genes that could be involved in prey selection and plant volatile recognition. The discovery of putative chemosensory genes gives way for further exploration into functional assessments regarding chemoreception association. Results {#Sec2} ======= Antennal transcriptome sequencing and sequence assembly {#Sec3} ------------------------------------------------------- E. balteatus and E. corollae antennal transcriptomes were sequenced using the Illumina HiSeq 2000 platform combined with Trinity assembly. Approximately 68.71 million and 77.28 million raw-reads were obtained from E. balteatus male and female antenna, respectively, reduced after filtering, to 65.69 and 74.25 million clean-reads. These were assembled into 57,950 unigenes for male and 68,165 for female. A final transcript dataset with 53,575 unigenes was obtained, consisting of 17,407 distinct clusters and 36,168 distinct singletons. The dataset was 47.61 megabase (Mb) in size with a mean length of 889 bp and N50 of 1724 bp (Additional file [1](#MOESM1){ref-type="media"}: Table S1). Parallel experiments generated 80.15 million and 77.38 million raw-reads in E. corollae male and female, respectively, and 65.69 and 74.25 million clean-reads. From these datas 54,116 and 61,220 unigenes were obtained for male and female, respectively. The final transcript dataset of E. corollae contained 50,942 unigenes with a mean length of 1039 bp and N50 length of 2104 bp, consisting of 18,054 distinct clusters and 32,888 distinct singletons (Additional file [1](#MOESM1){ref-type="media"}: Table S1). In addition, unigenes with a sequence length \> 500 bp accounted for 47.64% and 43.59% of the E. corollae and E. balteatus transcriptome assembly, respectively. Homology analysis and gene ontology (GO) annotation {#Sec4} --------------------------------------------------- A BLASTX homology search against the NCBI non-redundant protein database indicated 23,680 (44.2%) and 25,606 (50.3%) unigenes from E. balteatus and E. corollae, respectively, with sequence similarities to known proteins using a cut-off E-value of 10^−5^. For E. balteatus, the larger number of similar genes (27.4%) belonged to Ceratitis capitata followed by M. domestica (20.4%), D. melanogaster (7.4%), Drosophila willistoni (3.1%), Drosophila virilis (3.0%) and Drosophila mojavensis (2.5%). For E. corollae, again C. capitata was best represented (25.1%), followed by M. domestica (18.8%), D. melanogaster (7.1%), D. willistoni (2.9%), Acyrthosiphon pisum (2.8%), and D. virilis (2.5%) (Additional file [2](#MOESM2){ref-type="media"}:Fig. S1A). Gene ontology (GO) annotations were used to classify the transcripts into functional groups in accordance with specific GO categories. A total of 12,441 (23.22%) of all predicted proteins from E. balteatus and 12,425 (24.39%) predicted proteins from E. corollae were assigned to at least one GO term (Additional file [2](#MOESM2){ref-type="media"}: Fig. S1B). The GO terms distribution in the three categories were similar in the two species. In the "molecular function" category, the most abundant GO terms were "binding" and "catalytic activity". In the "biological process" category, "cellular process", "single-organism process" and "metabolic process" were the most represented. Finally, "cell", "cell part", and "organelle" were the most abundant GO terms in "cellular component" category (Additional file [2](#MOESM2){ref-type="media"}: Fig. S1B). GO terms associated with chemosensory genes were distributed in the "biological process" category (e.g. "cellular process", "developmental process", "response to stimulus", "establishment of localization", and "biological regulation", etc.), "molecular function" category (e.g. "molecular transducer activity", etc.) and "cellular component" category (e.g. "extracellular region", "membrane part", "membrane", etc.). Candidate ORs in E. balteatus and E. corollae {#Sec5} ------------------------------------------------- Based on our analysis of the antennal transcriptomes in the two species, 51 and 42 transcripts for candidate ORs were identified in the combined male and female data sets from E. balteatus and E. corollae, respectively (Additional file [3](#MOESM3){ref-type="media"}: Table S2). A total of 21 E. balteatus ORs (EbalORs) and 29 E. corollae ORs (EcorORs) contained full-length open reading frames (ORFs), whose translation products are predicted to possess 2--8 transmembrane domains (TMDs). Other partial length transcripts encoded proteins exhibiting overlapping regions with low identity and were classified as unique genes. After a more exhaustive comparison with OR genes from other insect species, we found that all putative EbalORs shared between 22% and 86% amino acid identity with other ORs, with almost identical values (22% to 87%) for EcorORs. Detailed information is reported in Additional file [3](#MOESM3){ref-type="media"} and Table S2. We next performed a phylogenetic analysis using our candidate ORs and the ORs from four other Diptera species including B. dorsalis, C. stygia, D. melanogaster and M. domestica (Fig. [1](#Fig1){ref-type="fig"}). Clustered with DmOR83b, the highly conserved co-receptor Orco, orthologous genes were identified in the antennal transcriptomes of both syrphid species, and named EbalOrco and EcorOrco. As expected, sequence identity between EbalOrco and EcorOrco is very high (97.27%). Among the other ORs, five EbalORs (EbalOR9, 16, 18, 22 and 37) and three EcorORs (EcorOR8, 24 and 29) clustered with DmelOR67d, the pheromone receptor from D. melanogaster. This OR67d specific clade also included the OR67d orthologues from M. domestica and B. dorsalis. Two of these genes, EbalOR16 and EcorOR24, are full-length transcripts with 71.65% amino acid identity. The remaining ORs in this group were highly divergent among different species. Within the Dipteran OR sequences, we found a species-specific clade including eight members from E. balteatus (EbalOR7, 10, 30, 32, 41, 46, 47 and 48) and seven from E. corollae (EcorOR9, 10, 11, 15, 16, 34 and 38) that shared low identities with other Dipteran ORs (Fig. [1](#Fig1){ref-type="fig"}).Fig. 1Phylogenetic tree of candidate E. balteatus and E. corollae ORs and other Dipteran ORs. The distance tree was rooted by the conservative Orco gene orthologous. Bootstrap values are shown. The Orco clade, OR67d clade and specific EbalORs and EcorORs clade are shown. Species in this phylogeny include E. balteatus (Ebal, red), E. corollae (Ecor, blue), Drosophila melanogaster (Dm, dark), Bactrocera dorsalis (Bdor, green), Calliphora stygia (Csty, magenta), and Musca domestica (Mdom, gray) Amino acid identity between gene products correlates with similarity between genes. An amino acid sequence comparison between EbalORs and EcorORs revealed 33 pairs of orthologous ORs (including Orco) with 73.86% identity and the amino acid identity of complete OR ORFs was 84.00%. Sequence similarity percentages of the 28 pairs of homologous ORs are greater than 60% (Additional file [4](#MOESM4){ref-type="media"}: Table S3). In addition to the Orco family, orthologous groups with identities higher than 90% include the EbalOR29/EcorOR36, EbalOR3/EcorOR5, EbalOR13/EcorOR19 EcorOR22 with/EcorOR22 with 94.88%, 92.20%, 91.46% and 91.30% sequence identities, respectively. In addition, 10 EbalOR sequences (EbalOR2, 5, 7, 8, 11, 14, 21, 23, 25 and 33) are closely related to EcorOR homologues (EcorOR3, 4, 7, 9, 13, 14, 23, 28, 31 and 35) with identities greater than 80% (Additional file [4](#MOESM4){ref-type="media"}: Table S3; Fig. [1](#Fig1){ref-type="fig"}). All of these highly homologous proteins may play important roles in olfactory recognition. Candidate GRs in E. balteatus and E. corollae {#Sec6} ------------------------------------------------- We have identified 14 and 16 candidate GR genes from E. balteatus and E. corollae transcriptomes, respectively (Additional file [3](#MOESM3){ref-type="media"}: Table S2). The majority of candidate EbalGRs and EcorGRs were partial fragments, with only three from E. balteatus and six from E. corollae encoding full-length proteins. These complete sequences all show six or seven TMDs with an intracellular N-terminus and extracellular C-terminus. Phylogenetic analysis with GRs from six Dipteran species suggest that Drosophila GR21a and GR63a, reported as carbon dioxide sensors \[[@CR42], [@CR56]\], clustered first with EbalGR2 and EcorGR2 and second with EbalGR1 and EcorGR1. In addition, EbalGR4, EbalGR13 and EcorGR4 showed high identities to thermoreceptor DmelGR28b responsible for rapid warmth avoidance \[[@CR57]\]. Several other GRs clustered with members of candidate sugar detection GR (GR5a, GR61a and GR64a-f) sub-family (Fig. [2](#Fig2){ref-type="fig"}) \[[@CR58]--[@CR62]\].Fig. 2Phylogenetic tree of candidate E. balteatus and E. corollae GRs and other Dipteran GRs. The distance tree was rooted by the conservative carbon dioxide GRs gene orthologous. Bootstrap values are shown. The carbon dioxide GRs clade, thermos-sensed GRs clade and sugar GRs clade are shown. This tree was constructed using the species E. balteatus (Ebal, red), E. corollae (Ecor, blue), D. melanogaster (Dm, dark), B. dorsalis (Bdor, green), C. stygia (Csty, magenta), and M. domestica (Mdom, gray) Candidate IRs in E. balteatus and E. corollae {#Sec7} ------------------------------------------------- We identified 32 transcripts for putative ionotropic receptors in E. balteatus and 23 in E. corollae. Of these, seven EbalIRs and 14 EcorIRs contained full-length ORFs, with two to five TMDs (Additional file [3](#MOESM3){ref-type="media"}: Table S2). Among these we found the common conserved co-receptors IR8a (EbalIR8a and EcorIR8a) and IR25a (EbalIR25a and EcorIR25a) in both species. Other candidate IRs were found as partial sequences (Fig. [3](#Fig3){ref-type="fig"}).Fig. 3Phylogenetic tree of candidate E. balteatus and E. corollae IRs and other Dipteran IRs. The distance tree was rooted by the conservative IR25a/IR8a gene orthologues. Bootstrap values are shown. The IR25a/IR8a clade, iGluRs clade and some antennal-associated orthologue clade are shown. This tree was constructed using the following species: E. balteatus (Ebal, red), E. corollae (Ecor, blue), D. melanogaster (Dmel, dark), Anopheles gambiae (Agam, green), and C. stygia (Csty, magenta) In order to further distinguish putative IRs from iGluRs, all EbalIRs and EcorIRs were aligned with IRs from A. gambiae, C. stygia and D. melanogaster, as well as some AgamiGluRs and DmeliGluRs for phylogenetic analysis. The results showed that the candidate EbalIRs and EcorIRs clustered with presumed "antennal" orthologues IR76b, IR93a, IR21a, IR68a, IR40a, IR75l, IR75d, IR64a, IR84a, IR31a and IR92a, and were well separated from the AgamiGluRs and DmeliGluRs clade (Fig. [3](#Fig3){ref-type="fig"}) \[[@CR63]\]. Interestingly, the conserved "antennal" orthologues, IR60a, was lacking from E. balteatus and E. corollae transcriptome assemblies, while IR68a was only absent from E. corollae. The sequences of E. balteatus clustering with DmelIR94d and DmelIR94e were quite divergent (Fig. [3](#Fig3){ref-type="fig"}; Additional file [5](#MOESM5){ref-type="media"}: Fig. S2). When compared to the orthologues within other species, these IRs may play different roles in olfaction. Candidate OBPs in E. balteatus and E. corollae {#Sec8} -------------------------------------------------- We identified 49 different transcripts encoding candidate OBPs in E. balteatus and 44 in E. corollae, numbers similar to the 52 OBPs of D. melanogaster \[[@CR64]\]. Of these, 38 transcripts of EbalOBPs and 31 EcorOBPs contained full-length ORFs with predicted signal peptide sequences (with EbalOBP31 as the only exception) (Additional file [3](#MOESM3){ref-type="media"}: Table S2). A phylogenetic tree was built with these sequences and those of orthologous from B. dorsalis, C. stygia, D. melanogaster and M. domestica. Among EbalOBPs, thirty-one showed the classic motif of six conserved cysteines, three were Plus-C (EbalOBP2, 3, 4) and fifteen were Minus-C (EbalOBP5, 7, 8, 9, 11, 12, 14, 15, 16, 18, 21, 25, 26, 44 and 45) (Fig. [4](#Fig4){ref-type="fig"}). For EcorOBPs, we found 30 classic, 5 Plus-C (EcorOBP1, 3, 4, 5, 41) and 9 Minus-C (EcorOBP6, 7, 8, 10, 11, 13, 18, 36 and 37) (Fig. [4](#Fig4){ref-type="fig"}) \[[@CR64], [@CR65]\]. One large group of classic OBPs, including 15 EbalOBPs and 14 EcorOBPs showed large differences compared to sequences of other species and could represent OBPs specific of syrphids (Fig. [4](#Fig4){ref-type="fig"}; Additional file [6](#MOESM6){ref-type="media"}: Fig. S3). We found the orthologue of DmelOBP-lush in both species, EbalOBP17 and EcorOBP14. These two proteins are 98.64% identical at the amino acid level between each other and 39.87% and 40.52% identity to DmelOBP-lush.Fig. 4Phylogenetic tree of candidate E. balteatus and E. corollae OBPs and other Dipteran OBPs. The distance tree was rooted by lush gene orthologous. Bootstrap values are shown. The classic OBPs clade, Plus-C OBPs clade and Minus-C OBPs clade are shown. The species used to construct tree include E. balteatus (Ebal, red), E. corollae (Ecor, blue), D. melanogaster (Dmel, dark), B. dorsalis (Bdor, green), C. stygia (Csty, magenta), and M. domestica (Mdom, gray) Candidate CSPs in E. balteatus and E. corollae {#Sec9} -------------------------------------------------- Through bioinformatic analysis, six and seven different transcripts encoding candidate CSPs were identified from E. balteatus and E. corollae transcriptomes, respectively. Five EbalCSPs and six EcorCSPs represented full-length proteins and only EbalCSP6 lacked a signal peptide (Additional file [3](#MOESM3){ref-type="media"}: Table S2). All of the identified amino acid sequences possessed the highly conserved four-cysteine profile. A phylogenetic tree was built with all the syrphid CSPs and those of A. gambiae, C. stygia, D. melanogaster (Fig. [5](#Fig5){ref-type="fig"}).Fig. 5Phylogenetic tree of candidate E. balteatus and E. corollae CSPs and other Dipteran CSPs. The distance tree was rooted by AgamCSP1/2/3 genes. Bootstrap values are shown. The species used to construct tree include E. balteatus (Ebal, red), E. corollae (Ecor, blue), D. melanogaster (Dmel, dark), A. gambiae (Agam, green), and C. stygia (Csty, magenta) Candidate SNMPs in E. balteatus and E. corollae {#Sec10} --------------------------------------------------- In both species, two SNMPs with full-length ORFs were identified possessing two TMDs (with EbalSNMP1 having a single TMD as an exception) (Additional file [3](#MOESM3){ref-type="media"}: Table S2). EbalSNMP1 and EcorSNMP1 are very similar to DmelSNMP1, a protein shown to be required for correct pheromone detection \[[@CR50], [@CR66]--[@CR68]\]. EbalSNMP2 and EcorSNMP2 are similar to DmelSNMP2, reported to be expressed in supporting cells (Fig. [6](#Fig6){ref-type="fig"}) \[[@CR27], [@CR69], [@CR70]\].Fig. 6Phylogenetic tree of candidate E. balteatus and E. corollae SNMP and other Dipteran SNMP. Bootstrap values are shown. The species used to construct tree including E. balteatus (Ebal, red), E. corollae (Ecor, blue), D. melanogaster (Dmel, dark) and A. gambiae (Agam, green) Differentially expressed genes (DEGs) analysis {#Sec11} ---------------------------------------------- Gene expression levels of all male and female antennae-associated chemosensory genes in both E. balteatus and E. corollae were assessed using fragments per kilobase per million fragments (FPKM) values, represented in a heatmap (Fig. [7](#Fig7){ref-type="fig"}). Normalised antennal expression levels of candidate E. balteatus and E. corollae ORs are shown in Additional file [7](#MOESM7){ref-type="media"}. Of all ORs, Orco had the highest expression level of transcripts in both sexes of each species. There were no significant differences of OR transcript abundances (FPKM value) in the respective male and female antenna, except for EcorOR14 (Additional file [7](#MOESM7){ref-type="media"}). A combined analysis of false discovery rate (FDR) ≤0.001 and \\|log2 Ratio\\| ≥ 1 showed that EcorOBP shared highest number of differentially expressed genes (DEGs), including eleven high-expression in male and six high-expression in female syrphids (Additional file [7](#MOESM7){ref-type="media"}). In addition, candidate carbon dioxide receptor GR1 and GR2, and SNMP1 in both sexes showed a high expression level (Fig. [7](#Fig7){ref-type="fig"}).Fig. 7Expression profiles of chemosensory genes in E. balteatus and E. corollae. a: CSPs; b: GRs; c: IRs; d: ORs; e: OBPs; and f: SNMPs Tissue- and sex- specific expression of candidate E. balteatus and E. corollae OR genes {#Sec12} ------------------------------------------------------------------------------------------- The expression of the candidate ORs in E. balteatus and E. corollae male and female antennae and legs (control sample) was analyzed using semi-quantitative reverse transcription PCR (RT-PCR). All 51 EbalORs and 42 EcorORs were detected in the antennae at high expressing level. Only EbalOR49 was found to be mainly expressed in legs. There were no significant differences of transcript abundances in the respective male and female antenna (Fig. [8](#Fig8){ref-type="fig"}). The Orco co-receptor gene also showed a high expression level in both syrphid species. This is consistent with DEGs analysis of OR transcript abundances using the FPKM value.Fig. 8Tissue- and sex- specific expression of candidate E. balteatus and E. corollae OR genes. M: male antennae, F: female antennae, L: legs (both sexes mixed) Discussion {#Sec13} ========== The syrphids E. balteatus and E. corollae are aphid-specific predators and predominately inhabit northern China wheat and cotton fields. Typical of most insects, chemical cues drive several aspects of their behavior, such as foraging on prey and choosing oviposition sites \[[@CR7], [@CR8], [@CR10]\]. Chemosensory proteins play an important role in this process. We analyzed antennal transcriptomes of E. balteatus and E. corollae and searched for chemosensory genes with the purpose of understanding chemical communication of tritrophic interactions among plants, herbivorous insects, and natural enemies. In our study, we sequenced E. balteatus and E. corollae antennal transcriptomes using next generation sequencing technology on the Illumina HiSeq 2000 platform. The total RNA was converted into a template library for high throughput DNA sequencing, allowing us to obtain all expressed transcripts. De novo assembly of transcripts using the Trinity method gives high-efficiency and reliable full-length transcripts across extensive expression levels, even without genome information \[[@CR71]\]. Our sequence assembly yielded a final transcript dataset of 50,942 unigenes from E. corollae and 53,575 E. balteatus unigenes. Total unigenes counts resulted in 44.2% unigenes from E. balteatus and 50.3% unigenes from E. corollae shared sequence similarities to known proteins using the BLASTX homology search from the NCBI non-redundant protein database. These percentages are very similar to other Dipteran species \[[@CR50], [@CR72]\]. Remaining transcripts without associated GO terms may represent species-specific genes. The antennal transcriptome analysis proved to be a powerful tool to identify chemosensory genes in insects without genome information. It has been successfully employed in many insect orders including Lepidoptera, Coleoptera, Hymenoptera and Hemiptera. In Diptera, the chemosensory genes were successfully identified in C. stygia, B. dorsalis and Scaeva pyrastri antennal transcriptomes \[[@CR50], [@CR72], [@CR73]\]. Here, we identified 154 and 134 candidate chemosensory genes in E. balteatus and E. corollae, respectively, a number similar to other Diptera antennal transcriptomes (e.g. 128 in C. stygia) \[[@CR50]\] but less than the chemosensory genes identified in D. melanogaster (254), M. domestica (386) and A. gambiae (292) genome \[[@CR30], [@CR33], [@CR47], [@CR49], [@CR63], [@CR65], [@CR74]\]. This could be the result of differential expression based on developmental stages of the insect larva or adult olfactory organ development such as maxillary palp and proboscis. All data shows that the chemosensory genes identified by antennal transcriptome sequencing are accurate and reliable. We identified 154 candidate chemosensory genes (51 ORs, 32 IRs, 14 GRs, 49 OBPs, 6 CSPs and 2 SNMPs) in E. balteatus and 134 (42 ORs, 23 IRs, 16 GRs, 44 OBPs, 7 CSPs and 2 SNMPs) were identified in E. corollae, numbers slightly different compared with those of other Dipteran species \[[@CR50], [@CR73], [@CR75], [@CR76]\]. Such differences could be due to sequencing methods, coverage and/or depth. The number of chemosensory genes is higher in E. balteatus than in E. corollae. However, assembling and splicing quality (unigene number and N50 length) in E. corollae is better than in E. balteatus. The differences in the number and quality of transcripts identified could arise from variations in sample preparation or could be due to evolution \[[@CR77]\] and adaptation to the environment (tritrophic interactions). A total of 49 and 44 OBPs were identified in E. balteatus and E. corollae transcriptomes, respectively. The number of OBPs is variable across species, with 52 members in D. melanogaster, 66 in A. gambiae (Diptera), 21 in Apis mellifera (Hymemoptera), 34 in Helicoverpa armigera, 29 in Helicoverpa assulta (Lepidoptera), 26 in Colaphellus bowringi, 46 in Tribolium castaneum (Coleoptera) and 15 in A. pisum (Hemiptera) \[[@CR27], [@CR74], [@CR77]--[@CR83]\]. Meanwhile, the OBPs of these two syrphid species are highly divergent with those of other insects. These evolutionary differences may result from different physiological functions or ecological niches. Compared with OBPs, only a small amount of CSPs were detected in Diptera. They are only 4 CSPs in D. melanogaster, 4 in C. stygia and 7 in A. gambiae (Diptera). These numbers are much lower than in other insect orders, such as 18 CSPs found in H. armigera and 17 in H. assulta (Lepidoptera) \[[@CR27], [@CR50], [@CR80], [@CR81]\]. In our study, six EbalCSPs and seven EcorCSPs are identified in transcriptome sequencing, revealing that the numbers of CSP gene family differ among species. CSPs show a high evolutionary diversity in insecta, probably related to different physiological functions. We identified 51 ORs from E. balteatus and 42 ORs from E. corollae, respectively. Compared with other Dipteran species, these numbers are similar to those identified in C. stygia (50) \[[@CR50]\] and G. morsitans morsitans (46) \[[@CR51]\] but lower than those of D. melanogaster (62), M. domestica (86), A. gambiae (79) \[[@CR47], [@CR49], [@CR84]--[@CR86]\], suggesting that sequencing method/depth may be different between studies yielding less genes that may be difficult to detect because of low expression \[[@CR77]\]. Here, we were able to detect species-specific OR transcripts in E. balteatus and E. corollae. This clade of ORs may have a greater impact on recognizing specific odors, particularly perception of aphids-derived volatiles and herbivore-induced plant volatiles granting syrphid localization access of its prey. The tissue- and sex-specific expression analysis showed no differences between male and female, which is consistent with DEGs analysis of OR transcript abundances using FPKM values. Lepidopteran ORs have shown male-specific expression that is usually involved in the detection of the sex pheromone \[[@CR19], [@CR26], [@CR27]\], but this does not seem to be the case in syrphids. Additional real time quantitative PCR, in situ hybridization and single-sensilla recordings would be required to validate OR expressions and functions. In D. melanogaster, males release the volatile sex pheromone cis-vaccenyl acetate (cVA) \[[@CR87]--[@CR89]\]. The perception of sex pheromone cVA is mediated by OR67d \[[@CR88]\], OR65a \[[@CR90], [@CR91]\], LUSH \[[@CR92]\], and SNMP1 \[[@CR34]\]. In our two syrphid species, EbalOBP17 and EcorOBP14 are the orthologues of the DmelOBP-lush gene, while EbalOR16 and EcorOR24 are the orthologues of DmelOR67d, and EbalSNMP1 and EcorSNMP1 are very similar to DmelSNMP1, suggesting that these proteins may be involved in detection of their yet unidentified pheromones. Therefore, further functional characterization of these candidate proteins will help reveal any mechanism associated with pheromone reception in E. balteatus and E. corollae. The E. balteatus and E. corollae IR family is relatively conserved, especially with respect to common receptors IR8a and IR25a, which are expressed in both olfactory and gustatory systems \[[@CR30], [@CR45]\]. The numbers of IRs identified in E. balteatus (32) and E. corollae (23) are similar to that of C. stygia (22) \[[@CR50]\], but lower than those of D. melanogaster (66) and A. gambiae (46) \[[@CR63]\]. It is possible that some IRs do not express in antennae tissues or perhaps the number of IRs varies between species and is dependent on natural habitats. A large number of EbalIRs and EcorIRs are clustered with "antennal" orthologues in Drosophila, indicating that IRs are highly conserved in Diptera. Furthermore, the IRs identified in these two species may be activated by acids, amines and other odorants that are not sensed by ORs \[[@CR30], [@CR45], [@CR46]\]. In the antennae of E. balteatus and E. corollae, we identified 14 and 16 candidate GRs, respectively. The total number of GRs in these two species may be much larger, because some members could be exclusively expressed in other gustatory organs, such as maxillary palps, proboscises and legs. However, the numbers are still lower than those reported in other Dipteran antennal transcriptomes \[[@CR50]\]. The conserved receptors identified in the two syrphid species may be involved in CO~2~ perception. However, we infer that the mechanism of CO~2~ perception is different from mosquitoes which concerns host-seeking \[[@CR43], [@CR44], [@CR48], [@CR93]\]. Some GRs may function as taste or contact receptors \[[@CR31]\], particularly with reference to their specific pollination behavior \[[@CR94], [@CR95]\]. Some GRs from these two species are clustered with thermos-sensing GRs and sugar-detecting GRs from Drosophila, indicating that they may perform similar functions. Functional analysis of the candidate E. balteatus and E. corollae chemosensory proteins is required to identify their physiological roles. Conclusions {#Sec14} =========== We have identified and annotated 154 transcripts encoding putative chemosensory proteins in antennal transcriptome of E. balteatus and 134 in E. corollae. Comparisons between the two syrphid species and among other Dipteran species were deduced using sequence information. This work gives a foundation for future studies aimed at understanding chemical communication in syrphids and tritrophic interactions between plants, herbivorous insects, and natural enemies in agricultural ecosystems. Methods {#Sec15} ======= Insect rearing and tissue collection {#Sec16} ------------------------------------ E balteatus and E. corollae larvae were fed with aphids (Aphis gossypii Glover) and maintained at 22 ± 1 °C with a 12 h light: 12 h dark photo-period at the Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing, China. Following eclosion, adult males and females were separated and provided with pollen and 10% honey solution. Antennae were excised from 2- to 5-day-old adult males and females respectively, and legs were collected together, then immediately frozen and stored in liquid nitrogen. cDNA library construction and Illumina sequencing {#Sec17} ------------------------------------------------- Total RNA of male and female antennae was extracted from E. balteatus and E. corollae using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). The method for RNA extraction followed in the manufacturer's instruction. Total RNA was dissolved in RNase-free water and RNA integrity was verified by gel electrophoresis. RNA concentration and purity were measured on a Nanodrop ND-2000 spectrophotometer (NanoDrop products, Wilmington, DE, USA). Ten micrograms total RNA of each sample was used to construct the cDNA library. The cDNA library construction and Illumina HiSeq 2000 (Illumina, San Diego, CA, USA) sequencing of the samples was performed at Beijing Genomics Institute (BGI, Shenzhen, China). The insert sequence length was around 200 bp and these libraries were pair-end sequenced using PE100 strategy \[[@CR22], [@CR27]\]. Assembly and function annotation {#Sec18} -------------------------------- Raw reads were pre-processed by filtering low quality reads, trimming low quality nucleotides at each ends and removing 3′ adaptors and poly-A/T tails. Each clean-read dataset of male and female antenna was fed to Trinity \[[@CR71]\]. The Trinity assembly procedure, including Inchworm, Chrysalis and Butterfly were followed using Grabherr et al., 2011 as a reference \[[@CR71]\]. In the first step of Trinity, Inchworm assembles reads into the unique sequences of transcripts using the default parameters (default k-mers = 25). Next, Chrysalis clusters related contigs that correspond to portions of alternatively spliced transcripts or otherwise unique portions of paralogous genes. Finally, Butterfly uses read sequences, read-pairings and Chrysalis' read mappings to select the paths that are best supported by read sequences \[[@CR71]\]. The Trinity outputs were clustered by TGICL \[[@CR96]\]. The consensus cluster sequences and singletons make up the unigenes dataset \[[@CR22]\]. The unigenes annotation was performed by NCBI BLASTX against a pooled database of non-redundant and SwissProt protein sequences with e-value \<1e-5. The BLASTX results were then imported into Blast2GO pipeline for GO annotation \[[@CR97]\]. Identification of chemosensory genes {#Sec19} ------------------------------------ Candidate unigenes encoding putative ORs, IRs, OBPs, CSPs, SNMPs and GRs were found by running Perl scripts against transcriptome assembly and annotation in the remote sever. Perl scripts were written to extract sequence from functional annotation results using olfaction keywords. Subsequently, all candidate chemosensory genes were manually checked by BLASTX against local non-redundant database with e-value \<1e-5. Using the BLASTX NCBI database, we manually performed alignments comparing transcripts against all known proteins to examine full-length coverage. The full-length transcripts contain start and termination codons. The ORFs of all putative chemosensory genes were predicted by using ExPASy (Expert Protein Analysis System) server version ([http://web.expasy.org/translate/](http://web.expasy.org/translate)) according to the BLASTX best hit result \[[@CR98]\]. Putative N-terminal signal peptide of OBPs and CSPs were predicted by SignalP 4.0 server version with default parameters \[[@CR99]\]. The TMDs of ORs, IRs and GRs were predicted using TMHMM server version 2.0 \[[@CR100]\]. Sequence and phylogenetic analysis {#Sec20} ---------------------------------- After removing redundancy, alignments of amino acid sequences were performed by MAFFT ([https://www.ebi.ac.uk/Tools/msa/mafft/](https://www.ebi.ac.uk/Tools/msa/mafft)). The phylogenetic trees of E. balteatus and E. corollae chemosensory genes were constructed by RaxML version 8 with Jones-Taylor-Thornton amino acid substitution model (JTT) \[[@CR101]\] with the putative chemosensory genes in other Dipteran species (Additional file [8](#MOESM8){ref-type="media"}: Table S4). Node support was assessed using a bootstrap method based on 1000 replicates. The OR data set contained OR sequences identified in Dipteran (51 from E balteatus, 42 from E. corollae, 62 from D. melanogaster \[[@CR85], [@CR86]\], 61 from B. dorsalis \[genome: assembly ASM78921v2\], 50 from C. stygia \[[@CR50]\] and 81 from M. domestica \[[@CR49]\]). The GR data set contained GR sequences identified in Dipteran (14 from E balteatus, 16 from E. corollae, 68 from D. melanogaster \[[@CR47]\], 40 from B. dorsalis \[genome: assembly ASM78921v2\], 21 from C. stygia \[[@CR50]\] and 43 from M. domestica \[[@CR49]\]). The IR data set contained IR sequences identified in Dipteran (32 from E balteatus, 23 from E. corollae, 76 from D. melanogaster \[[@CR30], [@CR63]\], 22 from C. stygia \[[@CR50]\] and 54 from A. gambiae \[[@CR63], [@CR84]\]). The OBP data set contained OBP sequences identified in Dipteran (49 from E balteatus, 44 from E. corollae, 71 from D. melanogaster \[[@CR64]\], 40 from B. dorsalis \[genome: assembly ASM78921v2\], 28 from C. stygia \[[@CR50]\] and 52 from M. domestica \[[@CR49]\]). The CSP data set contained CSP sequences identified in Dipteran (7 from E balteatus, 9 from E. corollae, 4 from D. melanogaster \[[@CR74]\], 4 from C. stygia \[[@CR50]\] and 8 from A. gambiae \[[@CR74]\]). The SNMP data set contained SNMP sequences identified in Dipteran (2 from E balteatus, 2 from E. corollae, 2 from D. melanogaster \[[@CR34], [@CR35]\] and 2 from A. gambiae \[[@CR35]\]). DEGs analysis {#Sec21} ------------- A mapping-based expression profiling analysis of the chemosensory genes was conducted to compare gene expression between male and female antennae. All of the clean reads were remapped onto the transcripts using SOAPaligner (<http://soap.genomics.org.cn> /soapaligner.html), allowing up to three base mismatches and a minimum length of 40 bp. The FPKM method was used for calculating unigene expression levels \[[@CR20], [@CR50], [@CR102], [@CR103]\]. The suitable P-values were calculated to identify differentially expressed genes according to the hypergeometric test \[[@CR103]\]. The FDR was a statistical method used in multiple hypothesis testing to correct for P-value. Criteria for estimating significant differential expression was set at FDR ≤ 0.001 and \\|log2 Ratio\\| ≥ 1. Heatmaps of differential gene expression between male antennae and female antennae in both species were generated by Heml 1.0 software \[[@CR104]\]. Expression analysis by semi-quantitative RT-PCR {#Sec22} ----------------------------------------------- Semi-quantitative RT-PCR was performed to verify the expression of candidate chemosensory genes. Male and female antennae and legs were collected from adult E. balteatus and E. corollae after eclosion. The extraction of total RNA followed the manufacturer's instruction \[[@CR27]\]. The cDNA was synthesized from total RNA using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Waltham, MA, USA). Gene specific primers were designed using PrimerQuest Tool (<http://sg.idtdna.com/Primerquest/Home/Index>) (Additional file [9](#MOESM9){ref-type="media"}: Table S5) and synthesized by Sangon Biotech Co., Ltd. (Shanghai, China). A Taq MasterMix (CWBIO, Beijing, China) was used for PCR reactions under the general three-step amplification of 94 °C for 30s, 55 °C for 30s, 72 °C for 30s. RT-PCR products were separated on 2% agarose gels, stained by ethidium bromide (EB), and photographed under UV light in Gel Doc XR+ Gel Documentation System with Image Lab Software (Bio-Rad, Hercules, CA, USA). Additional files ================ {#Sec23} Additional file 1: Table S1.Assembly summary of E. balteatus and E. corollae antennal transcriptome. (DOCX 16 kb) Additional file 2: Fig. S1.(A) Species distribution and annotation summaries in the E. balteatus (Ebal) and E. corollae (Ecor) antennal transcriptome assembly. (B) Gene ontology classifications of the E. balteatus and E. corollae unigenes with Blast2GO program, including categories with biological process, molecular function and cellular component. (TIFF 3397 kb) Additional file 3: Table S2.Candidate E. balteatus and E. corollae antennal chemosensory genes. Unigenes of candidate odorant receptors (2--1), gustatory receptors (2--2), ionotropic receptors (2--3), odorant binding proteins (2--4), chemosensory proteins (2--5) and sensory neuron membrane proteins (2--6) with gene name, length, ORF, best BLASTX hit and identity. (DOCX 112 kb) Additional file 4: Table S3.Comparison of homologous ORs in E. balteatus and E. corollae. (DOCX 61 kb) Additional file 5: Fig. S2.Protein domain analysis of the species-specific IR clade with Drosophila iGluRs and DmelIR94d /e. Amino acid alignments shows the ligand binding domains (S1 and S2), the ion channel pore (P), and TMD (M1, M2 and M3) of ionotropic receptors. The key ligand binding residues are marked in red box. (JPEG 4154 kb) Additional file 6: Fig. S3.Amino acid alignments of the species-specific OBPs clade in the E. balteatus and E. corollae. The motif of six conserved cysteines are marked with asterisks at the top. (JPEG 3067 kb) Additional file 7:Antennal expression levels of candidate E. balteatus and E. corollae odorant receptors. (XLSX 99 kb) Additional file 8: Table S4.GenBank accession numbers of chemosensory genes used in phylogenetic analyses. (XLSX 34 kb) Additional file 9: Table S5.Primers of candidate ORs in E. balteatus and E. corollae used for RT-PCR. (DOCX 20 kb) cDNA : Complementary DNA CSP : Chemosensory protein DEGs : Differentially expressed genes FDR : False discovery rate FPKM : Fragments per kilobase per million fragments GLV : Green leaf volatiles GO : Gene ontology GR : Gustatory receptor IR : Ionotropic receptor OBP : Odorant-binding protein OR : Odorant receptor ORF : Open reading frame PCR : Polymerase chain reaction RT-PCR : Semi-quantitative reverse transcription PCR SNMP : Sensory neuron membrane protein TMD : Transmembrane domain Electronic supplementary material The online version of this article (doi:10.1186/s12864-017-3939-4) contains supplementary material, which is available to authorized users. We thank Ph.D. students Kang He, Song Cao, Rui-Bin Zhang and M.S. student Yan-Wei Wang (Institute of Plant Protection, Chinese Academy of Agricultural Sciences) for help collecting insects. We thank Ms. Chun-Yan Wang and Ms. Li-Yan Yang (Institute of Plant Protection, Chinese Academy of Agricultural Sciences) for rearing insects. We are grateful to thank M.S. Xiang-Zhi Liang, Dr. Yong-Lei Liu and Dr. Ya-Nan Zhang for technical assistance. Thanks to Prof. Paolo Pelosi for editorial assistance and comments on the manuscript. Funding {#FPar1} ======= This project is supported by National Natural Science Foundation of China (31,621,064; 31,230,062&31,321,004 to GW, 31,402,023 to BW, 31,471,833 to YL), China Postdoctoral Science Foundation (2014 M550905 to BW), China Postdoctoral Science Foundation 2017T100121 to BW. Availability of data and materials {#FPar2} ================================== The clean reads of the four antennal transcriptomes in this study have been stored in the NCBI SRA database, under the accession number of SRX1812754 (E. balteatus male), SRX1812787 (E. balteatus female), SRX1829150 (E. corollae male), and SRX1829282 (E. corollae female). All authors contributed to research design and manuscript preparation. Conceived and designed the experiments: BW, YL, GW. Performed the experiments: BW. Analyzed the data: BW, YL. Contributed reagents/materials/analysis tools: BW, YL, GW. Wrote the paper: BW, YL, GW. All authors read and approved the final manuscript. Competing interests {#FPar3} =================== The authors declare that they have no competing interests. Publisher's Note {#FPar4} ================ Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.	{ "pile_set_name": "PubMed Central" }
INTRODUCTION ============ Psoriasis is a chronic immune-mediated inflammatory skin disease that is not explained simply by Mendelian inheritance. Genetic factors play a role in psoriasis development, although the exact causal mechanism of psoriasis is still unknown.^[@r1]^ Psoriasis is seen in nearly 1-3% of the human population all over the world and at every age. Its adverse effect on health is seriously aggravated by physical disruption of daily activities, shame, anxiety, guilt, social withdrawal, social exclusion, depression, and even suicidal tendencies. Psoriasis patients can also experience heart disease, diabetes mellitus, cancer, and depression. A chronic course involving unexplained remissions and exacerbations requires patients to undergo a treatment and follow-up process that extends throughout their lives.^[@r2]-[@r4]^ Psoriasis usually appears in the form of well-demarcated, scaly plaques and has been classified in the group of immune-mediated inflammatory diseases, alongside ulcerative colitis, Crohn\'s disease, rheumatoid arthritis, ankylosing spondylitis, and Behçet\'s disease.^[@r5],[@r6]^ It is known that antimicrobial peptides (AMPs), which provide high resistance to skin infection, show increased synthesis in psoriatic lesions.^[@r7]^ AMP production in the skin is the primary mechanism that protects skin against infections, and AMP levels are increased as a result of microbial infections.^[@r8]^ Granulysin is a member of the AMP family of cytolytic and proinflammatory peptides. The peptide exists in granules of T cells and natural killer (NK) cells and is secreted from these cells together with granzyme and perforin.^[@r9]^ Based on amino acid sequence homology, the 9-kDa granulysin protein belongs to the saposin-like protein (SAPLIP) family, which consists of activators of sphingolipid hydrolase in the central nervous system. Granulysin is tumoricidal and broadly antimicrobial, killing gram-positive and gram-negative bacteria, yeast, fungi, and parasites. Granulysin binds to the cell surface based on charge and appears to make a way through the cell membrane, causing ion fluxes.^[@r10]^ A recent immunohistochemistry study showed that granulysin expression is increased in psoriasis lesions.^[@r7]^ Granulysin encoding by the GNLY gene is located on 2 p11.2, and several polymorphisms have been identified in GNLY that contain 5 exons. C to T substitution in GNLY (rs10180391 (C/T) is located on 2: 85.699.222, and C to G substitution on GNLY (rs 7908 (C/G) is located on 2:85.698.633.^[@r11]^ It has also been shown that GNLY gene polymorphism is associated with infectious diseases such as chronic hepatitis B, but to our knowledge there has been no study on the relationship between GNLY gene polymorphisms and psoriasis.^[@r12]^ Our study investigated the role of GNLY gene polymorphisms in the etiopathogenesis of psoriasis and can help establish a psoriasis susceptibility profile. METHODS ======= Study subjects -------------- The study was conducted in 100 patients diagnosed with psoriasis in the Polyclinic for Skin and Venereal Diseases of Bulent Ecevit University Application and Research Center. A group of 100 individuals with no autoimmune or autoinflammatory disorders was used as the control group. Study participants were matched for gender and age distribution. The age and gender of all the subjects included in the study were recorded, and patients were divided into groups according to the Psoriasis Area and Severity Index (PASI) score. The study was approved by the Clinical Research Ethics Committee, Bulent Ecevit University (Date 28.05.2015 approval number 2015-24-26/05) and performed in accordance with the ethical standards established in the 1964 Declaration of Helsinki and later amendments. All participants were informed and signed consent forms before being included. DNA extraction and genotype analyses ------------------------------------ In order to study the GNLY rs7908 and rs10180391 gene polymorphisms, 2ml blood samples were collected from patients and healthy controls during routine follow-up. Genomic DNA was extracted from 250µL of peripheral blood with the E.Z.N.A.^®^ Blood DNA Mini Kit (Omega Bio-tek Inc., Atlanta GA), according to the manufacturer\'s protocol. Polymerase chain reaction (PCR) restriction fragment length polymorphism (RFLP)-based analysis was used to genotype GNLY rs7908 (C/G) and GNLY rs10180391(C/T) polymorphisms using PCR primers ([Tables 1](#t1){ref-type="table"} and [2](#t2){ref-type="table"}). ###### SNP ID, primer sequences, and annealing temperature ----------------------------------------------------------------------------- SNP ID Primer sequencing Annealing temp. (^0^C) ------------------------- -------------------------- ------------------------ rs7908 F: 5\'- TGT TCA GTA\ 56 0C GGG TCA GGT GG-3\' R: 5\'-GAT TCT GGA\ TCG AGG AAG CG -3\' rs10180391 F: 5\'- AAG CAA CAG\ 55 0C AAG TCT CAG CC-3\' R: 5\'-CTG CCA AGA\ GAG AAG CGA AC -3\' ----------------------------------------------------------------------------- ###### SNP ID, restriction enzymes for detecting each single-nucleotide polymorphism (SNP), cutting sequencing, working temperature, incubation time and allele size (bp) SNP ID Restriction enzymes Cutting sequencing Working temperature Incubation time Allele size (bp) ---------------- --------------------- -------------------- --------------------- ----------------- --------------------- 5\'-CC\^TCAGC-3\' GG: 225 bp rs7908 BbvC1 37°C 16 hours GC: 225, 183, 42 bp 3\'-GGAGT\^CG-5\' CC: 183, 42 bp 5\'-AT\^TAAT-3\' CC: 209 bp rs10180391 Ase1 37°C 16 hours CT: 209, 145, 64 bp 3\'-TAAT\^TA-5\' TT: 145,64 bp PCR was performed in 25µL containing 20 to 100ng genomic DNA, 1.5mM MgCl~2~, 1x PCR buffer with (NH~4~)~2~SO~4~, 0.25mM deoxynucleotide triphosphates (dNTPs), 10% dimethylsulphoxide, 0.5 units of Taq polymerase, and 20pmol of each primer. Amplification conditions for the GNLY rs7908(C/G) polymorphism were initial denaturation at 95°C for 3 minutes, then 35 amplification cycles of 60s at 95°C, 90s at 56°C, and 60s at 72°C, and a final 7-minute extension step at 72°C. The PCR products were checked on 1.5% agarose gel for assay completion, and then the PCR products with 225 base pairs (bp) were digested overnight with BbvCI restriction enzyme (New England Biolabs) ([Figure 1](#f1){ref-type="fig"}). Figure 1GNLY rs7908 gene poly morphism gel electrophoresis image. DNA marker is 100 bp. Samples with 3,4 are GC; samples with 1, 6 are CC and 2, 5 with GG genotype Amplification conditions for the GNLY rs10180391(C/T) polymorphism were initial denaturation at 95°C for 3 minutes, then 35 amplification cycles of 95°C for 60s, 55°C for 90s, and 72°C for 60s, and a final 7-minute extension step at 72°C. The PCR products were checked on 1.5% agarose gel for assay completion, and then the PCR products with 209 base pairs (bp) were digested overnight with AseI restriction enzyme (Thermo Scientific) ([Figure 2](#f2){ref-type="fig"}). Figure 2GNLY rs10180391 gene poly morphism gel electrophoresis image. DNA marker is100 bp. Samples 2,4,5 CT; samples 1 and 6 are in CC and 3 are in TT genotype All digestion products were electrophoresed on 3% agarose gel and viewed by staining with ethidium bromide and evaluated using a gel documentation system (Vilber Lourmat Bio-Print ST4). Statistical analyses -------------------- A case-control study was performed, and allelic frequencies of the polymorphisms were calculated both for cases and controls. The χ^2^ test was used to compare the genotype frequency of each gene polymorphism in psoriasis patients and controls. Odds ratios (OR) and 95% confidence intervals (CI) were calculated to compare the psoriasis risk for the alleles. Distribution of data was determined by the Shapiro-Wilks test. Continuous variables are expressed as mean and standard deviation (SD) or median, and categorical variables as frequency and percent. Continuous variables were compared with the independent-sample t-test or the Mann-Whitney U-test. P values \<0.05 were considered statistically significant. SPSS version 18.0 was used (SPSS Inc., Chicago, IL). RESULTS ======= The study enrolled 100 psoriasis patients and 100 healthy controls. There was no statistically significant difference in age distribution between patients and controls (p = 0.315). Mean age (SD) was 41.31 years (13.46) in patients and 39.49 years (12.06) in controls ([Table 3](#t3){ref-type="table"}). ###### Sex and age distribution of psoriasis patients and controls Healthy controls n (%) Psoriasis patients n (%) P value ------------ ------------------------ -------------------------- ----------- Male 43(43%) 44(44%) Female 57(57%) 5656(%) Age 39,49±12.06 41,31±13.46 0.315 As seen in [table 4](#t4){ref-type="table"}, the study showed a significant association between the GNLY rs7908 (C/G) polymorphism and psoriasis versus normal controls (OR=0.305 95% CI=0.121 - 0.773; p=0.033). The GNLY rs7908 CC genotype was significantly more frequent when compared to the control group (GG genotype) (p=0.033, OR=0.305 CI=0.121 - 0.773). Distributions of genotypes of the GNLY rs10180391 (C/T) polymorphism were similar in psoriasis patients and controls, and no relationship was detected between these groups (p \> 0.05). No statistically significant difference was found between patients and controls for GNLY rs10180391 allele frequencies (p = 0.604, OR = 1.114, CI = 0.741-1.627). Neither was there any statistically significant difference in GNLY rs7908 (C/G) and GNLY rs10180391 (C/T) gene polymorphisms in terms of psoriasis severity (PASI) ([Graph 1](#f3){ref-type="fig"}). In addition, there was no statistically significant association between GNLY rs7908 (C/G) and rs10180391 (C/T) gene polymorphisms in terms of haplotype ([Table 5](#t5){ref-type="table"}). ###### GNLY genotypes and alleles and the risk of developing psoriasis (n=100) All PASI\<3 3≤PASI\<1 PASI≥10 OR (95% CI) P ------------------------- ------------ ------------ ------------ ------------ ----------- -------------------- ------- Rs7908/Genotype GG 37 (37%) 51 (51%) 28(46.7%) 17 (53.1%) 6 (75%) Reference 0.033 GC 44 (44%) 41 (41%) 25(41.7%) 14 (43.8%) 2 (25%) 0.676(0.371-1.232) CC 19 (19%) 8 (8%) 7(11.7%) 1 (3.1%) 0 (0.0%) 0.305(0.121-0.773) Rs7908/Allele G 118 (59%) 143(71.5%) 81 (67.5%) 48 (75%) 14(87.5%) Reference 0.009 C 82 (41%) 57 (28.5%) 39(32.5%) 16 (25%) 2 (12.5%) 0.574(0.378-0.870) rs10180391/Genotype CC 44(44%) 41 (41%) 28(46.7%) 11 (34.4%) 2 (25%) Reference 0.886 CT 41 (41%) 42 (42%) 22(36.7%) 15 (46.9%) 5 (62.5%) 1.099(0.600-2.013) TT 15 (15%) 17 (17%) 10(16.7%) 6 (18.8%) 1 (12.5%) 1.216(0.539-2.746) rs10180391/Allele C 129(64.5%) 124 (62%) 78(65%) 37(57.8%) 9 (56.3%) Reference 0.604 T 71 (35.5%) 76 (38%) 42(35%) 27(42.2%) 7 (43.7%) 1.114(0.741-1.672) Graph 1The relationship between PASI score and GNLY genotypes ###### Haplotype distribution of GNLY rs7908 - rs10180391 gene polymorphisms ------------------------------------------------------------------------------- Haplotype Healthy controls Psoriasis patients P value OR (95% CI) ----------- ------------------ -------------------- ----------- --------------- GC 39 % 42.5 % 0.036 Reference GT 20 % 29 % 0.269 1.331\ (0.802-2.208) CC 25.5 % 19.5 % 0.180 0.702\ (0.418-1.178) CT 15.5 % 9 % 0.060 0.533\ (0.276-1.028) ------------------------------------------------------------------------------- DISCUSSION ========== Psoriasis is a dermatosis characterized by sharply demarcated erythematous, pearly-white scales whose etiology is not clearly known.^[@r5]^ According to current knowledge, genetic factors contribute to age at onset, clinical manifestations, and type and severity of the disease. It has been hypothesized that genetic differences may lead to the development of the disease (PSORS2) and/or affect the initial antigen presentation (PSORS1 associated with the HLA-Cw6 locus).^[@r13],[@r14]^ The most widely accepted mechanism in the etiopathogenesis of psoriasis involves chemokines, which play a key role in cytokine and lymphocyte activation as the fundamental step in this process, followed by keratinocyte hyperproliferation and inflammation.^[@r15],[@r16]^ Granulysin, a positively charged cytolytic antimicrobial peptide (AMP) class found in the granules of cytotoxic T lymphocytes (CTL) and natural killer cells (NK), was first described during the investigation of genes expressed in effector T lymphocytes.^[@r17]^ Granulysin has been associated with a number of diseases, including various infections, cancer, transplantation, skin disorders, and reproductive complications.^[@r18]^ We investigated the relationship between psoriasis, an autoinflammatory disease, and rs7908 (G/C) and rs10180391 (C/T) gene polymorphisms in the GNLY gene. We found no study in the literature on the relationship between psoriasis and GNLY gene polymorphisms. The role of GNLY gene polymorphisms has been investigated in chronic liver disease and HIV infection.^[@r12],[@r19]^ The difference between the two groups in the genotype distribution of GNLY rs7908 (C/G) polymorphism was statistically significant. CC genotype was significantly more frequent than GG genotype in the control group. In terms of allelic frequencies, the C allele has been found to have a protective effect against psoriasis. Due to the alternative splicing mechanism from the GNLY gene, four different protein variants are synthesized in the same cell. In GNLY rs7908 (C/G) variants 1, 2, and 3, this polymorphism occurs in the 3\'UTR region.^[@r20]^ Stabilization of the mRNA, exocytosis, and subcellular localization are affected by 3\'UTR. Therefore, the gene expression is post-transcriptionally regulated.^[@r21]^ Nucleotide changes in the 3\'UTR region lead to serious disruption.^[@r22]^ In the variant GNLY rs 7908, C/G nucleotide exchange is located in the protein-coding region, causing the amino acid transformation of leucine → valine (Leu150Val). Therefore, it is thought that GNLY rs7908 C allele reduces the amount or activity of granulysin protein, leading to the complete elimination of trigger factors causing psoriasis. The amino acid change imposed by the rs7908 SNP can be considered as a reason of activity decrease of the enzyme. Although both leucine and valine are nonpolar and hydrophobic, valine differs structurally from leucine, containing an extra methylene (-CH~2~) group. Most likely, the absence of the methyl group causes distortions in intramolecular interactions and decreases in enzyme activity. When the GNLY rs10180391 gene polymorphism genotype distribution and allele frequencies were analyzed, no statistically significant difference was found between the patient and control groups. In contrast, Hou et al. (2015) found a statistically significant association between GNLY rs1866139 and GNLY rs11127 and HBV infection.^[@r12]^ This may be due to the fact that the GNLY rs1866139 and GNLY rs11127 polymorphisms are not effective in protein activity or amount. When the patient group was classified as PASI\<3, 3 ≤ PASI\<10, PASI≥10 according to the PASI severity score, there was no statistically significant change in terms of the two polymorphisms. However, CC ratio was decreased in GNLY rs7908 polymorphism as the PASI score increased with GG and GC. In other words, CC genotype frequency decreased as disease severity increased. This decrease was not statistically significant because of the low number of cases. The study showed that GNLY rs7908 G allele is associated with disease severity and that the C allele protects against psoriasis. The increase in psoriasis severity associated with the GNLY rs7908 G allele in our study suggests that this allele may have an enhancing effect on granulase expression. Similarly, Elgarhy et al. (2015) found that the severity of psoriasis increased with the increase in granulysin expression. They corroborated our findings, and their immunohistochemical analyses showed that granulysin expression is increased in psoriasis lesions.^[@r7]^ Raychaudhuri et al. (2004) found increased granulysin levels in lesional T cells and dermal dendrocytes in psoriasis plaques.^[@r23]^ Pivarcsi et al. (2004) showed that lichen planus (LP) granulysin was expressed 100-200 times above the normal level.^[@r24]^ Granulysin has been identified as an important effector in both Stevens-Johnson syndrome (SJS) and diffuse keratinocyte death in toxic epidermal necrolysis (TEN).^[@r25]^ Tewary et al. (2010) demonstrated that granulysin promotes CTL proliferation by acting as an alarmin with chemoattractant and stimulatory effects on dendritic cells and macrophages.^[@r26]^ Our study suggests that the rs7908 C/G polymorphism in the GNLY gene may also reduce the alarmin function of granulysin, thus allowing the cells of the immunopathogenesis of psoriasis to continue to function normally. Our study analyzed GNLY rs7908 and GNLY rs10180391 SNPs and found that the C allele of rs7908 is protective against psoriasis and even inversely proportional to psoriasis severity. There are very few studies in the literature that examine the interactions between GNLY gene polymorphisms and immunological diseases. There is thus serious need for further research to identify the effect of granulysin in psoriasis. Psoriasis prevalence is reported at 1.5-2.8% in the United Kingdom, 2.5-3.5% in Germany, 3.15% in the United States, and 0.05-1.23% in China, and the highest prevalence in Asia was reported at 11.8%.^[@r27]^ Since prevalence of the disease differs between populations, our research in the Turkish population can be useful for elucidating the role of GNLY in psoriasis. CONCLUSION ========== We found that genotype GNLY rs7908 CC and C allele had a protective effect against psoriasis and reduced the severity of the disease. The rs10180391 SNP did not play a significant role in the pathogenesis of psoriasis. Investigation of the effect of GNLY rs7908 and GNLY rs10180391 gene polymorphisms on psoriasis risk and disease progression can lead to the development of new treatment modalities for psoriasis. Acknowledgments: The research was supported by the Unit of Scientific Research Projects of Bulent Ecevit University (BAP-2015-50737594-04), Turkey. Work conducted at the Bulent Ecevit University, Zonguldak, Turkey. Financial Support: The research was supported by the Unit of Scientific Research Projects of Bulent Ecevit University (BAP-2015-50737594-04), Turkey. Conflict of Interests: None. AUTHORS\'CONTRIBUTIONS Esra Ermis 0000-0001-6233-2420 Conception and planning of the study; Obtaining, analyzing, and interpreting the data; Elaboration and writing of the manuscript; Effective participation in research orientation; Critical review of the literature Sevim Karakas Celik 0000-0003-0505-7850 Statistical analysis; Approval of the final version of the manuscript; Conception and planning of the study; Critical review of the literature; Critical review of the manuscript Nilgun Solak 0000-0002-6572-9615 Conception and planning of the study Gunes Cakmak Genc 0000-0001-7222-0377 Conception and planning of the study; Obtaining, analyzing, and interpreting the data Ahmet Dursun 0000-0002-7625-837X Approval of the final version of the manuscript; Critical review of the literature	{ "pile_set_name": "PubMed Central" }
![](hosplond73813-0005){#sp1 .137} ![](hosplond73813-0006){#sp2 .138}	{ "pile_set_name": "PubMed Central" }
Related literature {#sec1} ================== For the synthesis and properties of quinolone derivatives, see: Anderson & Osheroff (2001[@bb1]); Ball et al. (1998[@bb2]); Hong et al. (1997[@bb5]); Ray et al. (2005[@bb6]); Wang et al. (2008[@bb9]). Experimental {#sec2} ============ {#sec2.1} ### Crystal data {#sec2.1.1} C~18~H~33~N~3~O~5~M ~r~ = 371.47Monoclinic,a = 28.867 (3) Åb = 6.1887 (13) Åc = 25.379 (3) Åβ = 112.769 (2)°V = 4180.6 (11) Å^3^Z = 8Mo Kα radiationμ = 0.09 mm^−1^T = 298 (2) K0.40 × 0.20 × 0.11 mm ### Data collection {#sec2.1.2} Bruker SMART CCD area-detector diffractometerAbsorption correction: multi-scan (SADABS; Sheldrick, 1996[@bb7]) T ~min~ = 0.963, T ~max~ = 0.99110032 measured reflections3699 independent reflections1915 reflections with I \> 2σ(I)R ~int~ = 0.060 ### Refinement {#sec2.1.3} R\[F ^2^ \> 2σ(F ^2^)\] = 0.050wR(F ^2^) = 0.144S = 1.023699 reflections244 parametersH-atom parameters constrainedΔρ~max~ = 0.23 e Å^−3^Δρ~min~ = −0.22 e Å^−3^ {#d5e367} Data collection: SMART (Bruker, 1998[@bb3]); cell refinement: SAINT (Bruker, 1999[@bb4]); data reduction: SAINT; program(s) used to solve structure: SHELXS97 (Sheldrick, 2008[@bb8]); program(s) used to refine structure: SHELXL97 (Sheldrick, 2008[@bb8]); molecular graphics: SHELXTL (Sheldrick, 2008[@bb8]); software used to prepare material for publication: SHELXTL. Supplementary Material ====================== Crystal structure: contains datablocks global, I. DOI: [10.1107/S1600536808044255/rk2125sup1.cif](http://dx.doi.org/10.1107/S1600536808044255/rk2125sup1.cif) Structure factors: contains datablocks I. DOI: [10.1107/S1600536808044255/rk2125Isup2.hkl](http://dx.doi.org/10.1107/S1600536808044255/rk2125Isup2.hkl) Additional supplementary materials: [crystallographic information](http://scripts.iucr.org/cgi-bin/sendsupfiles?rk2125&file=rk2125sup0.html&mime=text/html); [3D view](http://scripts.iucr.org/cgi-bin/sendcif?rk2125sup1&Qmime=cif); [checkCIF report](http://scripts.iucr.org/cgi-bin/paper?rk2125&checkcif=yes) Supplementary data and figures for this paper are available from the IUCr electronic archives (Reference: [RK2125](http://scripts.iucr.org/cgi-bin/sendsup?rk2125)). This work was supported by the IMB Research Foundation. Comment ======= Quinolones, a class of synthetic antibacterial compounds based on a 4--quinolone skeleton, have been the landmark discovery in the treatment of bacterial infections in both community and hospital setting (Ray et al., 2005; Ball et al., 1998;). The most intensive structural variations have been carried out on the basic group at the C--7 position, partially due to the ease of their introduction through a nucleophilic aromatic substitution reaction on the corresponding halide. Piperazine, piperidine, pyrrolidine and their derivatives have been the most successfully employed side chains, as evidenced by the compounds currently on the market (Anderson & Osheroff, 2001; Hong et al., 1997). Recently, as part of an ongoing program to find potent new quinolones displaying strong Gram--positive activity, we have focused our attention on introducing new functional groups to the piperidine ring (Wang et al., 2008). We report here the crystal structure of the title compound, which is a key intermediate of 4--methoxyimino--3--methylamino--3--methylpiperidine, a novel C--7 substituent of the quinolones. The oxime geometry of the title compound was confirmed to have the E--configuration. In the molecule of the compound (Fig. 1), the N1---C6 (1.352 (3) Å) and N2---C12 (1.359 (3) Å) bond lengths are significantly shorter than the normal C---N bond (1.47 Å), indicating some conjugation with the C6═O2 and C12═O4 carbonyl groups, respectively. The six--membered piperidine ring adopts a chair conformation with displacing N1 and C3 atoms (0.593 (3) Å and -0.654 (3) Å respectively) from the mean--plane (C1, C2, C4 and C5). Experimental {#experimental} ============ To a stirring solution of 1-N-tert-Butoxycarbonyl-3-(N-tert-butoxycarbonyl) amino-4-methoxyimino-3-methylpiperidine(2.4 g, 6.7 mmol) in dry tetrahydrofuran (40 ml) was added 70% sodium hydride (0.46 g, 13.4 mmol) at 273 K using an ice bath, and then stirred for 0.5 h at the room temperature. After addition of methyl iodide (0.84 ml, 13.4 mol), the reaction mixture was stirred at 313 K for 5 h and cooled to room temperature, adjusted to pH 7 with 1N HCl and then concentrated under reduced pressure. The residue was diluted with ethyl acetate (50 ml), washed with distilled water, dried over anhydrous sodium sulfate, and filtered. The filtrate was concentrated under reduced pressure, dried in vacuo to give the title compound as a white solid (2.37 g, 95.0%; mp: 380--382 K). Single crystals suitable for X--ray analysis were obtained by slow evaporation of a methanol/water solution (5:1 v/v). ^1^H NMR (CDCl~3~, δ): 1.34 (s, 3H, CH~3~), 1.41 (s, 9H, BOC), 1.46 (s, 9H, BOC), 2.24--2.25 (m, 1H, piperidine), 2.87--2.88 (m, 1H, piperidine), 2.91 (s, 3H, NCH~3~), 2.95--3.08 (m, 2H, piperidine), 3.82 (s, 3H, OCH~3~), 3.84--3.86 (m, 1H, piperidine), 4.30--4.31 (m, 1H, piperidine); MS (ESI, m/z): 372.2 m/z (M+1)^+^. Refinement {#refinement} ========== All H atoms were placed at calculated positions, with C---H = 0.96--0.97 Å, and included in the final cycles of refinement using a riding model, with U~iso~(H) = 1.2U~eq~(C) for methylene or 1.5U~eq~(C) for methyl H atoms. Figures ======= ![The molecular structure of the title compound with the atom--numbering scheme. Displacement ellipsoids are drawn at 40% probability level. H atoms are presented as a small spheres of arbitrary radius.](e-65-0o256-fig1){#Fap1} Crystal data {#tablewrapcrystaldatalong} ============ ------------------------ --------------------------------------- C~18~H~33~N~3~O~5~ F(000) = 1616 M~r~ = 371.47 D~x~ = 1.180 Mg m^−3^ Monoclinic, C2/c Mo Kα radiation, λ = 0.71073 Å Hall symbol: -C 2yc Cell parameters from 1707 reflections a = 28.867 (3) Å θ = 2.7--21.1° b = 6.1887 (13) Å µ = 0.09 mm^−1^ c = 25.379 (3) Å T = 298 K β = 112.769 (2)° Prism, colourless V = 4180.6 (11) Å^3^ 0.40 × 0.20 × 0.11 mm Z = 8 ------------------------ --------------------------------------- Data collection {#tablewrapdatacollectionlong} =============== --------------------------------------------------------------- -------------------------------------- Bruker SMART CCD area-detector diffractometer 3699 independent reflections Radiation source: Fine--focus sealed tube 1915 reflections with I \> 2σ(I) Graphite R~int~ = 0.060 φ and ω scans θ~max~ = 25.0°, θ~min~ = 1.5° Absorption correction: multi-scan (SADABS; Sheldrick, 1996) h = −33→34 T~min~ = 0.963, T~max~ = 0.991 k = −7→6 10032 measured reflections l = −30→23 --------------------------------------------------------------- -------------------------------------- Refinement {#tablewraprefinementdatalong} ========== ------------------------------------- ------------------------------------------------------------------------------------------------ Refinement on F^2^ Primary atom site location: structure-invariant direct methods Least-squares matrix: full Secondary atom site location: difference Fourier map R\[F^2^ \> 2σ(F^2^)\] = 0.050 Hydrogen site location: inferred from neighbouring sites wR(F^2^) = 0.144 H-atom parameters constrained S = 1.02 w = 1/\[σ^2^(F~o~^2^) + (0.058P)^2^ + 0.3657P\] where P = (F~o~^2^ + 2F~c~^2^)/3 3699 reflections (Δ/σ)~max~ = 0.001 244 parameters Δρ~max~ = 0.23 e Å^−3^ 0 restraints Δρ~min~ = −0.22 e Å^−3^ ------------------------------------- ------------------------------------------------------------------------------------------------ Special details {#specialdetails} =============== --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Geometry. All s.u.\'s (except the s.u. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell s.u.\'s are taken into account individually in the estimation of s.u.\'s in distances, angles and torsion angles; correlations between s.u.\'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell s.u.\'s is used for estimating s.u.\'s involving l.s. planes. Refinement. Refinement of F^2^ against ALL reflections. The weighted R--factor wR and goodness of fit S are based on F^2^, conventional R--factors R are based on F, with F set to zero for negative F^2^. The threshold expression of F^2^ \> σ(F^2^) is used only for calculating R--factors(gt) etc. and is not relevant to the choice of reflections for refinement. R--factors based on F^2^ are statistically about twice as large as those based on F, and R--factors based on ALL data will be even larger. --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å^2^) {#tablewrapcoords} ================================================================================================== ------ --------------- ------------- -------------- -------------------- -- x y z U~iso~\/U~eq~ N1 0.18007 (8) 0.2543 (3) 0.60235 (9) 0.0389 (6) N2 0.07900 (7) 0.1687 (4) 0.51326 (9) 0.0367 (6) N3 0.16321 (7) −0.1321 (4) 0.47432 (9) 0.0382 (6) O1 0.16324 (7) 0.3113 (3) 0.67977 (7) 0.0496 (5) O2 0.21149 (7) 0.5507 (3) 0.65653 (8) 0.0554 (6) O3 0.07112 (6) 0.1423 (3) 0.42229 (7) 0.0525 (6) O4 0.01093 (7) 0.3134 (3) 0.44214 (8) 0.0607 (6) O5 0.19807 (6) −0.0836 (3) 0.44880 (7) 0.0452 (5) C1 0.15544 (9) 0.0442 (4) 0.59370 (10) 0.0374 (7) H1A 0.1361 0.0344 0.6175 0.045\ H1B 0.1809 −0.0681 0.6057 0.045\* C2 0.12010 (9) 0.0043 (4) 0.53070 (10) 0.0335 (6) C3 0.15423 (9) 0.0354 (4) 0.49846 (10) 0.0325 (6) C4 0.17891 (10) 0.2510 (4) 0.50509 (11) 0.0387 (7) H4A 0.1536 0.3625 0.4899 0.046\* H4B 0.2006 0.2542 0.4839 0.046\* C5 0.20975 (10) 0.2938 (5) 0.56825 (11) 0.0459 (8) H5A 0.2390 0.2005 0.5813 0.055\* H5B 0.2213 0.4425 0.5732 0.055\* C6 0.18713 (10) 0.3856 (5) 0.64739 (11) 0.0399 (7) C7 0.15940 (11) 0.4430 (5) 0.72637 (12) 0.0495 (8) C8 0.21026 (13) 0.4764 (6) 0.77309 (13) 0.0738 (11) H8A 0.2261 0.3388 0.7856 0.111\* H8B 0.2066 0.5494 0.8046 0.111\* H8C 0.2306 0.5623 0.7589 0.111\* C9 0.13357 (14) 0.6544 (6) 0.70277 (15) 0.0871 (13) H9A 0.1561 0.7461 0.6934 0.131\* H9B 0.1241 0.7241 0.7309 0.131\* H9C 0.1041 0.6272 0.6690 0.131\* C10 0.12728 (14) 0.3024 (6) 0.74676 (14) 0.0903 (13) H10A 0.0957 0.2762 0.7158 0.135\* H10B 0.1216 0.3738 0.7773 0.135\* H10C 0.1441 0.1674 0.7603 0.135\* C11 0.09808 (10) −0.2224 (4) 0.52538 (12) 0.0432 (7) H11A 0.0769 −0.2306 0.5466 0.065\* H11B 0.1248 −0.3256 0.5403 0.065\* H11C 0.0786 −0.2537 0.4859 0.065\* C12 0.05026 (10) 0.2153 (5) 0.45784 (12) 0.0417 (7) C13 0.05602 (10) 0.2229 (5) 0.55381 (11) 0.0492 (8) H13A 0.0295 0.3257 0.5367 0.074\* H13B 0.0810 0.2841 0.5877 0.074\* H13C 0.0425 0.0944 0.5636 0.074\* C14 0.04913 (11) 0.1885 (6) 0.36115 (12) 0.0584 (9) C15 0.04440 (16) 0.4286 (7) 0.35091 (17) 0.1048 (14) H15A 0.0185 0.4844 0.3620 0.157\* H15B 0.0360 0.4573 0.3111 0.157\* H15C 0.0757 0.4972 0.3731 0.157\* C17 0.08735 (13) 0.0915 (7) 0.34093 (13) 0.0907 (13) H17A 0.1198 0.1530 0.3622 0.136\* H17B 0.0779 0.1217 0.3010 0.136\* H17C 0.0887 −0.0621 0.3467 0.136\* C16 −0.00059 (12) 0.0715 (7) 0.33491 (14) 0.0906 (13) H16A 0.0039 −0.0778 0.3460 0.136\* H16B −0.0129 0.0822 0.2940 0.136\* H16C −0.0244 0.1360 0.3481 0.136\* C18 0.20473 (11) −0.2741 (5) 0.42114 (13) 0.0557 (9) H18A 0.2217 −0.3815 0.4493 0.083\* H18B 0.2245 −0.2411 0.3992 0.083\* H18C 0.1725 −0.3283 0.3962 0.083\* ------ --------------- ------------- -------------- -------------------- -- Atomic displacement parameters (Å^2^) {#tablewrapadps} ===================================== ----- ------------- ------------- ------------- -------------- ------------- -------------- U^11^ U^22^ U^33^ U^12^ U^13^ U^23^ N1 0.0452 (13) 0.0367 (15) 0.0395 (13) −0.0085 (11) 0.0215 (11) −0.0060 (12) N2 0.0359 (12) 0.0426 (15) 0.0368 (13) 0.0076 (11) 0.0198 (11) 0.0002 (11) N3 0.0368 (12) 0.0427 (15) 0.0420 (13) 0.0011 (11) 0.0226 (11) −0.0024 (12) O1 0.0665 (13) 0.0501 (13) 0.0394 (11) −0.0098 (11) 0.0284 (10) −0.0081 (10) O2 0.0687 (14) 0.0486 (14) 0.0528 (13) −0.0193 (12) 0.0277 (11) −0.0143 (11) O3 0.0471 (11) 0.0771 (16) 0.0335 (11) 0.0188 (11) 0.0157 (9) 0.0062 (11) O4 0.0469 (12) 0.0716 (16) 0.0619 (14) 0.0252 (12) 0.0193 (10) 0.0080 (12) O5 0.0509 (11) 0.0453 (13) 0.0538 (12) −0.0003 (10) 0.0360 (10) −0.0070 (10) C1 0.0438 (16) 0.0357 (18) 0.0371 (16) −0.0012 (14) 0.0203 (13) −0.0006 (13) C2 0.0361 (15) 0.0331 (17) 0.0349 (15) 0.0017 (13) 0.0178 (12) 0.0011 (13) C3 0.0314 (14) 0.0359 (17) 0.0331 (15) 0.0034 (13) 0.0155 (12) 0.0006 (13) C4 0.0451 (16) 0.0350 (17) 0.0442 (17) −0.0010 (14) 0.0264 (13) −0.0012 (14) C5 0.0468 (17) 0.0457 (19) 0.0510 (18) −0.0088 (15) 0.0254 (15) −0.0077 (15) C6 0.0454 (17) 0.0398 (19) 0.0361 (16) 0.0002 (15) 0.0174 (14) −0.0027 (15) C7 0.066 (2) 0.050 (2) 0.0392 (17) 0.0003 (17) 0.0280 (16) −0.0052 (16) C8 0.090 (3) 0.084 (3) 0.0401 (18) 0.001 (2) 0.0172 (19) −0.0105 (19) C9 0.110 (3) 0.089 (3) 0.074 (3) 0.044 (3) 0.049 (2) 0.007 (2) C10 0.127 (3) 0.099 (3) 0.071 (2) −0.034 (3) 0.066 (2) −0.023 (2) C11 0.0486 (16) 0.0375 (18) 0.0494 (17) −0.0058 (14) 0.0252 (14) −0.0037 (15) C12 0.0406 (17) 0.0440 (19) 0.0439 (18) 0.0051 (15) 0.0201 (14) 0.0005 (15) C13 0.0461 (17) 0.058 (2) 0.0512 (18) 0.0062 (15) 0.0276 (15) −0.0066 (16) C14 0.054 (2) 0.082 (3) 0.0373 (17) 0.0154 (19) 0.0156 (15) 0.0125 (18) C15 0.120 (3) 0.109 (4) 0.092 (3) 0.012 (3) 0.048 (3) 0.044 (3) C17 0.079 (2) 0.151 (4) 0.046 (2) 0.030 (3) 0.0297 (19) 0.010 (2) C16 0.072 (3) 0.130 (4) 0.057 (2) 0.002 (3) 0.0116 (19) −0.011 (2) C18 0.065 (2) 0.052 (2) 0.065 (2) −0.0031 (17) 0.0409 (17) −0.0190 (17) ----- ------------- ------------- ------------- -------------- ------------- -------------- Geometric parameters (Å, °) {#tablewrapgeomlong} =========================== --------------------- -------------- ---------------------- ------------ N1---C6 1.352 (3) C8---H8B 0.9600 N1---C5 1.455 (3) C8---H8C 0.9600 N1---C1 1.457 (3) C9---H9A 0.9600 N2---C12 1.359 (3) C9---H9B 0.9600 N2---C13 1.463 (3) C9---H9C 0.9600 N2---C2 1.494 (3) C10---H10A 0.9600 N3---C3 1.280 (3) C10---H10B 0.9600 N3---O5 1.423 (2) C10---H10C 0.9600 O1---C6 1.342 (3) C11---H11A 0.9600 O1---C7 1.476 (3) C11---H11B 0.9600 O2---C6 1.210 (3) C11---H11C 0.9600 O3---C12 1.342 (3) C13---H13A 0.9600 O3---C14 1.459 (3) C13---H13B 0.9600 O4---C12 1.211 (3) C13---H13C 0.9600 O5---C18 1.423 (3) C14---C15 1.505 (5) C1---C2 1.548 (3) C14---C17 1.510 (4) C1---H1A 0.9700 C14---C16 1.513 (4) C1---H1B 0.9700 C15---H15A 0.9600 C2---C3 1.517 (3) C15---H15B 0.9600 C2---C11 1.525 (3) C15---H15C 0.9600 C3---C4 1.491 (3) C17---H17A 0.9600 C4---C5 1.525 (3) C17---H17B 0.9600 C4---H4A 0.9700 C17---H17C 0.9600 C4---H4B 0.9700 C16---H16A 0.9600 C5---H5A 0.9700 C16---H16B 0.9600 C5---H5B 0.9700 C16---H16C 0.9600 C7---C8 1.502 (4) C18---H18A 0.9600 C7---C10 1.502 (4) C18---H18B 0.9600 C7---C9 1.510 (4) C18---H18C 0.9600 C8---H8A 0.9600 C6---N1---C5 118.1 (2) C7---C9---H9C 109.5 C6---N1---C1 124.7 (2) H9A---C9---H9C 109.5 C5---N1---C1 115.2 (2) H9B---C9---H9C 109.5 C12---N2---C13 114.7 (2) C7---C10---H10A 109.5 C12---N2---C2 123.2 (2) C7---C10---H10B 109.5 C13---N2---C2 118.2 (2) H10A---C10---H10B 109.5 C3---N3---O5 110.9 (2) C7---C10---H10C 109.5 C6---O1---C7 121.1 (2) H10A---C10---H10C 109.5 C12---O3---C14 121.7 (2) H10B---C10---H10C 109.5 C18---O5---N3 107.7 (2) C2---C11---H11A 109.5 N1---C1---C2 112.7 (2) C2---C11---H11B 109.5 N1---C1---H1A 109.1 H11A---C11---H11B 109.5 C2---C1---H1A 109.1 C2---C11---H11C 109.5 N1---C1---H1B 109.1 H11A---C11---H11C 109.5 C2---C1---H1B 109.1 H11B---C11---H11C 109.5 H1A---C1---H1B 107.8 O4---C12---O3 123.7 (3) N2---C2---C3 111.2 (2) O4---C12---N2 124.5 (3) N2---C2---C11 110.1 (2) O3---C12---N2 111.8 (2) C3---C2---C11 113.9 (2) N2---C13---H13A 109.5 N2---C2---C1 109.1 (2) N2---C13---H13B 109.5 C3---C2---C1 103.36 (19) H13A---C13---H13B 109.5 C11---C2---C1 108.9 (2) N2---C13---H13C 109.5 N3---C3---C4 127.0 (2) H13A---C13---H13C 109.5 N3---C3---C2 116.7 (2) H13B---C13---H13C 109.5 C4---C3---C2 115.7 (2) O3---C14---C15 110.5 (3) C3---C4---C5 109.4 (2) O3---C14---C17 102.1 (2) C3---C4---H4A 109.8 C15---C14---C17 111.3 (3) C5---C4---H4A 109.8 O3---C14---C16 108.8 (3) C3---C4---H4B 109.8 C15---C14---C16 112.9 (3) C5---C4---H4B 109.8 C17---C14---C16 110.7 (3) H4A---C4---H4B 108.2 C14---C15---H15A 109.5 N1---C5---C4 110.9 (2) C14---C15---H15B 109.5 N1---C5---H5A 109.5 H15A---C15---H15B 109.5 C4---C5---H5A 109.5 C14---C15---H15C 109.5 N1---C5---H5B 109.5 H15A---C15---H15C 109.5 C4---C5---H5B 109.5 H15B---C15---H15C 109.5 H5A---C5---H5B 108.0 C14---C17---H17A 109.5 O2---C6---O1 124.7 (3) C14---C17---H17B 109.5 O2---C6---N1 123.8 (3) H17A---C17---H17B 109.5 O1---C6---N1 111.5 (3) C14---C17---H17C 109.5 O1---C7---C8 110.8 (2) H17A---C17---H17C 109.5 O1---C7---C10 101.8 (2) H17B---C17---H17C 109.5 C8---C7---C10 110.7 (3) C14---C16---H16A 109.5 O1---C7---C9 109.8 (2) C14---C16---H16B 109.5 C8---C7---C9 112.1 (3) H16A---C16---H16B 109.5 C10---C7---C9 111.3 (3) C14---C16---H16C 109.5 C7---C8---H8A 109.5 H16A---C16---H16C 109.5 C7---C8---H8B 109.5 H16B---C16---H16C 109.5 H8A---C8---H8B 109.5 O5---C18---H18A 109.5 C7---C8---H8C 109.5 O5---C18---H18B 109.5 H8A---C8---H8C 109.5 H18A---C18---H18B 109.5 H8B---C8---H8C 109.5 O5---C18---H18C 109.5 C7---C9---H9A 109.5 H18A---C18---H18C 109.5 C7---C9---H9B 109.5 H18B---C18---H18C 109.5 H9A---C9---H9B 109.5 C3---N3---O5---C18 −177.9 (2) C6---N1---C5---C4 −143.2 (2) C6---N1---C1---C2 139.0 (2) C1---N1---C5---C4 52.3 (3) C5---N1---C1---C2 −57.6 (3) C3---C4---C5---N1 −49.9 (3) C12---N2---C2---C3 48.3 (3) C7---O1---C6---O2 8.0 (4) C13---N2---C2---C3 −155.2 (2) C7---O1---C6---N1 −171.1 (2) C12---N2---C2---C11 −78.9 (3) C5---N1---C6---O2 10.3 (4) C13---N2---C2---C11 77.7 (3) C1---N1---C6---O2 173.3 (3) C12---N2---C2---C1 161.6 (2) C5---N1---C6---O1 −170.6 (2) C13---N2---C2---C1 −41.8 (3) C1---N1---C6---O1 −7.6 (4) N1---C1---C2---N2 −62.4 (3) C6---O1---C7---C8 −66.1 (3) N1---C1---C2---C3 56.0 (3) C6---O1---C7---C10 176.2 (3) N1---C1---C2---C11 177.4 (2) C6---O1---C7---C9 58.3 (3) O5---N3---C3---C4 −5.2 (3) C14---O3---C12---O4 4.0 (4) O5---N3---C3---C2 −176.31 (19) C14---O3---C12---N2 −175.0 (2) N2---C2---C3---N3 −129.7 (2) C13---N2---C12---O4 7.3 (4) C11---C2---C3---N3 −4.6 (3) C2---N2---C12---O4 164.6 (3) C1---C2---C3---N3 113.4 (2) C13---N2---C12---O3 −173.7 (2) N2---C2---C3---C4 58.2 (3) C2---N2---C12---O3 −16.5 (4) C11---C2---C3---C4 −176.7 (2) C12---O3---C14---C15 57.0 (4) C1---C2---C3---C4 −58.8 (3) C12---O3---C14---C17 175.5 (3) N3---C3---C4---C5 −113.4 (3) C12---O3---C14---C16 −67.5 (4) C2---C3---C4---C5 57.8 (3) --------------------- -------------- ---------------------- ------------	{ "pile_set_name": "PubMed Central" }
![A.C.H. Watson (1936-2018)](IJPS-51-260-g001){#d35e66} A CH Watson or Tony as he was commonly and affectionately called unobtrusively influenced the course of plastic surgery not only in his native Britain but also in India for over four decades. Born in 1936, he did his basic medical training in Edinburgh, Scotland, UK and graduated in 1960. His early surgical training started in The Royal Hospital for Sick Children or the 'sick kids' in Edinburgh and went on to the Edinburgh Royal Infirmary in General Surgery and allied branches until 1965. During this period, he qualified with an FRCS (Ed.) in 1964. Plastic surgery training was at the Bangaur General Hospital in the South East Scotland Regional unit under Mr. AB Wallace (of the rule of 9 fame), Mr. Buchan and Ms. Anne Sutherland from 1966 to 1970. The year 1971--1972 was spent as the Maytag-Macahill Fellow under Dr. Ralph Millard. I suppose this led to a lifelong love affair with cleft surgery. In 1972, he came back and became a consultant in the same unit at Bangaur, where he continued to work until his retirement in 1996. He contributed hugely to plastic surgery in general and cleft work in particular. A search on Google Scholar revealed over four pages of published papers and books. He was one of the earliest to adopt nasal endoscopy for diagnosis of velopharyngeal incompetence. Personally, I saw some beautiful cleft work while training under him in the 'sick kids', but what was more impressive was his dedication to the children, his patient manner in the out-patient clinic with fractious and to my eye, often spoilt kids and his very effective collaboration with the Speech therapist. Records were meticulously kept and proved a treasure trove for writing papers and drawing useful conclusions, which influenced care. The hallmark was, of course, the complete lack of fuss and pomp. He also contributed to breast cancer by being one of the early proponents of immediate reconstruction in Edinburgh. This work was later carried on by his colleague, James Watson. There are many other contributions, but they are too many to list here. Organisationally he played a huge role in the Royal College, as well as in the British Association of Plastic Surgeons (BAPS) as an office bearer, examiner and via the Specialist Advisory Committee which worked very hard to maintain and raise basic standards in teaching and training. He held multiple positions in BAPS, but the two most important were as Editor of the then BJPS and as President. His influence in India was through his numerous trainees, the author being one of them. Working with him taught us to be sincere, meticulous, balanced and most importantly well behaved. The last being the most challenging. Just as parents live on by spreading their genes to their offspring; I believe firmly that surgeons live through their students who have imbibed their skills and values. In that respect, Tony was a great success in the UK as well as in India and many other parts of the world. In 2010 he was honoured with the Honorary membership of the Association of Plastic Surgeons of India by the author who was then the President of APSI in recognition for his contribution to training Indian Surgeons. He also greatly helped the Indian Journal of Plastic Surgery after I became editor and guided me to make it a somewhat sensible scientific publication. As always this was done silently, with empathy and kindness. Over the last several years he has provided similar editorial input to 'The Short Notes in Plastic Surgery' a blog post run by Dr. RL Thatte with the active help of APSI. Tony lived life with grace and dignity and passed away in the same fashion. His passing is a great loss to plastic surgery in India for we have lost a revered Guru and a genuine friend of this country. Tony is survived by his wife Anne to whom he was married for five decades as well as three daughters and one son. Rest in peace Tony, we hope we can emulate some of your great qualities.	{ "pile_set_name": "PubMed Central" }
Related literature {#sec1} ==================== For the previous determination of the title compound at room temperature (Cambridge Structural Database refcode: JAMFEI), see: FitzRoy et al. (1989[@bb5]). For a description of the Cambridge Structural Database, see: Groom & Allen (2014[@bb6]). For previous studies investigating recovered molybdenum(VI) catalysts, see: Amarante et al. (2015[@bb1]); Lysenko et al. (2015[@bb7]). Experimental {#sec2} ============== Crystal data {#sec2.1} -------------- (C~5~H~9~N~2~)~4~\[Mo~8~O~26~\]·4H~2~OM ~r~ = 1644.15Triclinic,a = 10.1105 (9) Åb = 10.7469 (9) Åc = 11.9839 (10) Åα = 64.103 (3)°β = 84.272 (3)°γ = 75.826 (3)°V = 1135.67 (17) Å^3^Z = 1Mo Kα radiationμ = 2.24 mm^−1^T = 180 K0.20 × 0.14 × 0.01 mm Data collection {#sec2.2} ----------------- Bruker D8 QUEST diffractometerAbsorption correction: multi-scan (SADABS; Bruker, 2001[@bb3]) T ~min~ = 0.663, T ~max~ = 0.74525625 measured reflections4141 independent reflections3175 reflections with I \> 2σ(I)R ~int~ = 0.042 Refinement {#sec2.3} ------------ R\[F ^2^ \> 2σ(F ^2^)\] = 0.026wR(F ^2^) = 0.056S = 1.064141 reflections320 parameters7 restraintsH atoms treated by a mixture of independent and constrained refinementΔρ~max~ = 0.70 e Å^−3^Δρ~min~ = −0.54 e Å^−3^ {#d5e542} Data collection: APEX2 (Bruker, 2012[@bb4]); cell refinement: SAINT (Bruker, 2012[@bb4]); data reduction: SAINT; program(s) used to solve structure: SHELXS97 (Sheldrick, 2008[@bb8]); program(s) used to refine structure: SHELXL2014 (Sheldrick, 2015[@bb9]); molecular graphics: DIAMOND (Brandenburg, 1999[@bb2]); software used to prepare material for publication: SHELXL2014. Supplementary Material ====================== Crystal structure: contains datablock(s) I, New_Global_Publ_Block. DOI: [10.1107/S2056989015022823/hb7547sup1.cif](http://dx.doi.org/10.1107/S2056989015022823/hb7547sup1.cif) Structure factors: contains datablock(s) I. DOI: [10.1107/S2056989015022823/hb7547Isup2.hkl](http://dx.doi.org/10.1107/S2056989015022823/hb7547Isup2.hkl) ###### Click here for additional data file. . DOI: [10.1107/S2056989015022823/hb7547fig1.tif](http://dx.doi.org/10.1107/S2056989015022823/hb7547fig1.tif) Schematic representation of the molecular entities composing the asymmetric unit of the title compound. The β-octamolybdate anion has been completed by inversion symmetry for the sake of chemical accuracy. All non-hydrogen atoms are represented as displacement ellipsoids drawn at the 60% probability level and hydrogen atoms as small spheres with arbitrary radii. Non-hydrogen atoms belonging to the asymmetric unit have been labelled for clarity. Dashed green broken lines indicate N---H⋯O and O---H⋯O hydrogen-bonding interactions (see Table for geometrical details). ###### Click here for additional data file. \(a\) (b) . DOI: [10.1107/S2056989015022823/hb7547fig2.tif](http://dx.doi.org/10.1107/S2056989015022823/hb7547fig2.tif) Crystal packing of the title compound viewed in perspective along the (a) \[100\] (b) \[010\] directions of the unit cell emphasising the supramolecular N---H⋯O and O---H⋯O hydrogen-bonding interactions (dashed green lines) interconnecting the three types of chemical species present in the crystal structure of the title compound. ###### Click here for additional data file. . DOI: [10.1107/S2056989015022823/hb7547fig3.tif](http://dx.doi.org/10.1107/S2056989015022823/hb7547fig3.tif) Mixed polyhedral (for the β-octamolybdate anion), ball-and-stick (for the 3,5-dimethylpyrazolium cations) and space filling (for the water molecules of crystallisation) schematic representation of the crystal packing of the title compound viewed in perspective along the \[100\] direction of the unit cell. The Figure illustrates well how the inorganic component of the crystal structure is embedded into an organic matrix, with the entrapped water molecules of crystallization acting as molecular fillers interacting with the hybrid network through hydrogen bonds. CCDC reference: [1439409](http://scripts.iucr.org/cgi-bin/cr.cgi?rm=csd&csdid=1439409) Additional supporting information: [crystallographic information](http://scripts.iucr.org/cgi-bin/sendsupfiles?hb7547&file=hb7547sup0.html&mime=text/html); [3D view](http://scripts.iucr.org/cgi-bin/sendcif?hb7547sup1&Qmime=cif); [checkCIF report](http://scripts.iucr.org/cgi-bin/paper?hb7547&checkcif=yes) Supporting information for this paper is available from the IUCr electronic archives (Reference: [HB7547](http://scripts.iucr.org/cgi-bin/sendsup?hb7547)). Funding Sources and Entities Fundação para a Ciência e a Tecnologia (FCT, Portugal), the European Union, QREN, FEDER through Programa Operacional Factores de Competitividade (COMPETE), CICECO--Aveiro Institute of Materials (Ref. FCT UID/CTM/50011/2013) financed by national funds through the FCT/MEC and when applicable co-financed by FEDER under the PT2020 Partnership Agreement.Projects and Individual grants We wish to thank FCT for funding the R&D project FCOMP-01--0124-FEDER-041282 (Ref. FCT EXPL/CTM-NAN/0013/2013), and also CICECO for specific funding towards the purchase of the single-crystal diffractometer. FCT is gratefully acknowledged for the post-doctoral research grant No. SFRH/BPD/97660/2013 (to TRA). S1. Context and introduction {#d1e22} ============================ Research efforts in the development and application of new hybrid molybdenum(vi) heterogeneous catalysts require on a daily basis the recovery of used materials and their characterization in the solid state to check for structural modifications of the employed compound (Amarante et al., 2015; Lysenko et al., 2015). Often, the use of drastic experimental conditions leads to the formation of secondary products, which crystallize in the medium as trace amounts of impurity compounds. It is, thus, imperative that most of these possible products are fully described in the solid state in the most accurate fashion. The title compound, (C~5~H~9~N~2~)~4~\[Mo~8~O~26~\]^.^4(H~2~O), was previously reported by FitzRoy et al. (1989). Besides the fact that the authors did not fully elucidate the hydrogen bonding network of this material (hydrogen atoms were placed geometrically) and that only the metallic centers could be refined anisotropically (from a room-temperature determination), the unit-cell parameters reported in the main text and in the abstract do not match (viz. the c axis length). A search and match at the Cambridge Structural Database (Groom et al., 2014) also seems to ignore the presence of Mo(VI) metal centers. In this context, we decided to recollect the crystal structure of the title compound at low temperature to fully elucidate its finer structural details. S2. Structure description {#d1e82} ========================= The asymmetric unit of the title compound is composed of two 3,5-dimethylpyrazolium cations, (C~5~H~9~N~2~)^+^, one half of the β-octamolybdate anion, β-\[Mo~8~O~26~\]^4−^, and two water molecules of crystallization. Noteworthy, while one water molecule was fully located in its crystallographic position and even the associated hydrogen atoms found from difference Fourier maps, the other was found to be disordered over two distinct locations, O2W and O3W (Fig. 1). This feature was not disclosed in the previous structural determination by FitzRoy et al. (1989). The molecular geometrical parameters for the β-octamolybdate anion are typical, exhibiting the usual four families of Mo---O bonds: Mo---Ot to terminal oxido groups \[bond distances in the 1.691 (3)--1.715 (3) Å range\]; Mo---Ob to µ~2~-bridging oxido groups \[bond distances in the 1.754 (3)--2.268 (3) Å range\]; Mo---Oc to µ~3~-bridging oxido groups \[bond distances in the 1.953 (3)--2.351 (3) Å range\]; Mo---Oc to µ~5~-bridging oxido groups \[bond distances in the 2.146 (3)--2.435 (3) Å range\]. The four crystallographically independent Mo(VI) metal centers are thus hexacoordinated in a typical {MoO~6~} fashion resembling highly distorted octahedra: while this trans internal O---Mo---O octahedral angles were found in the 145.29 (11)--173.17 (12)° range, the cis angles refined instead in the 70.01 (9)--105.85 (14)° interval. We note that this wide dispersion for the internal octahedral angles is a notable and well known consequence of the marked trans effect created by the terminal oxido groups which displace the metal centers from the center of the octahedra. The β-octamolybdate anion, located in the center of the unit cell, interacts with the remaining chemical species through a series of both electrostatic interactions and hydrogen bonds (Fig. 2). One crystallographically independent 3,5-dimethylpyrazolium cation donates both the hydrogen atoms bound to nitrogen to form two strong \[D···A distances of 2.801 (4) and 2.869 (5) Å\] and relatively directional \[\<(DHA) angles of 120 and 152°\] charged N---H···O interactions with the polyoxoanion. The other cation has, however, a completely distinct behaviour: the same hydrogen atoms are instead donated in similar (strong and highly directional) interactions with neighbouring water molecules: while the D···A distances are 2.689 (7) and 2.776 (5) Å, the \<(DHA) interaction angles are close to linearity, being 171 and 164°. The water molecules are instead interacting with the β-octamolybdate anion as depicted in both Figs. 1 and 2. Indeed, as depicted in Fig. 3 the water molecules play a decisive role in the overall crystal packing, acting as molecular fillers to effectively occupy the available space left from the arrangement of inorganic anions and organic cations. S3. Synthesis and crystallization {#d1e175} ================================= All chemicals were purchased from commercial sources and used as received without additional purification steps. A Teflon-lined stainless steel vessel was charged with a reaction mixture composed of MoO~3~ (0.34 g, 2.43 mmol), 3,5-dimethylpyrazole (0.11 g, 1.21 mmol) and water (ca 25 ml) and heated in an oven at 160 °C for 26 h. The resultant blueish solid was filtered from the aqueous mother liquor and washed with an excess of water and (4×10 ml) diethyl ether, dried at ambient temperature and characterized in the solid state. Colourless plates of the title compound were directly harvested from the walls of the Teflon vessel. Selected FT---IR (KBr, cm^−1^): ñ = 948 (vs), 910 (vs), 843 (s), 733 (s), 714 (s), 663 (s). S4. Refinement details {#d1e222} ====================== Hydrogen atoms bound to carbon atoms were placed at idealized positions with C---H = 0.95 or 0.98 Å (for the aromatic and methyl groups, respectively), and included in the final structural model in riding-motion approximation with the isotropic thermal displacement parameters fixed at 1.2 or 1.5×U~eq~, respectively, of the carbon atom to which they are attached. Hydrogen atoms associated with nitrogen atoms have been directly located from difference Fourier maps and were included in the model with the N---H distances restrained to 0.95 (1) Å in order to ensure a chemically reasonable environment for these moieties. These hydrogen atoms were modelled with the isotropic thermal displacement parameters fixed at 1.5×U~eq~(N). A total of two water molecules of crystallization were directly located from difference Fourier maps. Though O1W was included in the final structural model by assuming full site occupancy and a typical anisotropic displacement behaviour, the second molecule was found to be disordered over two close crystallographic positions: O2W and O3W. These species were included in the structural model with linked site occupancy \[which ultimately refined to 0.65 (2) and 0.35 (2), respectively\] and by assuming an independent isotropic displacement behaviour. For O1W the two hydrogen atoms were markedly visible in difference Fourier maps and were included in the final model with the O---H and H···H distances restrained to 0.95 (1) and 1.55 (1) Å, respectively, in order to ensure a chemically reasonable geometry for this molecule. These hydrogen atoms were modelled with the isotropic thermal displacement parameters fixed at 1.5×U~eq~(O1W). Figures ======= ![Schematic representation of the molecular entities composing the asymmetric unit of the title compound. The β-octamolybdate anion has been completed by inversion symmetry for the sake of chemical accuracy. All non-hydrogen atoms are represented as displacement ellipsoids drawn at the 60% probability level and hydrogen atoms as small spheres with arbitrary radii. Non-hydrogen atoms belonging to the asymmetric unit have been labelled for clarity. Dashed green broken lines indicate N---H···O and O---H···O hydrogen-bonding interactions (see Table for geometrical details).](e-71-0m244-fig1){#Fap1} ![Crystal packing of the title compound viewed in perspective along the (a) \[100\] (b) \[010\] directions of the unit cell emphasizing the supramolecular N---H···O and O---H···O hydrogen-bonding interactions (dashed green lines) interconnecting the three types of chemical species present in the crystal structure of the title compound.](e-71-0m244-fig2){#Fap2} ![Mixed polyhedral (for the β-octamolybdate anion), ball-and-stick (for the 3,5-dimethylpyrazolium cations) and space filling (for the water molecules of crystallization) schematic representation of the crystal packing of the title compound viewed in perspective along the \[100\] direction of the unit cell. The Figure illustrates well how the inorganic component of the crystal structure is embedded into an organic matrix, with the entrapped water molecules of crystallization acting as molecular fillers interacting with the hybrid network through hydrogen bonds.](e-71-0m244-fig3){#Fap3} Crystal data {#tablewrapcrystaldatalong} ============ ---------------------------------------- --------------------------------------- (C~5~H~9~N~2~)~4~\[Mo~8~O~26~\]·4H~2~O Z = 1 M~r~ = 1644.15 F(000) = 796 Triclinic, P1 D~x~ = 2.404 Mg m^−3^ a = 10.1105 (9) Å Mo Kα radiation, λ = 0.71073 Å b = 10.7469 (9) Å Cell parameters from 9938 reflections c = 11.9839 (10) Å θ = 2.6--25.4° α = 64.103 (3)° µ = 2.24 mm^−1^ β = 84.272 (3)° T = 180 K γ = 75.826 (3)° Plate, colourless V = 1135.67 (17) Å^3^ 0.20 × 0.14 × 0.01 mm ---------------------------------------- --------------------------------------- Data collection {#tablewrapdatacollectionlong} =============== ------------------------------------------------------------ -------------------------------------- Bruker D8 QUEST diffractometer 4141 independent reflections Radiation source: Sealed tube 3175 reflections with I \> 2σ(I) Multi-layer X-ray mirror monochromator R~int~ = 0.042 Detector resolution: 10.4167 pixels mm^-1^ θ~max~ = 25.4°, θ~min~ = 3.7° ω / φ scans h = −12→12 Absorption correction: multi-scan (SADABS; Bruker, 2001) k = −12→12 T~min~ = 0.663, T~max~ = 0.745 l = −14→14 25625 measured reflections ------------------------------------------------------------ -------------------------------------- Refinement {#tablewraprefinementdatalong} ========== ------------------------------------- ------------------------------------------------------------------------------------------------- Refinement on F^2^ 7 restraints Least-squares matrix: full Hydrogen site location: mixed R\[F^2^ \> 2σ(F^2^)\] = 0.026 H atoms treated by a mixture of independent and constrained refinement wR(F^2^) = 0.056 w = 1/\[σ^2^(F~o~^2^) + (0.0188P)^2^ + 2.2481P\] where P = (F~o~^2^ + 2F~c~^2^)/3 S = 1.06 (Δ/σ)~max~ = 0.001 4141 reflections Δρ~max~ = 0.70 e Å^−3^ 320 parameters Δρ~min~ = −0.54 e Å^−3^ ------------------------------------- ------------------------------------------------------------------------------------------------- Special details {#specialdetails} =============== ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Geometry. All e.s.d.\'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.\'s are taken into account individually in the estimation of e.s.d.\'s in distances, angles and torsion angles; correlations between e.s.d.\'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.\'s is used for estimating e.s.d.\'s involving l.s. planes. ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å^2^) {#tablewrapcoords} ================================================================================================== ------ ------------- ------------- ------------- -------------------- ------------ x y z U~iso~\/U~eq~ Occ. (\<1) Mo1 0.44983 (3) 0.44979 (4) 0.75253 (3) 0.01430 (9) Mo2 0.45658 (3) 0.68126 (4) 0.47103 (3) 0.01229 (9) Mo3 0.23903 (3) 0.56541 (4) 0.38528 (3) 0.01415 (9) Mo4 0.23047 (3) 0.32798 (4) 0.67012 (3) 0.01511 (9) O1 0.3753 (3) 0.4944 (3) 0.5589 (2) 0.0142 (6) O2 0.3635 (3) 0.3005 (3) 0.7899 (2) 0.0177 (6) O3 0.5489 (3) 0.3881 (3) 0.8808 (3) 0.0235 (7) O4 0.5539 (3) 0.5839 (3) 0.6298 (2) 0.0145 (6) O5 0.3166 (3) 0.5748 (3) 0.7683 (3) 0.0208 (7) O6 0.5740 (3) 0.7845 (3) 0.3880 (2) 0.0168 (6) O7 0.3838 (3) 0.6770 (3) 0.3281 (2) 0.0153 (6) O8 0.3261 (3) 0.7924 (3) 0.5040 (3) 0.0190 (6) O9 0.1894 (3) 0.5797 (3) 0.2484 (3) 0.0232 (7) O10 0.1936 (3) 0.3896 (3) 0.4970 (2) 0.0160 (6) O11 0.1191 (3) 0.6842 (3) 0.4197 (3) 0.0220 (7) O12 0.1720 (3) 0.1773 (3) 0.7422 (3) 0.0255 (7) O13 0.1090 (3) 0.4568 (3) 0.6928 (3) 0.0213 (7) N1 0.7819 (4) 0.8121 (5) 0.9458 (3) 0.0294 (9) H1 0.865 (3) 0.747 (4) 0.979 (4) 0.044\ N2 0.6839 (4) 0.7657 (4) 0.9163 (4) 0.0281 (9) H2 0.692 (5) 0.6692 (19) 0.934 (5) 0.042\* N3 −0.0378 (4) 0.2498 (4) 0.3524 (4) 0.0295 (9) H3 −0.116 (3) 0.314 (4) 0.362 (5) 0.044\* N4 0.0827 (4) 0.2507 (4) 0.3913 (4) 0.0289 (9) H4 0.090 (5) 0.314 (4) 0.425 (4) 0.043\* C1 0.8229 (6) 1.0327 (6) 0.9381 (5) 0.0491 (15) H1A 0.8535 0.9870 1.0247 0.074\* H1B 0.7696 1.1293 0.9183 0.074\* H1C 0.9023 1.0365 0.8837 0.074\* C2 0.7364 (5) 0.9494 (5) 0.9191 (4) 0.0319 (11) C3 0.6048 (5) 0.9913 (5) 0.8718 (4) 0.0316 (11) H3A 0.5471 1.0837 0.8443 0.038\* C4 0.5743 (5) 0.8733 (5) 0.8726 (4) 0.0312 (11) C5 0.4486 (5) 0.8533 (6) 0.8341 (5) 0.0420 (14) H5A 0.4728 0.7778 0.8058 0.063\* H5B 0.4035 0.9420 0.7663 0.063\* H5C 0.3866 0.8271 0.9047 0.063\* C6 −0.1363 (5) 0.1101 (5) 0.2791 (5) 0.0329 (12) H6A −0.2222 0.1413 0.3150 0.049\* H6B −0.1209 0.0082 0.3021 0.049\* H6C −0.1412 0.1624 0.1885 0.049\* C7 −0.0218 (4) 0.1381 (5) 0.3270 (4) 0.0208 (10) C8 0.1130 (4) 0.0663 (5) 0.3502 (4) 0.0225 (10) H8 0.1545 −0.0174 0.3399 0.027\* C9 0.1768 (4) 0.1398 (5) 0.3917 (4) 0.0208 (10) C10 0.3202 (4) 0.1135 (5) 0.4290 (5) 0.0303 (11) H10A 0.3497 0.2034 0.3936 0.045\* H10B 0.3787 0.0479 0.3983 0.045\* H10C 0.3268 0.0717 0.5196 0.045\* O1W 0.7299 (3) 0.4764 (4) 0.9840 (3) 0.0276 (7) H1X 0.711 (5) 0.439 (5) 1.0698 (12) 0.041\* H1Y 0.676 (4) 0.449 (5) 0.943 (3) 0.041\* O2W 0.9897 (6) 0.5952 (10) 0.0682 (5) 0.031 (2)\* 0.65 (2) O3W 1.0101 (11) 0.6562 (19) 0.0507 (10) 0.033 (4)\* 0.35 (2) ------ ------------- ------------- ------------- -------------------- ------------ Atomic displacement parameters (Å^2^) {#tablewrapadps} ===================================== ----- -------------- -------------- -------------- --------------- --------------- --------------- U^11^ U^22^ U^33^ U^12^ U^13^ U^23^ Mo1 0.01445 (18) 0.0179 (2) 0.01272 (18) −0.00473 (15) 0.00061 (14) −0.00798 (16) Mo2 0.01153 (18) 0.01158 (19) 0.01458 (18) −0.00190 (14) −0.00116 (14) −0.00643 (15) Mo3 0.01088 (18) 0.0166 (2) 0.01578 (19) −0.00309 (15) −0.00217 (14) −0.00718 (16) Mo4 0.01333 (18) 0.0163 (2) 0.01719 (19) −0.00532 (15) 0.00086 (14) −0.00759 (16) O1 0.0124 (14) 0.0157 (15) 0.0150 (14) −0.0038 (12) 0.0002 (11) −0.0069 (12) O2 0.0179 (15) 0.0185 (16) 0.0157 (14) −0.0065 (12) 0.0021 (12) −0.0055 (13) O3 0.0232 (16) 0.0307 (19) 0.0185 (16) −0.0101 (14) −0.0013 (13) −0.0100 (14) O4 0.0142 (14) 0.0157 (15) 0.0179 (15) −0.0041 (12) −0.0013 (11) −0.0104 (13) O5 0.0194 (15) 0.0228 (17) 0.0223 (16) −0.0053 (13) 0.0038 (13) −0.0122 (14) O6 0.0184 (15) 0.0146 (15) 0.0171 (15) −0.0040 (12) −0.0022 (12) −0.0058 (13) O7 0.0128 (14) 0.0170 (16) 0.0150 (14) −0.0024 (12) −0.0010 (11) −0.0060 (12) O8 0.0186 (15) 0.0155 (16) 0.0227 (16) −0.0007 (12) −0.0028 (12) −0.0091 (13) O9 0.0201 (16) 0.0316 (19) 0.0194 (16) −0.0067 (14) −0.0042 (13) −0.0110 (14) O10 0.0146 (14) 0.0169 (16) 0.0199 (15) −0.0045 (12) −0.0005 (12) −0.0101 (13) O11 0.0159 (15) 0.0214 (17) 0.0273 (17) −0.0002 (13) −0.0014 (13) −0.0110 (14) O12 0.0243 (16) 0.0219 (17) 0.0313 (18) −0.0108 (14) 0.0025 (14) −0.0099 (15) O13 0.0171 (15) 0.0227 (17) 0.0232 (16) −0.0038 (13) 0.0018 (13) −0.0098 (14) N1 0.023 (2) 0.038 (3) 0.022 (2) −0.0038 (19) −0.0028 (17) −0.0087 (19) N2 0.027 (2) 0.029 (2) 0.024 (2) −0.0071 (19) −0.0013 (17) −0.0067 (19) N3 0.020 (2) 0.029 (2) 0.046 (3) 0.0003 (17) 0.0004 (18) −0.026 (2) N4 0.026 (2) 0.031 (2) 0.041 (2) −0.0094 (18) 0.0027 (18) −0.026 (2) C1 0.061 (4) 0.053 (4) 0.039 (3) −0.024 (3) −0.005 (3) −0.017 (3) C2 0.036 (3) 0.035 (3) 0.020 (2) −0.010 (2) 0.004 (2) −0.007 (2) C3 0.032 (3) 0.025 (3) 0.025 (3) −0.002 (2) −0.001 (2) −0.001 (2) C4 0.027 (3) 0.033 (3) 0.022 (2) −0.004 (2) −0.002 (2) −0.001 (2) C5 0.028 (3) 0.046 (4) 0.036 (3) −0.011 (2) −0.006 (2) 0.000 (3) C6 0.024 (3) 0.035 (3) 0.050 (3) −0.003 (2) −0.008 (2) −0.027 (3) C7 0.021 (2) 0.020 (2) 0.026 (2) −0.0049 (19) 0.0010 (18) −0.014 (2) C8 0.023 (2) 0.024 (3) 0.026 (2) −0.0022 (19) −0.0022 (19) −0.017 (2) C9 0.022 (2) 0.024 (2) 0.019 (2) −0.0054 (19) 0.0015 (18) −0.011 (2) C10 0.025 (2) 0.030 (3) 0.043 (3) −0.006 (2) −0.005 (2) −0.021 (2) O1W 0.0298 (18) 0.036 (2) 0.0240 (17) −0.0078 (15) −0.0003 (14) −0.0192 (16) ----- -------------- -------------- -------------- --------------- --------------- --------------- Geometric parameters (Å, º) {#tablewrapgeomlong} =========================== --------------------------- ------------- -------------------- ------------- Mo1---O3 1.702 (3) N2---C4 1.340 (6) Mo1---O5 1.715 (3) N2---H2 0.946 (10) Mo1---O2 1.875 (3) N3---C7 1.332 (5) Mo1---O4 1.982 (3) N3---N4 1.351 (5) Mo1---O7^i^ 2.307 (3) N3---H3 0.947 (10) Mo1---O1 2.320 (3) N4---C9 1.328 (6) Mo1---Mo2 3.2062 (6) N4---H4 0.947 (10) Mo2---O8 1.691 (3) C1---C2 1.488 (7) Mo2---O6 1.754 (3) C1---H1A 0.9800 Mo2---O7 1.953 (3) C1---H1B 0.9800 Mo2---O4 1.953 (3) C1---H1C 0.9800 Mo2---O1 2.146 (3) C2---C3 1.389 (7) Mo2---O1^i^ 2.341 (3) C3---C4 1.373 (7) Mo2---Mo3 3.2178 (5) C3---H3A 0.9500 Mo3---O9 1.695 (3) C4---C5 1.488 (7) Mo3---O11 1.695 (3) C5---H5A 0.9800 Mo3---O10 1.915 (3) C5---H5B 0.9800 Mo3---O7 1.998 (3) C5---H5C 0.9800 Mo3---O1 2.341 (3) C6---C7 1.488 (6) Mo3---O4^i^ 2.351 (3) C6---H6A 0.9800 Mo4---O12 1.692 (3) C6---H6B 0.9800 Mo4---O13 1.711 (3) C6---H6C 0.9800 Mo4---O10 1.928 (3) C7---C8 1.379 (6) Mo4---O2 1.943 (3) C8---C9 1.394 (6) Mo4---O6^i^ 2.268 (3) C8---H8 0.9500 Mo4---O1 2.435 (3) C9---C10 1.483 (6) O1---Mo2^i^ 2.341 (3) C10---H10A 0.9800 O4---Mo3^i^ 2.351 (3) C10---H10B 0.9800 O6---Mo4^i^ 2.268 (3) C10---H10C 0.9800 O7---Mo1^i^ 2.307 (3) O1W---H1X 0.946 (10) N1---C2 1.332 (6) O1W---H1Y 0.944 (10) N1---N2 1.350 (5) O2W---O3W 0.674 (13) N1---H1 0.943 (10) O3---Mo1---O5 104.71 (14) O12---Mo4---O1 162.52 (12) O3---Mo1---O2 102.12 (13) O13---Mo4---O1 91.62 (11) O5---Mo1---O2 100.72 (13) O10---Mo4---O1 74.97 (10) O3---Mo1---O4 99.75 (12) O2---Mo4---O1 73.62 (10) O5---Mo1---O4 97.01 (13) O6^i^---Mo4---O1 70.01 (9) O2---Mo1---O4 147.18 (11) Mo2---O1---Mo1 91.68 (9) O3---Mo1---O7^i^ 89.69 (12) Mo2---O1---Mo2^i^ 104.01 (10) O5---Mo1---O7^i^ 163.56 (12) Mo1---O1---Mo2^i^ 97.04 (9) O2---Mo1---O7^i^ 83.57 (11) Mo2---O1---Mo3 91.54 (10) O4---Mo1---O7^i^ 72.34 (10) Mo1---O1---Mo3 163.07 (12) O3---Mo1---O1 161.88 (12) Mo2^i^---O1---Mo3 98.29 (9) O5---Mo1---O1 93.03 (11) Mo2---O1---Mo4 163.99 (13) O2---Mo1---O1 77.65 (10) Mo1---O1---Mo4 86.35 (9) O4---Mo1---O1 74.01 (10) Mo2^i^---O1---Mo4 91.99 (9) O7^i^---Mo1---O1 72.23 (9) Mo3---O1---Mo4 85.98 (8) O3---Mo1---Mo2 134.82 (10) Mo1---O2---Mo4 116.94 (14) O5---Mo1---Mo2 84.85 (10) Mo2---O4---Mo1 109.13 (12) O2---Mo1---Mo2 119.63 (8) Mo2---O4---Mo3^i^ 110.39 (11) O4---Mo1---Mo2 35.14 (7) Mo1---O4---Mo3^i^ 103.30 (11) O7^i^---Mo1---Mo2 79.31 (7) Mo2---O6---Mo4^i^ 116.79 (14) O1---Mo1---Mo2 41.99 (7) Mo2---O7---Mo3 109.07 (13) O8---Mo2---O6 105.54 (13) Mo2---O7---Mo1^i^ 109.84 (11) O8---Mo2---O7 101.63 (12) Mo3---O7---Mo1^i^ 104.33 (12) O6---Mo2---O7 96.27 (12) Mo3---O10---Mo4 115.92 (13) O8---Mo2---O4 99.93 (12) C2---N1---N2 108.8 (4) O6---Mo2---O4 97.02 (11) C2---N1---H1 133 (3) O7---Mo2---O4 150.55 (11) N2---N1---H1 118 (3) O8---Mo2---O1 97.27 (12) C4---N2---N1 108.9 (4) O6---Mo2---O1 157.19 (11) C4---N2---H2 128 (3) O7---Mo2---O1 78.87 (10) N1---N2---H2 123 (3) O4---Mo2---O1 78.71 (10) C7---N3---N4 109.1 (4) O8---Mo2---O1^i^ 173.17 (12) C7---N3---H3 133 (3) O6---Mo2---O1^i^ 81.21 (11) N4---N3---H3 117 (3) O7---Mo2---O1^i^ 78.31 (10) C9---N4---N3 109.5 (4) O4---Mo2---O1^i^ 77.92 (10) C9---N4---H4 128 (3) O1---Mo2---O1^i^ 75.99 (10) N3---N4---H4 122 (3) O8---Mo2---Mo1 88.75 (10) C2---C1---H1A 109.5 O6---Mo2---Mo1 132.75 (9) C2---C1---H1B 109.5 O7---Mo2---Mo1 125.17 (8) H1A---C1---H1B 109.5 O4---Mo2---Mo1 35.73 (8) C2---C1---H1C 109.5 O1---Mo2---Mo1 46.33 (7) H1A---C1---H1C 109.5 O1^i^---Mo2---Mo1 85.79 (7) H1B---C1---H1C 109.5 O8---Mo2---Mo3 89.31 (9) N1---C2---C3 107.8 (4) O6---Mo2---Mo3 132.20 (9) N1---C2---C1 121.8 (5) O7---Mo2---Mo3 35.93 (8) C3---C2---C1 130.4 (5) O4---Mo2---Mo3 125.36 (8) C4---C3---C2 106.6 (4) O1---Mo2---Mo3 46.65 (7) C4---C3---H3A 126.7 O1^i^---Mo2---Mo3 86.76 (6) C2---C3---H3A 126.7 Mo1---Mo2---Mo3 91.724 (14) N2---C4---C3 107.8 (4) O9---Mo3---O11 105.85 (14) N2---C4---C5 121.2 (5) O9---Mo3---O10 100.45 (13) C3---C4---C5 130.9 (5) O11---Mo3---O10 101.76 (13) C4---C5---H5A 109.5 O9---Mo3---O7 100.71 (12) C4---C5---H5B 109.5 O11---Mo3---O7 97.58 (12) H5A---C5---H5B 109.5 O10---Mo3---O7 146.06 (11) C4---C5---H5C 109.5 O9---Mo3---O1 159.60 (12) H5A---C5---H5C 109.5 O11---Mo3---O1 94.36 (12) H5B---C5---H5C 109.5 O10---Mo3---O1 77.54 (10) C7---C6---H6A 109.5 O7---Mo3---O1 73.40 (10) C7---C6---H6B 109.5 O9---Mo3---O4^i^ 88.84 (12) H6A---C6---H6B 109.5 O11---Mo3---O4^i^ 163.17 (11) C7---C6---H6C 109.5 O10---Mo3---O4^i^ 83.12 (10) H6A---C6---H6C 109.5 O7---Mo3---O4^i^ 71.11 (10) H6B---C6---H6C 109.5 O1---Mo3---O4^i^ 70.77 (9) N3---C7---C8 107.4 (4) O9---Mo3---Mo2 135.69 (10) N3---C7---C6 121.8 (4) O11---Mo3---Mo2 85.49 (10) C8---C7---C6 130.8 (4) O10---Mo3---Mo2 119.34 (8) C7---C8---C9 107.0 (4) O7---Mo3---Mo2 35.00 (7) C7---C8---H8 126.5 O1---Mo3---Mo2 41.81 (7) C9---C8---H8 126.5 O4^i^---Mo3---Mo2 78.16 (6) N4---C9---C8 106.9 (4) O12---Mo4---O13 105.75 (14) N4---C9---C10 121.9 (4) O12---Mo4---O10 103.91 (13) C8---C9---C10 131.2 (4) O13---Mo4---O10 97.77 (13) C9---C10---H10A 109.5 O12---Mo4---O2 101.63 (13) C9---C10---H10B 109.5 O13---Mo4---O2 97.49 (13) H10A---C10---H10B 109.5 O10---Mo4---O2 145.29 (11) C9---C10---H10C 109.5 O12---Mo4---O6^i^ 92.59 (12) H10A---C10---H10C 109.5 O13---Mo4---O6^i^ 161.61 (12) H10B---C10---H10C 109.5 O10---Mo4---O6^i^ 78.80 (10) H1X---O1W---H1Y 109.9 (16) O2---Mo4---O6^i^ 76.95 (11) O3---Mo1---O2---Mo4 177.15 (15) N2---N1---C2---C1 −178.3 (4) O5---Mo1---O2---Mo4 69.40 (17) N1---C2---C3---C4 0.4 (5) O4---Mo1---O2---Mo4 −52.1 (3) C1---C2---C3---C4 179.0 (5) O7^i^---Mo1---O2---Mo4 −94.55 (15) N1---N2---C4---C3 1.4 (5) O1---Mo1---O2---Mo4 −21.39 (13) N1---N2---C4---C5 −179.1 (4) Mo2---Mo1---O2---Mo4 −20.75 (18) C2---C3---C4---N2 −1.1 (5) O8---Mo2---O6---Mo4^i^ 178.38 (13) C2---C3---C4---C5 179.4 (5) O7---Mo2---O6---Mo4^i^ −77.67 (14) N4---N3---C7---C8 −0.1 (5) O4---Mo2---O6---Mo4^i^ 75.99 (14) N4---N3---C7---C6 −178.6 (4) O1---Mo2---O6---Mo4^i^ −1.5 (4) N3---C7---C8---C9 0.4 (5) O1^i^---Mo2---O6---Mo4^i^ −0.57 (12) C6---C7---C8---C9 178.7 (5) Mo1---Mo2---O6---Mo4^i^ 75.31 (16) N3---N4---C9---C8 0.5 (5) Mo3---Mo2---O6---Mo4^i^ −78.01 (15) N3---N4---C9---C10 179.3 (4) C2---N1---N2---C4 −1.1 (5) C7---C8---C9---N4 −0.6 (5) C7---N3---N4---C9 −0.2 (5) C7---C8---C9---C10 −179.2 (5) N2---N1---C2---C3 0.4 (5) --------------------------- ------------- -------------------- ------------- Symmetry code: (i) −x+1, −y+1, −z+1. Hydrogen-bond geometry (Å, º) {#tablewraphbondslong} ============================= ------------------------ --------- --------- ----------- --------------- D---H···A D---H H···A D···A D---H···A N1---H1···O2W^ii^ 0.94 1.77 2.689 (7) 164 N2---H2···O1W 0.95 1.84 2.776 (5) 171 N3---H3···O13^iii^ 0.95 2.28 2.869 (5) 120 N4---H4···O10 0.95 1.93 2.801 (4) 152 O1W---H1X···O5^iv^ 0.95 1.88 2.785 (4) 160 O1W---H1Y···O3 0.94 1.91 2.848 (4) 172 ------------------------ --------- --------- ----------- --------------- Symmetry codes: (ii) x, y, z+1; (iii) −x, −y+1, −z+1; (iv) −x+1, −y+1, −z+2. ###### Hydrogen-bond geometry (Å, °) D---H⋯A D---H H⋯A D⋯A D---H⋯A ----------------------- --------- ------- ----------- ------------- N1---H1⋯O2W ^i^ 0.94 1.77 2.689 (7) 164 N2---H2⋯O1W 0.95 1.84 2.776 (5) 171 N3---H3⋯O13^ii^ 0.95 2.28 2.869 (5) 120 N4---H4⋯O10 0.95 1.93 2.801 (4) 152 O1W---H1X⋯O5^iii^ 0.95 1.88 2.785 (4) 160 O1W---H1Y⋯O3 0.94 1.91 2.848 (4) 172 Symmetry codes: (i) ; (ii) ; (iii) .	{ "pile_set_name": "PubMed Central" }
Background ========== MicroRNAs (miRNAs) are short (around 20-22 nucleotides) RNAs that post-transcriptionally regulate gene expression by base-pairing with complementary sequences in the 3\' untranslated regions (UTRs) of protein-coding transcripts and directing translational repression or transcript degradation \[[@B1]-[@B5]\]. There are currently 326 human miRNAs listed in the miRNA registry version 7.1 \[[@B6]\], but the total number of miRNAs encoded in the human genome may be nearer 1,000 \[[@B7],[@B8]\]. The function of most miRNAs is unknown, but many are clearly involved in regulating differentiation \[[@B9]\] and development \[[@B10]\]. It is estimated that up to 30% of human genes may be miRNA targets \[[@B11],[@B12]\]. miRNAs are transcribed by RNA polymerase II into long primary miRNA (pri-miRNA) transcripts which are capped and polyadenylated \[[@B13],[@B14]\]. Genomic analyses indicate that many miRNAs overlap known protein coding genes or non-coding RNAs \[[@B15]\], and that many are in evolutionarily conserved clusters with other miRNAs \[[@B16]\]. Furthermore, intronic miRNAs share expression patterns with adjacent miRNAs and the host gene mRNA indicating that they are coordinately coexpressed \[[@B17]\]. Pri-miRNAs contain a short double-stranded RNA (dsRNA) stem-loop formed between the miRNA sequence and its adjacent complementary sequence. In the nucleus, the ribonuclease III-like enzyme Drosha cleaves at the base of this stem-loop to liberate a miRNA precursor (pre-miRNA) as a 60-70-nucleotide RNA hairpin \[[@B18]\]. The pre-miRNA hairpin is exported to the cytoplasm by exportin-5 \[[@B19]-[@B21]\] where it is further processed into a short dsRNA molecule by a second ribonuclease III-like enzyme, Dicer \[[@B22]-[@B24]\]. A single strand of this short dsRNA, the mature miRNA, is incorporated into a ribonucleoprotein complex. This complex directs transcript cleavage or translational repression depending on the degree of complementarity between the miRNA and its target site. RNA editing is the site-specific modification of an RNA sequence to yield a product differing from that encoded by the DNA template. Most RNA editing in human cells is adenosine to inosine (A-to-I) RNA editing which involves the conversion of A-to-I in dsRNA \[[@B25],[@B26]\]. A-to-I RNA editing is catalyzed by the adenosine deaminases acting on RNA (ADARs). The majority of A-to-I RNA-editing sites are in dsRNA structures formed between inverted repeat sequences in intronic or intergenic RNAs \[[@B25],[@B27]-[@B30]\]. Therefore, the double-stranded precursors of miRNAs may be substrates for A-to-I editing. Indeed, it has recently been shown that the pri-miRNA transcript of human miRNA miR-22 is subject to A-to-I RNA editing in a number of human and mouse tissues \[[@B31]\]. Although the extent of A-to-I editing was low (less than 5% across all adenosines analyzed), targeted adenosines were at positions predicted to influence the biogenesis and function of miR-22. This raises the possibility that RNA editing may be generally important in miRNA gene function \[[@B31]\]. In this study we have systematically investigated the presence of RNA editing in miRNAs. Results ======= To search for RNA-editing sites in human miRNAs, PCR product sequencing was performed from matched total cDNA and genomic DNA isolated from adult human brain, heart, liver, lung, ovary, placenta, skeletal muscle, small intestine, spleen and testis. Primers were designed to amplify pri-miRNA sequences flanking all 231 human miRNAs in miRBase \[[@B6]\]. Of these, 99 miRNA containing sequences were successfully sequenced in both directions and from duplicate PCR products from total cDNA of at least one tissue. Total cDNA sequence traces were compared with genomic DNA sequence traces from the same individual, and A-to-I editing was identified as an A in the genomic DNA sequence compared with a novel G peak at the equivalent position in the total cDNA sequence. In total, 12 of the 99 miRNA-containing sequences (13%) were subject to A-to-I RNA editing according to A-to-G differences between matched genomic DNA and total cDNA sequence traces from at least one tissue (Figure [1](#F1){ref-type="fig"}). These sequences were next oriented with respect to the strand of transcription of the miRNAs. In six cases the A-to-G changes were in the same orientation as the miRNA, and overlap the stem-loop structure of the miRNA, consistent with RNA editing of the pri-miRNA precursor transcript. In an additional case, A-to-I editing was observed in a novel stem-loop structure in sequence adjacent to the unedited miRNA miR-374. This novel stem-loop structure may represent a novel miRNA (Figure [2](#F2){ref-type="fig"}, novel hairpin). In the remaining five cases, the A-to-G changes were from the opposite strand to the miRNA (that is, U-to-C changes in the miRNA sequence). Although U-to-C editing of miRNA sequences cannot be ruled out, no editing of this type has previously been observed and no enzymes capable of catalyzing this conversion are known. The most likely explanation is that these are A-to-I edits in a transcript derived from the DNA strand complementary to the annotated miRNA gene. Consistent with this hypothesis, all of these sequences overlap, or are adjacent to, genes transcribed from the opposite strand to the annotated miRNA gene. To distinguish these sequences from the edited pri-miRNAs, these sequences are referred to here as edited antisense pri-miRNAs (Table [1](#T1){ref-type="table"}). One of the antisense pri-miRNAs contains editing sites overlapping the intended miRNA (miR-144) and miR-451, a recently identified miRNA that was not deliberately included in our list of 231 miRNAs. Collectively the 13 sequences were edited at 18 sites. Ten out of the 13 were edited at a single site. miR-376a and antisense miR-451 were each edited at two sites, and antisense miR-371 was edited at four sites. The extent of editing varied with editing site and with tissue, ranging from around 10% (for example, miR-151 in multiple tissues) to around 70% (antisense miR-371 in placenta). Overall, the levels of RNA editing observed were considerably higher than the approximately 5% editing previously reported for the -1 position of miR-22 \[[@B31]\]. Editing of miR-22 was not detectable by our method, presumably because the low levels of editing of this miRNA fall below our limits of detection. All miRNAs were found to be edited in multiple tissues, with the extent of editing varying from tissue to tissue (Figure [1](#F1){ref-type="fig"}). All novel A-to-I editing sites were found within the dsRNA stems of the predicted stem-loop structures (Figure [2](#F2){ref-type="fig"}). Of the seven editing sites in pri-miRNAs, four were in the 22-nucleotide mature miRNA. Three of these were within nucleotides 2 to 7, which are thought to be important for conferring binding-site specificity between the miRNA and its target sites \[[@B3]\]. Five out of seven editing sites in pri-miRNAs were at single nucleotide A:C mismatches flanked by paired bases. Similarly, five out of seven editing sites were in 5\'-UAG-3\' trinucleotides. These results are consistent with local structural and sequence preferences of RNA editing determined from A-to-I editing sites in inverted repeat sequences \[[@B25]\]. Three of the ten editing sites in antisense pri-miRNAs were in 5\'-UAG-3\' trinucleotides. Six of the ten editing sites were at A:C mismatches. Only one was at a single A:C mismatch, however, with the remainder at extended mismatches involving more than one consecutive nucleotide. Discussion ========== We have identified novel A-to-I editing sites in six out of 99 pri-miRNAs, indicating that at least 6% of all human miRNAs may be targets of RNA editing. We were only able to detect relatively high levels of editing, as illustrated by our failure to detect editing of miR-22, so this estimate is probably a conservative one. Moreover, our method is not strand specific, and cannot distinguish multiple overlapping transcripts from the same genomic locus. Thus, in regions of transcriptional complexity, it is likely that the sensitivity of our assay will be reduced. For example, even miRNAs that are 100% edited would appear to be unedited if transcribed at low levels compared with an unedited overlapping transcript from the opposite strand. We may also be unable to detect RNA editing if it occurs subsequent to the processing of the pri-miRNA (for example, by splicing) such that the binding sites for the PCR primers are removed. In addition to the edited pri-miRNAs, six antisense pri-miRNA transcripts derived from the opposite strand to the annotated miRNA were subject to A-to-I editing. There are many potential explanations for apparent editing on the opposite strand to the annotated miRNA. One possibility is that these sequences are actually due to U-to-C editing of the pri-miRNA. There are, however, no known U-to-C RNA editing enzymes capable of catalyzing such a reaction, and despite extensive searches for RNA editing sites, only a single U-to-C RNA editing site has been reported \[[@B32]\]. It is therefore more likely that these sequences represent an edited transcript from the opposite strand to the annotated miRNA. These transcripts could be another miRNA transcribed and processed from the genomic strand opposite the annotated miRNA, or they could be some other class of transcript, for example the intron of a gene overlapping the annotated miRNA but transcribed from the opposite DNA strand. Alternatively, these may be pri-miRNAs that have been incorrectly annotated to the wrong strand of the genome. To evaluate the possibility that the edited antisense pri-miRNAs are due to incorrect annotation of miRNAs to the wrong genomic strand, we examined previous experimental data obtained for these miRNAs. One of the edited antisense pri-miRNA sequences is derived from the DNA strand opposite the computationally predicted miR-215 \[[@B33]\]. The method used to predict miR-215 successfully predicted 81 out of 109 known miRNAs from a reference set, but around 20% (17/81) were predicted on the wrong strand of the genome \[[@B33]\]. Our data and the direction of overlapping transcripts suggest that miR-215 may have been annotated to the wrong genomic strand. An edited antisense miRNA sequence was also derived from the DNA strand opposite experimentally verified miRNA miR-133a \[[@B34]\]. This miRNA is present in the genome in two copies (miR-133a-1 and miR-133a-2). Copy miR-133a-2 is hosted within a gene transcribed in the same direction as the annotated miRNA gene. In contrast, copy miR-133a-1 overlaps a gene transcribed from the opposite strand. Cloning and expression analysis of miR-133a \[[@B34]\] provides proof that at least one copy of miR-133a is transcribed. As a result of this finding, both copies of miR-133a have been annotated according to the sequence of the cloned copy. Given the direction of overlapping transcripts, however, it remains possible that miR-133a-1 is transcribed from the opposite strand to miR-133a-2, giving rise to a different miRNA. Indeed, our results suggest that miR-133a-1 may have been incorrectly annotated. Similarly, both copies of experimentally verified miR-194 (miR-194-1 and miR-194-2) have been annotated according to the sequence of a cloned copy \[[@B34]\]. Our data and the presence of overlapping transcripts on the opposite strand suggest that miR-194-1 may also have been incorrectly annotated to the wrong genomic strand. In the case of both mir-133a and mir-194, the two copies would generate miRNAs that are perfectly complementary to one another. It has previously been suggested that pairs of complementary miRNAs play a role in miRNA regulation by forming miRNA:miRNA duplexes \[[@B35]\]. Our results suggest that RNA editing may add a further layer of regulation by disrupting complementarity in miRNA:miRNA duplexes. A further two edited antisense miRNA sequences (antisense mir-144 and antisense mir-451) overlap miRNAs that are annotated on the basis of their similarity to mouse miRNAs, and have not been cloned or shown to be expressed by northern blotting in human tissues. The remaining antisense miRNA sequence overlaps mir-371, which has been validated by cloning and northern blotting in human tissues and is therefore correctly annotated. The presence of edited nucleotides in pri-miRNA transcripts indicates that RNA editing occurs early in miRNA biogenesis. Subsequent processes that recognize sequence or structural features of the miRNA precursor could therefore potentially be affected by RNA editing. These include cleavage of the pri-miRNA by Drosha, export of the pre-miRNA to nucleus by exportin-5, cleavage of the pre-miRNA by Dicer, and miRNA strand selection for inclusion in the microprocessor complex. Indeed, it has recently been demonstrated that RNA editing of pri-miRNAs can result in suppression of processing by Drosha, and subsequent degradation of the unprocessed edited pri-miRNA \[[@B36]\]. Although it is unclear whether a miRNA that base-pairs with its target through an I:U wobble would be functional, another possibility is that RNA editing may alter target site complementarity. To investigate the effect of RNA editing of miRNAs on target-site complementarity, we used the miRanda software \[[@B37]\] to predict binding sites of edited miRNAs in 3\' UTRs, and compared these with the predicted binding sites of the equivalent unedited miRNAs. For each of the four pri-miRNAs with an editing site in the mature 22mer, the set of predicted targets of edited miRNAs differs from the predicted targets of edited miRNAs (Table [1](#T1){ref-type="table"}). For the three miRNAs in which the edited adenosine is at a position two to seven bases from the 5\' end of the miRNA (miR-151, miR376a and miR-379) over half of the targets of the edited miRNA are unique to the edited miRNA. In the case of miR-99a the difference is small, with only 5/75 (6%) target predictions differing between edited and unedited miRNAs. In all cases, the top ten predicted targets of the edited miRNA differ from the top ten predicted targets of the unedited miRNA (data not shown). To gain further insight into the potential biological impact of miRNA editing, we identified Gene Ontology (GO) terms in the \'cellular process\' category \[[@B38]\] which were over-represented in the predicted targets of edited and unedited miRNAs compared with all Ensembl genes (Figure [3](#F3){ref-type="fig"}). For the three miRNAs that are edited in the 5\' seed region (miR-151, miR-376a and miR-379), comparison of over-represented GO terms associated with the predicted targets of edited and unedited copies reveals distinct differences (Figure [3](#F3){ref-type="fig"}). Of particular interest are the additional terms that become over-represented; these include regulation of programmed cell death, biosynthesis, RNA metabolism, cell proliferation and transcription (Figure [3](#F3){ref-type="fig"}). RNA editing may therefore contribute to miRNA diversity by generating multiple different miRNAs from an initial pool of identical miRNA transcripts. For example, the total number of predicted targets of Hsa-mir-151 increases from 143 to 229 when taking into consideration both edited and unedited miRNAs. Editing of miRNAs may simultaneously alleviate and augment the gene-regulation effects of miRNAs by changing the concentration of individual miRNAs. Conclusion ========== We have performed the first systematic survey of RNA editing of human miRNAs. We have identified RNA editing sites in at least 6% of human miRNAs that may impact on miRNA processing, including edits that alter miRNA binding sites and contribute to miRNA diversity. Furthermore, our results suggest that some miRNA genes may have been incorrectly annotated to the wrong strand of the genome. This has implications for the interpretation of existing miRNA experiment data and future experimental design. Materials and methods ===================== Total RNA, total cDNA and genomic DNA ------------------------------------- For the initial screen of RNA editing in ten human tissues, total RNA and matching genomic DNA from the same tissue sample was obtained for human brain, heart, liver, lung, ovary, placenta, skeletal muscle, small intestine, spleen and testis from Biochain (Hayward, USA). For each tissue, sequence data was obtained from one individual. The donor was different for each tissue type. Total cDNA synthesis was performed using random nonamers (200 ng per 20 μl reaction) with Superscript III (Invitrogen, Carlsbad, USA) according to the manufacturer\'s instructions. Sequencing of pri-miRNAs ------------------------ Primers were designed to the genomic sequence in the vicinity of all 231 miRNA sequences in the miRNA registry version 7.0 \[[@B6]\], using primer3 \[[@B39]\]. PCR primer design was optimized to give PCR products of approximately 500 bp with at least 75 nucleotides either side of the predicted stem-loop structure. PCR primers were used to sequence PCR products in both directions on ABI3700 DNA sequencers. Sequence traces were quality scored using phred. Sequences with less than 70% of bases having a quality score of 20 or more were rejected. In the first stage of sequencing, duplicate PCR and sequencing was performed for each miRNA from each tissue. A miRNA was considered to be successfully sequenced if the following minimum sequence requirements were met for at least one tissue: good-quality sequence from both strands of one PCR, and good-quality sequence from at least one strand of a second PCR. Successfully sequenced miRNAs that were found to be edited were submitted to a second confirmation stage of sequencing. In the second stage of sequencing, quadruplicate PCR and sequencing was performed for each miRNA from each tissue. For each tissue, a miRNA was considered to be successfully sequenced if the following minimum sequence requirements were obtained: good-quality sequence from both strands of one PCR, and good-quality sequence from at least one strand of a second PCR. See Additional data file 1; primary sequence data is available from \[[@B40]\]. Detection and quantification of RNA editing ------------------------------------------- Sequences were visualized and compared in a gap4 database. A-to-I editing was identified as a novel G peak and a drop in peak height at As in a cDNA sequence relative to the equivalent peak in the matching genomic DNA sequence. The extent of RNA editing was estimated using a modified version of the comparative sequence analysis (CSA) method \[[@B41]\]. Briefly, this program normalizes a cDNA sequence trace to a genomic DNA reference trace by comparison of peak heights at unedited nucleotides. The drop in peak height between the DNA reference trace and the cDNA trace at the edited nucleotide is then reported as a percentage of the peak height in the genomic DNA reference trace. For each edited miRNA, the mean extent of editing for each tissue is calculated from all cDNA sequences obtained for that tissue. Analysis of novel RNA editing sites ----------------------------------- miRNA structures were obtained from the miRBase database \[[@B6]\]. Stem-loop structures of antisense miRNAs were generated by folding the antisense of the miRNA stem-loop sequence obtained from miRBase using MFOLD \[[@B42]\]. To predict edited and unedited miRNA target sites, miRanda (v3.0) \[[@B32]\] was used to scan the edited and unedited miRNA sequences against all human 3\' UTR sequences available from Ensembl v34. The algorithm uses dynamic programming to search for maximal local complementarity alignments, which correspond to a double-stranded antiparallel duplex. The new version of the miRanda algorithm (AJ Enright, personal communication) assigns Pvalues to individual miRNA-target binding sites, multiple sites in a single UTR, and sites that appear, from a robust statistical model \[[@B43]\], to be conserved in multiple species. The resulting targets were filtered based on Pvalue (p\< 0.001) to ensure a high degree of confidence in the predicted target sites. GO analysis ----------- GO terms from level 4 of the \'cellular process\' category were obtained for each human transcript from Ensembl. Over-representation for each term (O~term~) in a group of sequences with C terms is calculated as follows: ![](gb-2006-7-4-r27-i1.gif) ![](gb-2006-7-4-r27-i2.gif) where F~1~is the frequency of a term in the group being considered, F~2~is the frequency of a term in the whole genome and tis the term at level L. GO terms with low transcript counts (\< 3.0) were excluded from further analysis. Additional data files ===================== The following additional data are available with this paper online. Additional data file [1](#S1){ref-type="supplementary-material"} contains examples of edited sequence traces for each of the edited sites identified in this survey, and the coordinates of edited bases. Additional data file [2](#S2){ref-type="supplementary-material"} contains PCR primer information, details of the initial screen of miRNAs and annotation of edited miRNAs. Supplementary Material ====================== ###### Additional data file 1 Figures containing examples of edited sequence traces for each of the edited sites identified in this survey, and the coordinates of edited bases. ###### Click here for file ###### Additional data file 2 PCR primer information, details of the initial screen of miRNAs and annotation of edited miRNAs. ###### Click here for file Acknowledgements ================ We would like to thank all members of the cancer genome project for technical help sequencing miRNAs and the Wellcome Trust for funding. Figures and Tables ================== ![A-to-I RNA editing of miRNA precursors in human tissues. The extent of A-to-I editing at each editing site is indicated by the color scale. Each colored box represents the average extent of editing calculated from at least two PCR product sequences, at least one of which was sequenced in both directions. Gray boxes indicate miRNAs that could not be amplified. The number in brackets after the miRNA name is the position of the edited adenosine from the 5\' end of the pre-miRNA or equivalent antisense pre-miRNA from the miRNA registry.](gb-2006-7-4-r27-1){#F1} ![Positions of edited adenosines in human pri-miRNAs and antisense pri-miRNAs. Folded pri-miRNA structures were taken from the miRNA registry \[6\]. Antisense pre-miRNA structures were generated from the reverse complement pri-miRNA sequence using MFOLD \[38\]. Mature miRNA sequences of around 22 nucleotides and antisense mature miRNA sequences of around 22 nucleotides are indicated by red letters. Edited adenosines are highlighted in yellow. In antisense Hsa-mir-371, edited adenosines were found to reside in base-paired sequence extending beyond the annotated hairpin. Additional bases are in gray.](gb-2006-7-4-r27-2){#F2} ![GO term comparison of edited and unedited miRNA target predictions. For each edited miRNA, GO terms from level 4 of the \'biological process\' category that are over-represented in predicted targets of the unedited or edited miRNA (indicated by +) compared with all Ensembl genes were identified. All values are normalized and colored in terms of significance, with bright red cells indicating that a miRNA specifically targets genes in that GO functional class.](gb-2006-7-4-r27-3){#F3} ###### Predicted targets of edited and unedited miRNAs MicroRNA Edited only Unedited only Edited and unedited -------------- ------------- --------------- --------------------- Hsa-miR-151 86 84 59 Hsa-miR-376a 74 58 71 Hsa-miR-379 79 70 75 Hsa-miR-99a 5 8 70 Target predictions were performed using the miRanda software using a probability score cut-off of p\< 0.001. For each miRNA, the number of targets predicted for both edited and unedited miRNAs is shown against the number of targets predicted exclusively for edited miRNAs, and the number of targets predicted exclusively for unedited miRNAs.	{ "pile_set_name": "PubMed Central" }
Dear Sir, We read with interest the article "Repeat gas insufflation for successful closure of idiopathic macular hole following failed primary surgery" by Rishi et al.\[[@ref1]\] We describe two cases with failed primary macular hole (MH) surgery that closed following repeat fluid-gas exchange (FGE) and make some additional observations. Case 1 was a 55-year-old male with a 2-month history of blurred vision in his left eye. Best-corrected visual acuity (BCVA) was 20/120 in this eye, with a stage IV MH. He underwent 23 gauge pars plana vitrectomy with internal limiting membrane (ILM) peeling and 14% C3F8 gas injection, with prone positioning. At 4 weeks postoperatively, the MH was open with a cuff of subretinal fluid (SRF). He underwent repeat FGE with 14% C3F8 gas and prone positioning. At 4 weeks following the second intervention, the MH was found to be closed, and BCVA increased to 20/40 \[[Fig. 1](#F1){ref-type="fig"}\]. Case 2 was a 60-year-old male with a BCVA of 20/80 in his right eye and a stage IV MH of 1-month duration. Primary MH surgery resulted in a smaller but open MH with an elevated inferior edge at 1-month follow-up. Repeat FGE with 14% C3F8 gas and prone positioning was carried out resulting in a closed MH and BCVA of 20/40 1 month later \[[Fig. 2](#F2){ref-type="fig"}\]. ![(a, b) Color fundus photograph and vertical spectral-domain optical coherence tomography scan of case 1 at presentation showing stage IV macular hole (MH). (c, d) One month following primary surgery, the MH was open with a cuff of subretinal fluid. (e, f) Following repeat fluid-gas exchange, type 1 closure occurred at 1-month follow-up](IJO-62-1104-g001){#F1} ![(a, b) Color fundus photograph and vertical spectral-domain optical coherence tomography scan of case 2 at presentation showing stage IV macular hole (MH). (c, d) Following primary MH surgery, the size of the MH decreased but the inferior edge was elevated. (e, f) Repeat fluid-gas exchange was performed resulting in type 1 closure noted 1-month postoperatively](IJO-62-1104-g002){#F2} Nonclosure or reopening of MH after a primary MH surgery can be managed by repeat ILM peeling,\[[@ref2]\] additional FGE done externally (as an outpatient procedure)\[[@ref3]\] or internally (repeat surgery).\[[@ref1]\] Once the gas bubble has absorbed enough to allow adequate examination of the macula, an open MH with elevated edges and a cuff of SRF is unlikely to close or flatten without additional intervention.\[[@ref3]\] The authors describe the MH following primary surgery as "type 2 closure of MH with cuff of SRF" \[[Fig. 2](#F2){ref-type="fig"}\] and mention that the cuff of SRF was a factor favoring re-surgery in their case.\[[@ref1]\] Type 2 closure is used to denote attachment of the entire rim of the MH to the retinal pigment epithelium (RPE) with flattening of the cuff of SRF, and persistence of a foveal defect of the neurosensory retina, while type 1 closure indicates that the foveal defect is closed.\[[@ref4]\] Thus, the appearance of their case following primary surgery is incorrectly labeled as type 2 closure. Postoperatively, careful inspection needs to be done to exclude any epiretinal membrane that may be preventing closure of the MH. If there is no such traction, then an early surgical failure should respond to FGE with a long-acting gas and prone positioning.\[[@ref3]\] A gas tamponade in the vitreous cavity is believed to limit the movement of fluid into the hole, maintain apposition of the edges of the MH to the RPE and facilitate glial proliferation to seal the macular defect. The time taken for this wound healing response that leads to MH closure is probably variable.\[[@ref5]\] So while some MH may close earlier, some may require renewal of the foveal tamponade that provides additional time for the process of MH closure.\[[@ref3][@ref5]\] The use of a long acting tamponade can keep the fovea isolated from the fluid without stringent prone positioning, which may not be possible in a subset of patients. Thus, postoperative FGE with a long-acting gas to provide additional macular tamponade along with repeat patient counseling to ensure compliance with prone positioning appears to be a viable option after failed MH surgery. Successful closure of the MH is associated with improvement in visual acuity in these cases.	{ "pile_set_name": "PubMed Central" }
Background ========== The Drosophilaadult brain is a highly organized and complex structure that contains thousands of neurons (in the order of 10^5^) \[[@B1]\] exhibiting multiple cell-type identities, as characterized by various morphological, electrophysiological and molecular features. All these neurons arise from the mitotic activity of a small number of progenitor cells (neuroblasts (NBs)), which generate lineages of clonally related neurons via two proliferative phases of neurogenesis \[[@B2],[@B3]\]. The first phase of neurogenesis takes place during embryogenesis and starts with the specification and delamination of the NBs from the procephalic neurogenic region. Between embryonic stages 8 and late 11, around 100 NBs delaminate from this region on either side of the embryo in a reproducible spatiotemporal pattern \[[@B4]\]. Each NB assumes a unique identity, as revealed by the expression of a specific set of marker genes such as the proneural genes, gap genes and segment polarity genes \[[@B5]-[@B7]\], and gives rise to an invariant cell lineage through multiple rounds of asymmetric cell divisions. In each cell division, the NB self-renews and generates a smaller daughter cell (named the ganglion mother cell), which divides only once to give rise to two post-mitotic neurons or glial cells (reviewed in \[[@B8]-[@B10]\]). The first neurogenic process terminates at the end of embryogenesis, when most NBs stop dividing and enter a dormant phase called quiescence \[[@B11]\]. The cells so far generated (primary neurons) wire the larval nervous system and eventually may remodel during metamorphosis to contribute to the adult brain \[[@B12],[@B13]\]. The second phase of neurogenesis starts during the late first (L1) and second (L2) instar larval stage, when the inactive NBs resume mitotic activity and, through rounds of asymmetric cell divisions, generate the population of secondary neurons and glial cells that accounts for more than 90% of the adult brain \[[@B11],[@B14]\]. Hence, a complete NB lineage can be divided into two discrete cell populations, each containing the cells generated by the NB during distinct developmental phases (primary neurons during embryogenesis and secondary neurons during larval development) and each harboring multiple neuronal cell types. It has been proposed that neuron identity within a NB lineage depends on a combination of spatial and temporal cues provided, firstly, by the unique identity the NB acquires during its specification/delamination time \[[@B7],[@B15]\] and, secondly, as a result of a birth time/order-dependent mechanism \[[@B16]\], whereby cell-type specification of the nascent post-mitotic neurons depends on the identity of the progenitor temporal transcription factor expressed by the NB at each particular time during lineage progression \[[@B17]\]. The homeobox gene orthodenticle(otd), as a cephalic gap gene, is expressed in broad domains of the procephalic ectoderm during early neurogenesis, covering most of the protocerebral anlage and the anterior part of the deutocerebral anlage \[[@B18]\]. Subsequently, its expression is also detected in NBs delaminating from these domains, where it plays instructive roles important for cell viability and spatial identity of the nascent NBs \[[@B5]\]. It has also been proposed that otdmight control brain NB formation by triggering proneural gene expression \[[@B19]\]. Inactivation of otdat this early embryonic phase impairs NB formation and leads to a gap-like phenotype in the anterior head that includes the deletion of the protocerebral anlage and part of the deutocerebral anlage \[[@B20]\]. Later in development, otdexpression is also detected in post-mitotic neurons of the developing brain and ventral nerve cord, not only in the embryo, but also in the larval and pupal brain and even in mature neurons of the adult brain. In this regard, flies homozygous for an otdviable hypomorphic mutation called ocelliless(oc) show developmental defects that affect the protocerebral bridge, an important neuropile structure in the adult brain that is part of the central complex \[[@B20]\]. However, whether this phenotype is due to the altered expression of otdin progenitor cells, post-mitotic cells or both is not known. A comprehensive study of gene function during neuronal cell fate specification requires a previous and thorough cell lineage analysis and demands cell type-specific molecular markers to trace the cells under study. In this paper, we have focused our attention on the array of dopaminergic (DA) neurons that populate the Drosophilacentral brain during larval development. We have used cell lineage tracing genetic techniques together with immunoreactivity against the enzyme tyrosine hydroxylase (TH), the rate-limiting enzyme in dopamine biosynthesis \[[@B21],[@B22]\], to study their development, phenotypic maturation and lineage relationships. Interestingly, one cluster of clonally related DA neurons expresses otdfrom early L1 to adulthood, allowing us to examine the post-mitotic role of otdin controlling identity and/or survival of DA neurons in the Drosophilalarval and adult brain. Results ======= Birth, clustering and phenotypic maturation of dopaminergic neurons in the larval central brain ----------------------------------------------------------------------------------------------- Using immunoreactivity against the enzyme TH as a DA neuron-specific molecular marker, an array of 21 DA neurons could be visualized in the central part of each brain hemisphere during the third instar larval stage (L3; Figure [1A](#F1){ref-type="fig"}). The stereotypical arrangement of their cell bodies in various groups has been used previously to define four clusters of DA neurons that occupy distinct anatomical positions in the central brain \[[@B23]-[@B25]\]: dorso medial 1 cluster (DM1: four cells), dorso medial 2 cluster (DM2: four cells), dorso lateral 1 cluster (DL1: seven cells) and dorso lateral 2 cluster (DL2: six cells). As judged by their projection and innervation patterns, DM2 and DL1 cell clusters contained apparently homogeneous populations of DA neurons. On the contrary, both DM1 and DL2 clusters could be further subdivided into two subclusters according to their differential projection patterns. A single neuron in the DM1 cluster (named DM1a) projected ventrally into the lower part of the ipsilateral brain lobe (Figure [1B](#F1){ref-type="fig"}), whereas the remaining three DM1 DA neurons (named the DM1b cell cluster) innervated more ventrally localized regions of the brain lobe (subesophagial ganglion) and further extended into the thoracic segments of the ventral ganglion (Figure [1C](#F1){ref-type="fig"}). DM2 neurons projected ipsilaterally into the anterior part of the protocerebrum (Figure [1D](#F1){ref-type="fig"}), whereas DL1 neurons were characterized by dorsally projecting neurites that bifurcated into dorsal and ventral branches before crossing the midline to reach the contralateral brain lobe (Figure [1E](#F1){ref-type="fig"}). Although described as homogeneous, DM2 and DL1 cell clusters are possibly heterogeneous at the single-cell level. Indeed, six different cell types with slightly different innervation patterns have been described within the DL1 cell cluster using single-cell labeling techniques \[[@B25]\]. Similarly to the DM1cell cluster, the six DL2 DA neurons displayed two distinct projection patterns: four neurons (named the DL2a cell cluster) projected their neurites dorsally into the anterior part of the ipsilateral brain lobe (Figure [1F](#F1){ref-type="fig"}). The remaining two cells (named the DL2b cell cluster) projected laterally and arborized in the ventral part of the brain lobe before crossing the midline to innervate the contralateral brain lobe (Figure [1G,G\'](#F1){ref-type="fig"}). ![Birth, clustering and differentiation of larval central brain dopaminergic neurons. (A)Third instar larval stage (L3) brain showing bilaterally symmetrical groups of TH-positive DA neurons clustered according to the position of the cell bodies within the cortex (the number of neurons per cluster is given in parentheses) and the neurite projection patterns. Scale bar: 50 μm. (B-G\')Neurite projection patterns. Wild-type cell clones were induced in progenitor cells during early embryogenesis (3 to 7 h after egg laying (AEL)) and analyzed during L3. The clones were labeled with membrane tethered green and red flourescent proteins (both in grey) using the Twin-spot MARCM technique together with the TH-gal4driver. When possible, neurite projection patterns of the entire cell cluster are shown (B,E,F); otherwise, just one DA neuron as a representative for the cell cluster is produced (C,D). (G,G\') The neurite projection pattern (red arrowheads) of DL2b DA neurons (yellow arrows) is shown in two consecutive confocal optical sections. (H)Early L1 brain (24 to 28 h AEL) showing clusters of TH-positive cells (the number of DA neurons per cluster is given in parentheses) Scale bar: 10 μm. (I)Most larval DA neurons in the central brain are born during early neurogenesis between 3 to 13 h AEL. Wild-type cell clones, labeled with the MARCM technique in combination with the TH-gal4driver, were induced in progenitor cells at different times during development as indicated and analyzed in L3 brains. Images, except for (G,G\'), represent Z projections of individual confocal optical sections. AEL, after egg laying; DA, dopaminergic; DL, dorso lateral; DM, dorso medial; L, instar larval stage; MARCM, mosaic analysis with a repressible cell marker; ML, midline; TH, tyrosine hydroxylase.](1749-8104-6-34-1){#F1} We next determined the time of birth of the larval central brain DA neurons by using the MARCM (mosaic analysis with a repressible cell marker) technique in combination with the TH-Gal4driver. Flippase recognition target (FRT)-mediated mitotic recombination was randomly induced, exposing progenitor cells to a one hour heat-shock treatment (37°C) at different developmental stages and L3 brains were assayed for the presence of green fluorescent protein (GFP)-labeled DA neurons. When mitotic recombination was induced during early L1, a time point commonly used to label neurons born during larval development, GFP-positive DA neurons were not detected in the resulting wild-type cell clones. This result agreed with the observation that THexpression was already detectable in early L1 brains (Figure [1H](#F1){ref-type="fig"}). On the contrary, a heat-shock treatment applied during early embryogenesis efficiently labeled DA neurons in the L3 central brain, with the highest labeling efficiencies achieved when cell clones were induced at early embryonic stages (between 3 and 7 and 7 and 11 hours after egg laying (AEL); Figure [1I](#F1){ref-type="fig"}). These results demonstrate that DA neurons present in the L3 central brain are primary neurons that arise during early embryogenesis. Despite their early embryonic origin, the larval central brain DA neurons did not express the cell type-specific marker gene THduring embryogenesis, and even at late embryonic stages (stage 17) anti-TH immunoreactivity in the central nervous system was restricted to the ventral nerve cord (data not shown and \[[@B24]\]). However, during early L1 (24 to 28 h AEL) most of the central brain DA neurons already displayed THexpression (Figure [1H](#F1){ref-type="fig"}), with two exceptions: the DM1a DA neuron started to show anti-TH labeling during mid-late L1, whereas a DL2a DA neuron only showed anti-TH immunoreactivity at mid-late L3 (data not shown). In summary, DA neurons present in the central brain of the Drosophilalarva at L3 are generated during early embryogenesis and most of them acquire a mature neurotransmitter phenotype during early L1. Dopaminergic neurons present in the central brain during larval development are generated by seven neuroblast lineages ---------------------------------------------------------------------------------------------------------------------- In Drosophila, clonally related neurons typically remain clustered in the mature brain and project their neurites into specific neuropile compartments \[[@B26],[@B27]\]. To address whether the anatomical clustering of DA neurons in the L3 central brain is due to a clonal origin, we analyzed lineage relationships among DA neurons in each cell cluster. For this purpose, we utilized the twin-spot MARCM technique in combination with the tubulin-Gal4driver. Wild-type cell clones were induced during early embryogenesis (3 to 7 h AEL) and assayed in early L1 brains (24 to 28 h AEL) for the presence of GFP- and red fluorescent protein (RFP)-labeled NB clones containing TH-positive neurons. Although at this developmental stage two DA neurons still did not express the cell type-specific marker gene THand thus could not be considered in our analysis, two reasons justified our decision. Firstly, the reduced number of total cells present in the larval brain during L1 facilitated an easier lineage analysis. Secondly, we observed that the tubulinpromoter underwent partial down-regulation in primary neurons during L3 (data not shown), impairing a reliable analysis of the lineage relationships among DA neurons. Our analysis revealed that all the DA neurons assigned to the DM1b (Figure [2A](#F2){ref-type="fig"}), DM2 (Figure [2B](#F2){ref-type="fig"}), DL2a (Figure [2D](#F2){ref-type="fig"}) and DL2b (Figure [2E](#F2){ref-type="fig"}) clusters were contained within individual NB clones and thus were clonally related. By contrast, we only found NB clones containing at most six out of the seven DL1 DA neurons (Figure [2C,C\'](#F2){ref-type="fig"}), indicating that two NBs generate the seven DL1 DA neurons. ![Twin-spot MARCM lineage analysis of larval central brain dopaminergic neurons at L1 reveals clonal relationship between dopaminergic neurons. (A-E)Membrane-tethered GFP- and RFP-labeled wild-type NB clones containing TH-positive (blue) DA neurons from different DA cell clusters: DM1b cluster (A), DM2 cluster (B), DL1 cluster (C,C\'), DL2a cluster (D) and DL2b cluster (E). All panels represent individual confocal optical sections (0.5 μm thick). The red asterisk in C~2~identifies the DL1 DA neuron that belongs to a different NB lineage (DL1b). The yellow asterisk in C~2~and C\'~2~identifies the same neuron in different optical sections. Scale bars: 10 μm. DA, dopaminergic; DL, dorso lateral; DM, dorso medial; GFP, green fluorescent protein; L, instar larval stage; NB, neuroblast; RFP, red fluorescent protein; TH, tyrosine hydroxylase.](1749-8104-6-34-2){#F2} As mentioned above, due to their delayed THexpression, two DA neurons were initially not included in the lineage analysis. In order to validate their allocation to the respective DA cell clusters, we induced MARCM-labeled wild-type NB clones during early embryogenesis (3 to 7 h AEL) and analyzed L3 brains. We tried to circumvent the down-regulation of the tubulinpromoter by including an additional copy of the UAS-CD8::GFPtransgene in the genotype of the analyzed larvae. Although this strategy did not always work reliably (for example, we did not observe single NB clones containing the group of six DL1 DA neurons), we were still able to assign the missing DM1 DA neuron to an independent NB lineage (named DM1a), as well as to confirm the allocation of the missing DL2 DA neuron to the DL2a NB lineage (Additional file [1](#S1){ref-type="supplementary-material"}). In summary, we found that seven NB lineages generate the DA neurons present in each hemisphere of the Drosophilalarval central brain and that their clustered appearance is, at least in part, a consequence of a clonal relationship. Moreover, the assignment of DA neurons to particular NB lineages provides access to study genetic mechanisms of DA neuron cell fate specification. orthodenticleis expressed in DL2a dopaminergic neurons -------------------------------------------------------- The homeobox containing gene otdis broadly expressed in the anterior ectoderm during the blastoderm stage and is necessary for the specification of most of the NBs that populate the protocerebrum and part of the deutocerebrum \[[@B5],[@B19]\]. Using Otd immunostainings, we found that otdwas expressed in DL2a DA neurons already at early L1 (Figure [3A](#F3){ref-type="fig"}). At mid-late L3, when the entire complement of DA neurons in the larval central brain was already visible with anti-TH labeling, Otd was specifically detected in the four DL2a DA neurons (Figure [3C](#F3){ref-type="fig"},E~1~-E~3~). ![**otdis expressed in DL2a dopaminergic neurons in the larval brain. (A,C)**otd(blue) is expressed in a specific cluster of TH-positive (red) DA neurons of the larval central brain during L1 (A) and L3 (C). (B,D)In oc^γa1^hemizygous larvae, two out of three DL2a DA neurons at L1 (B) and three out of four DL2a DA neurons at L3 (D) are not detected using anti-TH immunoreactivity. Scale bars: 10 μm (A,B); 20 μm (C,D). (E)The oc2enhancer is active in three out of the four DL2a DA neurons at L3. Scale bar: 10 μm. (E~1~-E~3~)otdexpression and (E\'~1~-E\'~3~)oc2enhancer activity in DL2a DA neurons are analyzed in individual optical sections. (E~4~,E\'~4~)DL2b DA neurons display neither otdexpression nor oc2enhancer activity. Scale bars: 5 μm (E~1~-E~4~). (F,F~1~)In L3 oc^γa1^hemizygous larvae, oc2enhancer activity in DL2a DA neurons is lost. Scale bar: 10 μm. (G)otd^YH13^mutant cell clones induced during early embryogenesis impair the development of DL2a DA neurons (white arrow). Yellow arrows point to the DL2b DA neurons in the left hemisphere Scale bar: 50 μm. All the panels, except for (E~1~-E~4~) and (E\'~1~-E\'~4~), represent Z projections of individual confocal optical sections. DA, dopaminergic; DL, dorso lateral; GFP, green fluorescent protein; L, instar larval stage; oc, ocelliless; otd, orthodenticle; TH, tyrosine hydroxylase.](1749-8104-6-34-3){#F3} To analyze the significance of otdexpression in the specification and/or survival of post-mitotic DL2a DA neurons, we made use of the hypomorphic otdallele oc. ochemizygous flies are viable, but they show a lack of ocelli and associated bristles in the head vertex. These flies also lack the protocerebral bridge, a neuropile structure present in the fly adult brain that is part of the central complex \[[@B20]\]. These phenotypes arise as a consequence of chromosomal rearrangements that remove cis-acting regulatory sequences (ocenhancer) important for otdexpression during ocelli and protocerebral bridge development \[[@B18],[@B28]\]. During early L1, otdexpression in ochemizygous larvae was downregulated and the number of TH-positive DL2a DA neurons was reduced to one or two neurons, as compared to the three DL2a DA neurons present in wild-type brains (Figure [3B](#F3){ref-type="fig"}). In L3 brains, the lack of anti-TH immunoreactivity affected three out of the four DL2a DA neurons (Figure [3D](#F3){ref-type="fig"}). To investigate whether this phenotype is due to the loss of DL2a DA neurons per seor to THdownregulation in these neurons, we made use of the ocenhancer and the oc2-Gal4driver \[[@B28]\] as an alternative way of labeling DL2a DA neurons. Detection of ocenhancer activity in DL2a DA neurons in ocmutant L3 brains would imply that otdis involved in the activation and/or maintenance of THexpression in these neurons. By contrast, the absence of ocenhancer activity in DL2a DA neurons in ocmutant L3 brains would indicate otdis primarily necessary for survival of DL2a DA neurons. Reporter gene expression under the control of the oc2-Gal4driver was detected in three out of the four DL2a DA neurons in wild-type L3 brains (Figure [3](#F3){ref-type="fig"}E\'~1~-E\'~3~). Interestingly, when oc2-gal4transcriptional activity was analyzed in ocmutant L3 brains, reporter gene expression was not detected in DL2a DA neurons (Figure [3F](#F3){ref-type="fig"},F~1~). We also analyzed the relevance of otdexpression during embryogenesis for the generation of DL2a DA neurons. We induced MARCM-labeled cell clones mutant for a null otdallele during early embryogenesis and assayed L3 brains for the presence of TH-positive cells within these otdmutant clones. Very few otd- NB clones were recovered in the central brain and none of them affected the DL2a DA neurons. However, L3 brains lacking some of the DL2a DA neurons were observed (white arrow in Figure [3G](#F3){ref-type="fig"}), though the corresponding NB clone was not detected. The simplest interpretation for this result is that Otd depletion during early neurogenesis induces NB cell death and, as a result, loss of DL2a DA neurons. Taken together, the lack of anti-TH labeling observed in three DL2a DA neurons in ocmutant L3 brains is not due to THdownregulation, but likely reflects the absence of these neurons. Thus, we conclude that otdis necessary for survival of the larval DL2a DA neurons. Adult PPL2 dopaminergic neurons derive from the larval DL2a cluster ------------------------------------------------------------------- Similarly to the larval brain, DA neurons also cluster in the adult Drosophilabrain and these clusters have been annotated according to their anatomical position \[[@B23],[@B24],[@B29]\]. In order to investigate whether adult DA neurons express otd, we assayed wild-type young adult brains (3 to 7 days old after eclosion) for the co-expression of otdand TH. We observed that seven DA neurons, assigned to the protocerebral posterior lateral 2 (PPL2) cluster, expressed otd(Figure [4A](#F4){ref-type="fig"},A~1~-A~4~) and five of them also showed oc2enhancer transcriptional activity (Figure [4A](#F4){ref-type="fig"},A\'~1~-A\'~4~). Interestingly, the entire PPL2 cluster was not detected in ocmutant hemizygous flies using anti-TH immunoreactivity (Figure [4B](#F4){ref-type="fig"}). Also, targeted knockdown of otdin DA neurons by RNA interference (RNAi) resulted in a partial phenocopy of this mutant phenotype (THexpression was detected in three out of the seven PPL2 DA neurons; Figure [4C](#F4){ref-type="fig"},C~1~-C~3~) and this effect could be rescued by the simultaneous expression of the anti-apoptotic gene P35(Figure [4](#F4){ref-type="fig"}D~1~-D~4~). Otd protein levels in the rescued DA neurons (cells 1, 2, 5 and 7) were indistinguishable from background levels or drastically reduced (Figure [4](#F4){ref-type="fig"}D\'~1~-D\'~4~); yet, cell viability and THexpression were recovered by P35co-expression. These results indicate that otdis dispensable for THexpression in PPL2 DA neurons and it is mainly required for their survival in the adult brain. ![**otdis expressed in PPL2 dopaminergic neurons in the adult brain. (A)**otd(blue) expression in the PPL2 cell cluster of TH-positive (red) DA neurons in the adult brain. The PPL2 cell cluster labeled with a yellow asterisk is magnified and otdexpression (A~1~-A~4~)and oc2enhancer activity (A\'~1~-A\'~4~)are analyzed in single optical sections. Scale bars: 10 μm. (B)In oc^γa1^hemizygous flies, the seven PPL2 DA neurons are not detected using anti-TH immunoreactivity. (C)Targeted expression of an otd-specific RNA interference construct in DA neurons (using the TH-gal4driver) impairs the viability of four PPL2 DA neurons in the adult brain of otd^YH13^heterozygous flies (7 days old). (C,C~1~-C~3~)THand otdexpression in the resulting PPL2 cluster (labeled with a yellow asterisk in (C)) is analyzed at higher magnification. Scale bars: 10 μm. (D~1~-D~4~)Targeted expression of P35in Otd-depleted DA neurons (using the TH-gal4driver) rescues cell viability and THexpression in PPL2 DA neurons in the adult brain of otd^YH13^heterozygous flies (7 days old). Pictures represent single optical sections. Scale bars: 10 μm. (D\'~1~-D\'~4~)Otd protein levels in four PPL2 DA neurons (1, 2, 5 and 7) are not recovered. Panels (A-C,C~1~-C~3~) represent Z projections of individual confocal optical sections. Scale bars in (A-C): 50 μm. DA, dopaminergic; GFP, green fluorescent protein; oc, ocelliless; otd, orthodenticle; PPL, protocerebral posterior lateral; RNAi, RNA interference; TH, tyrosine hydroxylase.](1749-8104-6-34-4){#F4} otdexpression and oc2enhancer activity in DA neurons of the larval and adult brain might establish an identity connection between the four larval DL2a DA neurons and the seven PPL2 DA neurons present in the adult brain. However, where did the additional three DA neurons present in the adult PPL2 cluster come from? A closer examination of L3 of the NB lineage that generates the DL2a DA neurons (Figure [5A-C](#F5){ref-type="fig"}) revealed that otdwas expressed not only in the four TH-positive neurons, but also in three adjacent TH-negative cells (arrows in Figure [5A,B](#F5){ref-type="fig"}). These three cells might represent neurons born during larval development that acquire anti-TH immunoreactivity during pupal stages or, alternatively, they might be undifferentiated embryonic neurons that undergo phenotypic maturation only during pupal development. To distinguish between these two possibilities, twin-spot MARCM-labeled wild-type cell clones were induced in combination with the TH-gal4driver during early L1 and adult brains were assayed for the presence of labeled DA neurons. In 13% of the analyzed adult brains (n = 30), one or two PPL2 DA neurons were labeled (Figure [5D-E](#F5){ref-type="fig"}), yet marker gene expression was not detected in three PPL2 DA neurons simultaneously. This result indicates that at least two PPL2 DA neurons in the adult brain are born during larval development. Conversely, the remaining five PPL2 DA neurons are likely to be of embryonic origin. Together, our data support the notion that the adult PPL2 cluster is derived from the larval DL2a cluster. ![PPL2 dopaminergic neurons in the adult brain are lineage related to the larval DL2a dopaminergic cell cluster. (A-C\')DL2a NB lineage (labeled with membrane tethered GFP) was analyzed during L3 using individual optical sections. Within the NB clone, seven cells are Otd-positive (blue). Three of them (arrows) only express otd, whereas the other four express both otdand TH(red). Scale bars: 10 μm. (D)Twin-spot MARCM-labeled wild-type cell clones were induced during early L1 and analyzed in the adult brain. Scale bar: 50 μm. (E-E\'\')Magnified view of the PPL2 cell cluster showing two PPL2 DA neurons (arrows) labeled with membrane tethered RFP. Two additional RFP-labeled neurons (arrowheads) do not express THand correspond to neurons in which the TH-gal4driver is ectopically activated. Scale bars: 10 μm. DA, dopaminergic; GFP, green fluorescent protein; L, instar larval stage; MARCM, mosaic analysis with a repressible cell marker; NB, neuroblast; otd, orthodenticle; PPL, protocerebral posterior lateral; RFP, red fluorescent protein; TH, tyrosine hydroxylase.](1749-8104-6-34-5){#F5} Discussion ========== Seven neuroblasts generate the larval central brain dopaminergic neurons ------------------------------------------------------------------------ The Drosophilalarval central brain contains 21 DA neurons per hemisphere during L3, which express the cell type-specific marker gene TH\[[@B21],[@B22]\]. Different methods have been proposed to classify and annotate these neurons according to anatomical criteria (position of the cell bodies within the cortex and/or projection pattern of the axonal tracts) \[[@B23]-[@B25]\]. In this paper, we have analyzed these neurons from a developmental point of view and classified them according to their lineage relationship (Figure [6](#F6){ref-type="fig"}). The MARCM technique is a powerful tool to study lineage progression and cellular pedigrees during Drosophilabrain development \[[@B30]\]. It allows the labeling of progenitor cells and their offspring at different times during development, depending on the timing of a heat-shock-induced flippase-mediated mitotic recombination event \[[@B31]\]. Using this technique, we have shown that the larval central brain DA neurons are primary neurons born during early embryogenesis. However, when analyzing the lineage relationship among these neurons, two major problems were encountered. Firstly, implicit in the technique is the fact that a labeled NB clone is accompanied by a non-labeled twin clone (two post-mitotic cells derive from the first ganglion mother cell born just after the mitotic recombination event). The exclusion of two cells from the lineage analysis is negligible when larval lineages are analyzed (the average size of a standard larval lineage at L3 is 120 cells \[[@B32]\]). However, embryonic lineages are small (on average between 10 and 20 cells at the end of embryogenesis \[[@B3]\]) and the exclusion of two cells can be significant. Secondly, MARCM-labeled NB clones induced during early embryogenesis can only be visualized with a considerable delay after their generation (from L2 onwards) due to the persistence of the Gal80 repressor protein \[[@B30]\]. These two problems have recently been circumvented by the development of the twin-spot MARCM technique \[[@B33]\]. This technique not only allows the visualization of cell clones earlier in development but also differentially labels the NB clone and the twin clone; thus, the study of the entire NB lineage is now possible. Using this technique, we have analyzed the lineage relationship among the DA neurons present in the Drosophilacentral brain during larval development. We found that seven NB lineages generate the 21 DA neurons present in the larval central brain (Figures [2](#F2){ref-type="fig"} and [6B](#F6){ref-type="fig"}; Additional file [1](#S1){ref-type="supplementary-material"}): DM1a (one DA neuron), DM1b (three DA neurons), DM2 (four DA neurons), DL1a (six DA neurons), DL1b (one DA neuron), DL2a (four DA neurons) and DL2b (two neurons). At large, the lineage analyses agree with the clustering of DA neurons according to anatomical criteria, supporting the general assumption that cell bodies arrangement and axonal projection patterns are reliable ways to classify neurons in Drosophila. Just in the case of the DL1 cell cluster was a discrepancy found. The cell bodies of the seven DL1 DA neurons are compactly arranged in a cell cluster that occupies medial-lateral positions in the L3 central brain and their neurites display similar projection patterns (Figure [1E](#F1){ref-type="fig"}). Yet, six DL1 DA neurons are clonally related (DL1a NB lineage; Figure [2C,C\'](#F2){ref-type="fig"}) and the remaining DL1 DA neuron (red asterisk in Figure [2](#F2){ref-type="fig"}C~2~) is generated by an additional NB (DL1b NB lineage). For future functional studies, it would be interesting to find molecular markers differentially labeling these two populations of DL1 DA neurons. ![Summary of the classification of dopaminergic neurons present in the larval central brain at L3, according to anatomical and lineage relationship criteria. (A)A cartoon depicting a dorsal view of a L3 larva central brain showing cell body locations and neurite projection patterns of the different types of DA neurons. For simplicity, just one neuron per cluster and per brain hemisphere is shown. The cartoon does not pretend to precisely reproduce the innervation pattern of every type of DA neuron, but just gives a general description. (B)Table showing the distribution of DA neurons in different NB lineages at different times during development. The lineage harboring Otd+ DA neurons is highlighted in blue. DA, dopaminergic; DL, dorso lateral; DM, dorso medial; L, instar larval stage; NB, neuroblast; PPL, protocerebral posterior lateral.](1749-8104-6-34-6){#F6} Otd acts as a survival factor in DL2a dopaminergic neurons ---------------------------------------------------------- Most studies involving the homeodomain transcription factor Otd in central nervous system development in Drosophilahave dealt with its role in the specification and proliferation of progenitor cells during early neurogenesis \[[@B5],[@B19]\], whereas a possible function in post-mitotic neurons has been largely overlooked. Our observation that otdis expressed in the DL2a DA neurons during larval development prompted us to investigate its role in the specification and/or survival of this DA cell cluster. According to anti-TH labeling, DL2a DA neurons mature mainly during early L1. Thus, null otdalleles, which are embryonic lethal, could not be used in our analysis. Therefore, we investigated the hypomorphic otdallele oc. We found that in ocmutant hemizygous larvae, otdexpression in dorsolateral regions of the central brain was reduced and, as a consequence, only one of the four DL2a DA neurons showed anti-TH labeling during L3. The failure to detect three of the four DA neurons can be due to a defect in the regulation of THexpression or to the loss of DA neurons per se. Several lines of evidence support the latter hypothesis. Firstly, a general regulator of THexpression would be expected to be present in all or most of the central brain DA neurons; yet, otdexpression during larval development is restricted to the DL2a DA cell cluster. Secondly, misexpression of otdin randomly induced cell clones in the central brain during larval development does not result in ectopic TH-expressing DA neurons (data not shown). Thirdly, labeling of DL2a DA neurons with the oc2-gal4driver shows that reporter gene expression is also abolished in ocmutant hemizygous larvae during L3. The oc2enhancer has been shown to be positively regulated by otdduring ocelli development \[[@B28]\] and might not, therefore, be suitable to label DL2a DA neurons in an otd-independent way. However, a minimal version of this enhancer harboring the characterized Otd binding site (oc7) was active in the ocelli primordium \[[@B28]\], but did not show enhancer activity in DL2a DA neurons during larval development (data not shown). This indicates that the oc2enhancer is differentially regulated in the ocelli primordium and in DA neurons during development and, hence, the oc2-gal4driver may be used to label DL2a DA neurons in an otd-independent fashion. Taken together, our observations support the hypothesis that otdexpression is required for survival of DL2a DA neurons during larval development. DL2a dopaminergic neurons survive into adulthood and participate in the PPL2 dopaminergic cell cluster ------------------------------------------------------------------------------------------------------ The wild-type Drosophilaadult brain is populated by about 200 DA neurons distributed in several bilaterally symmetric clusters \[[@B23],[@B24],[@B29]\]. The PPL2 cluster contains seven cells that express otdand five of them also show oc2enhancer activity in young adult brains (Figure [4](#F4){ref-type="fig"}A~1~-A~4~). Similarly to the larval brain, otdexpression in PPL2 DA neurons seems to be necessary for their survival, since neither anti-TH immunoreactivity nor transcriptional activity of the oc2enhancer is detected in ocmutant adult brains (Figure [4B](#F4){ref-type="fig"} and data not shown). Moreover, the effects of targeted depletion of Otd in PPL2 DA neurons (loss of cell viability and/or THexpression; Figure [4C](#F4){ref-type="fig"}) can be rescued by the simultaneous expression of the anti-apoptotic gene P35, pointing out a role in cell survival as the main function of otdin PPL2 DA neurons. Altogether, the simplest interpretation for these results would be that otdexpression labels homologous DA neuron populations in both the larval (DL2a cell cluster) and adult (PPL2 cell cluster) brains and, hence, both clusters contain the same DA neurons. The discrepancy in cell number between both clusters of DA neurons can be interpreted by analyzing the NB lineage responsible for the generation of the DL2a DA neurons. At L3, this lineage contains seven otdexpressing cells, four of them are primary neurons that have already undergone maturation and express TH. The other three cells might represent immature secondary neurons that differentiate during pupal stages to give rise to the additional three DA neurons present in the adult PPL2 cluster. The distinction between early-differentiating (four cells) and late-differentiating (three cells) PPL2 DA neurons finds support in the targeted depletion of Otd in DA neurons by RNAi. Expression of an otd-specific RNAi construct in DA neurons (using the TH-Gal4driver) has no effect on the larval brain (data not shown), but impairs the viability of four PPL2 DA neurons in the adult brain. Since these four cells differentiate during larval development, the RNAi machinery would have more time to completely deplete Otd than in the case of the late differentiating DA neurons. Further support for this interpretation also comes from the analysis in the adult brain of wild-type twin-spot MARCM cell clones induced during early L1. According to this analysis, at least two PPL2 DA neurons in the adult brain are secondary neurons, whereas the third DA neuron might represent an undifferentiated primary neuron that only matures during pupal development. Recently, the expression of Otx2, an otdortholog, in DA neurons in the mouse adult brain has also been reported \[[@B34]\]. It is selectively expressed in the central DA neurons of the ventral tegmental area, where it is cell autonomously required to antagonize identity features of the dorsal-lateral ventral tegmental area DA neurons \[[@B35]\]. Thus, contrary to Drosophila, depletion of Otx2 in these DA neurons does not induce cell death, but it changes neuron subtype identity. Interestingly, otx2expression in these DA neurons has been associated with their reduced vulnerability to Parkinsonian neurodegeneration \[[@B35]\]. Finally, in ocmutant adult flies most of the protocerebral bridge, a neuropile structure that is part of the central complex, is also missing \[[@B20]\]. In several behavioral paradigms, these mutant flies walk slowly and show altered orientation behavior toward visual objects \[[@B36],[@B37]\]. It has been recently proposed that the protocerebral bridge is an essential part of a visual targeting network that transmits directional clues to the motor output \[[@B37]\]. Thus, with regards to the data presented here, it would be interesting to analyze whether the lack of PPL2 DA neurons in ocmutant adult flies contributes to the behavioral phenotypes observed in these mutant flies. Conclusions =========== Using MARCM and twin-spot MARCM techniques together with anti-TH immunoreactivity, we have classified the 21 DA neurons present in the Drosophilalarval central brain into seven clusters of clonally related DA neurons. The homeobox gene otdis specifically expressed in DA neurons belonging to one of these clusters (DL2a cluster); thus, otdexpression differentially labels a subset of DA neurons. Furthermore, by taking advantage of an otdhypomorphic mutation and the oc2-Gal4reporter line, we have established a cell lineage relationship between the larval DL2a and the adult PPL2 DA cell clusters. We also studied the role of otdin the survival and/or cell fate specification of these post-mitotic neurons. Contrary to mice, where Otx2expression in DA neurons of the adult brain is necessary for neuron subtype identity, otdis required in the Drosophilalarval and adult brain for survival of DL2a and PPL2 DA neurons. These findings suggest that otdacts as a post-mitotic selector gene whose differential expression among DA neurons might help to establish functional differences. Materials and methods ===================== Fly strains, clonal analysis and RNAi experiments ------------------------------------------------- Flies were reared on standard medium at 25°C. The following transgene and reporter lines were used: UAS-P35(Bloomington Drosophila Stock Center, Bloomington, Indiana, USA), UAS-otd(J Blanco, unpublished), oc2-gal4\[[@B28]\], TH-gal4\[[@B24]\]. Mutant alleles used in this study: oc^γa1^, otd^YH13^\[[@B38]\]. Mitotic clones were generated and positively labeled (with membrane tethered CD8::GFP and CD2::RFP) according to the MARCM \[[@B30]\] and twin-spot MARCM \[[@B33]\] techniques. Unless indicated, recombination was induced 3 to 7 hours AEL by a one hour heat shock at 37°C and the larvae were dissected 21 hours later (early L1) or 110 hours later (L3). Genotypes of the analyzed larvae were as follows: otd^YH13^MARCM clones, w otd^YH13^FRT19A/w hs-FLP tubP-GAL80^LL1^FRT19A; tubP-GAL4UAS-mCD8::GFP^LL5^/+; wild-type MARCM clones, y w hs-FLP/+; FRT82/FRT82 tubP-GAL80^LL10^; tubP-GAL4^LL7^(or TH-GAL4) UAS-mCD8::GFP^LL6^/UAS-mCD8::GFP^LL6^; wild-type twin-spot MARCM clones, y w hs-FLP/+; FRT40A UAS-mCD8::GFP UAS-CD2-Mir/FRT40A UAS-rCD2::RFPUAS-GFP-Mir; tubP-GAL4^LL7^(or TH-GAL4)/+. Depletion of Otd by RNAi was carried out by targeted expression of an otd-specific RNAi construct (VDRC-105764) in DA neurons using the TH-gal4driver. To increase knockdown efficiency, the experiment was done at 29°C in otd^YH13^heterozygous flies. Immunohistochemistry -------------------- Antibody staining on brains was performed as previously described \[[@B39]\]. Primary antibodies were as follows: rabbit anti-Otd (1:250) \[[@B40]\], mouse anti-TH (1:100; Chemicon, Millipore AG, Temecula, California, USA), rabbit anti-TH (1:250) \[[@B41]\], rabbit anti-RFP (1:100; Abcam, Cambridge, UK). Secondary antibodies were Alexa488-, Alexa568- and Alexa647-conjugated antibodies generated in goat (1:200; Molecular Probes, Invitrogen, Paisley, Renfrewshire, UK). Fluorescent images were captured with an Olympus FV1000 confocal laser scanning microscope and analyzed in ImageJ \[[@B42]\]. Unless otherwise indicated, pictures correspond to single optical sections (1 μm thick). Figures were assembled using Adobe Illustrator and Photoshop. Abbreviations ============= AEL: after egg laying; DA, dopaminergic; DL: dorso lateral; DM: dorso medial; FRT: flippase recognition target; GFP: green fluorescent protein; L: instar larval stage; MARCM: mosaic analysis with a repressible cell marker; NB: neuroblast; oc: ocelliless; otd: orthodenticle; PPL: protocerebral posterior lateral; RFP: red fluorescent protein; RNAi: RNA interference; TH: tyrosine hydroxylase. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= JB and RP carried out all the experiments. JB, MW and GU conceptualized the project. JB and GU wrote the manuscript. All authors read and approved the final manuscript. Supplementary Material ====================== ###### Additional file 1 Supplementary Figure 1 - MARCM lineage analysis of late differentiating dopaminergic neurons in the larval central brain at L3. MARCM-labeled wild-type NB clones were induced during early embryogenesis (3 to 7 h AEL) and analyzed at L3. To circumvent the down-regulation of the tubulinpromoter, an additional copy of the UAS-CD8::GFPtransgene was included in the genotype of the analyzed larvae. (A-C)The DM1a DA neuron belongs to a NB lineage independent of the DM1b DA cell lineage. (D,E)The late differentiating DL2 DA neuron belongs to the DL2a DA cell lineage. Scale bars: 10 μm. All panels correspond to Z projections of individual confocal optical sections. AEL, after egg laying; DA, dopaminergic; DL, dorso lateral; DM, dorso medial; GFP, green fluorescent protein; MARCM, mosaic analysis with a repressible cell marker; NB, neuroblast. ###### Click here for file Acknowledgements ================ We acknowledge B Bello, C Desplan, T Lee, B Lu, U Walldorf and the Bloomington Stock Center for kindly providing fly strains and reagents. This work was supported by the Joint Singapore Bioimaging Consortium (SBIC)-Singapore Stem Cell Consortium (SSCC).	{ "pile_set_name": "PubMed Central" }
1. Introduction {#sec1} =============== Management of patients requiring mechanical ventilation is complex and poses a substantial patient safety risk for novice practitioners. Understanding the patient-ventilator interaction is essential to minimize the risk of iatrogenic barotrauma, ventilator-associated pneumonia (VAP), and death \[[@B1]\]. A lack of dedicated formal instruction on ventilator modes and patient-ventilator interaction leads to varying levels of comfort and ability to manage these patients. National work-hour restrictions, medicolegal concerns, and the increasing complexity and volume of critical care patients in the United States significantly limit the time allowed for learning key concepts and developing autonomy with mechanical ventilator management during ICU (Intensive Care Unit) rotations \[[@B2]\]. Simulation is effective as a training methodology within medical education \[[@B3], [@B4]\], including the development and refinement of critical care skills \[[@B5], [@B6]\], invasive procedures \[[@B7]--[@B11]\], and crisis resource management \[[@B12], [@B13]\]. Residents are evaluated on their clinical development by assessing their proficiency in the aforementioned skill sets. The Accreditation Council of Graduate Medical Education (ACGME) has established milestone guidelines to help residency programs standardize performance expectations, facilitate feedback for development, and identify areas needing further guidance. The ACGME mandates the use of simulation in emergency medicine, surgery, and internal medicine training \[[@B14]--[@B16]\]. Singer et al. demonstrated that first-year residents who underwent simulation-based education outperformed traditionally taught third-year residents in critical care medicine topics \[[@B17]\]. Frengley et al. reported that teamwork in critical care teams could be improved after simulated-based training intervention \[[@B18]\]. However, one area lacking sufficient study in the medical education literature is formal training in the management of mechanically ventilated patients. Even in today\'s medicolegal environment where "learning on patients" is no longer acceptable, many intensivists report a lack of standardization to their own mechanical ventilation training. The aim of this study was to develop and pilot test an interactive mechanical ventilation boot camp curriculum for first-year residents in surgery and emergency medicine. The objective of the training was to familiarize learners with common modes of ventilation, common etiologies of ventilator alarms and subsequent management strategies, and high-risk low-frequency presentations. We hypothesize that the use of a standardized curriculum would improve cognitive knowledge, performance, and participant confidence with ventilator management. 2. Materials and Methods {#sec2} ======================== 2.1. Study Location and Equipment {#sec2.1} --------------------------------- The study was performed at a tertiary-care university-affiliated teaching hospital simulation lab during July of 2015. Human-patient simulators in ICU beds were connected to ASL 5000 Breathing Simulators (IngMar Medical). Each of the five human-patient simulators were intubated and connected to a mechanical ventilator (Covidien Puritan Bennett™840). ASL 5000 Lung simulators were used to adjust the pulmonary mechanics (lung compliance, airway resistance, respiratory rate, and tidal volume) in real-time based on the actions of the resident. Each bay had a simulated patient monitor that demonstrated the patients\' vital signs, portable chest X-rays, electrocardiograms, and arterial blood gas measurements (if requested), as shown in [Figure 1](#fig1){ref-type="fig"}. This was a quality assurance project that did not meet the definition of human subject research. It was exempt from institutional review board review. 2.2. Curriculum Development and Outline {#sec2.2} --------------------------------------- A three-day pilot boot camp curriculum was developed to educate first-year residents on the management of patients requiring intubation and mechanical ventilatory support. The boot camp program required approximately 65 hours of initial preparation from the simulation and ICU faculty for the development of the curriculum, surveys, questionnaires, and simulation cases with accompanying didactic postsimulation lectures. After initial curriculum development, all participating staff performed a rehearsal of the boot camp to assure that simulation cases were executed without difficulties and all necessary equipment was available and working appropriately. This required an additional 3 hours. The boot camp was scheduled for 12 hours over three days. Overall, total preparation and execution of the curriculum took approximately 80 hours. The curriculum consisted of 4 parts (as shown inThree-Day Mechanical Ventilation Simulation Curriculum): preintervention evaluation, independent study, the intervention phase, and postintervention evaluation. Cognitive tests, critical action checklists, and confidence surveys were used to primarily assess the residents. The entire curriculum took participants approximately twelve hours to complete. The cognitive tests, critical action checklists, and confidence surveys were identical for both pre- and postintervention evaluation. Three-Day Mechanical Ventilation Simulation Curriculum Day 1 (Pretesting Evaluation): Pretest confidence survey (5 minutes). Pretest cognitive multiple-choice exam (25 minutes). Cases and evaluation by critical actions checklist (10 minutes each for 30 minutes total): ARDS. Complete lung atelectasis secondary to mucus plugging. Pneumothorax in a mechanically ventilated patient. Received supplemental readings. Independent study (estimated 4 hours of reading material provided for asynchronous education). Day 2 (Curriculum and Educational Intervention): Case structure: Two-three residents participated in the case (10 minutes). Evaluation by critical actions checklist (evaluated during case). Bedside debriefing by intensivists (20 minutes). Review of provided didactic PowerPoint (15 minutes). Pathology reviewed: ARDS. Complete lung atelectasis secondary to mucus plugging. Altered mental status secondary to overdose. Pneumothorax in a mechanically ventilated patient. Dynamic hyperinflation. Day 3 (Posttesting Evaluation): Posttest confidence survey (5 minutes). Posttest cognitive multiple-choice exam (25 minutes). Cases and evaluation by critical actions checklist (ten minutes each for 30 minutes total): ARDS. Complete lung atelectasis secondary to mucus plugging. Pneumothorax in a mechanically ventilated patient. 2.3. Participants, Faculty, and Staff {#sec2.3} ------------------------------------- First-year residents from three residency programs were invited to participate in the curriculum, which was conducted in July. Six intensivists were present for debriefing based on availability, as well as one emergency medicine attending physician and two respiratory therapists from a local teaching hospital. When present, the faculty debriefers were assigned to one station for the day and did not rotate through different cases. Each station required a simulation technician for all three days of curriculum. 2.4. Preintervention Evaluation {#sec2.4} ------------------------------- The preintervention stage assessed baseline knowledge and confidence on 3 simulated high-risk mechanical ventilation scenarios, including acute respiratory distress syndrome, complete lung atelectasis secondary to obstruction from a mucus plug (hereby referred to as mucus plugging), and pneumothorax. Participants filled out a pretest 12-question confidence survey (as shown inConfidence Survey) and a cognitive 20-question multiple-choice test. Each station accommodated one resident at a time. Residents transitioned through the stations based on availability. One intensivist was based at a station per day. The intensivists rated participants\' performance using a predetermined checklist of critical actions. The critical actions assessed for the mucus plugging case are given herein after. Feedback was not given to participants at any time in the preintervention phase. Confidence Survey Four digit identifying code: Surgery or Emergency Medicinevery uncomfortablesomewhat uncomfortableneutralsomewhat comfortablevery comfortable How comfortable do you feel distinguishing between the different modes of ventilation?How comfortable do you feel initially choosing a mode of ventilation?How comfortable do you feel switching between different modes of ventilation?How comfortable do you feel weaning mechanical ventilation?How comfortable do you feel addressing an alarming ventilator?How comfortable do you feel identifying causes of an elevated peak airway pressure?How comfortable do you feel identifying causes of an elevated plateau pressure?How comfortable do you feel managing the mechanical ventilation of a patient with ARDS?How comfortable do you feel managing the mechanical ventilation of a patient with decompensated CHF?How comfortable do you feel managing the mechanical ventilation of a patient with an acute exacerbation of COPD?How comfortable do you feel managing the mechanical ventilation of a patient with an acute exacerbation of asthma?How comfortable do you feel managing the mechanical ventilation of a patient with a traumatic brain injury? Critical Actions Checklist for Mucus Plugging Case Identifying mucus plugging□Yes□NoIdentifying increased airway resistance with high peak pressure□Yes□NoActions (suctioning, bronchoscopy, bronchodilators)Must include either suctioning or bronchoscopyYesNoTo increase FiO~2~ to maintain oxygenation□Yes□No (5)To decrease tidal volume while increasing rate to maintain minute ventilation□Yes□No(6)Reevaluate the patient after vent changes□Yes□No(7)Documenting ventilator setting changes□Yes□NoImmediately after initial testing, residents were provided with reading material. This asynchronous material consisted of approximately four hours of reading and included selections taken from book chapters, a review article, and a primer reviewing basic mechanical ventilation principles created by one of the intensivists. 2.5. Educational Intervention {#sec2.5} ----------------------------- The intervention phase was comprised of five 45-minute scenarios and took place six days after pretesting. These scenarios included a review of the basic tenets of mechanical ventilation, acute respiratory distress syndrome (ARDS), pneumothorax, mucus plugging, and dynamic hyperinflation. Residents were divided into groups of 1 to 3 participants for each station. Participants were given a brief clinical description and expected to manage the ventilated patient for the first 10 minutes. Afterwards, intensivists had approximately 35 minutes for bedside debriefing. The debriefing included individualized assessment and feedback, summation of clinical teaching points, and review of a focused didactic PowerPoint presentation, tailored to each case. 2.6. Postintervention Evaluation {#sec2.6} -------------------------------- During the postintervention stage, participants individually underwent the same three scenarios as the preintervention stage, with faculty grading their performance using the same predetermined critical action checklists. This occurred ten days after the initial preintervention evaluation. At the conclusion of the scenarios, participants completed the postintervention cognitive multiple-choice test and confidence survey, as well as a postcurriculum survey. 2.7. Postcurriculum Survey {#sec2.7} -------------------------- A postcurriculum survey was administered soliciting feedback on areas of strength and potential improvement of the curriculum using a 5-point Likert scale. 2.8. Data Analysis {#sec2.8} ------------------ Data were first imported into SPSS v22.0 software for analysis. Test results were summarized using mean (standard deviation), median, and range values for the percentage of correct answers at the pre- and postintervention study time points. The paired change in test results was determined as the postintervention test result minus the preintervention test result and then similarly summarized. The changes in test score were tested for median equality to zero using the nonparametric Wilcoxon signed rank test. Confidence was determined via 12 assessment questions measured on a Likert 1--5 ordinal scale at the pre- and postintervention study time points. Data were summarized using the aforementioned numeric measures of center of spread and the paired change data similarly tested for median equality to zero via Wilcoxon signed rank tests. Due to the potential for type I error rate inflation, p values were adjusted using Bonferroni adjustments to the p values to protect the overall type I error rate at 5%. Finally, for 3 of the 5 simulations, critical action performance was determined by expert review at the pre- and postintervention study time points. Paired dichotomous data (Yes/No) for each critical action was summarized with two-by-two bivariate frequency tables and tested for improvement in the discordant pairs with McNemar\'s tests. The paired changes, in total number of critical actions performed, were tested for median equality to zero using Wilcoxon signed rank tests for each of the 3 cases. All statistical testing was two-sided with p \< 0.05 considered statistically significant. 3. Results {#sec3} ========== 3.1. Demographics {#sec3.1} ----------------- Seventeen first-year residents participated in this study. The majority (64.7%, 11/17) were emergency medicine residents, while 23.5% (4/17) were general surgery residents, and 11.8% (2/17) were urology residents. All of the residents were present for the entirety of the boot camp. 3.2. Cognitive Knowledge Assessment {#sec3.2} ----------------------------------- Cognitive knowledge between the identical pre- and postintervention multiple-choice tests increased significantly ([Figure 2](#fig2){ref-type="fig"}). The mean (SD) preintervention score was 40.3% (9.76%) with a median (range) score of 40% (25%--55%), and the postintervention mean (SD) was 67.1% (9.53%) with a median (range) score of 70% (50%--80%). The mean (SD) change in test score from before to after intervention was +26.8% (11.31%), with a median (range) score increase of 25% (5%--45%). 3.3. Clinical Performance and Critical Actions Assessment {#sec3.3} --------------------------------------------------------- The increase in critical actions performed after intervention was significantly higher in the ARDS and mucus plugging cases but not significant in the pneumothorax case as shown in [Figure 3](#fig3){ref-type="fig"}. The ARDS case had a significant increase in 4 of the 5 critical actions. Participants had a mean of 1.5 critical actions met during preintervention evaluation and 4.1 during postintervention evaluation (p \< 0.001). The atelectasis and mucus plugging case had a significant increase in 5 out of 7 critical actions. The mean preintervention score was 1.24 and the mean postintervention score was 4.47 (p \< 0.001). The pneumothorax case had no significant increase in critical actions. The mean preintervention score was 3.06 out of 7 critical actions and the mean postintervention score was 3.5 (p = 0.123). There was no significant increase in the final critical action of all three cases. The final critical action was documentation of ventilator setting changes. 3.4. Confidence Assessment {#sec3.4} -------------------------- Participants felt more confident with ventilator management based on their pre- and postintervention confidence surveys. This was supported by a statistically significant confidence increase for all questions, with a mean pretest score of 1.56 and a posttest score of 3.64 (p = 0.049). Mean confidence gain was 2.1 with a 95% confidence interval (1.6--2.6). 4. Discussion {#sec4} ============= This study demonstrates the feasibility and effectiveness of a boot camp curriculum for residents on the basics of mechanical ventilation. This curriculum resulted in increased competency, knowledge, and confidence of seventeen first-year residents. Using an integration of bedside instruction, supplemental study, and simulation-based training, residents were exposed to several pathologic conditions of mechanically ventilated patients, as well as appropriate assessment and management techniques. A comprehensive ventilator management curriculum is long overdue and serves as a critically important patient safety initiative that may prevent iatrogenic morbidity. Residents often do not have uniform training in critical care before being expected to care for this patient population. Mismanagement of ventilated patients may result in significant morbidity and mortality \[[@B1]\]. It is no longer acceptable for residents and fellows to learn at the expense of the patient\'s health, especially in critical care environments. Medical education has subsequently evolved to include simulation as a way to educate, practice, and reinforce teaching points from cases that may otherwise only be seen rarely or in critically ill patients in the clinical setting. During this curriculum, residents were exposed to several teaching methodologies to accommodate different styles of learning. Not only did residents asynchronously read about critical care topics, but they were able to interactively experience and troubleshoot high-acuity scenarios in a high-fidelity simulated environment. Residents were engaged in a supportive learning environment under direct supervision of an intensivist, who was able to provide personalized feedback and answer any immediate questions. Afterwards, they received a focused didactic PowerPoint presentation on the particular pathology they were managing. Familiarity with the equipment was improved by having the residents physically touch and manipulate the ventilators utilized in the ICU at our institution, helping to ameliorate any anxiety working with unfamiliar equipment. We postulate that these factors contributed to the statistically significant improvement in the various categories evaluated in this study. If this curriculum motivated residents to seek out additional resources for their personal development, this could not be controlled for. Alternatively, we had no objective way of measuring if residents actually completed all provided readings. Our goal was to provide a novel and structured curriculum with dedicated time to improve the foundational knowledge in a skill set that typically did not have formal time set aside upon residents matriculation into residency. We did not intend to prove that any one method, such as simulation, was the sole contributing factor to improved performance but rather that the curriculum, as a whole, provided improved foundational knowledge. The mechanically ventilated pneumothorax case was the sole scenario where there were no statistically significant increases in critical actions when analyzing pre- and postintervention evaluations. This case was unique in that definitive treatment required a procedure and ventilator management was an ancillary measure. All participants performed needle decompression or tube thoracostomy placement in the preintervention sessions, but only five participants increased FiO~2~ to maintain oxygenation, and only one increased tidal volume or rate to maintain adequate ventilation. We postulate that participants were more focused on the definitive management of tube thoracostomy versus adjusting ventilator parameters. We also speculate that medical school curricula more commonly provide instruction on management of pneumothoraces and related cardiorespiratory physiology in comparison to the topics of ARDS, atelectasis, and dynamic hyperinflation. None of the three cases demonstrated a significant improvement in documenting ventilator changes. The importance of documentation was not covered in the supplemental reading material. Any emphasis of its importance would have been dependent on bedside instruction. Formal teaching on the necessity of proper documentation and its role within a patient safety context would likely reinforce appreciation and, therefore, compliance. To date, there are very few studies in the medical education literature addressing mechanical ventilation curriculum for residents. We postulate that the reason that there are so few studies and curricula developed for simulation-based mechanical ventilation is likely due to the significant time and cost commitments required. This curriculum was financially expensive, time-consuming, and human-resource intensive. The significant financial cost arose from renting the lung simulators (approximately \$5,000 USD). Our faculty and technicians needed eight hours of training on the lung simulators and a 6--25 hours\' time commitment from each of the six different intensivists for the boot camp\'s execution. This does not include the time dedicated to the curriculum development, where a multidisciplinary team of specialists developed curricular goals and testing materials. This included determining the selected readings and critical action lists, as well as creating the cognitive multiple-choice questionnaire and confidence surveys. Deadlines for goals and objectives were established as a team and shared throughout the curriculum development over a period of approximately 8 months. This study provides a curricular framework for other educators who desire to provide formal instruction to their residents on mechanical ventilation. Residency programs with an interest to develop and execute such a curriculum will improve their chances of success by pooling time and financial resources. Faculty members should work together to decide on goals and objectives early on in the process. Additionally, identifying how to finance renting necessary simulation equipment as well as how to divide up time for curricular and teaching responsibilities is paramount. Such cooperation will make the curriculum development less tedious and increase the likelihood of successfully executing such a curriculum. This curriculum only addressed pressure and volume control ventilation to focus on modes that residents will most commonly utilize during their training at our institution. Other modes of ventilation were not reviewed in order to prevent inundating residents with information unlikely to be used clinically. The expansion of this curriculum to include additional modes of ventilation and pathologies and the development of a longitudinal curriculum throughout residency training provide further opportunities for curriculum development, patient safety training, and research. This curriculum was well received by participants as supported by feedback from a postcurriculum survey. Participants expressed that hands-on participation and bedside teaching were highly beneficial for their learning. They commented that some additional articles should be included in the reading assignment that provide a basic review of pulmonary physiology and an overview of mechanical ventilation, to establish a stronger foundational knowledge. Additionally, residents did not feel overwhelmed by the amount of material covered, as it was predetermined to only cover the two modes of ventilation. Limitations. This was a small prospective pilot study from a community-teaching hospital. A randomized design with a control group of traditionally trained residents compared to residents completing this curriculum would have provided a more powerful study; however, residents and resources were limited. As the boot camp becomes further integrated into the residents\' curriculum, additional medical specialties may be recruited to participate, which may also serve to provide subgroup analyses with greater power. Additionally, this study was done one time and repeating it over several years would both increase the number of residents in the study and check its reproducibility. In regard to data collection, a nonvalidated confidence survey, critical action checklists, and multiple-choice tests were utilized. We also did not survey participants to determine if they completed the provided reading materials or if they used other sources as supplementation. 5. Conclusion {#sec5} ============= A twelve-hour boot camp pilot study demonstrated significantly improved confidence, knowledge, and performance of first-year residents\' ability to manage high-risk, low-frequency scenarios of mechanically ventilated patients. This study may serve as a model for other institutions that wish to implement a mechanical ventilation simulation curriculum within their training programs. Conflict of Interests ===================== There is no conflict of interests regarding the publication of this paper. ![Mechanical ventilation training and testing environment.](CCRP2016-4670672.001){#fig1} ![Pre- and postintervention knowledge assessment.](CCRP2016-4670672.002){#fig2} ![Critical actions assessment.](CCRP2016-4670672.003){#fig3} [^1]: Academic Editor: Djillali Annane	{ "pile_set_name": "PubMed Central" }
Introduction ============ Neuroblastoma is a common extracranial solid tumor that derives from neural crest progenitor cells \[[@r1],[@r2]\]. Neuroblastoma mostly takes place in children younger than 1 year, and the average diagnosis time is about 17 months of age \[[@r3]\]. Neuroblastoma is characterized by a wide range of variable prognosis, spanning from spontaneous regression without chemotherapy to life-threatening tumor progression despite intensive treatment \[[@r4]--[@r7]\]. Approximately 50% of neuroblastomas behave in highly malignant fashion, with distant metastasis at the time of diagnosis \[[@r8],[@r9]\]. Their 5-year survival rates remain less than 40% despite intensive, multi-modal therapy \[[@r10]\]. The affects of environmental factors on the risk of neuroblastoma have been investigated but remains undefined \[[@r11],[@r12]\]. Growing evidence has been directed to the genetic factors predisposing patients to neuroblastoma. Familial neuroblastoma is largely attributed to germline mutations in PHOX2B \[[@r13]\] or ALK \[[@r14],[@r15]\] gene. In contrast, the etiology of sporadic neuroblastoma, the most common type of neuroblastoma, remains partially unveiled. Several genome-wide association studies (GWASs) and the subsequent replication studies identified a number of neuroblastoma susceptibility alleles, including BARD1, LIN28B, HACE1, LMO1, MMP20 and CASC15 genes \[[@r16]--[@r23]\]. Moreover, candidate gene approaches also detected the genetic associations of NEFL \[[@r24]\] and CDKN1B \[[@r25]\] gene polymorphisms with neuroblastoma susceptibility. Ran (Ras-related nuclear protein) is a small Ras-related GTP-binding protein. Ran mainly locates in the nucleus and cycles between the GDP-bound inactive and the GTP-bound active state \[[@r26]\]. It facilitates the movement of molecules in and out of the nuclear-pore complexes \[[@r27]\]. Dysregulated protein level of Ran could cause aberrant nuclear-cytoplasmic transport of tumor suppressors and oncogenes, which might lead to the initiation of cancer \[[@r28]\]. Moreover, Ran also mediates several crucial functions, such as promoting spindle assembly, regulating cell cycle, and facilitating pre-mRNA generation \[[@r29]\]. RanBP2 (Ran-binding protein 2) is the largest protein of the nuclear pore complex (350 kDa). It contains rich FG-repeats, four Ran-binding domains and binds to Ran GTP with high affinity \[[@r30]\]. RanBP2 was initially described to be implicated in regulating nuclear transport due to its linkage with Ran \[[@r31]\]. It was further identified to regulate numerous cellular activities \[[@r32]--[@r34]\]. RAN/RANBP2 genes are reported to be associated with cancer development. However, the association of polymorphisms in the RAN/RANBP2 genes and neuroblastoma risk has yet to be elucidated. To address this issue, we conducted a three-center case-control study in a Chinese population. RESULTS ======= Characteristics of study population ----------------------------------- The detailed characteristics of subjects from Guangzhou and Zhengzhou were provided in the previous publications \[[@r35]--[@r37]\]. The detailed demographic characteristics in neuroblastoma patients and controls for Wenzhou, Guangdong and Henan subjects were presented in [Supplementary Table 1](#SD1){ref-type="supplementary-material"}. There were no significant differences between cases and controls from Wenzhou regarding age (20.25 ± 20.73 vs. 23.58 ± 15.36 months old, P=0.496) and gender (P=1.000). RAN/RANBP2 polymorphisms and neuroblastoma risk ------------------------------------------------- The genotype frequencies of RAN/RANBP2 genes polymorphisms ([Supplementary Table 2](#SD1){ref-type="supplementary-material"}) and neuroblastoma susceptibility between all cases and controls were presented in [Table 1](#t1){ref-type="table"} and [Supplementary Table 3](#SD1){ref-type="supplementary-material"}. All genotype frequencies in controls were in Hardy-Weinberg equilibrium (HWE) (rs56109543, P=0.587; rs7132224, P=0.289; rs14035, P=0.800; rs2462788, P=0.194). In single locus analysis, no statistically significant association were found regarding all the four SNPs and neuroblastoma risk. We further investigated the combined effect of protective genotypes of RAN in neuroblastoma risk. We observed that individuals with 3 protective genotypes were at significantly lower risk of developing neuroblastoma than those without protective genotypes \[adjusted odds ratio (OR)=0.33; 95% confidence interval (CI)=0.12-0.96; P=0.042\]. Moreover, subjects with 3 combined risk genotypes of RAN have a significant decreased risk of neuroblastoma (adjusted OR=0.33; 95% CI=0.11-0.94; P=0.038), compared with those with 0-2 protective genotypes. ###### Association of RAN and RANBP2 polymorphisms with neuroblastoma risk. ------------------------------------------------------- ------------- ------------- ----------- ---------------------- ----------- ---------------------- ----------- Genotype Cases\ Controls\ P ^a^ Crude OR\ P Adjusted OR\ P ^b^ (N=429) (N=884) (95% CI) (95% CI) ^b^ RAN rs56109543 (HWE=0.587) CC 304 (70.86) 620 (70.14) 1.00 1.00 CT 118 (27.51) 238 (26.92) 1.01 (0.78-1.31) 0.933 1.01 (0.78-1.31) 0.942 TT 7 (1.63) 26 (2.94) 0.55 (0.24-1.28) 0.165 0.55 (0.24-1.29) 0.168 Additive 0.363 0.93 (0.74-1.16) 0.504 0.93 (0.74-1.16) 0.502 Dominant 125 (29.14) 264 (29.86) 0.787 0.97 (0.75-1.24) 0.787 0.97 (0.75-1.24) 0.781 Recessive 422 (98.37) 858 (97.06) 0.155 0.55 (0.24-1.27) 0.161 0.55 (0.24-1.28) 0.164 RAN rs7132224 (HWE=0.289) AA 227 (52.91) 479 (54.19) 1.00 1.00 AG 170 (39.63) 335 (37.90) 1.07 (0.84-1.37) 0.581 1.07 (0.84-1.36) 0.596 GG 32 (7.46) 70 (7.92) 0.97 (0.62-1.51) 0.875 0.96 (0.62-1.51) 0.870 Additive 0.823 1.02 (0.85-1.22) 0.828 1.02 (0.85-1.22) 0.842 Dominant 202 (47.09) 405 (45.81) 0.665 1.05 (0.84-1.33) 0.665 1.05 (0.83-1.32) 0.680 Recessive 397 (92.54) 814 (92.08) 0.771 0.94 (0.61-1.45) 0.771 0.94 (0.61-1.45) 0.770 RAN rs14035 (HWE=0.800) CC 285 (66.43) 590 (66.74) 1.00 1.00 CT 135 (31.47) 263 (29.75) 1.06 (0.83-1.37) 0.635 1.06 (0.83-1.37) 0.641 TT 9 (2.10) 31 (3.51) 0.60 (0.28-1.28) 0.187 0.60 (0.28-1.29) 0.191 Additive 0.338 0.96 (0.78-1.19) 0.731 0.96 (0.78-1.19) 0.727 Dominant 144 (33.57) 294 (33.26) 0.912 1.01 (0.79-1.30) 0.911 1.01 (0.79-1.29) 0.918 Recessive 420 (97.90) 853 (96.49) 0.164 0.59 (0.28-1.25) 0.169 0.59 (0.28-1.26) 0.173 RANBP2 rs2462788 (HWE=0.194) CC 402 (93.71) 810 (91.63) 1.00 1.00 CT 27 (6.29) 74 (8.37) 0.74 (0.47-1.16) 0.187 0.74 (0.47-1.16) 0.188 TT 0 (0.00) 0 (0.00) / / / / Additive 0.185 0.74 (0.47-1.16) 0.187 0.74 (0.47-1.16) 0.188 Dominant 27 (6.29) 74 (8.37) 0.185 0.74 (0.47-1.16) 0.187 0.74 (0.47-1.16) 0.188 Combined effect of protective genotypes for RAN ^c^ 0 394 (91.84) 814 (92.08) 0.073 ^d^ 1.00 1.00 1 26 (6.06) 38 (4.30) 1.41 (0.85-2.36) 0.186 1.41 (0.84-2.36) 0.189 2 5 (1.17) 7 (0.79) 1.48 (0.47-4.68) 0.509 1.48 (0.47-4.70) 0.507 3 4 (0.93) 25 (2.83) 0.33 (0.11-0.96) 0.041 0.33 (0.12-0.96) 0.042 0-2 425 (99.07) 859 (97.17) 1.00 1.00 3 4 (0.93) 25 (2.83) 0.028 0.32 (0.11-0.94) 0.037 0.33 (0.11-0.94) 0.038 ------------------------------------------------------- ------------- ------------- ----------- ---------------------- ----------- ---------------------- ----------- OR, odds ratio; CI, confidence interval; HWE, Hardy-Weinberg equilibrium. ^a^ χ^2^ test for genotype distributions between neuroblastoma patients and controls. ^b^ Adjusted for age and gender. ^c^ Protective genotypes were rs56109543 TT, rs7132224 GG and rs14035 TT. ^d^ For additive model. Stratification analysis ----------------------- Stratification analysis was further adopted to assess the effects of the RAN polymorphisms on neuroblastoma risk among different strata ([Table 2](#t2){ref-type="table"}). However, we failed to detect significant association for any of the four polymorphisms in single locus analysis. Moreover, the cumulative effects of protective genotypes were also insignificant. ###### Stratification analysis for the association between RAN gene genotypes and neuroblastoma susceptibility. ----------------- ---------------- ------------------ ------------------ ---------------- ------------------ --------- ----------------------- ------------------ --------- ------ ------------------ ------- Variables rs56109543\ AOR (95% CI) ^a^ P ^a^ rs14035\ AOR (95% CI) ^a^ P ^a^ Protective genotypes\ AOR (95% CI) ^a^ P ^a^ (case/control) (case/control) (case/control) CC/CT TT CC/CT TT 0-2 3 Age, month ≤18 145/327 1/13 0.17 (0.02-1.32) 0.091 144/326 2/14 0.33 (0.07-1.45) 0.140 145/328 1/12 0.19 (0.02-1.45) 0.109 \>18 277/531 6/13 0.89 (0.33-2.35) 0.806 276/527 7/17 0.79 (0.32-1.92) 0.597 280/531 3/13 0.44 (0.12-1.55) 0.200 Gender Female 181/365 4/11 0.70 (0.22-2.24) 0.550 183/366 2/10 0.38 (0.08-1.77) 0.219 183/366 2/10 0.38 (0.08-1.77) 0.219 Male 241/493 3/15 0.42 (0.12-1.46) 0.170 237/487 7/21 0.68 (0.29-1.63) 0.392 242/493 2/15 0.28 (0.06-1.22) 0.090 Sites of origin Adrenal gland 163/858 1/26 0.22 (0.03-1.61) 0.134 160/853 4/31 0.71 (0.25-2.03) 0.518 164/859 0/25 / / Retroperitoneal 94/858 2/26 0.68 (0.16-2.92) 0.604 93/853 3/31 0.85 (0.25-2.83) 0.785 94/859 2/25 0.71 (0.16-3.03) 0.638 Mediastinum 119/858 4/26 1.08 (0.37-3.16) 0.887 121/853 2/31 0.46 (0.11-1.96) 0.295 121/859 2/25 0.56 (0.13-2.39) 0.432 Others 38/858 0/26 / / 38/853 0/31 / / 38/859 0/25 / / Clinical stage I+II+4s 175/858 4/26 0.73 (0.25-2.13) 0.567 176/853 3/31 0.47 (0.14-1.56) 0.217 177/859 2/25 0.38 (0.09-1.62) 0.190 III+IV 224/858 3/26 0.47 (0.14-1.59) 0.226 221/853 6/31 0.76 (0.31-1.86) 0.550 225/859 2/25 0.33 (0.08-1.39) 0.129 ----------------- ---------------- ------------------ ------------------ ---------------- ------------------ --------- ----------------------- ------------------ --------- ------ ------------------ ------- AOR, adjusted odds ratio; CI, confidence interval. ^a^ Adjusted for age and gender, omitting the corresponding stratification factor. DISCUSSION ========== In the current study, we performed the first investigation into the impact of SNPs in RAN/RANBP2 genes on the risk of neuroblastoma in Chinese Han children. Our data revealed that the single RAN or RANBP2 gene polymorphism might not be strong enough to confer the neuroblastoma susceptibility in Chinese children. However, three protective RAN genotypes were observed to cumulatively reduce the risk of neuroblastoma. Overexpression of Ran has been observed in several human malignancies, including lung, prostate, breast, colon cancer, and neuroblastoma \[[@r38],[@r39]\]. Conditional knockdown of RAN gene reduced the viability of activated K-Ras-transformed cells, through inducing S-phase arrest \[[@r40]\]. Barrès et al. found that Ran protein is highly expressed in invasive serous epithelial ovarian cancers and overexpression of Ran is associated with poor patient outcome \[[@r41]\]. They also detected that silencing Ran could impair tumor growth in vitro and in vivo \[[@r42]\]. Xia et al. showed that RNA interference-mediated knockdown of RAN induces aberrant mitotic formation and apoptosis in cancer cells \[[@r38]\]. Silencing RAN causes abnormal nucleocytoplasmic transportation of transcription factors in tumor cells \[[@r43]\]. RanBP2 protein also plays critical roles in cellular processes. Knockdown of RANBP2 results in an aberrant metaphase, mitotic arrest in G~2~/M phase and mitotic cell death \[[@r44]\]. A study by Dawlaty et al. demonstrated that RanBP2 acts as a novel tumor suppressor in lung cancer through regulating TopoII by sumoylation \[[@r45]\]. In addition, RanBP2 hypomorphic mice are more susceptible to spontaneous and carcinogen-induced lung tumors. Consistently, two independent studies also demonstrated that RanBP2 level was downregulated in human lung cancers \[[@r46],[@r47]\]. Herein, for the first time we investigated whether RAN/RANBP2 SNPs could contribute to the risk of neuroblastoma in Chinese children. However, our findings found no significant relationship between all the analyzed RAN/RANBP2 polymorphisms and neuroblastoma risk. Such null relationship might be attributed to the relatively small sample size, although we tried to expand the sample by recruiting subjects from three centers. To be highlighted, a study conducted by Luo et al. explored the association between sumoylation-related genes polymorphisms and risk of gastric cancer \[[@r48]\]. They are the first group investigating the role of RANBP2 gene polymorphism in cancer risk. Their study included 1021 gastric cancer cases and 1304 controls from Chinese population. However, they failed to obtain a significant association between RANBP2 gene intron variant rs12614691 and gastric cancer risk. In the combined analysis of our study, subjects carrying 3 protective genotypes tend to have decreased neuroblastoma risk in comparison to those without risk genotypes or those with 0-2 protective genotypes. This phenomenon was quite biologically plausible as each single variant in each gene might not be strong enough to influence the risk of cancer. The current study was the first investigation on the association of RAN/RANBP2 genes SNPs with neuroblastoma risk. Another merit of this study was that this is a three-center case-control study. Several limitations exist in the current study. First, because of the low incidence rate of neuroblastoma, the recruitment of eligible patients was a great challenge for us. Even though we enrolled participants from three hospitals, the sample size is still moderate. This limited sample size inevitably impaired the strength of the statistical power. Second, this study only incorporated four SNPs in the RAN/RANBP2 genes. Future studies should investigate more potentially functional polymorphisms in RAN/RANBP2 genes. Third, as all the participants included were of Chinese origin, conclusions should be taken with caution when extrapolated to other populations. Fourth, functional analysis is warranted to justify the described associations, which would illustrate the underlying mechanisms of how theses SNPs modify neuroblastoma susceptibility. Additionally, we only assessed the possible association of the SNPs with neuroblastoma risk. Other environmental factors, such as dietary habit, childhood exposure, and health situation, would help to provide further insight into the influence of RAN/RANBP2 polymorphisms on neuroblastoma risk. In all, here we demonstrate that common variants at the RAN/RANBP2 genes are associated with the risk of neuroblastoma in the Chinese children in a low-impact manner. Future larger-sample, functional studies are warranted to address the mechanism by which RAN/RANBP2 SNPs impacts tumorigenesis of neuroblastoma. MATERIALS AND METHODS ===================== Study populations ----------------- This case-control study was conducted in three centers: Guangzhou Women and Children's Medical Center, The First Affiliated Hospital of Zhengzhou University and The Second Affiliated Hospital and Yuying Children\'s Hospital of Wenzhou Medical University. The study was approved by the Institutional Review Board of the above three hospitals. In total, 429 neuroblastoma cases and 884 controls from three centers were included in this study. To be specific, 275 cases and 531 controls were enrolled from Guangzhou \[[@r35]--[@r37]\], 118 cases and 281 controls were recruited from Zhengzhou \[[@r49],[@r50]\], and 36 cases and 72 controls were enrolled from Wenzhou ([Supplementary Table 1](#SD1){ref-type="supplementary-material"}). The recruitment period lasts from December 2007 to June 2017. All the participants' parents provided signed informed consent before the study. Selection criteria of the included participants were accessible in our previous publication \[[@r51]\]. SNP selection and genotyping ---------------------------- We chose potentially functional polymorphisms in the RAN/RANBP2 genes from dbSNP database (<http://www.ncbi.nlm.nih.gov/>). An online tool, SNPinfo (<http://snpinfo.niehs.nih.gov/>) was used to predict the functions of SNPs. In brief, we searched the potentially functional candidate SNPs located in the 5'- flanking region, 5' untranslated region, 3' untranslated region, and exon of RAN/RANBP2 genes. Three potentially functional SNPs in RAN gene (rs56109543 C\>T, rs7132224 A\>G, rs14035 C\>T) and one SNP in RANBP2 (rs2462788 C\>T) were chosen for analysis that captured nine additional SNPs with LD\>0.8 ([Supplementary Table 2](#SD1){ref-type="supplementary-material"}). Three SNPs (rs56109543, rs7132224, rs2462788) are located in transcription factor binding sites (TFBS) and one SNP rs14035 might affect the microRNA binding site activity. As shown in [Supplementary Figure 1](#SD1){ref-type="supplementary-material"}, there was no significant LD (R^2^\<0.8) between each RAN SNP pair (R^2^=0.488 between rs56109543 and rs7132224, R^2^=0.582 between rs14035 and rs7132224), except for the rs56109543 and rs14035 (R^2^=0.838). The peripheral blood was used to extract genomic DNA. We genotyped the gene polymorphisms using Taqman real-time PCR \[[@r52]--[@r54]\]. On each 384-well plate, eight negative controls with water were used as quality control samples. The randomized and blinded process method was adopted to genotype all case and control samples. 10% random selection samples were re-genotyped and the genotype concordance rate was 100%. Statistical analysis -------------------- Departures from HWE for the selected SNPs in controls was evaluated using goodness-of-fit χ^2^ test. Allele frequencies and demographic variables between the two groups were assessed by chi-square test. The ORs, 95% CIs, and the corresponding P value for each SNP were calculated with adjustment for age and gender. Risk associations between genotypes and neuroblastoma were determined from logistic regression analysis. All calculations were performed using SAS software version 9.4 (SAS Institute, Cary, NC). All statistical tests were two-sided, and significant threshold was set using P\< 0.05. Supplementary Material ====================== CONFLICTS OF INTEREST: The authors have no competing interests to declare. FUNDING: This work was supported by grants from the Pearl River S&T Nova Program of Guangzhou (No: 201710010086), Scientific Research Foundation of Wenzhou (No: 2015Y0492), Zhejiang Provincial Medical and Health Science and Technology plan (No: 2009A148), and Zhejiang Provincial Science and Technology Animal Experimental Platform Project (No: 016C37113).	{ "pile_set_name": "PubMed Central" }
Introduction {#sec1_1} ============ Signet-ring cell carcinoma (SRCC) is a rare adenocarcinoma of epithelial origin, accounting for less than 0.5% of all malignancies \[[@B1]\]. It predominantly affects the stomach, with approximately a fourth of gastric cancers attributed to SRCC. However, SRCC has been known to affect other organs, such as the prostate, ovaries, pancreas, and large intestine to a much lesser extent \[[@B2], [@B3]\]. In addition, only a small number of cases have been reported in patients with diagnosed Crohn\'s disease \[[@B4]\]. It is generally known that inflammatory bowel disease (IBD) increases a patient\'s risk for developing cancer along the gastrointestinal tract (GI), though the exact mechanism is poorly understood \[[@B5]\]. Theories include chronic inflammation causing dysplasia, use of immunosuppressants and biologics, and manipulation of the gut flora \[[@B6]\]. Herein, we present a case of SRCC involving the terminal ileum with concurrent Crohn\'s flare. Case Presentation {#sec1_2} ================= A 44-year-old female with a history of Crohn\'s disease was admitted for progressive fatigue, chronic abdominal pain, and bright red blood per rectum. She was prescribed rivaroxaban approximately 2 weeks prior to this encounter for left calf tenderness determined to be secondary to a deep vein thrombosis. Physical examination on admission was notable for conjunctival pallor and right lower-quadrant tenderness on palpation of the abdomen with no distension; bowel sounds were present and normoactive. There was no appreciable fluctuation in weight in the past 12 months. Initial laboratory findings included a microcytic anemia (hemoglobin 6.6 g/dL), left-shift leukocytosis (14.3 count/μL), elevated sedimentation rate (ESR 45 mm/h), and elevated C-reactive protein (CRP 6.7 mg/L). Rivaroxaban was stopped, and the patient was transfused with packed red blood cells with a goal of hemoglobin \> 7.0. The patient was continued on 20 mg prednisone daily for Crohn\'s disease management with plans to start ustekinumab. Review of medical records showed an MRI enterography performed a month before, which suggested active Crohn\'s disease of the terminal ileum with increased dilatation of portions of the ileum proximal to the area of active disease; no fistula or abscesses were identified. In addition, a colonoscopy performed 1 year prior to this case presentation showed a stenotic but otherwise normally appearing ileocecal valve, preventing passage of the colonoscope and full visualization of the terminal ileum. Given her presentation and previous imaging findings, her symptoms were initially attributed to a lower GI bleed in the setting of rivaroxaban use. The patient had persistent anemia despite red blood cell transfusions prompting repeat colonoscopy. On visualization of the proximal cecum and terminal ileum, an ulcerating, partially obstructing, mass-like process was identified (Fig. [1](#F1){ref-type="fig"}). The mass was biopsied, which revealed SRCC with mucinous features. Additional laboratory tests included a carcinoembryonic antigen, which was within normal limits (CEA 1.8). Computed tomography of the chest, abdomen, and pelvis revealed no evidence of metastatic disease; radiographic measurement of the mass revealed approximately 8 cm in length at the distal ileum. The patient subsequently underwent laparoscopic-assisted right hemicolectomy with ileostomy, with the excised sample sent for gross pathology (Fig. [2a](#F2){ref-type="fig"}). Final pathological diagnosis was poorly-differentiated signet-ring cell adenocarcinoma (T4aN1) of the terminal ileum, ileocecal valve, and cecum (Fig. [2b](#F2){ref-type="fig"}). Molecular diagnostic was negative for RAS and BRAF genes. The patient was seen by medical oncology as an outpatient and was started on folinic acid, fluorouracil, and oxaliplatin (FOLFOX) adjuvant therapy. Discussion {#sec1_3} ========== Cancer of the small intestine is a rare diagnosis. However, in patients with Crohn\'s disease, the relative risk of having a malignancy of the small intestine increases up to 30 times compared to the general population. Undoubtedly, abnormal imaging and symptom presentations should raise the specter of malignancy in individuals with IBD if the clinical context is ripe. In our patient\'s case, it was difficult to distinguish between malignancy and inflammation. MRI enterography has been shown to be useful in the diagnosis and management of IBD, but its role in deciphering between inflammatory flares in IBD and new malignancies continue to evolve \[[@B7], [@B8]\]. Although it is consensual knowledge that IBD increases the risk of GI malignancies, a proper modality or algorithm of screening or follow-up is not present in standard of care guidelines. In addition, our patient had a previous colonoscopy, which showed abnormalities of the ileocecal valve most consistent with stricturing Crohn\'s disease. As mentioned above, inflammatory changes can be observed, but in our case, there was a quick development of malignancy between colonoscopy procedures. Such a rapidly developing entity is challenging to anticipate even with frequent procedures \[[@B9]\]. Given the various aggressiveness of GI malignancies that may develop in the setting of IBD, further investigation into improving the detection of malignancies early, either through more specific biomarker surveillance or increased sites of biopsy, is warranted. Additionally, our patient was treated with the appropriate medical therapy for her Crohn\'s disease. Our team questioned whether the number of inflammatory flares predisposed her for her diagnosed malignancy \[[@B10]\]. Certainly, SRCC of the small intestine carries a rare diagnosis. In all, the current literature present neither patient studies of long-term medical therapies nor the impact of other comorbidities to prevent the development of malignancies. Though there are studies of chemoprevention of malignancy in Crohn\'s disease, long-term data are limited, and further investigation is needed \[[@B11]\]. Statement of Ethics {#sec1_4} =================== The patient was contacted after her hospital stay. She was told of the team\'s desire to publish her medical case in an academic journal. She was told that all information would be anonymized. She agreed to have her medical case be published. All authors listed were involved in the patient\'s care. The authors have no ethical conflicts to disclose. Disclosure Statement {#sec1_5} ==================== The authors have no conflicts of interest to declare. ![The gross specimen displays an endophytic, ulcerating mass involving the terminal ileum, ileocecal valve, and cecum.](crg-0013-0085-g01){#F1} ![a Gross pathology of the resected bowel. The asterisk shows the endophytic mass at the distal ileum. b Tumor cells demonstrate signet-ring cell morphology with eccentrically displaced nuclei and abundant vacuolated cytoplasm. HE. ×200.](crg-0013-0085-g02){#F2}	{ "pile_set_name": "PubMed Central" }
Introduction {#s1} ============ Humans are very efficient at processing quantities. Assessing and evaluating the number of desirable or potentially dangerous entities---chunks of food, precious artifacts, or deadly animals---happens in fractions of seconds. Importantly, we are also able to share these rapid impressions with others: Concrete quantities or abstract symbolic representations can be translated into number words (cf. Dehaene et al., [@B4]) or, similarly, other number-describing expressions, so-called quantifiers (e.g., "many," "at least five," etc.). Quantifiers can take the form of different semantic classes. On the one hand, there are cardinal quantifiers ("three bottles") which contain an explicit numeric expression (in our example: three). There are also other classes of quantifiers that refer to quantity with an implicit notion of quantity that do not refer to an exact number or numeric interval. For instance, Aristotelian quantifiers (comprising existential quantifiers such as "all X are Y" and logical quantifiers like e.g., "some X are Y") or majority quantifiers ("at least half of the X are Y") refer to quantities in the environment in the absence of an explicitly defined number. Items in the environment may be defined by the total size of a set: If there are 12 oranges, for example, "at least half of the oranges" means more than 6. However, in other cases, the criterion may be defined more vaguely. For instance, "much chocolate" may mean one bar or five bars, depending on the individual\'s preference. Likewise, it may depend on the semantic context: "many lions" could be five, but "many ants" probably means hundreds or thousands. We focus our study on quantifiers because we can easily and precisely measure the meaning of this class of word, even one that does not explicitly mention a number. This is because the meaning of a quantifier is derived in part from number knowledge---to varying degrees depending on the exact nature of the quantifier. Majority quantifiers refer to a subset of items that require a computation based on an understanding of the entire set of items and thus tend to be more difficult than cardinal quantifiers where an explicit number is mentioned. For instance, children use number expressions much earlier, and far more correctly, than quantifiers (Hurewitz et al., [@B11]; Sullivan and Barner, [@B23]). In particular, the appropriate use of quantifiers is not always easy. Consider the sentence "All shirts are blue." If this sentence is true, then the sentence "Some shirts are blue" is, logically, also true (cf. Geurts, [@B5]). This phenomenon is called scalar implicature, indicating that if the truth value for a set of elements is 1, then it is also 1 for any given subset. However, as Grice ([@B6]) pointed out, this is not how we use quantifier expressions in every-day conversation. Rather, we use "conversational implicatures," trying to be as relevant as possible (for a discussion see Haugh, [@B7]). In the given example, this would mean that we distinguish between "some" and "all" such that "some" means "only some, not all." This Gricean use of quantifiers can be demonstrated empirically in the laboratory: Different quantifiers are used to refer to different ranges of numerosities (e.g., Oaksford et al., [@B17]). For a given set of objects, "none" refers to the lower end of the scale, followed by "few," "some," "some not," "most," and "all" in ascending order. This order was reliably observed with different tasks, indicating that a given quantifier is most appropriate, and thus most informative, for a certain fraction of the overall numerosity. However, the distinction between scopes of quantifiers is not completely clear-cut. Rather, scopes of quantifiers tend to overlap partially (e.g., Holyoak and Glass, [@B10]; Oaksford et al., [@B17]). This may lead to situations in which two quantifiers are (nearly) equally appropriate for the description of a given amount of entities. For instance, in the Oaksford et al. ([@B17]) study, participants were presented with images containing varying amounts of black or white squares, together with a statement like "Most of the squares are black." Their task was to decide whether the statement was an appropriate description of the visual scenario. When there were about 18% of black squares on the screen, subjects accepted the statement "Few of the squares are black" in about 75% of the cases. Interestingly, when presented with the same scenario but the sentence "Some of the squares are black," they would accept the statement in exactly the same proportion of cases. Compatible results were found in a memory experiment by Holyoak and Glass ([@B10]) who demonstrated that quantifiers like "many" and "a few" might be confused in memory in some proportion of cases, the chance being higher as the quantities and the referring quantifiers became more similar. To summarize, the evidence suggests that the selection of quantifiers for every-day use seems to follow the Gricean principle of relevance, maximizing the information expressed by this quantifier. Such specific selection is possible because each quantifier has a particular scope, i.e., it refers to a certain interval of probabilities distributed around one criterion value which best represents the semantics of the quantifier. These scopes overlap at least in part. Given this overlap of scopes and the fact that quantifiers may be exchanged under some circumstances, a question arises regarding the invariance of the criteria to which quantifiers refer in an individual. Given the close relationship between processing of numbers and numerosities on the one hand and quantifiers on the other hand, findings from studies of number processing may shed some light on this question. Both behavioral and neurophysiological studies (e.g., Nieder and Miller, [@B16]; Piazza and Izard, [@B18]) indicate that processing of numerosities is not an all-or-nothing phenomenon. Rather, even if a presented number of elements does not exactly match a target numerosity, neural assemblies may also fire to these if they are close enough to the target numerosity, resulting in the subject to show overt behavioral responses to this slightly deviant numerosity. For instance, Piazza and Izard ([@B18]) summarize a series of studies in monkeys and human subjects who performed same-different judgment tasks on stimuli with dot patterns. Intracortical recordings in parietal neurons in monkeys showed the firing pattern described above, which nicely corresponded to the distribution of the percentage of "same" responses in the monkeys. In human subjects, fMRI was used instead of intracortical recordings, yielding again a nice match between button press responses and the shape of the BOLD response in the left parietal cortex around the intraparietal sulcus. Most interestingly, these behavioral responses are not invariant. In his seminal work, Helson ([@B9]) demonstrated that the repeated exposure to a certain stimulus intensity, or numerosity, may result in a change of what he called "adaptation level." Responses to the former target intensities, or numerosities, are thus altered, leading to a so-called peak shift in the response curves. The maximum probability of a YES response is moved toward a new criterion. Following the argument above that the processing of quantifiers and numbers bears some similarities, one might assume that such a peak shift phenomenon can occur not only for number processing, but also for quantifier processing. In other words, some external learning influence is likely to change the internal criterion to which a quantifier refers, resulting in a shift of the scope of this quantifier. It is yet an open question whether such a phenomenon may occur. If so, another question follows: Does the scope of this one quantifier shift in isolation, or does it also affect the scope of other quantifiers? In particular, a change in the criterion for "many" affect the mental representation of the related quantifier "few?" If true, this would provide some support for the claim that quantifiers like "many" and "few" are linked in their representations in semantic memory even though these terms refer to different quantities (cf. Routh, [@B20]). The present study was thus conducted in order to address these two questions. For this purpose, we used the semantic truth value judgment (STVJ) task by Heim et al. ([@B8]) and combined it with explicit feedback trials as e.g., used by McMillan et al. ([@B13]). As a result, we were able to assess the initial criteria to which the quantifiers "many" and "few" refer, to train subjects to change their internal criteria for "many-ness" (Experiment 1) or "few-ness" (Experiment 2) independently from each other, and to test whether explicitly altering the one criterion (e.g., for "many") would also affect criterion for the other, untrained quantifier (in this case, for "few"). We used the quantifiers "many" and "few" because they are familiar, they do not differ in processing requirements (see above), and at the same time they are sufficiently distant from each other (Routh, [@B20]) to be not mistaken too easily. Experiment 1 ============ The purpose of Experiment 1 was to assess the subjects\' initial criteria for calling a certain amount of circles of a given color "many" or "few" (Block 1), then to shift the criterion for "many" explicitly to a lower proportion of circles (Block 2), and finally to test whether this shift also affected the criterion for "few" even though the criterion for this quantifier was not mentioned (Block 3). All subjects participated in an informed consent procedure that was approved by an Institutional Review Board convened at the University of Pennsylvania. Methods ------- ### Participants Twenty-one healthy volunteers (mean age 22.7 years, range 19--29 years; 13 women; average amount of education: 15.8 years) participated in this study. They were recruited at the University of Pennsylvania and received \$10 for their participation. ### Procedure Subjects were presented with pictures showing isoluminant blue and yellow circles on a gray background. The stimuli were taken from the study by Heim et al. ([@B8]). Each stimulus item included a picture containing 50 circles, with proportions of yellow circles ranging from 20/30/40/50/60/70% of the total number of circles. The remaining circles were blue. There were six different stimuli for each proportion. In order to minimize learning effects due to repetition of stimuli, each stimulus was presented in three different rotations (0°, 90°, 180°), a procedure that had previously been shown to reduce repetition memory effects for stimuli (Lassaline and Logan, [@B12]). Each picture was presented together with the sentence "Many of the circles are yellow" or "Few of the circles are yellow." The subjects performed a truth value judgment task, indicating whether they thought the sentence adequately described the picture or not by pressing the left or right response button[^1^](#fn0001){ref-type="fn"}. The experiment consisted of three blocks (cf. Figure [1](#F1){ref-type="fig"}). The first block served as a baseline block, in which the subjects\' internal criteria for "many" and "few" were assessed. To this end, subjects performed the STVJ task in a total of 378 trials counterbalanced for "few" and "many" trials along with the proportion of yellow circles. After this baseline block, an adaptation block followed in which subjects received feedback for their response. In this block, we provided feedback and only stimuli with the quantifier "many" were presented. Positive feedback was given if they responded YES to pictures containing 40% or more of yellow circles, or NO to pictures with less than 40% of yellow circles. Negative feedback was given if they responded NO to pictures with 40% or more, or YES to less than 40% yellow circles. Based on previous data (Heim et al., [@B8]), this feedback would effectively move the internal criterion from a presumed baseline of 50-60% down to 40%. For this adaptation block for "many," proportions of 20--70% yellow circles were used. ![Schematic of the trials in blocks 1 and 3 (top) and in the adaptation block 2 (bottom).](fpsyg-06-00441-g0001){#F1} The adaptation block consisted of 162 trials, which were organized to form three sub-blocks with equal amounts of trials of each proportion. Sub-dividing the adaptation block into these three homogeneous sub-blocks allowed us to assess the learning curve over the adaptation block with more precision by allowing us to analyze each sub-block independently. Proportions were distributed equally over the first, second, and third part of the adaptation phase, thus creating three parallel sub-blocks of 54 trials each (9 trials for each of the 6 proportions). After the adaptation block, the test block was administered. This was identical to the baseline block, i.e., sentences containing "many" or "few" were presented, and no feedback was given. The test block was run to assess whether and how the internal criterion for "many" had been changed during the explicit training block for "many," and whether the criterion for "few" had been altered in parallel even though "few" had never been presented in the adaptation block. Prior to the real experiment, all subjects were familiarized with the stimuli and task in two short blocks, one without feedback that resembled the baseline or test phase, and one with feedback as in the adaptation phase. The timing of a baseline/test trial is illustrated in Figure [1A](#F1){ref-type="fig"}. First, a written sentence ("Many of the circles are yellow" or "Few of the circles are yellow") was shown in the upper third of the screen for 3 s. The sentence stayed on for another 1.5 s while the stimulus picture containing yellow and blue circles appeared at the center of the screen for the same amount of time. The words "YES" and "NO" were written in the lower third of the screen to indicate which button to push for which decision. Finally, the screen went blank for another 1.5 s before the next trial began. Responses were recorded from the onset of the picture and through the entire blank period. The trials in the adaptation phase are illustrated in Figure [1B](#F1){ref-type="fig"} and were similar to those in the baseline/test phase, with the exception that feedback was given in these trials. If the subjects responded correctly, they received 10 points which were added to their score. False responses were penalized by deducting 10 points from their score. The feedback screen consisted either of a green arrow pointing upward and the information "+10" in green font, or a downward red arrow and the text "−10" in red font. The initial score was "+100." The actual score was always presented in the middle of the very upper part of the screen in black font. The adaptation phase consisted of 162 trials. Proportions were distributed equally over the first, second, and third portions of the adaptation phase, thus creating three parallel sub-blocks of 54 trials each (9 trials for each of the 6 proportions). These sub-blocks could then be compared in order to test how quickly the subjects learned. ### Data analysis Because of the balanced assignment of YES/NO responses to the left/right response button, the subjects\' button press responses were first recoded into "1" (YES) or "0" (NO) to make data sets comparable. Next, these acceptability ratings were aggregated per subject, experimental block, quantifier, and proportion of circles of the target color. Consequently, a 2 × 2 × 6 ANOVA with factors BLOCK (baseline/test), QUANTIFIER (many/few), and PROPORTION (20/30/40/50/60/70) was run to test for differential learning effects for the trained vs. untrained quantifier at trained vs. untrained proportions. Subsequently, planned contrasts (paired t-tests) were computed to compare the change in acceptability ratings at the critical proportion "40%" for the trained and the untrained quantifier. In a second analysis, we also tested how quickly the subjects learned during the adaptation phase. To this end, data of the adaptation block were collapsed over subject, sub-block, quantifier, and proportion. Paired t-tests (sub-blocks 1 vs. 2, 2 vs. 3, and 1 vs. 3) were run to compare the acceptability ratings at the critical proportion "40%." We used one-tailed t-tests because of our a priori predictions. Results ------- ### Adaptation effects The 2 × 2 × 6 ANOVA yielded significant main effects for BLOCK \[F~(1,\ 20)~ = 28.69; p \< 0.001\] and PROPORTION \[F~(5,\ 16)~ = 5.50; p = 0.004\] but not for QUANTIFIER \[F~(1,\ 20)~ \< 1\]. Moreover, all two- and three-way interactions were significant \[QUANTIFIER × BLOCK: F~(1,\ 20)~ = 34.39; p \< 0.001; QUANTIFIER × PROPORTION: F~(5,\ 16)~ = 1362.62; p \< 0.001; BLOCK × PROPORTION: F~(5,\ 16)~ = 4.24; p = 0.012; BLOCK × QUANTIFIER × PROPORTION: F~(5,\ 16)~ = 9.73; p \< 0.001\]. These effects demonstrate a strong adaptation effect for the trained quantifier "many," and most importantly, we also observed a transfer effect for the not trained quantifier "few" (Figure [2A](#F2){ref-type="fig"}). ![(A). Average acceptability ratings for a given proportion of circles of the mentioned color, plotted separately for the quantifiers "many" (black lines) and "few" (gray lines) in the baseline blocks (dashed lines) and the test blocks after adaptation (solid lines). (B) Average acceptability ratings for a critical proportion of circles of the mentioned color, plotted separately for "many" (black bars) and "few" (gray bars) in the baseline blocks (dashed bars) and the test blocks after adaptation (solid bars). ^\^p* \< 0.05, ^\\\^p* \< 0.001.](fpsyg-06-00441-g0002){#F2} Testing the adaptation effect for "many" and "few" at the critical proportion "40%" illustrates this overall effect (Figure [2B](#F2){ref-type="fig"}). We found a highly significant increase in acceptability for "many" \[t~(20)~ = −7.79; p \< 0.001\], and a decrease for "few" \[t~(20)~ = 2.58; p = 0.018\] (one-tailed t-tests, p-Bonferroni-corrected). ### Learning during the adaptation phase Testing the learning effect for "many" in block 2 yielded on-going adaptation (Figure [3](#F3){ref-type="fig"}). This effect was significant from sub-block 1 to sub-block 3 \[t~(20)~ = −7.75; p \< 0.001\] as a result of increases from sub-block 1 to sub-block 2 \[t~(20)~ = −5.03; p \< 0.001\] and from sub-block 2 to sub-block 3 \[t~(20)~ = −1.83; p = 0.042; all tests one-tailed, uncorrected for multiple comparisons\]. ![*Learning curves for the new meaning of "many"* during the adaptation phase (block 2) as a function of percent circles of the target color, plotted separately for each of the three sub-blocks*. The learning effect at the trained proportion "40" is visible.](fpsyg-06-00441-g0003){#F3} Discussion ---------- In this study we investigated whether individuals can flexibly adapt their interpretation of quantifier meaning. Moreover, we tested whether such change could affect the criterion of an untrained quantifier as well, thus indicating a shift in the entire semantic reference frame for numerosities. For both research questions we obtained positive evidence. Subjects were able to shift their internal criterion for "many"* toward the new reference criterion of "40%" following explicit training. A closer examination of the dynamics of this change (Figure [3](#F3){ref-type="fig"}) revealed that learning actually happened quite rapidly, with substantial changes in the acceptability ratings from the first third to the second third of the adaptation phase. Most interestingly, the comparison between baseline and test phase, which were absolutely identical with respect to quantifiers, stimuli, and their pairings, revealed that the internal criterion for the second quantifier ("few") had also shifted although it had not been mentioned in the adaptation phase. These findings have interesting implications. They suggest that contextual influence such as explicit reinforcement can be a driving force to shift the quantifier\'s scope (cf. Price et al., [@B19] for an extensive discussion of feedback mechanisms during the estimation of quantities). Moreover, there is a change in the reference of the related quantifier "few," even though this was not explicitly mentioned during the adaptation phase. This suggests that subjects\' learning was not limited to the superficial mapping of a word to a stimulus, but that the concept underlying this continuum of quantifier word meaning was changed. The data are consistent with the interpretation that the contextual appropriateness of a quantifier (such as the untrained "few" in the present study) depends on its relative position in the continuum of reference criteria: If a proportion of 40% circles of a given color is best described as "many," the initial tendency of a subject to call 40% "few" cannot be maintained. If it were maintained, the Gricean principle of maximized information would be violated because two different quantifiers that are partly exclusive and that are on different positions on the continuum of reference criteria (Oaksford et al., [@B17]) would describe the same semantic reality. Yet, the interpretation of this data set is limited in two ways. First, a direct shift for the trained quantifier and a transfer for the untrained quantifier were only observed for a decrease of the criterion (from 50 to 40%), i.e., leftwards on the continuum of reference criteria. Second, the trained quantifier "many" was positive. From earlier studies (e.g., Heim et al., [@B8]) it is known that positive quantifiers such as "many," "most," or "more than half" are easier to process than their negative counterparts, e.g., "few" or "less than half." In line with cognitive load theories (e.g., Bannert, [@B1]) one might assume that the learning transfer observed in Experiment 1 would diminish if the cognitive load associated with the quantifier would be higher, as it would be in the case of the negative quantifier "few." This is because there may be reduced potential mental flexibility associated with a word requiring increased cognitive load during processing. Therefore, we conducted Experiment 2 in which the setting was changed accordingly: Here we explicitly trained "few," and then examined whether a transfer effect would be observed for "many." Second, since the acceptability ratings for "few" in the baseline block of Experiment 1 indicate ceiling effects at low proportions, the criterion shift was now rightwards, i.e., toward the higher reference criteria 50 and 60%. Experiment 2 ============ Thus, the questions in Experiment 2 were the following. (1) Can the criterion of a negative quantifier like "few" (the scope of which is on the mental number line below the criterion, e.g., "less than half" or "fewer than five," not above as in the case of positive quantifiers such as "many" or "more than half") be changed in a manner that is analogous to that of "many" in Experiment 1? (2) Is there a generalization of this learning effect to the untrained quantifier "many"? Again, all subjects participated in an informed consent procedure that was approved by an Institutional Review Board convened at the University of Pennsylvania. Methods ------- The same subjects as in Experiment 1 completed another experiment that was almost identical to Experiment 1. The only difference was the following: In block 2 of Experiment 2, the subjects were now trained to call proportions of 50% and 60% of circles "few." YES responses were encouraged for 60% or less yellow circles, while NO responses were reinforced for proportions above 60% yellow circles. In Experiment 2, the proportion ranged between 30 and 80% of circles of the target color. The order of Experiments 1 and 2 were counterbalanced within a single experimental session such that potential carry-over effects from one experiment to the other could be controlled (10 subjects started with Experiment 1, 11 started with Experiment 2). Analyses were analogous to those in Experiment 1. Results ------- ### Adaptation effect The 2 × 2 × 6 ANOVA yielded significant main effects for BLOCK \[F~(1,\ 20)~ = 5.82; p = 0.026\] and QUANTIFIER \[F~(5,\ 16)~ = 78.67; p \< 0.001\] and a trend for PROPORTION \[F~(1,\ 20)~ = 2.71; p = 0.059\]. Moreover, the following interaction terms were significant (QUANTIFIER × BLOCK: F~(1,\ 20)~ = 25.61; p \< 0.001; QUANTIFIER × PROPORTION: F~(5,\ 16)~ = 763.71; p \< 0.001; BLOCK × QUANTIFIER × PROPORTION: F~(5,\ 16)~ = 6.92; p = 0.001\]. The interaction term for BLOCK × PROPORTION was marginally significant \[F~(5,\ 16)~ = 2.43; p = 0.081\]. These effects describe a strong adaptation effect for the trained quantifier "few" and a weaker transfer effect for the not trained quantifier "many" (Figure [4A](#F4){ref-type="fig"}). ![(A) Average acceptability ratings for a given proportion of circles of the mentioned color, plotted separately for the quantifiers "many" (black lines) and "few" (gray lines) in the baseline blocks (dashed lines) and the test blocks after adaptation (solid lines). (B) Average acceptability ratings for a critical proportion of circles of the mentioned color, plotted separately for "many" (black bars) and "few" (gray bars) in the baseline blocks (dashed bars) and the test blocks after adaptation (solid bars). ^\\^p \< 0.01, ^\\\^p* \< 0.001.](fpsyg-06-00441-g0004){#F4} Testing the learning/transfer effect for "many" and "few" at the critical proportion "50%" pinpoints this overall effect (Figure [4B](#F4){ref-type="fig"}), showing a highly significant increase in acceptability for "few" \[t~(20)~ = −5.49; p \< 0.001\] and a significant decrease for "many" \[t~(20)~ = 3.10; p = 0.006\] (one-tailed t-tests, p-Bonferroni-corrected). At proportion "60%," there was again a significant learning effect for "few" \[t~(20)~ = −2.90; p = 0.009\] but no effect for "many" \[t~(20)~ = 0.84; p = 0.411\]. ### Learning during the adaptation phase Testing the learning effect for "few" in the adaptation block yielded on-going learning (Figure [5](#F5){ref-type="fig"}). This effect was significant from sub-block 1 to sub-block 3 \[t~(20)~ = −3.57; p = 0.001\] as a result of a significant increase from sub-block 1 to sub-block 2 \[t~(20)~ = −2.55; p = 0.005\] and a strong trend from sub-block 2 to sub-block 3 \[t~(20)~ = −1.67; p = 0.056; all tests one-tailed, uncorrected for multiple comparisons\]. ![*Learning curves for the new meaning of "few"* during the adaptation phase (block 2), plotted separately for each of the three sub-blocks*. Learning effects for proportions 50 and 60 are visible.](fpsyg-06-00441-g0005){#F5} ### Stronger adaptation effect for "many" (Experiment 1) than for "few" (Experiment 2) Next, in order to test whether the direct adaptation effect for "few"* observed in Experiment 2 was equal in size to that for "many" in Experiment 1, we tested across experiments for the interaction of BLOCK and trained QUANTIFIER at the respective critical proportions "40%" (for Experiment 1) and "50%" (for Experiment 2). This interaction effect proved to be significant \[F~(1,\ 20)~ = 4.44; p = 0.048\], indicating that the adaptation effect for "many" was larger in size than that for "few." ### No effect of the order of Experiments 1 and 2 Finally, in order to test whether the order in which the subjects underwent the two experiments had any systematic influence on the data pattern, we re-ran this analysis adding ORDER as a between-subject factor. Neither the main effect of ORDER \[F~(1,\ 19)~ = 1.10; p = 0.308) nor any 2-way or 3-way interaction with ORDER was significant (all p \> 0.05). Discussion ---------- In Experiment 2, we tested whether the pattern of results from Experiment 1 would be mirrored in adaptation effects for the negative quantifier "few" with a generalization to the untrained quantifier "many." The findings were straight-forward. There was a clear learning effect for few at both trained proportions "50%" and "60%" as well as a transfer to the criterion for "many" at "50%." Moreover, the in-depth analysis of the learning curves in the three parts of the adaptation block revealed again relatively rapid learning of the new criterion from the first-third to the second-third of the adaptation phase. These data corroborate the initial findings and conclusions, i.e., that a change in meaning of a quantifier may be contextually induced and that this change in meaning may also affect related quantifiers even if they are not exact opposites or antonyms. The fact that the same subjects underwent both experiments precluded problems of between-group comparisons; the fact that the order of experiments was counterbalanced over subject ensured that the adaptation and transfer effects observed in Experiment 2 were not just a resetting to the subjects\' original criteria temporally shifted for the duration of Experiment 1. Still, we found that the training for "many" elicited a stronger adaptation effect than that for "few." The implications of the overall pattern of results will now be discussed in more detail. General discussion ================== In this paper we reported two experiments that tested whether the meaning of a quantifier such as "many" or "few" can be altered, and whether this would affect the meaning of another, untrained quantifier as well. The findings were straight-forward, with effects for explicit training as well as transfer effects to the untrained quantifier for both quantifiers. These findings emphasize the flexible representations of conceptual knowledge. These data have several implications for the field of quantification and reasoning. For one, they suggest a certain flexibility of our interpretation of quantifiers that depends in part on the context. As in the introductory example, "many" can imply different amounts of e.g., animals depending on whether they are rare ("many lions" = 10) or not ("many ants" = 10,000), i.e., the semantic context in which it is used. As Sanford and Moxey (e.g., in their [@B21] paper) point out, quantifier use may reflect the individual perspective on, or interpretation of, a quantity rather than an inflexible, fixed characterization of its magnitude. This is in line with the finding that quantifier scopes partly overlap, as could be shown by Holyoak and Glass ([@B10]). These authors asked subjects to provide not only their first choice of an appropriate quantifier but also their second choice. They found for this latter task preference for quantifiers with closest distance on the continuum (all---many---some---a few---none). However, the notion of a flexible, partly subjective perspective on quantifier meaning may not be true for quantifiers in general. According to Routh ([@B20]), one way of grouping quantifiers is with respect to the degree of variability of the criterion they refer to. Routh distinguished between "fuzzier quantifiers" such as "few," "several," or "many" that are determined in part by context, Aristotelean or logical quantifiers that are "more precise" such as "all," "none," or "each," and parity or majority quantifiers that are "comparative" like "more" or "fewer." This distinction is based in part on the flexibility of the criteria for the referents of a given quantifier. Whereas "many donuts" may be two for person A and five for person B, comparative quantifiers imply at least the imparity of sets, i.e., ordinal information about which set is bigger. Most objective are quantifiers for which the criterion is explicitly or implicitly named, as in the case of "at least 13" or "more than half" (the latter referring to an amount of more than 50% of the total set)---objective in the sense that they refer to an explicitly stated criterion or degree. Routh ([@B20]) presented evidence from clustering algorithms and multidimensional scaling that this classification of quantifiers has psychological and empirical reality and is thus effective in communication. Given this distinction, one could formulate the hypothesis that the observed changes for quantifier semantics in the present study might generalize to other "fuzzy" quantifiers such as "some," "a few," "several," "quite a few," etc. Interestingly, within the class of "fuzzy quantifiers" like the majority quantifiers "many" and "few" tested in the present study, there seem to be differences between individual quantifiers. Whereas we did observe adaptation of the meaning of both quantifiers, we also found that this adaptation effect was even more pronounced for "many" than for "few." This result corresponds to earlier reports in the literature (e.g., Routh, [@B20]; Geurts, [@B5]; Heim et al., [@B8]; Shikhare et al., [@B22]) that quantifiers with negative polarity are more difficult to process. There are various accounts for this effect. One stresses that the positive quantifiers are (unmarked) defaults whereas negative quantifiers are marked (Clark, [@B2]; Clark and Chase, [@B3]). Alternatively, it has been proposed that negative quantifiers refer more or less precisely to the complement of the set denoted by a positive quantifier. Thus, in order to obtain the correct representation of a negative quantifier, the expression about the original set has to be negated, implying extra processing costs[^2^](#fn0002){ref-type="fn"}. Much more work is required here. Still, in the context of the present study, it could well be that negation or markedness pose greater cognitive demands on the comparison of a negative polarity quantifier like "few," resulting in gradually smaller flexibility and thus a smaller adaptation effect. If the meaning of a "fuzzy quantifier" can be changed, the question arises in how far such flexibility can also be found for quantifiers that are less fuzzy, i.e., that have fixed criteria. While the current study suggests that quantity can modulate quantifier interpretation, there is also additional evidence suggesting that object size, or mass, can modulate quantifier interpretation (McMillan et al., [@B15]). The use of quantifiers with different degrees of fuzziness in one experiment may restrict the flexibility of e.g., many and few that was observed here. Alternatively, it could also be the case that shifts along the continuum of criteria might even affect quantifiers with a fixed criterion: For instance, in the case of "at least X," subjects might be more inclined to accept such a statement when the distance of the actual numerosity to the criterion increases (cf. Shikhare et al., [@B22], for first results on distance effects related to "at least" and "at most"). Our findings address the question of the flexibility of word meaning, such as deciding when the diameter of a cup becomes broad enough to be called a bowl. This has been a challenging issue to address because of the difficulty quantifying the associated object concepts. Here we addressed this issue by examining quantifiers, a class of words that is based in part on number knowledge. We found that word meaning is highly flexible. It is trivial to re-label a concept, as in our explicit training paradigm. However, we also found a shift in the meaning of untrained words. This suggests a modification of word meaning at a deeper, conceptual level. The mental flexibility allowing the development of a modified meaning was not transient and determined by the experimental paradigm, but appeared to be maintained throughout the test block. Thus, in contrast to models of word meaning that depend entirely on a set of fixed features (Locke) or reference to an expert (Goodman), our findings are compatible with the notion that a component of word meaning is associated in part with a central tendency (Wittgenstein) that is somewhat flexible in the reference of a word. Another question that needs to be addressed in future research is how long the criterion shift effect induced in the adaptation phase actually lasts, i.e., how long it takes for the learning curve to go back to baseline level. Any considerations of how to make use of the observed effect in therapeutic settings where magnitudes or amounts are being misevaluated by patients (as in additive disorders or body schema disorders) will depend on its natural decay. This decay may in part depend on working memory processes which have been discussed as essential for quantifier processing (McMillan et al., [@B14]; Zajenowski et al., [@B25]). The present findings and the paradigm used to produce them may be helpful in order to resolve a longer-standing debate about the nature of the brain network underlying quantifier processing. McMillan et al. ([@B14]) found increasing involvement of the left inferior frontal cortex in quantifiers with increasing processing demands ("at least 3," "all," "some" vs. "less than half," "odd," "even") and related this effect to aspects of working memory (see Wei et al., [@B24], for comparable findings in Chinese). More recently, Heim et al. ([@B8]) reported two fronto-parietal networks for quantifier processing in the left hemisphere that were neighboring but hardly overlapping. One of these networks was related to the initial estimation of numerosities, the other to the comparison with a criterion. Interestingly, only the frontal component also showed systematic effects of processing quantifier meaning, suggesting a functional segregation between the frontal and the parietal components of the two networks. The paradigm presented here may be useful to get deeper insights into the role of the frontal component in order to understand in how far it is involved in the processing of change of meaning during the adaptation phase and/or maintenance of the newly learned meaning in working memory and its decay. Thus, subsequent work both with functional neuroimaging and also with patients with frontal or parietal neurodegeneration is encouraged in order to identify the neural basis of quantifier processing from the initial percept to the final semantic evaluation. Some first steps into that direction have been done by McMillan et al. ([@B15]) and by Morgan et al. ([@B15a]) in studies of patients with fronto-temporal and/or posterior cortical disease. In accordance with Dehaene\'s triple-code model (Dehaene et al., [@B4]) it was again atrophy in the frontal cortex that had a severe effect on the verbal code of numerosities. It is yet an open question whether it is the dysfunction of parts of the frontal cortex itself, or rather dynamic diaschisis (i.e., alterations in the entire network of frontal and parietal areas, as could be supposed based on Troiani et al., [@B24a]) that thus affects the neurobiology of the language-number interface in the human brain. Conclusion {#s2} ========== The present study investigated quantifier processing in a setting in which their meaning could be experimentally altered for the course of the experiment. This novel paradigm yielded results suggesting that, in order to be most highly informative, a change in meaning in one quantifier generalizes also to that of other quantifiers, i.e., to a larger referential frame. Future studies are required to test the usefulness and limits of this paradigm behaviorally and with neuroimaging. Conflict of interest statement ------------------------------ The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The study was funded by the National Institutes of Health (NS044266, AG017586, AG032953, NS053488, and AG043503) and the Wyncote Foundation. SH received financial support for his lab visit at UPenn from the Department of Psychiatry, Psychotherapy and Psychosomatics (Frank Schneider) and the Section for Clinical and Cognitive Neuroscience (Ferdinand Binkofski), Uniklinik RWTH Aachen, as well as from the Institute for Neuroscience and Medicine (INM-1; Katrin Amunts) at the Research Centre Jülich. Thanks go to Yosef Grodzinsky and Klaus Willmes for ongoing discussion of quantifier processing, and especially to Birgit Heim for making this lab visit possible. ^1^The assignment of the YES and NO responses to the one or other button were kept constant for each single subject for the duration of the entire experiment. Across subjects, the assignment of YES or NO to the left or right button was balanced. Thus, 50% of the subjects pressed the left button for YES and the right for NO and vice versa for the other 50%. ^2^Note, however, that "few" is not exactly equivalent to "not many." This is due to the fact that these two quantifiers do not refer to the same reference, or criterion, whatever its value may be in a given context and for a given individual. In the example of peanuts, it is well possible that you neither eat few peanuts (i.e., a handful) nor many (i.e., the whole package), but just some intermediate amount. The situation might be different for quantifiers that are exact opposites not by virtue of explicit negation ("more than half"/"not more than half") but of semantic polarity ("more than half"/"less than half"). [^1]: Edited by: Sonja A. E. Kotz, Max Planck Institute Leipzig, Germany [^2]: Reviewed by: Anna J. Simmonds, Imperial College London, UK; John E. Drury, Stony Brook University, USA [^3]: This article was submitted to Language Sciences, a section of the journal Frontiers in Psychology	{ "pile_set_name": "PubMed Central" }
Introduction {#Sec1} ============ Hair cell function in the cochlea and the vestibular system is calcium (Ca^2+^)-dependent (Tanaka et al. [@CR39]; Ohmori [@CR27]; Kozel et al. [@CR18]) and requires the endolymphatic Ca^2+^ concentration (\[Ca^2+^\]~endolymph~) to be tightly controlled at a level that is unusually low (0.017--0.133 mmol/l (Salt et al. [@CR35])) for an extracellular fluid (e.g., \[Ca^2+^\]~blood\ plasma~ = 1.2 mmol/l (Diem and Lentner [@CR10])). To maintain this physiological \[Ca^2+^\]~endolymph~, numerous epithelial sites within the membranous labyrinth are engaged in Ca^2+^ transport across the endolymph-perilymph barrier, utilizing Ca^2+^-permeable channels (e.g., transient receptor potential cation channel subfamily V members \[TRPV channels\] (Takumida et al. [@CR38]; Wangemann et al. [@CR43]; Nakaya et al. [@CR25]; Ishibashi et al. [@CR17])), ion exchangers (e.g., sodium-calcium exchangers \[NCX\] (Oshima et al. [@CR28]; Yamauchi et al. [@CR44])), and Ca^2+^ ATPases (e.g., plasma membrane calcium ATPase \[PMCA\] (Crouch and Schulte [@CR8]; Ågrup et al. [@CR1])). Many of these transport proteins are directly regulated downstream targets of the calcium sensing receptor (CaSR), a G-coupled cell surface receptor that is prominently expressed in the primary tissues involved in whole-body Ca^2+^ homeostasis, including the kidney and the parathyroid and thyroid glands (Blankenship et al. [@CR4]; Topala et al. [@CR40]; Ranieri et al. [@CR30]). In those tissues, CaSR senses changes in Ca^2+^ levels in urine and plasma and modulates the activity and expression of Ca^2+^ transport proteins to increase renal Ca^2+^ reabsorption (Riccardi et al. [@CR31]) and the Ca^2+^-mediated release of calciotropic hormones (parathyroid hormone, calcitonin) (Brown and MacLeod [@CR5]), respectively. In the inner ear, the endolymphatic sac (ES), among other epithelial tissues, is a site of CaSR gene expression (Beitz et al. [@CR3]; Lin et al. [@CR19]) and critically contributes to endolymphatic Ca^2+^ homeostasis. This function was first indicated by in vivo animal experiments in which surgical separation of the ES resulted in a nonphysiological increase in \[Ca^2+^\]~endolymph~ and endolymphatic hydrops development in the cochlea (Ninoyu and Meyer zum Gottesberge [@CR26]; Salt and DeMott [@CR33]). Structural loss of the distal (extraosseous) ES, as induced in this animal model, is a histopathological hallmark and a presumed etiopathogenic factor in Meniere's disease (MD) (Eckhard et al. [@CR13]). Here, we investigated the immunolocalization patterns of CaSR and Ca^2+^ transport proteins in the mature murine ES epithelium to elucidate the molecular determinants of Ca^2+^ homeostasis in the ES and to better understand the pathophysiological consequences of the structural loss of the ES in MD. Materials and methods {#Sec2} ===================== Animals {#Sec3} ------- All animal experiments were performed according to Swiss animal welfare laws and were approved by the local veterinary authorities (Kantonales Veterinäramt Zürich; protocol no. ZH269/16). Male, 6- to 8-week-old C57BL/J6-Crl1 mice were purchased from Charles River Laboratories, Inc. (Sulzfeld, Germany). A total of 12 animals were used in this study. All animals were kept at an in-house animal facility with free access to a standard chow diet and water under a standard 12 h light/dark cycle. Tissue preparation, fixation, decalcification, embedding, and sectioning {#Sec4} ------------------------------------------------------------------------ Animals were sacrificed by CO~2~ inhalation. For tissue fixation, the thoracic cavity was opened, and a cannula was inserted in the left cardiac ventricle for transcardial perfusion with 10 ml phosphate buffered saline (PBS), followed by 10 ml of a fixative, i.e., (i) 10% neutral buffered formalin (F; Lucerna-Chem AG, Luzern, Switzerland), (ii) F with 1% acetic acid (FA; Sigma-Aldrich, Steinheim, Germany), (iii) F with 0.2% glutaraldehyde (GA; Sigma-Aldrich), or (iv) FA with 0.2% glutaraldehyde (FGA). The respective fixative used in each protocol is given in Table [1](#Tab1){ref-type="table"}. Then, the animals were decapitated, soft tissues were removed from the outer surface of the skull using a sharp scalpel, and the cranial part of the skull was opened to remove most of the cerebrum, while the cerebellum was left in situ. The tympanic bullae were opened to allow faster penetration of the fixative into the inner ear. The specimens were then immersed in the respective fixative used for transcardial perfusion for 6 h on a countertop shaker. Afterwards, the specimens were dehydrated in a graded series of ethanol solutions (50%, 70%, 95%, and 100%), embedded in paraffin and sectioned at 4 μm using an HM 355S Automatic Microtome (Thermo Fisher Scientific, Waltham, MA, USA). Sections were collected on SuperFrost Plus slides (Thermo Fisher Scientific), dried on a heating plate at 37 °C overnight and stored at room temperature.Table 1Primary antibodies used with DAB labeling. Dilutions indicate the highest dilution producing a sufficient signal. All antibodies were diluted in 1% normal horse serum (NHS). Form indicates 10% neutral buffered formalin. FA form + 1% acetic acid, FG form + 0.2% glutaraldehyde, FGA FA + 0.2% glutaraldehyde, HIER heat-induced epitope retrievalEpitopeCalbindin D-28 kCaSRNCX2ParvalbuminpPMCAPMCA1PMCA2SERCA1SERCA2TRPV5TRPV6Immunogen (peptide sequence)No infoADDYGRPGIEKFREEAEE RDIRVGDAQGMFEPDGGNo infoExact sequence not available^a^SGVKNSLKEANHDKEIPDPSSINAKTLETPDQVKRHLEKYGCTPNKPSRTSMSKGLNLSEGDGEEVYHFNRGLEDGESWEYQIFixativeFGFAFGFAForm, FGFGFGFAFAForm, FGForm, FGHIERNoYesYesYesNoNoNoNoNoYesYesPrimary antibodyMonoclonal anti-calbindin D-28 k (mouse)Monoclonal anti-CaSR (mouse)Polyclonal anti-NCX-2 (rabbit)Monoclonal antiparvalbumin (mouse)Monoclonal anti-pPMCA (mouse)Polyclonal anti-PMCA1 (rabbit)Polyclonal anti-PMCA2 (rabbit)Polyclonal anti-SERCA1 (rabbit)Polyclonal anti-SERCA2 (rabbit)Polyclonal anti-TRPV5 (rabbit)Polyclonal anti-TRPV6 (rabbit)Dilution1:5001:15′0001:5001:5001:40001:10001:10001:5001:5001:2001:200ManufacturerSwant, Martiny, SwitzerlandNovus Biologicals, Minneapolis, USAAlomone Labs, Jerusalem, IsraelSwant, Martiny, SwitzerlandAbcam, Cambridge, UKAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelCatalog number300NB120-19347SSANX-012235ab2825ACP-005ACP-002ACP-011ACP-012ACC-035ACC-036Secondary antibody (biotinylated)Anti-mouse IgG (donkey)Anti-mouse IgG (donkey)Anti-rabbit IgG (donkey)Anti-mouse IgG (donkey)Anti-mouse IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Dilution1:4001:4001:4001:4001:4001:4001:4001:4001:4001:10001:1000ManufacturerMilan Analytica AG, Rheinfelden,Milan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGCatalog number715-065-151715-065-151715-065-152715-065-151715-065-151715-065-152715-065-152715-065-152715-065-152715-065-152715-065-152^a^This antibody recognizes an epitope between amino acids 724--783 of the human erythrocyte calcium pump Immunohistochemical DAB labeling {#Sec5} -------------------------------- Sections were deparaffinized in Histo-Clear (National Diagnostics, Atlanta, GA, USA), rehydrated in a graded series of ethanol solutions (100%, 95%, 70%, and 50%), and rinsed in tap H~2~O. When required, heat-induced antigen retrieval (HIAR) with pressurized coverslipping of the mounted tissue sections was performed (Table [1](#Tab1){ref-type="table"}) according to a previously established protocol (Eckhard et al. [@CR12]). All subsequent steps were performed at room temperature. Nonspecific binding was blocked with 1% normal horse serum (NHS) for 15 min, followed by incubation with primary antibodies overnight. The sections were then incubated with biotinylated secondary antibodies for 1 h, followed by incubation with avidin-biotin-HRP complex (Vectastain ABC HRP Kit, Vector Laboratories, Burlingame, CA, USA). Visualization was performed with 3,3′-diaminobenzidine (DAB) (DAB Peroxidase \[HRP\] Substrate Kit, Vector Laboratories). All incubation steps were performed at room temperature, and each incubation step was followed by rinsing the sections in PBS for 5 min. All primary and secondary antibody combinations utilized in this study are listed in Table [1](#Tab1){ref-type="table"}. Negative controls for all immunolabeling experiments were obtained by omitting the primary antibody or after preabsorption of the primary antibodies with the corresponding commercially available control peptides (for primary antibodies against TRPV5/6, PMCA1/4, NCX2, and SERCA1/2). For nuclear counterstaining, the slices were incubated for 5 s with hematoxylin diluted in distilled water (1:6) and mounted with a permanent mounting medium (VectaMount, Vector Laboratories). Immunofluorescence double labeling {#Sec6} ---------------------------------- Deparaffinization, rehydration of the sections, and incubation with primary antibodies was performed as described for DAB-labeling experiments. All primary antibodies used for immunofluorescence double labeling are listed in Table [2](#Tab2){ref-type="table"}. The sections were then incubated with biotinylated antibodies raised in donkey and directed against mouse (1:400; 715-065-151, Milan Analytica AG, Rheinfelden, Switzerland) and Alexa Fluor 594-conjugated antibodies raised in goat and directed against rabbit (1:400; 111-585-144, Milan Analytica AG) for 1 h, followed by incubation with avidin-biotin-HRP complex (Vectastain ABC HRP Kit, Vector Laboratories). Biotinylated tyramine (10 min) and a subsequent second incubation with avidin-biotin-HRP complex were used for signal amplification. Visualization was performed with Alexa Fluor 488-conjugated streptavidin (1:800; Milan Analytica AG) and Alexa Fluor 594-conjugated anti-goat antibodies raised in donkey (1:400; 705-585-003, Milan Analytica AG). The sections were mounted with an aqueous mounting medium containing DAPI for nuclear counterstaining (H-1200, Vectashield, Vector Laboratories).Table 2Primary antibodies used with fluorescence labeling. Dilutions indicate the highest dilution producing a signal. All antibodies were diluted in 1% normal horse serum (NHS). If not otherwise specified, sodium citrate (10 mM) at pH 6.0 was used as the HIER buffer. Form indicates 10% neutral buffered formalin. FA form + 1% acetic acid, FG form + 0.2% glutaraldehyde, FGA FA + 0.2% glutaraldehyde, HIER heat-induced epitope retrieval, NHS normal horse serumEpitopeCaSRγENaCTRPV5V-ATPaseImmunogen (peptide sequence)ADDDYGRPGIEKFREEAEERDIAvailable upon request.GLNLSEGDGEEVYHFMAMEIDSRPGGLPGSSCNLGAAREHMQAVTRNYITHPRVTYRTVCSVNGPLVVLDR VKFAQYAEIVFixativeFormFormFGFAHIERYesYesYesYes (Tris-EDTA solution, 10 mM / 1 mM, at pH 9.0)Primary antibodyMonoclonal anti-CaSR (mouse)Monoclonal anti-γENaC (rabbit)Polyclonal anti-TRPV5 (rabbit)Polyclonal anti-Vacuolar-type H^+^ ATPase subunit B, kidney isoform (rabbit)Dilution1:2001:2001:1001:200ManufacturerNovus Biologicals, Minneapolis, USAGift from Prof. Johannes Loffing, University of Zurich, SwitzerlandAlomone Labs, Jerusalem, IsraelNovus BiologicalsCatalog numberNB120-19347SSn. a.ACC-035NBP2-33962 Microscopy analysis {#Sec7} ------------------- Images were acquired using either a Leica DMI6000 microscope (Leica, Wetzlar, Germany) or a Leica SP5 confocal laser scanning microscope (Leica) and processed with Adobe Photoshop CS5 software (Version 12.0, Adobe Systems, San Jose, CA, USA). Quantification of immunolabeled ES epithelial cells {#Sec8} --------------------------------------------------- DAB-labeled (DAB^+^) ES epithelial cells were quantified on light microscopic (× 40 air objective) images of hematoxylin-counterstained tissue sections. Cell counting was performed in at least three non-consecutive sections from two different axial planes of the intraosseous ES (iES) and the extraosseous ES (eES). Each section was divided into ten equally long segments along the longitudinal axis of the ES. The percentage of DAB^+^ cells among the epithelial cells with hematoxylin-stained nuclei was determined in each segment. Single- and double-immunofluorescence-labeled ES epithelial cells in the eES were quantified following double immunofluorescence labeling of CaSR with either V-ATPase, TRPV5, or γENaC. The percentages of fluorescence-labeled cells among the epithelial cells with DAPI-counterstained nuclei were determined. From each double immunofluorescence labeling experiment, at least 100 labeled cells were counted, and sections from at least 3 independent specimens were analyzed. Descriptive statistics {#Sec9} ---------------------- Statistical analyses were performed using Prism software (version 7.0, GraphPad Software). Cell counts are expressed as percentages of the number of epithelial cells in a segment of the ES. Mean values and standard deviations derived from six animals are shown. Results {#Sec10} ======= Based on the previously demonstrated similarities between the ES epithelium and the kidney distal tubular epithelium on the cellular and molecular levels, we mainly focused our immunohistochemical analysis on Ca^2+^ transport-associated proteins that are known to be crucial for the renal tubular handling of calcium. Cellular immunolocalization patterns of the CaSR and Ca^2+^ transport proteins {#Sec11} ------------------------------------------------------------------------------ DAB immunolabeling of the CaSR, as well as of TRPV5, TRPV6, SERCA1, SERCA2, NCX2, PMCA1, and PMCA4 was visible in the iES (proximal to the operculum) and eES (distal to the operculum; Fig. [1](#Fig1){ref-type="fig"} and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}) portions of the ES epithelium. Staining of proteins was absent or very weak in the endolymphatic duct (ED; Fig. [1](#Fig1){ref-type="fig"} and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}). Negative controls (omitted primary antibodies and preabsorption of primary antibodies with control peptides (TRPV5/6, NCX2, PMCA1/4, SERCA1/2) showed no specific staining in the murine ES epithelium.Fig. 1DAB-immunohistochemical staining of the calcium sensing receptor (CaSR) and selected calcium transport proteins along the murine endolymphatic sac. a, b Overviews of the iES emerging from the ED (a) and of the eES (b), demonstrating epithelial CaSR immunostaining along the ES. Boxed numbers indicate representative regions shown below in columns. c--f Representative images corresponding to the regions indicated by the boxed numbers in (a) and (b), i.e., the ED-iES transition (first column), iES (second column), and eES (third column). Scale bars: 100 μm (a), 20 μm (c). ED endolymphatic duct, eES extraosseous endolymphatic sac, iES intraosseous endolymphatic sac, sSCC superior semicircular canal CaSR immunolabeling was present in the iES and eES portions (Fig. [1a, b](#Fig1){ref-type="fig"}) and exhibited a scattered labeling pattern within the epithelium. In CaSR-labeled cells, strong cytoplasmic and membranous labeling was found, which in many cells was polarized to the basolateral membranes (Fig. [1c--c](#Fig1){ref-type="fig"}"). Immunolabeling signals for transient receptor potential cation channel subfamily V (TRPV) members TRPV5 and TRPV6 appeared to be cytoplasmic with strong polarization towards the apical cell pole, most clearly in epithelial cells in the eES (Fig. [1d--d](#Fig1){ref-type="fig"}" and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}A-A"). Homogeneous cytoplasmic immunolabeling of the sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA) subtypes SERCA1 and SERCA2 was found throughout the iES and eES epithelium (Fig. [1e--e](#Fig1){ref-type="fig"}" and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}B-B") and was consistent with their known localization in the endoplasmic reticulum (Strehler and Treiman [@CR37]). Labeling for the plasma membrane calcium ATPase (PMCA) subtypes PMCA1 and PMCA4, as well as for the sodium-calcium exchanger NCX2, was polarized in the basolateral membranes, most clearly in the eES (Fig. [1f--f](#Fig1){ref-type="fig"}" and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}). No immunolabeling in the ES and ED was found for PMCA2 and NCX1, nor for the Ca^2+^-binding proteins Calbindin D-28k and parvalbumin (Supplementary Fig. [2](#Fig6){ref-type="supplementary-material"}). All immunolabeling patterns and their cellular polarization in the ES epithelium were consistent with those previously described for the distal convoluted tubule and the connecting tubule epithelium (Loffing et al. [@CR21]; Loffing and Kaissling [@CR20]); positive control experiments are shown in Supplementary Fig. [3](#Fig7){ref-type="supplementary-material"}). Spatial (proximal-to-distal) immunolabeling gradients {#Sec12} ----------------------------------------------------- For each protein studied, DAB-labeled cells were quantified along the longitudinal (proximal-to-distal) axis of the ES epithelium, reaching from the proximal transition of the iES in the ED to the distal blind end of the eES. For all of the investigated proteins, a gradual increase in the immunolabeling signal in the proximal-to-distal direction was observed. At the very distal end of the eES portion, on average, more than 50% of the epithelial cells were positively labeled for most of the investigated proteins (Fig. [2](#Fig2){ref-type="fig"}).Fig. 2Quantification of DAB^+^ cells along the ES reveals staining gradients along the ES epithelium increasing from the iES to the eES. The y-axis shows the percentage of DAB^+^ cells relative to all ES epithelial cells. For cell counting, both the iES and eES were each divided into ten equal segments. The ticks on the x-axis correspond to each of the ten consecutive segments of the iES (left x-axis) and eES (right x-axis). The first two segments represent the endolymphatic duct epithelium. The values are the percentages of the total number of cells in the respective ES segment and are given as mean values ± SD from six animals. a Quantification of DAB^+^ cells for CaSR and the apical Ca^2+^ channels TRPV5 and TRPV6. b Quantification of DAB^+^ cells for the cytoplasmatic proteins SERCA1 and SERCA2. c Quantification of DAB^+^ cells for proteins located in the basolateral membrane Differential immunolocalization patterns in mitochondria-rich and ribosome-rich cells {#Sec13} ------------------------------------------------------------------------------------- We found scattered DAB-immunolabeling patterns within the ES epithelium for almost all of the investigated proteins (Fig. [1](#Fig1){ref-type="fig"}). We therefore investigated whether CaSR and Ca^2+^ transport proteins are differentially distributed among the two major ES epithelial cell types, i.e., mitochondria-rich cells (MRCs) and ribosome-rich cells (RRCs) (Lundquist et al. [@CR23]; Barbara et al. [@CR2]; Dahlmann and von Düring [@CR9]), using immunofluorescence double labeling. Labeling of the MRC-specific vacuolar-type H^+^-ATPase (V-ATPase; (Stanković et al. [@CR36]; Dou et al. [@CR11])) and CaSR in the eES portion showed a strict cellular colocalization of both proteins (Fig. [3a--a\'\'\'](#Fig3){ref-type="fig"}) in 45.0% of the cells (Fig. [3b](#Fig3){ref-type="fig"}), indicating the localization of CaSR in MRCs (vATPase^+^/CaSR^+^) but not in RRCs (vATPase^−^/CaSR^−^). Colabeling of TRPV5 and CaSR (Fig. [3c-c\'\'\', d](#Fig3){ref-type="fig"}) was found in 31.4% of the cells (TRPV5^+^/CaSR^+^), while the other cells showed exclusive labeling for either TRPV5 (TRPV5^+^/CaSR^−^, 19.3%) or CaSR (TRPV5^−^/CaSR^+^, 22.9%) or were devoid of labeling (TRPV5^−^/CaSR^−^, 26.4%). Colabeling of the gamma subunit of the epithelial sodium channel (γENaC) and CaSR (Fig. [3e-e\'\'\', f](#Fig3){ref-type="fig"}) was present in 28.1% of the cells (γENaC^+^/CaSR^+^), while 20.3% of cells showed labeling for only γENaC (γENaC^+^/CaSR^−^), 29.4% of the cells were labeled for only CaSR (γENaC^−^/CaSR^+^), and 22.2% of cells were devoid of labeling (γENaC^−^/CaSR^−^).Fig. 3Quantitative colocalization of CaSR with V-ATPase, TRPV5, and γENaC in the eES. Representative images and quantification of fluorescent immunohistochemical double labeling of CaSR and V-ATPase (a--a"', b), TRPV5 (c--c"', d), and γENaC (e--e"', f). Scale bar: 10 μm. N total number of cells counted CaSR immunolabeling within the murine inner ear is restricted to the ES epithelium {#Sec14} ---------------------------------------------------------------------------------- No specific CaSR immunoreactivity was found in other portions of the membranous labyrinth, including the cochlea, saccule, utricle, and semicircular canal cristae (Fig. [4](#Fig4){ref-type="fig"}).Fig. 4No CaSR immunolabeling was detected within the cochlea or the vestibular end organs. a Negative CaSR immunolabeling in the cochlea, in particular the stria vascularis, spiral ligament, Reissner's membrane, limbus, organ of Corti, and spiral ganglion. b Ampullary region of the lateral semicircular canal. No CaSR staining was observed in the crista or the SCC epithelium (membranous labyrinth). Adjacent to the crista, several dark cells (melanocytes) are interspersed in the membranous labyrinth. c, d No CaSR immunolabeling was found in the sensory epithelium or surrounding tissue of the utricle (c) and saccule (d). Several dark cells are interspersed in the membranous labyrinth covering the utricle (c) Discussion {#Sec15} ========== In the present study, the ES epithelium was identified as an exclusive site of CaSR immunolocalization within the mature murine inner ear. Several CaSR-regulated Ca^2+^ transport proteins (TRPV5, PMCA1/4), as well as non-CaSR-regulated Ca^2+^ channels/transporters (TRPV6, NCX2), and intracellular Ca^2+^ sequestering proteins (SERCA1/2) were localized in the ES epithelium, almost all of which exhibited their highest immunolabeling intensities in the distal eES portion. Consistent with the immunolocalization of CaSR in MRCs of the mature murine ES shown here, previous gene expression studies detected CaSR mRNA expression in MRCs of the early postnatal murine ES (Honda et al. [@CR15]), in the adult rat ES (Beitz et al. [@CR3]), and, moreover, in the adult rat organ of Corti, stria vascularis, Reissner's membrane, and vestibulum (Beitz et al. [@CR3]), as well as in neuromast hair cell stereocilia of zebrafish larvae (Lin et al. [@CR19]). The results of the present study indicate that the CaSR protein in the mature murine inner ear is exclusively expressed in the ES or that immunolabeling at other cellular sites was below the detection level. We propose that the CaSR may have a unique role in sensing and regulating \[Ca^2+^\]~endolymph~ in the ES, with significance for the overall endolymphatic Ca^2+^ homeostasis in the inner ear. However, it is unknown whether the CaSR enables ES epithelial cells to sense Ca^2+^ changes in the ES endolymph, the cerebrospinal fluid, or in both extracellular fluid compartments that face the ES epithelium. Moreover, the molecular signaling cascades by which the CaSR may regulate downstream effector proteins, such as TRPV5 and PMCA1/4 (Blankenship et al. [@CR4]; VanHouten et al. [@CR41]; Topala et al. [@CR40]), which are also localized in the ES epithelium (Ågrup et al. [@CR1]; Honda et al. [@CR15]; present study), require further exploration. Although CaSR function in the ES may contribute to maintaining Ca^2+^ within its narrow physiological range in the entire inner ear, its effects on \[Ca^2+^\]~endolymph~ in the vestibular and cochlear portions are probably slow and may have physiological relevance primarily on longer time scales. This assumption is supported by previous observations that (i) longitudinal fluid movements (between endolymph compartments) virtually do not occur, at least under physiological conditions (Salt et al. [@CR34]), and that (ii) surgical separation of the ES only leads to a slow rise of \[Ca^2+^\]~endolymph~ over a time course of several weeks (Ninoyu and Meyer zum Gottesberge [@CR26]; Salt and DeMott [@CR33]). This may be primarily due to local ion homeostatic mechanisms within the cochlea and the vestibular portions (Lundquist [@CR22]; reviewed in Salt [@CR32]), as well as anatomical diffusion barriers, such as the utriculosaccular duct, which (in the adult murine inner ear) largely prevents fluid exchange between the pars inferior (cochlea, saccule) and the pars superior (utricle, semicircular canals) (Cantos et al. [@CR6]). In addition, the CaSR may have local homeostatic functions in the ES, such as maintaining ES epithelial integrity or mediating immune responses (Cheng et al. [@CR7]). Among the two main epithelial cell types in the ES, i.e., MRCs and RRCs, we found that the CaSR was exclusively localized in vATPase^+^ MRCs (Dou et al. [@CR11]; Honda et al. [@CR15]). This cell-type-specific distribution of CaSR resembles that in intercalated cells in the renal cortical collecting duct (CCD; (Everett [@CR14])). On the subcellular level, CaSR was polarized towards both membrane domains (apical and basolateral) with additional strong cytoplasmic labeling. This finding is consistent with the distribution of CaSR in the distal renal tubule epithelium (distal convoluted tubule and collecting duct (Riccardi et al. [@CR31])), where CaSR regulates transepithelial Ca^2+^ reabsorption from the tubular fluid via TRPV5 (Topala et al. [@CR40]) and PMCA (Blankenship et al. [@CR4]). Notably, other shared molecular mechanisms of transepithelial ion (Na^+^) transport between the ES epithelium and the distal renal tubule epithelium have been reported (Mori et al. [@CR24]; Eckhard et al. [@CR13]). In contrast to CaSR, TRPV5 was differentially localized in subsets of MRCs and RRCs, thereby differing from its exclusive localization in principal cells in the renal CCD (Loffing and Kaissling [@CR20]). However, this non-cell-type-specific TRPV5 localization in the ES is consistent with the RNA sequencing analysis of the murine ES (Honda et al. [@CR15]). These and previously noted structural differences between the epithelia of the ES and the renal CCD (Wangemann and Marcus [@CR42]), despite their numerous shared cellular and molecular features, may be due to the different physiological requirements for maintaining extracellular Ca^2+^ homeostasis in the ES endolymph and the blood plasma. In addition to CaSR-regulated Ca^2+^ transporters, we identified TRPV6 and NCX2 in the ES epithelium, both of which are known to enable constitutively active Ca^2+^ transport in the renal tubular epithelium (Loffing and Kaissling [@CR20]). Loss of the proposed Ca^2+^-homeostatic function of the ES may be of significance in various pathological conditions, such as otoconial disorders and MD. The biogenesis of otoconial Ca^2+^ carbonate (CaCO~3~) crystals is believed to occur in the ES luminal microenvironment and is sensitive to changes in \[Ca^2+^\]~endolymph~ (Nakaya et al. [@CR25]). For example, disturbed formation of otoconia was observed in Foxi1 knockout (k.o.) mice and Efnb2 k.o. mice (Hulander [@CR16]; Raft et al. [@CR29]). In these models, the MRC population in the ES is developmentally lacking (Hulander [@CR16]) or mildly decreased and proximally mislocalized in the ES (Raft et al. [@CR29]), respectively. The loss of CaSR-regulated Ca^2+^ homeostasis in the ES is one possible explanation for the compromised otoconia formation in these transgenic mouse models. Another explanation is the loss of (MRC-specific) pendrin- and v-ATPase-mediated anion transport that leads to disturbances of endolymphatic pH, which in turn may inhibit the pH-sensitive TRPV5/6 channels in the ES epithelium and thereby disturb \[Ca^2+^\]~endolymph~ ((Nakaya et al. [@CR25]), present study). In MD, the structural loss of the distal ES epithelium is a consistent histopathological hallmark. Presuming that the human ES harbors Ca^2+^ transport mechanisms in proximal-to-distal expression gradients analogous to the murine ES, structural loss of the distal ES---a pathological hallmark in MD (Eckhard et al. [@CR13])---and in particular its Ca^2+^ transport functions may be crucial in the pathogenesis of endolymphatic hydrops (Ninoyu and Meyer zum Gottesberge [@CR26]; Salt and DeMott [@CR33]). In conclusion, the murine distal ES epithelium presumably has a "calciostatic" function for inner ear fluid homeostasis. Human pathology of the ES likely impairs Ca^2+^ homeostasis of the inner ear fluids and is therefore potentially a significant pathophysiological factor in various inner ear disorders. Electronic supplementary material ================================= {#Sec16} Figure S1DAB-immunohistochemical staining of selected calcium transport proteins along the murine endolymphatic sac. (A -- E) Representative images corresponding to the regions indicated by the boxed numbers in Fig. [1a and b](#Fig1){ref-type="fig"}, i.e., the ED-iES transition (first column), iES (second column) and eES (third column). Scale bar: 20 μm. (PNG 165460 kb) High resolution image (TIFF 3415 kb) Figure S2A lack of DAB-immunohistochemical staining in the ES epithelium was observed for calbindin-D28k, parvalbumin, PMCA2 and NCX1. (A) Negative DAB immunolabeling of calbindin-D28k in the ES epithelium. Scale bar (for A, D, G and I): 20 μm. (B -- C) Positive control DAB staining for calbindin-D28k in the distal nephron (B) and in Purkinje cells of the cerebellar cortex (C). Scale bar in (B): 10 μm (for B, E and J); scale bar in (C): 10 μm (for C, F and H). (D) Negative DAB immunolabeling of parvalbumin in the ES epithelium. (E -- F) Strong positive control DAB staining in parvalbumin-immunostained cells in the distal nephron (E) and weak positive DAB staining in Purkinje cells of the cerebellar cortex (F). (G) Negative DAB immunolabeling of PMCA2 in the ES epithelium. (H) Positive control staining of PMCA2, with a predominant apical staining pattern in Purkinje cells of the cerebellar cortex (F). (I) Negative DAB immunolabeling of NCX1 in the ES epithelium. (J) Positive control staining for NCX1, with a basolateral staining pattern in the distal nephron (J). All scale bars: 20 μm. (PNG 91011 kb) High resolution image (TIFF 1899 kb) Figure S3Positive DAB control staining in the kidney and in the trigeminal ganglion for proteins exhibiting positive immunostaining in the ES epithelium (Fig. [1](#Fig1){ref-type="fig"}, Fig. [S1](#Fig5){ref-type="supplementary-material"}). (A -- H) Positive DAB staining in sections along the distal nephron for CaSR, TRPV5, TRPV6, SERCA1, SERCA2, PMCA1, PMCA4 and pPMCA. (I) As NCX2 is not expressed in the kidney, somata of trigeminal ganglion neurons served as a positive control for NCX2 staining. Scale bar: 20 μm. (PNG 70669 kb) High resolution image (TIFF 1758 kb) Ca^2+^ : (Ionized) calcium \[Ca^2+^\]~endolymph~ : Endolymphatic Ca^2+^ concentration Calb : Calbindin D-28k CaSR : Calcium sensing receptor CCD : Cortical collecting duct ES : Endolymphatic sac ED : Endolymphatic duct MRC : Mitochondria-rich cell NCX1/2 : Sodium-calcium exchanger type 1/2 Parv : Parvalbumin PMCA1/4 : Plasma membrane calcium ATPase isoform 1/4 pPMCA : Pan-plasma membrane calcium ATPase RRC : Ribosome-rich cell SERCA1/2 : Sarco/endoplasmic reticulum Ca^2+^-ATPase paralog 1/2 TRPV5/6 : Transient receptor potential cation channel subfamily V member 5/6 V-ATPase : Vacuolar-type H^+^ ATPase Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. The authors thank Sabina Wunderlin for excellent technical assistance. The authors thank Prof. J. Loffing (Institute of Anatomy, University of Zurich) for providing the anti-γENaC antibody. The study was supported by grants from the British Ménière's Society and the Zürcher Stiftung für das Hören. DB is supported by a national MD-PhD scholarship from the Swiss National Science Foundation (SNSF). AHE was supported by a career development (Filling the gap) grant from the University of Zurich. The authors declare that they have no conflicts of interest. The animal research protocol was approved by the local veterinary authorities (permission number ZH269/16, Kantonales Veterinäramt, Zurich, Switzerland). All animal experiments were performed according to Swiss Animal Welfare laws and were carried out according to the approved animal research protocol.	{ "pile_set_name": "PubMed Central" }
![](hosplond70322-0005){#sp1 .185}	{ "pile_set_name": "PubMed Central" }
1. Introduction {#sec1} =============== Despite ongoing efforts to improve maternal and child health in developing countries, mortality rates remain much higher than in developed countries. Women in developing regions face a lifetime risk of maternal death of 1 in 160, as compared with 1 in 3700 for women living in developed regions \[[@B1]\]. These inequalities are driven by many causes, one of which is limited access to preventive services. For example, in low- and middle-income countries, only about 52% of pregnant women receive the World Health Organization- (WHO-) recommended minimum of four antenatal visits \[[@B2]\]. The postnatal period is also critical to the health of a mother and newborn, as the majority of postnatal maternal deaths happen during the first week after birth \[[@B3]\]. However, a recent analysis of Demographic and Health Surveys for 23 African countries found that, of the two-thirds of women giving birth at home, only 13% received a postnatal check-up within two days \[[@B3]\]. Immunization is another critical preventive service that can save the lives of many infants and children. Despite being one of the most cost-effective tools for saving lives, nearly one in five children globally did not receive their full package of immunizations in 2012 and 1.5 million children under the age of 5 died from vaccine-preventable diseases in 2008 \[[@B4], [@B5]\]. Antenatal care (ANC), postnatal care (PNC), and childhood immunization make up an important package of preventive services that can improve maternal and child health. Families tend to use medical services when someone is ill but frequently omit these beneficial preventive services that are essential to improve health. The field of mHealth, or mobile health, has been proposed as a potential solution to many of the challenges that developing countries face, including workforce shortages, lack of health information, limited training for health workers, and difficulty tracking patients. mHealth projects have been implemented all over the world, using mobile phones for record keeping, data collection, or patient communication \[[@B6]\]. Further, mHealth tools have been used to promote behavior change in health workers and/or patients. For example, text message reminders have been shown to increase care-seeking behavior or medication adherence in some patients and mobile data collection and communication tools for health workers have improved follow-up of patients and data reporting \[[@B7]--[@B9]\]. Though there are relatively few thorough evaluations of mHealth programs \[[@B6]\], some published studies do exist. Given that mHealth tools have shown some promise for behavior change more broadly, there is potential for this field to improve essential preventive maternal and child health services as well. Based on the existing evidence in peer-reviewed publications, this literature review aims to determine the effectiveness of mHealth tools to increase the coverage and use of antenatal care, postnatal care, and childhood immunizations through behavior change in low- and middle-income countries. 2. Materials and Methods {#sec2} ======================== 2.1. Information Sources {#sec2.1} ------------------------ This literature review was conducted through a keyword search of the following databases to identify relevant peer-reviewed articles: Google Scholar, PubMed, Embase, PsycINFO, and EBSCO Host. Keywords used in these searches includedmHealth,mobile health,mobile phone,reminder,recall,mobile medical records,antenatal care,postnatal care, andimmunization. 2.2. Inclusion Criteria {#sec2.2} ----------------------- In order to be included in the review, the article had to meet the following inclusion criteria:study was evaluating an mHealth intervention targeted at increasing antenatal care attendance, postnatal care attendance, or childhood immunization rates through behavior change;study was implemented in a low- or middle-income country;study included measurement of process, behavior change, health, or quality of care outcomes (i.e., studies were excluded that only evaluated willingness of participants to receive an mHealth intervention, without implementing it);study was a peer-reviewed article;study was available in English;study was published between January 1, 2000 and November 20, 2014.These criteria were selected to ensure that the included studies examined outcomes of existing mHealth interventions, not exploratory studies or protocols that have not been implemented yet. Low-, middle-, or high-income status for countries was determined using the World Bank\'s 2014 classification, which is based on estimates of the gross national income per capita for the previous year \[[@B10]\]. In addition, the inclusion of only peer-reviewed articles helped to ensure that higher quality studies were examined. Though there have been well-designed studies using mHealth for behavior change to support maternal and child health in high-income countries, these studies were excluded due to the resource disparities between high-income countries and others. Issues with prevalence of mobile phones and consistency of power and internet access are shared across many low- and middle-income countries and therefore these studies are more comparable than those conducted in high-income countries. No keywords for "low- or middle-income countries" were used in the searches, as these keywords might have excluded relevant results if the study was not specifically labeled as such. Instead, the authors screened manually for this criterion. The review was limited to studies available in English, though this is a limitation of this review, and future reviews should include additional languages, if feasible. Finally, studies only included those that were published after 2000, as mobile technologies were not widely available, especially in low- and middle-income countries, prior to that time. 2.3. Study Selection and Data Collection {#sec2.3} ---------------------------------------- The database searches were undertaken by two researchers (Jessica L. Watterson and Isheeta Madeka) between November 10, 2014 and January 18, 2015. Subsequent review of results was undertaken by one researcher (Jessica L. Watterson). The resulting articles were first screened by title, then by abstract, and finally by full text to progressively eliminate articles not meeting the inclusion criteria. Many systematic reviews of mHealth research were identified in the results (n = 26), so the included articles and reference lists of these reviews were all examined to ensure an exhaustive search. Finally, the references of all included articles were reviewed as well. The results of study screening and selection are illustrated in [Figure 1](#fig1){ref-type="fig"}. The database searches identified 1,899 articles initially. After removing duplicates, 508 records remained. Each of these records was screened by title and abstract (if necessary), and 455 records were excluded after this preliminary review. The full text of the remaining 53 articles was reviewed to determine if they met the inclusion criteria. 43 of the articles were excluded and the reasons for exclusion included study being conducted in a high-income country (n = 5); not studying antenatal care attendance, postnatal care attendance, or childhood immunization rates (n = 7); not studying an mHealth intervention (n = 2); or only providing program descriptions or a protocol but no evaluation data (n = 4). The other 25 articles that were excluded were mHealth literature reviews that did not identify any new articles for review. One article outlined a protocol for a study that will be very relevant once complete; however it was nevertheless excluded because no evaluation data is published yet \[[@B25]\]. 2.4. Quality Assessment {#sec2.4} ----------------------- Risk of bias was assessed for all included randomized controlled trials (RCTs) (n = 2) using the Cochrane Risk of Bias Assessment Tool \[[@B11]\]. This tool was introduced in 2008 by the Cochrane Collaboration and can be used to assess risk of bias in a study by evaluating a study\'s allocation sequence generation (randomization), allocation concealment, blinding, incomplete data, selective reporting, and other potential threats to the study\'s validity. The quality of the observational studies (n = 8) was assessed using the Newcastle-Ottawa Quality Assessment Scale \[[@B12]\]. This tool was developed by the universities of Newcastle, Australia, and Ottawa, Canada, and assessed the quality of nonrandomized studies by evaluating potential sources of bias in the selection and comparability of participants, the assessment of outcomes, and the duration and adequacy of follow-up. Scores are awarded out of 9 possible points, with higher scores indicating higher study quality. 2.5. Synthesis of Results {#sec2.5} ------------------------- The primary author (Jessica L. Watterson) extracted information from included articles for tabulation in an Excel spreadsheet. The information extracted included type of study, summary conclusions, methods used, intervention studied, health issue(s) studied, outcomes measured, sample size, intervention frequency, effectiveness of intervention, study quality, study location, clinical characteristics/setting, mHealth tools used, and project name (if any). 3. Results {#sec3} ========== Most articles examined process and behavior change outcomes and made recommendations for future mHealth programs and suggested further research. The study characteristics and key outcomes for each included article are outlined in [Table 1](#tab1){ref-type="table"}. 3.1. Characteristics of Studies {#sec3.1} ------------------------------- In total, ten articles satisfied the inclusion criteria. Of these, two studies were RCTs \[[@B13], [@B14]\] and the other eight were observational studies \[[@B15]--[@B22]\]. Four of the observational studies attempted to limit sources of bias (though not as rigorously as the RCTs) by using a historic control group \[[@B15], [@B16]\] or nonrandomized control group \[[@B18]\] or measuring outcomes before and after implementation of the mHealth intervention \[[@B17]\]. The remaining four observational studies did not use a control group \[[@B17]--[@B20]\], and as such the outcomes of these studies are less reliable. Seven (70%) of the articles studied antenatal care attendance \[[@B13], [@B14], [@B16]--[@B19], [@B21]\]; two (20%) studied postnatal care attendance \[[@B15], [@B19]\]; and four (40%) studied childhood immunization rates \[[@B17], [@B19], [@B20], [@B22]\]. Eight (80%) of the studies used an mHealth intervention that sent reminders to seek care directly to patients \[[@B13]--[@B17], [@B19], [@B20], [@B22]\] and five (50%) sent educational messages to patients \[[@B13], [@B14], [@B16], [@B18], [@B19]\]. Three (30%) studies sent reminders to health workers to follow up with patients \[[@B17], [@B20], [@B21]\] and three (30%) studies used an mHealth tool to improve patient records or identification \[[@B17], [@B20], [@B21]\]. The frequency of these interventions varied widely; educational messages were sent on schedules ranging from daily \[[@B18]\] to twice per month \[[@B14]\]. Some studies specified that appointment reminders were sent a few days in advance of a scheduled appointment \[[@B15], [@B17], [@B22]\] and others did not specify how far in advance patients or health workers were reminded. Eight (80%) of the studies were conducted in Africa \[[@B13]--[@B15], [@B18]--[@B22]\] and two (20%) were conducted in Asia \[[@B16], [@B17]\]. All included studies were published between 2010 and 2014, suggesting that the inclusion criterion of studies published after 2000 was sufficiently conservative and that it is unlikely that any relevant articles were missed from earlier publication dates. One study is taken from a literature review published in English on mHealth tools in China \[[@B16]\]; however, the original study was published in Chinese. Therefore, the information available on this study is less complete than that provided for the studies where the primary publication was included. 3.2. Findings by Intervention {#sec3.2} ----------------------------- All studies showed some evidence that the mHealth intervention implemented had a positive impact on patient or health worker behavior. However, the quality of the studies varied and some of these outcomes cannot be conclusively attributed to the mHealth intervention that was implemented from these studies alone. ### 3.2.1. Antenatal Care Attendance {#sec3.2.1} Two of the seven studies examining antenatal care attendance were RCTs \[[@B13], [@B14]\]. Both studies used text message reminders and education for pregnant women and one also provided the women with mobile-phone vouchers to contact their health worker, if needed \[[@B14]\]. Both studies found a statistically significant increase of over 10% in the proportion of women receiving at least four antenatal care visits between the intervention and control groups. Another study examined antenatal care attendance before and after implementation of an mHealth application for improved patient records and automated appointment reminders; this study similarly found a statistically significant improvement in on-time antenatal care attendance following implementation \[[@B17]\]. A study conducted in China sent text message reminders for antenatal care and health advice to an intervention group and found a statistically significant increase in antenatal care attendance, compared to a historic control group. The remaining studies examining antenatal care attendance found some self-reported behavior change from both patients and health workers \[[@B18], [@B19], [@B21]\]. ### 3.2.2. Postnatal Care Attendance {#sec3.2.2} One study examining postnatal care attendance used a historic control group from the previous 6 months in the same hospital and found that the intervention group, receiving text message appointment reminders, were 50% less likely to fail to attend their appointment (P = 0.002) \[[@B15]\]. Another study found that women self-reported intended or actual behavior change, including increased attendance to postnatal care, after receiving voice or SMS messages with education and reminders \[[@B19]\]. ### 3.2.3. Childhood Immunization {#sec3.2.3} A study examining childhood immunization found a statistically significant increase (from 34.5% to 44.2%, P \< 0.001) in the proportion of children receiving on-time vaccination after implementation of a mobile application for improved patient records and automated text message appointment reminders \[[@B17]\]. Another study found that mothers reported being influenced by a text message reminder (which were also tied to a conditional cash transfer, if child was vaccinated on time) to bring their child for immunization \[[@B22]\]. In one study, after receiving SMS or voice reminders and education, mothers self-reported intended or actual behavior change, including bringing their child for vaccines \[[@B19]\]. Finally, a study using an mHealth application to improve records and to send reminders during a mass vaccination campaign found that 92% of children visited at home following the campaign had received the measles vaccine \[[@B20]\]. 3.3. Findings on Cost {#sec3.3} --------------------- Two of the studies included information on the cost of their mHealth interventions. Adanikin et al. reported a total cost of only US\$21.12 to send 2252 SMS reminders for postnatal care during the six-month study in Nigeria \[[@B15]\]. Ngabo et al. cited initial investment cost as being considerable, largely due to the fact that they provided all community health workers in Rwanda with a mobile phone to "boost engagement and motivation of CHWs." However, ongoing costs were lowered by Ministry of Health negotiations with the private sector, reducing SMS costs from US\$0.05 to US\$0.005 per message \[[@B21]\]. 4. Discussion {#sec4} ============= Though all included studies showed some evidence that mHealth tools can be effective in changing patient and health worker behavior to increase antenatal care attendance, postnatal care attendance, and childhood immunization rates, the quality of the evidence varied widely. The strongest evidence exists for text message reminders and education delivered to pregnant women\'s mobile phones. The two RCTs that examined this intervention both found evidence of statistically significant increases in antenatal care attendance in their intervention groups, relative to their control groups \[[@B13], [@B14]\]. There is also some suggestion from the results that this intervention may also be effective when applied to the other health issues studied, such as postnatal care attendance and childhood immunization. Though no RCTs studied these health issues, two observational studies of high quality found evidence of effectiveness. Adanikin et al. found that intervention group receiving text message reminders for postnatal care were 50% less likely to fail to attend their appointments than a historic control group from the previous six months (P = 0.002) \[[@B15]\]. Kaewkungwal et al. found that after implementation of a smartphone application that supported record keeping and generated text message reminders for health workers and mothers, there was a 15% increase in women attending ANC on time (P \< 0.001) and a 10% increase in children receiving on-time immunizations (P \< 0.001) \[[@B17]\]. Beyond these findings, much of the evidence is based on self-reported behavior change from health workers and patients, which is not sufficiently reliable to draw any strong conclusions on the effectiveness of mHealth interventions \[[@B19], [@B21], [@B22]\]. Some other observational studies demonstrated good results of their programs, such as 92% confirmed coverage in a measles vaccine campaign \[[@B20]\]; however, it is impossible to determine which factors influenced the campaign\'s success and whether it was due to the use of an mHealth intervention or one of the other program components. In addition, several studies combined multiple mHealth interventions (e.g., text message reminders and conditional cash transfers via mobile phone \[[@B22]\]), making it impossible to determine to what degree each intervention influenced the resulting behavior change. As a result of these methodological limitations and the small number of studies meeting the inclusion criteria, further randomized controlled trials are needed to evaluate the effectiveness of mHealth tools for antenatal care, postnatal care, and childhood immunizations. By employing a multiarm or factorial design, researchers may be able to better ascertain which components of mHealth interventions are most effective. It is also worth noting that many of the mHealth tools studied focused on a single period of time on the maternal, neonatal, and child health (MNCH) continuum. For example, one study focused only on postnatal care, while five others focused only on antenatal care. Given the importance of continued follow-up of families during pregnancy, delivery, postnatal periods, and early childhood, it would be advisable for future mHealth interventions to consider expanding their tools to include more key events along the MNCH continuum \[[@B22]\]. Finally, the majority of these studies were conducted in Africa, suggesting that there is a need for future study of mHealth tools in broader contexts, including Asia and the Pacific, Central and South America, the Caribbean, and other regions. This literature review has provided us with the key knowledge that there is some existing evidence of the effectiveness of text message reminders for antenatal care, postnatal care, and immunizations and it has also helped to identify that this is an area where further research is needed. Given the limited, but largely positive, results of this literature review, researchers and public health practitioners should continue to implement mHealth tools for antenatal care attendance, postnatal care attendance, and childhood immunization. However, careful evaluation and further research are still needed to better determine how effective these tools are and in which settings. 5. Conclusions {#sec5} ============== Based on a systematic review of the literature, there is some evidence that mHealth tools may present an opportunity to influence behavior change and ensure that women and children in low-income countries are accessing prevention services, including antenatal care, postnatal care, and immunizations. Though mHealth programs have been implemented in low- and middle-income countries all over the world \[[@B23]\], there are few peer-reviewed studies and the majority of evaluations relating to maternal and child health have been conducted in Africa. Therefore, greater emphasis needs to be put on the evaluating mHealth tools and disseminating results to inform program design and policy making. In addition, many existing interventions focus on only one component of maternal and child health preventive services, rather than on design of an integrated system that follows women and children through the maternal, neonatal, and child health continuum \[[@B24]\]. The field of mHealth should continue to be supported and studied as it shows promise of improving the lives of women and children in low- and middle-income countries. Conflict of Interests ===================== The authors declare that there is no conflict of interests regarding the publication of this paper. ![PRISMA flow diagram \[[@B26]\].](BMRI2015-153402.001){#fig1} ###### Summary of included articles on mHealth interventions to increase use of antenatal care, postnatal care, and childhood immunization, classified by methods used. -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- First author,\ Title Health issue(s) studied Intervention studied and tools used Intervention frequency Key study outcomes Methods used Sample size Study location Study quality^1^ year ----------------------------------------------------------------- -------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------------------------------------------------- -------------------------------------------------------- ------------------------------------------------------------------------------ Randomized controlled trials (RCTs) Fedha, 2014 \[[@B13]\] "Impact of Mobile Telephone on Maternal Health Service Care: A Case of Njoro Division" Antenatal care attendance Text message reminders and educational messages for mother delivered to mobile phone.\ Appointment reminders every two weeks. Frequency of educational messages not specified 7.4% of women receiving SMS had less than 4 antenatal visits while 18.6% of those not receiving SMS had less than 4 visits (P = 0.002) Clinic attendance and antenatal service uptake compared for intervention and control groups Intervention group: 191\ Health facilities in Kenya RCT with low risk of bias No specific mHealth tools mentioned Control group: 206\ Total: 397 Lund, 2014 \[[@B14]\] "Mobile Phones Improve Antenatal Care Attendance in Zanzibar: A Cluster Randomized Controlled Trial" Antenatal care attendance Text message reminders and educational messages for mother delivered to mobile phone and mobile vouchers to contact health workers.\ Two messages per month before gestational week 36 and two messages per week after week 36 44% of women in the intervention group received the recommended four or more antenatal visits, compared with 31% in the control group. The odds for receiving four or more antenatal care visits were 2.39 (1.03--5.55) for women benefitting from the mobile phone intervention. 59% of intervention women stated that received text messages influenced the number of times they attended antenatal care Clinic attendance was compared for cluster randomized intervention and control groups Intervention group: 1311\ Urban and rural healthcare facilities in Zanzibar RCT with low risk of bias Tools used: custom Wired Mothers software Control group: 1239\ Total: 2550 Studies with nonrandomized control group or before/after design Adanikin, 2014\ "Role of Reminder by Text Message in Enhancing Postnatal Clinic Attendance" Postnatal care attendance Text message reminders for mother delivered to mobile phone.\ Two messages sent for each appointment: two weeks prior and 5 days prior Patients who received an SMS reminder were 50% less likely to fail to attend (FTA) their postnatal appointment (relative risk of FTA 0.50; 95% CI, 0.32--0.77; P = 0.002) Clinic attendance compared for intervention group and historic control group (from previous 6 months) Intervention group: 1126\ Teaching hospital in Nigeria 7/9 \[[@B15]\] No specific mHealth tools mentioned Control group: 971\ Total: 2097 Fang and Li, 2010 \[[@B27]\] from\ "Mobile Health in China: A Review of Research and Programs in Medical Care, Health Education, and Public Health" Antenatal care attendance Text message appointment reminders and antenatal health advice.\ Four appointment reminders per pregnancy. Frequency of health advice not specified The intervention group received 5.7 ± 1.8 antenatal visits, compared to 3.2 ± 1.1 antenatal visits in the control group (P \< 0.01) Clinic attendance compared for intervention group and historic control group (from previous year). Intervention group: 609\ China Unable to determine as not all info. on study design is available in English Corpman, 2013 \[[@B16]\] No specific mHealth tools mentioned Control group: 637\ Total: 1246 Kaewkungwal, 2010 \[[@B17]\] "Application of Smart Phone in "Better Border Healthcare Program": A Module for Mother and Child Care" Antenatal care attendance and childhood immunization (EPI) Smartphone application used by health workers to update antenatal and immunization status when outside clinic and SMS reminders for both health workers and mothers.\ Appointment reminders a few days prior to scheduled appointment 58.68% of pregnant women came to ANC on time after implementation as compared to 43.79% before (P \< 0.001). After adjusting for personal characteristics, sending appointment message increased odds of on-time visit by 2.97 (1.60--5.54). 44.22% of children received scheduled vaccines on time after implementation as compared to 34.49% before (P \< 0.001). After adjusting for personal characteristics, follow-up cases and updating immunization data on cell phones increased odds of on-time EPI by 2.04 (1.66--2.52). Sending appointment reminder increased odds of on-time EPI by 1.48 (1.09--2.03) Clinic attendance for ANC and EPI were compared before and after MCCM implementation ANC group: 280\ Rural border area in Thailand, near Myanmar 8/9 Tools used: custom Mother and Child Care Module (MCCM) EPI group:\ 544 Lau, 2014 \[[@B18]\] "Antenatal Health Promotion via Short Message Service at a Midwife Obstetrics Unit\ Antenatal care attendance Text messages with antenatal health information.\ Varied from three messages per week to daily messages 92% of participants in the intervention group reported not missing more than two antenatal visits. A focus group of intervention participants reported that they had improved health related behaviors, including attending the clinic regularly, as a result of the text messages. No statistically significant difference in knowledge was seen between the intervention and control groups at the exit interview Baseline questionnaire and exit interview were administered to convenience-sampled intervention and control groups to assess knowledge of antenatal health and clinic procedures. A focus group was conducted with a further convenience sample of the intervention group Intervention group: 102 but 45 were lost to follow-up\ Urban primary care facility in Cape Town, South Africa 4/9 in South Africa: A Mixed Methods Study" No specific mHealth tools mentioned Control group: 104 but 43 were lost to follow-up\ Total: 206 recruited, 118 included in analysis Studies with no control group Crawford, 2014\ "SMS versus Voice Messaging to Deliver MNCH Communication in Rural Malawi: Assessment of Delivery Success and User Experience" Antenatal care attendance, postnatal care attendance, and childhood immunization Text (SMS) or voice message reminders and educational messages for mother delivered to mobile phone or retrieved by calling a toll-free hotline.\ Once (voice) or twice (SMS) per week 91% of SMS enrollees surveyed reported that they had already changed or intended to change their behavior based on the messages, including attending more ANC/PNC or bringing their child for vaccines. SMS enrollees were significantly more likely to report intended or actual behavior change than voice enrollees Phone based surveys of participants. Participants in the pushed SMS and pushed voice groups were randomly sampled but participants in the retrieved voice group were convenience sampled Pushed SMS: 96\ Rural health centers in Malawi 2/9 \[[@B19]\] Tools used: Village-Reach custom application (SMS) and INTELLIVR software (voice messages) Pushed voice: 30\ Retrieved voice: 140\ Total: 266 Mbabazi, 2014\ "Innovations in Communication Technologies for Measles Supplemental Immunization Activities: Lessons from Kenya Measles Vaccination Campaign, November 2012" Childhood immunization Smartphone application used by volunteers to update immunization records when canvassing door-to-door and to provide text message and phone call reminders to caretakers.\ Varied/as needed In precampaign house-to-house visits, 25% of households had no plans to bring their children for the measles supplemental dose if they had not been contacted by the volunteers. Of the children found in the postcampaign house visits, 96% reported to have received a measles supplemental immunization dose, although only 92% had confirmation (finger mark) of vaccination Precampaign household canvassing and data collection for entire target population, followed by postcampaign verification of vaccine coverage Precampaign: 164,643 households with 161,695 children\ Urban areas in Kenya 5/9 \[[@B20]\] Tools used: EpiSurveyor Postcampaign: 17,627 households with 17,993 children Ngabo, 2012 \[[@B21]\] "Designing and Implementing an Innovative SMS-based Alert System (RapidSMS-MCH) to Monitor Pregnancy and Reduce Maternal and Child Deaths in Rwanda" Antenatal care attendance Electronic registration of pregnant women through text messages by community health workers (CHWs) and reminder text messages for antenatal care sent to CHWs\' mobile phones.\ As needed for upcoming antenatal visits and estimated delivery date 81% of the estimated annual pregnancies in the district were registered in the system. Reporting compliance among CHWs was 100%. CHWs reported being more proactive in finding new pregnant women and following up registered pregnant women as a result of reminders forwarded to their mobile phones Reporting compliance, system usage patterns, and error rates were monitored and feedback sessions were held with CHWs CHWs: 432 Rural district of Rwanda N/A, only process outcomes were studied Tools used: customized version of RapidSMS Wakadha, 2013\ "The Feasibility of Using Mobile-Phone Based SMS Reminders and Conditional Cash Transfers to Improve Timely Immunization in Rural Kenya" Childhood immunization Text message reminders for mother delivered to mobile phone and free airtime or mobile cash transfers for mothers that brought child in on time.\ Three days before vaccine due date and on due date 91% of mothers reported that the SMS reminders influenced their decision to come in for vaccination Enrolled mothers were randomized to receive either mMoney or airtime for on-time vaccinations.\ mMoney group: 48\ Rural district of Kenya 4/9 \[[@B22]\] Tools used: customized version of RapidSMS and mPESA Questionnaires were administered in home follow-up visits Airtime group: 24\ Total: 72 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ^1^Quality score assigned using the Cochrane Risk of Bias Assessment Tool (for RCTs) or the Newcastle-Ottawa Quality Assessment Scale (for observational studies). For RCTs, a low risk of bias is the best possible score and for observational studies the highest possible score is 9. Please see [Section 2](#sec2){ref-type="sec"} for more details. [^1]: Academic Editor: Pascale Allotey	{ "pile_set_name": "PubMed Central" }
1. Introduction {#sec1-genes-11-00418} =============== Lilies (Lilium spp.) have high commercial and ornamental value due to their various floral colors and coloration patterns. Anthocyanins are one of the major pigments affecting the coloration of the floral tepals in Lilium spp. \[[@B1-genes-11-00418],[@B2-genes-11-00418],[@B3-genes-11-00418],[@B4-genes-11-00418]\]. Anthocyanins are synthesized in the cytosol but their transfer to the vacuole is necessary for plant tissues to exhibit brilliant colors \[[@B5-genes-11-00418],[@B6-genes-11-00418]\]. The mechanism of anthocyanin transfer is still unclear, although anthocyanin biosynthesis at the molecular level has been studied extensively leading to the discovery of enzyme-coding structural and regulatory genes related to anthocyanin biosynthesis \[[@B7-genes-11-00418],[@B8-genes-11-00418]\]. To date, glutathione S-transferases (GSTs), multidrug and toxic compound extrusion (MATE) proteins, and the ATP-binding cassette (ABC) transporter have been proposed to be involved in anthocyanin transportation in many species \[[@B9-genes-11-00418],[@B10-genes-11-00418],[@B11-genes-11-00418]\]. MATE transporters perform various transport functions in plants, including the transportation of anthocyanin \[[@B12-genes-11-00418]\]. A MATE transporter related to flavonoid accumulation was isolated for the first time during screening of a TT12 Arabidopsis mutant with altered seed coloration \[[@B13-genes-11-00418]\]. Subsequently, orthologs of AtTT12 were obtained from many other species, such as upland cotton, blueberry, Medicago truncatula, apple, and grapevine, and orthologs of the TT12 gene were verified to be related to flavonoid transportation \[[@B6-genes-11-00418],[@B14-genes-11-00418],[@B15-genes-11-00418],[@B16-genes-11-00418],[@B17-genes-11-00418]\]. To date, the enzyme-coding structural genes and regulatory genes of anthocyanin biosynthesis in Lilium have been characterized thoroughly \[[@B2-genes-11-00418],[@B4-genes-11-00418],[@B18-genes-11-00418],[@B19-genes-11-00418],[@B20-genes-11-00418]\]; however, whether MATE transport is involved in anthocyanin transportation in Lilium remains unclear. In this study, based on our RNA-seq data published previously \[[@B21-genes-11-00418]\], we isolated a flavonoid transport-related MATE gene, LhDTX35, from Lilium 'Tiny Padhye'. Then, expression profiling, identification, and functional analysis of this gene in anthocyanin transportation were performed using 'Tiny Padhye' as a subject. This study demonstrates that a gene encoding MATE is critical for anthocyanin accumulation in Lilium, and promotes an understanding of the mechanism underlying MATE involvement in anthocyanin transportation in Lilium. 2. Materials and Methods {#sec2-genes-11-00418} ======================== 2.1. Plant Material {#sec2dot1-genes-11-00418} ------------------- The Asiatic lily cultivar 'Tiny Padhye' was grown in a greenhouse (25 °C, relative humidity 60%--70%, 16 h/8 h light/dark cycles) at the Chinese Academy of Agricultural Sciences (Beijing, China). The upper parts and bases of the inner tepals were collected at four different developmental stages. The tepal developmental stages were defined as described in Xu LF et al. \[[@B21-genes-11-00418]\]: Stage 1 (S1: no anthocyanin pigment is visible in tepals), stage 2 (S2: anthocyanin pigment becomes visible on tepals), stage 3 (S3: the day before anthesis, lower halves of tepals are fully pigmented), and stage 4 (S4: the first day after anthesis) ([Figure 1](#genes-11-00418-f001){ref-type="fig"}). Samples were obtained from 15 flowers and pooled together as one biological sample; three independent biological replicates were collected for each stage. Nicotiana tabacum was also grown in a greenhouse, as described above for environmental conditions. Seedlings at the four-true-leaf stage (6-week-old) were used for the subcellular localization experiment. 2.2. Isolation of a MATE-Like Gene {#sec2dot2-genes-11-00418} ---------------------------------- Six primers (LhMATE-YZ-F1, LhMATE-YZ-R1, LhMATE-3'RACE-GSP, LhMATE-3'RACE-nest, LhMATE-5'RACE, and LhMATE-5'RACE-nest) were designed based on unigenes encoding MATE-like proteins from our previous RNA-seq experiment ([Table 1](#genes-11-00418-t001){ref-type="table"}). First-strand cDNA was cloned using RNA isolated from the colored parts of the tepal at stage 2 as a template, according to the RACE kit manufacturer's instructions (Takara, Dalian China). To obtain the full-length MATE cDNA, the 5'-end and 3'-end sequences of the identified gene were identified using stage 2 total RNA from the basal (colored) tepal regions and a RACE kit, according to the recommendations of the manufacturer (Takara, Dalian, China). Five independent RACE cDNA clones were sequenced by Sangon Biotech (Shanghai, China). Based on the 5'-RACE and 3'-RACE sequencing results, the primers LhDTX35-F and LhDTX35-R ([Table 1](#genes-11-00418-t001){ref-type="table"}) were designed to amplify the entire gene region containing the open reading frame (ORF) from the first-strand cDNA. 2.3. Sequence Bioinformatics Analysis {#sec2dot3-genes-11-00418} ------------------------------------- A BLASTx search of MATE-like gene ORF sequences was conducted to analyze sequence similarity. Protein conserved domain prediction was performed by NCBI conserved domain online software \[[@B22-genes-11-00418]\]. The protein transmembrane domains were analyzed by TMHMM Server V.2.0 \[[@B23-genes-11-00418]\]. 2.4. Phylogenetic Analysis {#sec2dot4-genes-11-00418} -------------------------- The amino acid sequences of genes encoding MATEs from different plants were used for phylogenetic analysis. Sequence alignment was performed using CLUSTAL X. Based on this alignment, a Neighbor-Joining (NJ) tree was constructed using MEGA (version 7.0). Bootstrap values were calculated using 1000 replicate analyses. 2.5. Gene Expression Analysis {#sec2dot5-genes-11-00418} ----------------------------- Quantitative real-time polymerase chain reaction (qRT-PCR) was performed using SYBR Premix Ex Taq (Takara, Dalian, China). The primers used to amplify the MATE-like gene segment are shown in [Table 1](#genes-11-00418-t001){ref-type="table"}. The reaction parameters were as follows: (1) 95 °C for 1 min; (2) 40 cycles of 95 °C for 20 s, 60 °C for 10 s, and 72 °C for 25 s; and (3) a melt curve program (65 °C to 95 °C with an increment in temperature of 0.5 °C every 0.05 s). The signal was monitored using a CFX96 real-time system (Bio-Rad, CA, USA). The average Cq value was calculated from three biological and three technical replicates. To normalize the differences in the amounts of mRNA from other genes, the amount of Lhactin mRNA was determined for each sample and the relative expression level of LhDTX35 was analyzed using the 2^−△△CT^ method \[[@B24-genes-11-00418]\]. The error bars represented the ±SEs from three independent experiments. The data were analyzed by ANOVA using SAS software. 2.6. Subcellular Localization Analysis {#sec2dot6-genes-11-00418} -------------------------------------- Using the transient expression vector pCAMBIA2300-35s-GFP, the subcellular localization of the LhDTX35 gene was analyzed. To construct the expression vector 35S::LhDTX35-GFP, primers containing the XbaI and BamHI sites were designed to amplify the LhDTX35 gene ORF domain using the primer LhDTX35-DW-F/R ([Table 1](#genes-11-00418-t001){ref-type="table"}). The PCR products were separated on a 1% agarose gel and then purified. The empty vector pCAMBIA2300-35s-GFP was digested with the restriction enzymes XbaI and BamHI according to the manufacturer's instructions (Thermo Scientific FastDigest, America). Subsequently, the purified PCR product and cleaved empty vector were fused using T4 DNA Ligase (TransGen, China). Positive clones were selected and confirmed by sequencing and then transformed into Escherichia coli DH5α cells (TransGen, China). The sequencing primers (DW-YZ-F/R) are shown in [Table 1](#genes-11-00418-t001){ref-type="table"}. Finally, the plasmids pCAMBIA2300-GFP and pCAMBIA2300-GFP-LhDTX35 were each transformed into Agrobacterium tumefaciens (LBA4404) using the freeze-thaw method, as previously reported \[[@B25-genes-11-00418]\]. When N. tabacum had developed four true leaves, the underside of the leaves was infiltrated with Agrobacterium inocula using a 1 mL needleless syringe. The inoculated plants were grown in a phytotron for 24 h in the dark (25 °C, relative humidity 60%--70%). Then, the plants were grown under 16 h/8 h light/dark cycles for 5 days. Subsequently, the LhDTX35-GFP fusion protein in the N. tabacum epidermis cells was microscopically detected using a Leica confocal laser scanning microscope. The samples were illuminated with an argon ion laser using 488 nm light for GFP and a green HeNe laser using 543 nm light for chlorophyll autofluorescence. 2.7. Complementation Analysis {#sec2dot7-genes-11-00418} ----------------------------- The 35S::LhDTX35 vector was constructed by replacing the GUS gene in the pCAMBIA3301 vector with the coding sequence of LhDTX35. The LhDTX35 ORF was amplified by PCR with primers LhDTX35-noci-F and LhDTX35-Bahm-R ([Table 1](#genes-11-00418-t001){ref-type="table"}). The LhDTX35 ORF fragments were cloned into the linearized vector pCAMBIA3301 and digested with the NocII and HindII restriction enzymes using a one-step seamless cloning kit (Transgen, Beijing, China). The binary vectors were introduced into A. tumefaciens GV3101 and subsequently used in Arabidopsis ttDTX35 mutant transformation with the floral dip method \[[@B26-genes-11-00418]\]. Seeds of the Arabidopsis LhDTX35 mutant, transgenic lines, and wild-type Arabidopsis were germinated and grown on Murashige & Skoog Basic Medium 1/2 Macro (1/2 MS) medium. Transgenic plants were screened on 1/2 MS medium plates that contained 50 mg/L kanamycin. The expression of LhDTX35 was analyzed by RT-PCR with primers LhDTX35-YZ-F/R, and Arabidopsis Actin gene was control with primers Actin-F/R ([Table 1](#genes-11-00418-t001){ref-type="table"}). 3. Results {#sec3-genes-11-00418} ========== 3.1. Cloning and Sequence Analysis of the LhDTX35 Gene {#sec3dot1-genes-11-00418} ------------------------------------------------------ The full-length cDNA sequence of LhMATE-like was obtained by RACE and deposited in GenBank with the accession number MT001433. This sequence contained 1948 bp and encoded 507 amino acids, as predicted by the NCBI ORF Finder tool ([Figure 2](#genes-11-00418-f002){ref-type="fig"}). BLASTx results showed high homology to the DTX35 genes of other species, such as Dendrobium catenatum, Elaeis guineensis, Phoenix dactylifera, and Oryza sativa Japonica Group (ca. 71%--76.36% identity at the amino acid level). Therefore, the LhMATE-like cDNA isolated in this study was named LhDTX35. Conserved domain prediction revealed that LhDTX35 encoded functional domains typical of the MATE-like superfamily between 258 bp and 1562 bp ([Figure 3](#genes-11-00418-f003){ref-type="fig"}). Transmembrane domain analysis predicted that the protein encoded by LhDTX35 was a membrane protein with twelve typical transmembrane domains ([Figure 4](#genes-11-00418-f004){ref-type="fig"}). A phylogenetic analysis revealed that LhDTX35 was closely related to the DTX35 protein in Dendrobium catenatum, sharing 76.36% amino acid identity ([Figure 5](#genes-11-00418-f005){ref-type="fig"}). In conclusion, the isolated LhDTX35 gene was predicted to encode a membrane protein belonging to the MATE-like superfamily. 3.2. Expression Analysis of LhDTX35 {#sec3dot2-genes-11-00418} ----------------------------------- Anthocyanin accumulated in only the basal tepal regions of 'Tiny Padhye'. We previously showed that anthocyanin content increased gradually during tepal development ([Figure 6](#genes-11-00418-f006){ref-type="fig"}) \[[@B21-genes-11-00418]\]. Anthocyanin accumulation was the highest in tepals at stage 3 ([Figure 6](#genes-11-00418-f006){ref-type="fig"}) \[[@B21-genes-11-00418]\]. To exploit whether the LhDTX35 gene was involved in anthocyanin accumulation in Lilium, qRT-PCR was performed using the upper and basal tepal regions at stages 1--4. These results showed that LhDTX35 gene expression was positively correlated with anthocyanin accumulation, which occurred in the bases of the tepals ([Figure 7](#genes-11-00418-f007){ref-type="fig"}). 3.3. Subcellular Localization of LhDTX35 {#sec3dot3-genes-11-00418} ---------------------------------------- A subcellular localization experiment was performed to further identify whether the protein encoded by the LhDTX35 gene was a membrane protein. The results showed that GFP fluorescence was distributed in the cytoplasm of N. tabacum leaf epidermal cells treated with the empty GFP vector ([Figure 8](#genes-11-00418-f008){ref-type="fig"}a,b). In contrast, GFP fluorescence was located in the membrane of N. tabacum leaf epidermis cells treated with the LhDTX35 and GFP fusion protein vector ([Figure 8](#genes-11-00418-f008){ref-type="fig"}c,d). These results indicated that the LhDTX35 gene encoded a membrane protein. 3.4. Functional Analysis of the LhDTX35 Gene in the Arabidopsis DTX35 Gene Mutant {#sec3dot4-genes-11-00418} --------------------------------------------------------------------------------- An Arabidopsis genetic mutant lacking DTX35 was selected to investigate the functionality of the LhDTX35 gene. The coding sequence of the LhDTX35 gene was transferred into the Arabidopsis DTX35 mutant under the control of the cauliflower mosaic virus (CaMV) 35S promoter. The hypocotyls of wild-type plants and transgenic lines expressing LhDTX35 were red. In contrast, the hypocotyls of the Arabidopsis DTX35 mutant were green, except for the junctions between the cotyledons ([Figure 9](#genes-11-00418-f009){ref-type="fig"}B). The silique fertility of the Arabidopsis DTX35 mutant was low but normal fertility was restored in the complementation experiment ([Figure 9](#genes-11-00418-f009){ref-type="fig"}C). These results showed that the function of the LhDTX35 gene was similar to that of the DTX35 gene in Arabidopsis. Furthermore, the LhDTX35 gene may be involved in anthocyanin accumulation. 4. Discussion {#sec4-genes-11-00418} ============= Lilium has high ornamental value due to its wide variety of coloration patterns. Anthocyanins are important pigments responsible for Lilium flower pigmentation ranging from pink to purple. Therefore, the characterization of anthocyanins in Lilium is of considerable ongoing interest. In Lilium, the steps of anthocyanin biosynthesis and regulation are well established; however, the anthocyanin transportation mechanism remains unclear. In recent years, several transport proteins have been reported to be involved in anthocyanin transportation \[[@B9-genes-11-00418],[@B10-genes-11-00418],[@B11-genes-11-00418]\]. Among them, MATE is an important protein that transports anthocyanins. To date, whether the MATE protein is involved in anthocyanin transportation in Lilium is unknown. In the present study, a flavonoid transport-related MATE candidate gene, LhDTX35, was cloned successfully based on our previous transcriptome data. Phylogenetic analysis revealed that the LhDTX35 gene clustered with DTX35 genes from Dendrobium catenatum, Phoenix dactylifera, and Oryza brachyantha ([Figure 5](#genes-11-00418-f005){ref-type="fig"}). LhDTX35 is closely related to the DTX35 gene in Dendrobium catenatum, sharing 76.36% amino acid identity. Bioinformatics analysis revealed that the LhDTX35 gene contained sequences encoding the conserved domains typical of the MATE superfamily ([Figure 3](#genes-11-00418-f003){ref-type="fig"}), suggesting that the protein encoded by LhDTX35 belonged to the MATE family. In addition, MATEs are important transport proteins that possess 12 typical transmembrane domains, and transmembrane structure prediction showed that LhDTX35 also has 12 typical transmembrane domains ([Figure 4](#genes-11-00418-f004){ref-type="fig"}). These results indicated that LhDTX35 is a MATE protein in Lilium. The expression profiles of anthocyanin biosynthetic genes and regulatory genes during anthocyanin accumulation at stages 1--4 were previously analyzed in 'Tiny Padhye', and the results showed that anthocyanin accumulation in tepals increased gradually at stages 1--4, reaching its highest level at stage 3 \[[@B21-genes-11-00418]\]. In this study, we revealed that the expression level of LhDTX35 was consistent with anthocyanin accumulation at stages 1--4 ([Figure 7](#genes-11-00418-f007){ref-type="fig"}). Weak expression of LhDTX35 was detected in the tepals at stage 1, when the anthocyanin content was lowest. In conjunction with the increasing anthocyanin content in the tepals, the transcript levels of LhDTX35 were gradually upregulated at stages 1--4, and both anthocyanin content and LhDTX35 gene transcript levels peaked at stage 3 ([Figure 6](#genes-11-00418-f006){ref-type="fig"} and [Figure 7](#genes-11-00418-f007){ref-type="fig"}). All the results indicated that LhDTX35 expression was positively correlated with anthocyanin accumulation in Lilium, suggesting that the LhDTX35 gene may be involved in the transport of anthocyanin in Lilium. All the plant MATE proteins characterized to date have been localized to membranes \[[@B27-genes-11-00418]\]. To investigate whether LhDTX35 is involved in anthocyanin accumulation, its subcellular protein localization was examined. The results showed that LhDTX35 was localized to the membrane but not the cytoplasm, further suggesting that the LhDTX35 gene encoded a protein with the traits of MATE transport proteins. These results were also consistent with the results of the transmembrane structure prediction, that is, the protein encoded by LhDTX35 was predicted to have 12 typical transmembrane domains ([Figure 4](#genes-11-00418-f004){ref-type="fig"}). Complementation experiment is an effective method to analyze gene function because genetic transformation is a challenge for Lilium. Complementation has been used for this purpose in many species \[[@B9-genes-11-00418],[@B11-genes-11-00418],[@B15-genes-11-00418]\]. In this study, the silique-infertility phenotype of the DTX35 mutant was recovered when LhDTX35 was overexpressed in the DTX35 mutant of Arabidopsis, suggesting that the LhDTX35 gene has a similar function to that of DTX35 in Arabidopsis. The DTX35 gene is also called FFT (flower flavonoid transporter) \[[@B28-genes-11-00418]\]. The anthocyanin content of the FFT mutant was slightly less than that of the WT (Wile Type) in seedlings at one week of age \[[@B28-genes-11-00418]\]. Interestingly, the present study showed that the anthocyaninless phenotype of the DTX35 mutant was rescued by the complementation experiment ([Figure 9](#genes-11-00418-f009){ref-type="fig"}B). These results suggested that the LhDTX35 gene was required for anthocyanin transportation in Lilium. MATE genes are part of a gene superfamily, including 58 MATEs in Arabidopsis and 53 MATEs in rice \[[@B29-genes-11-00418]\]. In the present study, only one MATE gene was cloned successfully, and whether any other MATE genes are involved in anthocyanin transportation in Lilium should be further studied. 5. Conclusions {#sec5-genes-11-00418} ============== In conclusion, we first identified a flavonoid transport-related MATE candidate gene based on transcriptome data in Lilium 'Tiny Padhye' and found that this gene shared high similarity with published DTX35 genes. Bioinformatics and phylogenetic analysis revealed that LhDTX35 belongs to the MATE gene family. Expression profiling showed that the expression of LhDTX35 was positively correlated with anthocyanin accumulation in the tepals of 'Tiny Padhye'. Subcellular localization results showed that the LhDTX35-encoded protein was localized to the membrane, suggesting that the protein encoded by the LhDTX35 gene possessed the traits of a MATE transporter. Silique-infertility and anthocyaninless hypocotyl phenotypes were recovered after the complementation experiment, indicating that LhDTX35 has a similar function to that of the DTX35 gene in Arabidopsis and is involved in anthocyanin transportation in Lilium. Our results provide insight into the mechanism of anthocyanin transportation in Lilium. We are grateful to Haiwen Zhang (Beijing Academy of Agricultural and Forestry sciences) for kindly offering Arabidopsis DTX35 mutant. Conceptualization, J.M. and L.X.; methodology, H.X.; investigation, P.Y., Y.T., Y.C., and G.H.; writing---original draft preparation, H.X.; writing---review and editing, L.X. and J.M.; funding acquisition, J.M. and H.X. All authors have read and agreed to the published version of the manuscript. This research was funded by the National Natural Science Foundation of China, grant number 31672196, 31902048, and this research was also funded by the Fundamental Research Funds for Central Non-profit Scientific Institutions and the Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences. The authors declare no conflict of interest. ![Tepals of the Asiatic lily cultivar 'Tiny Padhye' at four different developmental stages.](genes-11-00418-g001){#genes-11-00418-f001} ![Nucleotide and deduced amino acid sequences of the LhDTX35 gene. The initiation codon is underlined, while the stop codon is labeled with an asterisk and box.](genes-11-00418-g002){#genes-11-00418-f002} ![Conserved domain prediction for the protein encoded by LhDTX35.](genes-11-00418-g003){#genes-11-00418-f003} ![Transmembrane prediction for the protein encoded by LhDTX35.](genes-11-00418-g004){#genes-11-00418-f004} ![Phylogenetic tree of the predicted amino acid sequences of the Lilium multidrug and toxic compound extrusion (MATE) protein and MATE proteins from other plants. The phylogenetic tree was constructed using MEGA 6.0 software and the NJ method. The red dot indicates the Lilium MATE-like protein identified in this study.](genes-11-00418-g005){#genes-11-00418-f005} ![Changes in anthocyanin content at different developmental stages in Lilium \[[@B21-genes-11-00418]\].](genes-11-00418-g006){#genes-11-00418-f006} ![Spatiotemporal analysis of LhDTX35 gene expression at different stages of tepal development in Lilium 'Tiny Padhye'. (a) Expression profiles of LhDTX35 at stages 1--4 in the basal tepal region in Lilium 'Tiny Padhye'; (b) tepal at stage 2; (c) expression profiles of LhDTX35 at stage 2 between the upper and basal tepal regions in Lilium 'Tiny Padhye'.](genes-11-00418-g007){#genes-11-00418-f007} ![Subcellular localization of the LhDTX35-GFP fusion protein. (a,c) Bright field images; (b,d) GFP fluorescence images; (a,b) cells expressing the LhDTX35-GFP fusion protein; (c,d) cells expressing 35S::GFP as a control.](genes-11-00418-g008){#genes-11-00418-f008} ![Functional complementation of the Arabidopsis DTX35 mutant with the LhDTX35 gene driven by the CaMV 35S promoter. (A) Expression of LhDTX35 in wild-type (a), mutant (b), and transgenic Arabidopsis seedlings (c--e). (B) Anthocyanin accumulation in wild-type (a), mutant (b), and transgenic Arabidopsis seedlings (c--e). (C) Silique fertility of wild-type (a), mutant (b), and transgenic Arabidopsis seedlings (c--e).](genes-11-00418-g009){#genes-11-00418-f009} genes-11-00418-t001_Table 1 ###### All primers used in this study. ---------------------------------------------------------------------- Primers Primer Sequence (5′-3′) --------------------- ------------------------------------------------ LhMATE-YZ-F1\ GTTGCGGTTATCACATCCCT\ LhMATE-YZ-R1 CTTCTTCCCCTCACAGTCTA LhDTX35-3'RACE-GSP\ CGGTGGGTGGCAAGGTCTGGTAGC\ LhDTX35-3'RACE-nest GGGCTATCCGTTACATTTGGGTGTGC LhDTX35-5'RACE,\ CCAAATCCCCTGCACACCCAAA\ LhDTX35-5'RACE-nest TGGTTCATTCTGTCCACCCCAC LhDTX35-F,\ AGTGGGAGAGAGAGAGAGGCGA\ LhDTX35-R CTTCTTCCCCTCACAGTCTATGC LhDTX35-YG-F\ CCAAGAGGATGACATTGCCGAG\ LhDTX35-YG-R TGGAGGATGAGGGTGAAGAAGC LhDTX35-DW-F\ CCCCCGGGATGGAAGATCCGCTTCTGAGAC\ LhDTX35-DW-R GCTCTAGACACTAACTTGACTTTGTTGGTT DW-YZ-F\ CGGGCTGTTGCGAAAATA\ DW-YZ-R TGCCGTTCTTCTGCTTGTC LhACT-YG-F\ GCATCACACCTTCTACAACG\ LhACT-YG-R GAAGAGCATAACCCTCATAGA Actin-F\ CGTGACCTTACTGATTACCT\ Actin-R AGCGATACCTGAGAACATAG LhDTX35-noci-F\ CGGGGGACTCTTGACCATGGAAGATCCGCTTCTGAGACAT\ LhDTX35-Bahm-R GGAAATTCGAGCTGGTCACCTGTAATCTAACTTGACTTTGTTGGTT LhDTX35-YZ-F\ CGGGCTGTTGCGAAAATA\ LhDTX35-YZ-R TGCCGTTCTTCTGCTTGTC ----------------------------------------------------------------------	{ "pile_set_name": "PubMed Central" }
1. Introduction =============== The use of artificial materials in medicine and biology has become very important, especially their application as substitutes for damaged tissues and organs, or as carriers for drug or gene delivery \[[@B1-materials-06-01632],[@B2-materials-06-01632],[@B3-materials-06-01632],[@B4-materials-06-01632]\]. These materials have to be biocompatible, i.e., they have to match the mechanical properties of the replaced tissue, and not act as cytotoxic, mutagenic or immunogenic. For specific applications, biocompatible materials can behave as bioinert and not as promoting cell adhesion and growth. For example, materials of these types have been applied in the construction of artificial eye lenses \[[@B5-materials-06-01632]\] or in the articular surfaces of joint prostheses \[[@B6-materials-06-01632]\], which are implants requiring transparency or smoothness, and are thus completely cell-free surfaces. Bioinert materials have also been used for fabricating polymeric vascular prostheses in order to prevent the adhesion and activation of thrombocytes and immunocompetent cells on the inner surface of these grafts \[[@B7-materials-06-01632]\]. However, adhesion, spreading, growth and phenotypic maturation of cells are required for constructing advanced bioartificial tissue replacements, including replacements for blood vessels or for bone. In these replacements, the artificial materials should act as analogs of the natural extracellular matrix, and thus they should support regeneration of the damaged tissue. For example, in vascular tissue engineering, this means that the material should enable reconstruction of the tunica intima, formed by a confluent layer of endothelial cells, and also reconstruction of the tunica media, which contain vascular smooth muscle cells. For these purposes, advanced artificial materials should not just be passively tolerated by the cells, but should act as bioactive or biomimetic, which means that they induce the required cell responses in a controllable manner (for a review, see \[[@B8-materials-06-01632],[@B9-materials-06-01632]\]). It is generally known that the cell--material interaction is strongly dependent on the physical and chemical properties of the material surface, such as wettability, roughness and topography, or the presence of various chemical functional groups and biomolecules. From this point of view, the surfaces of most materials currently used for constructing tissue replacements are not optimal for integrating with the surrounding tissues, and need further modification in order to improve their physicochemical surface properties and thus enhance the cell colonization and new tissue formation (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632]\]). A typical example is provided by synthetic polymers (polyethylene, polypropylene, polystyrene, polyethyleneterephthalate or polytetrafluoroethylene), which are biomaterials widely used in various biotechnologies, including cell cultivation and construction of tissue replacements. In their non-treated state, however, some materials usually form inadequate scaffolds for cell colonization, often due to their relatively high hydrophobicity (water drop contact angle in the range of about 90--120°). These materials can be rendered more hydrophilic by several techniques, particularly by irradiation with ultraviolet (UV) light \[[@B10-materials-06-01632]\], an ion beam \[[@B11-materials-06-01632],[@B12-materials-06-01632],[@B13-materials-06-01632]\] or exposure to plasma discharge \[[@B14-materials-06-01632],[@B15-materials-06-01632],[@B16-materials-06-01632]\]. In addition to the changes in surface wettability, these treatments also affect the roughness, morphology, electrical conductivity, stability, mechanical properties and chemical composition of the material surface. A common feature of irradiation with UV light, ion beam or plasma treatment is splitting of the bonds within the polymer molecules, which results in the creation of free radicals, double bonds and new functional groups (especially oxygen-containing groups) on the polymer surface. Oxidized groups increase the wettability of polymers, and this supports the adsorption of cell adhesion-mediating extracellular matrix (ECM) molecules in an appropriate spatial conformation, increasing the accessibility of specific sites in these molecules for cell adhesion receptors. Free radicals and unsaturated bonds can then be used for functionalizing the material surface with various biomolecules, e.g., grafting amino acids, oligopeptides or protein molecules, which can also influence (mediate or attenuate) the cell adhesion and growth (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632]\]). This grafting occurs spontaneously after exposure of the irradiated material to biological environments, including biological fluids such as blood, intercellular liquid or cell culture media. Polyethylene has been used in several tissue engineering applications. For example, composites of HDPE with hydroxyapatite \[[@B17-materials-06-01632]\] or tricalcium phosphate \[[@B18-materials-06-01632]\] have been developed for the construction of bone replacements, composites of HDPE with graphite for the construction of joint replacements \[[@B19-materials-06-01632]\], and copolymers of HDPE with hyaluronan for the reconstruction of osteochondral defects \[[@B20-materials-06-01632]\]. HDPE was also clinically applied for calvarial reconstruction \[[@B21-materials-06-01632]\]. LDPE is used for producing intravascular catheters for arteriography and angioplasty, due to its advantageous mechanical properties, such as elasticity and flexibility \[[@B22-materials-06-01632]\]. In our study, two types of polyethylene, namely high density polyethylene (HDPE) and low density polyethylene (LDPE), were modified by Ar^+^ plasma discharge of various exposure times (50, 100 and 300 s), and were subsequently exposed to solutions of two important components of fetal serum, namely fibronectin (Fn) and bovine serum albumin (BSA). We expected the biomolecules from these fluids to graft spontaneously to the plasma-activated polyethylene and to influence cell adhesion and growth, which were studied using rat vascular smooth muscle cells in cultures on these materials. Polyethylene was chosen as a model material for studies of cell--material interaction, due to its easy availability and particularly its relatively simple molecule, consisting only of carbon and hydrogen, which enables clear and reproducible results to be obtained. A comparison was also made for the cell behavior on HDPE and LDPE. 2. Results and Discussion ========================= 2.1. Physical and Chemical Properties of the Polyethylene Samples ----------------------------------------------------------------- ### 2.1.1. Surface Wettability The water drop contact angle decreased after plasma modification of the polymers, which means that the surface wettability increased ([Figure 1](#materials-06-01632-f001){ref-type="fig"}). The highest decrease in the contact angle was observed on samples irradiated for 50 s, and as the exposure time was prolonged the decrease was less apparent. On HDPE irradiated for 300 s, the contact angle even increased ([Figure 1](#materials-06-01632-f001){ref-type="fig"}A). A similar observation on HDPE was published earlier, and can be explained by the different rearrangement of HDPE after plasma irradiation in comparison with LDPE \[[@B23-materials-06-01632]\]. It is generally known that the increase in wettability is due to the formation of new oxidized groups on the polymer surface after plasma modification \[[@B4-materials-06-01632],[@B24-materials-06-01632]\]. However, surface wettability is also associated with rearrangement of the surface structure, mainly of the oxidized groups, during the aging period of the polymer after plasma exposure. During this period, the oxidized groups are reoriented from the surface to the inside of the polymer. This causes an increase in the contact angle, though in most cases the contact angle of a modified polymer after the aging period still remains lower than the contact angle of the non-treated polymer. The rearrangement of the oxidized groups during aging is influenced by the degree of crosslinking of the polymer structure and the time of plasma modification. A less crosslinked structure and longer modification times lead to greater mobility of these groups. The structure of HDPE is less crosslinked than the structure of LDPE, probably due to the fact that HDPE molecules are more linear and less branched than LDPE molecules, and these differences are further enhanced by plasma irradiation. Thus, the less crosslinked structure of HDPE and relatively long plasma irradiation (300 s) led to higher rotation of the oxidized groups into the polymer, their burrowing into the polymer, their lower exposure on the polymer surface, and thus to lower wettability of the polymer, manifested by a relatively high contact angle (119°), which became even higher than on non-treated HDPE. Moreover, the lesser branching of HDPE creates less opportunity for oxidized groups to occur on this polymer, which might be another important factor influencing the contact angle. ![Water drop contact angle on (A) HDPE; and (B) LDPE in non-treated form (N), treated with plasma for 50, 100 or 300 s (Plasma), and subsequently grafted with fibronectin (Fn) or bovine serum albumin (BSA). Mean ± S.D. from 30 measurements performed on 3 samples (10 measurements on each).](materials-06-01632-g001){#materials-06-01632-f001} Subsequent grafting of Fn or BSA onto the plasma-activated polymer surface usually resulted in a further decrease in the contact angle. This phenomenon was well apparent, particularly on HDPE and LDPE irradiated with plasma for 100 and 300 s. The increase in surface wettability could be explained by the polar groups present in fibronectin and albumin molecules, oxygen-containing and amine groups. However, on the polymers irradiated for 50 s, the presence of BSA and Fn on an HDPE surface and BSA on an LDPE surface did not have significant effects on the surface wettability, as indicated by the contact angles on these polymers, which were comparable with the values on the samples treated only with plasma. Similar results for BSA on HDPE and LDPE have been published earlier \[[@B25-materials-06-01632],[@B26-materials-06-01632]\]. In these studies focused on grafting bioactive substances onto HDPE and LDPE surfaces, only glycin and polyethylene glycol further decreased the contact angle on the plasma-irradiated polymers, while the contact angle remained similar after grafting of BSA, colloidal carbon particles and a combination of BSA and these particles. Similarly, in the present study, grafting the plasma-treated PE with Fn was more efficient in decreasing the contact angle than grafting with BSA, except in the case of HDPE irradiated for 50 s. ### 2.1.2. XPS Analysis and Immunofluorescence Staining of Grafted Biomolecules Fibronectin and albumin are proteins which contain nitrogen and disulfide bonds \[[@B27-materials-06-01632],[@B28-materials-06-01632]\]. The presence of these biomolecules on the PE surface was therefore tested using X-ray photoelectron spectroscopy analysis (XPS) by determining the concentration of nitrogen and sulfur on the polymer surface. XPS also determined the concentration of oxygen as evidence that oxidized polar groups were formed after plasma irradiation. XPS analysis showed that the concentration of nitrogen was significantly higher on HDPE and LDPE activated with plasma and subsequently grafted with Fn and BSA than on samples only treated with plasma ([Table 1](#materials-06-01632-t001){ref-type="table"}). The nitrogen found in a low amount on the plasma-treated samples was probably bound from the ambient atmosphere after plasma exposure. In addition, sulfur was determined on the samples grafted with Fn and BSA. Due to the higher presence of nitrogen and sulfur on the grafted samples than on the only plasma-treated samples, we can conclude that fibronectin and albumin were successfully bound on the plasma-treated polymer surface. XPS also indicates that the concentration of oxygen increased after the polymer was exposed to plasma, and this led to a decline in the contact angle ([Figure 1](#materials-06-01632-f001){ref-type="fig"}). The higher concentration of oxygen on the irradiated samples than on the non-treated samples indicated the formation of newly oxidized polar groups after plasma modification. materials-06-01632-t001_Table 1 ###### Concentration (in at. %) of carbon (C), oxygen (O), nitrogen (N) and sulfur (S), determined by XPS analysis, on HDPE and LDPE treated with plasma for 50 s (50), and subsequently grafted with fibronectin (Fn) and bovine serum albumin (BSA). Mean from three measurements for each experimental group; the measurement error did not exceed 5%. --------------------------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------------------- Material/Modification HDPE LDPE C (at. %) O (at. %) N (at. %) S (at. %) C (at. %) O (at. %) N (at. %) S (at. %) 50 82.2 14.2 2.5 0 83.0 14.8 1.5 0 50 Fn 72.9 20.2 6.7 0.2 82.2 13.6 2.8 at limit of detection 50 BSA 74.4 15.3 9.9 0.4 72.7 15.3 10.9 0.4 --------------------------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------------------- The presence of fibronectin and albumin was also tested by immunofluorescence staining ([Figure 2](#materials-06-01632-f002){ref-type="fig"}). The resultant fluorescence intensity was presented as the difference between the fluorescence intensity measured on samples treated with both primary and secondary antibodies and on samples treated with the secondary antibody only (i.e., the staining control). On HDPE and on LDPE grafted with Fn or BSA, the fluorescence intensity reached high positive values, indicating the presence of Fn and BSA on the polymer surface. By contrast, the fluorescence intensity had negative values on the non-treated polymers \[[Figure 2](#materials-06-01632-f002){ref-type="fig"}(A,B)\]. This was probably due to non-specific binding of the secondary antibody to the polymer surface, because the fluorescence intensity of the staining control, i.e., the sample incubated only in the secondary antibody, was higher than the fluorescence intensity after staining the non-treated polymer with both primary and secondary antibodies ([Figure 2](#materials-06-01632-f002){ref-type="fig"}C). The non-specific binding of the secondary antibody on samples grafted with Fn and BSA was significantly reduced due to the presence of these biomolecules. In this context it should be pointed out that in studies using immunocytochemical methods, BSA \[[@B25-materials-06-01632],[@B26-materials-06-01632],[@B29-materials-06-01632]\], the whole blood serum \[[@B30-materials-06-01632]\] or solutions of other biomolecules, e.g., gelatin \[[@B12-materials-06-01632]\] are commonly used for blocking non-specific binding sites for antibodies. In addition, the fluorescence intensity on HDPE grafted with Fn or BSA was higher than on LDPE, which showed that Fn and BSA were grafted in a higher concentration on the HDPE surface than on LDPE \[[Figure 2](#materials-06-01632-f002){ref-type="fig"}(A,B)\]. ![Immunofluorescence staining of (A) fibronectin; and (B) albumin and their microphotographs (C) on HDPE (H) and LDPE (L), in non-treated form (N), treated with plasma for 50 s and subsequently grafted with fibronectin (50 Fn) or bovine serum albumin (50 BSA). Olympus IX 51 microscope, obj. 10x, DP 70 digital camera. The measurements were performed at the same exposure time for all experimental groups (1.2 s).](materials-06-01632-g002){#materials-06-01632-f002} ### 2.1.3. Surface Roughness and Morphology Changes in surface roughness and morphology after plasma exposure and grafting with biomolecules were studied by the Atomic Force Microscopy (AFM) method ([Figure 3](#materials-06-01632-f003){ref-type="fig"}, [Table 2](#materials-06-01632-t002){ref-type="table"}). Plasma treatment led to ablation of the surface layer \[[@B23-materials-06-01632]\], which increased the nanoscale surface roughness in proportion to the exposure time. These events were more apparent on HDPE than on LDPE, and could be explained by the different structure of the two polymers. For example, LDPE is more crosslinked than HDPE, and thus it could be more resistant to ablation. In our earlier study, plasma treatment of PE for 400 s resulted in the removal of a surface layer of HDPE that was about 1 μm in thickness, but of a layer of LDPE that was only 0.6 μm in thickness \[[@B23-materials-06-01632]\]. ![AFM images of HDPE (H) and LDPE (L) in non-treated form (N), treated in plasma for 50 s, and subsequently grafted with fibronectin (Fn) or bovine serum albumin (BSA).](materials-06-01632-g003){#materials-06-01632-f003} materials-06-01632-t002_Table 2 ###### The mean surface roughness values (R~a~) of LDPE and HDPE in non-treated form (N), treated in plasma for 50, 100 or 300 s (50, 100, 300), and subsequently grafted with fibronectin (Fn) or albumin (BSA). Sample HDPE LDPE ------------- ------ ------ N 6.1 5.0 50 7.4 4.7 50 Fn 10.4 5.5 50 BSA 6.9 5.4 100 8.4 6.7 100 Fn 11.7 4.6 100 BSA 9.4 4.7 300 11.5 6.6 300 Fn 12.0 5.6 300 BSA 7.8 4.6 In addition, the surface morphology of both types of PE was changed after plasma modification. A lamellar structure appeared on the polymer surface, perhaps due to the different ablation rate of the amorphous and crystalline phases. The amorphous phase is ablated preferentially \[[@B23-materials-06-01632],[@B31-materials-06-01632]\]. The newly created lamellar structure was more apparent on the HDPE surface, while the LDPE surface appeared to be ablated more uniformly (homogeneously). The reason could be a higher proportion of the crystalline phase in HDPE, while the amorphous phase prevails in LDPE \[[@B23-materials-06-01632]\]. Subsequent grafting with Fn or BSA further changed the surface roughness (see [Table 2](#materials-06-01632-t002){ref-type="table"}) and the morphology of PE. Grafting with Fn strongly increased the surface roughness of HDPE, while BSA did not significantly increase, or even reduced, the surface roughness of the plasma-treated surface, although this roughness still remained higher than on the non-treated polymer. The lamellar structure was less apparent on the grafted surface than on the only-treated HDPE surface. Grafting LDPE with Fn or BSA produced rather small and non-significant changes in surface roughness and morphology. On LDPE irradiated for 100 or 300 s, both Fn and BSA showed a tendency to smoothen the material surface ([Figure 3](#materials-06-01632-f003){ref-type="fig"}). 2.2. Cell Adhesion and Growth ----------------------------- ### 2.2.1. Cell Adhesion and Growth on LDPE and HDPE Samples in a Standard Serum-supplemented Medium The proliferation of cells in DMEM with fetal bovine serum (FBS) had in general a growing tendency. On day 2 after seeding, there were similar numbers of initially adhered cells on the control polystyrene, on the samples exposed to a plasma discharge, and on all Fn-grafted and BSA-grafted samples ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). This cell behavior was observed on both LDPE and HDPE, and could be explained by a masking effect of the serum in the culture medium, i.e., secondary adsorption of serum components influencing cell adhesion, particularly vitronectin, fibronectin and albumin on the samples, which occurs within a few minutes after these substrates are exposed to the culture medium (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632],[@B9-materials-06-01632]\]). Significant differences were found only between the non-treated samples and some modified samples, usually with the exception of the BSA-grafted samples \[[Figure 4](#materials-06-01632-f004){ref-type="fig"}(A,B,D--F)\], and in two cases also with the exception of the samples only irradiated with plasma \[[Figure 4](#materials-06-01632-f004){ref-type="fig"}(B,D)\]. ![Growth curves of rat aortic smooth muscle cells cultured in a serum-supplemented medium on HDPE (A--C) or LDPE (D--F) in non-treated form (N), treated with plasma for 50, 100 or 300 s (50, 100, 300), and subsequently grafted with fibronectin (Fn) or bovine serum albumin (BSA). A standard cell culture polystyrene dish (PS) was used as a reference material. Mean ± S.E.M. from 20 measurements on microphotographs (day 2 and 4) or 200 measurements on a ViCell XR Analyser (Beckman Coulter, Brea, CA, USA) (day 6). ANOVA, Student--Newman--Keuls Method. Statistical significance: \* p ≤ 0.05 compared to non-modified HDPE or LDPE.](materials-06-01632-g004){#materials-06-01632-f004} From day 2 to day 4 after seeding, the cells on all modified PE samples showed an intense increase in their growth, while the cell growth on the non-treated PE samples was stagnant. As a result, the cell numbers on all modified samples on day 4 were significantly higher than the values on the non-treated samples. However, the differences among the groups of modified samples still remained non-significant ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). From day 4 to day 6 after seeding, the cells on most of the HDPE samples reached confluence, slowed down their proliferation and entered the stationary growth phase, i.e., their number did not increase as rapidly as between day 2 and day 4, and on some samples the number even decreased. By contrast, the cells on the LDPE samples usually continued their growth, and thus reached on an average higher final cell population densities. On day 6 after seeding, these densities ranged from about 78,000 to 111,000 cells·cm^−2^ on LDPE, while there were only 43,000 to 79,000 cells·cm^−2^ on HDPE ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). Particularly the cells on the BSA-grafted LDPE samples continued their exponential growth, and on day 6 after seeding their population densities were among the highest values reached on these samples, especially on LDPE activated with plasma for 300 s ([Figure 4](#materials-06-01632-f004){ref-type="fig"}F). These results can be considered as surprising, because albumin is known as a cell non-adhesive molecule \[[@B32-materials-06-01632]\], which has been used for creating bioinert cell-free surfaces, applicable e.g., in blood-contacting devices \[[@B33-materials-06-01632],[@B34-materials-06-01632]\]. In our study, the non-adhesive properties of albumin are slightly reflected by the relatively low numbers of initially adhered cells on the BSA-grafted samples, which often did not differ significantly from the values on the non-treated polyethylene ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). Nevertheless, the cell spreading areas were mostly larger than in the cells on non-treated samples, and usually not significantly smaller than on the control polystyrene dishes, except BSA-grafted HDPE treated with plasma for 100 s ([Figure 5](#materials-06-01632-f005){ref-type="fig"}, [Figure 6](#materials-06-01632-f006){ref-type="fig"} and [Figure 7](#materials-06-01632-f007){ref-type="fig"}). Albumin has been reported to promote adsorption of ECM molecules, e.g., vitronectin and fibronectin, in advantageous spatial conformations, supporting the accessibility of these molecules by cell adhesion receptors (for a review, see \[[@B25-materials-06-01632]\]). These molecules cannot only be adsorbed over the albumin from the serum of the culture medium, but are also synthesized and deposited by VSMC themselves \[[@B35-materials-06-01632]\], especially at later culture intervals. This may account for the relatively high proliferation activity of the cells on BSA-grafted PE ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). As for the non-treated HDPE and LDPE samples, the cell number increased from day 4 to 6 on these materials ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). However, the maximum population densities on these polymers still remained significantly lower (from about 24,000 to 29,000 cells·cm^−2^) than on the modified samples (from 43,000 to 111,000 cells·cm^−2^). In accordance with the weak cell proliferation on non-treated PE, the visualization of the cells on day 2 also showed a lower size of the cell spreading area on non-treated PE than on the modified samples ([Figure 5](#materials-06-01632-f005){ref-type="fig"}, [Figure 6](#materials-06-01632-f006){ref-type="fig"} and [Figure 7](#materials-06-01632-f007){ref-type="fig"}). This insufficient cell spreading can be explained by the relatively high hydrophobicity of non-treated HDPE and LDPE. It is known that, on highly hydrophobic surfaces, the cell adhesion-mediating proteins, e.g., fibronectin and vitronectin, present in the serum of the culture medium, are adsorbed in rigid, non-physiological and denatured form, less appropriate for binding the specific sites in the protein molecules by the cell adhesion receptors (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632],[@B9-materials-06-01632]\]). By contrast, the modified polymers were more hydrophilic, with an enhanced nanostructure or grafted with biomolecules. As mentioned above, BSA supports the adsorption of cell adhesion-mediating molecules in a near-physiological conformation, appropriate for binding to the cell adhesion receptors. In addition, the fibronectin grafted on some samples provided the possibility for direct binding of the cells, e.g., through specific RGD-containing amino acid sequences in the Fn molecules \[[@B28-materials-06-01632]\]. ![Morphology of rat aortic smooth muscle cells in two-day-old cultures in a serum-supplemented medium on HDPE in non-treated form (N), treated with plasma for 50, 100 or 300 s (50, 100, 300), and subsequently grafted with fibronectin (Fn) or bovine serum albumin (BSA). A standard cell culture polystyrene dish (PS) was used as a reference material. Cells stained with Texas Red C~2~-Maleimide and Hoechst \#33342. Olympus IX 51 microscope, obj. 20, DP 70 digital camera, bar 200 μm or 100 μm (Fn-grafted samples).](materials-06-01632-g005){#materials-06-01632-f005} ![Morphology of rat aortic smooth muscle cells in two-day-old cultures in a serum-supplemented medium on LDPE in non-treated form (N), treated with plasma for 50, 100 or 300 s (50, 100, 300), and subsequently grafted with fibronectin (Fn) or bovine serum albumin (BSA). A standard cell culture polystyrene dish (PS) was used as a reference material. Cells stained with Texas Red C~2~-Maleimide and Hoechst \#33342. Olympus IX 51 microscope, obj. 20, DP 70 digital camera, bar 200 μm or 100 μm (Fn-grafted samples).](materials-06-01632-g006){#materials-06-01632-f006} ![The size of the cell spreading area of rat aortic smooth muscle cells in two-day-old cultures in a serum-supplemented medium on (A) HDPE; or (B) LDPE in non-treated form (N), treated with plasma for 50, 100 or 300 s, and subsequently grafted with fibronectin (Fn) or bovine serum albumin (BSA). A standard cell culture polystyrene dish (PS) was used as a reference material. Mean ± S.E.M. from 32 to 95 measured cells for each experimental group. ANOVA, Student--Newman--Keuls Method. Statistical significance: ▪ p ≤ 0.05 compared to other experimental groups, indicated by the colors of these groups above the columns.](materials-06-01632-g007){#materials-06-01632-f007} ### 2.2.2. Cell Adhesion and Growth on LDPE and HDPE in a Serum-Free Medium In the serum-free medium, the differences in the behavior of VSMC on the PE samples with various modifications became more apparent than in the serum-supplemented medium, and the growth dynamics of the cells revealed a different trend. From days 2 to 4 after seeding, the cell number was mostly stagnant, and on some samples it even decreased. There was an increasing tendency only between days 4 and 6 ([Figure 8](#materials-06-01632-f008){ref-type="fig"}). This stagnancy in cell growth was most likely because there was a limited quantity of the ECM proteins normally present in fetal serum, mainly vitronectin and fibronectin, which ensure proper adhesion, spreading and subsequent growth of cells on the material surfaces. The serum-free medium used in this study did not contain these proteins. However, cell adhesion-mediating proteins can be present on cells in small quantities even after trypsinization \[[@B36-materials-06-01632]\], and could be used for cell adhesion after seeding cells on the tested materials, particularly on polystyrene, non-treated PE and plasma-treated PE (on Fn-grafted PE, the Fn molecules could be used directly for cell adhesion). In addition, VSMC are able to synthesize a wide spectrum of cell adhesion-mediating ECM molecules (for a review, see \[[@B35-materials-06-01632]\]). This takes place especially at later culture intervals. By way of these molecules, the cells adhere to the material surface through the bonds between specific amino acid sequences in these molecules and cell adhesion receptors on the cell membrane, e.g., integrins. Another way for cells to attach to the tested materials is by so-called weak chemical bonding, i.e., hydrogen bonding, electrostatic, polar or ionic interactions between various molecules on the cell membrane and functional chemical groups on the polymers, etc., but this type of binding does not transfer specific signals into the cells and does not ensure their survival, growth and other functions (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632],[@B9-materials-06-01632],[@B37-materials-06-01632],[@B38-materials-06-01632]\]). ###### Growth curves of rat aortic smooth muscle cells cultured in a serum-free medium on (A--C) HDPE; or (D--F) LDPE in non-treated form (N), treated with plasma for 50, 100 or 300 s (50, 100, 300), and subsequently grafted with fibronectin (Fn) and bovine serum albumin (BSA). A standard cell culture polystyrene dish (PS) was used as a reference material. Mean ± S.E.M. from 20 measurements (day 2 and 4) or 4 measurements (day 6). ANOVA, Student--Newman--Keuls Method. Statistical significance: \* p ≤ 0.05 compared to the non-modified LDPE or HDPE. ![](materials-06-01632-g008a) ![](materials-06-01632-g008b) As early as on day 2 after seeding, the cell non-adhesive properties of BSA were completely revealed in the serum-free medium. This effect of albumin was more apparent on HDPE, where the cell densities ranged from 37 to 367 cells·cm^−2^, while on LDPE the values were slightly higher (551 to 1470 cells·cm^−2^; [Figure 8](#materials-06-01632-f008){ref-type="fig"}). Accordingly, the cell spreading was also the lowest on the HDPE samples ([Figures 9](#materials-06-01632-f009){ref-type="fig"}). The cells mainly assumed a spherical shape on these surfaces, while on the other samples, even on the BSA-grafted LDPE samples ([Figure 10](#materials-06-01632-f010){ref-type="fig"}) or on non-treated LDPE and HDPE, the cell morphology was spindle-shaped or polygonal. From day 4 to 6, the cell numbers increased slightly, probably by preferential cell growth on sites with discontinuity of the albumin layer, or secondary deposition of cell adhesion-mediating ECM proteins synthesized by the cells themselves on the albumin layer. On day 6, this increase was slightly more apparent on BSA-grafted LDPE (from 2800 to 4900 cells·cm^−2^) than on BSA-grafted HDPE (from 400 to 1500 cells·cm^−2^). However, these samples still displayed the lowest cell population densities, irrespective of plasma exposure time ([Figure 8](#materials-06-01632-f008){ref-type="fig"}). The effects of the grafted fibronectin were controversial. We expected high adhesion and subsequent growth of cells on these samples, due to specific receptor-mediated adhesion of cells to Fn grafted on the plasma-activated polymer surfaces. However, the cell densities in general on all Fn-grafted samples were similar to the values on the plasma-irradiated surfaces, or on the control polystyrene culture dishes. Among all Fn-grafted samples, the weakest proliferation was observed on LDPE irradiated with plasma for 50 s (L 50Fn; [Figure 8](#materials-06-01632-f008){ref-type="fig"}B), where the cell population density on day 4 was lower than on non-treated LDPE, and as low as on the BSA-grafted samples. On day 6, the number of cells increased, but was still significantly lower than on PS or LDPE irradiated for 50 s ([Figure 8](#materials-06-01632-f008){ref-type="fig"}B). The relatively low cell proliferation activity on Fn-grafted polyethylene could be explained by relatively low quantity of the Fn molecules attached to the polymer surface. As indicated by XPS, the concentrations of nitrogen and sulfur, which are indicators of the presence of protein molecules, were relatively low on the Fn-grafted surfaces in comparison with the BSA-grafted surfaces, particularly on the LDPE samples. On LDPE, the concentration of S was even on the limit of detection ([Table 1](#materials-06-01632-t001){ref-type="table"}). In accordance with this, the intensity of fluorescence of antibody-labeled Fn grafted to plasma-activated surfaces is lower than in the case of albumin \[[Figure 2](#materials-06-01632-f002){ref-type="fig"}(A,B)\] In addition, the pictures indicate less homogeneous distribution of fibronectin on these surfaces compared to albumin ([Figure 2](#materials-06-01632-f002){ref-type="fig"}C). Also the geometrical conformation of Fn, which is more important for cell binding than the absolute amount of Fn immobilized on the material surface (\[[@B39-materials-06-01632]\], for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632],[@B9-materials-06-01632]\]) could be less appropriate for binding by the cell adhesion receptors. In spite of this, the cells on the Fn-grafted HDPE and LDPE samples were of similar polygonal morphology as the cells on the control polystyrene and plasma-irradiated samples ([Figure 9](#materials-06-01632-f009){ref-type="fig"} and [Figure 10](#materials-06-01632-f010){ref-type="fig"}), although their spreading area was often smaller ([Figure 11](#materials-06-01632-f011){ref-type="fig"}). ![Morphology of rat aortic smooth muscle cells in two-day-old cultures in a serum-free medium on HDPE in non-treated form (N), treated with plasma for 50, 100 or 300 s (50, 100, 300), and subsequently grafted with fibronectin (Fn) and bovine serum albumin (BSA). A standard cell culture polystyrene dish (PS) served as a reference material. Cells stained with Texas Red C~2~-Maleimide and Hoechst \#33342. Olympus IX 51 microscope, obj. 20, DP 70 digital camera, bar 200 μm or 100 μm (Fn-grafted samples).](materials-06-01632-g009){#materials-06-01632-f009} ![Morphology of rat aortic smooth muscle cells in two-day-old cultures in a serum-free medium on LDPE in non-treated form (N), treated with plasma for 50, 100 or 300 s (50, 100, 300), and subsequently grafted with fibronectin (Fn) and bovine serum albumin (BSA). A standard cell culture polystyrene dish (PS) was used as a reference material. Cells stained with Texas Red C~2~-Maleimide and Hoechst \#33342. Olympus IX 51 microscope, obj. 20, DP 70 digital camera, bar 200 μm or 100 μm (Fn-grafted samples).](materials-06-01632-g010){#materials-06-01632-f010} ![The size of the cell spreading area of rat aortic smooth muscle cells in two-day-old cultures in a serum-free medium on (A) HDPE; or (B) LDPE in non-treated form (N), treated with plasma for 50, 100 or 300 s, and subsequently grafted with fibronectin (Fn) or bovine serum albumin (BSA). A standard cell culture polystyrene dish (PS) was used as a reference material. Mean ± S.E.M. from 4 to 269 measured cells. ANOVA, Student--Newman--Keuls Method. Statistical significance: ▪ p ≤ 0.05 compared to other experimental groups, indicated by the colors of these groups above the columns.](materials-06-01632-g011){#materials-06-01632-f011} 3. Experimental Section ======================= 3.1. Preparation and Modification of Polymer Samples ---------------------------------------------------- The experiments were carried out on high-density and low-density polyethylene foils (HDPE and LDPE, both purchased from Granitol A. S., Moravsky Beroun, Czech Republic). The HDPE foils were of the Microten M\S type (thickness 40 μm, density 0.951 g·cm^−3^, melt flow index 0.14 g·10 min^−1^), and the LDPE was of the Granoten S\H type (thickness 40 μm, specific density 0.922 g·cm^−3^, melt flow index 0.8 g·10 min^−1^). Both types of polyethylene were cut into circular samples (diameter 2 cm) using a metallic perforator (punch). The samples were then treated by an Ar^+^ plasma discharge (gas purity 99.997%) using a Balzers SCD 050 device (BalTec Maschinenbau AG, Pfäffikon, Switzerland). The time of exposure was 50, 100 or 300 s, and the discharge power was 3 W. Immediately after the plasma treatment, the samples were immersed in solutions of two components of fetal bovine serum (FBS), namely 2 wt.% of fibronectin (Fn, Sigma-Aldrich, St. Louis, MI, USA, Cat. No. F2006) or 2 wt.% of bovine serum albumin (BSA, Sigma-Aldrich, Cat. No. A9418), both diluted in phosphate-buffered saline (PBS; Sigma-Aldrich, Cat. N° P4417). The samples were incubated for 24 h at room temperature (RT) in order to allow grafting of biomolecules to the plasma activated polymer surface. The samples were then rinsed in distilled water, air-dried at RT and stored in an air atmosphere for three weeks in order to reorganize and stabilize their surface structure, particularly the orientation of the oxidized structures on the material surface (the so-called aging period) \[[@B23-materials-06-01632]\]. 3.2. Characterization of the Physical and Chemical Properties of the Polymer Surface ------------------------------------------------------------------------------------ ### 3.2.1. Surface Wettability The sessile water drop contact angle was measured by See System reflection goniometry (Masaryk University, Brno, Czech Republic). The volume of the water drop on the polymer surface was 8 μL. The contact angle was measured after an aging period of three weeks. For each experimental group, three samples were used, and on each sample, 10 measurements were performed in different regions homogeneously distributed on the sample surface. The contact angle was presented as mean ± Standard Deviation (S.D.) from 30 measurements. Statistical analyses were performed using SigmaStat (Jandel Corp., Las Vegas, NV, USA). Multiple comparison procedures were performed by the One Way Analysis of Variance (ANOVA), Student--Newman--Keuls method. p values equal to or less than 0.05 were considered significant. ### 3.2.2. Chemical Composition of the Polymer Surface The presence of grafted biomolecules, i.e., fibronectin (Fn) and bovine serum albumin (BSA), on the HDPE and LDPE surface was determined by X-ray photoelectron spectroscopy (XPS, Omicron Nanotechnology ESCAProbeP spectrometer) and by immunofluorescence staining. XPS determined the concentration (at.%) of the main elements present in the grafted polymer surface (oxygen, nitrogen, sulfur and carbon). For immunofluorescence staining, specific primary antibodies against Fn (monoclonal anti-human fibronectin, Sigma-Aldrich, Cat. No. F0791) and BSA (monoclonal anti-bovine serum albumin, Sigma--Aldrich, Cat. No. B2901, St. Louis, MI, USA) were used. The primary antibodies were diluted in PBS to a concentration of 1:200 and were applied to HDPE and LDPE samples modified in plasma and subsequently grafted with Fn or BSA, and were also applied to the non-modified control samples. The samples were incubated with primary antibodies overnight at 4 °C. After rinsing with PBS, the secondary antibody, i.e., goat anti-mouse F(ab')2 fragments of IgG (H + L), conjugated with Alexa Fluor^®^ 488 (Molecular Probes, Invitrogen, Cat. No. A11017, Eugene, OR, USA), was diluted in PBS to a ratio of 1:400, and was added to the samples for 1 hour at RT. As a staining control, the secondary antibody was applied to samples which had not been treated with the primary antibodies before. The samples were then rinsed twice in PBS. The fluorescence intensity was evaluated from 10 randomly chosen fields homogeneously distributed on the material surface using an epifluorescence microscope (IX 51, Olympus, Tokyo, Japan, obj. 10×) equipped with a digital camera (DP 70, Olympus, Tokyo, Japan) at the same exposure time for all experimental groups (1.2 s). The data was presented as mean ± Standard Error of Mean (S.E.M.). Statistical analyses were performed using SigmaStat (Jandel Corp., Las Vegas, NV, USA). Multiple comparison procedures were made by the One Way Analysis of Variance (ANOVA), Student--Newman--Keuls method. p values equal to or less than 0.05 were considered significant. ### 3.2.3. Surface Morphology and Roughness The changes in surface morphology and roughness were determined by Atomic Force Microscopy (AFM), using a CP II device (VEECO, Santa Barbara, CA, USA) working in tapping mode. We used an RTESPA-CP Si probe, with spring constant 20--80 N/m. By repeated measurements of the same region (1 × 1 µm), it was proven that the surface morphology did not change after three consecutive scans. The mean roughness value (R~a~) represents the arithmetic average of the deviations from the center plane of the sample. 3.3. Cells and Culture Conditions --------------------------------- The samples were sterilized with 70% ethanol for 1 hour, placed into 12-well polystyrene multidishes (TPP, Switzerland, well diameter 2.2 cm) and then air-dried for 12 hours in a sterile environment. The materials were fixed to the bottom of the culture wells by plastic rings in order to prevent them floating in the cell culture media, and were seeded with vascular smooth muscle cells (VSMC), derived from the rat aorta by an explantation method \[[@B40-materials-06-01632]\]. The cells were used in passage 3. Each well contained 50,000 cells (i.e., about 14,000 cells·cm^−2^) and 3 mL of serum-supplemented or serum-free medium. The serum-supplemented medium was Dulbecco's modified Eagle's Medium (DMEM; Sigma, Ronkonkoma, NY, USA, Cat. N° D5648) with 10% of fetal bovine serum (FBS; Sebak GmbH, Aidenbach, Germany) and 40 µg/mL of gentamicin (LEK, Ljubljana, Slovenia), and the serum-free medium was SmGM^®^-2 Smooth Muscle Growth Medium-2 (SmBM; Lonza, Walkersville, MD, USA, Cat. N° CC-3182), supplemented with epithelial growth factor (EGF), fibroblast growth factor-B (FGF-B) and insulin according to the manufacturer's protocol. The serum-free medium was used in order to enhance the effects on cell adhesion and growth of the chemical functional groups and the biomolecules grafted onto the material surface. This influence can be masked, at least partly, by secondary adsorption of various biomolecules on the polymer surface from the serum supplement of the culture medium. The cells were then cultured for 2, 4 and 6 days in a cell incubator at 37 °C and in a humidified atmosphere with 5% of CO~2~ in the air. 3.4. Evaluation of the Cell Number and Cell Spreading Area ---------------------------------------------------------- On days 2 and 4 after seeding, the cells were rinsed with PBS, fixed with cold 70% ethanol (−20 °C, 5 min), and then stained for 2 hours at RT with a combination of the following fluorescent dyes diluted in PBS: Hoechst \#33342 nuclear dye (Sigma-Aldrich, 5 µg·mL^−1^) and Texas Red C~2~-maleimide membrane dye (Molecular Probes, Invitrogen, Carlsbad, CA, USA, Cat. No. T6008, 20 ng·mL^−1^). The cells were counted on microphotographs taken in 10 randomly chosen fields homogeneously distributed on the material surface, using an epifluorescence microscope (Olympus IX 51, Japan, obj. 20×) equipped with a digital camera (DP 70, Japan). This approach was not applicable on day 6, when the cells on some samples reached confluence and started to overlap each another. The cells were therefore detached from the materials by a trypsin-EDTA solution (Sigma-Aldrich, Cat. N° T4174) and were counted using a Vi-CELL XR Analyser (Beckman Coulter, Brea, CA, USA). The pictures taken on day 2 after seeding were also used for measuring the size of the cell spreading area using Atlas software (Tescan Ltd., Brno, Czech Republic). Two independent samples were used for each time interval and experimental group. Non-modified polyethylene (HDPE or LDPE) and standard cell culture polystyrene wells were used as reference materials. The quantitative data was presented as mean ± S.E.M. Statistical analyses were performed using SigmaStat (Jandel Corp., USA). Multiple comparison procedures were made by the One Way Analysis of Variance (ANOVA), Student--Newman--Keuls method. p values equal to or less than 0.05 were considered significant. The cell numbers obtained on day 2, 4 and 6 after seeding were expressed as the number of cells per cm^2^ (i.e., the cell population density), and were used for constructing growth curves. 4. Conclusions ============== Modification of polyethylene (HDPE and LDPE) with Ar^+^ plasma and grafting with fibronectin and albumin influenced the cell colonization of both polymers. In the serum-supplemented medium, all modifications improved the adhesion and growth of vascular smooth muscle cells in comparison with the unmodified polymer. The cells reached higher final population densities, on an average, on LDPE than on HDPE. In the serum-free medium, grafting with albumin did not support cell colonization due to its non-adhesive properties, which were not masked by the adsorption of cell adhesion-mediating molecules, normally present in the serum. The beneficial effect of the plasma treatment on cell adhesion and growth could be attributed to the formation of new-oxidized structures on the polymer surface, an increase in surface wettability, changes in surface morphology, and enhanced nanoscale roughness. Changes in the physical and chemical properties of the material surface were more apparent on HDPE than on LDPE. The intensity of the fluorescence indicated that HDPE also contained higher quantities of grafted fibronectin and albumin. This study was supported by the Grant Agency of the Czech Republic (grant No. P108/12/G108 "Center of Excellence"). Robin Healey (Czech Technical University, Prague) is gratefully acknowledged for his language revision of the manuscript. [^1]: These authors contributed equally to this work.	{ "pile_set_name": "PubMed Central" }
Background {#Sec1} ========== Hepatocellular carcinoma (HCC) is one of the most common malignant cancer worldwide and the second leading cause of cancer death for male in China, especially in China's rural areas, and it is also considered as the sixth leading cause of cancer-related mortality for male in developed countries. According to reports, the incidence and mortality of liver cancer in China occupies for nearly 50% of the global liver cancer population \[[@CR1], [@CR2]\]. Unfortunately, the pathogenesis of HCC is concealed, which is difficult to find in early stage. The rapid development and high rate of postsurgical relapse is a major challenge of HCC, owing to the rapid proliferation and differentiation of hepatoma cells \[[@CR3]\]. Thus, new antitumor agents and potential mechanisms are urgently needed to inhibit the proliferation of hepatoma cells. Sorafenib is considered to be the only effective treatment targeted drug for patients with advanced HCC, but it shows poor treatment effect for patients with extrahepatic spread or vascular invasion \[[@CR4]\]. Therefore, screening effective anti-hepatoma drugs from traditional Chinese medicine herbs is a necessary research field. In recent decades, some effective ingredients of Chinese medicinal herbs, such as icariin, psoralen, osthole, matrine, ginsenoside Rg1/Rh2 and celastrol, were reported to inhibit the proliferation of hepatocellular carcinoma cells \[[@CR5]--[@CR11]\]. Psoralea species possess unique bioactive compounds with anti-cancer properties \[[@CR12]\], and there are many tumor-suppression patents of traditional Chinese medicine which have psoralea as basic remedy \[[@CR13]--[@CR15]\]. Additionally, psoralen is the main active component of Psoralea corylifolia (L.) (Fig. [1](#Fig1){ref-type="fig"}a) and is used as marker to assess its quality \[[@CR16], [@CR17]\], which can inhibit the proliferation of adrenocortical tumors Y1 cells and pituitary tumor AtT20 cells \[[@CR18]\]. Psoralen is a kind of furocoumarin, the molecular formula of psoralen is C~11~H~6~O~3~, with molecular weight of 186.1 and dissolved in DMSO. The chemical structure of psoralen was shown in Fig. [1](#Fig1){ref-type="fig"}b. Furthermore, psoralen plays roles in anti-tumor, anti-oxidation and reversing the multidrug resistance, and it is concerned with cell cycle, apoptosis, calcium antagonism and estrogen-like effects \[[@CR19]--[@CR22]\]. However, the effect of psoralen on HCC is still unclear.Fig. 1Chemical structure of psoralen. a The plant diagram of psoralea. b The chemical structure of psoralen Endoplasmic reticulum stress (ER-stress) is the basic cells response against various extracellular factors, including UV, anoxia, oxidative stress, toxic substances, nutrient deficiency and drug agonists, which will result in the unfolded proteins accumulation in the endoplasmic reticulum and lead to Ca^2+^ homeostasis imbalance. According to previous studies, three major ER-spanning transmembrane proteins, PERK (protein kinase R-like ER kinase), ATF6 (activating transcription factor 6) and IRE1 (inositol-requiring enzyme 1) subsequently drive mutually reinforcing signaling pathways to correct the protein-misfolding stress \[[@CR23]\]. In this process, glucose-regulated protein of 78 (GRP78), also known as heavy chain-binding protein (Bip), plays an important role in detecting the accumulation of unfolded proteins in ER lumen and releasing the three sensors, so GRP78 is considered to be an ER homeostasis receptor \[[@CR24]\]. In the early stage of ER stress, the unfolded protein response (UPR) acts to help cells to cope with the stress by attenuating protein synthesis, clearing the unfolded/misfolded proteins and increasing the capacity of the ER to fold proteins, which restore the intracellular homeostasis and protect cell functions. However, strong and sustained ER stress induced by the activation of UPR will cause imbalance of ER homeostasis. Prolonged activation of the UPR can induces apoptosis pathway following ER stress, mainly involving in the CHOP/DITT3/GADD153 gene activation pathway, JNK activation pathway, ER-specific cysteine Caspase-12 activation pathway. In this study, we observed the proliferation inhibition effects of psoralen on hepatoma SMMC7721 cells and further investigated its relationship with ER stress. The results suggest that psoralen can induce cell cycle arrest and apoptosis by ER stress to inhibit malignant proliferation of hepatoma cells. Results {#Sec2} ======= Psoralen inhibits the proliferation of SMMC7721 cells {#Sec3} ----------------------------------------------------- We first examined whether psoralen was cytotoxic to hepatoma cell lines. Compared with the control group, the SMMC7721 cells treated with different dose psoralens for 24 h, 48 h and 72 h. We found that 20--80 μM psoralen inhibited cell proliferation and the inhibition rate was 17.50--25.10% in 24 h (Table [1](#Tab1){ref-type="table"}). More than 20 μM psoralen could effectively inhibit cell proliferation in 48 h (P \< 0.001) and 72 h (P \< 0.001) (Table [1](#Tab1){ref-type="table"}). In particular, more than 80 μM psoralen caused cell death (Fig. [2](#Fig2){ref-type="fig"}a, b). Next, we explored the effect of psoralen on L02 hepatocyte cell line and HepG2. We found that 10 μM, 20 μM and 40 μM PSO had little effects on the proliferation of L02, but 80 μM PSO could inhibit L02 proliferation (Fig. [2](#Fig2){ref-type="fig"}c). However, psoralen treatment showed little effect on proliferation of HepG2 cells (Fig. [2](#Fig2){ref-type="fig"}d). It is suggested that psoralen has a significant inhibitory effect on the proliferation of SMMC7721.Table 1Inhibitory effect of psoralen on proliferation of SMMC7721 hepatoma cells ($\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\overline{x}$$\end{document}$ ± s, n = 3)Dose/μMInhibition rate/%24 h48 h72 h0 (Con)0.00 ± 2.700.00 ± 8.900.00 ± 5.70108.40 ± 8.0019.60 ± 6.90\22.80 ± 10.9\2017.50 ± 1.9034.80 ± 6.90\\44.60 ± 7.80\\4022.90 ± 2.9054.80 ± 0.60\\\74.90 ± 2.80\\\8025.10 ± 6.4067.30 ± 3.50\\\96.60 ± 0.50\\\Compared with the control group, \P \< 0.05, \\P \< 0.01, \\\P \< 0.001. The data are from three independent experiments, four wells for each experiment Fig. 2Effects of psoralen on the proliferation of SMMC7721, L02 and HepG2. a Different doses of psoralen inhibit the proliferation of SMMC7721 for 24--72 h. The A570 value of SMMC7721 (b), L02 (c) and HepG2 (d) with the treatment of psoralen in 10--80 μM for 72 h. Values are mean ± SD (n = 4) and \* is means compared to the Con group, \P \< 0.05; \\P \< 0.01; \\\P \< 0.001 Psoralen causes shrinkage and endoplasmic reticulum dilation of SMMC7721 {#Sec4} ------------------------------------------------------------------------ The SMMC7721 cells were treated with 40 μM psoralen or 0.1 μM thapsigargin for 12 h, 24 h, and 48 h, respectively. The results showed that SMMC7721 cells appeared mild shrinkage and the cells gap increased. With the treatment prolonged, the SMMC7721 cells shrunk significantly, furthermore the dead and floated cells could be observed. The thapsigargin caused the cells shrinkage as a positive control, which showed significant inhibitory effect than that of psoralen (Fig. [3](#Fig3){ref-type="fig"}a). These results suggested that psoralen can induce the shrinkage and reduce adhesion of SMMC7721 cells. Next, we observed the intracellular structure of SMMC7721, which showed a large nucleus, few cytoplasm, scattered smooth endoplasmic reticulum and some mitochondria in transmission electron microscope. However, endoplasmic reticulum dilation appeared in SMMC7721 by 40 μM psoralen treatment for 24 h, and the cytoplasm presented loose and derangement distribution. Moreover, 0.1 μM thapsigargin can cause the endoplasmic reticulum dilation and cytoplasmic structure disorder (Fig. [3](#Fig3){ref-type="fig"}b). It is suggested that psoralen can cause abnormalities of endoplasmic reticulum and cytoplasmic structure in SMMC7721 hepatoma cells.Fig. 3Effects of psoralen on the morphology of SMMC7721. a SMMC7721 cells shrunk in the dose of 40 μM psoralen or 0.1 μM thapsigargin for 12--48 h. b Abnormalities of endoplasmic reticulum and cytoplasmic structure in SMMC7721 in the dose of 40 μM psoralen or 0.1 μM thapsigargin for 24 h. Yellow boxes represent the abnormal region Psoralen induces cell cycle arrest at G1 phase in SMMC7721 {#Sec5} ---------------------------------------------------------- We next evaluated whether psoralen modulated cell cycle progression to affect cell viability. PI staining analysis by flow cytometry demonstrated that 40 μM psoralen led to a significant increase in the percentage of cells at the G1 phase and a decrease in cells at the S phase. And 0.1 μM thapsigargin also caused a significant increase in the percentage of cells at the G1 phase (Fig. [4](#Fig4){ref-type="fig"}a). We also determined the effects of psoralen on the expression of G1/S transition-related cell cycle proteins. Compared with the control group, CyclinD1 protein expression was increased significantly by 40 μM psoralen, and CyclinE1 expression was suppressed significantly. However, psoralen showed little influence on the expression of CDK4 and p27Kip1 (CDKN1B). Additionally, the proteins expression of CyclinD1, CyclinE1, CDK4 and p27Kip1 were decreased significantly by 0.1 μM thapsigargin (Fig. [4](#Fig4){ref-type="fig"}b). These results suggested that psoralen could induces cell cycle arrest at the G1 phase by reducing the expression of CyclinE1 in SMMC7721, thus inhibiting the growth of hepatoma cells.Fig. 4Psoralen induces cell cycle arrest at G1 phase in SMMC7721. Flow cytometry (a) and Western blot (b) assays showing the effects of psoralen and thapsigargin on cell cycle progression and expression of G1/S transition-related proteins, respectively. Relative protein levels are showed on the right. ns no significant; \\\P \< 0.001 Psoralen induces apoptosis in SMMC7721 hepatoma cells {#Sec6} ----------------------------------------------------- We next determined the effects of psoralen on cellular apoptosis in SMMC7721 cells. Annexin-V staining revealed that psoralen treatment increased late apoptosis in SMMC7721 hepatoma cells and showed a significant dose-effect relationship (Fig. [5](#Fig5){ref-type="fig"}a). Western blot analysis also demonstrated that treated with 10 μM psoralens for 48 h could increase the level of Bax, and 40 μM psoralen could attenuate the expression of Bcl-2 (Fig. [5](#Fig5){ref-type="fig"}b). These results suggested that psoralen induces cell apoptosis in SMMC7721, which may be related to the inhibition of Bcl-2 protein expression.Fig. 5Psoralen induces apoptosis in SMMC7721 hepatoma cells. AnnexinV-PI staining (a) and western blot (b) assays showing the effects of different doses of psoralen on cell apoptosis and the expression of apoptosis-related proteins The effects of psoralen on ER-stress related gene expression in SMMC7721 {#Sec7} ------------------------------------------------------------------------ To investigate the effects of psoralen on the key gene expression of ER-stress, we treated SMMC7721 with different doses and time duration. Compared with the control group, GRP78 and SGK1 mRNA expression were increased significantly in the treatment with 20 μM and 40 μM psoralen for 48 h (P \< 0.05), DDIT3 and GDF15 mRNA levels were dose-dependent increased (P \< 0.01), and relative mRNA expression of GADD34 and FOXO3 were also increased by 40 μM psoralen (P \< 0.05). However, different doses of psoralen showed little influence on EIF2A mRNA expression (Fig. [6](#Fig6){ref-type="fig"}a). Next, we detected the mRNA levels with the treatment of 40 μM psoralen for 6 h, 12 h and 24 h. Compared with the control group, DDIT3, XBP1, and ATF4 mRNA expression were time-dependent increased (P \< 0.05), GRP78 and GADD34 mRNA expression were increased significantly with the treatment for 12 h and 24 h (P \< 0.05), and GRP94 and ATF6 mRNA levels were increased significantly after 24 h (P \< 0.05) (Fig. [6](#Fig6){ref-type="fig"}b). We also detected some gene involved in other biological processes on ER. Compared with the control group, GDF15, HERP, WARS, SYVN1, SCAP, ACAT1, DNAJC3, CALR, EDEM1, SEC61A1, and PDIA4 mRNA expression were increased significantly by 40 μM psoralen after treatment for 24 h (P \< 0.05) (Fig. [6](#Fig6){ref-type="fig"}c). However, psoralen treatment showed little effect on the mRNA levels of ER-stress gene in HepG2 (Additional file [1](#MOESM1){ref-type="media"}: Figure S1). These results showed that psoralen can induce unfolded proteins and cause endoplasmic reticulum stress in SMMC7721 hepatoma cells.Fig. 6Effects of psoralen on gene expression of ER-stress. a* The mRNA levels of key gene in ER-stress by different doses of psoralen treatment for 48 h. b The mRNA levels of key gene in ER-stress by 40Μm psoralen for 6 h, 12 h and 24 h. c The mRNA levels of ER functional gene by 40Μm psoralen for 6 h, 12 h and 24 h. Values are mean ± SD (n = 3) and \* is means compared to the Con group, \p \< 0.05; \\p \< 0.01; \\\p \< 0.001 The effects of psoralen on ER-stress related protein expression in SMMC7721 {#Sec8} --------------------------------------------------------------------------- 40 μM psoralen was used to explore the effect of psoralen on ER-stress related protein expression. Compared with the control group, GRP78 and XBP-1s protein expression were increased significantly after treatment for 24 h and 48 h (P \< 0.05). And the level of IRE1α was also up-regulated time-dependently (P \< 0.05). However, psoralen had no significant effect on the expression of CHOP protein. As a positive control drug, thapsigargin (TG) promoted the expression of GRP78 and XBP-1s protein only in 4 h (P \< 0.05) (Fig. [7](#Fig7){ref-type="fig"}a). To further evaluate the effects on ER-stress, the ER-stress inhibitor TUDC was added with the treatment of psoralen. Western blot assay showed that TUDC could not inhibit the increased expression of GRP78, XBP-1s, IREα, and CHOP proteins induced by psoralen (Fig. [7](#Fig7){ref-type="fig"}b). Similarly, TUDC had little effects on the psoralen-induced up-regulation of GRP78 and DDIT3 mRNA expression (Fig. [7](#Fig7){ref-type="fig"}c). The results suggested that psoralen can induce the endoplasmic reticulum stress in SMMC7721 hepatoma cells.Fig. 7Effects of psoralen on ER-stress related protein expression in SMMC7721. a Western blot assays showing the effects of psoralen and thapsigargin on ER-stress related protein expression. b Western blot assays showing the effects of TUDC on the expression of ER-stress related protein induced by PSO. c The mRNA levels of GRP78 and DDIT3 with TUDC under PSO and TG treatment. Values are mean ± SD (n = 3) and \* is means compared to the Con group, \\P \< 0.01 Discussion {#Sec9} ========== Psoralen is extracted from dried fruits of P. corylifolia (L.) from Leguminosae, which shows an active effect including anti-tumor, photosensitivity, cardiovascular protection, anti-histamine and estrogen-like pharmacological effects \[[@CR25]--[@CR29]\]. In the traditional Chinese medicine, the property and flavor of P. corylifolia (L.) are spicy, bitter and warm, and it has warming kidney and enhancing yang, preventing asthma, to warm the kidney and the spleen and check diarrhea, and removing beverage effect \[[@CR30]\], and P. corylifolia can be used as basic remedy to suppress tumor \[[@CR13]--[@CR15], [@CR31], [@CR32]\]. However, the roles of psoralen on the treatment of various tumors are not well understood. In our previous studies, we found that psoralen can inhibit the proliferation of Y1 adrenocortical tumor cells and AtT20 pituitary tumor cells. Based on these results, we explored the inhibition effects of psoralen on SMMC7721 human hepatoma cells for further study. Our results demonstrated that psoralen could significantly inhibit the proliferation of SMMC772 and induce apoptosis. In addition, we found that psoralen could induces cell cycle arrest at G1 phase and induce ER-stress to promote cell apoptosis, thus revealing the molecular mechanism of psoralen in anti-hepatoma effect in vitro. As the malignant proliferation of hepatoma cells is an important characteristic of the biological behavior of tumor cells, the inhibition effect of psoralen on the proliferation of SMMC7721 hepatoma cells has an objective prospect of research and development. Similarly, psoralen can inhibit the proliferation of BGC-823 human gastric cancer cells and the half inhibitory rate (IC50) was 5.82 μg/mL \[[@CR33]\]. Moreover, MG-63 osteogenic sarcoma cells were inhibited with 640 μmol/L psoralen treatment for 72 h, and the cell inhibition rate was 77.74%, and the cells became smaller, rounder, vacuoles and pseudopodia disappeared \[[@CR34]\]. Although there are lots of studies about the pro-apoptotic function of psoralen, however, recent study demonstrated that psoralen could inhibit the apoptosis and increase the proliferation of osteoporotic osteoblasts by modulating IRE1--ASK1--JNK pathway \[[@CR35]\]. Cell cycle disorder is an important feature of tumor cells, especially malignant transformation, so cyclin is an important target for the anti-cancer effect of psoralen. Among them, CyclinD1 and CyclinE1 play vital roles in the regulation of cell cycle, which are positive regulators of cyclin-dependent kinase (CDK). CyclinD1 associates with CDK4/6 and CyclinE1 binds to CDK2, which promote cells enter S phase from G1 phase, causing cell division and proliferation \[[@CR36]\]. Once the proteins are over-expressed, the cell will shorten G1 phase and advance into S phase, causing continued proliferation and increasing the risk of cancer \[[@CR37]\]. It has been reported that psoralen can form a complex with DNA in the form of non-covalent bond that cause DNA damage \[[@CR38]\], arrest cells in G1 phase, and repair damaged DNA. Psoralen can arrest cells in G1 and G2 phases, and significantly reduce the proportion of S phase cells thereby inhibiting cell proliferation of MCF-7 human breast cancer cells \[[@CR27]\]. However, another study reported that 10 μM and 30 μM psoralen can promote the progression of MCF-7 cells from G1 to S phase \[[@CR39]\]. It suggested that the regulation of psoralen on cell cycle in different tumors was not the same. Furtherly, psoralen can induce late apoptosis of SMMC7721 cells and show dose-effect relationship. According to recent studies, psoralen-induced apoptosis of K562 cells was observed under electron microscope \[[@CR40]\]. In gastric cancer, early apoptosis of BGC-803 cells could be obvious after psoralen treatment for 48 h \[[@CR41]\]. In this research, it suggests that psoralen has an exact apoptosis-inducing effect on SMMC7721 hepatoma cells, however, it is mainly not through the molecular mechanisms of the pro-apoptotic protein Bax and the apoptosis inhibitor protein Bcl-2. As well known, apoptosis may be involved in endoplasmic reticulum stress. SMMC7721 cells were observed under transmission electron microscopy after 24 h of psoralen treatment, and we found that the endoplasmic reticulum was dilatated. Then the mRNA and the protein levels of endoplasmic reticulum-associated functional molecules were detected, which also indicated that psoralen could induce ER-stress. The endoplasmic reticulum is a membranous intracellular organelle, including rough endoplasmic reticulum and smooth endoplasmic reticulum. It is mainly involved in the processes like synthesis and secretion of membrane proteins, the correct folding of proteins, the storage of Ca^2+^, the synthesis and the metabolism of lipids and cholesterol \[[@CR23]\]. Due to the influence of extrinsic or intrinsic factors, misfolded and unfolded proteins accumulate in the lumen of the endoplasmic reticulum as well as the disorder of the Ca^2+^ balance are known as ER-stress. The unfolded protein reaction in cells mainly involves three signaling pathways: double-strand RNA-activated protein kinase-like ER kinase (PERK) pathway, inositol-requiring enzyme 1 (IRE1) pathway, activating transcription factor 6 (ATF6) pathway. Under physiological conditions, PERK, IRE1 and ATF6 are all combined with glucose regulatory protein 78 (GRP78/Bip) in an inactive state \[[@CR42]\]. Once ER-stress, a large number of unfolded or misfolded proteins will snatch GRP78 that originally binds to three kinds of responsive proteins, making them exposed and activated. At present, it has been found that GRP78 is closely related to the occurrence and development of liver cancer \[[@CR43]\], breast cancer \[[@CR44]\] and lung cancer \[[@CR45]\]. The expression of GRP78 is increased when ER-stress occurs, which can promote the misfolded or unfolded protein to normal status, thereby reducing the endoplasmic reticulum load and maintaining the stability of the intracellular environment. Glucose regulatory protein 94 (GRP94) is mainly stored in the endoplasmic reticulum and is an important factor in the endoplasmic reticulum stress response. When stress occurs in the endoplasmic reticulum, GRP94 can induce the expression of C/EBP homologous protein, thereby down-regulating Bcl-2 and promoting apoptosis \[[@CR46]\]. This study showed psoralen plays a clear role in promoting the expression of GRP78 and GRP94, which can induce unfolded protein reaction and cause apoptosis. Because PERK is a type I endoplasmic reticulum transmembrane protein, its own activation of phosphorylation also results in phosphorylation of the catalytic substrate eIF2α. Phosphorylated eIF2α can inhibit the synthesis of protein and reduce the endoplasmic reticulum load in a short period of time, but long-term inhibition also leads to apoptosis. Furthermore, it can also promote the translation of activating transcription factor-4 (ATF-4) \[[@CR47]\]. Overexpression of ATF4 can activate the CHOP apoptotic pathway, which can activate GADD34 to cause the damage of oxygen free radicals. ATF6 is a type II transmembrane protein of the endoplasmic reticulum, which is separated from GRP78 and can be activated by S1P and S2P cleavage during ER-stress. Activated ATF6 promotes transcription and expression of CHOP in the nucleus, thereby inhibiting Bcl-2 and causing apoptosis \[[@CR48]\]. This results suggested that psoralen-induced endoplasmic reticulum stress mainly activates the IRE1 pathway and the ATF6 pathway. Once ER-stress occurs, phosphorylated eIF2α or IRE1 protein can activate GADD34 and ATF4. Activated GADD34 dephosphorylates p-eIF2α and restores protein synthesis \[[@CR49]\]. Overexpression of GADD34 and ATF4 induces CHOP expression, inhibits Bcl-2 expression and causes damage of oxygen free radicals which can induce apoptosis. These results showed that psoralen increased GADD34 and ATF4 genes expression, which participating in endoplasmic reticulum stress and promoting apoptosis. C/EBP homologous protein is known as DDIT3, which is in a low expression in physiological state, can activate apoptotic C/EBP homologous protein. During ER-stress, the PERK, IRE1 and ATF6 signaling pathways can promote CHOP expression. However, high expression of CHOP leads to the loss of Ca^2+^ in the endoplasmic reticulum, which increases mitochondrial permeability and induces apoptosis \[[@CR50]\]. On the other hand, it also reduces the expression of anti-apoptotic proteins and increases the expression of pro-apoptotic proteins, finally activating cellular mitochondrial apoptosis pathway \[[@CR51]\]. This study found that psoralen can promote significantly DDIT3 gene expression, but there is no significant effect on the CHOP protein level. Combined with the expression of apoptosis-related proteins Bcl-2 and Bax, it is speculated that psoralen may cause changes in Ca^2+^ concentration in the endoplasmic reticulum through highly expressed DDIT3 which triggers apoptosis. In general, for short-term ER-stress, cells can maintain the stability of the intracellular environment by enhancing the ability of protein folding, inhibiting protein synthesis and translation, and accelerating protein degradation. They are self-protective and belong to normal physiological phenomena. When ER-stress continues, a large amount of unfolded or misfolded protein will accumulate, which will induce unfolded protein response (UPR) and cause apoptosis. This study indicated that psoralen could cause a significant induction of ER-stress in dose- and time-effect relationship, which causing apoptosis of hepatoma cells. In addition, psoralen might regulate some physiological process in HCC by increased relevance mRNA levels on endoplasmic reticulum. Growth differentiation factor 15 (GDF15) is a divergent member of the BMP-subfamily of the TGF-β superfamily, which could inhibit the proliferation, migration and invasion, while promoting apoptosis of A549 cells \[[@CR52]\]. Induction of ER stress leads to upregulation of several genes such as WARS (tryptophanyl-tRNA synthetase), HERP (homocysteine-inducible ER protein with ubiquitin like domain 1), DNAJC3 (also called P58IPK), ER degradation-enhancing alpha-mannosidase-like 1 (EDEM1) and leads to caspase activation, release of mitochondrial intermembrane proteins and dissipation of mitochondrial transmembrane potential (ΔΨm) \[[@CR53]\]. Synovial apoptosis inhibitor 1 (SYVN1), an ER-associated degradation (ERAD) E3 ubiquitin ligase, could inhibit the breast cancer cell growth and metastasis through the miR-96-5p/SYVN1 axis \[[@CR54]\]. Site 1 protease (S1P), SREBP cleavage-activating protein (SCAP) and acetyl-CoA acetyltransferase 1 (ACAT1) were involved in the lipid metabolism \[[@CR55]\]. ER chaperone calreticulin (CALR), protein transport protein Sec61 subunit alpha isoform 1 (SEC61A1), protein disulfide-isomerase A4 (PDIA4) were up-regulated in the ER stress response for binding to misfolded proteins \[[@CR56]--[@CR58]\]. Besides, it is worth noting that TUDC, an ER-stress inhibitor, did not inhibit the expression of GRP78, DDIT3, XBP-1s, IREα, and CHOP, which were increased expression by psoralen. It suggested that psoralen-induced SMMC7721 cells have a strong ER stress response, and the damage to cells is irreversible and is not easily corrected by ER-stress inhibitor. It suggested that the psoralen-induced ER-stress promotes apoptosis of SMMC7721 hepatoma cells with specific characteristics. Conclusion {#Sec10} ========== Psoralen inhibits proliferation of SMMC7721 cells by arresting cells at G1 phase. Moreover, it can also cause the endoplasmic reticulum expansion and dysfunction, so as to induce the ER-stress persistently, leading to the apoptosis of hepatoma cells. The study enriched the pharmacological mechanism of psoralen in the anti-hepatoma effect. Methods {#Sec11} ======= Materials {#Sec12} --------- Psoralen (with the chemical structure illustrated in Fig. [1](#Fig1){ref-type="fig"}) was obtained from the Shanghai Institute for Food and Drug Control (NO. 110739, Shanghai, China; purity: more than 98% HPLC) and dissolved in dimethyl sulfoxide (DMSO, ST038, Biyotime, China). The final concentration of DMSO was 0.1% in all psoralen groups and had no effect on cell viability. Thapsigargin was obtained from Sigma-Aldrich (T9033, St. Louis, MO, USA), TUDC was purchased Selleck Chemicals (S3654, Houston, TX, USA). DMSO and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) were obtained from Sigma-Aldrich (M2128, St. Louis, MO, USA). Fetal bovine serum and RIPM1640 were purchased from Gibco (Grand Island, NY, USA), penicillin and streptomycin and 0.25% Trypsin (0.02% EDTA) were purchased from Hyclone (Marlborough, MA, USA). Trizol was obtained from Invitrogen Life Technologies (Grand Island, NY, USA), PrimeScript^®^RT reagent Kit, SYBR^®^ Premix Ex Taq™ (Tli RNaseH Plus) II were purchased from TaKaRa Bio Inc (Otsu, Japan). RIPA lysis buffer, BCA kit, IgG(H + L) and IgG(H + L) second antibodies were purchased from Beyotime Biotechnology (Shanghai, China), Annexin V-FITC Apoptosis Analysis Kit and PI/RNase Staining Solution were purchased from Sungen Biotech (Tianjing, China). CyclinD1 (ab134175), CyclinE1 (ab33911), CDK4 (ab68266), Bax (ab182733) and Bcl-2 (ab182858) antibodies were purchased from Abcam (Cambridge, MA, USA). GRP78 (3177s), GRP94 (2104s), IRE1α (3294s), XBP-1S (12782s), PERK (3192s) and p-eIF2α (9721s) antibodies were purchased from Cell signaling technology. β-actin (A5441) and p27 (SAB4500067) antibodies were purchased from Sigma-Aldrich (St. Louis, MO, USA). CHOP (NBP2-13172) antibody was purchased from Novus Biologicals (Littleton, CO, USA). Cell culture and treatment {#Sec13} -------------------------- The human hepatoma SMMC7721 cells were cultured in RIPM1640 medium containing 1% penicillin--streptomycin and 10% FBS. According to different experimental purposes, drug treatment was as follows: SMMC7721 cells were plated at a density of 5000 cells/well in 96-well plates with 24 h incubation, and treated with 10--80 μM psoralen for 24 h, 48 h and 72 h, then the proliferation of cells was detected by MTT assay. The cells were plated at a density of 1 × 10^4^ cells/well in 12-well plates with 24 h incubation, and treated with 40 μM psoralen and 0.1 μM thapsigargin for 12 h, 24 h and 48 h, then the cell morphologic change was observed by inverted microscope. The cells were plated into 6-well plates at a density of 1 × 10^5^ cells per well in 6-well plates with 24 h incubation, and treated with 40 μM psoralen and 0.1 μM thapsigargin for 24 h, then the ultrastructure of cells was observed by transmission electron microscope; or the cells were treated with 10--50 μM psoralen for 48 h, then the cell apoptosis and cell cycle were tested by flow cytometry; or the cells were treated with 40 μM psoralen for 6 h, 12 h, 24 h or 48 h, then the expression of mRNA and protein expression were detected by RT-qPCR or Western blotting. Additionally, 0.1% DMSO was as a control group in all experiments. MTT assay {#Sec14} --------- Cell viability was tested using the MTT assay according to the manufacturer's instructions. SMMC7721 cells were plated into 96-well plates at a density of 5000 cells/well for 24 h. Following treatment with psoralen for 24 h to 72 h, the medium was removed and cells were washed with PBS. MTT (0.5 mg/mL) was then added to each well and the mixture was incubated for 4 h at 37 °C. MTT reagent was then replaced with DMSO (150 μL per well) to dissolve formazan crystals. After the mixture was shaken at room temperature for 10 min, absorbance was measured at 490 nm using a microplate reader (Bio-Tek, Winooski, VT, USA). Routine transmission electron microscopy (TEM) of adherent cells {#Sec15} ---------------------------------------------------------------- SMMC7721 cells were plated at a density of 1 × 10^5^ cells/well in 6-well plates and treated with 40 μM psoralen or 0.1 μM Thapsigargin for 24 h. Cells were fixed in situ in 2.5 % glutaraldehyde in 0.05 M phosphate buffer (pH 7.4) for 2 h at 4 °C, then cells were scraped down and to centrifuge for 15 min in 2000 rpm, then discarding the supernatant and resuspending in the 2.5% glutaraldehyde buffer. After fixation, the samples were post-fixed in 1% OsO~4~ for 4 h and dehydrated in increasing concentrations of ethanol and propylene oxide series. Subsequently, the samples were embedded in Epon 618 resin and polymerized for 48 h at 60 °C. Then, the ultrathin sections (90--100 nM) samples were mounted on the copper grids and were stained with 0.5% aqueous uranyl acetate and lead citrate. Subsequently, the grids were air-dried and examined under a transmission electron microscope (Philips Tecnai-12, Netherlands) at 25 kV. Flow cytometry for cell cycle and apoptosis {#Sec16} ------------------------------------------- SMMC7721 cells were plated into 6-well plates at a density of 1 × 10^5^ cells per well in 6-well plates, after 24 h incubation at 37 °C, cells were treated with 40 μM psoralen for 48 h. Cycle arrest and apoptotic cells were detected by flow cytometric analysis. After the treatment, cells were collected by trypsinization and washed twice with PBS. For cell cycle assay, the collected cells were stained with propidium iodide (PI) (Sanjian, Tianjin, China). Cellular apoptosis was determined with Annexin V-FITC and PI using FITC Apoptosis Detection kit (Sanjian, Tianjin, China). Stained cells were assessed by BD FACSalibur flow cytometry (Franklin Lakes, NJ, USA), and the data were analyzed by FlowJo software (TreeStar, Ashland, OR, USA). Quantitative real-time reverse-transcription PCR (qRT-PCR) {#Sec17} ---------------------------------------------------------- Total RNA was extracted using TRIzol (Invitrogen) according to protocols from the manufacturers. The purity and integrity of the RNA were checked spectroscopically using a NanoDrop 2000/c spectrophotometer (Thermo). Then, for each sample, 2 μg RNA was reverse transcribed to obtain the cDNA template using PrimeScript^®^RT reagent Kit (TaKaRa). Each cDNA sample was diluted 5 times for qRT-PCR amplification; qRT-PCR was performed using the fluorescent dye SYBR^®^ Premix Ex Taq™ (Tli RNaseH PlusII, TaKaRa) with a 7500 Fast Real-Time PCR System. Amplification was performed with the following fast time course: 95 °C 30 s, 95 °C 5 s, 65 °C for 30 s for 40 cycles. Relative mRNA expression values were determined by the 2−ΔΔCt method using human β-actin as the normalization control. Western blotting {#Sec18} ---------------- SMMC7721 cells were lysed with RIPA lysis buffer (Beyotime, Haimen, China) supplemented with PMSF. The proteins were separated through 10% SDS-PAGE gels electrophoresis, and transferred to polyvinylidene difluoride membranes, and incubated in block buffer and then incubated with primary anti-bodies, including Bax (1:2000), Bcl-2 (1:2000), CDK4 (1:1000), CyclinD1 (1:30,000), CyclinE1 (1:1000), P27 antibody (1:1000), GRP78 (1:1000), CHOP (1:1000), IRE1α (1:1000), XBP-1S (1:1000), and β-actin antibody (1:20,000), followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibody. Protein expression was visualized by enhanced chemiluminescence assay and signal Signals were detected using a Fluorchem E system (Protein-Simple, USA). Statistical analysis {#Sec19} -------------------- Data were expressed as the mean ± SD. Significant differences were accepted at the 0.05 level of probability and were statistically determined with ANOVA followed by a Newman--Keuls post hoc test using GraphPad Prism6 (San Diego, CA). Additional file =============== {#Sec20} Additional file 1: Figure S1. Effects of psoralen on gene expression of ER-stress in HepG2. The mRNA levels of key gene in ER-stress in the dose of 40 μM psoralen or 0.1 μM thapsigargin for 24 h. Values are means ± SD (n = 3) and \* is means compared to the Con group, \\, p \< 0.01; \\\, p \< 0.001. PSO : psoralen HCC : hepatocellular carcinoma ER : endoplasmic reticulum TG : thapsigargin PERK : protein kinase R-like ER kinase ATF6 : activating transcription factor 6 IRE1 : inositol-requiring enzyme 1 GRP78 : glucose-regulated protein of 78 UPR : unfolded protein response CHOP : C/EBP homologous protein XBP1 : X-box binding protein 1 TUDC : tauroursodeoxycholic acid qRT-PCR : quantitative real-time PCR Publisher\'s Note* Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Xiaomin Wang and Peike Peng contributed equally to this work Not applicable. All the experiments were carried out by XW, the manuscript was written by PP, all data were analyzed by ZP and ZF. We are thankful to WL and XL for providing her valuable inputs for English. All authors read and approved the final manuscript. This work was supported by the National Natural Science Foundation of China (No. 81473562). All data generated or analysed during this study are included in this published article. This study was approved by the Ethics Committee of Shanghai University of Traditional Chinese Medicine, and the written informed consent was also obtained from each participant in advance. Not applicable. The authors declare that they have no competing interests.	{ "pile_set_name": "PubMed Central" }
Glioblastomas (GBM) are the most aggressive tumors accounting for 45.6% of primary malignant brain tumors[@b1]. The standard therapy for GBM comprises surgery followed by radiation therapy with adjuvant chemotherapy[@b2][@b3][@b4]. Despite advances in multimodal therapies, the median survival of patients with GBM is approximately 15 months with a 5-year survival rate of 5.0% after diagnosis[@b5]. The highly infiltrative, heterogeneous and mutable nature of GBM[@b6] often contributes to tumor recurrence and resistance to therapies. Current cytotoxic chemotherapeutic agents used to treat GBM include carmustine, lomustine, and carboplatin[@b2]. A recent trial of combining bevacizumab with standard chemotherapy and radiation therapy only marginally improved overall survival[@b7]. Therapy regimens are currently being developed to target EGFR, VEGFR, PDGFR, Ras pathway, mTOR, histone acetylation and integrins[@b8], and thus far these molecular-targeted therapies have produced poor-to-modest clinical responses[@b6]. Identification of more effective therapeutic agents that work as a single agent or in combination with existing drugs are clearly needed. The incidence rate of GBM and other glial tumors is higher in males than females[@b1]. Females of reproductive ages demonstrate a survival advantage over both males and post-menopausal women. Usage of exogenous hormones reduces the risk of glioma development[@b9][@b10][@b11][@b12]. Estrogen improves the survival in a glioblastoma orthotopic animal model[@b13]. Estrogen mediates its effects through the estrogen receptor α (ERα) and estrogen receptor β (ERβ). ERβ has quite different function than ERα, and ERβ is considered as a tumor suppressor. Recent studies demonstrated that ERβ reduces proliferation and induces apoptosis in several cancer cells[@b14][@b15][@b16][@b17][@b18][@b19][@b20] and that its expression declines during tumor progression[@b21][@b22][@b23][@b24]. Recent studies including ours demonstrated that GBM cells express ERβ, and high expression of ERβ was an independent favorable prognostic factor[@b25][@b26][@b27]. Collectively, these correlative findings suggest that estrogen and ERβ play a significant role in suppression of GBM; however, the mechanisms are poorly understood. Furthermore, estrogen as potential therapy for GBM has limited therapeutic application due to the risk of breast and uterine cancers in women, prostate cancer and feminization in men, and may increase risk of heart disease in both sexes[@b28][@b29][@b30][@b31]. Even though ERα and ERβ are structurally similar, their ligand-binding domains differ enough to be selective for different ligands[@b32]. Recent studies identified a number of selective synthetic and natural ERβ agonists which are currently being investigated for therapeutic use[@b33]. One synthetic compound LY500307 is a highly potent and selective ERβ agonist; it has a 12-fold higher affinity for ERβ than ERα and exhibits 32-fold more functional potency. Further, preclinical studies showed that LY500307 treatment dose dependently reduced the prostate weight in a mouse model of benign prostatic hyperplasia[@b34]. LY500307 was well tolerated in BPH patients with no side effects[@b35] and more importantly, LY500307 is currently being tested in phase 2 clinical trials for improving negative symptoms and cognitive impairment associated with Schizophrenia. However, it remains unknown whether LY500307 has efficacy in treating GBM. Here, we tested the efficacy of the ERβ agonist LY500307 as a novel therapeutic agent for treating GBM using in vitro and in vivo preclinical models. Our results demonstrate that LY500307 selectively kills established and patient-derived primary GBM cells with minimal toxicity on normal cells. Mechanistic studies showed that LY500307 modulates cell cycle, apoptosis and DNA damage response pathways and sensitizes GBM cells to current chemotherapeutic agents. Further, LY500307 reduced GBM growth in vivo in orthotopic models and prolonged the survival of tumor-bearing mice. This represents the first report demonstrating specificity of ERβ ligand LY500307 on GBM cells and suggests that LY500307-- ERβ-mediated inhibition may be an effective strategy for targeted therapy. Results ======= Selective ERβ agonist LY500307 reduces the cell viability and survival, and induces apoptosis of GBM cells ---------------------------------------------------------------------------------------------------------- To test whether LY500307 reduces cell viability of GBM cells, MTT cell viability assays were performed. Treatment with LY500307 significantly reduced the viability of various GBM cell lines in a dose-dependent manner. In contrast, viability of normal astrocytes was not affected at the tested doses, suggesting that LY500307 has tumor cell--specific activity ([Fig. 1a](#f1){ref-type="fig"}). Further the effect of LY500307 on cell viability of several patient derived GBM cells was examined. As shown in [Supplementary Fig. 1a](#S1){ref-type="supplementary-material"}, all patient derived primary GBM cells tested expressed ERβ but not ERα. Cell Titer-Glo luminescent cell viability assays revealed that LY500307 reduced the viability of various patient-derived GBM cells in a dose dependent manner ([Fig. 1b](#f1){ref-type="fig"}). We next examined the effect of LY500307 on survival of GBM cells using colony formation assays. As shown in [Fig. 1c](#f1){ref-type="fig"}, LY500307 significantly reduced the colony formation of U87 and U251 GBM cells. We then tested whether LY500307 induces apoptosis of GBM cells using the Annexin V assay. LY500307 significantly induced Annexin V-positive apoptotic cells in U87, U251 ([Fig. 1d,e](#f1){ref-type="fig"}) and patient derived primary GBM (GBM10) cells ([Supplementary Fig. 1b](#S1){ref-type="supplementary-material"}). Collectively, these results suggested that LY500307 has potential to selectively reduce cell viability, decrease survival and induce apoptosis of GBM cells. LY500307 enhances ERβ signaling in GBM cells -------------------------------------------- To determine whether LY500307 promotes activation of classical ERβ--ERE signaling, U87 GBM cells were transfected with the ERE-luciferase reporter and treated with LY500307 for 24 h. LY500307 significantly increased the ERE luciferase activity in GBM cells ([Fig. 2a](#f2){ref-type="fig"}, left panel). In addition, LY500307 stimulation enhanced the expression of ERβ and its target gene MDA7/IL-24 in GBM cells ([Fig. 2a](#f2){ref-type="fig"}, middle and right panel). To determine whether LY500307 modulates the non-classical ERβ signaling, GBM cells were transfected with AP-1, SP-1 and NF-κB-Luc luciferase reporter plasmids followed by treatment with vehicle or LY500307. As shown in [Fig. 2b](#f2){ref-type="fig"}, LY500307 significantly increased AP-1 and SP-1 luciferase activities and reduced NF-κB-Luc activity in GBM cells. To determine whether LY500307 has an effect on ERβ-mediated rapid extra-nuclear signaling, GBM cells were treated with LY500307 for short periods of time (5, 15, and 30 min) and the phosphorylation status of p38MAPK, JNK, ERK1/2 and Akt was examined. Western blot analysis showed that LY500307 significantly increased the phosphorylation of proapoptotic stress activated kinases p38MAPK and JNK in GBM cells. However, LY500307 treatment reduced phosphorylation of ERK1/2, and Akt phosphorylation was not affected ([Fig. 2c](#f2){ref-type="fig"}). To examine the ERβ selectivity of LY500307, cell viability assays were performed using ERβ knockdown in GBM cells. LY500307-mediated reduction in cell viability was significantly compromised in ERβ shRNA cells than in control shRNA cells, suggesting the specific requirement of ERβ for LY500307 actions ([Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). We also confirmed the ERβ-mediated GBM suppression using GBM cells that overexpress ERβ. The effect of ERβ expression on GBM cell proliferation was studied using Cell Titer-Glo assay. ERβ overexpression significantly reduced the proliferation of U87 and U251 GBM cells ([Supplementary Fig. 3](#S1){ref-type="supplementary-material"}). Collectively, these results suggest that LY500307 has potential to modulate both classical and non-classical ERβ signaling and that LY500307 can enhance ERβ expression in GBM cells. Analysis of LY500307 induced global transcriptional changes in GBM cells ------------------------------------------------------------------------ To identify the global changes in gene expression following LY500307 treatment in GBM cells, we performed global transcriptome analysis. U87 cells were treated with either vehicle or LY500307 for 48 h, and the isolated RNA was subjected to RNA-seq analysis. Genes that had at least a 1.5-fold change in expression (p \< 0.01) were chosen for analysis. Overall, 3204 genes were differentially expressed in LY500307-treated U87 cells; 1568 genes were upregulated and 1636 genes were downregulated. A representative heat map was shown in [Fig. 3a](#f3){ref-type="fig"}. The complete list of differentially expressed genes is given in [Supplementary Table 1](#S1){ref-type="supplementary-material"}. To further examine the biological significance of the differentially expressed genes, IPA analysis was performed. The top networks that were differentially expressed after LY500307 treatment were related to tissue development, cell cycle, cellular movement, cellular assembly and organization, cellular function and maintenance ([Fig. 3b](#f3){ref-type="fig"}). Analysis of molecular and cellular functions of differentially expressed genes revealed that they are involved in cellular growth, proliferation, cell death and survival ([Supplementary Fig. 4](#S1){ref-type="supplementary-material"}). The canonical pathways that were modulated by LY500307 include cell cycle, DNA damage, p53 signaling, and checkpoint regulation ([Fig. 3c](#f3){ref-type="fig"}). In addition, pathways related to cancer such as molecular mechanism of cancer, GBM signaling, Wnt signaling and glioma invasion were also altered. Ability of LY500307 to modulate these pathways were independently validated by using qRT-PCR on selective genes in U87 ([Fig. 3d,e](#f3){ref-type="fig"}), U251 ([Fig. 3f](#f3){ref-type="fig"}) and patient derived GBM (GBM10) cells ([Supplementary Fig. 5](#S1){ref-type="supplementary-material"}). Collectively, these results suggest that LY500307 modulates the expression of genes involved in cell cycle, cell death, survival, and DNA damage response. LY500307 induces G2/M cell cycle arrest --------------------------------------- Since RNA-seq results demonstrated cell cycle as the top network modulated by LY500307, we examined the effect of LY500307 on cell cycle distribution of GBM cells. Flow cytometry analysis of PI-stained cells revealed that LY500307 treatment significantly increased the percentage of cells in G2/M phase in U87, U251, LN229 and GBM10 cells when compared to vehicle ([Fig. 4](#f4){ref-type="fig"}), further supporting the RNA-seq pathway analysis of cell cycle and G2/M DNA damage checkpoint regulation. LY500307 sensitizes GBM cells to chemotherapeutic agents -------------------------------------------------------- Emerging data provide the evidence that using a multi-targeted approach has an advantage over using a single agent for GBM therapy, and that sensitizer drugs that enhance the utility of chemotherapy are advantageous. We performed an initial in vitro screen of 119 FDA-approved drugs in combination with LY500307 on the cell viability of U87 cells. Our results demonstrated that LY500307 sensitized U87 cells to several FDA approved drugs and nine drugs that showed synergism with LY500307 are presented in [Fig. 5a](#f5){ref-type="fig"}. Interestingly, many of the chemotherapeutic agents that LY500307 sensitized cause DNA damage and apoptosis. We independently validated some of the potent compounds from the initial screen including cisplatin, bleomycin. We also tested potent GBM chemotherapeutic agents including lomustine, temozolomide. Our results revealed that LY500307 sensitized GBM cells to cisplatin, bleomycin and lomustine ([Fig. 5b--d](#f5){ref-type="fig"}). Importantly, we also found that LY500307 significantly sensitized temozolomide-resistant cells U138 to temozolomide-mediated reduction in cell viability ([Fig. 5e](#f5){ref-type="fig"}). LY500307 reduces GBM progression in an orthotopic model ------------------------------------------------------- To evaluate the effect of LY500307 on in vivo tumor growth, U251 cells that express luciferase reporter were injected orthotopically into the brain of mice. After tumors were established, the mice were randomized into a control group, which received vehicle, and a treatment group, which received LY500307 through oral gavage. GBM progression was measured by monitoring luciferase intensity using Xenogen-IVIS imaging system weekly. Compared to vehicle, LY500037 treatment significantly reduced the GBM progression in the U251 tumor model ([Fig. 6a](#f6){ref-type="fig"}). Immunohistochemical analysis revealed that LY500307 treatment significantly reduced the number of proliferation marker Ki-67-positive cells ([Fig. 6b](#f6){ref-type="fig"}) and increased the number of TUNEL-positive apoptotic cells ([Fig. 6c](#f6){ref-type="fig"}). To further confirm the effect of LY500307 on apoptosis, the expression of Cleaved Caspase-3 and Bcl-2 was examined. As shown in [Supplementary Fig. 6a](#S1){ref-type="supplementary-material"}, LY500307 treatment significantly increased the expression of Cleaved Caspase-3 and reduced the Bcl-2 expression in tumors compared to controls. Collectively, these results suggest that LY500307 inhibits the GBM progression in vivo and induces apoptosis. LY500307 increases survival in syngeneic glioma mouse model ----------------------------------------------------------- Immune effects during tumor progression play key roles in GBM progression, and GBM-mediated immunosuppression acts as a barrier to the efficacy of chemotherapeutic agents[@b36][@b37]. To study the anti-tumor activity of LY500307 in the presence of an intact immune system, GL26 cells stably expressing luciferase were implanted into C57BL/6 mice. After establishment of tumors, the mice were randomized into two groups based on luciferase intensity and treated with either vehicle or LY500307 daily by oral gavage. Survival was calculated using Kaplan Meier analysis. Compared to vehicle, LY500307 treatment significantly increased the overall survival of the mice ([Fig. 7a](#f7){ref-type="fig"}). Further, immunohistochemical analysis of tumor sections revealed that LY500307 treatment significantly reduced the expression of the proliferation marker Ki-67 and increased the number of TUNEL-positive apoptotic cells ([Fig. 7b,c](#f7){ref-type="fig"}). Further, the expression of Cleaved Caspase-3 was increased and Bcl-2 expression was decreased significantly in LY500307 treated tumors compared to controls ([Supplementary Fig. 6b](#S1){ref-type="supplementary-material"}). Discussion ========== ERβ was initially discovered as second receptor of estrogen[@b38], and several studies demonstrated that ERβ expression was downregulated during tumor progression. Ligands that increase ERβ expression or activity will have therapeutic utility. Recently, several groups including ours showed that ERβ expression is reduced during the progression of gliomas and that plant-derived ligands of ERβ exhibit anti-tumor activities[@b25][@b39][@b40]. However, plant-derived ligands have low efficacy and are difficult to synthesize in large quantities. Thus, more potent synthetic ERβ ligands that work more effectively are urgently needed for clinical application. Our work provides the evidence that LY500307, a synthetic ERβ agonist has potential to specifically reduce the proliferation of GBM cells with high potency, induce apoptosis, promote G2/M cell cycle arrest and sensitize GBM cells to chemotherapeutic agents. Further, our results also demonstrated that LY500307 has the ability to reduce GBM progression in vivo and improve the survival of mice. Selective killing of cancer cells without affecting normal cells is critical for successful chemotherapy. Our results demonstrated that LY500307 selectively reduces the viability of GBM cells with little effect on normal astrocytes. Furthermore, our in vivo studies showed no LY500307-related toxicities in various organs (data not shown). Previous studies demonstrated that overexpression of ERβ or activation with ligands results in a decrease in proliferation and induction of apoptosis of cancer cells, and that depending on the cell type, activation of ERβ can promote either G~2~ or G~1~ arrest [@b14][@b41]. Our results suggested that activation of ERβ with LY500307 induces apoptosis of GBM cells. Earlier studies showed that overexpression of ERβ or activation with ligands causes G2/M arrest in several cancer cells, and that ERβ increases the expression of several genes, including p21 and GADD45A[@b14][@b41][@b42][@b43]. Our results demonstrate that LY500307 treatment causes cell cycle arrest of GBM cells in G2/M phase and increases the expression of p21 and GADD45A in GBM cells also confirms the earlier observations in other model cells. ERβ functions as a transcription factor that modulates both classical estrogen response element (ERE)--containing genes and non-ERE genes via non-classical signaling by interacting with AP1, SP1, NF-κB and KLF5 transcription factors[@b44][@b45][@b46]. Our RNA-seq analysis revealed that LY500307 modulates several genes involved in cell cycle regulation, apoptosis and DNA damage response. Further, LY500307 treatment also down regulated several genes involved in tumor development and progression including Wnt signaling, GBM signaling and glioma invasion signaling. Since ERβ regulates multiple pathways directly or indirectly by binding to the key mediators, it is not surprising to observe the modulation of multiple pathways by LY500307 that includes induction of cell cycle arrest and cell death molecules and suppression of oncogenic signaling molecules. In addition, our results showed that LY500307 promotes ERβ extra-nuclear signaling, leading to activation of the proapoptotic p38 and JNK pathways. Our results showing ERβ knockdown reduced the efficacy of LY500307 to inhibit GBM proliferation further confirms that specific role of ERβ in LY500307 actions. The current treatment options for GBM are poor, and the mortality rates are very high. Multiple challenges remain, including difficulty of complete resection of GBM, rapid and aggressive tumor relapse and resistance to external radiation and chemotherapy. Therefore, sensitizer drugs that enhance the utility of chemotherapy are urgently needed. Recent studies showed that selective targeting of ERβ with agonists can sensitize malignant pleural mesothelial cells to cisplatin toxicity[@b19] and that inhibition of ERβ increased DNA repair that contributes to cisplatin resistance in medulloblastoma cells[@b47]. Further, ERβ expression influences the malignant pleural mesothelial cell responsiveness to gefitinib[@b19]. Our results using 119 FDA-approved drugs in combination with a low dose of LY500307 revealed sensitization of GBM cells to several chemotherapeutic agents, including axitinib, doxorubicin, cisplatin, bleomycin and etoposide. Our RNA-seq studies revealed potential of ERβ agonists to down-regulate a number of genes involved in DNA repair and DNA damage response. Most importantly downregulation of DNA repair genes such as DDR1 and DDR2 may provide mechanistic explanation for LY500307-mediated sensitization effects. Further, our studies also revealed that LY500307 sensitizes GBM cells to currently used chemotherapeutic agents temozolomide and lomustine. ERβ agonists' ability to suppress pathways involved in DNA repair can be exploited in future to promote apoptosis of GBM cells. We previously reported that the plant-derived ERβ agonist liquiritigenin reduces the growth of subcutaneous glioma xenograft tumors[@b25]. Recently, salicylketoxime-based estrogen receptor β agonists also reduced the glioma growth in subcutaneous models[@b48]. However, lack of drug testing using an orthotopic model is a limitation of these studies. Further, the tumor microenvironment and the presence of intact immune system must be considered in efficacy testing of chemotherapeutic drug on GBM growth. In this study, we have tested the effect of LY500307 using both orthotopic tumor models and syngeneic model with intact immune system. Our results demonstrate that the selective ERβ agonist LY500307 reduced GBM progression as well as enhanced survival in syngeneic mouse models. Since LY500307 readily crosses the blood--brain barrier, it is currently being tested in clinical trials and it is well tolerated, it can be readily transferred to clinical use with current chemotherapies, thereby providing an additional tool for enhancing survival in GBM patients with limited toxicity. Materials and Methods ===================== Cell culture, reagents and generation of stable ERβ overexpression and ERβ shRNA cells -------------------------------------------------------------------------------------- Human glioblastoma cell lines U87, U251, LN229 and T98G and the mouse glioma cell line GL26 were obtained from the American Type Culture Collection (ATCC). Cell lines were maintained in DMEM supplemented with 10% fetal bovine serum (Sigma Chemical Co, St. Louis, MO). LY500307 was purchased from Selleckchem (Houston, TX) and Apex Biosciences (Durham, NC). ERβ antibodies were obtained from Santa Cruz Biotechnology (Dallas, TX),GeneTex (Irvine, CA) and Millipore (Billerica, MA), Ki-67 was obtained from Abcam (Cambridge, MA) and Bcl-2 was purchased from Dako (Carpinteria, CA) and Santa Cruz Biotechnology (Dallas, TX). The p-ERK1/2, ERK1/2, p-AKT, AKT, p-p38MAPK, p38MAPK, p-JNK,JNK, Cleaved Caspase-3 and GAPDH antibodies were obtained from Cell Signaling Technology (Beverly, MA). ERβ-specific short hairpin RNA (shRNA) lentiviral plasmids, β-actin and all secondary antibodies were purchased Sigma from Chemical Co (St. Louis, MO). Normal human astrocytes were obtained from ScienCell Research Laboratories (Carlsbad, CA) and maintained in astrocyte medium. Stably ERβ-expressing GBM cells were generated by infecting them with pLenti6/V5-D-FLAG ERβ and empty control vectors and positive cells were selected with blasticidine (5 μg/mL). GBM cells stably expressing ERβ-shRNA were generated by infecting cells with human-specific lentiviral ERβ-shRNA particles and selected with puromycin (1 μg/mL), and pooled clones were used for all the studies. Lentiviral particles expressing nontargeted shRNA were used to generate control. Primary GBM cells ----------------- Primary GBM cells were isolated from discarded GBM tumor specimens after approval from Office of the Institutional Review Board (IRB) at UT Health Science Center at San Antonio. All the discarded tumor specimens were obtained from patients undergoing surgery after informed consent, and no clinical linkers or codes for the specimens accessible to any research personnel or the PI. All the methods were carried out in accordance with IRB approved guidelines. Briefly, fresh tumor tissue was collected intraoperatively, dispersed into single cells, and cultured briefly in a neurobasal media for expansion and purification followed by intracranial injection in nude mice. Patient-derived GBM neurosphere lines were maintained in Neurobasal Medium supplemented with B27 serum-free supplement and growth factors EGF (20 ng/mL), bFGF (20 ng/mL) and LIF (10 ng/mL) as described[@b49]. The resulting tumors were characterized for consistency by histology with the primary tumor, growth attributes including time to animal demise, and for gene expression profiling for the molecular subtype. GBM line 101310 (GBM10) used in the assays is grade IV GBM, MGMT hypomethylated and average survival time in mice is 108 days. Cell viability and colony formation assays ------------------------------------------ The effect of LY500307 on cell viability of GBM cells and normal astrocytes was assessed by using MTT assays. For LY500307 treatment, GBM cells were maintained in phenol red--free DMEM containing 5% dextran-charcoal treated FBS. Briefly, GBM cell lines and normal astrocytes were seeded in 96 well plates (2 × 10^3^ cells/well). After overnight incubation, cells were treated with either vehicle or varying concentrations of LY500307 for 72 h. MTT was added to each well and incubated for 4 h. Formazan crystals were solubilized in DMSO and the optical density was measured using micro-plate reader. The effect of LY500307 on the viability of patient-derived primary GBM lines and the cell proliferation rates of control and ERβ-overexpressing GBM cells were measured using CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI) in 96-well, flat, clear-bottom, opaque-wall microplates according to manufacturer's instructions. Total ATP content as an estimate of total number of viable cells was measured on an automatic Fluoroskan Luminometer. For colony formation assays, U87 and U251 cells (500 cells/well) were seeded in 6 well plates. After overnight incubation, the cells were treated with vehicle or LY500307 (5 μM) for 72 h. Then, cells were washed with PBS and the cells were allowed to grow for additional 7 days. The cells were fixed in ice-cold methanol and stained with 0.5% crystal violet solution. Colonies that contained ≥50 cells were counted. Annexin V apoptosis assay ------------------------- The apoptotic effect of LY500307 on GBM cells was analyzed using the Annexin V/PI kit as per the manufacturer's instructions (BioLegend, San Diego, CA). Briefly, U87 and U251 GBM cells were seeded in 60-mm culture plates and treated with either vehicle or LY500307 (5 μM) for 48 h. Cells were harvested at a density of 2.5 × 10^6^ cells/mL in Annexin V binding buffer and 100 μL of cell suspension was incubated with Annexin V FITC and propidium Iodide (PI) for 15 min at room temperature in the dark. Then, 400 μL of Annexin V binding buffer was added to each sample and stained cells were analyzed using flow cytometry. Cell cycle analysis ------------------- U87, U251, LN229 and GBM10 (101310) cells were seeded in 100-mm culture plates, and after overnight incubation, cells were treated with either vehicle (0.1% DMSO) or LY500307 (5 μM) for 24 h. Cells were then trypsinized and harvested in PBS, followed by fixation in ice-cold 70% ethanol for 30 min at 4 °C. Cells were washed again with PBS and stained with a mixture of 50 μg/mL propidium iodide and 50 μg/mL RNase A. The PI-stained cells were subjected to flow cytometry using a FACS Calibur (BD Biosciences). Reporter gene assays -------------------- GBM cells were maintained in phenol red--free DMEM supplemented with 5% dextran-charcoal--stripped serum for 48 h prior to transfection. Cells were transiently transfected with 250 ng of ERE-Luc, AP1-Luc, SP1-Luc, or NF-κB-Luc reporter plasmids using Turbofect transfection reagent (Thermo Scientific, Waltham, MA). After 24 h, cells were treated with either vehicle or LY500307 (1 μM) for additional 24 h. β-galactosidase reporter (50 ng) plasmid was co-transfected and used for data normalization. Cells were lysed in Passive Lysis Buffer, and Luciferase activity was measured using the luciferase assay system (Promega, Madison, WI) in luminometer. Western blotting ---------------- Whole cell lysates were prepared from GBM cells using RIPA buffer (Sigma, St Louis, MO) containing protease and phosphatase inhibitors. Total proteins (50 μg) were mixed with SDS sample buffer and separated on SDS-PAGE gels. Resolved proteins were then transferred onto nitrocellulose membranes, blocked with 5% non-fat dry milk powder for 1 h at room temperature and incubated with respective primary antibodies over night at 4 °C followed by secondary antibody incubation for 1 h at room temperature. Blots were developed using the ECL kit (Thermo Scientific, Waltham, MA). In vitro screening of FDA-approved drugs in combination with LY500307 on GBM cells ------------------------------------------------------------------------------------ All the FDA-approved anti-cancerous drugs were obtained from NCI at 10 mM concentration (<https://dtp.cancer.gov/organization/dscb/obtaining/available_plates.htm>). The initial screen of the combination treatment of the 119 FDA-approved drugs and LY500307 was performed on U87 cells using MTT assay. U87 cells (1 × 10^3^/well) were seeded in 96 well plates. After an overnight incubation, the cells were treated with LY500307 (1 μM) and varying doses of each FDA-approved drug at final concentrations of 0.1 μM, 1 μM, and 10 μM for 72 h in phenol red--free DMEM that contains dextran-charcoal treated FBS. Some of the compounds that sensitized GBM cells with LY500307 were further validated using U251 cells. U251 cells were seeded in 96 well plates (1 × 10^3^ cells/well) in phenol red--free DMEM containing dextran-charcoal treated FBS. After an overnight incubation, the cells were pretreated with vehicle or LY500307 (1 μM) for 48 h followed by wash-off and replenished with varying concentrations of each drug alone or in combination with LY500307 for 96 h. Cell viability was then measured using the MTT assay as described above. The temozolomide-resistant cell line U138 was subjected to pretreatment with LY500307 for 72 h followed by varying doses of temozolomide for 5 days, and then cell viability was measured using the MTT assay. RNA sequencing and qRT-PCR -------------------------- U87 GBM cells were treated with either vehicle or LY500307 (5 μM) for 48 h, and total RNA was isolated using RNAesy mini kit (Qiagen) according to the manufacturer's instructions. The purity of prepared RNA was determined using Agilent 2100 BioAnalyzer. Illumina TruSeq RNA Sample preparation was performed following manufacturer's protocol, and the samples were run on an Illumina HiSeq 2000 in duplicates. The combined raw reads were aligned to UCSC hg19, and the genes were annotated by Tophat. Genes were annotated and quantified by HTSeq-DESeq pipeline. Differential expression analysis was performed by DEseq and significant genes with at least 1.5-fold change with p \< 0.01 were chosen for analysis. The interpretation of biological pathways using RNA-seq data was performed with ingenuity pathway analysis (IPA) software using all significant and differentially expressed genes. To validate the selected genes, reverse transcription (RT) reactions were performed by using SuperScript III First Strand kit (Invitrogen, Carlsbad), according to the manufacturer's protocol. Real-time PCR was done using SYBR Green (Thermo Scientific) on an Illumina Real-Time PCR system with the following primers. Results were normalized to β-actin transcript levels, and the difference in fold expression was calculated using delta-delta-CT method. Actin-F-5′- GTGGGCATGGGTCAGAAG-3′; Actin-R-5′- TCCATCACGATGCCAGTG-3′ ESR2-F-5′- GGC AGA GGA CAG TAA AAG CA -3′; ESR2-R-5′- GGA CCA CAC AGC AGA AAG AT -3′ IL24-F-5′- CTTTGTTCTCATCGTGTCACAAC-3′; IL24-R-5′- TCCAACTGTTTGAATGCTCTCC-3′ CDKN1A-F-5′- CTGGAGACTCTCAGGGTCGAAA-3′; CDKN1A-R-5′-GATTAGGGCTTCCTCTTGGAGAA-3′;GADD45A-F-5′- GGTGTACGAAGCGGCCAA-3′; GADD45A-R-5′- GCAGGCACAACACCACGTTA-3′ PUMA-F-5′- ATGCCTGCCTCACCTTCATC-3′; PUMA-R-5′- TCACACGTCGCTCTCTCTAAACC-3′ CTNNB1-F-5′- AAAATGGCAGTGCGTTTAG-3′; CTNNB1-R-5′- TTTGAAGGCAGTCTGTCGTA-3′ SFN-F-5′- TGACGACAAGAAGCGCATCAT-3′; SFN-R-5′- GTAGTGGAAGACGGAAAAGTTCA-3′ AKT3-F-5′-TGTGGATTTACCTTATCCCCTCA-3′; AKT3-R-5′-GTTTGGCTTTGGTCGTTCTGT-3′ PML-F-5′- CGCCCTGGATAACGTCTTTTT-3′; PML-R-5′- CTCGCACTCAAAGCACCAGA-3′ CHEK1-F-5′- ATATGAAGCGTGCCGTAGACT-3′; CHEK1-R-5′- TGCCTATGTCTGGCTCTATTCTG-3′ CHEK2-F-5′- TGAGAACCTTATGTGGAACCCC-3′; CHEK2-R-5′- ACAGCACGGTTATACCCAGC-3′ GADD45B-F-5′- TACGAGTCGGCCAAGTTGATG-3′; GADD45B-R-5′- GGATGAGCGTGAAGTGGATTT-3′ WT1-F-5′- CACAGCACAGGGTACGAGAG-3′; WT1-R-5′- CAAGAGTCGGGGCTACTCCA-3′ WNT5A-F-5′- ATTCTTGGTGGTCGCTAGGTA-3′; WNT5A-R-5′- CGCCTTCTCCGATGTACTGC-3′ SOS2-F-5′- ATGTAGAGGAGCGAGTTCAGAA-3′; SOS2-R-5′- ATGGTAGTCCACTTTGTACCCT-3′ FZD5-F-5′- CATGCCCAACCAGTTCAACC-3′; FZD5-R-5′- CGGCGAGCATTGGATCTCC-3′ FZD7-F-5′- GTGCCAACGGCCTGATGTA-3′; FZD7-R-5′- AGGTGAGAACGGTAAAGAGCG-3′ NTRK3-F-5′- ACGAGAGGGTGACAATGCTG-3′; NTRK3-R-5′- CCAGTGACTATCCAGTCCACA-3′ DDR1-F-5′- AAGGGACATTTTGATCCTGCC-3′; DDR1-R-5′- CCTTGGGAAACACCGACCC-3′ DDR2-F-5′- GCTATATGCCGCTATCCTCTGG-3′; DDR2-R-5′- ACTCTGACCACTGACTGGAAG-3′ MAGI2-F-5′- TCCGGCTCAAGTGTGTCAAG-3′; MAGI2-R-5′- AGGTTGTCACGAATGATTTGCT-3′ MRAS-F-5′ TTCCTCATCGTCTACTCCGTC-3′; MRAS-R-5′ AGGATCATCGGGAATGACTCC-3′ Immunohistochemistry -------------------- Immunohistochemical studies were performed as described previously[@b25]. Coronal brain sections were incubated in xylene and passed through series of graded alcohols and then subjected to antigen retrieval using the antigen retrieval solution (Vector Lab, Inc. CA). Tissue sections were incubated in 3% H~2~O~2~ solution for 20 min and then subjected to blocking using the vector lab blocking kit. Tissue sections were incubated overnight with Ki-67 (1:100), Cleaved Caspase-3 (1:200) and Bcl-2 (1:50) and then with secondary antibodies for 45 min at room temperature. Immunoreactivity was visualized by using the DAB substrate and counterstained with haematoxylin (Vector Lab, Inc. CA). The proliferative index was calculated as percentage of Ki-67-positive cells in five randomly selected microscopic fields at 20X per slide. TUNEL analysis was performed using the In situ Cell Death Detection Kit (Roche, Indianapolis, IN) as per the manufacturer's protocol, and five randomly selected microscopic fields in each group were used to calculate the relative ratio of TUNEL-positive cells. DAPI was used to visualize the nuclei. In Vivo Orthotopic Tumor Models --------------------------------- All animal experiments were performed after obtaining UTHSCSA IACUC approval, and all the methods were carried out in accordance with the IACUC approved guidelines. U251 and GL26 cells labelled with GFP-Luciferase (1 × 10^6^) were injected orthotopically into the mouse brain at 2 mm right, 2 mm posterior and 3 mm deep from bregma into the striatum. GL26 mice were ovariectomized bilaterally to avoid the effects of circulated estrogen. Five days after injection, the mice were imaged using Xenogen IVIS system and randomized to receive either control or treatment. The control group received vehicle (30% captisol) and the treatment group received LY500307 in 30% captisol (5 mg/Kg body weight/day) orally. Tumor growth was measured weekly using Xenogen IVIS system. After treatment, the mice were euthanized, and their brains were isolated and processed for histological studies. Statistical analyses -------------------- Statistical differences between groups were analyzed with either t-test or ANOVA as appropriate using GraphPad Prism 6 software (GraphPad Software, SanDiego, CA). Student t-test was used to assess the statistical difference between control and LY500307-treated groups. Mice survival was determined using Kaplan-Meier survival curve. All the data represented in bar graphs are shown as means ± SE. A value of p \< 0.05 was considered as statistically significant. RNA-seq data was analyzed using IPA software. Additional Information ====================== How to cite this article: Sareddy, G. R. et al. Selective Estrogen Receptor β Agonist LY500307 as a Novel Therapeutic Agent for Glioblastoma. Sci. Rep. 6, 24185; doi: 10.1038/srep24185 (2016). Supplementary Material {#S1} ====================== ###### Supplementary Information This study was supported by the NIH/NCI grant NIH-CA178499 (R.K.V; A.B.); ABTA Discovery Grant (GRS), CPRIT training grant (RP140105) (GRS) and the Cancer Therapy and Research Center at the University of Texas Health Science Center at San Antonio through the NCI Cancer Center Support Grant P30CA054174-17. Jan-Ake Gustafsson thanks the Robert A Welch Foundation (grant number E-0004) for support. Author Contributions G.R.S., A.B., R.R.T. and R.K.V. designed the experiments; G.R.S., X.L. and S.V. performed in vitro and IHC studies; G.R.S., J.L., X.L., L.G., A.G., D.C. and M.G. performed Animal experiments and imaging studies; A.M.S. and J.G. provided reagents and interpreted the data; G.R.S., A.B., R.R.T. and R.K.V. analyzed data; and G.R.S. and R.K.V. wrote the paper. All authors reviewed the manuscript. ![ERβ agonist L500307 reduces proliferation and induces apoptosis in GBM cells.\ (A) U87, U251, T98G and normal astrocytes were treated with either vehicle or LY500307 for 72 h and the cell viability was measured using an MTT assay. (B) Patient-derived primary GBM cells were treated with either vehicle or LY500307 for 72 h and the cell viability was measured using the CellTiter-Glo luminescent assay. (C) U87 and U251 cells were treated with either vehicle or LY500307 for 72 h and then cultured for 7 subsequent days. The number of colonies for each group was counted. (D--E) U87 and U251 cells were treated with either vehicle or LY500307 for 48 h followed by Annexin V-FITC and Propidium Iodide (PI) staining for 15 min. The Annexin V--positive apoptotic populations were determined using flow cytometry. Data are represented as mean ± SE. \p \< 0.05, \\p \< 0.01, \\\p \< 0.001.](srep24185-f1){#f1} ![ERβ agonist LY500307 activates ERβ signaling in GBM cells.\ (A) U87 GBM cells were transfected with ERE-Luc plasmid. After 24 h, the cells were treated with either vehicle or LY500307 for additional 24 h, and then reporter activity was measured (left panel). U87 cells were treated with vehicle or LY500307 for 24 h, and the mRNA expression of ERβ (ESR2) and MDA7/IL-24 was measured using qRT-PCR (middle and right panels). (B) U87 cells were transfected with AP-1 luc, SP-1 luc and NF-kB luc plasmids and 24 h after transfection, the cells were treated with either vehicle or LY500307 for 24 h and reporter activity was then measured. (C) U87 cells were treated with either vehicle or LY500307 for 5 min, 15 min and 30 min and lysed in RIPA buffer. The lysates were subjected to Western blotting with the indicated antibodies, and β-actin used as loading control. Data are represented as mean ± SE. \p \< 0.05, \\p \< 0.01, \\\p \< 0.001.](srep24185-f2){#f2} ![Analysis of global transcriptome changes modulated by LY500307.\ Total RNA was isolated from U87 cells that were treated with either vehicle or LY500307 for 48 h using RNeasy mini kit and the quality of RNA was tested using a Bio-analyzer and subjected to RNA sequencing. (A) heat map of differentially expressed genes between vehicle and LY500307 is shown. (B) Differentially expressed genes were subjected to pathway analysis using IPA software and the top five associated network functions of differentially expressed genes are shown. (C) The selected top canonical pathways are shown. (D--E) U87 cells that were treated with either vehicle or LY500307 for 48 h and differentially expressed genes that were upregulated or downregulated were validated using qRT-PCR. F, U251 cells were treated with either vehicle or LY500307 for 48 h and differentially expressed genes were validated using qRT-PCR.](srep24185-f3){#f3} ![LY500307 promotes G2/M accumulation of GBM cells.\ U87, U251, LN229 and primary GBM10 cells were treated with either vehicle or LY500307 for 24 h, and the cells were fixed in 70% ethanol and subjected to PI staining for 20 min. Cell cycle distribution was analyzed using flow cytometry. Data are represented as mean ± SE. \\p \< 0.01, \\\p \< 0.001.](srep24185-f4){#f4} ![LY500307 sensitizes GBM cells to chemotherapeutic agents.\ (A) U87 GBM cells were treated with either vehicle or LY500307 for 24 h and then treated with one of 119 different FDA-approved drugs for an additional 72 h. Cell viability was determined using an MTT assay, and the list of the drugs that had synergistic activity in the presence of LY500307 are represented in [Fig. 5A](#f5){ref-type="fig"}. U251 GBM cells were pretreated with LY500307 for 48 h followed by treatment with varying doses of cytotoxic drugs cisplatin (B) lomustine (C) or bleomycin (D) for an additional 96 h. Cell viability was determined using MTT assay. (E) U138 cells were pretreated with LY500307 for 72 h followed by treatment with varying doses of temozolomide for additional 5 days. Cell viability was determined using an MTT assay.](srep24185-f5){#f5} ![LY500307 reduces GBM progression and induces apoptosis in vivo.\ (A) U251 GBM-luc cells were implanted intracranially into the right striatum of nude mice. The mice were then randomized to either the control or the treatment group and received either vehicle or LY500307 (5 mg/kg body weight/day), respectively, for 28 days. Tumor growth in terms of luciferase intensity was measured using Xenogen IVIS imaging (n = 5). (B) Mouse brains collected from both the control and LY500307-treated mice were fixed in formalin and subjected to immunohistochemical staining for Ki-67. For quantitation, Ki-67-positive cells from five different fields were counted and plotted as histogram. (C) Brain sections were subjected to a TUNEL assay, and the number of TUNEL-positive cells was counted in five different fields and plotted as histogram. DAPI was used to visualize the nuclei. Data are represented as mean ± SE. \p \< 0.05, \\p \< 0.01, \\\p \< 0.001.](srep24185-f6){#f6} ![LY500307 prolongs survival of tumor-bearing mice in a syngeneic glioma model.\ (A) GL26 glioma cells were implanted orthotopically into C57BL6 mice that were treated with either vehicle or LY500307 (5 mg/kg body weight/day) orally. Survival of the mice was plotted using Kaplan-Meier curve (n = 4). (B) Mouse brains collected from the control and LY500307-treated mice were fixed in formalin and processed for immunohistochemical staining for Ki-67. The number of Ki-67--positive cells from five different fields were counted and plotted as histogram. (C) Brain sections were subjected to a TUNEL assay, and number of TUNEL-positive cells was counted for five different fields and plotted as histogram. DAPI was used to visualize the nuclei. Data are represented as mean ± SE. \\p \< 0.01, \\\p \< 0.001.](srep24185-f7){#f7}	{ "pile_set_name": "PubMed Central" }
Introduction {#S1} ============ All organisms depend on ring- and spiral-shaped ATPase assemblies to carry out essential processes ranging from proteolysis and membrane trafficking, to signaling events and nucleic acid transactions. DNA replication onset in cells reflects one such process, employing ATP-dependent initiation factors to coordinate replisome assembly^[@R1],\ [@R2]^. Replication initiators of eukaryotes and prokaryotes contain AAA+-family ATPase domains, whose activity is augmented by duplex DNA-binding domains (DBDs) and specialized protein-protein interaction elements that assist with origin recognition and recruit specific replication factors^[@R3],\ [@R4]^. AAA+ enzymes share a common structural core with RecA-type ATPases, together forming the [A]{.ul}dditional [S]{.ul}trand [C]{.ul}atalytic glutamat[E]{.ul} (ASCE) supergroup of P-loop NTPases^[@R5]^; the molecular logic that allows a common nucleotidyl-hydrolase module to control the disparate activities of replication initiators, and ASCE proteins in general, is not understood. In bacteria, replication initiation relies on the DnaA protein^[@R6]--[@R8]^. In Escherichia coli, multiple DnaA molecules bind to the replication origin, oriC, through several duplex DNA-binding sites, forming a large nucleoprotein complex in the presence of ATP^[@R9]--[@R11]^. With the aid of of appropriate architectural proteins (such as Integration Host Factor) and negatively-supercoiled DNA, this complex subsequently melts an AT-rich, DNA-unwinding element (DUE) located adjacent to the duplex DnaA binding sites^[@R12],\ [@R13]^. ATP also activates a secondary DNA-binding site within DnaA, postulated to reside within the AAA+ domain, which engages single-stranded regions of the DUE to form a stable open complex^[@R12],\ [@R14]--[@R16]^. DnaA then collaborates with the bacterial helicase loader (DnaC in E. coli), to recruit two hexamers of the DnaB helicase to the origin and promote replisome assembly^[@R17]--[@R19]^. Although most AAA+ enzymes form closed-ring assemblies^[@R20],\ [@R21]^, structural studies have indicated that initiators and polymerase clamp-loaders form openring structures^[@R14],\ [@R22]--[@R24]^. Among initiator/loader systems, DnaA is particularly unusual in that it has been seen to oligomerize into a right-handed, spiral filament^[@R14]^. Two models have been proposed to explain how this structure might aid origin melting ([Fig S1](#SD1){ref-type="supplementary-material"}). In one, the wrapping of duplex DNA about a DnaA super-helix would constrain a positive supercoil, generating compensatory negative writhe that could aid opening of the neighboring DUE. In the other, the wrapped DnaA/DNA complex would serve as a nucleation center, allowing DnaA protomers to directly engage and melt the DUE, possibly through the initiator\'s ATPase elements. Thus far, experimental evidence has supported both models^[@R9],\ [@R14]--[@R9],\ [@R25]^, leaving open the question as to how DnaA catalyzes origin melting. The relationship of this mechanism to other initiation systems, or to AAA+/ASCE proteins overall, is also unclear. A DnaA-ssDNA crystal structure {#S2} ============================== To examine these issues, we set out to determine the structure of DnaA bound to single-stranded DNA. Employing a truncation of Aquifex aeolicus DnaA consisting of the AAA+ and duplex-DNA-binding domains (which, like its E. coli counterpart^[@R16],\ [@R17]^, is active for both ATP-stimulated assembly and single-stranded-DNA (ssDNA) binding^[@R25]^), we first grew DNA-free crystals in the presence of Mg^2+^ and the non-hydrolyzable ATP mimic AMPPCP^[@R14]^. DNA substrates were then soaked into these crystals under low-salt conditions (Methods). Data collection and phasing by molecular replacement revealed four DnaA protomers per asymmetric unit, arranged in a spiral configuration that propagates into a continuous protein helix by the action of crystal-symmetry elements ([Fig 1a, 1b](#F1){ref-type="fig"}), along with bound single-stranded DNA. Of the multiple substrates screened (Methods), dA~12~ yielded the highest-quality density ([Fig S2a, S2b](#SD1){ref-type="supplementary-material"}), and served as the best target for model building and refinement. The final structure, containing a DnaA:AMPPCP:Mg^2+^:dA~12~ stoichiometry of 4:4:4:1, was refined to an R~work~/R~free~ of 24.9/26.8% at 3.35 Å resolution ([Table S1](#SD1){ref-type="supplementary-material"}). DnaA-ssDNA interactions {#S3} ======================= The overall arrangement of DnaA subunits in the helical assembly is highly similar to a DNA-free form reported previously (0.7 Å r.m.s.d. between all Cα positions)^[@R14]^. AMPPCP•Mg^2+^ binds at the interface between neighboring subunits, with the γ-phosphate of AMPPCP coordinated by catalytic amino acids from pairs of adjoining AAA+ domains. Single-stranded DNA associates exclusively with the AAA+ elements of the initiator, with each protomer binding three nucleotides of the dA~12~ strand ([Fig 1a](#F1){ref-type="fig"}). Almost all contacts are made through the phosphodiester backbone, exposing the DNA bases to solvent. Each trinucleotide segment adopts a B-form DNA conformation ([Fig S3](#SD1){ref-type="supplementary-material"}) with the bases between consecutive segments separated by large (\~10 Å) gaps, extending the substrate by \~50% ([Table S2; Supplemental Material](#SD1){ref-type="supplementary-material"}). DnaA binds single-stranded DNA using just two pairs of helices, α3/α4 and α5/α6, both of which line the central channel of the protein assembly ([Fig 1c](#F1){ref-type="fig"}). The geometry of these two elements creates a single conduit along the length of the DnaA superhelix that allows substrate to traverse consecutive DnaA protomers. Interestingly, helices α3/α4 also comprise the [I]{.ul}nitiator [S]{.ul}pecific [M]{.ul}otif (ISM), which both promotes filament formation^[@R14],\ [@R19]^ and distinguishes DnaA as a member of initiator clade of the AAA+ superfamily^[@R26],\ [@R27]^. DnaA uses a simple network of interactions to coordinate ssDNA. The ISM forms a shelf for each trinucleotide, in which a conserved hydrophobic residue, Val156, forms van der Waal contacts with the sugar and base of the first nucleotide in the triplet ([Fig 1d](#F1){ref-type="fig"}). The central phosphate of each trinucleotide is bound by the electropositive, N-terminal helix dipole of α6 and hydrogen bonded by Thr191 ([Fig 1c, 1d](#F1){ref-type="fig"}). These contacts are flanked by two positively charged residues, Arg190 and Lys188, which make salt-bridge interactions with the phosphates of nucleotides 1 and 3, respectively. Importantly, mutant initiators containing substitutions in these observed DNA-binding residues show reduced affinity for ssDNA in solution ([Fig S4](#SD1){ref-type="supplementary-material"}) confirming that the crystals captured a physiologically-meaningful initiator state. Moreover, mutations of the same positions in E. coli DnaA (amino acids Arg245, Lys243 and Val211) also disrupt ssDNA binding and origin melting^[@R15]^. Thus, the single-stranded DNA engagement strategy seen here appears conserved across bacterial species. Structural similarities between DnaA and RecA {#S4} ============================================= In considering the assembly patterns of oligomeric ATPases, we were struck by the similarity of DnaA to one system in particular: the homologous recombination protein, RecA. Although the cellular functions of these two proteins are fundamentally different (catalysis of DNA strand-exchange reactions versus replication origin melting and coordination of replisome assembly), both RecA and DnaA are predicated upon an ASCE ATPase fold^[@R27],\ [@R28]^. Like DnaA, RecA (and its Rad51/RadA orthologs) forms a helical assembly that engages DNA with its pore regions^[@R28]--[@R32]^. These shared physical properties led us to undertake a more detailed comparison of RecA and DnaA. Of the multiple models available, the structure of a RecA oligomer bound to single-stranded DNA^[@R33]^, representing the presynaptic complex formed during the initial stages of homologous recombination, is globally most similar to the DnaA state we observe ([Fig 2a, 2b](#F2){ref-type="fig"}). As with DnaA, RecA contacts DNA almost exclusively through the phosphodiester backbone, which sits in the interior of a positively-charged filament pore. Each RecA protomer binds three nucleotides in a B-DNA conformation, with the base stacking between each triplet interrupted such that single-stranded DNA is extended \~1.5 fold compared to a B-form duplex ([Fig 2c](#F2){ref-type="fig"}). RecA and DnaA also exhibit some interesting and significant differences. A visual examination of each triplet shows that RecA uses a more extensive network of contacts for engaging single-stranded DNA than does DnaA ([Fig 2d, 2e](#F2){ref-type="fig"}), burying twice as much surface area per triplet (318 Å^2^ and 639 Å^2^ for DnaA and RecA, respectively). This difference derives largely from an additional β-hairpin in RecA that fills the gap between each triplet and reinforces each three-base stack^[@R33]^. Moreover, while two of the three nucleotides within each RecA triplet (positions 1 and 2) align well with those seen in DnaA, position 3 of the DnaA trinucleotide rotates away from the pore axis by \~50° ([Fig 2f](#F2){ref-type="fig"}). This difference skews consecutive DnaA triplets away from one another, disrupting the formation of a smoothly spiral arrangement as seen in RecA ([Fig 2c](#F2){ref-type="fig"}). DNA extension is ATP- and assembly-dependent {#S5} ============================================ The ability of RecA to stretch DNA to the extent observed crystallographically has been amply substantiated by a wealth of methodologies^[@R34]--[@R36]^. Using these efforts as a guide, we set out to determine whether the DNA conformation we observe bound to DnaA accurately represents the state of the substrate in solution. To accomplish this, we employed a bulk-phase Fluorescence Resonance Energy Transfer (FRET)-based ssDNA extension assay analogous to single-molecule approaches applied to RecA^[@R37]^. Using a poly-thymine DNA labeled with Cy3 and Cy5 (FR-dT~21~) ([Table S3](#SD1){ref-type="supplementary-material"}), we monitored changes in the length of single-stranded DNA resulting from DnaA binding ([Fig 3a](#F3){ref-type="fig"}). Analogous studies were performed with RecA as a control. As both RecA and DnaA require ATP for formation of the oligomers observed in the structural models, we expected ATP-dependent extension to lead to a loss of FRET signal. We tested for extension both in the presence of ADP•BeF~3~, to avoid complications that might arise from nucleotide hydrolysis, and in the presence of ADP, which is known to promote DnaA disassembly. Pronounced extension was observed only in the presence of the ATP analog ([Fig 3b and 3c](#F3){ref-type="fig"}), and not in the presence of ADP. The lengths of single-stranded DNA in the ATP-assembled states of both proteins, as calculated from the FRET data, were in close agreement with those observed in the crystal structures ([Table S6](#SD1){ref-type="supplementary-material"}). Likewise, mutations in ssDNA-binding amino acids and residues required for DnaA assembly all significantly reduced ssDNA extension ([Fig S6](#SD1){ref-type="supplementary-material"}), demonstrating that this activity depends on substrate binding to the pore of an initiator oligomer that forms only when activated by ATP. DnaA directly catalyzes duplex melting {#S6} ====================================== How replication origins are opened for replisome assembly is an important, unanswered question. Given the similarities between the ssDNA binding and extension activities of DnaA and RecA, we reasoned that the initiator might directly destabilize and disrupt DNA duplexes. This activity is a known property of RecA^[@R38]^, albeit one that permits the recombinase to actively exchange DNA strands between target substrates^[@R30],\ [@R33]^. To test this idea, we developed a DNA strand-displacement assay for DnaA. First, the initiator was incubated with a short duplex containing one fluorescently-labeled strand. Unlabeled competitor strand was then added to capture any unwound species ([Fig 4a](#F4){ref-type="fig"}). Both ADP and ADP•BeF~3~ were tested to determine whether initiator assembly affected the outcome of the experiment, as were DNAs of different lengths and stabilities. Analysis of the resultant products by gel electrophoresis shows that DnaA readily unwinds a 15mer duplex DNA of moderate stability (T~m~=43°C) in the presence of the ATP mimic ([Fig 4b](#F4){ref-type="fig"}). By contrast, increasing the stability of the DNA substrate by \~30% (using a 20mer, T~m~=55°C) weakens the unwinding activity of DnaA ([Fig 4b](#F4){ref-type="fig"}), whereas increasing DNA stability even further (30mer, T~m~=62°C) abrogates melting completely ([Fig S7a](#SD1){ref-type="supplementary-material"}). Importantly, ADP did not support strand displacement, nor did ssDNA binding and DnaA assembly mutants ([Fig S7b, S7c](#SD1){ref-type="supplementary-material"}). These controls indicate that dsDNA melting is dependent not only upon formation of an assembled DnaA oligomer, but that the initiator is fine-tuned to specifically disrupt DNAs of modest stability. One significant functional difference between RecA and DnaA is that the recombination protein can drive a true strand-exchange reaction; that is, in addition to displacing one strand of a duplex, RecA can also pair homologous single-stranded DNA segments into a double-stranded molecule. By contrast, DnaA\'s function is to separate double-stranded origin regions. Inspection of the RecA and DnaA complexes reveals a physical basis for these differing properties: in DnaA, successive trinucleotide elements are arranged in a state incompatible with the formation of a continuous duplex, whereas ssDNA bound to RecA adopts a smoothly spiraled arrangement permitting the contiguous pairing of a complementary strand ([Fig 4c](#F4){ref-type="fig"}). This distinction arises primarily from the 50° rotation between the nucleotides at the third position of each triplet seen in the RecA and DnaA models ([Fig 2f](#F2){ref-type="fig"}). In DnaA, the orientation of this nucleotide appears to be stabilized by base stacking, whereas in RecA the β-hairpin insertion helps sculpt the configuration of the DNA to create a contiguous base-pairing surface. Implications for origin melting {#S7} =============================== Together, our findings present the strongest evidence yet that DnaA melts replication origins by directly assisting with the separation and sequestration of duplex DNA strands ([Fig S1c](#SD1){ref-type="supplementary-material"}). Notably, this activity does not contradict the demonstrated need for other factors capable of reshaping and/or destabilizing DNA (e.g., IHF and negative supercoiling) during initiation^[@R12],\ [@R13]^. Rather, these elements likely help promote DnaA assembly and prime the origin for melting by what otherwise would be an inefficient unwindase. In this view, the AAA+ domains of DnaA may first engage only one of the two strands of duplex DNA with their ssDNA binding elements (possibly at reported ssDNA or ATP-DnaA binding sites^[@R15],\ [@R16]^). In the presence of ATP, which triggers initiator assembly, subunit/subunit interactions would help restructure the DNA backbone, stretching the contacted strand to facilitate melting. Reannealing would be disfavored by the non-contiguous arrangement of base triplets in the extended state ([Fig 2d](#F2){ref-type="fig"}). Future studies will be needed to define the specific order and effect of these events further. We envision that the propensity of DnaA to open DNA could be adjusted in other bacterial species by strengthening or weakening the association of its ATPase domains with DNA and/or each other. An attractive feature of such a mechanism is that it is amenable to additional layers of control by changes to DUE sequence, superhelical density, and co-resident architectural factors to ensure that a replication origin fires only when DnaA is both present and assembled properly. Such flexibility may have played a role in allowing DnaA to persist as the primary initiator in bacteria that have adapted to markedly different environmental niches. Thematic patterns of substrate recognition in ASCE ATPases {#S8} ========================================================== The mechanism by which DnaA coordinates ssDNA also comports well with findings in other replication initiation systems and with ASCE ATPases in general. For instance, many oligomeric RecA and AAA+ enzymes bind substrate in the interior pore of a closed- or cracked-ring particle^[@R33],\ [@R39]--[@R41]^. DnaA follows this pattern. A comparison of DnaA to other, disparate nucleic acid-dependent AAA+ systems -- e.g., polymerase clamp loaders and processive helicases -- further shows that these factors also associate with client substrates in a remarkably analogous manner, using the same face of the core αβα ATP-binding fold to engage a short backbone stretch of their target DNAs ([Fig 5](#F5){ref-type="fig"}). For AAA+ proteins involved in initiation, these similar contact mechanisms have been differentially co-opted to assist with specific protein functions, ranging from the control of origin recognition (as seen in archaeal Orc1 proteins^[@R26],\ [@R42]^), to mediating processive DNA unwinding (viral superfamily 3 helicases^[@R40],\ [@R43]^). DnaA, with its ability to melt (but not translocate along) DNA, appears to employ an intriguing mix of some of the activities exhibited by related initiation systems. Future efforts will be needed to determine how subtle differences in the position and nature of substrate-binding surfaces, combined with specific alterations in the assembly patterns of central AAA+ domains, endow such molecular motors and switches with their distinct biochemical properties. Methods Summary {#S9} =============== Detailed information regarding experimental methods, substrate sequences, binding constants, and FRET efficiencies and distances can be found in the [Online Methods](#SD2){ref-type="supplementary-material"} section and in the [Supplementary Material](#SD1){ref-type="supplementary-material"}. Supplementary Material {#SM} ====================== We would like to thank Karl Drlica, James Keck, Thomas Murray and the Berger laboratory for helpful comments, and Michael M. Cox for his generous contribution of RecA protein. This work was supported by the NIGMS (GM071747) and the National Institute of Health Molecular Biophysics Training Grant T32 GM008295. Author Information Coordinates have been deposited in the RSCB PDB under the accession number 3R8F. Author Contributions K.E.D. and J.M.B. designed the experiments, analyzed the data and wrote the paper. Protein purification, crystallization and ssDNA binding assays were performed by K. C. and K.E.D. K.E.D. performed the other experiments. ![The ATPase pore of assembled DnaA binds ssDNA\ a, Side view of the asymmetric unit, with DnaA subunits differentially colored. Single-stranded DNA is displayed as red sticks. AMPPCP and Mg^2+^, bound to chain A, are shown as spheres colored by element and in magenta, respectively; AMPPCP•Mg^2+^ bound to chains B-D are occluded in this view. b, Side and top views of oligomerized DnaA, reconstructed through crystal packing, showing twelve DnaA subunits and three strands of ssDNA. Coloring as per panel A. c, Side view of the DnaA tetramer with helices α3/α4 and α5/α6 highlighted in orange and yellow, respectively ("ISM" -- initiator specific motif). Single-stranded DNA is shown as a transparent stick-and-surface representation colored by element; phosphates are further highlighted as red spheres. d, Protein-DNA contacts. Protein chains B (left) and C (right) are displayed with the same coloring as in c. Single-stranded DNA is colored by element.](nihms-318339-f0001){#F1} ![DnaA engages ssDNA in a manner similar to RecA\ a, View of a DnaA-AMPPCP-ssDNA pentamer (consisting of one full tetramer, as well as chain A (DnaA~A\'~) and its associated triplet from the adjacent asymmetric unit). AMPPCP•Mg^2+^ is shown as spheres colored by atom, ssDNA as red sticks. b, View of a RecA-ADP-AlF~4~-ssDNA pentamer (PDB ID 3CMW)^[@R33]^. ADP•AlF~4~•Mg^2+^ is shown as spheres colored by atom, ssDNA as red sticks. c, Comparison of ssDNA bound to DnaA (orange), RecA (green) and a strand of B-DNA (yellow). d, Close-up view of triplet bound to DnaA (chain C) with magenta dashed lines indicating key contacts. e, Close-up view of triplet bound to RecA (protomer 2) with magenta dashed lines indicating key contacts. f, Side (left) and top (right) views of the triplets displayed in d and e aligned with each other.](nihms-318339-f0002){#F2} ![DnaA extends ssDNA in solution\ a, Cartoon of ssDNA extension assay. b, Emission scan (donor excitation) of FR-dT~21~ in the presence of 10 μM DnaA with either ADP•BeF~3~ (top) or ADP (bottom). c, Emission scan (donor excitation) of FR-dT~21~ in the presence of 10 μM RecA with either ATPγS (top) or ADP (bottom). Reported transfer efficiencies and distances were calculated using donor emission as described in the Methods.](nihms-318339-f0003){#F3} ![DnaA directly melts duplex DNA\ a, Schematic of strand displacement assay. The green circle represents the Cy3 fluorescent end-label used to follow the status of one DNA strand. Complementary strands of duplex substrates are colored grey and black. b, Strand displacement assay conducted with 15 and 20mer duplex substrates (C3--15mer and C3--20mer) in the presence and absence of different nucleotides. DnaA concentrations used are indicated above each lane. c, (left) Cartoon model showing how complementary base triplets (yellow) would pair (in a B-DNA manner) with ssDNA bound to DnaA (red). The orientation of successive DnaA-bound triplets is such that it prevents the formation of a continuous base-paired strand favoring duplex separation. (right) Same DNA view, but as seen in RecA, where triplets are oriented to allow pairing of an extended complementary strand to promote duplex formation and strand exchange (PDB ID 3CMX)^[@R33]^.](nihms-318339-f0004){#F4} ![Common DNA recognition strategies of AAA+ proteins\ Structures of DNA-bound assemblies (top) and individual domains (bottom) for AAA+ proteins involved in replication. All recognize DNA using the same face of the AAA+ fold (violet) (bottom). a, Bacterial clamp-loader (γδδ′) complex (AAA+ domains -- differentially colored) bound to primer-template DNA (PDB ID 3GLF)^[@R41]^. b, Archaeal initiators Orc1--1 (gray) and Orc1--3 (AAA+ domain - green) bound to origin DNA (PDB ID 2QBY)^[@R26]^. c, Bacterial initiator DnaA (AAA+ domains -- gray/blue) bound to ssDNA. d, Viral initiator/helicase E1 (AAA+ domains -- orange/gray) bound to ssDNA (PDB ID 2GXA)^[@R40]^. For all panels, DNA is shown as either red spheres (top), or as a red/grey cartoon (bottom).\ Nucleotide co-factors bound to AAA+ domains (bottom) are represented as spheres colored by atom.](nihms-318339-f0005){#F5}	{ "pile_set_name": "PubMed Central" }
Introduction {#Sec1} ============ Innate immunity is a ubiquitous system that widely protects organisms from infectious pathogens as a front-line host defense mechanism. The immune response is triggered by the recognition of broadly conserved microbial components, known as pathogen-associated molecular patterns, by germline-encoded pattern recognition receptors of the host cells^[@CR1]^. As an early defense system against RNA viruses in mammals, the innate immune response is precisely controlled by two distinct signal transduction pathways mediated by the pattern recognition receptors Toll-like receptor 3 (TLR-3) and retinoic acid-inducible gene I (RIG-I)-like receptors (RLR) that respond to virus-derived RNAs^[@CR2],\ [@CR3]^ (i.e., pathogen-associated molecular patterns). Although the two pathways differ with respect to the initial activation of their downstream effectors, they converge at the point of activation of the transcriptional factors interferon regulatory factor 3 (IRF-3) and nuclear factor κB (NF-κB), which results in the rapid production of type I interferons (IFN-α and -β) and other proinflammatory cytokines to establish adaptive antiviral immunity^[@CR4]^. Mitochondria, eukaryotic cell powerhouses, are crucially involved in numerous cellular processes, including apoptosis^[@CR5]^ and calcium homeostasis^[@CR6]^. Mitochondria also have a unique role in innate immunity against RNA viruses^[@CR7]^. Mitochondrial-mediated antiviral immunity depends on activation of the RLR signaling pathway, and mitochondrial antiviral signaling (MAVS), a downstream adaptor of RLR at the mitochondrial outer membrane (MOM), has a key role in the signal transduction^[@CR8],\ [@CR9]^. Upon viral infection, MAVS recruits various types of effectors at the MOM, and the orchestrated "MAVS signalosome", including the mitochondrial membrane potential (Δψ~m~), is the primary unit governing antiviral innate immunity^[@CR10],\ [@CR11]^. Although the role of the MAVS signalosome in mitochondria with its dynamic morphologic properties^[@CR12]^ to provide a molecular platform that facilitates signal transduction is well characterized, insight into how the organelle functions to facilitate antiviral immunity through the activity of oxidative phosphorylation (OXPHOS) has remained unclear. Results and Discussion {#Sec2} ====================== Cultured cells rely on mitochondrial respiratory activity {#Sec3} --------------------------------------------------------- To evaluate the functional coordination of mitochondrial-mediated antiviral immunity and OXPHOS activity, we first sought to determine optimal cell culture conditions in which the cellular bioenergetics would rely on mitochondrial respiratory activity. We used a fluorescence resonance energy transfer (FRET)-based assay to visualize metabolized intracellular adenosine 5′-triphosphate (ATP) at the single-cell level in human embryonic kidney 293 (HEK293) cells expressing an ATP probe, ATeam1.03^[@CR13]^. The biosensor assay performed with cells cultured under our customized medium containing galactose (10 mM) as the carbon source revealed high FRET signal \[based on an emission ratio of 527/475 nm (denoted YFP/CFP)\] in individual cells \[Fig. [1A](#Fig1){ref-type="fig"}, galactose panels, (−)\], indicating that cells maintained adequate intracellular ATP levels. The intracellular ATP level, however, which affects the YFP/CFP ratio, was dramatically decreased (\~2.5-fold) by the addition of electron transport chain (ETC) inhibitors (rotenone and antimycin A), an ATP synthase inhibitor (oligomycin), or a protonophore \[carbonyl cyanide m-chlorophenylhydrazone (CCCP)\] to the media \[Fig. [1A](#Fig1){ref-type="fig"}, galactose panels, (+), and B\], demonstrating that energy production in the cells critically depends on mitochondrial OXPHOS. In contrast, cells maintained in a customized medium containing glucose (10 mM) exhibited no significant decrease in the intracellular ATP level, even after treatment with the pharmacologic drugs (Fig. [1A,B](#Fig1){ref-type="fig"}). These results clearly demonstrated that the cellular bioenergetics are less dependent on OXPHOS activity in the presence of glucose, and that cellular ATP production predominantly relies on glycolysis.Figure 1Bioenergetic profiling of cells under oxidative conditions. (A) HEK293 cells stably expressing the ATeam1.03 were cultured in customized media containing either galactose (right) or glucose (left) at 37 °C, and then the indicated mitochondrial inhibitors were added to each medium. The time-course of the fluorescence emission ratio (YFP/CFP) was monitored to visualize cytosolic ATP levels in each living cell, and ratiometric pseudocolor images of cells pre- (−) or post-treated (+) with inhibitors are shown. The images were processed in MetaMorph (Molecular Devices), and the blue color indicates less cytoplasmic ATP. Scale bar, 20 μm. (B) Quantification of the YFP/CFP ratio calculated from images in (A). Number of cells (n) used for the quantification is shown at the top of the graph. Error bars indicate SD (Unpaired t-test; \\\*P \< 0.001). Validation of the RLR signaling pathway under oxidative conditions {#Sec4} ------------------------------------------------------------------ We examined RLR-mediated signal transduction under the galactose-containing oxidative condition. Under oxidative conditions, similar to the glucose condition, ectopic expression of MAVS in HEK293 cells potently activated both IFN-β and NF-κB luciferase reporters in a dose-dependent manner (Supplementary Fig. [S1A](#MOESM1){ref-type="media"}). The OXPHOS-dependent RLR-mediated signal transduction was also confirmed by stimulating cells transfected with either a plasmid encoding RIG-I(1--250) (Supplementary Fig. [S1B](#MOESM1){ref-type="media"}) or a synthetic analog of viral double-stranded RNA \[dsRNA; poly(I:C)\] (Supplementary Fig. [S1C and D](#MOESM1){ref-type="media"}), both of which are upstream factors of MAVS^[@CR3]^. The kinetic response of IRF-3 phosphorylation (a hallmark of IRF-3 activation) in cells against infection by Sendai virus (SeV), a negative-stranded RNA virus of the Paramyxoviridae family, in the OXPHOS-dependent condition was also similar to that in the glucose condition (Fig. [2A](#Fig2){ref-type="fig"}). The RIG-I-mediated activation of the IFN-β reporter in the OXPHOS-dependent condition was sufficiently impaired by co-expression of a hepatitis C virus serine protease NS3/4A (an inhibitor of the RLR signaling pathway)^[@CR14],\ [@CR15]^ \[Fig. [2B](#Fig2){ref-type="fig"}, wild-type (WT)\], whereas its inactive mutant (S139A) had no functional effect, indicating that the observed signal transduction occurred via RLR-dependent signaling pathway.Figure 2RLR-mediated signal transduction under oxidative conditions. (A) The kinetic profile of IRF-3 activation in oxidative or glycolytic medium-cultured HEK293 cells that were challenged with SeV (4 HA units/mL). Each cell lysate was collected at the indicated time-points (3, 6, 9 and 12 h) and analyzed by Western blotting with antibodies against the specific antibody (pIRF-3; phosphorylation of Ser386). Anti-β-actin was used as the loading control and anti-SeV as the infection control. U.I., uninfected. (B) Oxidative medium-cultured HEK293 cells were transfected with 50 ng of empty vector (Mock) or expression plasmid for Myc-tagged RIG-I(1--250) (control) together with the IFN-β reporter plasmid. The two right lanes (+NS3/4A) indicate that 100 ng of FLAG-tagged WT or inactive (S139A) NS3/4A serine protease expression plasmids were also co-transfected with the RIG-I(1--250) plasmid. The immunoblot on the top represents an expression profile of Myc-tagged RIG-I(1--250) and FLAG-tagged NS3/4A mutants as well as the loading control of endogenous β-actin. Error bars indicate SD (n = 3; Unpaired t-test; \\P \< 0.01 and \\\*P \< 0.001, respectively). (C) BRET saturation assay of MAVS oligomerization in glycolytic versus oxidative media. HEK293 cells were co-transfected with 5 ng NLuc-MAVS expression plasmid and increasing amounts (0--200 ng) of Venus-tagged MAVS plasmid along with 200 ng of WT (circles) or S139A (squares) FLAG-tagged NS3/4 A plasmids, and analyzed 24 h later using a BRET saturation assay. Closed and open symbols represent glycolytic and oxidative conditions of HEK293 cells, respectively, and inset blots show Western blots from the BRET saturation point of each curve by immunoblotting with the indicated antibodies. Error bars indicate SEM (n = 3). (D) Heat maps of microarray analysis. Total RNAs were isolated from glycolytic and oxidative cultured conditions of primary MEFs that were unchallenged (−) or challenged (+) with SeV (30 HA units mL^−1^) for 6 h, and microarray analysis was performed. The heat map was generated by MeV software^[@CR42]^, and the color indicates the distance from the median of each row. (E) Similar to (A), except that J774A.1 macrophages were challenged with SeV (2 HA units/mL). U.I., uninfected. We next analyzed the structural features of MAVS activation accompanied by RLR-mediated signal transduction, homotypic oligomerization at the MOM^[@CR11]^, when cells were dependent on the oxidative condition. The interaction between ectopically expressed Venus- and NanoLuc luciferase (NLuc)-tagged MAVS with co-expression of the S139A mutant was successfully monitored, and a bioluminescence resonance energy transfer (BRET)-based assay^[@CR16]^ revealed a hyperbolic saturation curve characteristic of a specific interaction (Fig. [2C](#Fig2){ref-type="fig"}, squares). Importantly, the specificity of the observed MAVS-MAVS interaction was verified by the disappearance of the saturated curve when the WT NS3/4A protease was co-expressed in cells due to MAVS cleavage^[@CR14]^ (Fig. [2C](#Fig2){ref-type="fig"}, circles, also see immunoblot). We further verified the antiviral immune response of primary mouse embryonic fibroblasts (MEFs) against SeV infection in the OXPHOS-dependent condition. Gene expression profiling of MEFs showing its dependence on the oxidative condition (Supplementary Fig. [S2](#MOESM1){ref-type="media"}, upregulation of OXPHOS genes) revealed the expected induction of numerous IFN-stimulated and antiviral-signaling genes in response to the viral infection, similar to that observed in the glucose condition (Fig. [2D](#Fig2){ref-type="fig"}). Consistent with the aforementioned assays, we also observed that the antiviral immune response of macrophages cultured under the oxidative conditions was intensively upregulated upon SeV infection (Fig. [2E](#Fig2){ref-type="fig"}). RLR-mediated antiviral innate immunity requires OXPHOS activity {#Sec5} --------------------------------------------------------------- Because altering the energy metabolism in cells through the activity of OXPHOS could accomplish the antiviral immune response, we next asked whether arresting OXPHOS activity would disrupt RLR-mediated signal transduction. IRF-3 phosphorylation in cells responding to SeV infection was completely suppressed by the addition of ETC inhibitors to the medium under oxidative conditions (Figs [3A](#Fig3){ref-type="fig"} and [S3](#MOESM1){ref-type="media"}). The functionality of the cells treated with the pharmacologic drugs under these conditions, however, was unaffected when the cells were either stimulated with extracellular poly(I:C) (Supplementary Fig. [S4A](#MOESM1){ref-type="media"}) or infected with a double-stranded DNA virus, Herpes simplex virus 1 (HSV-1) (Supplementary Fig. [S4B](#MOESM1){ref-type="media"}), which activates the TLR-3 or cGAS/STING cytosolic DNA sensing pathways, respectively. Influenza A virus (IAV; PR8 strain), an RNA virus of the Orthomyxoviridae family, encodes the viral protein PB1-F2 that translocates into mitochondria and induces mitochondrial depolarization^[@CR17]--[@CR20]^. Using the viral strain, we observed severe PB1-F2--induced inhibition of RLR signaling in cells that depended on OXPHOS (Fig. [3B](#Fig3){ref-type="fig"}, WT); the inhibitory effect was attenuated, however, when we used a recombinant virus with genetic ablation of the PB1-F2 gene (ΔF2)^[@CR21]^. These observations suggest that the severity of the mitochondrial dysfunction correlates with the severity of the defect in the RLR-induced antiviral response.Figure 3OXPHOS activity couples with the RLR pathway to execute antiviral signal transduction. (A) HEK293 cells cultured in glycolytic or oxidative media were infected with SeV (4 HAU/mL) for 2.5 h, and the infected cells were further incubated with the indicated mitochondrial inhibitors for 2.5 h (total 5 h infection). The activation of endogenous IRF-3 was analyzed by Western blotting with antibodies against the specific antibody (pIRF-3; Ser386). Anti-β-actin was used as the loading control. U.I., uninfected. (B) Comparison of the IRF-3 activation between cells infected with recombinant influenza A/PR8 viruses (WT versus ΔF2, each used 2 HAU/mL) cultured under oxidative conditions. The ΔF2 strain is a mutant strain with genetic removal of the PB1-F2 gene from the viral genome^[@CR21]^. The graph on the right shows the quantification of pIRF-3 bands analyzed by densitometry. Error bars indicate SD (n = 3; Unpaired t-test; \*P \< 0.05). (C,D) The B82 WT cybrids and ρ~0~ cells were infected with (C) SeV (4 HAU/mL) for 18 h or (D) HSV-1 (1 × 10^5^ PFU) for 24 h, and the cell-free supernatants were analyzed by ELISA to measure the secreted amounts of IFN-β (left panel) and IL-6 (right panel), respectively. Error bars indicate SD (n = 3; Unpaired t-test; \\P \< 0.01 and \\\*P \< 0.001, respectively). U.I., uninfected. N.D., not detected. (E) Similar to (C), except that the mtDNA-less J774A.1 and its parental macrophages were infected with SeV (2 HA units/mL). Inset panel: relative mtDNA copy number was confirmed by qPCR. In (C--E), cells were maintained in ρ~0~ medium. We next used a genetic approach to examine whether mitochondrial DNA (mtDNA), which is essential for mitochondrial OXPHOS activity^[@CR22]^, modulates RLR-mediated signal transduction. Mouse fibrosarcoma B82 cells devoid of mtDNA (ρ~0~ B82)^[@CR23]^ (Supplementary Fig. [S5A](#MOESM1){ref-type="media"}) produced less IFN-β and IL-6 against SeV infection, although the immune response was fully rescued in a cell cybrid harboring WT mtDNA (Fig. [3C](#Fig3){ref-type="fig"}), despite the fact that both cell types had adequate energy due to the use of a high glucose medium (25 mM) with pyruvate and uridine (Supplementary Fig. [S5B](#MOESM1){ref-type="media"}) or the same infection level (Supplementary Fig. [S5C](#MOESM1){ref-type="media"}). Conversely, ρ~0~ B82 cells had substantial antiviral responses against HSV-1 infection (Fig. [3D](#Fig3){ref-type="fig"}), consistent with a previous observation^[@CR24]^. To confirm that these observations were not due to specific characteristics of the cell type we used, we performed the same experiment in HeLa cells lacking mtDNA^[@CR25]^ (ρ~0~ HeLa), and found that ρ~0~ HeLa cells also exhibited defective antiviral innate immune responses against SeV infection (Supplementary Fig. [S5D and E](#MOESM1){ref-type="media"}). In addition, transient depletion of mtDNA in macrophages (mtDNA-less J774A.1) treated with rhodamine 6G exhibited an impaired immune response against SeV infection (Fig. [3E](#Fig3){ref-type="fig"}). Taken together, these findings indicated that RLR-mediated signal transduction requires OXPHOS activity to execute antiviral immune responses. Pathogenic mtDNA mutations cause defects in antiviral innate immunity {#Sec6} --------------------------------------------------------------------- The importance of mtDNA in the regulation of RLR-mediated signal transduction prompted us to explore the phenotypic effects of pathogenic mtDNA mutations on antiviral immunity. We used human osteosarcoma cybrid 143B cells harboring homoplasmic pathogenic mtDNA mutations: the ND1 cybrid^[@CR26]^ (complex I dysfunction), which is associated with Leber hereditary optic neuropathy, carrying a missense mutation at 3460/ND1 (Supplementary Fig. [S6A](#MOESM1){ref-type="media"}); and the COX cybrid^[@CR27]^ (missing complex IV), which is associated with mitochondrial encephalopathy, carrying a stop codon mutation at 6930/COXI (Supplementary Fig. [S6B](#MOESM1){ref-type="media"}), both of which exhibit severe respiratory defects. As described, ρ~0~ cells from 143B cell lines also exhibited an impaired immune response against SeV infection (Fig. [4A,B](#Fig4){ref-type="fig"}). Remarkably, activation of the RLR-mediated antiviral signaling in the mutant cells (ND1 and COX) was severely disrupted (Fig. [4A,B](#Fig4){ref-type="fig"}) despite the indistinguishable MAVS pattern and energy supply (Supplementary Fig. [S7A, B and C](#MOESM1){ref-type="media"}). Consistent with these results, both pathogenic mutants were more susceptible to viral infection (recombinant IAV-expressing GFP)^[@CR28]^ than WT cells (Fig. [4C](#Fig4){ref-type="fig"}), due to defective mitochondrial-mediated immunity. Most importantly, the immunodeficiency observed in the ND1 cybrid, but not in the COX cybrid, was significantly recovered by the addition of succinate, a complex II substrate, to the media, suggesting that OXPHOS activity in ND1 mutant cells was driven by the substrate through complex II, which consequently restored its phenotype (Fig. [4D](#Fig4){ref-type="fig"}). Together, these results indicated that cells with defective OXPHOS activity due to mtDNA mutations have severely impaired RLR-mediated antiviral signaling.Figure 4Defects in mtDNA cause malfunction in antiviral innate immunity. (A,B) The 143B cybrids and ρ~0~ cells were infected with SeV (5 HAU/mL) for 18 h, and (A) activation of both IRF-3 and IκBα was analyzed by Western blotting with antibodies against its specific phosphorylated-detection antibodies (Ser386 for IRF-3 and Ser32/36 for IκBα) or (B) the cell-free supernatants were analyzed by ELISA to measure the secreted amounts of IFN-β (top panel) and IL-6 (bottom panel). U.I., uninfected. Error bars indicate SD (n = 3; Unpaired t-test; \\\*P \< 0.001). (C) Fluorescence microscopy of 143B cybrid cells infected with IAV-GFP (0.6 HAU/mL) for 24 h. In (A--C), cells were maintained in ρ~0~ medium. (D) Similar to (B), except that succinate (suc), the complex II substrate with ADP, was added to the oxidative medium as indicated. Error bars indicate SD (n = 3; Unpaired t-test; \\\*P \< 0.001; N.D., not detected; N.S., not significant). Trans-mitochondrial mice (Mito-miceΔ) are susceptible to IAV infection {#Sec7} ---------------------------------------------------------------------- To evaluate whether our in vitro observations were physiologically relevant, we investigated the relationship between OXPHOS and antiviral innate immunity in vivo. We used 28-week-old trans-mitochondrial mice carrying mtDNA with a large-scale deletion (Mito-miceΔ)^[@CR29]^ and challenged the mice with the IAV PR8 strain. Mito-miceΔ in a resting state had high blood lactate levels (7.2 ± 2 mmol/L) relative to those in WT mice (3.2 ± 0.4 mmol/L), showing lactic acidosis in the mutant mice. Although the mortality of Mito-miceΔ (n = 7) infected with IAV was not much worse than that of WT mice (n = 9) (Supplementary Fig. [S8A](#MOESM1){ref-type="media"}), the Mito-miceΔ exhibited significant weight loss during the monitoring period (Fig. [5A](#Fig5){ref-type="fig"}; 17% of Mito-miceΔ versus 4% of WT on day 7). Given the higher pathogenicity in Mito-miceΔ following IAV-infection, we examined the physiologic effects of the infection on the lung tissue. In the lungs of WT mice infected with IAV, mild hyperemia and perivascular inflammatory cells infiltration around the bronchus were minimally observed, but few alveoli were injured by the infection (Fig. [5B](#Fig5){ref-type="fig"}, left panels). On the other hand, Mito-miceΔ infected with the virus had significant inflammation in whole lung tissues with desquamation of the bronchial epithelium (Fig. [5B](#Fig5){ref-type="fig"}, right panels, see arrowheads) and alveolar pneumocyte hyperplasia. Notably, virus-induced expression of endogenous IFN-β was impaired in the lungs of the Mito-miceΔ (Supplementary Fig. [S8B](#MOESM1){ref-type="media"}), explaining the histopathologic observation of enhanced pathogenicity. We thus concluded that defective OXPHOS enhances the susceptibility to IAV infection in vivo, emphasizing the importance of mitochondrial respiration for modulating innate immunity.Figure 5OXPHOS-defective mice are highly susceptible to viral-infection. (A) Mito-miceΔ (n = 7) or WT mice (n = 9) were challenged with IAV (1 × 10^3^ PFU) and mouse weight loss was monitored for 14 days. In the graph, the percentage change from the initial weight of the mice is shown. Error bars indicate SEM \[Dunnett's test; \\P \< 0.01 and \\\\P \< 0.0001 (versus day 0), respectively)\]. ^\#^Dunnett's test could not be applied because dead mice appeared from day 9. (B) In a separate experiment, the lungs were obtained from each infected mouse on day 8 post-infection, sectioned, and analyzed for histopathology following staining with hematoxylin and eosin. Enlarged boxes are depicted in lower images (400× magnification). Labels A and B in the images indicate alveoli and bronchus, respectively, and arrowheads indicate desquamation of the bronchial epithelium. Scale bars, 100 μm (top) and 50 μm (bottom), respectively. OPA1-mediated execution of antiviral innate immunity {#Sec8} ---------------------------------------------------- To elucidate the molecular basis of OXPHOS activity coupled with the RLR signaling pathway, we speculated that optic atrophy 1 (OPA1), a mediator of mitochondrial fusion^[@CR30]^, has a role in antiviral innate immunity^[@CR12]^ because of its stabilizing effects on mtDNA^[@CR31]^ and the cristae structure^[@CR32],\ [@CR33]^, both of which contribute to regulating mitochondrial respiration. As reported previously, we observed that OPA1 ^−/−^ cells contained less mtDNA (Fig. [6A](#Fig6){ref-type="fig"}), completely lacked mitochondrial fusion (Supplementary Fig. [S9A](#MOESM1){ref-type="media"}), and had a disrupted cristae morphology (Supplementary Fig. [S9B](#MOESM1){ref-type="media"}). These phenotypic defects were restored when a WT OPA1 isoform gene (L-OPA1; variant 1), but not an inactive mutant K301A or short isoform (S-OPA1), was re-introduced into null cells. Strikingly, although the OPA1 ^−/−^ cells exhibited severely impaired SeV-induced production of IFN-β and IL-6 (Fig. [6B](#Fig6){ref-type="fig"}), the immune responses were substantially restored by recovery of the mtDNA level in rescued cells \[Figs [6A,C](#Fig6){ref-type="fig"} and [S9C](#MOESM1){ref-type="media"}; V1(WT)\]. Given that OPA1 is important for mtDNA stability and the antiviral immune response, we used histochemical staining for cytochrome c oxidase (complex IV) activity to directly assess the OPA1 function involved in mitochondrial respiration. As expected, the COX staining pattern of the V1(WT)-rescued OPA1 ^−/−^ cells showed full recovery of the respiration defect, whereas K301A and S-OPA1 failed to rescue the COX activities (Fig. [6D](#Fig6){ref-type="fig"}). Taken together, these results indicate that functional actions of OPA1 are linked to the improvement of OXPHOS activity and the induction of mitochondrial-mediated antiviral innate immunity.Figure 6OPA1 contributes to mitochondrial-mediated antiviral signaling through stabilizing mtDNA. (A) Analysis of mtDNA copy number per nuclear DNA in WT and mutant MEFs. Restoration of mtDNA in OPA1 null cells was measured in mutant cells infected with a retrovirus-expressing variant 1 WT OPA1 \[V1(WT)\], its K301A mutant \[V1(K301A)\], and short isoform OPA1 (S-OPA1). Error bars indicate SEM (n = 3; Unpaired t-test; \\P \< 0.01). (B) SeV-induced antiviral innate immune response in OPA1-null MEFs. The WT and OPA1-null MEFs were either uninfected (U.I.) or infected with SeV (4 HAU/mL) for 18 h, and the cell-free supernatants were analyzed by ELISA to measure the secreted amounts of IFN-β (left panel) and IL-6 (right panel). Error bars indicate SD (n = 3; Unpaired t-test; \\P \< 0.01 and \\\*P \< 0.001, respectively). N.D., not detected. (C) Similar to (B), except that the immune response in mutant OPA1 MEFs was monitored. Error bars indicate SD (n = 3; Unpaired t-test; \\\*P \< 0.001). (D) Cytochemical analysis of COX activity. Cells expressing COX activity were indicated by a brown color. Scale bar, 20 μm. The findings of the present study using both in vitro and in vivo approaches provided new insight into how mitochondria facilitate antiviral immunity through OXPHOS activity. Mitochondria are believed to have evolved from organisms such as α-proteobacterium, and the discovery of their role in host-cell defense was unexpected. It is not surprising, however, that cells gaining respiratory function through a symbion could greatly advance not only through enhanced energy production, but also through safeguarded host defense against other infectious pathogens, especially in vertebrates. Materials and Methods {#Sec9} ===================== Reagents {#Sec10} -------- CCCP, rotenone, oligomycin, and antimycin A were purchased from Sigma-Aldrich (St. Louis, MO). Cytochrome c from horse heart was supplied by Nacalai Tesque Inc. (Kyoto, Japan) and 3,3′-diaminobenzidine was purchased from Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan). Furimazine was supplied by Promega (Madison, WI) and poly(I:C) was purchased from InvivoGen (San Diego, CA). Enzyme-linked immunosorbent assay (ELISA) kits for human and mouse IFN-β were supplied by Kamakura Techno-Science Inc. (Kanagawa, Japan) and PBL Assay Science (Piscataway, NJ), respectively. All other reagents were of biochemical research grade. Antibodies {#Sec11} ---------- A polyclonal antibody against IRF-3 (FL-425; 1:2000) and monoclonal antibody against β-actin (C4; 1:4000) were purchased from Santa Cruz Biotechnology (Dallas, TX), and the rabbit monoclonal antibody against phosphorylated IRF-3 (pIRF-3) Ser^386^ (EPR2346; 1:2000) and polyclonal antibody against MAVS (ab25084; 1:1000) were from Abcam (Cambridge, MA). Monoclonal antibodies against Myc (9E10; 1;1000) and FLAG (M2; 1:1000) were obtained from Covance (Princeton, NJ) and Sigma-Aldrich, respectively. Monoclonal antibodies against phosphorylated IκBα (pIκBα) Ser^32/36^ (5A5; 1:1000), phosphorylated NF-κB p65 (pNF-κB p65) Ser^536^ (93H1; 1:2000) and NF-κB p65 (D14E12; 1:3000) were supplied by Cell Signaling Technology (Danvers, MA), and the monoclonal antibody against OPA1 (1:2000) was from BD Biosciences (San Jose, CA). The anti-human and anti-mouse IL-6 antibodies used for ELISA were obtained from eBioscience, and a polyclonal antibody against SeV (PD029; 1:1000) was purchased from MBL (Nagoya, Japan). Monoclonal antibodies against mtCO1 (1D6E1A8; 1:1000) and mitochondrial heat shock protein 70 (mtHsp70; 1:2000) were purchased from Molecular Probes/Invitrogen (Carlsbad, CA) and Affinity BioReagents (Golden, CO), respectively. The Alexa Fluor 488--conjugated polyclonal antibody against rabbit immunoglobulin G (IgG) and Alexa Fluor 568--conjugated polyclonal antibody against rabbit IgG (1:500) were obtained from Molecular Probes-Invitrogen, and the Cy3-conjugated sheep anti-mouse IgG (1:1000) monoclonal antibody was purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). A monoclonal antibody against PB1-F2 (A/PR8 strain; 1:10) was a kind gift from Viktor Wixler (Münster University Hospital Medical School). Plasmids {#Sec12} -------- Plasmids encoding human MAVS variants, hRIG-I(1--250), NS3/4A variants, and the ATeam1.03 were described previously^[@CR13],\ [@CR16],\ [@CR20],\ [@CR34]^. Plasmid encoding rat OPA1 was a gift from Naotada Ishihara (Kurume University, Japan). To generate the retroviral expression constructs, each cDNA was recloned into the retroviral vector pMXs-Puro (Cell Biolabs, San Diego, CA). The retroviral expression vectors were then transfected into the platinum packaging cell lines (Cell Biolabs), and the retroviral supernatant was harvested 48 h post-transfection and used to infect cells. Cell lines and viruses {#Sec13} ---------------------- HEK293, MEF, and J774A.1 cells were maintained in Dulbecco's modified Eagle medium (high glucose; 4,500 mg/L; GIBCO BRL) supplemented with 1% GlutaMAX, penicillin (100 U/mL)-streptomycin (100 μg/mL), and 10% fetal bovine serum at 5% CO~2~ and 37 °C. The ρ~0~ cells (lacking mtDNA) from HeLa, B82 (mouse fibrosarcoma), and 143B (human osteosarcoma) and its WT or mutant mtDNA cybrids were similarly maintained, except with the addition of sodium pyruvate (110 mg/L) and uridine (50 μg/mL; ρ~0~ medium). The mtDNA-less J774A.1 macrophages were pretreated with rhodamine 6 G (0.06 μg/mL in ρ~0~ medium) for 48 h to transiently eliminate endogenous mtDNA. For glucose-free culture experiments, glycolytic (glucose) and oxidative (galactose) media were customized according to previously published protocols^[@CR35]^ with slight modifications; the glycolytic medium was distinct from the high glucose medium used to maintain the cultures. The customized medium was based on glucose and glutamine-free Dulbecco's modified Eagle medium (GIBCO; A14430--01) externally supplemented with 10% dialyzed fetal bovine serum (GIBCO; 26400--044), 2% GlutaMAX, penicillin (100 U/mL)-streptomycin (100 μg/mL), and either 10 mM glucose (glycolytic) or 10 mM galactose (oxidative), respectively. Sendai virus Cantell strain was purchased from the American Type Culture Collection, and the recombinant influenza viruses A/PR8 (ΔF2)^[@CR21]^ and IAV-GFP^[@CR28]^ were a generous gift from Jonathan A. McCullers (St. Jude Children's Research Hospital) and Adolfo García-Sastre (Icahn School of Medicine at Mount Sinai), respectively. Viral infection in vivo {#Sec14} ------------------------- Twenty eight-week-old C57BL/6 mice were anesthetized with isoflurane and inoculated intranasally with 50 μL containing 1 × 10^3^ PFU of A/PR8. One group of mice was mock-infected with sterile phosphate buffered saline (PBS) as a control. Mice were observed daily for weight loss and mortality for 14 days. For histopathologic analysis, dissected lung tissues were fixed in 3.5% formaldehyde and paraffin-embedded. Tissues were sliced into 6-μm-thick paraffin sections and stained with hematoxylin and eosin. To determine the level of IFN-β mRNA, mice were euthanized 4 days post-infection, and the total RNAs were purified from 30 mg of lung tissue homogenate using an RNeasy Kit (Qiagen). All animal experiments were carried out in a humane manner after receiving approval from the Institutional Animal Care and Use Committee of the University of Tsukuba, and in accordance with the regulation for Animal Experiments in University of Tsukuba and Fundamental Guideline for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions under the jurisdiction of the Ministry of Education, Culture, Sports, Science, and Technology. Luciferase assays {#Sec15} ----------------- HEK293 cells were plated in 24-well plates (2.5 × 10^5^ cells per well) in the customized media. The following day, the cells were co-transfected with 100 ng luciferase reporter plasmid (p125luc or pELAM), 2.5 ng Renilla luciferase internal control vector phRL-TK (Promega), and each of the indicated expression plasmids with Lipofectamine 2000 reagent (Invitrogen). Empty vector \[pcDNA3.1(−)\] was used to maintain equivalent amounts of DNA in each well. Cells were harvested 24 h after transfection and analyzed by a dual-luciferase reporter assay on the GloMax 20/20n luminometer (Promega). Each experiment was repeated at least three times. BRET assay {#Sec16} ---------- All BRET saturation assays were performed as previously described^[@CR16]^ with slight modifications. In brief, HEK293 cells (5 × 10^5^ cells per well) cultured under customized media were co-transfected with a constant amount (5 ng) of NLuc-tagged MAVS plasmid and increasing amounts of Venus-tagged MAVS constructs using Lipofectamine 2000 reagent. Empty vector \[pcDNA3.1(−)\] was used to maintain equivalent amounts of DNA in each well. The cells were harvested at 21 h post-transfection and transferred to each well of white 96-well microplates. NLuc substrate (furimazine; 5 μM) was added, and the plates were analyzed via a BRET saturation assay using a Flexstation 3 Microplate Reader (Molecular Devices, Sunnyvale, CA) at 37 °C. ELISA {#Sec17} ----- Measurements of species-specific IFN-β and IL-6 production were performed as described previously^[@CR20],\ [@CR36]^. Microarray analysis {#Sec18} ------------------- Total RNA was isolated from each cultured condition of primary MEFs using TRIzol Reagent (Invitrogen) and purified using SV Total RNA Isolation System (Promega) according to the manufacturer's protocol. RNA samples were quantified using an ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE) and the quality was confirmed with an Experion System (Bio-Rad Laboratories, Hercules, CA). The cRNA was amplified, labeled (Low Input Quick Amp Labeling Kit), and hybridized to the SurePrint G3 Mouse Gene Expression 8 × 60 K microarray according to the manufacturer's instructions (Agilent Technologies, Santa Clara, CA). All hybridized microarray slides were scanned by an Agilent scanner (Agilent Technologies). Relative hybridization intensities and background hybridization values were calculated using Agilent Feature Extraction Software (9.5.1.1). Raw signal intensities and flags for each probe were calculated from the hybridization intensities (gProcessedSignal), and spot information (gIsSaturated, etc.), according to the procedures recommended by Agilent. The raw signal intensities of four samples were normalized by a quantile algorithm with the 'preprocessCore' library package^[@CR37]^ of the Bioconductor software^[@CR38]^. To identify upregulated or downregulated genes, we calculated the Z-scores^[@CR39]^ and ratios (non-log scaled fold-change) from the normalized signal intensities of each probe to compare between control and experiment samples, and established criteria for regulated genes: Z-score ≥2.0 and ratio ≥1.5-fold for upregulated genes, and Z-score ≤−2.0 and ratio ≤0.66 for downregulated genes, respectively. Visualization of cytoplasmic ATP in living cells {#Sec19} ------------------------------------------------ Imaging of cytoplasmic ATP levels in living cells was performed as previously described^[@CR13],\ [@CR40]^. Briefly, cells expressing ATeam1.03 were plated on collagen-coated 35-mm glass-bottom dishes (MatTek, Ashland, MA). The following day, the cells were washed once with the customized media and the cells were subjected to ATP imaging. Observations of cells that were maintained at 5% CO~2~ and 37 °C using a stage-top incubator (Tokai Hit, Shizuoka, Japan) were performed using a Nikon ECLIPSE Ti-E microscope with an oil-immersion type CFI Plan Apo VC 60x lens (Nikon Instruments Inc, Tokyo, Japan). The filters, 438/24-nm BrightLine single-band bandpass filter (FF01--438/24--25), 458-nm edge BrightLine single-edge dichroic mirror (FF458-Di02), and two emission filters (FF01--483/32 and FF01--542/27) used for dual-emission ratio imaging of ATeam1.03 were obtained from Semrock Inc. (Rochester, NY). Fluorescence emission from the ATeam1.03 was imaged by altering the emission filters with a filter changer and a scientific CMOS camera (Zyla 4.2, Andor Technology), and the imaging analysis was performed using MetaMorph (Molecular Devices). The YFP/CFP emission ratio was calculated by dividing the YFP intensity by the CFP intensity of each cell. Confocal microscopy {#Sec20} ------------------- Cells were plated on coverslips in 12-well plates (1.5 × 10^5^ cells per well). The following day, cells were fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.2% Triton X-100 in PBS (pH 7.4), and blocked with 5% bovine calf serum. Endogenous MAVS was detected with the polyclonal primary and AlexaFluor488 secondary antibodies, and mitochondria were stained with anti-mtHsp70 primary antibody followed by the Cy3-conjugated secondary antibody. Cells were imaged with the C2^+^ confocal microscope (Nikon Instruments Inc). Electron microscopy analysis {#Sec21} ---------------------------- For electron microscopy, cells were fixed in 100 mM cacodylate buffer (pH 7.4) containing 2% paraformaldehyde and 2% glutaraldehyde (GA) for 1 h, washed and fixed again with 100 mM cacodylate buffer containing 2% GA overnight at 4 °C. After fixation, the cells were washed four times with 100 mM cacodylate buffer for 20 min each, and postfixed with 2% osmium tetroxide (OsO~4~) in cacodylate buffer at 4 °C for 1 h. Cells were then dehydrated in a graded series of ethanol (50%, 70%, 90%, and 100%), embedded in resin (Quetol 812, Nisshin EM Co., Japan), and polymerized at 60 °C for 48 h. The polymerized resins were cut in ultra-thin sections at 70 nm with an LEICA UTC ultramicrotome, mounted on copper grids, and stained with 2% uranyl acetate and lead solution (Sigma-Aldrich). Images were collected with a transmission electron microscope (JEM-1400Plus; JEOL Ltd, Japan) operating at 80 kV and equipped with a CCD camera (VELETA; Olympus Soft Imaging Solutions GmbH, Germany). Quantification of mtDNA {#Sec22} ----------------------- Total genomic DNA was extracted and purified from cells, and diluted to 5 ng/μL. To quantify the amount of mtDNA per nuclear DNA, quantitative real-time polymerase chain reaction (qPCR) was performed using FastStart Essential DNA Green Master (Roche). Quantification of relative copy number differences was performed based on the difference in the threshold amplification between mtDNA and nuclear DNA \[ΔΔC(t) method\]. We used the following sets of primers for mtDNA: TK702 (5′-cctatcacccttgccatcat, forward) and TK703 (5′-gaggctgttgcttgtgtgac, reverse); and for nuclear DNA: TK704 (5′-atggaaagcctgccatcatg, forward) and TK705 (5′-tccttgttgttcagcatcac, reverse). Cytochemical analysis of cytochrome c oxidase (COX) activity {#Sec23} -------------------------------------------------------------- COX staining was performed as previously described^[@CR41]^ with slight modifications. MEFs plated on 18-mm coverslips were fixed with 2% paraformaldehyde in PBS for 15 min and washed twice with PBS for 5 min each. To visualize COX activity, cells were stained with 100 mM phosphate buffer (pH 7.4) containing 0.6 mg/mL 3′,3′-diaminobenzidine, 0.3 mg/mL cytochrome c, and 45 mg/mL sucrose at 37 °C for 2 h. Cells were then washed twice with 0.1 M Tris-HCl buffer (pH 8.0) for 5 min each and analyzed by microscopy. Statistical analysis {#Sec24} -------------------- An analysis of variance test (GraphPad QuickCalcs) was used for the statistical analyses. A P-value of less than 0.05 was considered statistically significant. Electronic supplementary material ================================= {#Sec25} Supplementary Information Electronic supplementary material Supplementary information accompanies this paper at doi:10.1038/s41598-017-05808-w Accession code: The GEO accession number for the microarray data reported in this paper is GSE75399. Publisher\'s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. We are grateful to Thomas Langer (University of Cologne) for helpful discussions regarding the study. We also thank Katsuyoshi Mihara (Kyushu University), Yuichi Michikawa (National Institute of Radiological Sciences, Japan), and Mitsuyasu Kato (University of Tsukuba) for their valuable comments on the study. We appreciate Yuko Fuchigami and Norio Sasaoka for their technical assistance with DNA cloning, sequencing, establishing the cell line expressing ATeam1.03, and technical comments on the figure preparation. We are also grateful for technical support regarding the fluorescence measurements (BRET assay) from the Research Support Center, Graduate School of Medical Sciences, Kyushu University. The 143B COX and ND1 cybrids were a generous gift from Giovanni Manfredi (Cornell University) and Valerio Carelli (University of Bologna, Italy), respectively, and OPA1-null MEFs were obtained from David Chan (HHMI and Caltech). This work was supported by the JSPS KAKENHI Grants (No. 15H01265, 17H03667, 26291032, and 26620135 to T.K.), the Daiichi Sankyo Foundation of Life Science, the Japan Foundation for Applied Enzymology, the Takeda Science Foundation, the Naito Foundation, the NOVARTIS Foundation (Japan) for the Promotion of Science, the Life Science Foundation of Japan, and KANEKA Corporation to T.K, and partly supported by a Grant-in-Aid for JSPS Fellow (T.Y.). T.Y. and T.K. performed most of the in vitro experiments. H.I. established the biosensor ATeam1.03 and analyzed the cellular ATP level. K.N. and K.I. established the Mito-miceΔ and ρ~0~ cells from cultured human and murine cells. A.K., T.T., and T. Kuroki performed the mouse infection experiments and K.I. analyzed the histopathologic analysis. T.Y., H.I., A.K., T.T., T. Kuroki, K.I., K.N., and T.K. analyzed the data and interpreted all the experimental results. T.K. designed the research program and contributed to writing the paper. Competing Interests {#FPar1} =================== The authors declare that they have no competing interests.	{ "pile_set_name": "PubMed Central" }
Sir, Central venous pressure (CVP) monitoring is a simple, relatively inexpensive method of assessing a patient's circulating blood volume, cardiac status and vasomotor tone. It is essential to be aware of the inherent fallacies and inadequacies of the information derived. Inaccurate measurements are often obtained by the aberrant lodgement of the central venous catheter (CVC) tip. In continuation of the previously published letter to editor concerning misdirected CVC, we describe an unusual case of CVC coiling in the superior vena cava (SVC) leading to falsely high CVP measurement. A 55 year old male patient was brought to the emergency room (ER) with head injury, blunt trauma abdomen with haemodynamic instability. He was further posted for an emergency laparotomy. In view of the clinical condition of the patient and the need to know intravascular volume status, a 7 French triple lumen CVC was inserted in the right internal jugular vein (IJV) in the operating room (OR). All the three ports were checked for free flow of blood and the CVC was fixed at 11cm at skin level. On connecting the transducer to the monitor, ideal waveform was absent. Intra operatively CVP tracing was suboptimal despite the change of transducer, the cable, flushing the unit and repeated zeroing. Post operatively the patient was shifted to intensive care unit on ventilator support for further management. Chest radiograph revealed coiled CVC in the SVC \[[Figure 1](#F0001){ref-type="fig"}\]. Hence it was removed and right subclavian vein was cannulated. ![Coiled CVC in superior vena cava](IJA-54-351-g001){#F0001} The correct placement of the CVC tip is an important factor in obtaining accurate CVP measurements. Malposition of a CVC may occur at the time of insertion or later as a result of spontaneous migration due to anatomic positioning or pressure changes within the thoracic cavity.\[[@CIT1]\] There has been case report on CVC folding back during guide-wire removal inside IJV.\[[@CIT2]\] In our case CVC coiling inside SVC was unusual as it is a large calibre vessel with high flows. There was no anatomical vascular abnormality and no manufacturing defect in CVC. The abutting of the guide wire against the wall of the SVC probably caused the coiling of CVC in the SVC. The J tip of guide wire probably was unknowingly directed cephalad while insertion, which could have caused CVC to further angulate in the upward direction over the guide wire. In conclusion, inaccurate CVP measurements or inability to obtain an ideal wave from tracing are suggestive of an undesirable location of the catheter tip. Awareness of this possibility and careful review of the CVC tip position on X-ray pictures in suspicious cases are important. Inaccurate CVP readings lead to improper assessment of the intravascular status of the patient. Careful clinical co-relation under such circumstances is essential. Roentgenograms after insertion of CVC are essential to eliminate this problem, which is often encountered in clinical practice.	{ "pile_set_name": "PubMed Central" }
Introduction ============ The effectiveness of drug eluting stents for the prevention of restenosis after percutaneous coronary interventions (PCI) relies on the potent effects of cell cycle inhibitors on vascular smooth muscle cell (VSMC) proliferation ([@B2]). Intimal hyperplasia, however, remains a significant clinical problem in for example coronary artery bypass graft failure ([@B17]). In addition to direct effects on VSMCs, intimal hyperplasia is also driven by inflammation, by means of neutrophil, and macrophage infiltration, as well as cytokine and matrix metalloproteinase (MMP) release ([@B17]). In particular, a failure in the resolution of the acute inflammatory response to vascular injury prevents re-endotheliazation and promotes VSMC proliferation and migration ([@B18]). Omega-3 fatty acids decrease inflammation ([@B15]), and inhibit intimal hyperplasia in mice ([@B7]). Other studies have shown similar effects after administration of the pro-resolving lipid mediators resolvins and maresins, which are enzymatically formed from omega-3 fatty acids ([@B5]; [@B1]; [@B8]; [@B19]). The latter studies provided the initial evidence that stimulating the resolution of inflammation by means of lipid pro-resolving mediators would promote an adequate healing of vascular injury. However, the receptor(s) involved in this protective effects in vivo remain unknown. We recently established a protective role for the G protein-coupled receptor ChemR23 for the omega-3-derived pro-resolving lipid mediator resolvin E1 in atherosclerosis ([@B6]), but its implications for intimal hyperplasia have remained hitherto unexplored. The aim of the present study was therefore to establish the role of ChemR23 in the downstream signaling of omega-3 fatty acids in intimal hyperplasia, in a pro-inflammatory vascular injury murine model ([@B20]). Methods ======= Carotid Ligation ---------------- The study was approved by the Regional Ethical Review Board in Stockholm. All animals used were male and on a C57BL/6J background. ChemR23^-/-^ mice were obtained from Deltagen. Mice expressing the Caenorhabditis elegans fat-1transgene were bred as previously described ([@B9]). The two mice strains were then crossbred to generate four groups; ChemR23^+/+^, ChemR23^-/-^, fat-1/ChemR23^+/+^, and fat-1/ChemR23^-/-^. 10 weeks old littermates (n = 6--8/group) were subjected to a complete left carotid ligation as previously described ([@B12]) and indicated in Figure [1A](#F1){ref-type="fig"}. In brief, the four groups were fed in a 10% v/w Omega-6 (Sigma-aldrich, S8281) enriched diet to increase the fat-1 desaturase substrate. EPA and DHA were quantified by gas chromatography as a control as previously described ([@B6]), and exhibited the expected increase in fat-1 transgenic mice independently of ChemR23 expression (data not shown). After 7 days, mice were anesthetized with isoflurane/O~2~ (2:1) followed by 0.1 mg s.c. injection of buprenorphine for pain relief. Left common carotid artery was exposed, followed by a complete ligation at the bifurcation level with a 7-0 suture. After 28 days from ligation, mice were euthanized by CO2, PFA fixated, and the ligated carotid was collected in PFA, and paraffin embedded. For carotid intimal hyperplasia evaluation, 8 sections of 10 um each every 100 um, were collected. Next, H&E staining was performed to assess neointimal growth area at the site of ligation. Furthermore, sections were stained with antibodies (Supplementary Table [1](#SM1){ref-type="supplementary-material"}) for rabbit anti α-SMA (Abcam), rat anti CD206 (Serotec), rat anti Mac2 (Cedarlane), and rat anti Ly6G (BD) for macrophage and neutrophil content determination, respectively. Negative and isotype controls are shown in Supplementary Figure [2](#SM1){ref-type="supplementary-material"}. Staining in the neointima was assessed using the automated software Leica QWin Standard Y 2.8 (Leica Microsystems) and normalized to neointima area. Rat and rabbit isotype controls were purchased from R&D, and Abcam, and secondary antibodies from Vector. ![ChemR23 deletion promotes intimal hyperplasia under pro-inflammatory conditions. (A) Schematic representation of in vivo experimental procedure. (B) Mouse intima hyperplasia quantification in: ChemR23^+/+^ n = 7, ChemR23^-/-^ n = 7, fat-1^tg^ ChemR23^+/+^ n = 6, fat-1^tg^ ChemR23^-/-^ n = 8; after left common carotid ligation, and representative H&E stained photomicrographs. (C) Ly6G^+^ (neutrophil) ChemR23^+/+^ n = 5, ChemR23^-/-^ n = 7 and Mac-2 (macrophage) ChemR23^+/+^ n = 4, ChemR23^-/-^ n = 7 immunohistochemistry quantification, and representative photomicrographs. Data represent mean ± SEM. P-values derive from (B) 2-way ANOVA, (C) Student's t-test.](fphar-09-01327-g001){#F1} VSMC Isolation and Evaluation of Proliferation ---------------------------------------------- Abdominal aortas from ChemR23^+/+^ and ChemR23^-/-^ mice (n = 3/group) were isolated, fat and adventitia removed, and digested in a sterile mixture of 1 mg/mL collagenase type II (Worthington) and 0.3 mg/mL elastase (Sigma, E0127) in DMEM with 10% FBS for 90 min at 37°C and 5% CO~2~. Cell suspension was spun down, resuspended in complete medium (DMEM, 10% fetal calf serum, 100 units/ml penicillin, 100 μg/ml streptomycin, 1 mM sodium pyruvate, 10 mM HEPES, and 2 mM [L]{.smallcaps}-glutamine) and plated. Cells were passaged using trypsin when they reached 80% confluency. Proliferation was assessed by WST-1 reagent (Roche) according to manufacturer's protocol. Peritoneal Macrophages Conditioned Media Generation --------------------------------------------------- Peritoneal macrophages from ChemR23^+/+^ and ChemR23^-/-^ mice (n = 4/group) were obtained as previously described ([@B6]). Macrophages were treated for 24 h with complete medium supplemented with LPS (100 ng/ml), washed and followed by a 24 h incubation in complete medium without LPS. After that, cell supernatant was collected and frozen. After thawing, supernatants were diluted 1:10 in complete medium and transferred to ChemR23^+/+^ VSMCs. RNA Extraction and Real-Time PCR -------------------------------- RNA from ChemR23^+/+^ and ChemR23^-/-^ peritoneal macrophages was isolated after 24 h of LPS (100 ng/ml) stimulation using the RNeasy Mini Kit (Qiagen). RNA concentration was quantified by Nanodrop (Thermo Scientific). Relative gene expression was assessed using Taqman assays from Life Technologies: GAPDH as endogenous control (Mm99999915), TNF-α (Mm00443258), IL-6 (Mm00443258), MMP9 (Mm00442991), IL-1β (Mm00434228), and IL-10 (Mm01288386). Statistics ---------- Results are expressed as mean ± S.E.M. Statistical significance was assigned at p \< 0.05 as assessed with Student t-test when comparing two groups, and with two-way ANOVA as appropriate followed by recommended post hoc tests, for multiple comparisons. All analyses were performed using GraphPad Prism 7 (GraphPad Software Inc., CA, United States). Results ======= We here report for the first time that genetic disruption of ChemR23 significantly increased intimal hyperplasia (Figure [1B](#F1){ref-type="fig"}). Furthermore, and as predicted, mice expressing the Caenorhabditis elegans fat-1 transgene, which enables the endogenous production of omega-3 fatty acids, exhibited reduced intimal hyperplasia (Figure [1B](#F1){ref-type="fig"}). Crossbreeding of the two models also allowed us to determine the interaction between the observed effects attributed to omega-3 fatty acids and ChemR23, respectively. Unexpectedly, the fat-1 transgene was protective in both ChemR23^+/+^ and ChemR23^-/-^ mice (Figure [1B](#F1){ref-type="fig"}). Immunohistochemistry in intimal lesions revealed an infiltration of neutrophil granulocytes (Ly6G) and macrophages (Mac-2), which were not significantly different between ChemR23^+/+^ and ChemR23^-/-^ mice (Figure [1C](#F1){ref-type="fig"}), whereas no T-lymphocyte infiltration (CD3) was detected (data not shown). Analysis of CD206, as a marker of M2 macrophages, revealed no significant differences between ChemR23^+/+^ and ChemR23^-/-^ mice in the percentage of CD206 positive cells in the intima (1,52 ± 0,47, 1.26 ± 0.38, p = 0.682), nor in the ratio between M2 and total macrophages (CD206/Mac-2) (31.17 ± 9.66, 40.98 ± 12.30, p = 0.538) (Supplementary Figure [1A](#SM1){ref-type="supplementary-material"}). Finally, intimal lesions stained positive for α-smooth muscle actin, but revealed no differences in the percentage of α-smooth muscle actin in the intima between ChemR23^+/+^ and ChemR23^-/-^ mice (Supplementary Figure [1B](#SM1){ref-type="supplementary-material"}). To decipher the mechanisms involved in the phenotype associated with ChemR23-deficiency, we subsequently isolated abdominal aortic VSMCs from ChemR23^+/+^ and ChemR23^-/-^ mice. ChemR23^-/-^ VSMCs exhibited significantly less proliferation compared with ChemR23^+/+^ VSMCs (Figure [2A](#F2){ref-type="fig"}). These in vitro findings were hence in sharp contrast to the increased intimal hyperplasia observed in ChemR23-deficient mice in vivo. To further characterize the ChemR23-dependent phenotype, we subsequently studied the inflammatory response in LPS-activated peritoneal macrophages, isolated from ChemR23^+/+^ and ChemR23^-/-^ mice. ChemR23^-/-^ macrophages exhibited a more pro-inflammatory phenotype, with significantly higher mRNA levels of TNFα, and MMP9 (Figure [2B](#F2){ref-type="fig"}). These results indicated that in intimal hyperplasia, ChemR23^-/-^ VSMCs may have a protective phenotype (less proliferation), whereas ChemR23^-/-^ macrophages would be detrimental (increased inflammation). To assess the consequences of the differential macrophage phenotypes on VSMC proliferation, VSMCs were treated with CM derived from LPS-activated ChemR23^+/+^ and ChemR23^-/-^ macrophages, respectively. ChemR23^-/-^ macrophage CM significantly increased VSMC proliferation as compared with VSMC treated with ChemR23^+/+^ media (Figure [2C](#F2){ref-type="fig"}). ![ChemR23 deletion alters the Vascular smooth muscle cells (VSMC) and macrophage phenotype. (A) Basal ChemR23^+/+^ and ChemR23^-/-^ VSMC proliferation (48h) in vitro, assessed by WST-1. (B) mRNA expression of LPS-activated (100 ng/mL, 24 h) peritoneal macrophages in vitro. (C) VSMC proliferation (48 h) treated with conditioned media (CM) derived from LPS-activated ChemR23^+/+^ and ChemR23^-/-^ macrophages. For VSMC n = 3/group, for macrophages n = 4/group. Data represent mean ± SEM. P-values derive from Student's t-test.](fphar-09-01327-g002){#F2} Discussion ========== Three major observations emerge from the present study. First, our results revealed a protective role of ChemR23 in intimal hyperplasia, which was unrelated to the beneficial effects of enriching tissues with omega-3. Secondly, we identify ChemR23 as a transducer of opposing effects in different cell types, in terms of suppressing inflammatory activation of macrophages, but stimulating proliferation of VSMCs. Third, ChemR23^-/-^ macrophages in turn promoted VSMC proliferation. Taken together, these results suggest that ChemR23 signaling toward the resolution of inflammation protects from intimal hyperplasia by means of reducing macrophage activation. The observation that genetic ChemR23 deletion increased the inflammatory response in macrophages supports previous reports ([@B3]; [@B10]) and is consistent with our recent findings that ChemR23 deletion in hyperlipidemic mice accelerated atherosclerosis ([@B6]). The present study extends that observation by showing that a failure in the resolution of macrophage-induced inflammation by means of ChemR23 deficiency increased VSMC proliferation. This effect was independent of an increased total number of macrophages or M2 macrophage subtype. In accordance with this findings, in atherosclerotic lesions ChemR23 expressing macrophages do not correspond to an M2 phenotype ([@B6]). Indeed, previous results indicate that ChemR23 expression in M1 macrophages acts to promote the resolution of inflammation ([@B4]). Altogether the expression of ChemR23 in both macrophages and smooth muscle cells in human atherosclerotic lesions ([@B6]), support an extrapolation of these results to human pathophysiology. Previous reports have shown that pro-resolving lipid mediators decrease VSMC proliferation ([@B19]), and that VSMC lacking the lipoxin, and D-resolvin receptor ALX/FPR2 exhibit decreased migration ([@B12]). In the present study, however, ChemR23 deficient VSMCs proliferated less compared with wild-type VSMCs. Hence, in the absence of a macrophage-derived response, ChemR23 signaling may in contrast promote VSMC proliferation. These observations suggest that while ChemR23 signals to limit macrophage-induced inflammation, ChemR23 may transduce deleterious effects on VSMCs under non-inflammatory conditions. Whether different agonists transduce those differential responses, however, remains to be established. In general, pro-resolving agonists are associated with a beneficial smooth muscle response in vascular injury ([@B5]; [@B19]), abdominal aortic aneurysms ([@B14]; [@B13]) and atherosclerosis ([@B16]; [@B11]). The reduced intimal hyperplasia by ChemR23 appeared independently of the beneficial effects of omega-3 fatty acids, and suggest ChemR23 expression being directly coupled to the VSMC, and macrophage phenotypes. Other studies have indeed implicated signaling through the free fatty acid receptor-4 to mediate beneficial effects of omega-3 fatty acids in and murine models of intimal hyperplasia ([@B7]). In conclusion, the opposing effects of ChemR23 on VSMC and macrophages reported in the present study raise a novel notion for intimal hyperplasia pathophysiology, namely that the same receptor may transduce both protective and deleterious effects, which may vary over time depending on different stages in the resolution of inflammation. Ethics Statement ================ This study was carried out in accordance with the recommendations and guidelines of the Regional Ethical Review Board in Stockholm. The protocol was approved by the "Regional Ethical Review Board in Stockholm." Author Contributions ==================== GA, MC, and AL-F designed and performed the experiments, and analyzed the data. MB, GA, MC, and AL-F conceived the study. GA, MC, and MB wrote the manuscript. All authors participated in the interpretation of the data and provided critical review of the manuscript. Conflict of Interest Statement ============================== The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Funding. This work was the supported by grants from the Swedish Research Council (Grant No. 2014--2312), the Swedish Heart and Lung Foundation (Grant Nos. 20150600 and 20150683), King Gustaf V and Queen Victoria Freemason Foundation, the Stockholm County Council (Grant No. 20170365), and Marianne and Marcus Wallenberg Foundation (2015.0104). Supplementary Material ====================== The Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fphar.2018.01327/full#supplementary-material> ###### Click here for additional data file. [^1]: Edited by: Lucy V. Norling, Queen Mary University of London, United Kingdom [^2]: Reviewed by: Asif Jilani Iqbal, University of Birmingham, United Kingdom; Soon Yew Tang, University of Pennsylvania, United States [^3]: ^†^These authors have contributed equally to this work [^4]: This article was submitted to Inflammation Pharmacology, a section of the journal Frontiers in Pharmacology	{ "pile_set_name": "PubMed Central" }
Background {#S0001} ========== Female lower abdominal pain has long posed diagnostic difficulties for clinicians, especially when they have had little more than the history and physical examination to guide them. Even first-trimester pregnancy diagnoses were expensive and time consuming until disposable qualitative immunological β-HCG assays became widely available in the 1980s. More accurate diagnoses have resulted from using transabdominal and transvaginal ultrasonography, computerized tomography, and diagnostic laparoscopy. These modalities, however, usually are not available in resource-poor areas. When they are lacking, as in the case presented here, clinicians may continue to follow the wrong diagnostic path, leading to unnecessary surgical procedures and potentially devastating treatment delays. Patient and case report {#S0002} ======================= A 19-year-old sexually active high school student with a history of regular menses believed she was pregnant after missing her period for three months. Still in school and not ready to have a child, she wanted to terminate the pregnancy but was reluctant to visit a healthcare professional. She could not pay for such a visit and even though she had government-sponsored medical insurance, it did not cover abortion services. Her girlfriend, the only person in whom she confided, suggested misoprostol for termination; she had never been pregnant or previously used misoprostol. A Licensed Chemical Seller (LCS) sold her misoprostol without a prescription. The fourth of 6 siblings, the patient had been sent to school from a rural farming community and lived near the school with her two younger sisters. Her first coitus was at 16 years, but she had never used birth control or condoms, claiming that she did not believe that she would get pregnant or contract sexually transmitted diseases (STDs). Her religious beliefs did not influence her behaviour. Vaginal bleeding began a few hours after taking 200 mcg misoprostol orally and the same dose vaginally. She bled for four days, accompanied by severe constant lower abdominal pain. Nevertheless, she remained at home. Five days later, the bleeding subsided but right sided pelvic pain persisted. Two days later, she went to a hospital near her home, where she was given the clinical diagnosis of septic incomplete abortion with severe anaemia (Hgb 7.3 g/dl). They had no ultrasound or transfusion capability. Twenty-four hours after performing a manual vacuum aspiration (MVA) with no observable products of conception, they transferred the patient with a diagnosis of pelvic abscess. Despite her level of illness, she came to our district hospital, about 20 minutes away by taxi, the common transport for ambulatory patients. On arrival, the thin and small-for-her-age woman looked pale and obviously ill. Her vital signs were temperature 37.8°C, BP 100/60 mm/Hg, pulse 100/min, regular and good volume. Her abdomen was diffusely tender with peritoneal signs. Since the one surgeon for the facility was operating, she was sent to the maternity ward, where the nurses, immediately recognizing her level of illness, started an IV, drew labs, and notified the surgeon. She was found to have a Hgb 6.1 g/dl, WBC 15.9 x109/L, with 10.8% lymphocytes, 9.0% mononuclear cells, and 80.2% neutrophils. A physician-performed ultrasound transabdominal examination showed an empty but bulky uterus with fluid in the dependant parts of the abdominal cavity. The presumptive diagnosis was pelvic abscess, although other processes were considered that also required surgery. She received two units of whole blood, additional crystalloids, ciprofloxacin and metronidazole. Under general anesthesia, exploration through a Pfannenstiel incision demonstrated that the uterus, fallopian tubes and both ovaries were normal, but that the peritoneal cavity contained a large quantity of serosanguinous fluid. The incision was extended in an inverted "T" to the epigastrium for a more extensive laparotomy. After draining the fluid, an inflamed appendix was found adhering to the cecum and no other pathology was identified. An appendectomy was performed ([Figure 1](#F0001){ref-type="fig"}), the peritoneal cavity was lavaged and the abdomen was closed primarily. ![Inflamed appendix in a girl treated for missed appendicitis after self-induced abortion](PAMJ-10-38-g001){#F0001} Post-operative course {#S20001} --------------------- Post-operative management included prolonged use of parenteral antibiotics. During the post-operative healthcare education, the patient stated that she did not intend to use either contraception or STD protection in the future, relying instead on sexual abstinence. Discussion {#S0003} ========== Presentation {#S20002} ------------ This patient presented to a relatively resource-poor hospital after treating herself for a self-diagnosed pregnancy. Although she never took a pregnancy test, her self-diagnosis was reasonable given her menstrual and sexual history. Doing a pregnancy test at the initial hospital would have been positive if she were still pregnant or had recently aborted. The persistent right lower quadrant abdominal pain should, however, have steered those clinicians in the direction of immediate referral, rather than first subjecting the patient to an MVA and then delaying her transfer by 24 hours. While lower abdominal pain, usually colicky, is often associated with misoprostol use, the cramps would have dissipated long before this patient presented to the hospital. Misoprostol {#S20003} ----------- Misoprostol is a prostaglandin analogue with uterotonic properties. It is 94% to 98% effective in inducing abortions when used orally, vaginally, or both. The failure rate is greater if used after 9 weeks gestation \[[@CIT0001],[@CIT0002]\]. The drug commonly causes vaginal bleeding within 12 hours of administration, with vaginal bleeding and uterine cramps being the most common reasons for women to seek medical care\[[@CIT0003]\]. In this case, misoprostol use may have complicated and delayed appropriate treatment. Under Ghanaian law, however, the patient should never have obtained misoprostol without first seeing an experienced physician. Misoprostol is a class C medication that may only be prescribed by doctors at a District Hospital or higher level institution, and only dispensed by registered pharmacists. In this case, an LCS illegally provided her with misoprostol. LCS\'s usually have only a few weeks of training and staff small retail outlets that sell over-the-counter medication. The LCS could be fined or their outlet closed down for selling misoprostol, although it is unclear whether this occurred. ### Diagnostic Process {#S30001} The differential diagnosis of persistent abdominal pain, especially following menstrual irregularities in a 19-year-old woman, has the potentially life-threatening ectopic pregnancy near the top of the list. It should also include all other causes of abdominal pain in an adult woman, including acute appendicitis. Acute appendicitis is the most common general surgical problem encountered during pregnancy \[[@CIT0004]\]. The incidence averages 1 in 550 pregnancies \[[@CIT0005]\]. With a recently pregnant patient, the potential dangers inherent in missing an ectopic pregnancy diagnosis would have been serious. It\'s unclear why that possibility wasn\'t considered, with the patient transferred immediately for diagnosis and treatment. When the patient arrived at our hospital, her clinical picture plus free fluid on the ultrasound made it clear that she required surgery, probably for an ectopic pregnancy or, given her lack of STD protection, a pelvic abscess. Appendicitis was a lesser consideration but, as Hickam\'s Dictum tells us, patients can have as many diseases as they wish \[[@CIT0006]\]. This contrasts with the more commonly used and cited principle of Ockam\'s Razor, which says that "of two explanations, each capable of explaining a given phenomenon, one should choose the simpler of the two" \[[@CIT0007]\]. Or, in other words, the hoof beats you hear probably mean horses, not unicorns. Even in non-pregnant women, however, it can be very difficult to differentiate appendicitis from pelvic inflammatory, gastrointestinal, and urinary tract disorders. Up to 1/3 of these patients are initially misdiagnosed. Such misdiagnosis is associated with an increased incidence of perforation and abscess formation, as well as prolonged hospitalization \[[@CIT0008]\]. Differentiating appendicitis from gynecological disease often requires adequate ultrasound capability (the hospital had only transabdominal probes on a non-moveable machine) and laparoscopy (not available)\[[@CIT0009]\]. One would imagine that this brush with death would alter the patient\'s view of safe sexual activity. Yet, when receiving education from nurses on these practices, this relatively well-educated patient demonstrated a cavalier attitude toward both birth control and sexually transmitted diseases. She asserted that she would rely on abstinence. Unfortunately, history shows that this will most likely lead to both pregnancy and sexually transmitted diseases. Conclusion {#S0004} ========== System problems in resource-poor areas can limit access to simple laboratory and healthcare services due to their unavailability in healthcare facilities or patients' lack of healthcare education or poverty. One problem that is difficult to remedy is unsafe sexual practices, even among better educated young women. In many countries, those selling medications often subvert official regulations and dispense potentially dangerous medications without a clinician\'s authorization. This leads to many problems, not the least of which is obscuring or delaying serious diagnoses and appropriate treatment. Clinicians must be wary of relying on patients' self-diagnoses while disregarding others that may better explain their symptomatology. The differential diagnosis for a woman\'s lower abdominal pain is extensive; neglecting other potential diagnoses risks increased morbidity (as in this case) or mortality. Sometimes clinicians must make more than one diagnosis to explain temporally and anatomically related symptoms. Authors' contributions {#S0005} ====================== Both authors participated in diagnosing and treating the patient in this case, drafting this article, and approving the final version for publication. Competing interests {#S0006} =================== The authors declare no competing interests.	{ "pile_set_name": "PubMed Central" }
Introduction ============ Elevated blood glucose concentration is an important factor for mortality and morbidity in critically ill patients. Previously, a randomized controlled trial showed that a blood glucose level of about 6 mmol/l is associated with less multiorgan failure and a significantly higher survival rate compared with levels above 8 mmol/l. Platelets play a crucial role in hemostasis and inflammation. However, the effect of short-term elevated blood glucose levels on platelet activation has not yet been evaluated systematically. Objective ========= To evaluate the influence of blood glucose levels on platelets in vitro. Methods ======= Citrated blood samples were drawn from healthy blood donors (40% male, age 38 ± 13 years \[mean ± SD\]). Exclusion criteria were smoking and the use of drugs interfering with platelet function. Blood samples were adjusted with glucose (Sigma, Taufkirchen, Germany) to final concentrations of 5 mmol/l (control group), 10 mmol/l (group 1) and 15 mmol/l (group 2), respectively. Samples were incubated for 10 min at 37°C with fluorescence-labeled monoclonal antibodies against CD62P, CD41, CD36, or CD42b (all: Beckman-Coulter, Krefeld, Germany). To evaluate platelet reactivity 2 and 6 μM thrombin-receptor-agonist-peptide-6 (TRAP-6; Bachem, Heidelberg, Germany) or 5 and 10 μM adenosine-di-phosphate (ADP; Sigma, Taufkirchen, Germany) were added. Analyses were performed in a flow-cytometer (EpicsXL; Beckman-Coulter). The mean fluorescence intensity was calculated. Determination of platelet aggregation was performed by the turbidimetric procedure (BCT, Dade Behring, Marburg, Germany). Aggregation was induced with ADP (200 μM/l), collagen (2 mg/l) and epinephrine (100 μM/l; all Dade Behring). Statistics for intergroup differences were performed by one-way ANOVA. Results ======= The initial blood glucose concentration was 5.0 ± 1.1 mmol/l. The blood glucose level had no significant influence on the expression of CD36 and CD62P, with and without stimulation. By contrast we observed a significant decrease in expression of CD42b in group 2 compared with the control group (unstimulated, P\< 0.001; TRAP-6, P= 0.005; ADP, P= 0.012). A similar observation was made for CD41 expression (unstimulated, P\< 0.001; TRAP-6, P\< 0.05; ADP, P\< 0.001). Also in group 1, a significant decrease in CD41 expression was observed after ADP stimulation. No significant differences were seen by aggregometry with either agonist. Conclusion ========== This in vitrostudy demonstrates that elevated blood glucose levels reduce expression of platelet receptors CD41 and CD42b. In contrast, platelet function measured by aggregometry showed no impairment. We conclude that acute hyperglycemia does not lead to lowered platelet function.	{ "pile_set_name": "PubMed Central" }
1. Introduction {#sec1} =============== Pluripotent stem cells are defined by their ability to undergo self-renewal and capability of differentiating into all three germ layers (mesoderm, endoderm, and ectoderm). Even though during development pluripotency is a temporary condition, ad hoc cultures can sustain indefinite self-renewal of cells isolated from the inner cell mass (embryonic stem cells, ESCs) in vitro \[[@B1]\]. Also, it is now well assessed that somatic cells can be reprogrammed back to the pluripotent state, obtaining the so-called induced pluripotent stem cells (iPSCs), thus reverting the physiological differentiation process \[[@B2]\]. Given the potential of pluripotent cells in regenerative medicine applications, in the recent years, a big research effort has been put in understanding the molecular mechanisms behind ESCs decision-making. Outcomes of this research are crucial to define optimal culture conditions to push cells into the desired pluripotent or differentiated state, to optimise somatic cell reprogramming, to better understand in vivo development, and to guide the use of reprogrammed cells for regeneration purposes. High variability in terms of functionality, gene expression, and epigenetic signature has been highlighted as a peculiar feature of both ESCs and iPSCs \[[@B3]\]. Focusing on gene expression variability in undifferentiated, isogenic mouse ESCs (mESCs), a number of pluripotency-related genes have been shown to be expressed heterogeneously and to present temporal fluctuations in mESCs cultured in standard serum/leukemia inhibitory factor (LIF) medium. Nanog, a master regulator of pluripotency and development \[[@B4]--[@B6]\], was the first pluripotency gene for which heterogeneity and temporal fluctuations were observed \[[@B7], [@B8]\]. This was followed by the discovery of heterogeneous expression of other pluripotency factors, such as T-box 3 \[[@B9]\], zinc-finger protein 42 (also known as Rex1) \[[@B10]\], Klf4 \[[@B10]\], Stella \[[@B11]\], Esrrb \[[@B12]\], and β-catenin \[[@B13]\]. Importantly, the mosaic and interconvertible distribution of pluripotency genes often correlates with different degrees of potency; at population level, mESCs are fully pluripotent, but subpopulations show a different predisposition towards differentiation. Alternative media to serum/LIF have been proposed \[[@B14]\]. The so-called 3i/LIF medium is serum free and contains 3 small molecules: PD0325901 (hereafter named PD), a MEK inhibitor \[[@B15]\]; SU5402, a FGF receptor inhibitor \[[@B16]\]; and CHIR99021 (Chiron), a glycogen synthase kinase-3 (Gsk3) inhibitor \[[@B17]\]. Intriguingly, most pluripotency genes become overall homogeneous in mESCs cultured in the 3i/LIF medium, even in the absence of FGF inhibitor (i.e., in 2i/LIF medium), and the "ground state" of pluripotency is achieved \[[@B14], [@B18], [@B19]\]. Is heterogeneity of Nanog and other pluripotency factors an inherent property of mESCs, fundamental for their competence to choose different fates \[[@B20]\], or is it an impediment to achieve standardizes cultures \[[@B21]\]? How do heterogeneity and complex temporal patterns of Nanog originate, and how do different culture media regulate them? In the recent years, both experimental and computational studies have tried to address these questions. In vitro cultures are an artificial approximation of in vivo systems and self-renewal is not part of in vivo development; still, the range of pluripotent states that exists in vitro as a response to different culture conditions \[[@B1]\] mimics different stages of development (preimplantation embryo and late blastocyst for 2i/LIF and serum/LIF cultures, resp. \[[@B22], [@B23]\]); thus, experimental and computational studies of gene expression patterns and dynamics in both media are relevant. Notably, many of the metastable mESC genes are, directly or indirectly, regulated by Nanog \[[@B20]\]; thus, in this review, we provide an updated overview of the major recent computational attempts to explain origins and functions of Nanog dynamics in mESCs, alongside with a review of available experimental data. We show that mathematical models can both aid in elucidating the mechanisms behind complex temporal gene expression dynamics and generate testable predictions. We start reviewing works that use small transcriptional gene regulatory networks (GRNs) to explain Nanog dynamics and their role in mESC fate decision; then, we report non-GRN approaches; finally, we discuss open questions in the field and possible future research directions. 2. Dynamics of mESCs Explained through Transcriptional Regulatory Networks: Multistable Models for mESC Pluripotency {#sec2} ==================================================================================================================== Notable is the work of Peterson\'s group in developing computational models to understand the role of transcription factor dynamics in mESC decision-making. We firstly review two early Peterson\'s works which consider the mESC switch between the pluripotent and the differentiated states a bistable system; Nanog heterogeneity is not considered, but the basis for formalisms developed later by the same authors to recapitulate more complex dynamics are given. In 2006 \[[@B24]\], a nonlinear ordinary differential equation (ODE) model, based on the Shea-Ackers formalism \[[@B25]\], was developed to describe mESC decision-making in view of the dynamics of the core pluripotency network, formed by Oct4, Sox2, and Nanog \[[@B26]\]. The topology is composed of nested positive feedback loops, which respond to environmental signals (inputs) and act on target genes (outputs), generically indicated as stemness and differentiation genes. Oct4 and Sox2 form a heterodimer, which positively regulates both genes and Nanog \[[@B27]\]; also, Nanog activates itself, Oct4, and Sox2 ([Figure 1()](#fig1){ref-type="fig"}). The ODE model describing such network predicts correlated dynamics of Nanog and Oct4-Sox2. Particular input signals (such as addition of LIF or Wnt to culture) can switch both on, pushing cells to a pluripotent state, while different inputs (such as p53) can switch them off. The system shows bistability; the pluripotent and differentiated steady states are stable and mutually exclusive, meaning that the system can converge to one of the two, depending on the initial conditions and inputs. Model analysis highlights a central role of Nanog autoregulation for the core pluripotency network maintenance. Also, the model predicts that, if Nanog is high enough, pluripotency is maintained even in the absence of Oct4 and Sox2 induction by external factors. Intriguingly, this prediction has been partly confirmed experimentally. Nanog can compensate for loss of self-renewal genes such as Esrrb, Tbx3 or Tcl1 \[[@B28]\], and Oct4^+/−^ mESCs lack Nanog-low, undifferentiated cells \[[@B29]\]. The model presents some limitations: it is unable to describe an effect of Oct4 overexpression in initiating differentiation \[[@B4], [@B30], [@B31]\], Nanog heterogeneity in the pluripotency state is not considered, and noise is neglected. In 2008, the same authors proposed an extension of the above model to better recapitulate mESC differentiation dynamics \[[@B32]\]. While the topology of the core pluripotency network is kept identical to the previous work, two new mutual antagonistic interactions are introduced to describe differentiation: Cdx2/Oct4 (responsible for trophectoderm lineage specification) \[[@B33]\] and Nanog/Gata6 (responsible for endoderm lineage specification) \[[@B34]\]. In addition, Cdx2 and Gata6 activate the orphan nuclear receptor germ cell nuclear receptor (Gcnf), a transcriptional repressor of pluripotency genes \[[@B35]\], which in turn represses Oct4. Finally, in order to account for evidence of cell tendency to specify towards an endoderm state when Oct4 is overexpressed \[[@B36]\], Gata6 activation by Oct4 is considered; this interaction had not been reported in the literature, so it is inserted as a model hypothesis, which is needed to recapitulate dynamics ([Figure 2()](#fig2){ref-type="fig"}). The authors developed an ODE formalism, similar to the one in their previous work \[[@B24]\]; the model reproduces the coexistence of three steady states, corresponding to pluripotent, trophectoderm, and endoderm cellular fates. In the absence of an external factor that forces Oct4 expression, the cells stay in the pluripotent state; the switch to a lineage-specific steady state is possible only by adding such external factor. The main achievement of this model is its ability to explain the biphasic response ("bell/inverse bell shaped") of TFs as functions of Oct4 (i.e., they can be expressed in low/high Oct4 ranges or at intermediate Oct4 levels, only), a mechanism not easy to explain by intuition. Interestingly, Gata6 activation by Oct4 hypothesized by the authors has been later proved, experimentally, in early blastocyst development \[[@B37]\]. As in the 2006 model \[[@B24]\], Nanog is expressed homogeneously in all cells in the pluripotent state. 3. Experimental Evidence of Nanog Heterogeneity in Serum/LIF-Cultured mESCs {#sec3} =========================================================================== In 2005, the first experimental evidence of Nanog multimodal distribution was reported \[[@B7]\]. Immunofluorescence experiments showed in serum/LIF-cultured mESCs the coexistence of two subpopulations, one positive for both Oct4 and Nanog and one positive for Oct4 only. In 2007, two different studies confirmed these results and highlighted new features of Nanog heterogeneity. Singh and colleagues \[[@B38]\], using a cell line previously developed by Mitsui\'s group (β-geo cells, in which the beta-galactosidase-neomycin fusion gene is inserted on one allele under the endogenous Nanog promoter \[[@B6]\]), confirmed heterogeneity of Nanog in contrast to homogeneous Oct4 expression. Analyzing by microarray the two Nanog subpopulations, a clear pluripotency signature characterized high Nanog (henceforth called HN) cells, while low Nanog (henceforth called LN) cells showed pronounced expression of mesodermal genes. Also in 2007, Chambers and colleagues \[[@B8]\] generated a mESC reporter cell line for Nanog, named TNGA, obtained by inserting a GFP-IRES-Puro-pA cassette at the Nanog start codon on one allele. This study provided evidence of both multimodal Nanog distribution and its temporal fluctuations between the two states; at steady state, two distinct (GFP negative and GFP positive, corresponding to LN and HN, resp.) subpopulations appeared clearly separated and could interconvert in time. In the following years, a number of additional reporter cell lines have been generated, and different experimental approaches, including flow cytometry and fluorescence-activated cell sorting (FACS), time-lapse microscopy, single-cell sequencing and single-molecule RNA-FISH, have been used to study Nanog distribution and dynamics at the levels of both single cells and cell populations \[[@B20]\]. 4. An Excitable Model for Nanog Heterogeneity in mESCs Grown in Serum/LIF {#sec4} ========================================================================= The first mathematical model accounting for Nanog heterogeneity in mESCs grown in serum/LIF medium was presented in \[[@B39]\]. Kalmar and colleagues combined experiments and mathematical modeling and proposed that Nanog dynamics arise from a noise-driven excitable system. Experiments were performed analyzing Nanog levels with the mentioned TNGA mESC cell line \[[@B8]\] both at population level by FACS and at single-cell level by time-lapse imaging. Nanog was proved to change dynamically, as sorted LN and HN mESCs were able to reproduce the original bimodal distribution, both at population level and at single-cell level; although with different timescales, sorted LN cells tended to rapidly switch to the HN state (within 24 hours), while sorted HN cells showed a slower switch off. At single-cell level, LN mESCs were able to switch Nanog on in 2-day time-lapse experiments (time lapses for the HN to LN transition were not included in the paper). Following these experimental observations, the authors derived a mathematical model showing that the HN state is stable for the system, with noise driving sporadic excursions of cells in the LN state. The system is excitable, thus highly sensible to perturbations (transcriptional noise, in this case) beyond a certain noise threshold. In detail, nonlinear ODEs describe the underlying GNR network, in which Oct4 and Nanog autoactivate their own and also each other expression and Oct4 inhibits Nanog, resulting in an overall negative feedback loop topology ([Figure 3(a)](#fig3){ref-type="fig"}). The authors assumed that Oct4 activates Nanog to saturated levels (thus this interaction is neglected in the equations) and only high levels of Oct4 can inhibit Nanog (a full biological motivation for such negative interaction, crucial for the system dynamics, is left by the authors for further investigation). The equations, based on Hill-like kinetics, consider protein dynamics only (mRNA steady-state assumption); Gaussian noise is added to the Nanog equation, in order to describe the stochastic excursions from the unique steady state of the system (HN). The model predicts that Oct4 is more homogenous in HN than in LN cells; this prediction is validated experimentally by immunofluorescence in FACS-sorted mESCs. While interesting, this result is not entirely surprising, taking into account the network topology. Finally, a stochastic model is derived and simulated using Gillespie\'s first reaction method. Both single-cell dynamics and population steady-state distribution of Nanog are reproduced, showing a good agreement with experimental data, and indicating that the regulation of occupancy of the two Nanog states is noise driven. The authors showed that the excitable regime is robust to parameter variations and concluded that Nanog cooperative autoregulation is crucial for the system\'s dynamics. 5. Multistable Models for Nanog Heterogeneity in mESCs Grown in Serum/LIF {#sec5} ========================================================================= In 2010, Glauche and colleagues proposed an alternative GRN-based formalism to recapitulate again experimental data obtained with the TNGA cell line. The modeled GRN includes Oct4-Sox2 heterodimer autoactivation and its activation on Nanog; also, as in Kalmar\'s model, Nanog autoactivation is considered ([Figure 3(b)](#fig3){ref-type="fig"}). This latter positive feedback loop, for proper parameter choices, can give raise to bistability. The stochastic differential equation (SDE) model, based on Hill kinetics, describes only protein dynamics. At steady state, the Nanog bimodal distribution observed experimentally in \[[@B8], [@B39]\] is reproduced; in time course simulations, fluctuations between the two states randomly happen due to the introduction of Gaussian noise (only in the Nanog equation). The main difference with the Kalmar model is that, in this case, both the LN and HN states are stable, with noise allowing the switch between them. Altering the noise changes the transition probability of the two Nanog states and, consequently, the residence times within each state. Interestingly, the authors simulate also the redistribution of TNGA cells upon sorting, performed experimentally in \[[@B39]\], matching the reestablishment of both populations on a timescale of circa 10 days. The model is finally analyzed to study the effect of Nanog heterogeneity on differentiation; the authors also consider an external differentiation signal, which inhibits Oct4 and is inhibited by Nanog, generating a double negative feedback loop. As a result of such loop, only LN mESCs respond to the differentiation signal with a decrease in Oct4-Sox2 complex level, indicating that Nanog has a "gatekeeper" role in mESC pluripotency maintenance. Of note, the differentiation signal is external; thus, no mESC spontaneous cell differentiation is conjectured. In the same paper, the authors also consider an alternative scenario in which, due to the introduction of a negative feedback loop through a not specified gene "X" that is activated by Nanog and represses it ([Figure 3(c)](#fig3){ref-type="fig"}), Nanog shows sustained oscillations, even in the absence of noise. Notably, such Nanog-mediated autorepression, hypothesized in the paper for modeling aims, has been later experimentally proven \[[@B40]\]. While still being able to recapitulate the experimental steady-state Nanog distribution, the oscillation scenario significantly differs in the sorting simulations: while the system is predicted to continuously shift from one steady state to the other in the fluctuation case, damped oscillations lead to the reestablishment of the steady-state distribution in the oscillation scenario. The same authors, more recently, tested experimentally these two discordant predictions, showing higher plausibility of the noise-induced bistable dynamics \[[@B41]\]. More recently, Lakatos and colleagues proposed a different formalism to describe Nanog dynamics arising from a bistable system \[[@B43]\]. The authors compared dynamics of five GRN topologies compatible with experimental evidence of interactions in the core pluripotency network. In all the analyzed GRNs, the genes considered are Oct4, Sox2, and Nanog; Oct4-Sox2 is a common transcription factor for the three genes and Nanog activates Sox2 ([Figure 4()](#fig4){ref-type="fig"}, blue arrows). In addition, the following interactions are included or not: Nanog activation on Oct4 (as in earlier Chickarmane\'s works \[[@B24], [@B32]\]), repression of Oct4 on Nanog, and autoinhibition of Nanog (experimental evidence in \[[@B44]\] and \[[@B40]\], resp.). Of note, in all topologies, Nanog autoactivation is neglected. Quasi-steady-state approximation is used for mRNA levels, proteins are modeled using ODEs based on the Shea-Ackers representation for transcriptional regulations, and variations to parameters are introduced to account for cell-to-cell diversity. Intriguingly, modeling shows that all the topologies considered can give rise to bistability, due to the Oct4-Sox2 autoregulation and dimerization. Due to the feedback loop between Nanog and Oct4, original parameter values predict the coexistence of two steady states for both Nanog and Oct4, contradicting experimental evidence of Nanog heterogeneity in contrast to Oct4 homogeneity \[[@B45]\]. However, parameter manipulation can shift nullclines and reproduce the mentioned experimental data. The authors also considered an extended topology ([Figure 4](#fig4){ref-type="fig"}, additional interactions indicated with black arrows), which includes the downstream pluripotency genes Klf4 and Esrrb (the latter acting as an amplifier \[[@B46]\]), fibroblast growth factor (FGF) signaling as an autocrine module, and a negative feedback through the MAP kinase/ERK kinase (MEK) cascade \[[@B47]\], which affects the cooperativity of the Oct4/Sox2-Nanog interaction \[[@B48]--[@B50]\]. Of note, FGF and MEK signaling have been proved experimentally to drive differentiation of mESCs, with a potential role in mESC heterogeneity \[[@B47]\]. Simulations of this extend topology show that the system allows the existence of distinct substates of Nanog and downstream transcription factors, still within the HN state; LN mESCs have higher Nanog expression than cells fully committed to differentiation, in agreement with experimental evidence \[[@B45]\]. Noise induces stochastic fluctuations between the multiple steady states: Nanog fluctuates in phase with Esrrb and Klf4, and in antiphase with FGF5, with no significant changes in Oct4 levels. Interestingly, multiple substates of the ON state of Nanog are predicted, suggesting the coexistence of multiple subpopulations of mESCs, responding differently to external signaling and having different pluripotency signature. The authors confirm experimentally that several components of the pluripotency network present a range of expression levels in mESC subpopulations, as also previously reported \[[@B45]\]. 6. Nanog Dynamics in Serum-Free Cultures {#sec6} ======================================== The development of culture conditions that enable mESC ground state pluripotency (2i/LIF medium) prompted the development of new mathematical formalisms to understand the mechanisms behind the overall loss of pluripotency gene heterogeneity. In 2012, the Peterson group proposed a novel model \[[@B51]\] to recapitulate distribution and dynamics of Nanog in different culture conditions (heterogeneity and fluctuations in LIF/bone morphogenetic protein (BMP4), homogeneity in 2i/LIF-3i/LIF) and their impact on somatic cell reprogramming (the BMP4/LIF culture condition was considered equivalent to serum/LIF \[[@B52]\]). In the network topology, the authors included Nanog and Oct4 autoactivations, and the activation of Oct4-Sox2 on Nanog ([Figure 5(a)](#fig5){ref-type="fig"}), as in their aforementioned previous works \[[@B24], [@B32]\]. In order to represent cells with high Oct4 levels while low Nanog \[[@B39]\], the previously considered Nanog activation on Oct4 and Sox2 is now removed. Also, the network includes a differentiation gene (loosely indicated as "G" in [Figure 5(a)](#fig5){ref-type="fig"}), which autoactivates itself, is activated by Oct4-Sox2 and mutually represses Nanog. Importantly, medium-regulated external factors are included: LIF/BMP4, which activates Nanog, the FGF4 receptor signaling, and Gsk3. FGF4 and Gsk3, lumped into a single variable, inhibit Nanog, are activated by the Oct4-Sox2 complex, and are blocked by 2i/3i medium ([Figure 5(a)](#fig5){ref-type="fig"}). Thus, in the overall topology, Nanog is regulated by a feed-forward loop including Oct4-Sox2 direct activation and inhibition via FGF4. An important assumption is that only when the Oct4-Sox2 heterodimer is bound to Nanog promoter additional TFs can bind (experimental evidence in \[[@B53]\]). A nonlinear ODEs model based on the Shea-Ackers formalism \[[@B25]\] is used, assuming steady-state levels for mRNA concentrations. To account for stochasticity originated within the network, the authors also formulated a master equation, simulated using the Gillespie algorithm. In BMP4/LIF medium, the system is bistable, and stochastic Nanog fluctuations arise. When Nanog levels are very low, cells are irreversibly pushed into a differentiation state, indicating again that Nanog has a pluripotency gatekeeper function. Also, the model predicts that suppression of FGF4/Gsk3 by 2i/3i can push cells into the monostable state (HN); in single-cell simulations, Nanog fluctuates around the unique steady state due to noise, but heterogeneity at population level is lost. Notably, the Nanog steady state in 2i/3i is higher than the HN steady state in BMP4/LIF medium. Of note, the noise is kept at the same level, both in BMP4/LIF and 2i/3i simulations; although LIF is not present in the 2i/3i simulations, the stem cell state is initialized with low "G," and high Nanog and Oct4-Sox2 levels. Interestingly, the model is also used to predict somatic cell reprogramming dynamics, performing simulations upon overexpression of Oct4 and Sox2: model predictions match the experimental observation of the importance of reprogramming factor dosage for the efficiency of the process \[[@B54]\]. If cells are cultured in 2i, an increase in reprogramming efficiency is predicted, with synchronous Oct4 and Nanog switch-on dynamics; these predictions are in agreement with experimental evidence of an increased efficiency in the late reprogramming phase if partially reprogrammed cells are cultured in 2i/LIF medium \[[@B55], [@B56]\]. In 2014, Herberg and colleagues proposed another GRN-based modeling framework to describe medium-dependent Nanog dynamics and the resulting mESC pluripotency signature \[[@B57]\]. The authors extended the network topology modeled in 2010 \[[@B42]\] introducing (i) the pluripotency marker Rex1 \[[@B58]\], directly activated by both Nanog and the Oct4-Sox2 heterodimer, in order to reproduce experimental data generated using the Rex1GFPd2 cell line \[[@B59]\]; (ii) FGF/Erk signaling, which inhibits Nanog and is activated by Oct4; and (iii) a phenomenological external differentiation signal, which inhibits the Oct4-Sox2 complex and is inhibited by Nanog, to recapitulate differentiation dynamics ([Figure 5(b)](#fig5){ref-type="fig"}). Of note, Rex1 is only an output of the system, as it is not involved in any feedback. Overall, the topology is comparable to the one modeled by Chickarmane in 2012 \[[@B51]\], but here, spontaneous differentiation arising from Nanog heterogeneity is not considered: only the influence of the cellular environment is investigated. Hill-based ODEs are derived for all genes; Gaussian noise is added to Oct4-Sox2, Nanog, and Rex1 equations; finally, steady state for mRNAs is assumed. As in the Chickarmane et al. 2012 paper, the system presents bistability, abrogated when culture medium is switched to 2i (LIF is implicitly considered here). In particular, in serum/LIF (FGF/Erk active), Nanog shows bistability due to its positive autofeedback loop, which strength is adjusted in order to have a steady-state distribution of circa 20% LN and 80% HN cells, as observed in TNGA experiments \[[@B8], [@B39]\]; the model can also reproduce Rex1 distribution measured culturing Rex1GFP2 mESCs. In time simulations, noise allows the switch between the HN and LN states, which is predicted to happen in the timescale of days. Increasing the level on noise results in more frequent switch and raised proportion of LN cells. Parameters of the Oct4-Sox2 complex are adjusted to keep its level constantly high and homogeneous. When simulating the 2i condition, the block of Erk (due to the presence of PD in the medium) reduces the inhibition of Nanog, pushing the system outside the bistability region into the "ground pluripotency" state, characterized by high Nanog, Oct4-Sox2, and Rex1 and low FGF/Erk. Of note, the effect of Chiron in the 2i medium is not considered. The model is used to predict differentiation dynamics: after the first 24 hours of 2i withdrawal, Nanog decreases with fast dynamics, followed by Rex1 and, ultimately, Oct4-Sox2, matching experimental observations \[[@B60]\]. Experimental data in the paper validate model predictions about the asynchronicity of the differentiation process, possibly due to individual cell variability arising when the inhibitor is removed from the medium. A comparison with differentiation dynamics of cells cultured in serum/LIF (thus, starting from bimodal Nanog and Rex1 distribution) is not reported. 7. A Probabilistic Model for Nanog Multistationary Dynamics {#sec7} =========================================================== Luo and colleagues \[[@B61]\] proposed an alternative approach to describe medium-depended Nanog dynamics in mESCs. Differently from the above formalisms, the authors did not rely on a regulatory network, to avoid introducing assumptions in case of lacking or contrasting data about underlying regulations, and possible overparameterization. Instead, they used a statistical model, based on probability density functions and Gaussian noise. The model has just one variable (representing Nanog), for which the existence of a stationary distribution is assumed (represented as a mixture of Gaussian distributions). Using nested sampling, model parameters are fitted on novel experimental data presented in the paper: the authors identify in TNGA mESCs the distribution of Nanog as a mixture of high, medium, and low (HN, MN, and LN, resp.) subpopulations. Then, they focus on the MN state, which is shown by experiments to have the highest dynamical changes. Nanog dynamics are measured in sorted MN cells replated in different culture conditions for four days (LIF/BMP4, PD only, Chiron only, and 2i/LIF); Nanog distribution shifts towards the HN state in the 2i/LIF condition and these data are used to validate model predictions about sorting dynamics. Intriguingly, Chiron widens the LN and MN states, while PD narrows them. One interesting result of the fitting is a competitive effect of PD and Chiron: while the first tends to add noise to the system, the second can filter noise. The approach used in this paper is definitively interesting: it allows inferring directly the shape of the potential function from the data, without a priori knowledge of the underlying signaling network. The experiments were performed 4 days after the sorting; it could be of interest to perform both experiments and simulations on a longer timescale, to check if the distributions reported in the paper are or not at steady state. 8. Nanog Heterogeneity Arising from Allelic Switching {#sec8} ===================================================== Miyanari and Torres-Padilla \[[@B62]\] recently suggested an alternative mechanism contributing to Nanog heterogeneity based on allelic regulation. The authors firstly analyzed Nanog nascent transcription by RNA fluorescent in situ hybridization (RNA-FISH) in vivo, observing monoallelic and biallelic expression in preimplantation embryo and in late blastocyst, respectively. To study the dynamic switch of Nanog in vitro, the authors generated a novel reporter cell line based on a dual-reporter system (named NGR mESCs), in which two distinct fluorescent reporters (destabilized TurboGFP and mCherry) are inserted downstream of the NANOG-coding region in the two alleles. In NGR mESCs cultured in serum/LIF, Nanog transcriptional firing was observed, with an odd distribution of allelic firing: only a very low percentage of cells showed Nanog biallelic expression, in contrast to biallelic expression of other pluripotency genes such as Oct4. Consequently, in vitro cultures showed distinct populations for Nanog: monoallelic, biallelic, and no expression; similar results were obtained using single-molecule RNA-FISH (smRNA-FISH) \[[@B63]\]. In contrast, biallelic expression was reported in 2i/LIF. These results suggested a crucial role of allelic expression, and the possibility of achieving ground state pluripotency by activating the second Nanog allele. In light of such results, Wu and Tzanakakis proposed a mathematical model to recapitulate medium-dependent Nanog dynamics \[[@B64]\]. The mESC population is divided into 4 groups (biallaleic, monoallelic, and both alleles inactive), with percentages of residence in each group and probabilities of transition among groups extrapolated from RNA-FISH and allele-specific RT-PCR experiments in \[[@B62]\] ([Figure 6](#fig6){ref-type="fig"}). The stochastic allelic switching is modeled using a homogeneous Markov chain, assuming that the next state does solely depend on the current state; asynchronous proliferation rates are considered. The model is simulated using a Monte Carlo algorithm, setting the half-life of endogenous Nanog to 2 hours, as measured in \[[@B65]\]. In the absence of noise, at population level, Nanog distribution shows 3 peaks (low for both alleles inactive, middle for monoallelic, and high for biallelic cells), while addition of noise leads to a bimodal distribution (LN and HN only). In single-cell-simulated dynamics, HN cells show a slow (circa 20 hours) switch off, while LN cells switch on in circa 5 hours. The predicted dynamics are much faster than the ones previously observed with TNGA mESCs; thus, the authors simulate the expected output of the latter cells, in which the reporter is inserted only in one allele and has half-life longer than endogenous Nanog. In this case, even in the absence of transcriptional noise, only 2 peaks in the reporter distribution are predicted, confirming experimental observations in \[[@B8], [@B39]\]. Also, the predicted correlation between endogenous Nanog and the reporter dynamics is very high (Pearson coefficient = 1) if both alleles are tagged and the half-life of the reporter is comparable to the one of Nanog, while correlation is lost (Pearson coefficient = 0.06) using a TNGA-like tagging approach. Finally, the authors use the model to simulate dynamics of Nanog^+/−^ cells, predicting that Nanog deletion on one allele can impact its distribution at population level. The paper does not include results about Nanog dynamics in 2i/LIF, but the authors mention that the model could easily be extended to describe the latter case by properly changing the proportion of cells in each subpopulation. Simulation results about the impact of differences between endogenous and reporter kinetics could explain mismatches in steady-state distributions and dynamics observed using different Nanog-tagged cell lines. Such predictions have been partly corroborated recently: the group of Henrique generated a novel reporter cell line (Nd mESCs), which contains a transgenic bacterial artificial chromosome (BAC) with a destabilized Venus reporter protein (Venus-Nuclear-PEST) under the control of Nanog regulatory regions \[[@B66]\]. Nd mESCs differ from TNGA cells for two main reasons: the two endogenous Nanog alleles are kept intact and the half-lives of the reporter mRNA and protein are comparable to the endogenous Nanog ones (circa 4 and 2 hours, resp.). In Nd mESCs grown in serum/LIF, as compared to TNGA cells, Nanog expression is again mosaic but within an overall narrower distribution; also, fluctuations between the LN and the HN state and restoration of the original distribution from LN- and HN-sorted cells occur in a shorter timescale (circa 4 hours and 2--4 days of culture, resp. \[[@B66]\]). 9. Nanog Dynamics Arising from a Protein Interaction Network {#sec9} ============================================================ Recently, Muñoz-Descalzo and colleagues \[[@B67]\] considered an alternative model to describe Nanog dynamics in mESCs focusing on posttranscriptional interactions. In view of recent data generated by the same research group about the key role of protein balance for mESC pluripotency maintenance \[[@B68]\], the authors reanalyzed the excitable model of Kalmar and colleagues and pointed out its failure in reproducing (i) the increased correlation between Oct4 and Nanog observed by single-cell immunofluorescences in 2i/LIF medium and (ii) a critical region characterized by low levels of both Nanog and Oct4, in which pluripotency is lost. The authors started considering a minimal model, named NOC ([Figure 7(a)](#fig7){ref-type="fig"}), which encompasses only Oct4 and Nanog, and assume that the correlation between the two proteins results from the formation of a protein complex (O:N) previously described experimentally \[[@B69]--[@B71]\]. In the model, Oct4 and Nanog proteins exist either free or bound together in a complex; transcriptional regulations are neglected. Importantly, the model takes into account different stabilities for proteins (high for Oct4 and Nanog/Oct4 complex and low for Nanog), measured experimentally in the paper. To account for transcriptional heterogeneity of Nanog, its transcriptional bursting \[[@B40], [@B62]\] is considered. The resulting discrete stochastic formalism is simulated for both serum/LIF and 2i/LIF conditions by tuning Nanog expression (infrequent transcriptional bursting in serum/LIF and high frequency bursting in 2i/LIF). The model correctly matches Nanog and Oct4 correlation in the two media, and Nanog mRNA distribution (unimodal in both media, but shifted towards high levels in 2i/LIF) observed experimentally by mRNA-FISH. Still, the model has some pitfalls: it predicts that knock-out of Nanog results in increased Oct4 levels, the latter experimentally associated with differentiation, contradicting experimental evidence about mESC ability to maintain pluripotency even in absence of Nanog \[[@B8]\]. Thus, the authors developed a refined model (named TBON), which also includes Tcf3 and β-catenin, two Wnt pathway proteins associated with pluripotency \[[@B59], [@B72], [@B73]\] and involved in a protein-complex with Oct4 \[[@B74], [@B75]\]. The new topology ([Figure 7(b)](#fig7){ref-type="fig"}) considers, in addition to the Oct4-Nanog complex (O:N), the inducers (PD and Chiron) and three additional complexes: β-catenin with Oct4 (β:O) \[[@B74], [@B75]\], β-catenin with Tcf3 (β:T) \[[@B76]\], and β-catenin with Oct4 and Nanog (β:O:N). Stochastic simulations of the model match protein distributions and correlations of protein levels in both serum/LIF and 2i/LIF. The model shows that, in 2i/LIF, ground state pluripotency is achieved by attenuation of free Oct4, thus limiting its effect in promoting differentiation. Also, the model predicts that, in absence of Nanog, the β:O complex becomes stronger, enabling cells to maintain pluripotency, as confirmed in experiments reported in the paper. In case of lack of β-catenin, the model predicts that the correlation between Nanog and Oct4 is unchanged, but their levels are lower, due to Tcf3 increase. These model predictions are in agreement with experimental evidence about the not absolute requirement of β-catenin for pluripotency \[[@B77]\], although its lack makes mESCs more prone to differentiate \[[@B59]\]. Conversely, abrogation of Tcf3 sustains pluripotency through increased Nanog, decreased Oct4 and β-catenin, and impaired Oct4-Nanog correlation. Finally, the model predicts that the removal of Oct4 from the system causes a drop in Nanog levels (as, if not bound in protein complex, Nanog has faster degradation), in agreement with loss of pluripotency in Oct4^−/−^ cells \[[@B31]\]. 10. Limitations of Existing Formalisms for Ground State Pluripotency Cultures {#sec10} ============================================================================= All the reviewed models but the Muñoz-Descalzo et al. one just consider the molecular effect of the MEK inhibitor (PD), neglecting effects of the Gsk3 inhibitor (Chiron) in ground state pluripotency cultures. Given the failure of PD alone in sustaining mESC clonal propagation \[[@B1], [@B59]\] and the crucial functions of the Wnt/β-catenin pathway in pluripotency maintenance and successful reprogramming of somatic cells to pluripotency \[[@B13], [@B78]--[@B82]\], the role of the Wnt pathway in cellular heterogeneity and mESC plasticity should be better characterized. Also, possible effects of nonhomogenous, or cell density-dependent, drug uptake in cellular cultures on the system dynamics might be considered. Finally, no current model accounts for a certain level of Nanog heterogeneity and temporal fluctuations persistent in 2i/LIF recently observed experimentally \[[@B83], [@B84]\]. It remains to be shown whether long-term cultures in 2i/LIF could impair mESC karyotypic and epigenetic stability, given prolonged exposure of cells to Gsk3 inhibition \[[@B85]\]. Recently, R2i, an alternative serum-free, chemical media, has been proposed to sustain mESC ground state pluripotency \[[@B86]\]. Of note, also R2i eliminates mESC heterogeneity, but it acts on the pluripotency circuit through other routes (it contains inhibitors of TGFβ and FGF signaling pathways); this provides a good indication that mosaic expression of pluripotency genes in serum/LIF might originate from multiple sources, which mathematical formalisms should consider. 11. Crosstalk between Gene Expression Dynamics and the Cell Cycle {#sec11} ================================================================= Mathematical models could be particular useful in elucidating the complex interconnection between cell cycle, the pluripotency network, and cellular fate. In the experimental works mentioned in this review, fluctuations of Nanog reporter genes were observed within a mESC cell cycle \[[@B62], [@B84]\]; however, there is also evidence of coupling between pluripotency network gene dynamics and the cell cycle, as recently reviewed in \[[@B87]\]. In an early work reporting Nanog heterogeneity \[[@B38]\], FACS-sorted HN cells were found to upregulate cell cycle genes characteristic of the S-G2 phases, while LN cells expressed genes characteristic of the G1 phase. Also, MacArthur and colleagues, using an inducible system, found a correlation between Nanog and cell cycle checkpoints genes \[[@B60]\]. Recently, Nanog but not Oct4 expression was proved to oscillate in mESCs synchronized for the cell cycle \[[@B88]\]. In a recent work \[[@B41]\], Herberg and colleagues extended their previous GRN model \[[@B57]\] to include mESC proliferation; the resulting agent-based model predicts an effect of different cell cycle times in the proportion of cells in low-high Nanog subpopulations in serum/LIF; analysis in ground state pluripotency cultures (in which mESC cell cycle is known to differ \[[@B89]\]) was not performed. Multiscale modeling approaches, able to account simultaneously for processes at subcellular, intercellular, and population levels, could be highly informed by quantitative single-cell measurements (such as live imaging and sequencing), and be able not only to reproduce experimental data but also to generate useful predictions, usable for targeted control of mESC fate, both in pluripotency maintenance and in differentiation. 12. Conclusions and Future Perspectives {#sec12} ======================================= Here, we have reviewed recent computational/experimental results about mechanisms and consequences of Nanog dynamics in populations of isogenic mESCs. We reported main assumptions, results and predictions of mathematical models based on regulatory networks (in which Nanog dynamics result from its interactions with other pluripotency genes, signaling pathways, and drugs at transcriptional and posttranslational levels \[[@B39], [@B42], [@B43], [@B51], [@B57], [@B67]\]), statistical models \[[@B61]\], and unbalanced Nanog allelic expression \[[@B64]\]. A common conclusion is that Nanog acts as a molecular gatekeeper, fine-controlling cell fate in response to pluripotency and differentiation genes\' regulations, internal noise, and external stimuli. We showed that different formalisms are able to reproduce Nanog dynamics observed experimentally with reporter mESCs cultured in both serum/LIF and chemical, serum-free culture media, and generate testable predictions. Still, many open questions remain regarding both discordant experimental results and the validity of modeling assumptions. How can models based on different GRN topologies reproduce the same experimental data? One critical step in deriving mathematical models is parameterization. Notably, given the same system of equations, different set of parameters and timescales of the variables involved can result in completely different dynamic scenarios; conversely, different sets of genes can be included in GRNs to reproduce specific dynamics, as far as the topology encompasses key elements (e.g., positive feedback loop(s) for bistability). It is therefore crucial, as far as possible, to use parameters directly measured and rely only on interactions unambiguously identified experimentally. Also, when comparing simulations to experimental data, it is vital to critically account for the experimental settings and the timescales of the considered variables. Among the models reviewed here, only Wu and Tzanakakis explicitly took into account the differences in the degradation rates of endogenous Nanog and its reporter; their results suggest modeling as a powerful tool for experimentalists to infer endogenous dynamics from fluorescent reporter data. The Wu and Tzanakakis paper, as the Luo and colleagues one, is for sure interesting also for being able to describe Nanog dynamics at population level starting from single-cell dynamics, without relying on a GRN model with consequent assumptions on the network topology. However, the Luo et al. model, while capable of reproducing Nanog heterogeneity, cannot explain its source. Focusing on timescales, the commonly used steady-state assumption for mRNA dynamics, while reducing the parameter space, might lead to misrepresentative results. Indeed, in 2014, two independent reports \[[@B90], [@B91]\] focused on Nanog distribution at the protein level and found that mESCs cultured in serum/LIF do not show protein biallelic expression, in contrast with mRNA firing measured in \[[@B62]\]. In particular, in \[[@B91]\], Filipczyk and colleagues generated mESCs encoding dual fusion proteins to the Nanog C-terminus, with green and red fluorescent reporters placed on the two alleles. At cell population level, FACS analysis showed bimodal distribution for Nanog in mESCs cultured in serum/LIF, although with a narrower expression range as compared to TNGA cells (as with Nd cells), but correlated expression of the two fusion proteins, suggesting biallelic expression of Nanog protein. These results are in agreement with examination of mature cytoplasmic RNA transcript, which was found to be biallelic regardless allelic Nanog firing \[[@B63]\]. Given that the fusion fluorescent proteins generated by Filipczyk and colleagues have an half-life comparable to the endogenus Nanog\'s one, protein biallelic expression can be compatible with the mRNA allelic switching reported in \[[@B62]\] only if different timescales for mRNA and protein are considered \[[@B20]\]. Taken together, these results suggest that the quasi-steady-state assumption for Nanog mRNA dynamics might be revised and that GNR feedback mechanisms are still needed to explain heterogeneous Nanog protein dynamics in serum/LIF. Also, in our view, the impact of intrinsic and extrinsic noise on mESC fate should be better addressed: molecular instability is often represented incorporating a stochastic term in a set of differential equations, but whether it is a good representation has not been addressed. It would be extremely interesting to combine quantitative estimate of noise (as in \[[@B92]\]) with GRN-based formalisms. Scaling up the reviewed mathematical models for small GRNs, involving other key genes involved in pluripotency and differentiation identified with bioinformatics and statistical methods \[[@B60], [@B93]--[@B95]\] and unraveling their role in mESC temporal dynamics, remains an open challenge. Also, the descriptive and predictive power of mathematical models could be significantly increased accounting also for the combined role of noncoding RNAs \[[@B96]--[@B99]\], epigenetic mechanisms \[[@B100]\], metabolism \[[@B101]\], and posttranscriptional modifications \[[@B102]\]. From a more broad prospective, we do believe that the questions about the role of heterogeneity in mESC decision-making stated at the beginning of this review have not yet been fully addressed. Xenopoulos et al. \[[@B103]\], using high-resolution live cell imaging, reported "salt and pepper" Nanog pattern in early blastocyst, but lack of temporal fluctuations, and irreversible commitment to epiblast and primitive endoderm. If and how heterogeneity can be advantageous for mESC plasticity is currently not known; the principles behind the use of noise and fluctuations by genetic circuits, also in an evolutionary prospective, are beginning to be elucidated only recently \[[@B104]\]. The hypothesis that pluripotency is an emergent property of a population of cells rather than a characteristic of a single cell \[[@B60]\], with noise and cellular heterogeneity conferring stem cell high entropy, and thus the potential to choose a number of specialized, differentiated fates \[[@B105], [@B106]\], still needs to be more extensively characterized not only for embryonic but also induced and adult stem cells. Also, the "exploratory hypothesis" for pluripotent cells, which conjectures stem cell decision-making as a two-step process in which firstly stochasticity induces a critical state that primes diverse transcriptional programs, and then one particular fate is chosen via interaction with external inputs \[[@B107], [@B108]\], represents, in our view, an interesting research avenue for further investigation. This work was funded by MRC (MR/N021444/1) (Lucia Marucci) and BrisSynBio, a BBSRC/EPSRC Synthetic Biology Research Centre (BB/L01386X/1) (Lucia Marucci). Conflicts of Interest ===================== The author declares that there is no conflict of interests regarding the publication of this paper. ![Core pluripotency regulatory network modeled in \[[@B24]\]. The core network includes the transcription factors Oct4, Sox2, and Nanog and the heterodimer Oct4-Sox2. Arrow- and bar-headed lines represent activation and inhibition, respectively, among genes (green ovals) and inputs/outputs (grey rectangles).](SCI2017-7160419.001){#fig1} ![Extended regulatory network for pluripotency modeled in \[[@B32]\]. The network includes, in addition to the core pluripotency factors Oct4, Sox2, and Nanog, Cdx2 and Gata6 to represent trophectoderm and endoderm commitment through Oct4-Cdx2 and Nanog-Gata6 interactions, respectively, and the nuclear receptor Gcnf. Green and purple ovals represent pluripotency and differentiation genes, respectively.](SCI2017-7160419.002){#fig2} ![Pluripotency regulatory networks to describe excitable, bistable and oscillatory Nanog dynamics in serum/LIF. (a) Regulatory network proposed in \[[@B39]\] to describe Nanog dynamics in serum/LIF as excitable; additive noise is considered for Nanog and allows the system to transiently escape the HN stable steady state. (b) and (c) Networks proposed in \[[@B42]\] to reproduce Nanog dynamics as bistable (b) or oscillatory (c). The noise term, summed to Nanog\'s equation, is indispensable to reproduce Nanog heterogeneity and dynamics in (b) only. In (c), the negative feedback loop between Nanog and a not specified gene ("X") can generate oscillatory dynamics.](SCI2017-7160419.003){#fig3} ![Extended regulatory network to describe Nanog multistable dynamics in serum/LIF \[[@B43]\]. Blue and black arrows indicate interactions of the core and extended network, respectively. The extended network interactions include additional pluripotency genes (Klf4, Esrrb), and differentiation signaling pathway (FGF and MAPK) genes.](SCI2017-7160419.004){#fig4} ![Regulatory network to recapitulate Nanog dynamics in 2i/LIF. Networks proposed and modeled in \[[@B51]\] (a) and \[[@B57]\] (b). Both systems include Nanog, the Oct4-Sox2 heterodimer, a module for FGF signaling, and the effect of the medium (2i/3i) on network regulations. In addition, in \[[@B51]\] (a), the interactions with LIF/BMP4 and a differentiation gene ("G") are included, while Herberg and colleagues (b) include Rex1 as a system output only and a differentiation signal ("Y," similar to "G" in (a)). In both models, the effects of the two inhibitors in 2i are lumped.](SCI2017-7160419.005){#fig5} ![Schematic of Nanog allelic expression, as modeled in \[[@B64]\]. mESCs are grouped into 4 types given different patterns of Nanog allelic expression; the percentage fraction of mESCs residing in each group is indicated as well as the transition probability (in percentage).](SCI2017-7160419.006){#fig6} ![Pluripotency protein interaction networks, as modeled in \[[@B67]\]. Scheme of the minimal (a) and refined (b) protein interaction networks. O:N, β:T, β:O, and β:O:N indicate Oct4-Nanog, β-catenin-Tcf3, β-catenin-Oct4, and β-catenin-Oct4-Nanog complexes, respectively.](SCI2017-7160419.007){#fig7} [^1]: Academic Editor: Gary E. Lyons	{ "pile_set_name": "PubMed Central" }
Background ========== The effectiveness of conservative treatment of scoliosis is controversial. Some studies suggest that brace is effective in stopping curve progression, whilst others did not report such an effect. The purpose of the present study was to effectiveness of PASB in the correction of lumbar curves, in agreement with the SRS Committee on Bracing and Nonoperative Management Standardisation Criteria \[[@B1]\]\[[@B2]\]. Materials and methods ===================== Fourty adolescent females (mean age 12.95 ± 1.72 years) with lumbar curve and a pretreatment Risser score ranging from 0 to 2 have been enrolled. The minimum duration of follow-up was 24 months (mean: 41.75 ± 34.47 months). Antero-posterior radiographs were used to estimate the curve magnitude (CM) and the torsion of the apical vertebra (TA) at 5 time points: beginning of treatment (t1), one year after the beginning of treatment (t2), intermediate time between t1 and t4 (t3), end of weaning (t4), 2-year minimum follow-up from t4 (t5). Three situations were distinguished: curve correction, curve stabilisation and curve progression. Results ======= CM mean value was 26.43 ± 2.77 SD at t1 and 13.80 ± 7.94 SD at t5. TA was 10.83 ± 3.74 SD at t1 and 7.88 ± 4.24 at t5. The variation between measures of Cobb and Perdriolle degrees at t1,2,3,4,5 and between CM t5-t1 and TA t5-t1 were significantly different. Curve correction was accomplished in 82.5% of patients, whereas a curve stabilisation was obtained in 17.5% of patients. Conclusions =========== The PASB, due to its peculiar biomechanical action on vertebral modelling, is highly effective in correcting lumbar curves.	{ "pile_set_name": "PubMed Central" }
Published: March 17, 2020 Introduction {#sec1} ============ The rodent hippocampus coordinates a wide array of behaviors, from spatial navigation ([@bib34]) to decision making under approach-avoidance conflict ([@bib17], [@bib22]) and reward processing ([@bib17]). A central hypothesis of how the hippocampus might participate in such diverse behaviors is the presence of heterogeneous principal neurons that differ widely in their gene expression ([@bib10], [@bib42]), electrophysiological properties ([@bib23]), and behavioral function ([@bib11], [@bib12], [@bib17], [@bib42]). In particular, the main ventral hippocampal output region, the ventral subiculum (vS), is composed of multiple neuronal populations that send parallel, long-range projections to distinct areas, including prefrontal cortex (PFC; vS^PFC^), lateral hypothalamus (LH; vS^LH^), and nucleus accumbens shell (NAc; vS^NAc^) ([@bib32]). These populations are proposed to integrate a myriad of local and long-range inputs ([@bib4], [@bib42], [@bib51]) to perform their unique behavioral functions ([@bib1], [@bib11], [@bib17], [@bib22], [@bib41]). However, to date, knowledge of the input connectivity of vS output neurons is lacking. Further, vS populations are spatially patterned, in particular along the proximal-distal (PD) axis (ranging from the CA1 to the presubiculum borders) ([@bib11]), and synaptic input varies dramatically across different spatial locations in vS ([@bib2], [@bib11], [@bib25], [@bib30], [@bib46]). Based on this, we hypothesized that different vS output populations receive distinct upstream inputs, and we reasoned that these inputs may in turn depend on the spatial location of postsynaptic neurons ([@bib10], [@bib11]), their downstream target ([@bib17], [@bib22], [@bib32]), or a combination of these two factors. To address this hypothesis, we studied the anatomical organization of projection-defined neurons in vS; we then applied rabies tracing across different vS projections and postsynaptic cell locations, and obtained a brain-wide map of inputs to vS subpopulations. We identified quantitative differences in multiple long-range input regions to vS that depended to different extents on the spatial location and projection target of vS neurons. Results {#sec2} ======= Hippocampal Projection Populations Are Topographically Organized in vS {#sec2.1} ---------------------------------------------------------------------- Each vS projection population is thought to occupy a unique spatial distribution in the hippocampus ([@bib23]), and long-range input into hippocampus has been shown to be highly topographical ([@bib2], [@bib11], [@bib25], [@bib30], [@bib46]). Therefore, we first wanted to determine the spatial distribution of the different projection populations within vS, because this spatial distribution could be an important determinant for differential input connectivity. To do this, we stereotaxically injected the retrograde tracer cholera toxin subunit-B (CTXβ) into PFC, LH, or NAc in a pairwise manner ([Figure 1](#fig1){ref-type="fig"}A); collected coronal sections that spanned the hippocampus; conducted whole-hippocampus cell counts; and registered the data to the Allen Brain Atlas (ABA; see [STAR Methods](#sec4){ref-type="sec"}; [Figures 1](#fig1){ref-type="fig"}B--1F). We first confirmed that vS^PFC^, vS^NAc^, and vS^LH^ neurons resided predominantly in vS ([Figure 1](#fig1){ref-type="fig"}C), and that the fraction of colocalized cells (neurons that projected to more than one injection site) ranged from 2% to 6% ([@bib32]). Using this approach, we found that vS projection populations occupied spatially distinct locations within vS ([Figures 1](#fig1){ref-type="fig"}D and 1E). The locations of these neurons varied across all three axes (anterior-posterior \[AP\], medial-lateral \[ML\], and dorsal-ventral \[DV\]), but most notably along the AP axis ([Figures 1](#fig1){ref-type="fig"}E, 1F, 1H--1J), whereas vS^PFC^ neurons were located at more anterior locations, vS^LH^ neurons were located at more posterior locations, and vS^NAc^ neurons were spread across the entire range of vS. Using correlation analysis, we confirmed that projection type covaried across all axes, but most dramatically with AP position ([Figure 1](#fig1){ref-type="fig"}F), and that the spatial position of neurons along AP was most predictive of their output target (see [STAR Methods](#sec4){ref-type="sec"}; [Figure 1](#fig1){ref-type="fig"}F).Figure 1Non-overlapping vS Neurons Occupy Distinct Spatial Sites(A) Schematic of pairwise CTXβ injections.(B) Left: side view of the hippocampus with the plane of coronal sectioning illustrated. Right: example schematic of a coronal section. Boxed area is the approximate location of the example images in (C).(C) Left: examples of stitched coronal sections of vS with retrogradely labeled cells. Scale bar: 500 μm. Middle: zoom-in images of boxed area from left. Scale bar: 50 μm. Right: proportion of single- and dual-labeled hippocampal neurons.(D) 3D whole-brain diagrams with CTXβ-labeled hippocampal neurons (n = 10,8291 hippocampal neurons from 6 brains; see [Video S1](#mmc4){ref-type="supplementary-material"}).(E) Cumulative distributions of CTXβ-labeled cells in ventral hippocampus along AP, ML, and DV axes (vS^LH^: n = 7,717 cells, 8 injections; vS^NAc^: n = 9,989 cells, 6 injections; vS^PFC^: n = 9,287 cells, 6 injections).(F) Left: Pearson correlation coefficient was highest along AP compared with ML and DV (n = 10 injections). Right: logistic regression analysis predicts projection type based on spatial location. Removal of AP as a predictor led to the largest decrease in accuracy (see [STAR Methods](#sec4){ref-type="sec"}). Data are summarized as mean ± SEM. Individual injections are overlaid in gray.(G) Right: side view of the hippocampus, with horizontal cutting plane illustrated. Middle: example stitched horizontal brain section. Boxed area indicates location of the left ventral hippocampus. Scale bar: 500 μm. Right: schematic of ventral hippocampus in the horizontal plane, illustrating PD axis. Boxed area is the approximate location of the example images in (H).(H) Example stitched horizontal sections of retrogradely labeled vS neurons. Scale bar: 100 μm.(I) Cumulative distributions of cell counts along the proximal-distal (PD), deep-superficial (DS), and DV axes (vS^LH^: n = 5,785 cells, 7 injections; vS^NAc^: n = 3,405 cells, 6 injections; vS^PFC^: n = 3,509 cells, 7 injections; see [STAR Methods](#sec4){ref-type="sec"}).(J) Pearson correlation (left) and logistic regression (right) analyses analogous to (F). The PD axis captures the most variation and best predicts the projection type of vS neurons (n = 10 injections).(K) LDA of registered neurons in vS unbiasedly identifies the anatomical plane that best discriminates between projection populations. Left: cell positions projected onto the first and second linear discriminant axes. Cumulative distributions of cells along the first (middle) and second (right) linear discriminant demonstrate overlap of projection populations.(L) Schematic of vS topography, where vS neurons are intermingled, vS^PFC^ neurons are located most proximally, vS^LH^ neurons most distally, and vS^NAc^ neurons span the area between both vS^PFC^ and vS^LH^.Video S1. Rotating 3D Whole-Brain View of CTXβ-Labeled Hippocampal Neurons Registered to the Allen Brain Atlas, Related to Figure 1Data from 108,921 registered, CTXβ-labeled hippocampal neurons are presented in a virtual glass brain (color code: vS^PFC^ in yellow, vS^NAc^ in magenta, vS^LH^ in cyan). Each dot represents one labeled hippocampal neuron. We reasoned that the marked distribution of projection populations across the AP axis might be a reflection of the known PD distribution of projection populations along the pyramidal cell layer of the hippocampus ([@bib11]), which in ventral hippocampus is oriented approximately along the AP axis ([Figures 1](#fig1){ref-type="fig"}G and 1H). We confirmed that this was the case by making horizontal slices of labeled vS ([Figure 1](#fig1){ref-type="fig"}G). This experiment showed that the variance in AP is well explained by a PD gradient of neurons in vS, where vS^PFC^ neurons were located at more proximal locations near the CA1 border, vS^LH^ neurons were located at more distal locations near the presubiculum border, and vS^NAc^ neurons were spread across the entire PD axis ([Figures 1](#fig1){ref-type="fig"}H and 1I). In keeping with previous observations, correlation and regression analyses also indicated that projection type varies most along the PD compared with the deep-superficial (DS) and DV axes ([Figure 1](#fig1){ref-type="fig"}J). Interestingly, the spatial distributions of the different projection populations were highly overlapping across each of the AP, ML, DV, and PD axes. This is notably different from that observed in dorsal subiculum ([@bib11]), where there is a sharp PD border separating distinct projection populations. To control for the possibility that this sharp border was not simply obscured by our coronal or horizontal slicing angles, we carried out linear discriminant analysis (LDA) on the registered whole-hippocampus neuronal distributions ([Figure 1](#fig1){ref-type="fig"}K). This method allowed us to unbiasedly find a virtual plane that best separates the three populations of projection neurons. By examining the distribution of neurons in this plane, we observed that, although there was a clear subregion organization, the spatial distribution of the three populations in vS still remained highly overlapping ([Figure 1](#fig1){ref-type="fig"}K). Thus, in contrast with dorsal subiculum, vS appears to be organized as a gradient, especially with vS^NAc^ neurons being present at almost every point along the AP axis. Overall, our data indicate that vS projection populations are segregated cell types and occupy overlapping yet distinct locations within vS, best separated across the AP axis. Labeling of Hippocampal Input Dependent on Spatial Location and Projection Target {#sec2.2} --------------------------------------------------------------------------------- Next, to directly assess the organization of presynaptic inputs onto these vS projections, we applied tracing the relationship between input-output (TRIO) ([@bib6], [@bib8], [@bib37], [@bib39]) to vS projections ([Figure 2](#fig2){ref-type="fig"}A; [Figures S1](#mmc1){ref-type="supplementary-material"}A--S1E). This approach involved injecting AAV2-retro-Cre ([@bib45]) into the output target region to retrogradely express Cre recombinase in vS neurons that projected to the target site. In the same surgery, we injected a single Cre-dependent helper construct (AA2/1-synP-FLEX-split-TVA-2A-B19G \[TVA-G\]) into vS. After at least 2 weeks of TVA-G expression in projection-specific neurons, G-deleted, pseudotyped rabies virus (EnvA-RVΔG-H2B-mCherry) harboring nuclear-localized mCherry was injected into vS to infect TVA-G^+^ cells.Figure 2TRIO of vS Output Neurons(A) Schematic of TRIO.(B) Left: example stitched sagittal section of the brain with rabies-labeled cells (black) in the hippocampus. Right: zoom-in of boxed area from left. Starter cells colocalize for TVA-G (cyan) and rabies (red). Scale bars: 1,000 μm (right); 200 μm (left).(C) Overall input from coarse anatomical structures normalized to the total number of inputs counted in a single brain (n = 20 brains). Note the break in the y axis. Data are summarized as mean ± SEM.(D) Representative stitched images of long-range monosynaptic inputs to vS neurons by projection type and the center of mass (COM) of the starter cells along the AP axis. Classification of starter cells into anterior and posterior groups was based on the rank order of COM along the AP axis. Approximate sagittal planes are displayed with red-shaded boxes indicating the estimated locations of the corresponding images. For images of MPO input, three consecutive images (spanning 180 μm) were projected to produce the displayed representative images. 3V, third ventricle; ac, anterior commissure. Scale bars: 200 μm (all input regions except PVT); 100 μm (PVT).See also [Figure S1](#mmc1){ref-type="supplementary-material"}. Importantly, we systematically varied the injection site of TVA-G and rabies within vS ([Figure 2](#fig2){ref-type="fig"}A). This method ensured that for each projection target, we sampled distributions of starter cells (i.e., cells from which rabies virus begins the monosynaptic retrograde tracing) from different locations within vS ([Figures 2](#fig2){ref-type="fig"}A and 2B; [Figure S1](#mmc1){ref-type="supplementary-material"}F). This provided us with experimental control over both starter cell location (by injection location of TVA-G and rabies within vS) and output projection (by injection location of AAV2-retro-Cre in target sites). This approach allowed us to assess how variations in spatial position of vS neurons (across the AP, ML, and DV axes), projection target (across PFC, NAc, and LH), and their combination influence input size and identity. In parallel, we performed control experiments where either Cre, TVA-G, or both Cre and TVA-G were excluded, to show that input labeling by rabies infection required the presence of both Cre and TVA-G ([Figures S1](#mmc1){ref-type="supplementary-material"}G--S1K). Brain-wide Rabies Tracing Reveals Biased Connectivity of vS Projection Neurons {#sec2.3} ------------------------------------------------------------------------------ To quantify input to vS, we conducted brain-wide cell counts of rabies-labeled neurons in sagittal sections and registered the data to the ABA ([@bib20], [@bib33]). Consistent with previous studies, the majority (∼90%) of direct inputs to all vS projection types and spatial locations arose locally within the hippocampal formation ([Figure 2](#fig2){ref-type="fig"}C; [Figures S2](#mmc1){ref-type="supplementary-material"}A--S2E). In addition, there were numerous long-range inputs arising from ∼35 brain regions spanning thalamic, striatal, pallidal, cortical, hypothalamic, and amygdalar regions ([Figures 2](#fig2){ref-type="fig"}C and 2D; [Figure S2](#mmc1){ref-type="supplementary-material"}E), including the nucleus of diagonal band, medial septum, nucleus reuniens (RE), posterior amygdala (PA), and preoptic area. Further, we observed that there were variations in labeling from these regions when we compared tracing from different projection targets or starter cell position (center of mass \[COM\]) along the AP axis ([Figure 2](#fig2){ref-type="fig"}D). Next, we wanted to quantitatively investigate whether long-range input onto vS neurons depended on either projection target (PFC, LH, or NAc; [Figure 3](#fig3){ref-type="fig"}A), spatial location (COM in AP, ML, and DV; [Figures 3](#fig3){ref-type="fig"}A and 3B), or their combination (see [STAR Methods](#sec4){ref-type="sec"}; [Figure 3](#fig3){ref-type="fig"}). For each brain region that contributed more than 1% of extrahippocampal input, we built a linear regression model with both COM and projection information as predictors ([Figure 3](#fig3){ref-type="fig"}C) and tested the ability of the model to predict the percentage of total extrahippocampal input from that region. Using this strategy, we found that input from medial preoptic area (MPO), PA, RE, and paraventricular thalamus (PVT) provided quantitatively different input sizes to vS depending on either COM, output projection, or both predictors ([Figure 3](#fig3){ref-type="fig"}D). Importantly, these results were independent of normalization method, because we found similar projection and spatial dependence using alternative normalization procedures ([Figures S3](#mmc1){ref-type="supplementary-material"}A and S3B).Figure 3TRIO of vS Output Neurons Reveals Projection and Spatial Bias of Inputs(A) Schematic of TRIO protocol for assessing projection (via retro-Cre injection in the output site) and spatial dependence (via varying AP injection sites of TVA-G and rabies) of inputs.(B) 3D scatterplots (left) and geometric COM of individual brain samples (right) of starter cells. COM of individual brains are represented as ellipsoids (center defined as the mean position and radii as 1 standard deviation of the cell distribution in AP, ML, and DV). The spatial distribution of starter cells along the three brain axes was comparable (standard deviation of COM: AP = 0.23 mm; ML = 0.19 mm; DV = 0.28 mm).(C) Schematic of linear regression analysis. Linear models were constructed with COM (in all three brain axes), and output type as independent variables and the extrahippocampal input fraction (log-transformed) as the dependent variable (n = 20 brains).(D) Full model fits assessed by computing the F-statistic of full models with COM and projection as predictors. Inputs from PVT, MPO, PA, and RE (red) exceed the threshold of p = 0.05 after correction for multiple comparisons.(E) Likelihood ratio tests (see [STAR Methods](#sec4){ref-type="sec"}) show that MPO and PA inputs are projection dependent, whereas PVT and RE are spatially dependent. p values have been corrected for all 15 input regions tested, but only the significant hits shown in (D) are illustrated. Dashed lines indicate p = 0.05.(F) Quantification of extrahippocampal inputs. The number of inputs from a brain region is expressed as a percentage of the total number of extrahippocampal inputs (vS^NAc^: n = 8 brains; vS^PFC^: n = 5 brains; vS^LH^: n = 7 brains). Shaded boxes indicate projection-dependent input regions with significant model fits, followed up with post hoc pairwise Tukey multiple comparisons (^∗^p \< 0.05). See also [Video S2](#mmc5){ref-type="supplementary-material"}. Data are summarized as mean ± SEM.(G) Same dataset as in (E) where the input fractions are plotted as a function of COM along the AP axis. Shaded continuous line represents smoothed input density. Inputs from RE and PVT (∗) resulted in a significant model fit with a statistically significant COM AP coefficient.See also [Figures S2](#mmc1){ref-type="supplementary-material"} and [S3](#mmc1){ref-type="supplementary-material"} and raw cell count data in [Tables S1](#mmc1){ref-type="supplementary-material"} and [S2](#mmc1){ref-type="supplementary-material"}.Video S2. Rotating 3D Whole-Brain Views of Rabies-Labeled Inputs Registered to the Allen Brain Atlas, Related to Figure 3Example data for each condition (vS^PFC^, vS^NAc^, and vS^LH^ TRIO in posterior and anterior locations) illustrated as rotating virtual glass brains. Each dot represents one rabies-labeled input neuron. We next sought to directly investigate the relative contribution of projection type or COM on the amount of RE, PVT, MPO, and PA inputs. To do this, we compared single-predictor (projection or COM) models with combined (projection and COM) models (see [STAR Methods](#sec4){ref-type="sec"}; [Figure 3](#fig3){ref-type="fig"}E). Using this approach, we found that MPO and PA innervated vS dependent solely on the projection target of the postsynaptic neuron, whereas RE and PVT inputs were dependent on COM ([Figure 3](#fig3){ref-type="fig"}E). Utilizing a similar approach, we next investigated whether AP, ML, or DV information was important for the spatial dependence of PVT and RE input, and found that for both RE and PVT, input was most dependent on starter cell location along the AP axis ([Figure S3](#mmc1){ref-type="supplementary-material"}C), consistent with our predictions from [Figure 1](#fig1){ref-type="fig"}. Post hoc testing revealed that MPO input selectively innervated vS^NAc^ neurons, and PA input selectively targeted vS^LH^ and vS^NAc^ neurons, whereas RE and PVT input targeted vS^LH^ and vS^NAc^ neurons only at posterior locations within vS ([Figures 3](#fig3){ref-type="fig"}F and 3G). Subsequently, we obtained a qualitative description of all extrahippocampal inputs (35 regions) and their relative dependence on COM or projection, which revealed a wide range in the relative dependence of different synaptic input on both COM and projection identity ([Figures S3](#mmc1){ref-type="supplementary-material"}D and S3E). Finally, we confirmed the spatial predictions of the rabies tracing data by analyzing anterograde labeling experiments available publicly from the ABA ([Figures S3](#mmc1){ref-type="supplementary-material"}F and S3G), which showed strong spatial dependence of axon innervation across the AP axis for RE and PVT, but not for MPO or PA input. Overall, our rabies tracing dataset identified brain-wide regions that project to vS, including quantitatively biased inputs from MPO, RE, PVT, and PA that depend differentially on the location and projection identity of the postsynaptic neuron. Biased Nucleus RE Input to Hippocampal Projection Neurons {#sec2.4} --------------------------------------------------------- A surprising finding in our dataset was that RE input was anatomically biased to avoid vS^PFC^ neurons ([Figure 4](#fig4){ref-type="fig"}A). This finding runs counter to classic models of the vS-PFC-RE-vS circuit where RE functions as a relay between PFC and hippocampus via long-range input to vS^PFC^ projections ([@bib19], [@bib47]). We thus sought to confirm these anatomical data using channelrhodopsin-assisted circuit mapping (CRACM) to ensure that this result was not due to methodological constraints such as viral tropism ([@bib28]) or activity dependence of viral tracing ([@bib7]) (see [Discussion](#sec3){ref-type="sec"}). From our tracing data, we hypothesized that RE input was spatially biased, i.e., RE input targets posterior areas where vS^PFC^ neurons are not abundant (COM). In addition, we wanted to ask whether RE input does not form synaptic connections even with less abundant vS^PFC^ neurons in the most posterior locations (projection).Figure 4Nucleus Reuniens Inputs to vS Projection Neurons Are Functionally Biased(A) Extrahippocampal fractions of RE inputs as a function of COM and projection population. Non-PFC projectors (i.e., pooled vS^NAc^ and vS^LH^) have overall higher RE inputs than vS^PFC^ (n = 15 non-PFC projectors, n = 5 PFC-projectors; Mann-Whitney U test, U = 12.0, p = 0.013). Data are summarized as median and interquartile range.(B) AAV expressing ChR2 under the synapsin promoter was injected bilaterally into RE.(C) Stitched confocal image of a horizontal section of hippocampus. RE axons expressing ChR2 target distal regions of vS. Scale bar: 500 μm.(D and H) In the same surgery after ChR2 injection into RE, red and green retrobeads were injected into either LH and PFC (D) or NAc and PFC (H).(E and I) Light-evoked excitatory postsynaptic currents in pairs of (E) vS^LH^ and vS^PFC^ or (I) vS^NAc^ and vS^PFC^. All data represent responses to 10 ms of blue light. Solid line: mean response; shaded region: 95% confidence interval.(F and J) Scatterplots of light-evoked photocurrents from vS^LH^ and vS^PFC^ cell pairs (F) and vS^NAc^ and vS^PFC^ cell pairs (J).(G and K) Normalized EPSCs from vS^LH^ (G) and vS^NAc^ (K) neurons scaled to the photocurrent elicited in neighboring vS^PFC^ neurons; the relative amplitudes of light-evoked photocurrents are higher in vS^LH^ (Wilcoxon signed-rank test, n = 18 pairs from 3 animals, V = 4.0, p = 3.86 × 10^−4^) and vS^NAc^ (Wilcoxon signed-rank test, n = 11 pairs from 3 animals, V = 6.0, p = 0.016). Data are summarized as median ± 95% confidence intervals.(L--N) Models of vS input connectivity based purely on (L) spatial location or (M) projection identity. (N) Our dataset supports a combined model from (L) and (M), where inputs can be either projection or COM dependent.See also [Figure S4](#mmc1){ref-type="supplementary-material"}. We injected adeno-associated virus (AAV) to express ChR2 in RE ([Figure 4](#fig4){ref-type="fig"}B) and found that ChR2^+^ axons emanating from RE were observed most densely in the distal, posterior region of vS ([Figure 4](#fig4){ref-type="fig"}C), consistent with the spatial dependence of our rabies tracing data. In these slices, we recorded light-evoked postsynaptic currents from pairs of retrogradely labeled neurons within axon-rich distal vS. By recording from pairs of neighboring neurons within the same slice in the presence of tetrodotoxin (TTX) and 4-aminopyridine (4-AP), this approach allowed us to directly compare the relative monosynaptic RE input strength across projection populations while controlling for PD and AP position, ChR2 expression, and light intensity. We observed that excitatory RE input was indeed much weaker onto vS^PFC^ neurons, whereas it strongly targeted neighboring vS^LH^ and vS^NAc^ neurons ([Figures 4](#fig4){ref-type="fig"}D--4K; [Figure S4](#mmc1){ref-type="supplementary-material"}). Overall, these results demonstrate a functional bias of RE inputs away from vS^PFC^ and toward both vS^LH^ and vS^NAc^. Discussion {#sec3} ========== Using a combination of retrograde tracing, conditional rabies tracing, and optogenetics with whole-cell electrophysiology, we demonstrated that the vS output circuitry is composed of projection-specific, topographically organized populations that receive a range of local and long-range inputs. In turn, the targeting of these inputs depends to different degrees on the spatial position and projection target of the postsynaptic vS neuron. Topography of vS Projections {#sec3.1} ---------------------------- We used CTXβ retrograde tracing to reveal the distribution of neurons in vS that project to PFC, NAc, and LH, and confirmed previous findings that suggest that there are unique populations of neurons in vS that project to each downstream region ([@bib32]). However, due to the efficiency of retrograde labeling, significant proportions of neurons projecting to multiple downstream sites cannot be definitively ruled out. Using single-neuron tracing, it was recently shown that many neurons in dorsal subiculum do indeed project to multiple downstream areas ([@bib50]). In the future, it will be interesting to look at individual neurons in vS and directly compare the extent of projection specificity and how this varies with spatial location. Our data, however, show that a large proportion of neurons in vS appear to project to unique downstream areas. Consistent with previous findings, we found that different vS projections are organized topographically. vS^PFC^ neurons were located most anteriorly (along AP), vS^LH^ were found posteriorly, and vS^NAc^ were widely distributed throughout vS ([@bib23], [@bib11]). We demonstrated that this AP gradient is most likely due to the previously described divergence along the PD axis of the pyramidal cell layer. The spatial variation of projection neurons within vS prompted us to experimentally vary our injection coordinates during rabies tracing, which enabled us to sample starter cells from different COM positions ([Figure 3](#fig3){ref-type="fig"}B). Thus, we were able to investigate the relative contribution of both projection identity and spatial location of vS neurons on the type and amount of rabies-labeled inputs. Biased Input to vS Based on Projection Identity and Spatial Location {#sec3.2} -------------------------------------------------------------------- Rabies tracing identified many inputs to vS, including MPO, PA, RE, and PVT, that targeted distinct populations in vS differentially; MPO selectively innervated vS^NAc^, PA selectively innervated vS^NAc^ and vS^LH^, and RE and PVT innervated vS^NAc^ and vS^LH^ only in posterior vS. Notably, the consistent MPO inputs to vS^NAc^ have not, to our knowledge, been previously described in the literature ([@bib51], [@bib9]). The specificity in input labeling across the different vS projections raises interesting questions regarding the function of these upstream neurons. For example, vS is important for social memory, and this behavior may be controlled in part by MPO input to vS^NAc^ ([@bib31], [@bib35]). More generally, the rabies tracing data support a model of combined topographical and output-defined connectivity of vS inputs ([Figures 4](#fig4){ref-type="fig"}L--4N) where, depending on the upstream region, vS inputs are biased according to the location, projection type, or both of these attributes. This variation in inputs across space and projection type is in keeping with the known spatial and projection-specific functions of subiculum; across PD subdivisions, proximal vS is involved in sensory encoding ([@bib25]) and distal vS supports path integration ([@bib11], [@bib25]), whereas across projection populations, vS^PFC^ and vS^LH^ encode innate threat ([@bib1], [@bib17], [@bib22]) and vS^NAc^ encodes social memory ([@bib35]). Crucially, the fact that RE inputs target vS^NAc^ and vS^LH^ neurons in posterior vS also support the existence of spatial- and projection-specific function. For example, goal-directed locomotion has been proposed to be specific to both vS^NAc^ neurons and distal subiculum ([@bib11], [@bib17], [@bib35]). Biased Connectivity of RE Input Away from vS^PFC^ Neurons {#sec3.3} --------------------------------------------------------- A surprising result from our dataset was that the RE input did not innervate vS^PFC^ neurons. RE is essential for bidirectional communication between hippocampus and PFC. This thalamic region is proposed to form an anatomical link between hippocampus and PFC, thereby closing a PFC-RE-vS-PFC loop ([@bib19], [@bib21], [@bib47]). Little information is present about the extent to which the projection-defined vS populations are involved in this circuit loop, and it is generally assumed that RE input is integrated by vS^PFC^ neurons to relay signals from hippocampus to PFC. We found that vS^PFC^ neurons receive few and functionally weak RE inputs. Thus, although vS^PFC^ neurons do indeed receive input from RE, our data suggest that the main effect of this input is by integration upstream of vS^PFC^ neurons (i.e., through dorsal hippocampus, entorhinal cortex, local interneurons, or other pyramidal populations) before being transmitted back to PFC. It will be necessary for future work to investigate in more detail the hippocampal microcircuitry that is involved in integrating RE input, the details of other multi-synaptic routes that may allow reciprocal connectivity between the hippocampus and PFC, and how this circuit organization contributes to behavior. Interestingly, we also show that RE input is anatomically and functionally biased toward vS^NAc^ and vS^LH^ neurons. This observed circuit connectivity complements previous data that demonstrated key roles for RE in goal-directed planning ([@bib21]) and fear generalization ([@bib52]), functions that may be crucial for the proposed roles of vS^NAc^ in reward seeking and vS^LH^ in anxiety ([@bib11], [@bib17], [@bib22]). Therefore, it will be important to elucidate the role of RE input in these vS-related behaviors, especially given the key role of this circuit in preclinical models of disorders such as schizophrenia, depression, and Alzheimer's disease ([@bib19], [@bib21], [@bib47]). Limitations of the Study {#sec3.4} ------------------------ Although conditional rabies tracing is now a widely used technique that enables systematic brain-wide mapping of synaptic input ([@bib6], [@bib8], [@bib18], [@bib37], [@bib3], [@bib44]), it is important to note the many caveats to this technique when considering the biological meaning of such tracing data. For instance, the mechanism of transsynaptic retrograde spread is almost completely unknown ([@bib28]). This raises potential confounds; for example, the efficiency of rabies viral spread may depend on uncontrolled variables such as tropism for certain cell types and synapses ([@bib28]), the level of upstream circuit activity ([@bib7]), or even non-synaptic transfer ([@bib28]). In addition, there is conflicting evidence as to whether the quantity of rabies-labeled input neurons correlates with functional synaptic strength ([@bib28]). Although the number of input neurons has been shown to match the connection probability and synaptic strength in multiple studies across many brain regions ([@bib27], [@bib43], [@bib26]), other studies did not reach similar conclusions and showed marked divergence ([@bib40], [@bib15]). To mitigate the potential shortcomings of the technique, it is therefore important to conduct complementary experiments and analyses. One method is to combine findings from rabies tracing experiments with analysis of publicly available anterograde tracing experiments, such as those from the ABA Mouse Connectivity Atlas ([@bib8]). For example, we used this approach to confirm direct input to vS from MPO, and also the spatial targeting of RE and PVT input ([Figure S3](#mmc1){ref-type="supplementary-material"}). A second, more powerful method is to combine tracing with follow-up experiments that directly address synaptic strength, such as CRACM ([@bib27]). We used this approach to confirm that the RE input is stronger at synapses onto vS^NAc^ and vS^LH^ and weaker onto vS^PFC^ ([Figure 4](#fig4){ref-type="fig"}). In this instance, the functional strength of input correlated with the number of rabies-labeled inputs, but this is not always guaranteed. Finally, there are other technical considerations that are important to highlight with the use of whole-brain mapping. First, the plane of sectioning is an important consideration and should be motivated by the purpose of the experiment. This is because there is potential for reduced sampling resolution across the axis of slicing, although error associated with this was recently shown to be minimal when detailed registration is carried out ([@bib20]). In our study, we used sagittal sectioning in order to allow accurate estimates of AP position within the brain ([@bib48], [@bib38], [@bib5]), but both horizontal and coronal sections have also been repeatedly and successfully used ([@bib9], [@bib11], [@bib43], [@bib6], [@bib49]). In addition, the use of stereotaxic injections means that despite small injection volume and controlled rate of injection, we cannot completely exclude the possibility that small amounts of virus or tracer might have leaked into neighboring regions. Finally, the success of tracing experiments is dependent on the efficiency of the tracing system, and novel variants of the rabies system, such as more efficient glycoproteins ([@bib24]), or more stable rabies virus ([@bib36], [@bib16], [@bib14]) could enhance the power of whole-brain anatomy experiments in the future. Overall, our study has revealed a basis for the selective control of vS projection neurons through the biased organization of brain-wide input connectivity. Further work will be required to comprehensively delineate the functional connectivity and behavioral relevance of these dedicated circuits in the execution of adaptive behavior. STAR★Methods {#sec4} ============ Key Resources Table {#sec4.1} ------------------- REAGENT or RESOURCESOURCEIDENTIFIERBacterial and Virus StrainspAAV2-retro-CAG-Cre[@bib45], UNC Vector CoreN/ApENN-AAV-hSyn-Cre-WPRE-hGHa gift from James M. WilsonAddgene viral prep \# 105553-AAV1; <http://addgene.org/105553>; RRID:Addgene_105553pAAV-synP-FLEX-splitTVA-EGFP-B19Ga gift from Ian WickershamAddgene viral prep \# 52473-AAV1; <http://addgene.org/52473>; RRID:Addgene_52473RabiesΔG-EnvA-H2B-mCherry-2A-CLIPa gift from Marco Tripodi, MRC LMBN/ApAAV-hSyn-hChR2(H134R)-EYFPa gift from Karl DeisserothAddgene viral prep \# 26973-AAV1; <http://addgene.org/26973>; RRID:Addgene_26973Chemicals, Peptides, and Recombinant ProteinsRed retrobeadsLumafluorItem \#: R180Green retrobeadsLumafluorItem \#: G180Cholera Toxin Subunit B (Recombinant), Alexa Fluor 647 ConjugateInvitrogenCat \#C34778Cholera Toxin Subunit B (Recombinant), Alexa Fluor 555 ConjugateInvitrogenCat \#C34776ProLong Gold Antifade Mountant with DAPIInvitrogenCat \#P36930ProLong Glass Antifade Mountant with DAPIInvitrogenCat \#P36984IsofluranePiramal Critical CareCarprofenNorbrookTetrodotoxinHello BioCat \#HB10354-aminopyridineHello BioCat \#HB1073Experimental Models: Organisms/StrainsMouse: C57BL/6Charles RiverN/ASoftware and AlgorithmsImageJ (Fiji) Software<https://fiji.sc/>N/APython 3.6<https://www.python.org/>N/AJupyter Notebook<https://jupyter.org/>N/AR<https://www.r-project.org/>N/AWholeBrain<http://www.wholebrainsoftware.org/>[@bib20]N/A Lead Contact and Materials Availability {#sec4.2} --------------------------------------- Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Andrew MacAskill (<a.macaskill@ucl.ac.uk>). This study did not generate new unique reagents. Experimental Model and Subject Details {#sec4.3} -------------------------------------- ### Mice {#sec4.3.1} Young adult C57BL/6 male mice (CTXβ and rabies tracing: at least 7 weeks old; physiology: 7 -- 9 weeks old) provided by Charles River were used for all experiments. All animals were housed in cages of 2 to 4 in a temperature- and humidity-controlled environment with a 12 h light-dark cycle (lights on at 7 am to 7 pm). Food and water were provided ad libitum. All experiments were approved by the UK Home Office as defined by the Animals (Scientific Procedures) Act, and strictly followed University College London ethical guidelines. Method Details {#sec4.4} -------------- ### Stereotaxic surgery {#sec4.4.1} Stereotaxic surgeries were performed according to previously described protocols ([@bib13]). Mice were anaesthetised with isoflurane (4% induction, 1.5 to 2% maintenance) and secured onto a stereotaxic apparatus (Kopf). A single incision was made along the midline to reveal the skull. AP, ML and DV were measured relative to bregma, and craniotomies were drilled over the injection sites. Long-shaft borosilicate pipettes were pulled and backfilled with mineral oil, and viruses were loaded into the pipettes. Viruses were injected with a Nanoject II (Drummond Scientific) at a rate of 13.8 nL every 10 s. Following infusion of the virus, the pipette was left in place for an additional 10 mins before being slowly retracted. The following coordinates (in mm) were used:Injection siteMLAPDVMedial prefrontal cortex:0.4+2.3−2.4Lateral hypothalamus:0.9−1.3−5.2Nucleus accumbens (medial shell):0.9+1.1−4.6Ventral subiculum (anterior):3.4−3.2−4.3Ventral subiculum (posterior):3.4−3.7−4.3Nucleus reuniens0.25−0.7−4.4 Following injection of substances into the brain, animals were sutured and recovered for 30 mins on a heat pad. Animals received carprofen as a peri-operative s.c. injection (0.5 mg / kg) and in their drinking water (0.05 mg / mL) for 48 hours post-surgery. The titers of viruses used are the following:•RabiesΔG-EnvA-H2B-mCherry-2A-CLIP, 1.8 × 10^8^ genome copies (gc) / mL;•pAAV-synP-FLEX-splitTVA-EGFP-B19G, 3.9 × 10^12^ gc / mL;•pENN-AAV-hSyn-Cre-WPRE-hGH, 1.3 × 10^13^ gc / mL;•pAAV-hSyn-hChR2(H134R)-EYFP, 2.5 × 10^13^ gc / mL;•pAAV2-retro-CAG-Cre, 5.3 × 10^12^ gc / mL ### Retrograde tracing {#sec4.4.2} For CTXβ retrograde tracing, 150 nL of Alexa 555- or 647-tagged CTXβ was injected into one of three output regions (PFC, NAc or LH). After at least 14 days post-surgery, animals were sacrificed for histology. For rabies monosynaptic tracing, experiments were done according to previously described protocols ([@bib6]). Adult male mice were injected with 100 nL of AAV2-retro-CAG-Cre or AAV2-retro-synP-Cre into one of three output regions (PFC, NAc or LH), and in the same surgery, 250 nL of AA2/1-synP-FLEX-split-TVA-EGFP-B19G was injected into ventral subiculum (anterior) or ventral subiculum (posterior). In a second surgery after at least 2 weeks, 300 -- 400 nL of EnvA-RVΔG-H2B-mCherry was injected into vS. After 7 days of rabies expression, animals were sacrificed for histology. ### Histology and imaging {#sec4.4.3} Animals were deeply anaesthetised with a lethal dose of ketamine and xylazine (100 mg/kg) and perfused transcardially with phosphate-buffered saline (pH = 7.2) followed by 4% paraformaldehyde. Brains were dissected and post-fixed overnight at 4°C prior to sectioning. For CTXβ tracing analysis, 60-μm sections in the horizontal plane were prepared using a vibratome (Campden Instruments). For brain-wide rabies tracing analysis, 60-μm sections in the sagittal plane were prepared with a supporting block of agar, and every 2^nd^ section was mounted and analyzed. Sections were mounted on Superfrost Plus slides with ProLong Gold or ProLong Glass antifade mounting medium (Molecular Probes) and imaged with a 5x objective using a Zeiss Axioscan Z1 or 10x objective using a Zeiss SLM 800, using standard filter sets for excitation/emission. Tiled images were stitched automatically during image acquisition by Zen software to generate the full images displayed. ### Whole-hippocampus cell quantification and analysis {#sec4.4.4} For quantification of CTXβ-labeled hippocampal neurons, consecutive 60-μm coronal sections spanning the hippocampus (approximately AP −4.3 mm to −1.0 mm) were collected. Cell counting of retrogradely labeled neurons was conducted using custom written scripts in R based around the WholeBrain package ([@bib20]), a recently developed automatic segmentation and registration workflow in R. Only sections containing labeled neurons in the hippocampus (∼-4.1 mm to −2.7 mm) were analyzed. Segmentation was performed using wavelet multiresolution decomposition on the WholeBrain platform in R, and the segmentation parameters (pixel threshold, soma size, brain outline etc.) were adjusted slice by slice to achieve accurate segmentation of neurons. For the registration of coronal sections, six to eight random brain sections were sampled and manually annotated with approximate AP coordinates, and the remaining sections were assigned AP coordinates based on interpolation. For a given experiment, two projection-specific hippocampal populations (pairs of NAc, LH or PFC) were labeled with two fluorophores (Alexa Fluor 555 and Alexa Fluor 647) in a counterbalanced order. First, segmentation was conducted for each channel individually. Subsequently, registration was conducted on one image data from one channel in a semi-automated manner; this involved an automatic registration of the tissue section by the WholeBrain software, and then refining the registration through manually adding, subtracting or changing correspondence points at clear anatomical landmarks. The same anatomical registration was applied to register the tissue section image obtained from the other channel. Cell count data were saved as RData files, imported into Python and analyzed using Python 3.6. For spatial distribution, correlation and logistic regression analyses, individual sections were manually assigned AP coordinates by first estimating the posterior-most AP coordinate using the ABA as a reference, and then labeling the next anterior coronal section at 60-μm increments until the last section in the coronal section series. ML and DV positions were calculated through registration of individual neurons to the ABA (see above). Only hippocampal cells in subiculum ('SUB') and CA1 residing in ventral hippocampus (∼DV −3.5 and −4.5 mm) were analyzed. Cell distributions across AP, ML and DV were determined by kernel density estimation using a Gaussian kernel function and bandwidth estimated by Scott's rule. Correlation analysis was conducted to determine which brain axis covaries most with the projection type of hippocampal neurons. For each pair of injections in a given hemisphere, the covariation of spatial position along each axis with projection type was computed using Pearson correlation using the scipy function scipy.stats.pointbiserialr. The absolute value of the correlation coefficient indicates the degree of covariation, where \\|r\\| = 1 indicates perfect correlation, and \\|r\\| = 0 indicates no correlation; the absolute correlation coefficient was then compared across the different brain axes. To identify which axis contributes most information to predict the projection class of hippocampal neurons, we conducted logistic regression analysis using the sklearn package in Python. For this analysis, the total ventral hippocampal cell counts from each hemisphere were divided into 80% training dataset to train the multinomial logistic regression model, and 20% test dataset to examine the performance of the model. The train-test-split was crucial to assess how well the linear classifiers generalized to unseen data. The 10-fold cross-validated accuracy of the models was calculated using the LogisticCV function, and the classifiers were assessed for how much their performance degraded after removal of each brain axis (AP, ML or DV) as a predictor in the model. This reduction in accuracy after the removal of one predictor (cv. Δ accuracy) indicates the unique information that spatial position along one axis contributes to their performance. Note that 2 hemispheres out of 12 total hemispheres from 6 animals were excluded from the dataset due to poor hippocampal labeling. ### Analysis of spatial positions of vS projections along the PD axis {#sec4.4.5} For a subset of CTXβ injected brains, brain sections were prepared in the horizontal plane to analyze the spatial positions of retrogradely labeled vS projections to PFC, NAc and LH along the PD axis. Six to eight sections spanning vS (approximately DV −3.5 to −4.5 mm) were analyzed per brain hemisphere. Using ImageJ, horizontal sections of the hippocampus were digitally straightened from the dentate gyrus to the end of subiculum using the Straighten function on ImageJ to approximate the PD axis. Labeled cells in each slice were manually counted and registered to this axis using the ImageJ CellCounter plugin. The radial (y-coordinate) and PD (x-coordinate) positions of each cell were used for spatial position analysis. The DV positions of each cell was manually annotated for each hippocampal slice based on the Paxinos atlas. The CA1/subiculum border occurs approximately at 0.7 within this normalized PD axis range and was anatomically defined as the disappearance of stratum oriens and the fanning out of the pyramidal cell layer. The spatial positions were analyzed with custom Python routines. ### Linear discriminant analysis (LDA) of registered hippocampal neurons {#sec4.4.6} LDA was used to unbiasedly identify the plane which most optimally separates the clusters of CTXβ-labeled, projection-specific hippocampal neurons. LDA is a dimensionality reduction method that identifies the subspace which maximizes the ratio of the between-class over the within-class variability. The between- and within-class variability were calculated as scatter matrices S~B~ and S~W~, respectively, where S~B~ and S~W~ are 3x3 matrices and the number of rows or columns corresponds to each brain axis (AP, ML and DV).The predictor variables for the LDA analysis were the registered spatial positions in AP, ML and DV, and the target variable was the projection type (encoded labels of NAc, PFC or LH). We focused on hippocampal cells labeled between DV −4.5 mm and −4.0 mm, and this dataset (n = 14,527 ventral hippocampal neurons counted from 10 experiments; 6 animals) was split into an 80% training and 20% held-out test dataset. The projection matrix used to transform the retrogradely labeled neurons to the subspace that best maximizes discriminability was solved by matrix factorisation using singular value decomposition (SVD) based on the LinearDiscriminantAnalysis function from the sklearn package. The spatial positions of the held-out test dataset were then projected onto the subspace by matrix multiplication of the spatial position and projection matrices:$$X' = X\varphi$$where $\varphi$ is the eigenvector (projection) matrix whose columns correspond to the eigenvectors (linear discriminant vectors), and X is the matrix of spatial positions of each registered neuron. The transformed spatial positions were plotted in the LD subspace, and the cell distributions in the first and second linear discriminants were determined by kernel density estimation. ### Mapping and analysis of rabies-labeled inputs {#sec4.4.7} Cell counting of rabies-labeled inputs was conducted using WholeBrain ([@bib20]). After acquiring the imaged sections and exporting them as 16-bit depth image files, images in the rabies mCherry channel were manually assigned a bregma coordinate (ML −4.0 to 0.0 mm) and processed using WholeBrain ([@bib20]) and custom cell counting routines written in R. The workflow comprised of (1) segmentation of cells and brain section, (2) registration of the cells to the ABA and (3) analysis of anatomically registered cells. As tissue section damage impairs the automatic registration implemented on the WholeBrain platform, sections with poor registration were manually registered to the atlas plate using corresponding points to clear anatomical landmarks. Once all cells have been registered, the cell counts were further manually filtered from the dataset to remove false-positive cells (e.g., debris). Virtually all cells were detected in the injected hemisphere, apart from a consistent set of contralateral CA3 inputs ([@bib9]). Therefore, only the injected hemisphere up to the midline was used for cell quantification analysis. Each cell registered to a brain region was classified as belonging to an anatomically defined region as defined by the ABA brain structure ontology. Information on the ABA hierarchical ontology was scraped from the ABA API (link: <http://api.brain-map.org/api/v2/structure_graph_download/1.json>) using custom Python routines. The coarse-level (or parent) structures to which each cell belonged were defined a priori and comprised the following: Hypothalamus, Isocortex, Hippocampal formation, Thalamus, Cortical subplate, Pallidum, Striatum, Midbrain, Pons and Medulla. All cells falling under these parent structures were analyzed. The fine-level (or child) structures represent all brain regions existing as subcategories of the corresponding coarse-level structure, e.g., nucleus reuniens and paraventricular thalamus are fine-level (child) structures relative to Thalamus (parent). For quantification of input fractions, cells residing in different layers within the same structure, e.g., CA1 stratum oriens (CA1so) and stratum lacunosum-moleculare (CA1slm), or subdivisions of nuclei, e.g., basomedial amygdala, posterior division (BMAp) and anterior division (BMAa), were agglomerated across layers and subdivisions and counted as residing in one single region (CA1 and BMA, respectively). Note that lateral and medial entorhinal cortex (LEC and MEC, respectively) were analyzed as separate structures. ### Starter cell center of mass (COM) quantification {#sec4.4.8} To determine the starter cell COM, every 2^nd^ sagittal section from both the rabies mCherry+ and TVA-G GFP+ channels that spanned the extent of the TVA-G GFP+ expression in vS was obtained and analyzed. Images were collected in order from lateral to medial. Colocalized cells representing starter cells were manually counted and registered onto digital plates from the Paxinos atlas using the ImageJ ROI Manager function. Cell positions in AP and DV were calculated by first obtaining the spatial position of cells in pixel space. Then, the location of bregma in pixel space was determined by placing an ROI at the approximate site of bregma in the Paxinos atlas. Finally, a scaling factor was calculated by estimating the change in pixel space with a unit change in ML and DV. The final positions of cells were calculated by offsetting the positions of the cells in pixel space to that of bregma and normalizing the positions by the scaling factor. As the TVA-G construct expressed TVA and G bicistronically (i.e., the exon for TVA and G are linked by a self-cleaving 2A peptide), all cells were assumed to express TVA and G in a 1:1 stoichiometry and unlikely to express one gene without the other. Therefore, all colocalized cells were treated as starter cells. ### Analysis of COM versus projection dependence of inputs {#sec4.4.9} Inputs arising from within the hippocampal formation were analyzed as input fractions of the total inputs counted in the same brain, while inputs arising from outside the hippocampal formation were analyzed as input fractions normalized to the total number of extrahippocampal inputs. The dataset containing cell counts from n = 20 brains were analyzed according to projection or COM. For extrahippocampal inputs, only fine-level structures exceeding 1% of extrahippocampal input were assessed (15 brain regions). Multiple linear regression modeling was performed to compare the relative influence of COM and projection identity on the amount of rabies-labeled inputs in an input region of interest. For each input brain region, a multiple regression model -- in which the predictor variables were the COM and projection identity -- was constructed using the ols function from the statsmodel package in Python. The overall statistical significance of full models (containing COM and projection as predictors) were assessed by computing the F-statistic values. All p values generated from multiple comparisons were corrected using the Benjamini-Hochberg method (false-discovery rate \< 0.05). The models with adjusted p values \< 0.05 were further analyzed for statistically significant coefficients using the Wald test and followed up with post hoc pairwise Tukey multiple comparisons between vS projection populations (see [Figures S3](#mmc1){ref-type="supplementary-material"}C and [3](#fig3){ref-type="fig"}F). To further assess the importance of each predictor to the model, the likelihood ratio (LR) test was used to compare the full model to a reduced model containing only one predictor -- either COM or projection. The reduced models containing either COM or projection identity as a predictor were built using the same ols function. The LR test was computed using the following formula from the statsmodel package function compare_lr_test:$$D = - 2\ \log\left( \frac{\mathcal{L}_{restricted}}{\mathcal{L}_{full}} \right)$$where $\mathcal{L}$ is the likelihood of the model, and D is a test statistic that follows a $\chi^{2}$ distribution with degrees of freedom (df) equal to the difference in the number of predictors between the full and restricted (reduced) model. The p values generated from multiple LR tests were corrected using the Benjamini-Hochberg method (FDR \< 0.05). As complementary analyses, the projection-dependence of extrahippocampal inputs from [Figure 3](#fig3){ref-type="fig"}F was further analyzed using multiple one-way ANOVAs, while the spatial-dependence of these inputs from [Figure 3](#fig3){ref-type="fig"}G was assessed using multiple Spearman rank correlation tests of input fractions against starter cell COM. This analysis revealed similar patterns of biased connectivity to the multiple linear regression analysis ([Table S2](#mmc1){ref-type="supplementary-material"}), where MPO and PA were detected as projection-dependent inputs, but RE was detected as being COM-dependent. For spatial-dependence analysis, the input density for each brain region as a function of COM was visualized by first sorting the input fractions of each brain by the COM in the posterior-to-anterior direction. The array of input fractions was then interpolated to produce 500 data points, smoothed with a Savitzky-Golay filter (window size = 51 points, order = 3), and normalized by dividing each data point by the total area under the curve. To assess the relative goodness-of-fit of non-nested COM models or projection models, we performed linear regression models with only one predictor by using the ols function from the Python package statsmodels. Linear models were built for each brain region, where the target variable was the input fraction observed in that brain region normalized to the total number of inputs, and the predictor variable was either the COM or projection. For each of 35 extrahippocampal inputs in total, there were 20 observations (n = 20 brains). The models were fitted, and goodness-of-fit was measured using the Bayesian information criterion (BIC). The BIC was computed using the following formula:$$BIC = k\ \mspace{9mu}\log\ n - 2\ \log\widehat{\mathcal{L}}$$where k is the number of model parameters, n is the observation number and $\widehat{\mathcal{L}}$ is the maximum likelihood of the model. Finally, to compare model fits, the difference between the BIC values obtained from COM and projection models were computed to obtain ΔBIC. ### ABA Mouse Brain Connectivity analysis of axonal projections {#sec4.4.10} To validate the spatial targeting of rabies-labeled inputs as predicted by the multiple linear regression analysis, analysis of axonal input density arising from the input region 'hits' (i.e., MPO, RE, PA and PVT) was conducted. For MPO, PVT and RE, three separate experiments were analyzed. For PA, the only two experiments available were analyzed. All images were downloaded from the ABA API using the Python package Allen SDK (2015). The following experiments were used:InputExperiment IDTransgenic lineImage sections used for analysisMPO158738180C57BL/6158738374, 158738370, 158738366, 158738362, 158738358, 158738354119846838C57BL/6119847348, 119847344, 119847340, 119847336, 119847332, 119847328182459635Gal-Cre_KI87182460001, 182459997, 182459993, 182459989, 182459985, 182459981RE204832205Adcyap1-2A-Cre204832574, 204832570, 204832566, 204832562, 204832558, 204832554175374982C57BL/6175375180, 175375176, 175375172, 175375168, 175375164, 175375160174957972C57BL/6174958286, 174958282, 174958278, 174958274, 174958270, 174958266PA304721447Rbp4-Cre_KL100304721806, 304721802, 304721798, 304721794, 304721790, 304721786545415593Npr3-IRES2-Cre545415944, 545415940, 545415936, 545415932, 545415928, 545415924PVT278510903Ppp1r17-Cre_NL146278511259, 278511255, 278511251, 278511247, 278511242, 278511238183225830Grm2-Cre_MR90183226060, 183226056, 183226052, 183226048, 183226044, 183226040301209502Efr3a-Cre_NO108301209844, 301209840, 301209836, 301209832, 301209828, 301209824 For each experiment, six consecutive coronal images (spanning approximately −3.7 to −3.1 mm AP relative to bregma) of ventral hippocampus were downloaded and analyzed using ImageJ. Only one hemisphere per brain was analyzed. The brightness and contrast values were fixed for all sections obtain from a single experiment to allow comparison of the relative fluorescence of axonal projections across slices. Images were downloaded with a downsample factor of 8, ROIs were manually drawn around the vS (defined as the region below the rhinal sulcus, between the alveus and the boundary of the CA1 stratum lacunosum moleculare and adjacent dentate gyrus or CA3 stratum radiatum border, see also [Figures S3](#mmc1){ref-type="supplementary-material"}F and S3G), and the mean pixel intensity within the ROIs were measured. The pixel intensity values were normalized to that from the posterior-most section, and the relative fluorescence intensities were compared across the AP axis. ### CRACM experiments {#sec4.4.11} #### Surgeries {#sec4.4.11.1} CRACM of projection-defined vS neurons was done according to previously described protocols ([@bib29]). 7- to 9-week old animals were injected with 250 nL of red (1:10 dilution in sterile saline) or green (undiluted) retrobeads in a counterbalanced order into two of three output regions (PFC, NAc or LH). In the same surgery, 250 nL of AAV1-synP-ChR2-YFP was injected into RE bilaterally. After at least 14 days of ChR2 expression, animals were sacrificed for electrophysiological recording. #### Slice preparation {#sec4.4.11.2} Acute, transverse hippocampal slices were used for all electrophysiological recordings. Mice were deeply anaesthetised with a lethal dose of ketamine and xylazine (100 mg / kg), and perfused transcardially with ice-cold sucrose solution containing (in mM): 190 sucrose, 25 glucose, 25 NaHCO~3~, 1.2 NaH~2~PO~4~, 10 NaCl, 2.5 KCl, 1 Na^+^ ascorbate, 2 Na^+^ pyruvate, 7 MgCl~2~ and 0.5 CaCl~2,~ bubbled continuously with 95% O~2~ / 5% CO~2~. Following perfusion, mice were decapitated, and their brains were rapidly dissected. The dissected brains were then placed in ice-cold sucrose solution and hemisected. The cerebellum was removed, and transverse slices were prepared using a vibratome (VT1200S, Leica), with a ∼10° angle along the ventromedial plane to obtain sections that were orthogonal to the long-axis of the hippocampus. The thickness of hippocampal sections was 300 μm. Slices were transferred to a bath containing artificial cerebrospinal fluid (aCSF) and recovered first for 30 mins at 37°C, and subsequently for 30 mins at room temperature. The aCSF solution contained (in mM): 125 NaCl, 2.5 KCl, 1.25 NaH~2~PO~4~, 22.5 glucose, 1 Na^+^ ascorbate, 3 Na^+^ pyruvate, 1 MgCl~2~, 2 CaCl~2~. All recordings were performed at room temperature (22 -- 24°C). All chemicals were from Sigma or Tocris. #### Whole-cell electrophysiology and optogenetics {#sec4.4.11.3} Whole-cell recordings were performed on retrogradely labeled hippocampal pyramidal neurons with retrobeads visualized by their fluorescent cell bodies and targeted with Dodt contrast microscopy. For sequential paired recordings, neighboring neurons were identified using a 40x objective at the same depth into the slice. The recording order of neuron pairs was alternated to avoid complications due to rundown. Borosilicate recording pipettes (3 -- 5 MΩ) were filled with a Cs-gluconate internal solution containing (in mM): 135 Gluconic acid, 10 HEPES, 10 EGTA, 10 Na-phosphocreatine, 4 MgATP, 0.4 Na~2~GTP, 10 TEA and 2 QX-314. Presynaptic glutamate release was elicited by illuminating ChR2 expressed in the presynaptic terminals of long-range inputs into the slice, as previously described ([@bib29]). Wide-field illumination was achieved through a 40x objective with brief 10 ms pulses of blue light over the recorded neuron from an LED centered at 473 nm (CoolLED pE-4000, with corresponding excitation-emission filters). Light intensity was measured as 4--7 mW at the back aperture of the objective and was constant between all recorded cell pairs. When recording a cell pair, the LED power was adjusted until responses were ∼200 pA in a connected neuron or set to maximum in an unconnected neuron, and fixed at the same level when recording light-evoked responses in the other cell. In all experiments, the aCSF contained 1 μM TTX and 100 μM 4-AP to isolate monosynaptic connectivity and increase presynaptic depolarisation, respectively. Recordings were conducted using a Multiclamp 700B amplifier (Axon Instruments), and signals were low-pass filtered using a Bessel filter at 1 kHz and sampled at 10 kHz. Data were acquired using National Instruments boards and WinWCP (University of Strathclyde) and analyzed using custom routines written in Python 3.6. For synaptic connectivity analysis, six recording sweeps were obtained for each optical pulse duration. The signals were preprocessed by baselining the signals to the first 100 ms, low-pass filtering using a Bessel filter (cutoff = 2 Hz, order = 2), averaging across the six recording sweeps for each optical pulse duration, and decimating the averaged signal to 1 kHz. The peak amplitude response of light-evoked EPSCs was measured as averages over a 2-ms time window around the peak compared to a 2-ms baseline period preceding the optical pulse. Only paired data in which at least one cell received \> 5 pA were included for analysis for [Figures 4](#fig4){ref-type="fig"}D--4K, while all cells irrespective of connectivity were included for the analysis described in [Figure S4](#mmc1){ref-type="supplementary-material"}. We note that due to the jitter in release as a result of the use of presynaptic ChR2, across-cell averages of postsynaptic currents ([Figures 4](#fig4){ref-type="fig"}D--4K) may appear relatively slow. For each recorded neuron, the spatial position was obtained by manually registering the cell to a digital atlas depicting a horizontal section of the vS. The long axis of the CA1 and subiculum area (i.e., the y-coordinates) of each registered cell was used to approximate the anterior-posterior position. Input connectivity was determined by a threshold light-evoked response of \> 5 pA. Quantification and Statistical Analysis {#sec4.5} --------------------------------------- All statistics were calculated using Python's scipy and statsmodels packages, and R. Summary data are reported throughout the text and figures as mean ± sem unless otherwise stated. For [Figures 1](#fig1){ref-type="fig"} and [4](#fig4){ref-type="fig"}, normality of data distributions was determined by visual inspection of the data points. For [Figures 3](#fig3){ref-type="fig"} and [S2](#mmc1){ref-type="supplementary-material"}, the rabies tracing dataset was assessed for normality with the Jarque-Bera test and equal variance with the Levene's test. The input fractions for multiple regions showed non-normality and heteroscedasticity in the input fractions across projection populations. Therefore, all analysis for [Figure 3](#fig3){ref-type="fig"} were conducted after log-transformation of input fractions. Distributions of the data became more Gaussian-like after log-transformation, as assessed again by the Jarque-Bera test and Levene's test. All data were analyzed using statistical tests described in the figure legends and in [Table S2](#mmc1){ref-type="supplementary-material"}. The alpha level was defined as 0.05. No power analysis was run to determine sample size a priori. The sample sizes chosen are similar to those used in previous publications. Data and Code Availability {#sec4.6} -------------------------- Programme code, original data and scripts for statistical tests and analysis are available upon request. Raw data are available at <https://github.com/macaskill-lab>. Supplemental Information {#app2} ======================== Document S1. Figures S1--S4 and Tables S1 and S2Data S1. Raw Cell Counts in Each Condition of the TRIO Experiments (vS^PFC^, vS^NAc^, and vS^LH^ Projections with Associated COM Information), Related to Figure 3Data S2. Abbreviations of the 115 Brain Regions with Raw Cell Counts Shown in Data S1, Related to Figure 3Document S2. Article plus Supplemental Information We thank Marco Tripodi and Fabio Morgese for the unpublished pseudotyped rabies virus, Francesca Cacucci for assistance with surgery, Harsha Gurnani for discussions about LDA, and Rachel Wee for help with illustrations. We are grateful to Larry Swanson, Andy Murray, Tara Keck, Kenneth Harris, and the MacAskill lab for helpful comments on the manuscript. A.F.M. was supported by a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society (grant number 109360/Z/15/Z) and by a UCL Excellence Fellowship. R.W.S.W. was supported by a UCL Graduate Research Scholarship and a UCL Overseas Research Scholarship. Author Contributions {#sec5} ==================== Conceptualization, R.W.S.W. and A.F.M.; Methodology, R.W.S.W. and A.F.M.; Investigation, R.W.S.W. and A.F.M.; Formal Analysis, R.W.S.W. and A.F.M.; Writing -- Original Draft, R.W.S.W. and A.F.M.; Writing -- Review & Editing, R.W.S.W. and A.F.M.; Funding Acquisition, A.F.M.; Supervision, A.F.M. Declaration of Interests {#sec6} ======================== The authors declare no competing interests. Supplemental Information can be found online at <https://doi.org/10.1016/j.celrep.2020.02.093>. [^1]: Lead Contact	{ "pile_set_name": "PubMed Central" }
SpecificationsOrganism/cell line/tissueNicotiana tabacumSequencer or array typeIllumina HiSeq™ 2000-RNA sequencingData formatRaw and processedExperimental factorsWo^v^ transgenic and wild-type tobaccoExperimental featuresRNA-seq dataset for gene expression profiling in young primary leaf in Wo^v^ transgenic and wild-type tobaccoSample source locationWuhan, China 1. Direct link to deposited data {#s0005} ================================ Deposited data can be found here: <http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE72310>. 2. Experimental design, materials and methods {#s0010} ============================================= 2.1. Sample collection and RNA isolation {#s0015} ---------------------------------------- Wo^v^ transgenic tobacco plants and their WT species (N. tabacum) were grown in a naturally illuminated glasshouse in the experimental field of the Huazhong Agricultural University. Environmental control was implemented based on standard protocols [@bb0005]. Two independent Wo^v^ transgenic tobacco plants and two negative regenerated plants were used for transcriptomic analysis. All samples were immediately frozen in liquid nitrogen, and total RNA was isolated using TRIzol reagent (Invitrogen). The integrity of all RNA samples was examined by 1.2% agarose gel electrophoresis after treatment with RQ1 DNase (Promega). The quality and quantity of these RNA samples were further determined by measuring the absorbance at 260/280 and 260/230 nm by using SmartSpec Plus spectrophotometer (Bio-Rad). 2.2. RNA-seq and data processing {#s0020} -------------------------------- The materials used for RNA-seq analyses were three young primary leaves from the transgenic and control plants. Up to 10 μg of total RNA was sent to ABlife Inc. (Wuhan, China) where the libraries were produced. The cDNA libraries were then sequenced in BGI Inc. (Shenzhen, China) by using Illumina HiSeq™ 2000 by 100 nt pair-end sequencing. RNA-seq was performed as previously described [@bb0010]. Whole RNA-seq data were submitted to the Gene Expression Omnibus (series accession number [GSE72310](ncbi-geo:GSE72310){#ir0015}). The clean reads were aligned to the N. tabacum genome (<ftp://anonymous@ftp.solgenomics.net/genomes/Nicotiana_tabacum/assembly/Ntab-K326_AWOJ-SS.fa.gz>) using TopHat (2.0.12) software [@bb0015]. Based on the length of the gene and read counts uniquely mapped to this gene, gene expression levels were calculated using FPKM method. 2.3. Functional analysis of differentially expressed genes {#s0025} ---------------------------------------------------------- We used the DESeq software to identify the differentially expressed genes between samples, which were specific for differential expression analysis of the RNA-Seq data with biological replicates [@bb0020]. Genes with P-value ≤ 0.05 and fold change ≥ 2 were regarded as differentially expressed. To characterize the putative functions of differentially expressed genes, Gene Ontology (GO) term analysis was also performed using GOseq based Wallenius non-central hyper-geometric distribution [@bb0025]. For further identification of pathways significantly influenced by the Wo^v^ gene, KEGG enrichment pathways analysis of DEGs was implemented by the KOBAS (2.0) software by a hypergeometric test and the Benjamini--Hochberg FDR correction (FDR ≤ 0.05) [@bb0030]. [^1]: These two authors contributed equally to this work.	{ "pile_set_name": "PubMed Central" }
INTRODUCTION {#s1} ============ Notch signaling is involved in various cellular processes, including cell fate specification, differentiation, proliferation, and apoptosis. Abnormal Notch signaling is oncogenic in several cancers, including hematologic malignancies \[[@R1]\]. Disregulated Notch signaling has also been associated with chronic lymphocytic leukemia (CLL), a frequent adult leukemia characterized by the accumulation of CD19^+^/CD5^+^ B lymphocytes resistant to apoptosis \[[@R2]\]. The involvement of Notch in CLL has been recently demonstrated by findings that a NOTCH1 PEST domain mutation, generating a highly active truncated protein, and affecting up to 10-15% of patients, is associated with poor prognosis, disease progression and refractoriness to chemotherapy \[[@R3]-[@R7]\]. Previous evidence that constitutive activation of Notch1 and Notch2 signaling contributes to apoptosis resistance in CLL also underscores the importance of Notch in this leukemia \[[@R8], [@R9]\], encouraging further investigation of its therapeutic potential. Indeed, a better understanding of the mechanisms involved in the anti-apoptotic signaling of Notch in CLL cells may provide insight for designing future Notch-targeted therapies. In cancer, Notch signaling prevents apoptosis through different networks, involving cell cycle and survival pathways, and interactions with mitochondria. Notch suppresses p53 \[[@R10]\] or JNK function \[[@R11]\] as well as the expression of the pro-apoptotic proteins Bax, Bim and Noxa \[[@R12], [@R13]\]. Notch increases the activation of the pro-survival PI3K/AKT \[[@R14]\] and NF-kB pathways \[[@R15]\] and the expression of the anti-apoptotic proteins Bcl-2 and Bcl-xL \[[@R12]\], stabilizes the apoptosis inhibitor protein XIAP \[[@R16]\] and induces mitochondrial integrity and functions \[[@R17]\]. Some of these functions are mediated by the transcriptional activity of Notch-intracellular domain (ICD) \[[@R10], [@R14], [@R15]\], which, after Notch-ligand interactions, is released from the membrane to the nucleus. Other functions involve a non-canonical Notch-ICD-activated signaling which operates in the cytoplasm \[[@R11], [@R16]\] and can also converge on the mitochondria by promoting cell survival \[[@R17]\]. A crucial role in controlling mitochondrial integrity and apoptosis is played by the balance between pro-apoptotic and anti-apoptotic Bcl-2 family members \[[@R18]\]. In CLL as well as in other hematologic malignancies, the over-expression of the anti-apoptotic Mcl-1 and Bcl-2 proteins is one of the major causes of apoptosis resistance \[[@R19], [@R20]\], poor prognosis \[[@R21], [@R22]\] and chemoresistance \[[@R23]-[@R25]\]. Mcl-1 and Bcl-2 bind and sequester the pro-apoptotic proteins Bax and Bak blocking their ability to form pores in the mitochondrial membrane with the consequent release of cytochrome c into the cytoplasm. Degradation of Mcl-1 frees Bax and Bak allowing their polymerization and activating apoptosis \[[@R18]\]. Mcl-1 is a short-lived protein tightly regulated by transcriptional \[[@R26], [@R27]\], translational \[[@R28], [@R29]\], and degradation mechanisms \[[@R30], [@R31]\]. Interestingly, Mcl-1 mRNA translation is highly dependent on the eukaryotic initiation factor 4E (eIF4E), a key component of the mRNA cap-binding complex, which preferentially enhances translation of a subset of mRNAs with complex 5^\'^ untranslated regions, such as those of Mcl-1 and several other transformation-related and survival proteins \[[@R32]-[@R34]\]. eIF4E has been associated with cancer development and progression, and proposed as an important therapeutic target \[[@R35], [@R36]\]. Recently, it has been demonstrated that CLL cells also express high levels of eIF4E, and that its pharmacologic targeting increases in vitro fludarabine cytotoxicity, suggesting an involvement of eIF4E in chemoresistance of these cells \[[@R37]\]. Considering the critical role of Mcl-1 and eIF4E in CLL and in other malignancies, in this study, we investigated whether Mcl-1 and eIF4E are targets of the anti-apoptotic Notch signaling in CLL. RESULTS {#s2} ======= Notch1 and Notch2 downregulation decreases viability of CLL cells from different patient subgroups {#s2_1} -------------------------------------------------------------------------------------------------- We used small-interfering RNA (siRNA) and nucleofection to silence expression of Notch1 and Notch2 in all 22 patients included in the study. Table [1](#T1){ref-type="table"} gives clinical and biological characteristics of CLL patients. Downregulation of the expression of each Notch receptor, achieved at different levels in all samples examined (Table [2](#T2){ref-type="table"}), did not affect the levels of the other receptor (Figure [1A](#F1){ref-type="fig"}), suggesting that the expression of each of them is independent of the other. As previously reported \[[@R8], [@R9]\] and shown in Table [2](#T2){ref-type="table"} and Figure [1B](#F1){ref-type="fig"}, silencing of either Notch1 (siNotch1) or Notch2 (siNotch2) decreased, to a similar extent, CLL cell viability compared with cells transfected with control siRNA (siCtrl). This effect was observed in 18 of 22 samples (Table [2](#T2){ref-type="table"}), suggesting that both receptors contribute to CLL cell survival in the vast majority of patients (81.8%). The four patients with Notch-independent viability (CLL4, 17, 20, 22) did not belong to any specific subgroup regarding clinical and biological characteristics (Table [1](#T1){ref-type="table"}). Decrease in cell viability induced by silencing of each receptor varied among the different CLL samples, ranging from 17.7 to 65.2% for Notch1 and from 16.1 to 51.4% for Notch2 (Table [2](#T2){ref-type="table"}). However, similar responses were observed irrespective of Binet stage, previous therapy, IgVH mutational status and ZAP70 and CD38 expression (Table [3](#T3){ref-type="table"}), suggesting that Notch targeting is effective in CLL cells despite the presence of adverse prognostic factors. Even in the three samples with NOTCH1 PEST domain mutation (CLL1, 7, 11; Table [1](#T1){ref-type="table"}), either Notch1 or Notch2 downregulation reduced CLL cell viability at levels similar to those observed in NOTCH1-unmutated samples (Table [3](#T3){ref-type="table"}). These results suggest that NOTCH1 mutation does not influence the sensitivity of CLL cells to Notch targeting, at least when it is harboured by a small fraction of leukemic cells, as indicated by the low NOTCH1 mutant allele burden detected in all three mutated samples examined (Table [1](#T1){ref-type="table"}). ###### Characteristics of CLL patients ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Patients Binet stage Previous treatment^[a](#tfn_001){ref-type="table-fn"}^ IgVH status^[b](#tfn_002){ref-type="table-fn"}^ ZAP70 expression^[c](#tfn_003){ref-type="table-fn"}^ CD38 expression^[d](#tfn_004){ref-type="table-fn"}^ NOTCH1 status(% mutant allele burden)^[e](#tfn_005){ref-type="table-fn"}^ Cytogenetic alterations^[f](#tfn_006){ref-type="table-fn"}^ ----------- ------------- -------------------------------------------------------- --------------------------------------------------- ------------------------------------------------------ ----------------------------------------------------- ----------------------------------------------------------------------------- ------------------------------------------------------------- CLL1 B no Unm \+ − Mut\ Normal (1.9) CLL2 C yes Unm \+ − Unm del 11q22-23\ del 13q14 CLL3 A no Unm \+ ND Unm ND CLL4 C yes Unm \+ \+ Unm del 11q22-23\ del 13q14 CLL5 C yes Unm \+ \+ Unm del 13q14\ del 14q32 CLL6 A no Mut \+ − Unm Normal CLL7 B yes Unm \+ \+ Mut\ del 11q22-23 (2.4) CLL8 B yes Unm \+ \+ Unm ND CLL9 C yes Unm − − Unm ND CLL10 C yes Mut − − Unm del 13q14\ del 17p13 CLL11 B yes Unm − − Mut\ del 11q22-23\ (1.3) del 13q14\ del 14q32 CLL12 C yes Unm \+ − Unm del 11q22-23\ del 13q14 CLL13 A yes Unm \+ \+ Unm Normal CLL14 C yes Unm \+ − Unm del 14q32 CLL15 B no Mut − − Unm Normal CLL16 C yes Mut \+ \+ Unm ND CLL17 A no Mut − − Unm del 13q14 CLL18 A no Mut − − Unm del 13q14 CLL19 C yes Mut \+ − Unm ND CLL20 B yes Mut − \+ Unm Normal CLL21 B yes Unm \+ − Unm del 13q14\ del 14q32 CLL22 B no Mut \+ − Unm Normal ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Mut, mutated; Unm, unmutated; ND, not determined. Treated patients had not received treatment for at least 3 months before the study. Mutated was defined as having a frequency of mutations \>2% from germline VH. Positivity refers to detection of \>20% ZAP70+/CD19+. Positivity refers to detection of \>20% CD38+/CD19+. c.7544_7545delCT in NOTCH1 exon 34. Assessed by FISH. ###### Effect of Notch1 and Notch2 silencing on CLL cell viability and Mcl-1 protein expression Patients siRNA Receptor \% reduction in Notch expression \% reduction in CLL cell viability \% reduction in Mcl-1 protein expression ---------- ---------- ---------- ---------------------------------- ------------------------------------ ------------------------------------------ ------ CLL 1 siNotch1 Notch1 55.0 67.0 32.6 39.0 siNotch2 Notch2 51.2 61.1 17.9 38.0 CLL 2 siNotch1 Notch1 76.0 53.8 19.8 none siNotch2 Notch2 30.2 38.1 16.2 none CLL 3 siNotch1 Notch1 98.0 99.3 65.2 70.0 siNotch2 Notch2 71.0 64.0 51.4 29.8 CLL 4 siNotch1 Notch1 87.0 69.5 none none siNotch2 Notch2 84.0 64.4 none none CLL 5 siNotch1 Notch1 57.0 80.1 32.7 36.9 siNotch2 Notch2 43.7 65.3 35.1 51.0 CLL 6 siNotch1 Notch1 61.5 52.8 38.1 44.4 siNotch2 Notch2 84 78 43.7 50.9 CLL 7 siNotch1 Notch1 73.2 89.1 29.7 44.2 siNotch2 Notch2 69.0 56.7 23.1 32.6 CLL 8 siNotch1 Notch1 78.2 72.1 29.3 36.0 siNotch2 Notch2 52.0 47.3 24.8 23.2 CLL 9 siNotch1 Notch1 67.0 73.0 33.8 27.7 siNotch2 Notch2 81.0 72.5 39.5 34.0 CLL 10 siNotch1 Notch1 78.0 58.0 25.0 none siNotch2 Notch2 75.0 51.0 28.4 none CLL 11 siNotch1 Notch1 48.0 59.0 20.8 22.9 siNotch2 Notch2 41.0 54.0 38.2 41.7 CLL 12 siNotch1 Notch1 51.0 65.0 31.5 none siNotch2 Notch2 35.3 40.1 17.1 none CLL 13 siNotch1 Notch1 31.0 35.2 17.7 23.0 siNotch2 Notch2 75.7 61.2 16.1 48.0 CLL 14 siNotch1 Notch1 79.0 72.0 26.1 92.0 siNotch2 Notch2 88.0 98.3 20.7 80.0 CLL 15 siNotch1 Notch1 44.4 52.0 58.4 41.9 siNotch2 Notch2 72.7 59.8 34.9 26.0 CLL 16 siNotch1 Notch1 58.0 69.0 43.9 59.4 siNotch2 Notch2 65.0 71.0 36.1 50.0 CLL 17 siNotch1 Notch1 47.1 53.0 none none siNotch2 Notch2 69.0 71.0 none none CLL 18 siNotch1 Notch1 89.0 74.5 23.9 52.9 siNotch2 Notch2 73.9 52.8 31.3 71.6 CLL 19 siNotch1 Notch1 36.0 31.3 18.4 32.0 siNotch2 Notch2 56.4 45.1 20.5 35.4 CLL 20 siNotch1 Notch1 95.3 71.0 none none siNotch2 Notch2 50.7 45.4 none none CLL 21 siNotch1 Notch1 68.3 75.3 33.6 21.0 siNotch2 Notch2 57.3 61.3 44.3 44.0 CLL 22 siNotch1 Notch1 45.6 61.0 none none siNotch2 Notch2 81.9 82.3 none none Expression of Notch1 and Notch2 and Mcl-1 was analyzed by western blot in CLL cells transfected with Notch1 (siNotch1), Notch2 (siNotch2) or control siRNA (siCtrl). The band intensities of Notch1 and Notch2 (TM and IC) and Mcl-1 were quantified by densitometric analysis and normalized to GAPDH. Values represent the percentage reduction of Notch1 and Mcl-1 expression in siNotch1 cells, and of Notch2 and Mcl-1 expression in siNotch2 cells, compared with siCtrl cells. Viable CLL cells were quantified by flow cytometry using Annexin V/PI staining, and the percentage reduction in viability of siNotch1 and siNotch2 cells was calculated relative to siCtrl. None indicates reductions in cell viability or Mcl-1 expression lower than 5%. ![Notch1 and Notch2 silencing decreases cell viability and expression of Mcl-1 protein in CLL cells\ CLL cells were transfected with control siRNA (siCtrl), Notch1 siRNA (siNotch1), Notch2 siRNA (siNotch2) or combined siNotch1 and siNotch2 (siNotch1/2) as described in "siRNA transfection" and then cultured in complete medium for 72 hours. A., C. Western blot analysis of Notch1, Notch2, Mcl-1 and Bcl-2 expression was performed on 15 μg whole-cell lysates separated on a 7.5% SDS-PAGE. The antibodies used for Notch1 and Notch2 recognized the 120-kDa transmembrane/cytoplasmic subunit (TM) and the 100-kDa active intracellular domain (IC). Protein loading was assessed by reprobing the blots with an anti-GAPDH antibody. Vertical line inserted in Notch2 blot of CLL7 indicates a repositioned gel lane. Data of CLL7 and 18 are representative of six samples. D. The blots of Mcl-1 were subjected to densitometric analysis and densitometry units (U) were calculated relative to GAPDH. Data are the mean ± SD of six samples. \*P \< 0.05, \\P \< 0.01 (each siNotch transfection condition versus siCtrl) according to Student t test. B. Cell viability was evaluated by flow cytometric analysis of Annexin V/PI (An V/PI) staining. Viability (An V^−^/PI^−^) of siCtrl cells was set to 100%. Data are the mean ± SD of six samples. \\P \< 0.01 (each siNotch transfection condition versus siCtrl) according to Student t test.](oncotarget-06-16559-g001){#F1} ###### Notch1 and Notch2 silencing decreases CLL cell viability independently of clinical characteristics and prognostic factors Binet stages and prognostic factors Number of patients \% Viability related to siCtrl (mean ± SD)^[a](#tfn_007){ref-type="table-fn"}^ ---------------------------------------------------------- -------------------- -------------------------------------------------------------------------------- ---------------- ------------- ---------------- Binet stage A 5 70.5 ± 23.7 0.778 (A vs B) 71.3 ± 20.4 0.594 (A vs B) B 8 73.8 ± 18.2 0.974 (B vs C) 76.9 ± 16.1 0.924 (B vs C) C 9 74.1 ± 11.8 0.707 (C vs A) 76.2 ± 12.2 0.580 (C vs A) Previous treatment^[b](#tfn_008){ref-type="table-fn"}^ No 7 68.1 ± 24.6 74.2 ± 20.0 Yes 15 75.5 ± 11.4 0.337 75.9 ± 13.1 0.814 IgVH status^[c](#tfn_009){ref-type="table-fn"}^ Unm 13 71.2 ± 14.1 73.5 ± 14.2 Mut 9 76.1 ± 20.0 0.505 78.0 ± 16.9 0.501 ZAP70 expression^[d](#tfn_010){ref-type="table-fn"}^ Negative 7 76.1 ± 19.2 75.0 ± 16.6 Positive 15 71.8 ± 15.6 0.579 75.4 ± 15.0 0.954 CD38 expressione^[e](#tfn_011){ref-type="table-fn"}^ Negative 14 73.8 ± 14.2 74.7 ± 14.3 Positive 7 77.4 ± 15.8 0.599 80.4 ± 14.4 0.402 NOTCH1 status^[f](#tfn_012){ref-type="table-fn"}^ Unm 19 73.3 ± 17.7 75.6 ± 15.9 Mut 3 72.3 ± 6.2 0.925 73.6 ± 10.5 0.839 Mut, mutated; Unm, unmutated. Percentage of viable cells determined by Annexin V/PI staining in siNotch1 or siNotch2 transfected CLL cells related to siCtrl cells as 100% viability. Treated patients had not received treatment for at least 3 months before the study. Mutated was defined as having a frequency of mutations \>2% from germline VH. Positivity refers to detection of \>20% ZAP70+/CD19+. Positivity refers to detection of \>20% CD38+/CD19+. c.7544_7545delCT in NOTCH1 exon 34. The evidence that Notch1 and Notch2 exert redundant effects in promoting CLL cell survival prompted us to examine the effect of combined Notch1 and Notch2 silencing. We simultaneously transfected CLL cells with Notch1 and Notch2 siRNA (siNotch1/2), and performed these experiments in six CLL samples (patients 6, 7, 9, 11, 16, 18), selected to include patients with different clinical and biological characteristics. Results showed that although the combined siNotch1/2 transfection efficiently downregulated the expression of both receptors (Figure [1A](#F1){ref-type="fig"}), it did not further decrease CLL cell viability with respect to transfection of each single receptor (Figure [1B](#F1){ref-type="fig"}). Notch1 and Notch2 silencing decreases Mcl-1 but not Bcl-2 protein expression in CLL cells {#s2_2} ----------------------------------------------------------------------------------------- Based on the evidence that Mcl-1 and Bcl-2 proteins are highly expressed in CLL cells and play a crucial role in apoptosis resistance and CLL pathogenesis \[[@R19], [@R23]\], we analyzed the effect of Notch downregulation on the expression of these proteins (n = 22). We found that in 15 of the 18 samples where Notch1 and Notch2 downregulation reduced CLL cell viability, there was a decrease in Mcl-1 levels compared with siCtrl cells, whereas in the four samples with Notch-independent viability, Mcl-1 levels remained unchanged (Table [2](#T2){ref-type="table"}, Figures [1C and 1D](#F1){ref-type="fig"}). When we transfected CLL cells with combined siNotch1/2 (n = 6), we observed that reduction of Mcl-1 levels was more pronounced (Figures [1C and 1D](#F1){ref-type="fig"}). In contrast, in all 22 patients, Bcl-2 levels were not affected by either each single or combined Notch receptor downregulation (Figure [1C](#F1){ref-type="fig"} and data not shown). Altogether, these data indicate that apoptosis of CLL cells induced by Notch silencing involves downregulation of Mcl-1 but not Bcl-2 expression. Combined Notch1 and Notch2 silencing prevents the increase in Mcl-1 levels and cell viability induced in CLL cells by IL-4 {#s2_3} -------------------------------------------------------------------------------------------------------------------------- Several micro-environmental stimuli have been shown to promote ex vivo CLL cell survival by increasing Mcl-1 expression \[[@R38], [@R39]\]. The involvement of Notch signaling in Mcl-1-mediated CLL cell survival induced by the microenvironment has never been explored. We examined whether the T-cell derived cytokine IL-4, known inducer of cell survival and Mcl-1 expression in CLL \[[@R39]\], enhanced Notch expression in promoting these effects, and if so, whether this increase was required for Mcl-1-mediated CLL cell survival induced by IL-4. CLL cells, transfected with siCtrl or combined siNotch1/2, were cultured for 72 hours with or without IL-4, and then examined for Notch1, Notch2 and Mcl-1 expression, and cell viability/apoptosis (n = 6). In agreement with previous studies \[[@R39]\], we found that in siCtrl cells, IL-4 increased both Mcl-1 levels (Figures [2A and 2B](#F2){ref-type="fig"}) and cell viability (Figure [2D](#F2){ref-type="fig"}). In siCtrl cells, IL-4 significantly upregulated also Notch1 and Notch2 expression (Figures [2A and 2C](#F2){ref-type="fig"}), and interestingly, combined Notch1/2 silencing partially abrogated the increase in both Mcl-1 expression and cell viability induced by the cytokine (Figures [2A-2D](#F2){ref-type="fig"}). These results indicate that in CLL cells, IL-4 enhances Notch expression and that this event is required to induce the increase in Mcl-1-mediated cell survival. ![Combined Notch1/2 silencing prevents the increase in Mcl-1 levels and cell viability induced by IL-4 in CLL cells\ CLL cells, transfected with control siRNA (siCtrl) or combined Notch1 and Notch2 siRNA (siNotch1/2) as described in "siRNA transfection", were cultured for 72 hours in complete medium with or without 25 ng/ml IL-4 (n = 6). A. Expression of Notch1, Notch2 and Mcl-1 was analyzed as described in Figure [1A,C](#F1){ref-type="fig"}. Vertical line inserted in Notch1 blot of CLL18 indicates a repositioned gel lane. The blots of Mcl-1 B. and those of Notch1 and Notch2 C. were subjected to densitometric analysis, and densitometry units (U) were calculated relative to GAPDH. A. Data of CLL7 and 18 are representative of six samples. B., C. Data are the mean ± SD of six samples. \*P \<0.05, \\P \< 0.01 calculated by Student t test. D. Cell viability was evaluated by flow cytometric analysis of Annexin V/PI (An V/PI) staining. Viability (An V^−^/PI^−^) of IL-4-untreated siCtrl cells was set to 100%. Data are the mean ± SD of six samples. \\P \< 0.01 calculated by Student t test.](oncotarget-06-16559-g002){#F2} Mcl-1 downregulation by Notch silencing is not due to transcriptional control or degradation by caspases but in part to degradation by proteasome {#s2_4} ------------------------------------------------------------------------------------------------------------------------------------------------- Mcl-1 protein expression is regulated at multiple levels, including transcription, translation and degradation \[[@R26]-[@R31]\]. To define whether the reduced Mcl-1 expression induced by Notch targeting was due to transcriptional inhibition, we analyzed Mcl-1 mRNA expression by real-time PCR (n = 6). In siNotch1 as well as in siNotch2 cells, Mcl-1 mRNA levels were similar to those observed in siCtrl cells (Figure [3A](#F3){ref-type="fig"}), suggesting that Notch silencing reduces Mcl-1 expression at posttranscriptional level. ![Mcl-1 downregulation by Notch silencing partially depends on proteasome degradation\ A. Mcl-1 downregulation by Notch silencing is independent of reduced transcription. CLL cells were transfected with control siRNA (siCtrl), Notch1 siRNA (siNotch1) or Notch2 siRNA (siNotch2) as described in "siRNA transfection" and then cultured in complete medium for 72 hours (n = 6). Mcl-1 mRNA levels were evaluated by real-time PCR, normalized to GAPDH and represented as fold change with respect to siCtrl cells. Data are the mean ± SD of six samples. Differences between each siNotch transfection and siCtrl were not significant. B., C. Mcl-1 downregulation by Notch silencing is independent of degradation by caspases. siCtrl, siNotch1 or siNotch2 transfected cells were cultured for 72 hours in complete medium with 50 μM pan-caspase inhibitor z-VAD-fmk or 0.005% DMSO as control (n = 6). B. PARP cleavage, indicator of caspase activity, and Mcl-1 expression were analyzed by western blot on 15 μg whole-cell lysates. Protein loading was assessed by reprobing the blots with an anti-GAPDH antibody. Data of CLL18 are representative of six samples. C. The blots of Mcl-1 were subjected to densitometric analysis, and densitometry units (U) were calculated relative to GAPDH. Data are the mean ± SD of six samples. ^\^P* \< 0.05; ns, not significant (z-VAD-fmk-treated cells versus DMSO-treated cells in each transfection condition) according to Student t test. D., E. Mcl-1 downregulation by Notch silencing partially depends on degradation by proteasome. siCtrl, siNotch1 or siNotch2 transfected cells were cultured for 72 hours in complete medium additioned, during the last 4 hours, with 2.5 μM proteasome inhibitor MG132 or 0.001% DMSO as control (n = 6). D. Accumulation of polyubiquitinated proteins, indicator of proteasome inhibition, and Mcl-1 expression were analyzed by western blot on 15 μg whole-cell lysates. Protein loading was assessed by reprobing the blots with an anti-GAPDH antibody. Data of CLL18 are representative of six samples. E. The blots of Mcl-1 were subjected to densitometric analysis, and densitometry units (U) were calculated relative to GAPDH. Data are the mean ± SD of six samples. \*P \< 0.05, \\P \< 0.01 (MG132-treated cells versus DMSO-treated cells in each transfection condition) according to Student t test.](oncotarget-06-16559-g003){#F3} It has been shown that Mcl-1 can be degraded by caspases during apoptosis \[[@R40]\]. To test whether Notch silencing decreased Mcl-1 levels through degradation by caspases, we cultured siNotch1, siNotch2 and siCtrl cells with or without the pan-caspase inhibitor z-VAD-fmk (n = 6). Results showed that whereas in siCtrl cells, z-VAD-fmk increased Mcl-1 levels, in siNotch1 as well as in siNotch2 cells treated with z-VAD-fmk, Mcl-1 levels continued to be downregulated, although PARP cleavage, an indicator of caspase activation, continued to be inhibited as in siCtrl cells (Figures [3B and 3C](#F3){ref-type="fig"}). These results suggest that reduction in Mcl-1 expression induced by Notch targeting was independent of caspase activation. Another mechanism controlling Mcl-1 protein expression is proteasomal degradation \[[@R26], [@R30], [@R31]\]. To determine whether the decrease in Mcl-1 levels induced by Notch targeting was due to proteasomal degradation, we tested the effect of the proteasome inhibitor MG132 (n = 6). It was added to siNotch1, siNotch2 and siCtrl cells during the last 4 hours of the 72-hour post transfection culture, and its action was demonstrated by the accumulation of polyubiquitinated proteins, an indicator of proteasome inhibition (Figure [3D](#F3){ref-type="fig"}). In siCtrl cells, MG132 increased Mcl-1 levels, indicating that in CLL cells cultured ex vivo, Mcl-1 is degraded by proteasome. In both siNotch1 and siNotch2 cells, MG132 prevented the loss of Mcl-1 protein which returned to levels observed in siCtrl cells (Figures [3D and 3E](#F3){ref-type="fig"}). These results suggest that Mcl-1 downregulation by Notch targeting depends, at least in part, on increased degradation by proteasome. Mcl-1 downregulation by Notch silencing is accompanied by reduced eIF4E phosphorylation {#s2_5} --------------------------------------------------------------------------------------- Mcl-1 mRNA translation is highly dependent on the activity of the eukaryotic initiation factor 4E (eIF4E) \[[@R33]\], a key translation regulator associated with tumorigenesis \[[@R35], [@R36]\]. eIF4E is indeed frequently over-expressed and over-activated in human cancers, and acts at a converging point of relevant oncogenic pathways. One pathway regulating eIF4E activity is mediated by the MAPK-interacting kinases 1 (MNK1) and MNK2, which are targets of the Ras/Raf/MAPK signaling and directly phosphorylate eIF4E at Ser209, a crucial event for its oncogenic activity \[[@R41]\]. Another pathway is mediated by the eIF4E-binding protein 1 (4E-BP1), a target of the PI3K/AKT/mTOR signaling, which, when is hypophosphorylated, prevents eIF4E activity through inhibitory interactions \[[@R42]\]. Based on these observations, we examined whether in CLL cells, Mcl-1 downregulation by Notch silencing was accompanied by effects on the expression and phosphorylation of eIF4E (Ser209), MNK1 (Thr197/202), and 4E-BP1 (Thr37/46 and Ser65). Consistent with previous studies \[[@R37]\], our results showed that all examined CLL cells (n = 6) expressed high levels of both total and phosphorylated eIF4E forms. eIF4E phosphorylation, but not expression, was reduced by either Notch1 or Notch2 targeting, and to a greater extent, by combined Notch1/2 targeting (Figures [4A and 4B](#F4){ref-type="fig"}). In contrast, expression and phosphorylation levels of MNK1 and 4E-BP1 were not affected by either each single or combined Notch receptor downregulation (Figure [4A](#F4){ref-type="fig"}). These results indicate that Notch signaling controls eIF4E activity, but in doing this, it does not involve MNK1 and 4E-BP1 regulation. These results, along with recent evidence that eIF4E contributes to CLL cell survival \[[@R37]\], also suggest that eIF4E activity is another target of the anti-apoptotic Notch signaling in CLL, in addition to Mcl-1. ![Mcl-1 downregulation by Notch silencing is associated with reduced phosphorylation of eIF4E\ CLL cells were transfected with control siRNA (siCtrl), Notch1 siRNA (siNotch1), Notch2 siRNA (siNotch2) or combined siNotch1 and siNotch2 (siNotch1/2), as described in "siRNA transfection" and then cultured in complete medium for 72 hours (n = 6). A. Expression and phosphorylation of eIF4E (Ser209), 4E-BP1 (Thr37/46 and Ser65) and MNK1 (Thr197/202) were analyzed by western blot on 15 μg whole-cell lysates, using antibodies able to detect total and phosphorylated forms. Protein loading was assessed by reprobing the blots with an anti-GAPDH antibody. Data of CLL7 and 18 are representative of six samples. B. The blots of phosphorylated eIF4E (Ser209) were subjected to densitometric analysis, and densitometry units (U) were calculated relative to total eIF4E. Data are the mean ± SD of six samples. \*P \< 0.05, \\P \< 0.01 (each siNotch transfection condition versus siCtrl) according to Student t test.](oncotarget-06-16559-g004){#F4} DISCUSSION {#s3} ========== In the present study, we have identified in the anti-apoptotic Mcl-1 protein and in the key translational regulator eIF4E, two targets of the pro-survival activity of Notch signaling in CLL cells. These findings are interesting because Mcl-1 and eIF4E are oncogenic in several malignancies \[[@R26], [@R35]\], including CLL. In this leukemia, both proteins are important mediators of cell survival \[[@R19], [@R20], [@R37]\], with Mcl-1 which is closely associated with adverse prognosis \[[@R21], [@R22]\] and chemoresistance \[[@R23], [@R24]\]. Specifically, we have shown that the increase in CLL cell apoptosis induced by either Notch1 or Notch2 genetic inhibition is accompanied by Mcl-1 protein downregulation. In contrast, the levels of Bcl-2 protein remained unaffected indicating that Bcl-2 is not a target of the anti-apoptotic signaling of Notch in CLL cells and that Mcl-1 reduction is not due to a general effect of Notch silencing. This observation along with the evidence that the decrease in Mcl-1 levels induced by Notch silencing was observed in the vast majority of CLL cells with Notch-dependent viability (83.3%), but not in CLL cells whose viability was independent of Notch, suggest that Mcl-1 is important for Notch-mediated CLL cell survival. Thus, one of the mechanisms by which Notch1 and Notch2 sustain CLL cell survival is by maintaining the constitutive high levels of Mcl-1. In CLL cells, several pro-survival signals have been implicated in promoting Mcl-1 expression, including those mediated by STAT3 \[[@R27]\], NF-kB \[[@R43]\], and Syk/AKT pathways \[[@R20], [@R44]\]. This is the first evidence that in CLL cells, Mcl-1 expression is also sustained by Notch signaling. This effect of Notch is not due to transcriptional regulation because reduction in Mcl-1 levels induced by Notch1 and Notch2 targeting is not accompanied by changes in Mcl-1 mRNA expression. Mcl-1 downregulation by Notch targeting is also independent of cleavage by caspases, but, as shown in studies of pharmacologic proteasome inhibition, it is partially due to increased Mcl-1 proteasome-mediated degradation, suggesting that Notch signaling controls this process, and contributes to stabilize Mcl-1 levels by interfering with it. Whether in CLL cells, Notch receptors act directly on Mcl-1 protein to stabilize it, as Notch1 does to stabilize XIAP protein \[[@R16]\], or indirectly, by influencing some of the pathways which regulate Mcl-1 proteasome-mediated degradation in these cells, including the Syk/PKCδ \[[@R45]\] and AKT/GSK3 \[[@R46]\] pathways, remains to be defined. Furthermore, the evidence that Mcl-1 downregulation by Notch targeting is accompanied by a decreased activity of eIF4E, an essential factor for Mcl-1 translation, also suggests that Notch signaling may control Mcl-1 expression by regulating its biosynthesis. Further studies will be needed to define this point and to clarify the mechanisms whereby Notch signaling sustains eIF4E activity in CLL cells, given that Notch targeting does not have any effect on the phosphorylation levels of 4E-BP1 and MNK1, which in other cell types, are two key upstream regulators of eIF4E activity \[[@R41], [@R42]\]. The impact of Notch signaling on Mcl-1 expression in CLL cells is also supported by the evidence that the combined Notch1/2 downregulation induces a higher reduction of Mcl-1 levels than that induced by downregulating each single receptor. However, these results seem discordant with the observation that combined Notch1/2 silencing does not enhance CLL cell apoptosis induced by downregulating each single receptor. A possible explanation is that the anti-leukemic activity of a strong Notch downregulation is limited by compensatory survival mechanisms, suggesting that inhibition of Notch signaling alone is not sufficient to kill all leukemic cells. This is consistent with the evidence that in several malignancies, the best clinical activity of Notch-targeted therapies was observed when the specific Notch inhibitors, including γ-secretase inhibitors (GSI) or monoclonal antibodies to Notch receptors or Notch ligands, were administered in combination with conventional chemotherapy or other targeted agents \[[@R47]\]. In line with this evidence, even in CLL cells, it has been recently demonstrated that the clinically relevant GSI PF-03084014 improves the pro-apoptotic effect of fludarabine \[[@R48]\]. This occurs because Notch inhibition overcomes the resistance mechanisms activated in CLL cells by fludarabine \[[@R23]\], including increased NF-kB activation \[[@R48]\] and Mcl-1 expression \[[@R21]\]. In this context, a critical role of Notch in regulating NF-kB pathway has been previously described in T-cell leukemia \[[@R15], [@R49]\] and a role of Notch in controlling Mcl-1 expression is here demonstrated in CLL. Another important aspect of this study is that Notch signaling also contributes to Mcl-1 accumulation induced by survival micro-environmental stimuli. Specifically, we demonstrate that IL-4, known to induce in CLL cells an increased Mcl-1-mediated cell survival \[[@R39]\], also enhances Notch1 and Notch2 activation. Interestingly, combined Notch1/2 downregulation partially prevents the increase in both CLL cell survival and Mcl-1 expression, suggesting that Notch targeting, in addition to reducing the constitutive Mcl-1 levels, is also able to prevent Mcl-1 accumulation induced by micro-environmental stimuli. This effect of Notch downregulation is important because in CLL, the major resistance mechanisms to current chemotherapy, including Mcl-1 expression, are highly favored by the microenvironment \[[@R38]\]. Overall, Notch signaling sustains CLL cell survival by promoting Mcl-1 expression and eIF4E activity. These findings along with the evidence that both Mcl-1 and eIF4E contribute to survival and chemotherapy resistance of CLL cells highlight the importance to target Notch signaling for CLL treatment, especially in combination with agents whose poor efficacy is mainly due to the elevated Mcl-1 expression and eIF4E activity detected in these leukemic cells. MATERIALS AND METHODS {#s4} ===================== Patients {#s4_1} -------- Twenty-two CLL patients entered this study. Diagnoses of CLL were based on Stanford criteria defined by the National Cancer Institute-sponsored Working Group \[[@R50]\], and clinical staging was based on the Binet classification \[[@R51]\]. This study was approved by the local Ethics Committee, and all patients signed informed consent in accordance with the Declaration of Helsinki. CLL cell isolation {#s4_2} ------------------ Peripheral blood mononuclear cells were isolated from heparinized blood of CLL patients by Ficoll density-gradient centrifugation (Nycomed, Oslo, Norway). Monocytes were removed by plastic adherence, and T cells by sheep erythrocyte rosetting. All CLL samples contained more than 96% CD19^+^/CD5^+^ CLL cells, as assessed by flow cytometry (EPICS-XL-MCL; Beckman Coulter, Fullerton, CA). CLL clinical laboratory characteristics {#s4_3} --------------------------------------- IgV~H~ mutations, CD38 surface and ZAP70 intracellular expression were analyzed as previously reported \[[@R8]\]. Cytogenetic abnormalities were examined by fluorescent in situ hybridization (FISH) using probes for chromosomes 11, 12, 13, 14 and 17. NOTCH1 exon 34 mutations were analyzed as previously reported and the percentage of mutant allele burden was determined using a semi-quantitative assay on the basis of Genescan analysis \[[@R4]\]. Table [1](#T1){ref-type="table"} gives clinical and biological characteristics of CLL patients. siRNA transfection {#s4_4} ------------------ CLL cells (12 × 10^6^) were resuspended in 100 μl Cell Line Solution Kit V (Lonza Group Ltd, Basel, Switzerland) with ON-TARGETplus SMARTpool small interfering RNA (siRNA) to Notch1 (0.5 μM), Notch2 (0.5 μM) or ON-TARGETplus siCONTROL nontargeting pool as negative control (all from Dharmacon RNA Technologies, Lafayette, CO). In the case of combined Notch1 and Notch2 silencing, 0.25 μM per siRNA was added. Cells were then transfected with the Amaxa Nucleofector II device (program U-013) and cultured for 72 hours in 12-well plates in complete medium consisting of RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum (Hyclone Laboratories, Logan, UT), 2 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin (all from Invitrogen, Milan, Italy). In some experiments, transfected cells were incubated with 25 ng/ml recombinant human IL-4 (Immunotools, Friesoyte, Germany), 50 μM pan-caspase inhibitor z-VAD-fmk or 2.5 μM proteasome inhibitor MG132 (both from Calbiochem, La Jolla, CA). The inhibitors z-VAD-fmk and MG132 were dissolved in DMSO and diluted in complete medium at the used concentrations. DMSO concentrations, which did not exceed 0.005%, did not affect CLL cell responses. IL-4 and z-VAD-fmk were added at the beginning of the 72-hour post transfection culture and maintained until cell collection, whereas MG132 was added during the last 4-hour culture. Analysis of cell viability and apoptosis {#s4_5} ---------------------------------------- Cell viability and apoptosis were assessed by flow cytometric analysis (EPICS-XL-MCL) after Annexin V-fluorescein isothiocyanate/propidium iodide staining, performed using a commercial kit (Immunotech, Beckman Coulter) according to the manufacturer\'s instructions. Western blot analysis {#s4_6} --------------------- Whole-cell lysates were extracted in RIPA buffer. Equal amounts of proteins were separated by 7.5% SDS-PAGE and transferred to nitrocellulose membranes, which, after blocking, were incubated with primary antibodies to: Notch1 (clone bTAN20) and Notch2 (clone C651.6DbHN), developed by Spyros Artavanis-Tsakonas, obtained from DSHB developed under the auspices of the NICHD, and maintained by Iowa University; Mcl-1 (Santa Cruz Biotechnology, Santa Cruz, CA); Bcl-2 (DakoCytomation, Milan, Italy); phospho-eIF4E (Ser209), total eIF4E, phospho-4E-BP1 (Ser65), phospho-4E-BP1 (Thr37/46), total 4E-BP1, phospho-MNK1 (Thr197/202), total MNK1 and PARP (Cell Signaling Technology, Beverly, MA); ubiquitinated proteins (BIOMOL Research Laboratories, Plymouth Meeting, PA) and GAPDH (Sigma-Aldrich, St. Louis, MO). Signals were detected using appropriate horseradish peroxidase-conjugated secondary antibodies and the ECL system (GE Healthcare, Milan, Italy). Densitometric analysis was performed using Quantity One software (Bio-Rad, Milan, Italy). Real-time quantitative PCR {#s4_7} -------------------------- Total RNA was isolated using Trizol (Invitrogen) and 1 μg reverse-transcribed using RT-kit plus (Nanogen Advanced Diagnostics, Milan, Italy). Real-time quantitative PCR was performed with PCR Master Mix Power SYBR Green (Applied Biosystems, Warrington, UK), using the 7900HT Fast Real-Time PCR System (Applied Biosystems). The primer sequences used for MCL-1 were forward: 5′-GAG ACC TTA CGA CGG GTT-3′ and reverse: 3′-TTT GAT GTC CAG TTT CCG-3′ (Invitrogen). Relative fold change was normalized to GAPDH and calculated using the 2^−ΔΔCt^ method. Statistical analysis {#s4_8} -------------------- Statistical differences between mean values were evaluated using the Student t test. The minimal level of significance was P \< 0.05. This work was supported by grants from Progetti di Ricerca di Interesse Nazionale (PRIN 2010-2011; 2010MCLPLB_005 to E.R.), and the Associazione Italiana per la Ricerca sul Cancro (IG13314 to I.S.). CONTRIBUTION F.D.F. conceived and performed experiments and discussed data; R.S., B.D.P. and S.B. performed experiments and interpreted the data; F.F., M.D.I. and P.S. provided patient samples, patient clinical and laboratory data, and critical suggestions; I.S. and P.M. supervised the study and provided critical suggestions; E.R. planned every aspect of the proposal, conceived and supervised the research, analyzed and interpreted data, and wrote the manuscript. All authors revised the manuscript and approved the final version. CONFLICTS OF INTERESTS The authors declare no conflict of interest.	{ "pile_set_name": "PubMed Central" }
Murray SR, Stock SJ, Cowan S, Cooper ES, Norman JE. Spontaneous preterm birth prevention in multiple pregnancy. The Obstetrician & Gynaecologist 2018;20;57--63. DOI: [10.1111/tog.12460](10.1111/tog.12460). The copyright line for this article was changed on 10 May 2018 after original online publication. Introduction {#tog12460-sec-0003} ============ Twin pregnancies are high‐risk and associated with increased perinatal morbidity and mortality.[1](#tog12460-bib-0001){ref-type="ref"} Multiple pregnancy is associated with adverse maternal outcomes including increased rates of pre‐eclampsia, pregnancy‐induced hypertension, maternal anaemia and venous thromboembolism.[1](#tog12460-bib-0001){ref-type="ref"} Compared to singleton newborns, newborn infants of twin pregnancies are also at increased risk of adverse outcomes including congenital anomalies, cerebral palsy, intrauterine growth restriction and stillbirth. Although only 3% of all live births are twin pregnancies,[1](#tog12460-bib-0001){ref-type="ref"} twin babies account for up to 15% of special care unit admissions.[2](#tog12460-bib-0002){ref-type="ref"} A 2009 survey conducted by the UK charity the Twin and Multiple Birth Association (TAMBA) found that in 44% of all twins born to the 1298 mothers interviewed, at least one baby entered special care.[3](#tog12460-bib-0003){ref-type="ref"} According to the 2014 MBRRACE‐UK (Mothers and Babies: Reducing Risk through Audits and Confidential Enquiries across the UK) report,[4](#tog12460-bib-0004){ref-type="ref"} the perinatal mortality rate in twins was three times higher than that of singleton pregnancies.[4](#tog12460-bib-0004){ref-type="ref"} Much of this perinatal mortality is driven by prematurity. In Scotland, 50% of twins are delivered preterm (at fewer than 37 weeks of gestation), with around 20% delivering before 34 weeks of gestation.[5](#tog12460-bib-0005){ref-type="ref"} Figures from the USA are similar: a 12‐fold higher preterm birth (PTB) rate of 56.6% was found in twins compared to 9.7% in singletons (odds ratio \[OR\] 12.8, 95% confidence interval \[CI\] 12.6--12.9).[6](#tog12460-bib-0006){ref-type="ref"} Despite attempts by the UK\'s National Institute for Health and Care Excellence (NICE) to reduce multiple pregnancy with the universal use of single embryo transfer (SET) during assisted reproduction technologies (ARTs),[7](#tog12460-bib-0007){ref-type="ref"} the UK twin pregnancy rate remains high among women undergoing ART, with up to 24% of successful in vitro fertilisation (IVF) procedures resulting in multiple pregnancy.[1](#tog12460-bib-0001){ref-type="ref"} As well as the increased risk of perinatal morbidity and mortality, the economic cost of healthcare provision for a twin baby is significantly higher than for singletons, and is quoted to be twice as high in the first 5 years.[8](#tog12460-bib-0008){ref-type="ref"} The James Lind Alliance (JLA) Preterm Birth Priority Setting Partnership (PSP) has prioritised the following question as the number one uncertainty in preterm birth: which interventions are most effective in predicting or preventing PTB? The JLA brings together clinicians, patients and carers to set research priorities in several aspects of obstetrics and gynaecology, including stillbirth and endometriosis.[9](#tog12460-bib-0009){ref-type="ref"} Aetiology {#tog12460-sec-0004} ========= The aetiology of spontaneous PTB in multiple pregnancies is likely to be multifactorial, different from singletons, and remains largely unknown. Different proposed pathophysiology mechanisms include intrauterine infection, cervical insufficiency and increased uterine stretch/distension. There is also increased secretion of mediators such as corticotrophin‐releasing hormone (CRH) from the larger placental mass, and factors produced by the maturing fetal lung such as surfactant protein‐A, which stimulates myometrial contractility and may contribute to preterm parturition.[10](#tog12460-bib-0010){ref-type="ref"} Twins are associated with a higher risk of obstetric intervention and therefore iatrogenic PTB is higher. Approximately one‐third of all premature deliveries in multiple pregnancies are medically indicated.[11](#tog12460-bib-0011){ref-type="ref"} This review focuses on the prevention of spontaneous PTB in twin pregnancies. Prediction of preterm labour in multiple pregnancy {#tog12460-sec-0005} ================================================== Detailed descriptions of the methods of prediction of PTB in multiple pregnancy are out of the scope of this review, but the evidence surrounding the two most researched methods, cervical length and fetal fibronectin, is summarised below. Cervical length measurement {#tog12460-sec-0006} --------------------------- ### In asymptomatic women with a twin pregnancy {#tog12460-sec-0007} A 2010 meta‐analysis of 16 cohort/cross‐sectional studies (n = 3213)[12](#tog12460-bib-0012){ref-type="ref"} showed that in asymptomatic women with twins, a cervical length of \<25 mm was associated with a 25% risk of delivery before 28 weeks of gestation. This review also found a cervical length of \<20 mm at 20--24 weeks of gestation to be associated with a 42.4% risk of birth before 32 weeks, and a 62% risk of birth before 34 weeks. A subsequent systematic review[13](#tog12460-bib-0013){ref-type="ref"} upheld these findings and concluded that in asymptomatic women with a twin pregnancy, a cervical length measurement at 20--24 weeks of gestation was a good predictor of spontaneous PTB. A recent systematic review of 1024 women with twins[14](#tog12460-bib-0014){ref-type="ref"} assessed repeated measures of a change in cervical length as a predictor of PTB, and found that the shortening of cervical length over time had a low predictive accuracy for preterm birth at fewer than 34 weeks of gestation. The largest and most up‐to‐date individual patient data (IPD) meta‐analysis on the effect of gestational age and cervical length measurements in the prediction of PTB in twin pregnancies was published by Kindinger et al.[15](#tog12460-bib-0015){ref-type="ref"} in 2015 (n = 4409 twin pregnancies). The benefit of an IPD meta‐analysis is that it is a more robust method of combining and easily comparing studies, easily allowing subgroup analysis; for example, a short cervix group. This analysis consisted of 12 twin cohorts and found that when cervical length was \<30 mm at 18 weeks of gestation, it was most predictive of birth at ≤28 weeks. Prediction of later spontaneous PTB (28--34 weeks) improved with cervical length measurements taken at later gestations (≥22 weeks). The authors concluded by recommending cervical length screening in twin pregnancies as a predictor of spontaneous PTB from 18 weeks of gestation. In summary, in asymptomatic women with a twin pregnancy, current evidence supports the use of measuring cervical length from 18 weeks of gestation as a predictor of spontaneous PTB, but does not support repeated measures of cervical length. ### In symptomatic women with a twin pregnancy {#tog12460-sec-0008} In a meta‐analysis of five combined cohort/cross‐sectional studies (n = 310), Conde‐Agudelo et al.[12](#tog12460-bib-0012){ref-type="ref"} showed that, in women with twin pregnancies and symptoms of spontaneous PTB, cervical length measurement had a low predictive accuracy for PTB at \<34 weeks of gestation. However, the review concluded that the sample size was small and there is a paucity of evidence in the area of predicting PTB in twin pregnancies. Fetal fibronectin {#tog12460-sec-0009} ----------------- ### In asymptomatic women with a twin pregnancy {#tog12460-sec-0010} NICE does not recommend the use of fetal fibronectin (fFN) in multiple pregnancies. A 2010 meta‐analysis[16](#tog12460-bib-0016){ref-type="ref"} summarising 11 studies of fFN use in asymptomatic twin pregnancy to predict PTB suggested only limited prediction accuracy, and better negative predictive rates than positive (6% risk of PTB before 34 weeks of gestation with a negative test compared to a 33% risk of PTB before 34 weeks with a positive test). ### In symptomatic women with a twin pregnancy {#tog12460-sec-0011} The meta‐analysis by Conde‐Agudelo et al.[16](#tog12460-bib-0016){ref-type="ref"} synthesised the results of five studies using fFN to predict PTB in symptomatic women with a twin pregnancy. The authors concluded that fFN testing in multiple pregnancy was most accurate in women with symptoms of PTB (positive and negative likelihood ratios 85% and 75%, respectively, within 7 days of testing). However, although fFN testing is recommended for use in singleton pregnancies with symptoms of PTB if cervical length measurement is not available,[17](#tog12460-bib-0017){ref-type="ref"} it is not recommended in the NICE guideline for multiple pregnancy. A combination of cervical length measurements and fFN may be a more accurate predictor of PTB. One small study[18](#tog12460-bib-0018){ref-type="ref"} (n = 155 twin pregnancies) reported that if fFN was positive and cervical length was \<20 mm, then 54.4% of twins would deliver before 34 weeks of gestation, and this was significantly higher than the overall rate of PTB. Methods of preterm birth prevention {#tog12460-sec-0012} =================================== Cervical cerclage {#tog12460-sec-0013} ----------------- Cervical cerclage is a surgical technique to prevent PTB and has been described in the UK literature since 1902.[19](#tog12460-bib-0019){ref-type="ref"} Although cerclage is endorsed by NICE for use in singleton pregnancies, it is not recommended for use in multiple pregnancy as its effectiveness remains controversial.[11](#tog12460-bib-0011){ref-type="ref"} A subgroup of twins (n = 49 women) in a 2005 meta‐analysis[20](#tog12460-bib-0020){ref-type="ref"} investigating the use of cerclage showed an increase in prematurity and a trend towards harm. In multiple pregnancies, cerclage was associated with an increased risk of premature delivery (relative risk \[RR\] 2.15, 95% CI 1.15--4.01) and a trend towards an increased risk of perinatal mortality. However, this finding was not statistically significant, with wide confidence intervals likely reflecting the small sample size (RR 2.66, 95% CI 0.83--8.54). A subsequent Cochrane review published in 2014[21](#tog12460-bib-0021){ref-type="ref"} examined five trials, of which two (n = 73 women) assessed history‐indicated cerclage and three (n = 55 women) assessed ultrasound‐indicated cerclage. This review found no benefit of cervical cerclage in reducing preterm delivery in twin pregnancies at fewer than 34 weeks of gestation (RR 1.16, 95% CI 0.44--3.36, four trials, n = 98 women). It also showed a trend towards harm with an increased risk of perinatal death, although this was not statistically significant (RR 1.74, 95% CI 0.92--3.28). There was no reduction in a composite of adverse neonatal outcome (RR 1.54, 95% CI 0.58--4.11). In the pre‐specified subgroup, ultrasound‐indicated cerclage was associated with an increased risk of low birthweight (RR 1.39, 95% CI 1.06--1.83, three trials, n = 98 women) and respiratory distress syndrome (RR 5.07, 95% CI 1.75--14.70, three trials, n = 98 women). The authors concluded that there was no evidence of the usefulness of cerclage in reducing the risk of PTB in twins, but that more research was needed because of a small number of trials, which each had a small number of patients. It is also important to note that no trials were identified that reported long‐term infant neurodevelopmental outcomes following cervical cerclage. ### Ultrasound‐indicated cerclage in twin pregnancies {#tog12460-sec-0014} Following the Cochrane review, a further IPD meta‐analysis was published looking specifically at ultrasound‐indicated cerclage (cervical length ≤25 mm before 24 weeks of gestation) in twin pregnancies.[22](#tog12460-bib-0022){ref-type="ref"} Three randomised controlled trials (RCTs; n = 49 women) were identified. No significant differences in PTB rates before 34 weeks of gestation were found between cerclage and no cerclage in twin pregnancies with a trend towards harm (RR 2.19, 95% CI 0.72--6.63). Similar to the Cochrane review, the authors concluded that further large trials were necessary to determine the effectiveness of cerclage in twin pregnancies. The latest retrospective cohort study was published in 2015 and involved 140 women with twin pregnancies.[23](#tog12460-bib-0023){ref-type="ref"} This showed that ultrasound‐indicated cerclage with a cervical length of ≤25 mm did not reduce the risk of preterm delivery before 34 weeks of gestation (adjusted OR \[aOR\] 0.37, 95% CI 0.16--1.1) nor was it associated with increased neonatal morbidity (aOR 0.7, 95% CI 0.4--1.2). In the pre‐specified subgroup analysis of women with a cervical length of ≤15 mm (n = 32 women), there was a significant reduction in the rate of PTB before 34 weeks of gestation (OR 0.42, 95% CI 0.24--0.81). In summary, unlike singleton pregnancies, there appears to be no benefit of cervical cerclage in reducing PTB rates in multiple pregnancies. Although the study by Roman et al.[22](#tog12460-bib-0022){ref-type="ref"} reported a benefit in those with a short cervix (≤15 mm), the numbers in the study were small and further large trials are needed to address this question adequately. A search on <https://clinicaltrials.gov/> identified two currently recruiting studies comparing ultrasound and emergency cerclage in twin pregnancies with expectant management.[24](#tog12460-bib-0024){ref-type="ref"}, [25](#tog12460-bib-0025){ref-type="ref"} One trial[24](#tog12460-bib-0024){ref-type="ref"} is a multicentre, international RCT of physical exam‐indicated cerclage in twin gestations with a primary outcome measure of preterm delivery at fewer than 34 weeks of gestation. The other study[25](#tog12460-bib-0025){ref-type="ref"} is a single‐centre RCT of cervical cerclage versus expectant management for women with a twin pregnancy and a short cervix (≤25 mm), with a primary outcome of pregnancy prolongation. The 'C‐STICH' RCT[26](#tog12460-bib-0026){ref-type="ref"} of monofilament versus braided sutures for insufficient cervix is currently recruiting in multiple centres in the UK. Although this study only looks at singleton pregnancies, a similar trial in twin pregnancies would be appropriate after the results are published. Progesterone {#tog12460-sec-0015} ------------ ### Is it biologically plausible? {#tog12460-sec-0016} The use of progesterone in the prevention of PTB in both singleton and multiple pregnancies has been extensively investigated. Arguably, the use of progesterone is biologically plausible given that uterine quiescence is maintained throughout pregnancy by progesterone and progesterone receptor‐mediated inhibition of inflammation, which causes suppression of the contractile genes.[27](#tog12460-bib-0027){ref-type="ref"} Labour is thought to occur as a result of a functional withdrawal of progesterone.[28](#tog12460-bib-0028){ref-type="ref"} Anti‐progesterones such as mifepristone are used to stimulate abortion and induce labour. However, others have argued that progesterone levels are high during pregnancy, and progesterone receptors are fully occupied, hence the therapeutic benefit of adding further progesterone is unclear. Progesterone is available as an intramuscular injection of 17α‐hydroxylase caproate (only licensed in the USA), or vaginal progesterone (the only available progesterone product in the UK, but not licensed for prevention of PTB in the USA or Europe). ### Use of progesterone in unselected twin pregnancies {#tog12460-sec-0017} A 2012 IPD meta‐analysis by Romero et al.,[27](#tog12460-bib-0027){ref-type="ref"} which investigated the use of vaginal progesterone in singletons, also included a subgroup of women with twins (n = 52 women). This revealed no reduction in PTB before 34 weeks of gestation when vaginal progesterone was used in twin pregnancies (RR 0.7, 95% CI 0.34--1.44). A subsequent IPD meta‐analysis by Schuit et al. (2015)[29](#tog12460-bib-0029){ref-type="ref"} upheld these main conclusions. The review included 13 trials (3768 pregnancies) of the use of progestogens in unselected twin pregnancies and the primary outcome was neonatal morbidity. Treatment with vaginal progesterone did not reduce the risk of adverse perinatal outcome (RR 0.97, 95% CI 0.77--1.2), and there were no significant differences for delivery before 32 weeks of gestation between the progesterone and the control groups (RR 0.91, 95% CI 0.68--1.2).[29](#tog12460-bib-0029){ref-type="ref"} A recent systematic review and meta‐analysis comparing progesterone, cerclage and the cervical pessary for prevention of PTB in unselected twin pregnancies found no reduction in PTB rates with any of the interventions.[30](#tog12460-bib-0030){ref-type="ref"} Some secondary outcomes were reduced with vaginal progesterone only (very low birth weight, \[RR 0.71, 95% CI 0.52--0.98\], and need for mechanical ventilation, \[RR 0.61, 95% CI 0.45--0.82\]). One of the larger RCTs in the Schuit et al.[29](#tog12460-bib-0029){ref-type="ref"} IPD meta‐analysis, the 'STOPPIT' study (n = 500 women),[31](#tog12460-bib-0031){ref-type="ref"} performed a baby follow‐up study of the effect of prophylactic progesterone in twin pregnancy on childhood outcome.[32](#tog12460-bib-0032){ref-type="ref"} Performed through record linkage of childhood records, this study found no increased incidence of perinatal death (15 twins in the progesterone group versus 11 in the placebo group), congenital anomalies (OR 1.04, 95% CI 0.49--1.21) or hospitalisation (OR 0.97, 95% CI 0.71--1.33) of those exposed to progesterone versus placebo. The authors concluded there was no evidence of a detrimental or beneficial impact on health and developmental outcomes at 3 and 6 years of exposure to progesterone in utero. ### Use of progesterone in twin pregnancies in women with a short cervix {#tog12460-sec-0018} Debate still exists as to whether progesterone is effective in singleton pregnancies of women with a short cervix. The Schuit et al.[29](#tog12460-bib-0029){ref-type="ref"} IPD meta‐analysis published in 2015 included a subgroup analysis of women with a twin pregnancy and a short cervix of ≤25 mm (n = 116 women). Among this subgroup, vaginal progesterone provided a protective effect over the control with regards to adverse perinatal outcome (RR 0.57, 95% CI 0.47--0.70). However, the number of women in the subgroup was small, thus limited conclusions can be inferred. The primary outcome in this study was neonatal morbidity, therefore differences in PTB in the subgroup of women with a short cervix ≤25 mm were not investigated in this review. In summary, there is no benefit of universal vaginal progesterone to reduce PTB rates in multiple pregnancies. One meta‐analysis showed a benefit in adverse perinatal outcome in a subgroup of women with a short cervix ≤25 mm, suggesting it may be useful in this group, but the numbers in the study were small and further research is needed. There appears to be no long‐term harm caused to infants exposed to progesterone in utero.[32](#tog12460-bib-0032){ref-type="ref"} The NICE guidelines for multiple pregnancy currently followed by UK practitioners do not promote the routine use of cervical cerclage or progesterone for the prevention of PTB in multiple pregnancies (see Box [1](#tog12460-fea-0001){ref-type="boxed-text"}). ###### Preterm birth prevention in multiple pregnancy from the National Institute for Health and Care Excellence guideline 2011[1](#tog12460-bib-0001){ref-type="ref"} #### {#tog12460-sec-0019} Do not use the following interventions (alone or in combination) routinely to prevent spontaneous preterm birth in twin or triplet pregnancies: Bed rest at home or in hospitalIntramuscular or vaginal progesteroneCervical cerclageOral tocolytics Cervical pessary {#tog12460-sec-0020} ---------------- The Arabin pessary is a cervical pessary used to prevent PTB. It is a flexible silicon ring with a smaller inner diameter that encompasses the cervix, aiming to tilt it posteriorly and provide cervical support.[33](#tog12460-bib-0033){ref-type="ref"} It is usually inserted at around 18--22 weeks of gestation, and in twin pregnancies is removed before 36 weeks. The Arabin cervical pessary and its correct positioning are shown in Figure [1](#tog12460-fig-0001){ref-type="fig"}. Available evidence surrounding the use of the cervical pessary in twin pregnancies is conflicting and is discussed here in terms of unselected twin pregnancies and twins in women with a short cervix. ![a) The Arabin cervical pessary; b) the position of the Arabin cervical pessary.](TOG-20-57-g001){#tog12460-fig-0001} ### Use of the cervical pessary in unselected twin pregnancies {#tog12460-sec-0021} Two RCTs looking at the use of the cervical pessary in unselected twin pregnancies have been performed. The largest, by Nicolaides et al.[34](#tog12460-bib-0034){ref-type="ref"}, which involved 1180 women who received a cervical pessary versus expectant management, found no difference in the rate of PTB before 34 weeks of gestation (RR 1.05, 95% CI 0.79--1.41). The ProTWIN study[35](#tog12460-bib-0035){ref-type="ref"} was an RCT of 808 unselected twin pregnancies. This study found no difference in PTB before 32 weeks of gestation (10% versus 12%, RR 0.86, 95% CI 0.65--1.15) and no reduction in adverse perinatal outcome (RR 0.98, 95% CI 0.69--1.39). ### Use of the cervical pessary in twin pregnancies with a short cervix {#tog12460-sec-0022} The PECEP‐Twins trial,[36](#tog12460-bib-0036){ref-type="ref"} published in 2016, is the only RCT to focus on twins in women with a short cervix (defined as ≤25 mm). In this study, 137 women with a short cervix were randomly assigned to cervical pessary or expectant management groups. The rate of PTB before 34 weeks of gestation was reduced in the pessary group compared with the expectant management group (RR 0.41, 95% CI 0.22--0.76). In a post hoc subgroup analysis of the study by Nicolaides et al.,[34](#tog12460-bib-0034){ref-type="ref"} the incidence of PTB before 34 weeks of gestation among 106 women with a cervical length of \<25 mm was not significantly different between the pessary group and the expectant management group (31% versus 26%, RR 1.2, 95% CI 0.8--1.8). In contrast, subgroup analysis of the ProTWIN study,[35](#tog12460-bib-0035){ref-type="ref"} which used a cervical length cut‐off of \<38 mm, reported a reduction in PTB before 32 weeks of gestation (16.2% versus 39.4%; RR 0.41, 95% CI 0.22--0.76). In a systematic review and meta‐analysis, Saccone et al.[37](#tog12460-bib-0037){ref-type="ref"} combined the results of the three trials of cervical pessary use for the prevention of PTB of twins of women with a short cervical length, and a forest plot was produced (Figure [2](#tog12460-fig-0002){ref-type="fig"}). It is important to note here that because the Nicolaides et al. and ProTWIN studies used different cervical length cut‐offs (\<25 mm and \<38 mm, respectively), they were not directly comparable in the meta‐analysis and therefore were not powered to detect a difference in PTB in these subgroups. The review concluded that use of the Arabin pessary in twin pregnancies of women with a short cervix may not prevent PTB or improve perinatal outcome. ![Risk of preterm birth \<34 weeks of gestation in twin pregnancies with a cervical length \<25 mm.[37](#tog12460-bib-0037){ref-type="ref"}](TOG-20-57-g002){#tog12460-fig-0002} In summary, evidence for the use of the cervical pessary for the prevention of PTB in twins is conflicting, though there is some evidence to suggest it may be useful in twin pregnancies of women with a short cervix. The STOPPIT 2 RCT[38](#tog12460-bib-0038){ref-type="ref"} is currently recruiting in the UK to help address this paucity of evidence. This RCT aims to resolve the uncertainty surrounding whether or not the Arabin pessary reduces spontaneous PTB in twins of women with a short cervix (\<30^th^ centile, which equates to 35 mm[15](#tog12460-bib-0015){ref-type="ref"}). The cervical pessary is not currently routinely used in clinical practice outside of the research setting. Other methods of preterm birth prevention {#tog12460-sec-0023} ----------------------------------------- As stated in Box 1, the following methods of PTB prevention in multiple pregnancies are not endorsed by NICE. The evidence is summarised below: ### Bed rest and uterine monitoring {#tog12460-sec-0024} In the past, bed rest was used as a method of PTB prevention for both singleton and twin pregnancies. A 2010 Cochrane review[39](#tog12460-bib-0039){ref-type="ref"} comparing hospitalisation and bed rest against expectant management summarised the results of six trials (n = 600 women). It found no benefit of this intervention, but did find an increased risk of PTB before 34 weeks of gestation (OR 1.84, 95% CI 1.01--3.34). Similarly, a 1995 meta‐analysis[40](#tog12460-bib-0040){ref-type="ref"} of six trials on the use of home uterine monitoring showed no reduction in PTB. Thus, neither of these interventions have been endorsed by NICE. ### Prophylactic tocolytics {#tog12460-sec-0025} For singleton pregnancies, the RCOG\'s Green‐top Guideline[41](#tog12460-bib-0041){ref-type="ref"} (GTG) supports the use of tocolysis for the completion of corticosteroids or an in utero transfer, but not for PTB prevention. Similarly, for multiple pregnancy, the GTG states that there is insufficient evidence for its use in the prevention of preterm labour. A 2005 Cochrane review[42](#tog12460-bib-0042){ref-type="ref"} of five trials (n = 344 women) found no reduction in PTB before 34 weeks of gestation with the use of tocolytics (RR 0.47, 95% CI 0.15--1.50). In summary, the use of prophylactic tocolytics to prevent preterm labour in multiple pregnancy is not recommended. Conclusion {#tog12460-sec-0026} ========== There is a lack of effective, evidence‐based interventions for the prevention of PTB in twin pregnancies. There is limited evidence for the use of vaginal progesterone and cervical cerclage, and the cervical pessary is currently only used within a research setting. There are no reported trials comparing the effectiveness of each of these interventions against each other, whether in isolation or in combination.[43](#tog12460-bib-0043){ref-type="ref"} Likewise, although fFN and cervical length scanning may be beneficial in predicting PTB in twins either alone or in combination, no reliable evidence supports the use of any one predictor. Research is needed to further evaluate the benefit of the cervical pessary and the use of cervical cerclage in twins of women with a short cervix. A recent article by Stock et al.[43](#tog12460-bib-0043){ref-type="ref"} concludes by advising clinicians to share with women the uncertainty of methods to prevent PTB in multiple pregnancy, and offer the opportunity to participate in clinical trials. Disclosure of interests {#tog12460-sec-0027} ----------------------- JEN has research grants from government and charitable organisations (Tommy\'s, the baby charity, the Medical Research Council and National Institute of Health Research \[NIHR\]) for preterm birth prevention (including in twins). The University of Edinburgh receives also funding for her contribution to a data monitoring committee for a preterm birth study run by GlaxoSmithKline. SRM, SJS, SC and ESC have no conflicts of interest. Contributions to authorship {#tog12460-sec-0028} --------------------------- SRM and JEN instigated the article. SRM researched and wrote the article. SJS, ESC, JEN and SC edited the article. All authors approved the final version.	{ "pile_set_name": "PubMed Central" }
Introduction {#sec1-1} ============ Dental caries is still a major concern in public health management.\[[@ref1]\] Establishment of a productive dental health philosophy, especially for the low socioeconomic levels of the society, where the lack of development endangers the dental health, even more, seems to be necessary.\[[@ref2]\] Microbial nature of caries and the contemporary concept of treatment based on the medical model necessitate the consideration of chemical control plans along with the traditional mechanical means of caries control.\[[@ref3]\] Innovations in dental material science are come to the point of developing smart materials.\[[@ref4]\] glass-ionomer cement (GIC), a tooth-colored acid-base material, with the capability of fluoride release in an aqueous environment is the first of this category.\[[@ref3][@ref4]\] It can be used as base, liner, or direct restorative material and is also considered as the material of choice in atraumatic restorative technique (ART). The ART is a minimal intervention approach, particularly beneficial for pediatric and elderly patients as well as those with dental anxiety or learning difficulties.\[[@ref3]\] While the bulk of evidence supports the protective nature of fluoride in public water and oral health products; however, the data available does not endorse the anti-caries ability of GIC or any useful outcome of fluoride-releasing restorative materials.\[[@ref3][@ref5]\] To address the issue, multiple modifications are being under investigation to develop a GIC with direct antimicrobial effect based on its ability to participate in ion-exchange reactions with the oral environment.\[[@ref3][@ref5]\] Such a material would be of tremendous clinical benefits. Providing antibacterial seal leads to suppression of remaining bacteriae under the restoration, reduced postoperative sensitivity, improved pulpal and periodontal health, and finally reduction in recurrent caries. Different studies are designed attempting to develop a modified GIC by incorporation of antimicrobial agents such as chlorhexidine hydrochloride, cetylpyridinium chloride, cetrimide, and benzalkonium chloride.\[[@ref6]\] Hook et al. proposed a GIC functionalized with chlorhexidine-hexametaphosphate particles. The suggested material released chlorhexidine in a dose-dependent manner and showed to have antimicrobial properties.\[[@ref3]\] In another study, de Castilho et al. added doxycycline hyclate as an antibiotic with local antimicrobial and antimatrix metalloproteinase activities into the GIC. The modified material showed inhibitory effects on caries producing bacteria.\[[@ref7]\] Deepalakshmi et al. incorporated chlorhexidine and cetrimide as wide-spectrum antiseptics into GIC and developed an antibacterial material.\[[@ref6]\] Herbal extracts have recently gained even more attention as active agents to be incorporated into oral care products and dental materials. Many of them have potential antimicrobial activities which could lead to the production of safe, economical, and efficient alternative materials for use in caries management.\[[@ref8]\] The Iranian herbal medicine with a long history of effectiveness and safety is a rich source of such agents. Salvia officinalis, which is also called Sage, is a perennial evergreen plant native to Mediterranean region\[[@ref8]\] and is also naturalized in Iran.\[[@ref9]\] It has woody stems, grayish leaves with blue-to-purplish flowers. A long ancient and contemporary history of medicinal and pharmacological applications supports its use. S. officinalis extract has antimicrobial, analgesic, anti-inflammatory and antioxidant properties.\[[@ref8][@ref10]\] The essential oil of S. officinalis consists of alpha- and beta-thujone, borneol, camphor, and cineole. In contemporary European herbal medicine, a type of S. officinalis tea is commonly advocated to control a sore throat, inflammatory oral lesions, and gingivitis.\[[@ref8][@ref10]\] Kermanshah et al. showed Sage mouthwash to have an inhibitory effect on caries-producing oral microbial flora.\[[@ref9]\] In another study, Beheshti-Rouy et al. proved the S. officinalis extract to be effective in reducing the number of Streptococcus mutans colonies in bacterial plaque.\[[@ref11]\] Different other studies also reported the inhibitory effect of S. officinalis extract on collagenolytic activity of Porphyromonas gingivalis,\[[@ref12]\] and also on the attachment of Candida albicans to denture surfaces.\[[@ref8]\] In an attempt to develop a GIC with direct antimicrobial properties, this study is conducted to investigate any possible inhibitory effects S. officinalis modified GIC may have on S. mutans and Lactobacillus casei as the main bacteria involved in caries initiation and progression process. The experiment is performed for S. officinalis weight concentration levels of 0.5%, 0.75%, 1%, and 1.25%. Methods {#sec1-2} ======= Some high-quality dried S. officinalis is purchased from a local herbal medicine retail store (Parand herbal products, Isfahan, Iran). Leaves of the plant are chopped and fragmented into small pieces and filtered through a \#40 mesh (Sina Lab. Inst., Tehran, Iran). Each 50 g of leaves are soaked in 1500 ml of solvent (50% water, 50% ethanol \[96%\]) in a shaker apparatus (Heidolph Unmax, Schwabach, Germany) at 90 rpm for 48 h. Thereafter, the solution is passed through a strainer and then transferred to a rotary evaporator apparatus (Heidolph WD2000; Schwabach, Germany) to separate the solvent from the extract.\[[@ref11]\] The purified extract is then dried by applying freeze-drying technique in three stages over a 1-week period. The final extract powder is stored in sealed vial at low temperature to be used in the next steps. To filter the particles, the same size as range of GIC powder which is to be \<50 μ, the powder is grinded and again filtered through a \#270 laboratory mesh (Sina Lab. Inst., Tehran, Iran). The procedure is performed under the supervision of a pharmacology professor at the main laboratory of the School of Pharmacy, Isfahan University of Medical Sciences and Health Services, Isfahan, Iran. A conventional powder and liquid Fuji IX GIC (GC Corporation, Tokyo, Japan) is used as control group named Group I. Experimental GIC samples are prepared by incorporating S. officinalis extract powder into the powder component of Fuji IX GIC (GC Corporations, Tokyo, Japan) at 0.5% (Group II), 0.75% (Group III), 1% (Group IV), and 1.25% (Group V) weight concentration levels using a digital weight scale (CAS MWP Micro Weighing Balance, CAS Scales, New Zealand). Thus, four groups of experimental powder formulations are obtained, each sample weighs 1000 mg according to the procedure outlined in.\[[@ref6]\] The concentrations used in this research are based on the minimal inhibitory concentration and minimal bacterial assays suggested in.\[[@ref9]\] The groups are presented in [Table 1](#T1){ref-type="table"}. ###### Test groups assigned to each bacterial category based on the weight concentration of Salvia officinalis extract powder Test group number\* Weight percentage concentration of S. officinalis extract powder^\^^ --------------------- ------------------------------------------------------------------------ I (control) 0 II 0.5% III 0.75% IV 1% V 1.25% \These groups are assigned to each bacterial category (S. mutans* and L. casei), ^\^^Weight concentration of S. officinalis extract powder added to the powder component of Fuji IX Glass-ionomer cement (GC Corporation, Tokyo, Japan) to form test groups. S. mutans=Streptococcus mutans, L. casei=Lactobacillus casei, S. officinalis=Salvia officinalis Five specimens are prepared from each experimental group by a specialist in restorative dentistry. Powder/liquid is dispensed on a mixing pad according to the manufacturer recommendations and mixed for 30 s with a sterile plastic spatula, and then inserted into the stainless steel molds of 10 mm in diameter and 2-mm thickness (Abzar Iran, Arak, Iran) within 1 min with a sterile dental instrument and allowed to set for 30 min at room temperatures, same as the procedure in.\[[@ref6]\] Stock cultures of S. mutans (UA159) and L. casei (ATCC \#193) obtained from the Microbiology and Immunology Laboratory of the School of Medicine, Isfahan University of Medical Sciences and Health Services, Isfahan, Iran are used by an expert laboratory technician. For each bacterial category, cells are cultured freshly from frozen stock on brain--heart infusion broth (BHI; DIFCO Laboratories, Detroit, MI, USA) for 24 h at 37°C in a 10% CO~2~ incubator. The viability and absence of contamination is confirmed by plating in a specific medium and using Gram procedures. Cultures are again grown in BHI for 20 h at 37°C and the inoculums for subsequent testing are obtained according to.\[[@ref7]\] In each sterilized Petri dish (15 mm × 90 mm), a base layer containing 15 ml of BHI agar mixed with 300 ml of each inoculums (S. mutans and L. casei) was prepared according to.\[[@ref7]\] The set disk-shaped specimens are placed on a BHI agar plate inoculated with bacterial strain and left at 37°C for 48 h. Zones of inhibition are then measured in millimeters using a digital caliper at three different points. Size of the inhibition zones is calculated by subtracting the diameter of the specimen from the average of the three measurements of the halo according to.\[[@ref6]\] Statistical analysis {#sec2-1} -------------------- Data are analyzed using Statistical Package for the Social Sciences, SPSS software version 22 for Windows (IBM Inc., NY, USA). SPSS is a software package used in statistical analysis of data. This software package was developed by SPSS incorporation and acquired by IBM in 2009. In 2014, the software was officially renamed IBM SPSS statistics. Kolmogorov--Smirnov test is used to test the normality of the data. The test is rejected at significance level of 5%, and the data follow distributions different than normal. Thus, nonparametric Kruskal--Wallis, Mann--Whitney, and Wilcoxon signed-rank sum tests are performed to compare the test groups and also the test groups with the control group. Data are analyzed using Statistical Package for the Social Sciences (SPSS, version 22) for Windows (IBM Inc., NY, USA). Kolmogoro v--Smirnov test is used to test the normality distribution of the data. The nonparametric Kruskal--Wallis and Mann--Whitney tests are then employed to compare the test groups and also the test groups with the control group. The statistical analyses are all performed at significance level of 5%. Results {#sec1-3} ======= According to Kruskal--Wallis test, there was a significant difference between the mean diameter of the inhibition zone for Streptococcus mutans (P \< 0.001) and Lactobacillus casei (P \< 0.001) in different concentrations of salvia extract \[[Table 2](#T2){ref-type="table"}\]. ###### The mean diameter of the inhibition zone for Streptococcus mutans and Lactobacillus casei in different concentrations of salvia extract Concentration Mean±SD P ------------------------ --------- ----------- ----------- ----------- ----------- --------- Streptococcus mutans 0 0.26±0.05 0.02±0.34 0.02±0.62 0.07±1.1 \<0.001 Lactobacillus casei 0 0.05±0.04 0.02±0.47 0.05±0.56 0.05±1.06 \<0.001 According to Mann-Whitney test, the mean diameter of the inhibition zone for Streptococcus mutans was significantly different between all groups with each other and with the control group (P \< 0.05) and the mean diameter of the inhibition zone for Lactobacillus casei was significantly different between all groups with each other and with the control group (P \< 0.05), except between the control group and the group with the concentration of 0.5% (P = 0.134). \[Tables [3](#T3){ref-type="table"} and [4](#T4){ref-type="table"}\]. ###### Comparison of the mean diameters of the inhibition zones between the test groups in the Streptococcus mutans category\* Test group Group I Group II Group III Group IV Group V ------------ --------- ---------- ----------- ---------- --------- Group I \- Group II 0.005 \- Group III 0.004 0.014 \- Group IV 0.005 0.007 0.006 \- Group V 0.005 0.007 0.006 0.007 \- \Wilcoxon signed-rank sum test as a nonparametric test was used for comparing the means. The numbers show P* - value within groups. Significant level was considered at 5% ###### Comparison of the mean diameters of the inhibition zones between the test groups in the Lactobacillus casei category\* Test group Group I Group II Group III Group IV Group V ------------ --------- ---------- ----------- ---------- --------- Group I \- Group II 0.134 \- Group III 0.005 0.007 \- Group IV 0.005 0.007 0.02 \- Group V 0.005 0.007 0.007 0.007 \- \Wilcoxon signed-rank sum test as a nonparametric test was used for comparing the means. The numbers show P* - value within groups. Significant level was considered at 5% Pairwise comparison of the mean diameter of inhibition zones of samples with similar concentrations in the Streptococcus mutans and Lactobacillus acidophilus showed that only the inhibitory effect of the concentration of 0.5% on Streptococcus mutans was significantly higher than that of Lactobacillus acidophilus \[[Table 5](#T5){ref-type="table"}\]. ###### Comparison of the mean diameters of the inhibition zones of the same group between two bacterial categories of Streptococcus mutans and Lactobacillus casei\* Test group Mean diameter of the inhibition zones P ------------ --------------------------------------- --------- ----------- Group II 0.26 mm 0.04 mm 0.006 Group III 0.34 mm 0.47 mm 0.058^\^^ Group IV 0.62 mm 0.56 mm 0.339^\^^ Group V 1.10 mm 1.06 mm 0.99^\^^ \Wilcoxon signed-rank sum test as a nonparametric test was used for comparing the means. Significant level was considered at 5%, ^\^^P-value indicates no statistically significant difference. S. mutans=Streptococcus mutans, L. casei=Lactobacillus casei* Discussion {#sec1-4} ========== In this research, the possibility of producing an herbally modified GIC with direct antibacterial properties is investigated. A distinguishing feature of this study is the use of an herbal extract with a long history of safety and effectiveness instead of chemical agents to produce a modified GIC. The results showed that the incorporation of S. officinalis extract into the powder component of a routinely used GIC produces a dental material with direct inhibitory properties against the most important bacterial causes of dental caries. It is also shown that this ability has a dose-response nature. The statistically nonsignificant difference between Group I and Group II (control group) in the L. casei category shows a lower potential in S. officinalis to inhibit the L. casei at its lowest concentration. The statistically significant higher value of the mean inhibition zone diameter in Group II of the S. mutans category compared to the Group II of the L. casei category points also to the same finding which should be considered cautiously. The findings of this study challenge the traditional idea that the dental materials must be passive without any interactions with surrounding oral environment\[[@ref4]\] and also recommends the use of materials with active capabilities as a possible approach in the management of different carious states. It seems that the ability of such a modified GIC to inhibit the bacterial activity depends on its interactions with the surrounding aqueous environment.\[[@ref3][@ref5]\] It is shown that bacteria responsible for caries may survive up to 2 years even under regular GICs, which many believe to have anticarious properties.\[[@ref13]\] This means that any intentional (in case of close proximity to pulp) or unintentional residual caries under the restorations may progress and encroach on pulp and produce discomfort for patient and finally endanger the pulp vitality.\[[@ref14][@ref15][@ref16]\] This also questions the accuracy of ART, in which multiple carious teeth are only excavated with a hand instrument and filled with a regular GIC, especially in low socioeconomic areas of society.\[[@ref2][@ref16][@ref17]\] A modified GIC with direct antimicrobial properties addresses the situation here. Different investigators incorporated different antimicrobial agents in different concentrations into GIC in an attempt to find a novel solution.\[[@ref1][@ref3][@ref6][@ref7][@ref18]\] Wide-spectrum antimicrobials such as chlorhexidine,\[[@ref1][@ref3][@ref18]\] cetrimide,\[[@ref6]\] and antibiotics\[[@ref7][@ref6]\] were at the center of the scope of those studies. The agent(s) were added into the powder or liquid component of the GIC, utilizing different incorporation techniques. However, none of them are incorporated any herbal agent as a safer source with fewer or no possible complications such as mucosal or dental staining seen in chlorhexidine products.\[[@ref19]\] The reasons for selecting the S. officinalis in this research are the long history of successes and no history of any negative side effect documented for, in traditional and contemporary medical literature.\[[@ref12]\] It is also reported that S. officinalis extract has superior efficiency compared to the antibiotics and has also less chance of drug resistance when used in high concentrations.\[[@ref20]\] In addition, researchers are recently investigating possible applications of S. officinalis essential oil extract in oral care products. The studies reported inhibitory effect of the extract on many oral bacteria and fungi responsible for carious and periodontal infections.\[[@ref8][@ref9][@ref11][@ref12]\] They also reported positive effect on reducing the number of plaque bacterial colonies,\[[@ref11]\] impairing plaque formation,\[[@ref21][@ref22]\] and fungi attachment to prosthetic surfaces.\[[@ref8]\] Based on the available evidence, which recognizes the anti-inflammatory properties in S. officinalis extract,\[[@ref9][@ref12][@ref20][@ref22][@ref23]\] one may suggest the use of the S. officinalis modified GIC may alleviate and control the inflammatory status of the pulp or periodontium. It is recommended to conduct experiments to investigate such abilities. While the presence of S. officinalis extract as an antibacterial agent in GIC is effective against oral pathogens, investigations to evaluate other important properties such as mechanical, physical, and biocompatibility properties of the material are highly recommended before any routine clinical applications. Considerations should be taken to improve the color of S. officinalis modified GIC since the current product has an intense yellow color. Conclusions {#sec1-5} =========== Within the limitations of this study, it is concluded that a S. officinalis modified GIC has a direct antibacterial property against S. mutans and L. casei. This, in turn, provides many advantages for use as a base, liner, or restorative material in operative procedures and management of the complications of carious teeth. It also emphasizes the importance of the herbal agents in developing new dental materials with astonishing properties. Financial support and sponsorship {#sec2-2} --------------------------------- Nil. Conflicts of interest {#sec2-3} --------------------- There are no conflicts of interest. The authors are thankful to the staff of the Department of Microbiology, School of Medicine, and the Main Laboratory of the School of Pharmacy -- Isfahan University of Medical Sciences and Health Services, Isfahan, Iran, for their support and friendly cooperation during every step of this investigation. We also would like to thank Mr. Tavangar, School of Dentistry -- Isfahan University of Medical Sciences and Health Services, Isfahan, Iran, for his kind help in organizing and analyzing the data.	{ "pile_set_name": "PubMed Central" }
INTRODUCTION {#sec1-1} ============ The development of whole slide imaging (WSI) has made it possible to capture images of an entire histological slide. While different hardware platforms support the image capture process, it is the development of user-friendly software interfaces that has enabled pathologists to begin to seriously contemplate integration of digital pathology activities into their routine clinical practice. The user-friendly interfaces make it possible to navigate to various regions of the scanned slide and change magnification for diagnosis (digital pathology), to view the image remotely by different pathologists in real time (telepathology), to capture static images of the slide for reporting or archiving, and to perform computer-aided analysis (digital image analysis). Roles for WSI in surgical pathology to support off-site frozen section interpretation;\[[@ref1]--[@ref4]\] primary diagnosis, not only as telepathology consultations,\[[@ref5]--[@ref7]\] but also to improve service to underserved areas;\[[@ref8][@ref9]\] educational activities;\[[@ref10]--[@ref14]\] and laboratory quality assurance (QA) activities are all well established.\[[@ref15]--[@ref18]\] While the advantages of WSI for all of these applications are well established, the criteria for evaluating the cost of WSI-based activities have traditionally been based solely on direct costs (hardware and software) and diagnostic accuracy.\[[@ref6][@ref19]--[@ref22]\] However, these types of analyses largely ignore a spectrum of workflow and cost issues that arise when WSI is integrated into routine diagnostic surgical pathology as a result of redundancy (due to the fact that histological sections on glass slides are an intrinsic component of surgical pathology, and the addition of WSI necessarily involves duplication of the primary diagnostic material), the need for additional staffing (so that WSI does not introduce delays in diagnosis), and customized software development (often required to address unique aspects of specific pathology practices). The pre-scan, scan, and post-scan costs; quality control and QA costs; and IT process costs that result when WSI is integrated into routine surgical pathology can be significant.\[[@ref6][@ref21][@ref23][@ref24]\] Consequently, most pathology groups find it difficult to perform a realistic cost--benefit analysis of adding WSI to their practice, or are uncertain as to the process they should use to decide whether WSI will have utility in their spectrum of patient care activities. In this paper, we report a "value added" approach developed to guide our decisions regarding integration of WSI into our surgical pathology practice.\[[@ref25]--[@ref28]\] The approach focuses on specific operational measures (cost, time, and enhanced patient care) and the settings in which they can provide enhancement (patient care, education, and research). Application of this value added approach provided the focus to identify several routine activities in which the addition of WSI could improve our practice, and resulted in expanded and improved surgical pathology operations. SETTING {#sec1-2} ======= Our group is a large surgical pathology practice (33 attending pathologists, 35 residents, and 16 subspecialty fellows from 10 different pathology subspecialty fellowship programs) at an academic tertiary care medical center and affiliated medical school. The practice has a subspecialty emphasis model with 12 subspecialty sections and multiple sign-out areas (individual pathologist\'s offices, common sign-out rooms, two frozen section areas that are in different buildings one city block apart); handles a large volume of high complexity cases (approximately 55,000 cases per year, including in-house cases and consultations); staffs over 40 patient care conferences per month; supports clinical, translational, and basic research activities of the department and medical school; and is involved in medical student teaching. In support of the diagnostic activities of the group, the histology laboratory processes approximately 200,000 blocks per year (approximately 800 blocks per day) and produces about 380,000 slides per year (approximately 1500 slides per day) including over 28,200 immunohistochemical stains per year, over 1600 in situ hybridization (ISH) slides per year, and over 3100 immunofluorescence stains per year. Value added in our practice is defined in purely operational terms in three categories \[[Table 1](#T1){ref-type="table"}\], specifically cost savings, time savings, and improvements in patient care.\[[@ref1]--[@ref3][@ref5]--[@ref12][@ref15][@ref16]\] Although added value can be assessed on a number of different axes within each of these three categories, analysis was limited to support of patient care, educational activities, and research. Some components of the analysis are quantifiable and objective (see below), other components of the analysis are subjective. Added value can be achieved overall despite negative impacts in some categories and/or along some axes within each category. Finally, it should be emphasized that some aspects of WSI are specifically not included in a value added approach; for example, the mere capability to produce a digital image that can be used for diagnosis is, in and of itself, not value added simply because it is technically feasible or novel. ###### Operational definition of "Value added" ![](JPI-2-39-g001) VALUE ADDED ANALYSIS {#sec1-3} ==================== [Table 2](#T2){ref-type="table"} presents the operational costs associated with WSI in our practice. Note that the time to scan slides in routine surgical pathology workflows is significantly higher than advertised by vendors, which reflects the operational realities of whole slide scanning as part of daily practice. The capital and personnel costs to incorporate WSI into our practice are summarized in [Table 2](#T2){ref-type="table"}. In the most recently completed calendar year (2010), we scanned 2.7% of the slides from our laboratory, for which we needed one scanner (Aperio ScanScope, Aperio Technologies Inc., Vista, CA, USA) at approximately 30% maximal utilization, one digital imaging technician (at approximately 0.5 FTE) and one information technology (IT) support person (at approximately 0.3 FTE). Were we to scan every slide in our practice, it would require an additional initial capital investment of approximately \$2,000,000 (for hardware and software), and a yearly indirect cost of approximately \$650,000 for support personnel and \$10,000 for data storage. ###### Operational costs for incorporating WSI into routine surgical pathology workflows ![](JPI-2-39-g002) Value added: Enhanced patient care {#sec2-1} ---------------------------------- There are five areas in which WSI provides capabilities which are superior to current options, or for which other options are simply not available \[[Table 3](#T3){ref-type="table"}\]. Although these areas do not improve diagnostic accuracy of the specimen of origin of the scanned slides, they provide operational advantages that improve patient care in our practice in four ways. First, the use of WSI of selected slides from cases sent in consultation (the medicolegal climate in the United States dictates return of all diagnostic materials to the referring institution) provides the opportunity to enhance patient care by providing an immediately available permanent record of the slides to guide frozen section diagnosis at the time of subsequent definitive excision; for comparison at sign-out of subsequent excision or post-therapy specimen; for presentation at patient care conferences; in QA activities; and so on. Second, WSI of selected slides sent to other institutions as requested or required by their policies for patient care, or slides encumbered by medicolegal proceedings, provides us with a permanent record for use in patient care activities even though we lose control of the original glass slides. Third, WSI of slides that will be destroyed as part of ancillary testing makes it possible to retain the diagnostic content of the slides; given the demonstration that molecular tests can be performed on nucleic acids collected from diagnostic areas microdissected from glass slides,\[[@ref29]\] the electronic record of slides produced by WSI will likely become more important. Fourth, WSI of slides for digital image analysis, for example, HER-2/neu analysis, supports emerging slide-based diagnostic paradigms.\[[@ref30][@ref31]\] ###### Value added in enhanced patient care ![](JPI-2-39-g003) Value added: Educational programs {#sec2-2} --------------------------------- The use of WSI to support educational activities is well established.\[[@ref10]--[@ref14]\] At our institution, the added value was provided by the opportunity to produce virtual study sets. WSI also made it possible to view selected slides remotely from anywhere in the institution, which can be used to support medical lectures, medical student labs, and both intramural and extramural departmental educational activities. Value added: Research {#sec2-3} --------------------- WSI makes it possible to produce virtual study sets that can be viewed for expert panel review, eliminating the need to produce multiple recuts of the study sets, mailing the various panel numbers, and even travel to a common site for real-time synchronous viewing by panel members. WSI is also used to support various clinical trials (in which digital images can substitute for glass slides for consensus for centralized review, e.g., by the Gynecologic Oncology Group), as well as to create a permanent record of slides that will be destroyed by the ancillary testing that is part of a clinical trial. Not value added {#sec2-4} --------------- In our practice, WSI does not provide added value in many areas classically associated with digital imaging, although it must be emphasized that aspects of WSI that are not value added for our group may well provide a benefit in other practice settings. Since we have a large laboratory, we already have on site the glass slides from both routine histochemical and immunohistochemical stains, and so WSI does not provide the opportunity to save money or time in primary diagnosis. Also, since our practice has a subspecialty emphasis model with experts on site in virtually every organ system (quite literally across the hallway, if not one block away), there is no value added component for WSI to save time or increase diagnostic accuracy. Similarly, since we already have a well-developed "bricks and mortar" consultation service, there is no immediate advantage to be gained in our practice by WSI for telepathology or branding. In fact, several impediments were identified that actually impede the capture of value added components of WSI in our consultation practice. First, among our referral base, the 1-day delay introduced by overnight express shipping of slides for primary or subspecialty review is not viewed as a significant detriment to patient care (in fact, overnight shipping of slides for early morning delivery integrates nicely into the routine workflow of our group). Second, although overnight express shipping is somewhat cumbersome and introduces the risk of slide loss or damage, it is extremely cost-effective (at current rates, the cost of express overnight shipping one 18 × 12.5 × 3 inch box that can hold up to 100 slides and 100 blocks, every day for 1 year, is less than \$4000; in contrast, the current cost of entry level WSI hardware and software is approximately \$135,000--\$160,000 for a 5--10 slide scanner). Third, significant IT resources are required to develop, implement, and support the HIPAA compliant processes necessary to transfer the images; to support automated reporting to numerous outside client lab information systems; and to bill for the patient care activities (to say nothing of navigating the quagmire of state licensure requirements). COSTS ASSOCIATED WITH IMPLEMENTATION {#sec1-4} ==================================== It became clear early in our evaluation of WSI that the faculty and trainees at our institution varied in their experience with the software for viewing digital slide images, their willingness to incorporate WSI into their routine practice, and in the increases in time they were willing to tolerate (even among faculty who were willing to include WSI into their workflow). The faculty and trainees were vocal in their unwillingness to incorporate a WSI process that required them to move back and forth between different software packages (with concomitant multiple log-ons using different usernames and passwords, after time-outs due to inactivity of the different programs, and so on), and most were unenthusiastic about a need to have several computer monitors so that multiple software packages could be open at the same time (notably, faculty that already had two computer screens running various software packages simultaneously were reluctant to add a third computer monitor to accommodate WSI). Space constraints limited the feasibility of multiple computer screens at all work stations, and the department was reluctant to purchase additional monitors. Finally, from a purely operational perspective, we were reluctant to pursue a model incorporating a requirement for multiple monitors that might work well at the medical center, but would not support remote faculty sign-out (either at home or while traveling), remote faculty and trainee research activities, or remote faculty and trainee educational activities. Consequently, in collaboration with several vendors, we pursued a model of "one-stop-shopping" in which there was an automatic in-house interface between our lab information system (Cerner CoPath Plus, Cerner Corporation, Kansas City, MO, USA) and the imaging software (Aperio Spectrum). Development of this new functionality took approximately 11 months overall, including 9 months to design the system architecture and write the associated software code and 2 months to implement within our existing workflow. Implementation of the automatic interface had an overall cost of approximately \$70,000 \[for software development for the Aperio and the CoPath HL7 interfaces, and for purchase of the underlying CoPath Advanced Bar Coding and Tracking (AB and T) module\]. The "one-stop-shopping" solution is illustrated in [Figure 1](#F1){ref-type="fig"}. ![One-stop-shop solution. In CoPath, clicking on the Digital Imaging tab (upper red oval, panel a) immediately opens an Aperio Spectrum WSI window (panel b) that automatically displays the scanned slides from that case. Clicking on the Image Gallery tab (lower red oval, panel a) immediately opens a static image gallery (panel c) that includes the scanned paperwork from the case, gross images, and microscopic images. Secure remote access via the Citrix solution opens a window (panel d) with links (red ovals) to CoPath (for access to clinical WSI, as above) or directly to Spectrum (for access to research and educational WSI)](JPI-2-39-g004){#F1} The need for the automatic interface emphasizes the types of additional costs that are often overlooked in evaluations of the utility of WSI in routine pathology workflow. Specifically, so-called off-the-shelf hardware and software packages, regardless of the vendor, have generic functionality. Integration of WSI into specific pathology practices often requires hardware and software changes to the standard products which can be costly and time consuming to develop and implement. OUTCOMES {#sec1-5} ======== We also sought to evaluate whether the value added approach was successful in guiding integration of WSI into our clinical practice in ways that enhanced our patient care, research, and education activities. Our metrics in this analysis included changes in the number of scans, changes in faculty and trainee acceptance, and utilization \[[Table 4](#T4){ref-type="table"}\]. ###### Outcomes of a value added approach ![](JPI-2-39-g005) Number of scans {#sec2-5} --------------- The number of cases scanned per year has shown consistent growth (at least 33% per year over the last 3 years). Of note, the average number of slides scanned per cases has decreased, which we interpret as evidence that the faculty and trainees are gaining experience with WSI and are learning which slides contribute the most to patient care. This change in the number of slides scanned per case suggests that the economic model presented above may require revision as we gain more experience with WSI. Acceptance {#sec2-6} ---------- Measurement of faculty and trainee acceptance is not a simple endeavor. However, we interpret faculty and trainee demands for remote access to clinical WSI activities via laptop computer, tablet personal computers, and smart phones as evidence that the faculty and trainees are integrating WSI into their routine workflows. In this regard, our pathologists are voicing an expectation in line with use by their clinical colleagues of mobile devices to access lab test results\[[@ref32]\] and imaging studies.\[[@ref33]\] We have implemented the same one-stop approach to support the demands for remote access through development of a secure infrastructure in which all data are housed in an HIPAA approved environment. Access to the environment is via a Citrix application (Citrix Systems, Fort Lauderdale, FL, USA), specifically a Citrix Presentation server 4.5 installed on a Windows 2003 server-based system, which permits delivery of applications as services and thus provides on-demand secure access for users. Although this solution gives us the capability to access patient reports as well as view whole slide images in a HIPAA compliant environment, our use of Citrix\'s Independent Computer Architecture (ICA) protocol has clear limitations when accessing and scrolling through large amounts of graphical data; since ICA protocols were not expressly designed for this functionality, current user experience can be choppy. Expanded utilization {#sec2-7} -------------------- Our initial value added approach identified WSI of slides seen in consultation as an enhancement to patient care. We interpret faculty interest in extending WSI to include select in-house cases as evidence of increasing recognition of a role for WSI in patient care activities. We likewise interpret faculty and trainee interest in extending WSI into a broader range of educational and research roles as evidence that our value added approach was successful in demonstrating key areas in which WSI could provide utility. Available FREE in open access from: <http://www.jpathinformatics.org/text.asp?2011/2/1/39/84232>	{ "pile_set_name": "PubMed Central" }

1.. Introduction ================ Cirrhotic cardiomyopathy (CCM) was first described by Kowalski and Abelmann in 1953. The estimated incidence rate is about 56% of patients waiting for orthotopic liver transplantation (OLT) without a previous history of cardiac disease \[[@b1-ijms-15-08037]--[@b4-ijms-15-08037]\]. The CCM mechanism has been discussed relative to hyperdynamic circulation, but even after liver transplantation 7%--15% of deaths are related to cardiac-related dysfunction \[[@b5-ijms-15-08037]\]. Humoral factor abnormalities are suggested to play important roles in cirrhosis, finally causing blunted cardiac response or cardiac dysfunction \[[@b6-ijms-15-08037]\]. In liver fibrosis the transforming growth factor (TGF-β) is required and TGF-β signaling blunting can reduce fibrogenesis \[[@b7-ijms-15-08037]--[@b9-ijms-15-08037]\]. TGF-β production occurs in non-parenchymal liver cells when the liver is damaged, especially by Kupffer cells and hematopoietic stem cells (HSC) rather than fully differentiated epithelial cells \[[@b10-ijms-15-08037]\]. In cardiac remodeling TGF-β leads in modulating fibroblast phenotype and gene expression \[[@b11-ijms-15-08037]\]. TGF-β also promotes extracellular matrix deposition by upregulating collagen and fibronectin synthesis in the infarct \[[@b12-ijms-15-08037]\]. In addition, dilative ventricular remodeling by inducing interstitial fibrosis is also mediated through the TGF-β signaling pathway \[[@b12-ijms-15-08037]\]. Up to now, there has been no standard treatment because CCM does not have a solidly established diagnosis and is based on high clinical suspicion \[[@b13-ijms-15-08037]\]. In the therapeutic strategy for fibrogenesis prevention, TGF-β signaling should also be reduced using any treatment for CCM patients \[[@b14-ijms-15-08037]--[@b16-ijms-15-08037]\]. Here, silymarin and a traditional Chinese herbal medicine formula (San Huang Shel Shin Tang, SHSST) and its water-soluble beta-cyclodextrin (β-CD) complex modification compound (SHSSTc) were used in treating CCM rats, which were induced from carbon tetrachloride (CCl~4~) intraperitoneal (IP) injection induced cirrhosis model \[[@b17-ijms-15-08037]--[@b19-ijms-15-08037]\]. Silymarin is a well known drug against cirrhosis and is a cocktail-like herbal liver-protective drug with four flavonolignan isomers, silybin, silychristin, silydianin and isosilybin \[[@b20-ijms-15-08037],[@b21-ijms-15-08037]\]. SHSST is also a cocktail-like traditional herbal decoction used for liver and heart protection in China. SHSST is composed of 50% *Rheum officinale* Baill, 25% *Scutellaria baicalnsis* Geprgi and 25% *Coptis chinensis* Franch in weight \[[@b22-ijms-15-08037],[@b23-ijms-15-08037]\]. Rheum was reported to have a liver protection effect that can protect the liver in CCl~4~-induced injury treatment in rats. *Scutellaria* and *Coptis* were also reported to have similar liver protection effects in acute hepatotoxicity. The liver protection effects between *Rheum*, *Scutellaria* and *Coptis chinensis* are due to the same bioactive compounds, baicalein and other flavonoids \[[@b24-ijms-15-08037]--[@b28-ijms-15-08037]\]. Both SHSST and silymarin are potential liver protection drugs, but both present poor water solubility and poor bioavailability \[[@b17-ijms-15-08037]\]. A formulation approach is necessary to increase the solubility of these liver protection drugs. β-CD modification can increase the solubility and spectral properties of guest molecules, especially hydrophobic drugs, without changing their intrinsic property to permeate the cell membranes. Thus, SHSST was modified into SHSST-β-CD-complex (SHSSTc) and evaluated for its therapeutic effects in a CCM animal model. 2.. Results and Discussion ========================== The heart phenomena changes in each group were measured and presented in [Table 1](#t1-ijms-15-08037){ref-type="table"}. The average heart weight in CCl~4~-induced CCM groups was higher than the control. The average heart weights in silymarin, baicalein, SHSST, and SHSSTc low and high dose treatments were reduced and similar to the control group average heart weight. The ratio between left ventricular weight (LVW) and tibia length (TL) is a particularly accurate indicator for cardiomegaly. The average LVW/TL of CCl~4~-induced CCM groups are higher, and silymarin and baicalein can slightly reduce cardiac hypertrophy. Interestingly, the SHSST, SHSSTc low and high dose treatments can significantly reverse cardiac hypertrophy in CCl~4~-induced CCM rat hearts. In CCl~4~-induced CCM heart proteins analysis, the high level expression of brain natriuretic peptide (BNP) presents a stress load on the beating heart. ([Figure 1](#f1-ijms-15-08037){ref-type="fig"}) After 4 weeks of silymarin, baicalein and SHSST treatments, BNP levels were slightly reduced and lower than that in the CCM only group. Low and high dose SHSSTc treatments efficiently inhibited the BNP secretion in a dose dependent manner. The heart biopsy showed a large area of collagen accumulation in the CCl~4~-induced CCM rat heart. The hematoxylin and eosin (H&E) staining assay also showed cardiomyocytes presenting a disordered arrangement with more inter-space between the CCM heart cells ([Figure 2](#f2-ijms-15-08037){ref-type="fig"}). After 4 weeks of silymarin, baicalein and SHSST treatments, the cell arrangement became neat and close. Low and high dose SHSSTc treatments caused the cell arrangement to present just like the control and the collagen accumulation areas disappeared. The protein analysis results show that CCl~4~-induced CCM efficiently increased the TGF-β, phoshphorylated mothers against decapentaplegic homolog 3 (p-Smad3) and CTGF protein levels. ([Figure 3](#f3-ijms-15-08037){ref-type="fig"}) The silymarin, baicalein and SHSST treatments can only slightly reduce the TGF-β signaling pathway. However, low and high SHSSTc dose treatments significantly reduced TGF-β pro-form (monomer) and mature-form (dimer) expression in a dose dependent manner. Cirrhosis usually causes a hyper-dynamic state and this syndrome leads to increased cardiac output and decreased systemic vascular resistance \[[@b29-ijms-15-08037]\]. In some CCM cases the cardiac dysfunction in cirrhosis is not associated with the liver disease severity \[[@b30-ijms-15-08037]\]. On the contrary, one in five CCM patients suffers death after liver transplantation \[[@b31-ijms-15-08037]\]. The death risk might be caused by the maladaptive cardiac re-circulation system with a new liver. The traceable biomarker is BNP, which is directly secreted from the cardiac response to output stress \[[@b32-ijms-15-08037],[@b33-ijms-15-08037]\]. BNP expressions were indeed increased in the CCl~4~-induced CCM animal model in this research. ([Figure 1](#f1-ijms-15-08037){ref-type="fig"}) Cardiac hypertrophy was also induced in this animal model ([Table 1](#t1-ijms-15-08037){ref-type="table"}). In the results of this study cardiac fibrosis was serious in the CCM rat heart. ([Figure 2](#f2-ijms-15-08037){ref-type="fig"}) The protein analysis results also show a high mature-form of TGF-β expression in CCM. ([Figure 3](#f3-ijms-15-08037){ref-type="fig"}) Such results suggest the high expression might cause cardiac fibrosis or remodeling \[[@b11-ijms-15-08037],[@b12-ijms-15-08037]\]. Thus, in addition to support therapy for OLT patient treatment for TGF-β expression reduction might work and needs further clinical research to confirm. It is interesting that BNP expression levels after low and high dose SHSSTc treatments are comparable in the heart, while the fibrosis parameters (TGF-β pathway activation) are different (less activated in the high dose group). This might suggest that high dose SHSSTc treatment could provide a stronger effect on the TGF-β pathway. Another possible mechanism might exist in BNP expression reduction in CCM rat hearts after both low and high dose SHSSTc treatments. Silymarin and baicalein (one of the bioactive compounds in SHSST) were reported to reduce TGF-β expression in CCl~4~-induced acute liver injury in a rat model \[[@b34-ijms-15-08037],[@b35-ijms-15-08037]\]. This study also found a similar therapeutic effect can be used in CCM. SHSST treatment in CCM rats also inhibited TGF-β expression and its downstream proteins p-Smad3 and CTGF, similar to silymarin and baicalein. Although, SHSST and silymarin are effective, they are still limited by their hydrophobic characteristics. Low drug water solubility often presents poor bioavailability, preventing cirrhosis patients from accepting the normal use dose. This study used β-CD modified SHSSTc, which requires only minimal dose treatment to suppress TGF-β expression in the CCM rat heart. SHSST was reported to have an anti-atherogenic effect (0.2 mg/mL) in human aortic smooth muscle cells and suggested to present cytokine production inhibition effects \[[@b36-ijms-15-08037]\]. SHSST and SHSSTc were used for the first time and their cardiac protect effects against CCM evaluated in this research. The experimental results from this research show that low dose (30 mg/kg/day) SHSSTc can provide a protective effect in the CCM animal model through reducing TGF-β. SHSSTc does not just inhibit TGF-β dominated fibrosis in CCM rat hearts, but also slows CCM rat cardiac hypertrophy in a dose dependent manner. Thus, TGF-β might play an important role in heart remodeling in CCl~4~-induced CCM rat hearts. Silymarin, SHSST and baicalein treatments can reduce TGF-β expression and the downstream pathway, but SHSSTc treatment leads to a significantly improved result. 3.. Experimental Section ======================== 3.1.. Preparation of SHSST-β-CD (San Huang Shel Shin Tang-β-cyclodextrin) Complex and Drugs ------------------------------------------------------------------------------------------- SHSST-β-CD complex was prepared by co-precipitation. β-CD (70.0 g) was dissolved in distilled water (85 mL) at 70 °C in a water bath for 1 h. SHSST (10.0 g) in ethanol (15 mL) was slowly added to the β-CD solution with continuous agitation for 6 h. Forty milliliter of ethanol was then added drop wise to regulate the hydrophobic solute solubility in β-CD solution. Afterwards the solution was refrigerated overnight at 4 °C. The precipitated SHSSTc (SHSST:β-CD = 1:9 in weight) was recovered by filtration and washed with ethanol to remove unencapsulated SHSST. This residue was dried in a vacuum oven at −20 °C for 48 h. The final powder was stored at 4 °C until used. The silymarin and baicalein were purchased from Sigma (St. Louis, MO, USA). The silymarin, baicalein, SHSST and SHSSTc stock solutions for treatments were prepared by dissolving in distilled deionized water at 100 mg/mL each. CCl~4~ was dissolved in olive oil at the concentration of 4% *v*/*v*. 3.2.. Animal Model ------------------ The animal use protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of China Medical University (No. 100-3-B, date: 1 September 2010). There were 42 Sprague-Dawley (SD) rats purchased from BioLASCO Taiwan Co., Ltd. (Taipei, Taiwan) and divided into 7 groups (*n* = 6 each). CCl~4~ intraperitoneal (IP) injection to rats for cirrhosis induction was applied twice with 0.2 mL/kg at the first and fourteenth days. After 4 weeks from the final CCl~4~ IP injection treatments, all CCl~4~-induced CCM rat hearts were confirmed by echocardiography with the ejection fraction (EF) \<50% and fractional shortening (FS) \<85% parameters (control group EF \> 50% and FS \> 85%). After CCM induction, 4 more weeks of drug treatments were applied through gavage assay to each rat group. Groups were designated control, CCl~4~-induced CCM, silymarin (100 mg/kg/day) oral treatment, baicalein (30 mg/kg/day) oral treatment, SHSST (30 mg/kg/day) oral treatment, SHSSTc (30 mg/kg/day) oral treatment, SHSSTc (300 mg/kg/day) oral treatment. 3.3.. Cardiac Echocardiography ------------------------------ M-mode echocardiographic examination was performed using a 6--15 MHz linear transducer (15--6 L) via a parasternal long axis approach. Left ventricular (LV) M-mode measurements at the papillary muscles level included left ventricular internal end-diastolic dimensions (LVIDd), left ventricular internal end-systolic dimensions (LVIDs). Fractional shortening (FS%) was calculated according to the following equation: FS% = \[(LVIDd − LVIDs)/LVIDd\] × 100%. Ejection fraction (EF) is defined as the ratio between the volume of blood pumped out of the LV and total volume of blood in LV. 3.4.. Hemotoxylin and Eosin Staining ------------------------------------ The rat hearts from each group were soaked in 10% formalin, dehydrated through graded alcohols and embedded in paraffin wax. The 0.2 μm-thick paraffin sections were then cut into slices from these paraffin-embedded tissue blocks. The tissue sections were deparaffinized by immersing in xylene and rehydrated. All slices were dyed with hematoxylin and eosin (H&E) and then rinsed with water. Each slide was dehydrated through graded alcohols. Heart sections were finally soaked in xylene twice. Photomicrographs were obtained using Zeiss Axiophot microscopes (Taiwan Instrument Co., Taipei, Taiwan). 3.5.. Masson's Trichrome Staining --------------------------------- The rat hearts from each group were soaked in 10% formalin, dehydrated through graded alcohols and embedded in paraffin wax. The 0.2 μm-thick paraffin sections were then cut into slices from these paraffin-embedded tissue blocks. The tissue sections were deparaffinized by immersing in xylene and rehydrated. Samples were then stained with Masson's trichrome staining to investigate liver histo-logical and fibrotic changes. Photomicrographs were obtained using Zeiss Axiophot microscopes. 3.6.. Tissue Protein Extraction ------------------------------- Heart tissue extracts from 6 rats in each group were obtained by homogenizing in a lysis buffer (0.05 M Tris--HCl, pH 7.4, 0.15 M NaCl, 0.25% deoxycholic acid, 1% nonyl phenoxypolyethoxylethanol, 1 mM EDTA) at a ratio of 100 mg tissue/1 mL buffer. The homogenates were placed on ice and then centrifuged at 13,000 rpm for 40 min. The supernatants were collected and stored at −80 °C for further experiments. 3.7.. Western Blot Assay ------------------------ Heart tissue protein concentration extracts were determined using the Lowry protein assay. Protein samples were separated in a 12% SDS polyacrylamide gel electrophoresis (SDS-PAGE) with a constant voltage of 75 V for 120 min. Proteins were then transferred to Hybond-C membranes (GE healthcare UK Ltd., Little Chalfont, Buckinghamshire, UK) using 50 V for 3 h. Polyvinylidene difluoride (PVDF) membranes were incubated in 3% bovine serum albumin (BSA) in tricine buffer solution. Primary antibodies including brain natriuretic peptide (BNP, SC-18818, Santa Cruz Biotechnology, Dallas, TX, USA), TGF-β (SC-31609, Santa Cruz Biotechnology), Smad-3 (SC-8332, Santa Cruz Biotechnology), α-tubulin (SC-5286, Santa Cruz Biotechnology), CTGF (SC-14939, Santa Cruz Biotechnology), were added into the membranes for recognizing the fitted proteins. Horseradish peroxidase-labeled antibodies were finally used and pictures were then taken with Fujifilm LAS-4000 (GE healthcare UK Ltd.). 3.8.. Statistical Analysis -------------------------- The results shown are the means ± SD of three independent experiments. Statistical analysis was performed using one-way analysis of variants. The Student's *t* test was applied for paired samples. 4.. Conclusions =============== The experimental evidence reported here suggests that SHSSTc treatment can improve the present clinical treatment for non-cirrhotic heart failure. Present clinical treatment includes bed rest, oxygen and diuretics. TGF-β expression elimination in CCM through SHSSTc treatment can improve heart function in CCM and may also improve the survival rate for OLT patients after clinical liver transplantation surgery. This study is supported in part by Taiwan Ministry of Health and Welfare Clinical Trial and Research Center of Excellence (DOH102-TD-B-111-004). **Author Contributions** The authors' contributions were as follows: F.-J.T, Y.T and C.-Y.H. designed the experiments; C.-H.Y., W.-J.T., D.-T.J. and C.-H.D. acquired and analyzed the results; C.-H.Y., W.-J.T., D.-T.J., Y.-L.Y. and C.-Y.H. interpreted the results; L.-C.C., C.-H.T., F.-J.T., Y.T. and C.-Y.H. prepared and edited the manuscript. All authors read and approved the final manuscript. The authors declare no conflict of interest. ![Phoshphorylated GATA binding protein 4 (p-GATA4) and brain natriuretic peptide (BNP) expressions in cirrhotic cardiomyopathy (CCM) hearts. (**A**) The p-GATA4 and BNP expressions (*n* = 6 in each group) were increased in CCM and reduced by silymarin, baicalein, SHSST (San Huang Shel Shin Tang), and SHSSTc low dose and high dose treatments; (**B**) The normalized protein expression folds of p-GATA4 with GATA4; (**C**) The normalized protein expression folds of BNP with α-tubulin (the scale bars is also presented as the normalized folds with α-tubulin); (**D**) Echocardiography analysis images (*n* = 6 in each group) of the heart function is compared by the left ventricular systolic and diastolic distance (cm). \*\* *p* \< 0.01, \*\*\* *p* \< 0.001 compared with control group; \# *p* \< 0.05, \#\# *p* \< 0.01, \#\#\# *p* \< 0.001 compared with CCM group.](ijms-15-08037f1){#f1-ijms-15-08037} ![Morphology assessments by hematoxylin and eosin (H&E) stain, Massion's trichrome stain (MS) assay in cirrhotic cardiomyopathy (CCM) rat hearts. In H&E stain slides, cell nuclei are stained with blue color, other intracellular or extracellular protein are stained with pink color. The heart fibrosis can be assessed using the collagen accumulation (indicated by blue color). Normal cells are indicated by pink color in the MS assay. All heart sections were obtained from the ventricular septal of each rat.](ijms-15-08037f2){#f2-ijms-15-08037} ###### The transforming growth factor (TGF)-β pathway protein expression level analysis. (**A**) TGF-β/phoshphorylated mothers against decapentaplegic homolog 3 (Smad-3)/connective tissue growth factor (CTGF) expression levels were increased in CCM and reduced by silymarin, baicalein, SHSST, and SHSSTc low and high dose treatments; (**B**) The normalized TGF-β protein expression folds with α-tubulin; (**C**) The normalized protein expression folds of p-Erk with Erk; (**D**) The normalized protein expression folds of p-Smad3 with Smad3; (**E**) The normalized protein expression folds of CTGF with α-tubulin. *n* = 6; \* *p* \< 0.05, \*\*\* *p* \< 0.001 compared with control group; \#\#\# *p* \< 0.001 compared with the CCM group. ![](ijms-15-08037f3a) ![](ijms-15-08037f3b) ###### Cardio vascular structure physiological characteristics assessment. Groups Control CCM CCM silymarin CCM Baicalein CCM SHSST CCM SHSSTc (low dose) CCM SHSSTc (high dose) ---------------- -------------- ----------------------------------------------------------------- ----------------------------------------------------------------- ----------------------------------------------------------------- ----------------------------------------------------------------- ----------------------------------------------------------------- ----------------------------------------------------------------- BW (g) 323.0 ± 8.9 352.3 ± 14.0 [\*](#tfn2-ijms-15-08037){ref-type="table-fn"} 343.7 ± 6.1 [\*](#tfn2-ijms-15-08037){ref-type="table-fn"} 325.7 ± 11.6 366.0 ± 22.3 [\*](#tfn2-ijms-15-08037){ref-type="table-fn"} 349.7 ± 25.8 316.7 ± 9.1 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} TL (cm) 3.80 ± 0.15 4.10 ± 0.00 3.97 ± 0.06 4.10 ± 0.00 4.10 ± 0.10 4.03 ± 0.06 4.00 ± 0.10 HW (mg) 936.7 ± 53.3 1146.0 ± 21.9 [\*\*](#tfn3-ijms-15-08037){ref-type="table-fn"} 927.3 ± 49.8 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 921.7 ± 17.8 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 940.3 ± 25.1 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 944.0 ± 20.1 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 965.7 ± 11.2 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} LVW (mg) 681.7 ± 15.0 922.7 ± 11.9 [\*\*\*](#tfn4-ijms-15-08037){ref-type="table-fn"} 711.3 ± 62.5 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 720.0 ± 62.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 683.7 ± 52.5 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 687.3 ± 54.7 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 683.0 ± 22.7 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} HW/BW (mg/g) 2.90 ± 0.10 3.26 ± 0.15 [\*](#tfn2-ijms-15-08037){ref-type="table-fn"} 2.70 ± 0.19 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 2.83 ± 0.13 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 2.85 ± 0.22 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 2.71 ± 0.25 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 3.15 ± 0.12 HW/TL (mg/cm) 238.6 ± 20.5 279.5 ± 5.3 [\*](#tfn2-ijms-15-08037){ref-type="table-fn"} 233.9 ± 16.0 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 224.8 ± 4.3 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 22.9 ± 11.7 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 234.1 ± 8.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 241.6 ± 8.3 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} LVW/BW (mg/g) 2.1 ± 0.1 2.6 ± 0.1 [\*\*](#tfn3-ijms-15-08037){ref-type="table-fn"} 2.1 ± 0.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 2.2 ± 0.3 1.9 ± 0.3 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 2.0 ± 0.3 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 2.2 ± 0.0 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} LVW/HW (mg/mg) 0.73 ± 0.03 0.81 ± 0.01 [\*](#tfn2-ijms-15-08037){ref-type="table-fn"} 0.77 ± 0.03 0.78 ± 0.05 0.73 ± 0.04 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 0.73 ± 0.04 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 0.71 ± 0.03 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} LVW/TL (mg/cm) 173.5 ± 8.1 225.0 ± 2.9 [\*\*\*](#tfn4-ijms-15-08037){ref-type="table-fn"} 179.5 ± 15.1 [\#](#tfn5-ijms-15-08037){ref-type="table-fn"} 175.6 ± 15.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 167.0 ± 16.9 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 170.6 ± 15.8 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 170.7 ± 2.4 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} EF (%) 80 ± 2.6 68.5 ± 1.0 [\*\*\*](#tfn4-ijms-15-08037){ref-type="table-fn"} 72.8 ± 3.3 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 72.3 ± 1.5 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 74 ± 1.0 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 76.7 ± 1.5 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 77 ± 0.3 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} FS (%) 43.7 ± 2.1 33.7 ± 1.4 [\*\*\*](#tfn4-ijms-15-08037){ref-type="table-fn"} 38.8 ± 6.1 36.5 ± 1.0 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 38.3 ± 0.6 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 40.7 ± 1.5 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} 41.0 ± 0.7 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} LVIDd (mm) 8.5 ± 0.1 8.0 ± 0.2 [\*\*\*](#tfn4-ijms-15-08037){ref-type="table-fn"} 8.2 ± 0.4 8.1 ± 0.3 8.1 ± 0.3 8.3 ± 0.2 [\#\#](#tfn6-ijms-15-08037){ref-type="table-fn"} 8.4 ± 0.1 [\#\#\#](#tfn7-ijms-15-08037){ref-type="table-fn"} BW, body weight; TL, tibia length; HW, whole heart weight; LVW, left vestibular weight; EF, ejection fraction; FS, fractional shortening; LVIDd, left ventricular interior dimention; *p* \< 0.05, *p* \< 0.01, *p* \< 0.001 compared with control group; *p* \< 0.05, *p* \< 0.01, *p* \< 0.001 compared with cirrhotic cardiomyopathy (CCM) group.

{ "pile_set_name": "PubMed Central" }

1. Introduction {#sec1-jcm-09-01661} =============== Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection \[[@B1-jcm-09-01661]\] that may culminate in organ failure and death. Despite continuous efforts to understand and improve outcomes of sepsis, it remains a commonly fatal disease \[[@B2-jcm-09-01661],[@B3-jcm-09-01661],[@B4-jcm-09-01661]\]. In addition, the current definition of sepsis identifies a heterogeneous population of individuals with diverse patterns of immune response, organ dysfunction, and clinical outcomes \[[@B1-jcm-09-01661]\]. Therefore, early diagnosis, precise stratification of severity, and accurate outcome prediction are critical in managing patients with sepsis. Various biomarkers, including C-reactive protein and procalcitonin, have been investigated to date \[[@B5-jcm-09-01661],[@B6-jcm-09-01661]\]; however, no marker has demonstrated sufficient discriminatory power \[[@B7-jcm-09-01661]\]. Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) is a cytokine and member of the TNF superfamily. TRAIL initiates apoptosis of transformed cells or tumor cells by binding to death receptors (DR) 4 or 5 \[[@B8-jcm-09-01661]\] and functions as an immune response regulator in sepsis \[[@B8-jcm-09-01661],[@B9-jcm-09-01661]\]. More recently, however, it was discovered that TRAIL also functions as a trigger for necroptosis, a specific form of programmed cell necrosis characterized by its caspase-independent activation and release of damage-associated molecular patterns (DAMPs) which convey highly proinflammatory properties \[[@B10-jcm-09-01661],[@B11-jcm-09-01661]\]. In a recent study, plasma TRAIL was related to poor outcomes in patients with sepsis \[[@B12-jcm-09-01661]\]. A subsequent multicenter study suggested a relationship between the plasma level of TRAIL with necroptosis in sepsis; however, its association with mortality demonstrated a contradictory result \[[@B13-jcm-09-01661]\]. Therefore, further studies are necessary to clarify the relationship between plasma TRAIL and predicting outcomes of patients with sepsis. In this study, we analyzed the association between the plasma level of TRAIL and severity and outcomes of sepsis. We also compared the plasma level of TRAIL to that of receptor-interacting protein kinase-3 (RIPK3), a well-known necroptosis mediator \[[@B14-jcm-09-01661]\], to identify the correlation between TRAIL and necroptosis. 2. Materials and Methods {#sec2-jcm-09-01661} ======================== 2.1. Study Design and Registry {#sec2dot1-jcm-09-01661} ------------------------------ This was a prospective observational study of the Samsung Medical Center Registry of Critical Illness (SMC RoCI), which is an on-going single-center prospective registry of the Samsung Medical Center (1989-bed, university affiliated, tertiary referral hospital in Seoul, South Korea) initiated in April 2014 for the purpose of establishing a human sample repository and developing new biological markers for critical illness \[[@B15-jcm-09-01661]\]. The study was approved by the institutional review board of Samsung Medical Center. Written informed consent was obtained from patients or their legally authorized representative prior to enrollment. 2.2. Study Patients {#sec2dot2-jcm-09-01661} ------------------- Critically ill adult (≥19 years old) patients admitted to the medical intensive care unit (ICU) of Samsung Medical Center were considered eligible for inclusion in the registry. Exclusion criteria were as follows: (1) cognitive impairment, (2) inability to provide informed consent, (3) ICU admission for a simple procedure or postsurgical care, (4) transfer from other hospitals, (5) end-of-life decision or admission to facilitate comfort care, (6) hemoglobin \< 8 g/dL upon admission or persistent bleeding, and (7) discharge within 24 h of admission to ICU. Screening and enrollment were completed within 24 h of ICU admission. Patients registered between April 2014 and December 2016 were included in the analysis. Some clinical data and RIPK3 levels from patients enrolled until August 2016 were reported in a previous study \[[@B15-jcm-09-01661]\]. 2.3. Data Collection {#sec2dot3-jcm-09-01661} -------------------- A trained study coordinator used hospital records for each patient to prepare a standardized case report form. Clinical data consisting of patient demographics, reason for ICU admission, severity of illness scoring, and laboratory data were obtained at the time of enrollment. Illness severity was assessed by the Acute Physiology and Chronic Health Evaluation II (APACH II) \[[@B16-jcm-09-01661]\], Simplified Acute Physiology Score 3 (SAPS 3) \[[@B17-jcm-09-01661]\], and Sequential Organ Failure Assessment (SOFA) scores \[[@B18-jcm-09-01661]\]. The Revised Trauma Score was used to determine the severity of trauma patients \[[@B19-jcm-09-01661]\]. The primary outcome was 28-day mortality. Secondary outcomes were in-hospital and 90-day mortality. Sepsis was defined according to the third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) \[[@B1-jcm-09-01661]\]. Since enrollment for the registry began in April 2014, patients enrolled before release of the new definition were reclassified. Patients in the registry who did not meet the definition of either sepsis or septic shock were defined as controls. 2.4. Measurement of Plasma TRAIL and RIPK3 {#sec2dot4-jcm-09-01661} ------------------------------------------ Along with clinical data, 19 mL of whole blood was drawn from each patient within 48 h of study enrollment. When possible, additional blood samples were collected at day 3 (15 mL) and day 7 (15 mL). Blood samples were centrifuged within 4 h of collection. Plasma was separated and stored at −80 °C until further analysis. Plasma TRAIL level was measured from stored aliquots using commercially available TRAIL Human ELISA kits according to the manufacturer's recommendations (R & D systems, Minneapolis, MN, USA). Plasma RIPK3 level was measured using a commercially available ELISA kit as per the manufacturer's recommendations (CUSABIO, Houston, TX, USA) \[[@B15-jcm-09-01661]\]. 2.5. Statistical Analysis {#sec2dot5-jcm-09-01661} ------------------------- Data are presented as numbers (percentages) for categorical variables, and as the median and interquartile range (IQR, 25th--75th percentiles) for continuous variables. Categorical variables were compared using the Chi-square test or Fisher's exact test, while continuous variables were compared using the Mann--Whitney *U* test. Differences in plasma TRAIL level across control, sepsis, and septic shock groups were assessed with the Kruskal--Wallis test. Baseline characteristics, clinical status at ICU admission, illness severity, and mortality were compared between patients with low and high TRAIL level divided by the median level of plasma TRAIL. To evaluate the association between plasma TRAIL level and patient outcomes, a Kaplan--Meier curve was used to determine the 90-day survival curves according to plasma TRAIL. These were then compared using the log-rank test. Linear regression was applied to assess the association between plasma levels of TRAIL and RIPK3. The Friedman test was performed to determine differences among serial levels of TRAIL. The Wilcoxon signed rank test with Bonferroni correction was used to compare plasma TRAIL levels between days 0 and 3 and between days 3 and 7. All tests were two-sided, and a *p* value \< 0.05 was considered significant. Data were analyzed using IBM SPSS Statistics 20.0 (IBM, Chicago, IL, USA). 3. Results {#sec3-jcm-09-01661} ========== Over the study period, 1419 patients were admitted to the ICU. After excluding 1224 patients who met the exclusion criteria and one patient who missed screening, 194 patients were enrolled in the registry. Since four patients withdrew their consent, 190 patients were finally included in the analysis ([Figure 1](#jcm-09-01661-f001){ref-type="fig"}). The baseline characteristics of 190 patients at the time of ICU admission are summarized in [Table 1](#jcm-09-01661-t001){ref-type="table"}. The median age was 64 (IQR, 54--73) years, and 121 (63.7%) patients were male. Of 62 (32.6%) patients with cancer, 32 patients had hematologic malignancy. The reason for ICU admission was sepsis in 59 (31.1%) patients and septic shock in 84 (44.2%) patients. For 47 patients of the control group, the most common cause of ICU admission was pulmonary edema in nine patients, followed by airway stenosis in six, interstitial lung disease in five, multiple trauma in five, and acute exacerbation of chronic obstructive pulmonary disease or asthma in four patients. The median revised trauma score for five patients with multiple trauma was 13.26 (IQR, 7.19--22.15). No patients in the control group were with infection. Ninety-five (50.0%) patients were on mechanical ventilator support, and 116 (61.1%) patients required vasopressor support upon ICU admission. The median plasma TRAIL level was 34.48 pg/mL in the study patients. While patients on mechanical ventilatory support were more common in the control group, patients requiring vasopressor support were more prevalent in the sepsis group. Furthermore, severity scores of SAPS 3 and SOFA, as well as the C-reactive protein level were higher in the sepsis group. However, no difference in 28-day ICU mortality or in-hospital mortality was observed. The plasma TRAIL level for 143 patients with either sepsis or septic shock was 31.55 pg/mL, which was lower compared to that of 52.00 pg/mL in the control group. There was a statistically significant trend of decreased plasma TRAIL level across control, sepsis, and septic shock groups (52.00 \[35.66--74.41\] pg/mL vs. 35.83 \[28.07--60.82\] pg/mL vs. 26.08 \[15.29--40.97\] pg/mL; *p* \< 0.001) ([Figure 2](#jcm-09-01661-f002){ref-type="fig"}). When correlation between TRAIL and RIPK3 was assessed, the plasma level of TRAIL was inversely related to the plasma level of RIPK3 in patients with sepsis and septic shock (r = −0.172, r^2^ = 0.02973, *p* = 0.039) ([Figure 3](#jcm-09-01661-f003){ref-type="fig"}). To evaluate the association between characteristics of patients with sepsis and plasma TRAIL level, 143 patients who were diagnosed with either sepsis or septic shock were divided into two groups---low and high plasma TRAIL level---according to the median TRAIL level ([Table 2](#jcm-09-01661-t002){ref-type="table"}). Septic shock and patients requiring vasopressor support were more prevalent in the group with low plasma TRAIL. Furthermore, lactic acid, the SAPS 3 score, APACHE II score, and SOFA score at initial ICU admission were significantly higher in patients with low plasma TRAIL. However, statistical differences in 28-day ICU mortality, in-hospital mortality, and 90-day mortality were not observed between the two groups. In addition, Kaplan--Meier survival estimation did not demonstrate a difference in 90-day survival between patients with high and low plasma TRAIL level (*p* = 0.419) ([Figure 4](#jcm-09-01661-f004){ref-type="fig"}). Finally, serial levels of plasma TRAIL were analyzed in 143 patients with sepsis. Blood samples at days 3 and 7 were available for TRAIL measurement in 59 and 32 patients, respectively. Plasma TRAIL levels increased over time, with median levels of 31.55 pg/mL (19.29--47.56) at day 0, 40.00 pg/mL (27.00--56.00) at day 3, and 40.50 pg/mL (25.00--53.75) at day 7 (*p* \< 0.001) ([Figure 5](#jcm-09-01661-f005){ref-type="fig"}A). Although the TRAIL level on day 3 was significantly higher compared to that of day 0 (*p* = 0.036), there was no difference in levels between days 3 and 7. Plasma TRAIL levels were also assessed according to 28-day ICU mortality ([Figure 5](#jcm-09-01661-f005){ref-type="fig"}B). While serial levels of TRAIL significantly increased over time in survivors (*p* = 0.005), TRAIL levels in non-survivors did not demonstrate any differences (*p* = 0.102). The TRAIL level in survivors was higher on day 3 compared to day 0 (*p* = 0.039); however, TRAIL levels of day 3 and day 7 did not differ. Conversely, there was no difference in levels of TRAIL between day 0 and day 3, or between day 3 and day 7 in non-survivors. Finally, when plasma TRAIL levels on days 3 and 7 were compared between survivors and non-survivors, there was a tendency for a higher TRAIL level only on day 3 in survivors compared to non-survivors \[42.00 (32.00--57.00) vs. 26.00 (17.50--53.50), *p* = 0.0502\]. Nonetheless, the day 7 levels demonstrated no difference between survivors and non-survivors \[40.00 (25.00--52.50) vs. 53.00 (23.00--75.00), *p* = 0.801\]. 4. Discussion {#sec4-jcm-09-01661} ============= The aim of this prospective observational study was to investigate the association between plasma TRAIL level and sepsis severity, as well as to determine the role of TRAIL in predicting mortality. In addition, we planned to analyze the correlation between plasma levels of TRAIL and RIPK3 to identify the potential relationship of TRAIL with necroptosis. Our study of 190 consecutively enrolled critically ill patients demonstrated that the plasma TRAIL level was inversely associated with sepsis severity and plasma level of RIPK3; however, it was not predictive of mortality. Until now, only two studies have investigated the significance of plasma TRAIL in sepsis. Tian et al. reported that the level of soluble TRAIL was lower in patients with septic shock (defined by an old definition) compared to healthy controls or patients with sepsis; soluble TRAIL was also associated with mortality \[[@B12-jcm-09-01661]\]. Very recently, a multicenter study demonstrated an inverse correlation between the level of circulating TRAIL and both organ dysfunction and RIPK3; \[[@B13-jcm-09-01661]\] however, the association of TRAIL with mortality was portrayed in only two of three cohorts. The inverse relationship between plasma TRAIL level and sepsis severity observed in our study concurs with the results of two previous studies, strengthening the growing evidence of the potential role of TRAIL in sepsis. The reasons for the inverse relationship between TRAIL and sepsis severity and the role of TRAIL in sepsis pathogenesis are not fully understood. Nonetheless, there are several suggested mechanisms to explain this phenomenon. One possible explanation is based on the anti-inflammatory effect of TRAIL. TRAIL is thought to facilitate and accelerate apoptosis of leukocytes, contributing to resolution of inflammation \[[@B20-jcm-09-01661]\]. Animal studies have shown that TRAIL is associated with improved survival in murine models of infection and autoimmune diseases \[[@B21-jcm-09-01661],[@B22-jcm-09-01661],[@B23-jcm-09-01661]\]. However, this protective anti-inflammatory effect remains controversial. In a study by Unsinger et al. on murine cecal ligation and puncture models, TRAIL-deficient mice were better able to control bacterial infection due to TRAIL-dependent immune suppression and immune unresponsiveness \[[@B24-jcm-09-01661],[@B25-jcm-09-01661]\]. In addition, plasma TRAIL level was measured at the early phase of sepsis in our studies, as well as previous studies. Since TRAIL-dependent apoptosis affects immune suppression and immune unresponsiveness, which plays a role in the relatively later course of sepsis \[[@B25-jcm-09-01661]\], changes in plasma TRAIL level may not be reflective of immune status. Considering those facts, the anti-inflammatory effect of TRAIL may not fully explain its association with the pathogenesis of sepsis. Another possible mechanism is infection clearance by the TRAIL-dependent pathway. In vitro studies have shown that TRAIL contributes to infection control by mediating elimination of infected cells or restriction of viral replication in influenza or encephalomyocarditis viral infection models \[[@B26-jcm-09-01661],[@B27-jcm-09-01661]\]. However, this observation is also conflicting. Cardoso Alves et al. demonstrated in their study that TRAIL depletion resulted in reduced NK-cell mediated antiviral CD8^+^ T cell killing, thereby leading to faster pathogen clearance in a mouse model of viral infection \[[@B28-jcm-09-01661]\]. Moreover, since pathogen clearance by T cell mediated cell death is crucial in viral infection but not in bacterial infection, it may not fully explain the results of our finding. Finally, the capability of TRAIL to initiate necroptosis, a programmed cell death mechanism, may be another possible theory \[[@B10-jcm-09-01661]\]. Although the precise function and effect of necroptosis in sepsis has yet to be elucidated, the detrimental contribution of necroptosis is thought to lie in cell death, leading to local tissue injury and excess release of pro-inflammatory DAMPs. DAMPs stimulate the innate immune response and trigger an inflammatory cascade, resulting in organ damage \[[@B15-jcm-09-01661],[@B29-jcm-09-01661],[@B30-jcm-09-01661],[@B31-jcm-09-01661]\]. The relationship between sepsis and necroptosis has been observed in murine models, as well as in critically ill patients. Recent studies examining the necroptosis regulators of RIPK1, RIPK3, and MLKL in patients with sepsis reported that their levels were associated with disease severity and/or mortality \[[@B32-jcm-09-01661],[@B33-jcm-09-01661]\]. In certain pathologic environments, such as acidic states, TRAIL triggers necroptosis via RIPK1 and RIPK3 to hinder caspase-dependent apoptosis \[[@B10-jcm-09-01661],[@B34-jcm-09-01661]\]. Our study, as well as the study by Schenck et al. \[[@B13-jcm-09-01661]\], illustrated a significant correlation between the plasma levels of TRAIL and RIPK3, proposing the possibility that TRAIL may have an impact on sepsis through the necroptosis pathway. Nonetheless, it should be noted that this correlation between the TRAIL and RIPK3 proteins may not provide direct evidence that TRAIL reflects the process of necroptosis of sepsis. Although we have identified in our preliminary analysis that the plasma level of TRAIL also correlates with that of mitochondrial DNA (a well-known DAMP) in sepsis ([Figures S1--S3](#app1-jcm-09-01661){ref-type="app"}), supporting the hypothesis that TRAIL is associated with necroptosis, further research to investigate the pathogenesis of TRAIL in sepsis would be reasonable. Contrary to previous studies, the plasma TRAIL level was not associated with either 28-day ICU mortality or in-hospital mortality in our study. Interestingly, in the study by Schenck et al. \[[@B13-jcm-09-01661]\], the association between the plasma level of TRAIL with mortality was depicted only in two American cohorts of the United States, but not in a Korean cohort. The authors speculated that different baseline characteristics or processes of care might have contributed to this discrepancy. However, definite causes could not be determined. No relationship between TRAIL and mortality was observed in our cohort, in addition to another Korean cohort studied by Schenck et al. \[[@B13-jcm-09-01661]\]. This raises an ethnic difference as one of the many possibilities. Since the mortality of patients with sepsis may be attributable to various factors, further studies are warranted to confirm the relationship between mortality and TRAIL level and to further investigate the reason for the current discrepancy. One noteworthy finding of this study is that serial change in plasma TRAIL levels was available for analysis. To our knowledge, this is the first study to provide trends of the plasma TRAIL in sepsis. In particular, two discoveries require attention. First, TRAIL levels increased over time; however, the levels between days 3 and 7 did not differ. Second, when survivors and non-survivors were divided, the plasma TRAIL level on day 3 compared to day 0 was higher in survivors, but not in non-survivors. That there was no difference between levels on days 3 and 7 suggests that TRAIL, or the mechanism underlying the decrease in TRAIL, stabilizes or recovers within the initial days of sepsis. This suggests that TRAIL may be important in the relatively early phase of sepsis. Furthermore, because the TRAIL level on day 3 increased in survivors but not in non-survivors, perhaps it is not only the initial TRAIL level, but also the early serial changes of TRAIL that are crucial in understanding the patient's course and predicting mortality. However, since the day 3 and 7 samples were available only in some of the 143 patients with sepsis, we could not draw a definitive conclusion. As we have limited understanding of the role of TRAIL in sepsis, changes in TRAIL levels may be a consequence of sepsis or organ failure, rather than demonstration of a key process underlying sepsis. Comprehension of the pathogenesis and kinetics of TRAIL in sepsis is required to determine the significance of serial changes in TRAIL. To fully appreciate the results of our study, several limitations should be acknowledged. First, our study was conducted at a single referral center, which may limit the generalizability of our data. Second, follow-up data on TRAIL were not available in some patients due to death, recovery, and refusal to collect additional samples. Third, based on our study protocol, enrollment of a study patient and sampling of blood were done within 24 h of ICU admission and 48 h of enrollment, respectively. Therefore, patients with high severity and early mortality may not have been included in the study. Interpretation of results requires caution, since there may be a selection bias. Fourth, the correlation between TRAIL and RIPK3 and its association with severity of sepsis does not provide sufficient evidence that TRAIL functions as an initiator of necroptosis in sepsis or necroptosis is activated in sepsis. Further studies regarding gene expression and quantification are necessary to confirm this relationship. 5. Conclusions {#sec5-jcm-09-01661} ============== In conclusion, plasma TRAIL level was inversely associated with sepsis severity and plasma level of RIPK3. Plasma TRAIL levels increased following the course of recovery. However, they were not predictive of mortality. The following are available online at <https://www.mdpi.com/2077-0383/9/6/1661/s1>, Association of plasma level of TNF-related apoptosis inducing ligand with mitochondrial DNA in sepsis. Figure S1: Correlation between plasma level of TRAIL and mtDNA in patients with 49 patients with sepsis, Figure S2: Correlation between plasma level of RIPK3 and mtDNA in patients with 49 patients with sepsis, Figure S3: Plasma levels of mtDNA in sepsis and septic shock. ###### Click here for additional data file. H.Y. collected and analyzed the data and drafted this manuscript. J.Y.L. conducted experiments and analyzed the data. J.P. collected and analyzed the data. J.H.Y. and G.Y.S. analyzed the data and revised the manuscript. K.J. conceived and designed this study, analyzed the data, and wrote the final manuscript. All authors have read and agreed to the published version of the manuscript. This work was supported by a Samsung Medical Center grant (SMO1190131). The authors disclose that they do not have any conflict of interest. APACHE II: Acute physiology, age, chronic health evaluation II; DAMPs: Damage-associated molecular patterns; ICU: Intensive care unit; RIPK3: Receptor-interacting protein kinase-3; SAPS3: Simplified acute physiology score 3; SOFA: Sequential organ failure assessment; TRAIL, Tumor necrosis factor related apoptosis-inducing ligand. ![Study flow diagram.](jcm-09-01661-g001){#jcm-09-01661-f001} ![Plasma levels of Tumor necrosis factor TNF-related apoptosis-inducing ligand (TRAIL) in control, sepsis, and septic shock (*p* for trend \< 0.001).](jcm-09-01661-g002){#jcm-09-01661-f002} ![Correlation between plasma level of TRAIL and RIPK3 in patients with sepsis (n = 143). Slope: −8.713 (95% CI: −0.426--−17.000), r^2^: 0.02973, Pearson's r: −0.172 (*p* = 0.039).](jcm-09-01661-g003){#jcm-09-01661-f003} ![Kaplan--Meier survival analysis comparing patients with sepsis or septic shock with high and low plasma level of TRAIL (n = 143). The 30-day and 90-day survival estimates are 76.6% and 60.5%, respectively for patients with high TRAIL (solid line), while it is 80.2% and 66.9% for patients with low TRAIL (dotted line) (*p* = 0.419, log-rank test).](jcm-09-01661-g004){#jcm-09-01661-f004} ![(**A**) Serial plasma levels of TRAIL on days 0, 3, and 7 of 143 patients with sepsis. (**B**) Comparison of serial plasma TRAIL levels in 28-day intensive care unit survivors (solid line) and non-survivors (dotted line). Upper and lower bars represent the 75th and 25th interquartile ranges, respectively.](jcm-09-01661-g005){#jcm-09-01661-f005} jcm-09-01661-t001_Table 1 ###### Baseline characteristics of study patients (n = 190). Characteristics Total (N = 190) Control (n = 47) Sepsis (n = 143) *p* Value ---------------------------------------- ------------------------- ------------------------- ------------------------- ----------- Age, years 64 (54--73) 60 (54--68) 67 (53--74) 0.086 Gender, male 121 (63.7) 25 (46.8) 96 (67.1) 0.085 Co-morbidities Solid tumor 62 (32.6) 11 (23.4) 51 (35.7) 0.120 Hematologic malignancy 32 (16.8) 7 (14.9) 25 (17.5) 0.681 Diabetes 55 (28.9 12 (25.5) 43 (30.1) 0.552 Chronic obstructive pulmonary disease 18 (9.5) 7 (14.9) 11 (7.7) 0.157 Chronic kidney disease 15 (7.9) 4 (8.5) 11 (7.7) 1.000 Myocardial infarction 11 (5.8) 4 (8.5) 7 (4.9) 0.470 Congestive heart failure 9 (4.7) 4 (8.5) 5 (3.5) 0.228 Cerebrovascular disease 6 (3.2) 0 6 (4.2) 0.339 Chronic liver disease 1 (0.5) 0 1 (0.7) 1.000 Charlson Comorbidity Index 2 (1--3) 2 (1--3) 2 (1--3) 0.554 Clinical status on ICU admission Need for MV 95 (50.0) 32 (68.1) 66 (46.2) 0.009 Need for vasopressor support 116 (61.1) 15 (31.9) 101 (70.6) \<0.001 Laboratory findings PaO~2~/FiO~2~ ratio 139.3 (125.2--298.1) 255.1 (125.4--362.1) 181.3 (124.7--280.5) 0.112 CRP, mg/dL 10.0 (3.0--19.0) 5.4 (0.8--9.1) 12.5 (4.1--22.5) \<0.001 Lactic acid, mg/dL 3.0 (2.0--4.0) 2.4 (1.2--4.3) 2.7 (1.8--4.0) 0.230 Severity of illness SAPS 3, points 52 (44--59) 48 (35--56) 53 (46--59) 0.008 APACHE II score 23 (18--28) 22 (15--27) 23 (19--29) 0.184 SOFA score, initial 8 (5--11) 5 (3--9) 6 (6--11) \<0.001 Plasma TRAIL, pg/mL 34.482 (23.446--56.946) 52.000 (35.657--74.410) 31.545 (19.289--47.564) \<0.001 Outcome 28-day ICU mortality 39 (20.5) 10 (21.3) 29 (20.3) 0.883 In-hospital mortality 48 (25.3) 12 (25.5) 36 (25.2) 0.961 Note: APACHE II, acute physiology and chronic health evaluation II; CRP, C-reactive protein; ICU, intensive care unit; MV, mechanical ventilation; SAPS 3, simplified acute physiology score 3; SOFA, sequential organ failure assessment; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand. jcm-09-01661-t002_Table 2 ###### Characteristics of patients with sepsis stratified according to the median level of plasma TRAIL (n = 143). Characteristics Low TRAIL (n = 72) High TRAIL (n = 71) *p* Value ---------------------------------- ---------------------- ---------------------- ----------- Age, years 67 (55--73) 62 (51--74) 0.418 Gender, male 40 (69.4) 46 (64.8) 0.553 Co-morbidities Solid tumor 10 (13.9) 16 (22.5) 0.448 Hematologic malignancy 14 (19.4) 11 (15.5) 0.534 Diabetes 23 (31.9) 20 (28.2) 0.623 Chronic kidney disease 11 (11.6) 7 (7.4) 0.322 Myocardial infarction 8 (11.1) 4 (5.6) 0.217 Congestive heart failure 3 (4.2) 3 (4.2) 1.000 Charlson Comorbidity Index 2 (1--3) 2 (1--3) 0.668 Septic shock 52 (72.2%) 32 (45.1%) 0.001 Clinical status on ICU admission Need for MV 36 (50.0) 30 (42.3) 0.353 Need for vasopressor support 61 (84.7) 40 (56.3) \<0.001 Laboratory findings PaO~2~/FiO~2~ ratio 191.7 (120.2--282.6) 180.6 (130.5--277.5) 0.436 CRP, mg/dL 13.7 (4.2--13.7) 12.0 (4.0--22.5) 0.305 Lactic acid, mg/dL 3.1 (2.1--4.5) 2.3 (1.7--3.3) \<0.001 Severity of illness SAPS 3, points 56 (50--68) 50 (42--56) \<0.001 APACHE II score 24 (19--30) 22 (19--27) 0.007 SOFA score, initial 9 (7--11) 7 (5--11) \<0.001 Outcome 28-day ICU mortality 12 (16.7) 17 (23.9) 0.279 In-hospital mortality 19 (26.4) 17 (23.9) 0.736 90-day mortality 23 (31.9) 26 (36.6) 0.556 Note: APACHE II, acute physiology and chronic health evaluation II; CRP, C-reactive protein; ICU, intensive care unit; MV, mechanical ventilation; SAPS 3, simplified acute physiology score 3; SOFA, sequential organ failure assessment; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand.

{ "pile_set_name": "PubMed Central" }

![](indmedgaz72376-0058){#sp1 .638} ![](indmedgaz72376-0059){#sp2 .639} ![](indmedgaz72376-0060){#sp3 .640} ![](indmedgaz72376-0061){#sp4 .641} ![](indmedgaz72376-0062){#sp5 .642} ![](indmedgaz72376-0063){#sp6 .643} ![](indmedgaz72376-0064){#sp7 .644} ![](indmedgaz72376-0065){#sp8 .645} ![](indmedgaz72376-0066){#sp9 .646}

{ "pile_set_name": "PubMed Central" }

INTRODUCTION ============ Prompt detection of stimuli whose novelty or familiarity is unexpected in the current environment (i.e., contextual novelty) provides an evolutionary advantage, enabling adaptive behavior and flexible learning of the new information (Kakade and Dayan, [@b24]). The ability to discriminate potentially salient familiar and novel information, in a context abundant with old and new stimuli, becomes crucial for ensuring effective contextual learning. Efficient discrimination may trigger an orienting response toward the new unexpected information and may therefore enhance memory formation (Tulving and Kroll, [@b53]). Indeed, the brain\'s dopaminergic system, including striatal and midbrain structures, has been shown to respond to reward anticipation and to prediction errors as well as having an effect on memory formation (Adcock et al., [@b1]; Wittmann et al., [@b56]). It has also been proposed (Lisman and Grace, [@b29]; O\'Carroll et al., [@b33]; Shohamy and Adcock, [@b47]) that interaction between the midbrain dopaminergic regions and the hippocampus is crucial for the construction of new memories, especially when novel information is involved (Bunzeck and Duzel, [@b3]; Bunzeck et al., [@b4]). Drawing on the measures of eye tracking, subsequent memory, and functional magnetic resonance imaging (fMRI), we further explored this proposal by manipulating the contextual novelty of novel and familiar visual stimuli in three separate experiments. The response of the dopaminergic midbrain and striatum to unexpected information may be associated with increased exploration of such a stimulus (Düzel et al., [@b9]), perhaps to establish its *absolute* novelty or familiarity status. Dopamine release, from the midbrain to the evaluation centers in the ventral striatum and from there to the hippocampus, may therefore be manifested behaviorally through an enhanced visual exploration of the presented information, leading to enhanced memory formation for this information. Eye movements have been found to be the markers of the increased visual exploration that leads to enhanced encoding into long-term memory (Kafkas and Montaldi, [@b19]), whereas pupil responses are argued to reflect resource allocation, cognitive processing (Granholm et al., [@b13]; Kahneman and Beatty, [@b23]; for a recent review, see Laeng et al., [@b28]), and more recently also encoding and retrieval processes (Kafkas and Montaldi, [@b19],[@b20]; Papesh et al., [@b36]). The current research explored eye movement and pupil response patterns as well as striatal/midbrain and medial temporal lobe (MTL) blood oxygenation level-dependent (BOLD) responses while unexpected novel and familiar stimuli were presented. Critically, in contrast with the previous studies that have focused on the encoding of rewarding information, the current experiments focused on retrieval when familiar and novel, expected and unexpected stimuli were (re-)encoded. The aim of this design was to further characterize the role of the dopaminergic system in memory at retrieval, for both familiar and novel stimuli, when these are either expected or unexpected, as determined by the experimental context. Retrieval provides a setting in which to vary the probability of old and new stimuli, so as to explore how memory is updated depending on both the novelty and the familiarity status of a stimulus, and the interaction of that status with contextual expectation. This investigation promises to provide a more comprehensive, but so far, lacking (Scimeca and Badre, [@b45]), understanding of the role of the dopaminergic system in memory. Specifically, the purpose of this research was to investigate the degree to which the detection of contextual novelty (for both novel and familiar stimuli) would trigger an increased response in the striatal and midbrain dopaminergic system and whether this would, in turn, modulate the eye behavior accompanying visual processing, and in so doing, enhance further encoding of contextual information. We hypothesized that if, indeed, striatal and midbrain activity accompanies the processing of contextual novelty this would drive exploratory behavior, reflected in eye movements and pupil responses. Furthermore, we hypothesized that functional connectivity of the striatal/midbrain structures with the MTL, in response to the processing of unexpected (compared to expected) information, would explain the enhanced learning. These hypotheses were explored in two separate, but matched, experiments drawing on fMRI BOLD responses and eye measures (pupil response and fixation patterns). Finally, subsequent memory for unexpected familiar and novel stimuli was explored in a third experiment using a surprise follow-up memory test. MATERIALS AND METHODS ===================== Participants ------------ In the fMRI study 17 (10 females) right-handed volunteers with a mean age of 22.6 years (SD = 4.2) gave informed consent and participated. Remuneration was given at a rate of £20 per session. Another 37 (27 females) undergraduate psychology students, with a mean age of 20.3 years (SD = 2.0) took part in the eye-tracking experiment in return for course credits. In both experiments, volunteers with normal or corrected (with contact lenses) to normal vision were recruited. No history of psychiatric or neurological disorder and no systematic use of psychotropic medicines or drugs were reported by participants in either experiment. Finally, 27 participants (24 females) with a mean age of 20 years (SD = 1.97) participated (receiving £5) in Experiment 3. The experiments were approved by the Research Ethics Committee of the University of Manchester (eye tracking and behavioral) and the National Research Ethics Service (fMRI study). Stimulus Materials ------------------ The pictures of manmade and natural objects, normalized to a mean gray level, were used in all three experiments. In the fMRI study, 318 of these stimuli were used (18 for practice), whereas the eye**-**tracking experiment used 220 stimuli (20 stimuli for practice), each subtending a visual angle of 18.69° horizontally and 14.03° vertically at presentation. In the behavioral subsequent memory study (Experiment 3), 280 stimuli (220 from the eye**-**tracking experiment plus 60 new foils) were presented. Owing to the sensitivity of pupillary responses to light intensity (Cheng et al., [@b5]), as well as to chromatic changes (Tsujimura et al., [@b52]), stimulus properties such as ambient light, brightness, contrast, and the color of the presented visual items were controlled (for further details regarding the development and standardization of these stimuli, see Kafkas and Montaldi, [@b19]). Experimental Design and Task ---------------------------- In the fMRI and eye-tracking experiments, participants completed two phases: an encoding phase and a retrieval phase, with the fMRI and eye-tracking data recorded during the retrieval phase. At encoding, a series of visual stimuli were encoded using a perceptual-matching-to-sample task with each stimulus appearing for 4 s. In each trial, participants were presented with triplet images of the same object and were asked to decide which of the two bottom images was identical to the target image presented on top ([Fig. 1](#fig01){ref-type="fig"}; for this task, see Montaldi et al., [@b32]; Kafkas and Montaldi, [@b20],[@b21]). After a short filled interval (∼10 min), participants were presented with two recognition conditions, a novelty task (NT) emphasizing the detection of novel stimuli, and a familiarity task (FT) emphasizing the detection of familiar stimuli. Critically, 70% of the items presented in the two tasks were the "target" stimuli (i.e., novel in the NT condition and familiar in the FT condition), whereas 30% of the items in each task were foils (familiar stimuli in the NT condition and novel stimuli in the FT condition; [Fig.1](#fig01){ref-type="fig"}). These foil stimuli, which were presented with considerably lower probability, are classified here as unexpected items, compared to the target familiar and novel stimuli, which are classified as expected. For each target stimulus, participants provided a rating of how strongly familiar (FT) or novel (NT) they felt they were. In response to the unexpected stimuli (foils), on the other hand, participants were only required to report them as familiar or novel. The comparison of the BOLD response (Experiment 1) and the eye-tracking measures (Experiment 2) produced by unexpected versus expected novel and familiar stimuli is therefore the main focus of this study. ![Experimental design of the fMRI (Experiment 1) and eye-tracking (Experiment 2) studies.](hipo0025-1262-f1){#fig01} All stimuli were presented for a period of 3 s during which participants responded. In the fMRI experiment, 6 alternating blocks of familiarity (FT) and novelty (NT) conditions were presented, whereas in the eye-tracking experiment the two tasks were presented sequentially in an order that was counterbalanced across participants. In the fMRI study, 60 null events (providing an implicit baseline) were intermixed with the 300 visual stimuli, pseudorandomly allocated across the six blocks with the only restriction being that null events were never presented consecutively. These null events involve a fixation cross presented centrally and constitute therefore slightly extended presentations of the interstimulus fixation cross. This ensures that the null events are implicitly processed as they are not distinguished from standard interstimulus presentations by participants. A debriefing procedure after the end of the fMRI session indicated that no participant perceived the null events as unexpected in any way. The design of the behavioral follow-up study (Experiment 3) was identical to the eye-tracking experiment with the only difference being the addition of a surprise retrieval task at the end in which memory for all the stimuli presented in the two main recognition conditions (FT and NT) was tested. Due to the addition of the surprise recognition procedure, the number of stimuli presented in the two main recognition conditions (FT, NT) was reduced maintaining, however, the same probability (30%) for new and old foils in FT and NT, respectively. The completion of the familiarity and novelty tasks was followed by a 10-min distracter task, after which participants were presented with the surprise recognition memory test in which 120 stimuli from the FT and NT conditions were presented together with 60 new foils. Participants were instructed to provide familiarity (F) or recollection (R) responses to each stimulus if they believed they had seen it at some point in the experiment, or to report it as new, if they felt they were encountering the stimulus for the first time (remember/know procedure; for the instructions used see Migo et al., [@b31]). fMRI and Eye-Tracking Data Acquisition -------------------------------------- MR scanning was carried out on a 3T MRI scanner (Philips, Achieva) using a gradient echo-planar imaging (EPI) sequence to obtain T2\* images, employing the BOLD contrast. Forty slices positioned parallel to the AC--PC line and covering the whole brain (matrix size, 96 × 96; voxel size, 2.5 × 2.5 × 3.5 mm) were acquired for each volume (TR = 2.5 s; TE = 35 ms). In total, 726 volumes were recorded per participant across two sessions (363 per session). High-resolution T1 anatomical images (size, 256 × 256; 180 slices; size: 1 mm isotropic) were acquired for each participant before the two functional sessions. In the eye-tracking study, pupillary responses and eye movements were recorded using an ASL infrared eye-tracking system (Applied Science Laboratories, Bedford, MA; Model Eye-Trac 6000; sampling rate 60 Hz). A nine-point standard calibration procedure was carried out separately for each participant prior to data collection. For each trial, the eye-tracking data were recorded and analyzed from stimulus onset until the point of behavioral response. Blinks and other losses in the raw eye signal were identified by the eye-tracking software and were discarded from further analyses. Pupil recordings departing more than 3 SDs from the mean of each trial were discarded, as such abrupt pupil dilations or constrictions are considered noise in the pupil signal (Granholm et al., [@b13]; Beatty and Lucero-Wagoner, [@b2]). Trials containing \<60% of valid recordings were excluded (\<5% of total). The peak pupil size for each trial was calculated as the average of three pupil recordings preceding and three recordings following the maximum pupil value of the corresponding trial. The resulting peak pupil responses for each trial of the experiment were expressed as the deviation from the baseline pupil size, recorded during a period of 1,000 ms preceding each trial (Beatty and Lucero-Wagoner, [@b2]), consisting of a gray background (RGB level, 130). Finally, from the raw eye movement signal, a measure of spatial fixation (i.e., the number of fixations) was extracted. Data Analyses ------------- ### fMRI preprocessing The fMRI data were preprocessed using SPM8 (Statistical Parametric Mapping, Wellcome Department of Cognitive Neurology, University College London, London, United Kingdom; <http://www.fil.ion.ucl.ac.uk/spm/>). The EPI data were realigned to the first image and resliced using a six-parameter rigid body transformation and sinc interpolation in space. Slice-timing correction was also applied to the resliced images, which were then spatially normalized to the Montreal Neurological Institute (MNI) EPI template. Smoothing was performed using an isotropic (FWHM, 6 mm) Gaussian kernel. A canonical hemodynamic response function (Friston et al., [@b12]) was used to model the event-related responses for each participant. Correctly recognized expected and unexpected familiar and novel stimuli were modeled as separate conditions, whereas regressors of no interest (e.g., cue screens preceding each block, trials with no response, etc.) were also included in each individual model. The six movement parameters for each of the two sessions of the experiment were also modeled as regressors to capture residual movement-specific variance in the time series. ### fMRI data analyses Contrasts of the *t-statistic* for the four conditions of interest (expected novelty, unexpected novelty, expected familiarity, and unexpected familiarity) were produced for each participant in the first-level analysis, and carried forward to a second-level group analysis treating participants as a random (mixed) effect. A two-way ANOVA was conducted in SPM with the stimulus type (familiar or novel) and expectation (expected or unexpected) as the within-subjects factors. Two direct *t* contrasts (unexpected novelty \> expected novelty and unexpected familiarity \> expected familiarity) were used to further explore the functional data. Activations surviving a cluster-wise *P* \< 0.05 significance level (cluster FWE-corrected) are reported unless otherwise stated. MNI coordinates are reported for the activation data. ### Psychophysiological interaction analyses Effective connectivity was explored using psychophysiological interaction (PPI) analysis in SPM8 (Friston et al., [@b11]). PPI models were used to explore the hypothesis that the detection of unexpected familiar or novel stimuli would be associated with increased connectivity between the midbrain/striatal areas and the MTL. A PPI analysis explores the effect of activity in one brain region, on the activity in another region under the influence of a psychological contrast (i.e., unexpected vs. expected events). A significant PPI between two brain regions indicates that the slope of the regression characterizing their activity changes owing to the effect of the psychological context (Friston et al., [@b11]). Two areas, within the globus pallidus (GP) and the substantia nigra/ventral tegmental area (SN/VTA), showing selective responses to unexpected familiarity and novelty respectively (Results section) were chosen as the two functional seeds. Spheres of 6 mm radius within the GP (centered at *x* = 17, *y* = 1, *z* = −5) and SN/VTA (centered at *x* = 7, *y* = −20, *z* = −8) were created for each participant and the deconvolved BOLD activation data were extracted from these volumes of interests. Two separate PPI analyses were set up for the two seed areas, one for unexpected familiarity (contrast: unexpected familiar \> expected familiar stimuli; seed area: GP) and another for unexpected novelty (contrast: unexpected novel \> expected novel stimuli; seed area: SN/VTA). At the first-level analysis, three conditions were entered for each participant: (a) the BOLD signal time series of the seed region (*physiological variable*), (b) the contrast of interest (unexpected \> expected familiarity or novelty; *psychological variable*), and (c) the interaction term expressed as the multiplication of the psychological and physiological variables (*PPI*). A general linear model, with these three regressors, was estimated and *t*-contrasts of the interaction term were created for each participant and analyzed at the group level using a random-effects model (for further details, see Friston et al., [@b11]). The same statistical inference criteria as those applied to the main fMRI analyses were used for identifying significant activations. The MTL regions are reported as significant when the activation survived a FWE small-volume correction for the volume of the MTL or when eight or more contiguous voxels are active at *P* \< 0.001. This level was established as being equivalent to a corrected probability level of *P* \< 0.05 for the volume of the MTL using a Monte Carlo simulation (implemented in the AlphaSim tool in AFNI). ### Eye-tracking data analysis The fixation and pupil data were analyzed using a 2 × 2 repeated measures ANOVA with stimulus type (familiar, novel) and expectation (expected, unexpected) as the within-subject factors. Direct pairwise comparisons are also reported, where relevant, adopting a significance level of *P \<* 0.05 (Bonferroni corrected). ### Follow-up surprise recognition memory experiment analysis Data from the follow-up recognition memory experiment (Experiment 3) were analyzed to determine whether the unexpected novel and familiar stimuli were later remembered better than the expected novel and familiar stimuli. Using signal detection, *d*′ scores were calculated for familiarity and recollection responses provided at later test for the originally expected and unexpected novel and familiar stimuli. Two ANOVAs were run separately for the novel and familiar stimuli, with response in the follow-up task (familiar or recollected) and expectation status in the original task (expected or unexpected) as the within-subjects factors. Paired *t*-tests were also employed to further explore the significant interactions using a significance level of *P* \< 0.05 (Bonferroni corrected). RESULTS ======= Experiment 1: fMRI Results -------------------------- Behavioral data acquired during the fMRI study showed that the participants were able to accurately categorize novel and familiar stimuli in both tasks with well above chance accuracy (0.76 ± 0.08 and 0.77 ± 0.06). The analysis of the response times (RTs) for novel and familiar expected and unexpected stimuli showed a significant main effect of expectation (*F*~1,16~ = 5.50, *P* = 0.03). Expected stimuli (familiar~expected~: *M* = 1,749 ms, SD = 236 ms, and novel~expected~: *M* = 1,839, SD = 198 ms) had longer RTs than the unexpected ones (familiar~unexpected~: *M* = 1,679 ms, SD = 193 ms, and novel~unexpected~: *M* = 1,714 ms, SD = 220 ms). The fMRI analyses revealed striatal, midbrain, and neocortical activations for the main effect of expectation as well as for the expectation by stimulus-type interaction ([Fig. 2](#fig02){ref-type="fig"}). Four sets of neural response emerged. First, one region within the left caudate nucleus (caudate body, [Fig. 2a](#fig02){ref-type="fig"}) responded selectively to expected familiar and novel stimuli compared to unexpected stimuli. A second set included areas within the midbrain/brainstem ([Fig. 2c](#fig02){ref-type="fig"}) and the inferior temporal gyrus (BA 37; *x* = −53, *y* = −52, *z* = −19; 37 voxels, *P* \< 0.001 uncorrected), which responded to both familiar and novel unexpected stimuli relative to expected ones. A third set of regions, within the right GP ([Fig. 2b](#fig02){ref-type="fig"}), the left inferior parietal lobe (extending into the occipito-parietal junction; [Fig. 2d](#fig02){ref-type="fig"}), and the left precuneus ([Fig. 2e](#fig02){ref-type="fig"}) showed selective involvement in the processing of unexpected familiarity when compared to expected familiarity, but no modulation for novelty. Finally, the interaction effect revealed a set of regions in the right caudate (caudate head, [Fig. 2f](#fig02){ref-type="fig"}), the left middle frontal gyrus ([Fig. 2g](#fig02){ref-type="fig"}), the right inferior parietal lobe ([Fig. 2h](#fig02){ref-type="fig"}), and the right precuneus ([Fig. 2i](#fig02){ref-type="fig"}) that respond to unexpected familiarity (relative to expected familiarity) and to expected novelty relative to unexpected novelty. Critically, in these latter regions unexpected familiarity responses were significantly higher than unexpected novelty responses, whereas there were no significant differences between expected familiar and expected novel stimuli. This pattern of results suggests that the regions isolated in the interaction analysis have a predominant role in responding to unexpected familiarity. ![fMRI results from Experiment 1. A main effect of expectation status (expected/unexpected) was observed within (a) the left caudate body *x* = −21, *y* = 6, *z* = 20 (b) the right GP *x* = 15, *y* = −7, *z* = −1, (c) the left midbrain *x* = −3, *y* = −27, *z* = −26, (d) the left occipito-parietal junction *x* = −36, *y* = −70, *z* = 31; BA39 and BA19 and (e) the left precuneus *x* = −11, *y* = −75, *z* = 45; BA7. A significant interaction between expectation status (expected/unexpected) and stimulus type (familiar/novel) was observed in (f) the right caudate head *x* = 5, *y* = 11, *z* = 6, (g) the left middle frontal gyrus *x* = −31, *y* = 16, *z* = 34; BA8/9, (h) the right inferior parietal lobe *x* = 42, *y* = −72, *z* = 48; BA39 and (i) the right precuneus *x* = 12, *y* = -57, *z* = 41; BA7. Parameter estimates are plotted for each region separately. Error bars show one standard error of the mean. Asterisks denote statistically significant effects; \**P* \< 0.05; \*\**P* \< 0.01; \*\*\**P* \< 0.001. \[Color figure can be viewed in the online issue, which is available at [wileyonlinelibrary.com](http://wileyonlinelibrary.com).\]](hipo0025-1262-f2){#fig02} To further constrain these effects, direct contrasts between unexpected and expected familiar and novel stimuli were run separately. The unexpected novelty versus expected novelty revealed significant activity in the right SN (*x* = 7, *y* = −20, *z* = −8, *P* \< 0.001 uncorrected, 21 voxels, *Z* = 4.17; [Fig. 3a](#fig03){ref-type="fig"}). The same contrast (unexpected vs. expected) for familiar stimuli revealed significant activity in the bilateral lentiform nucleus (including GP and putamen; [Fig. 4a](#fig04){ref-type="fig"}; Right: *x* = 17, *y* = 1, *z* = −5, cluster FWE-corrected; Left: *x* = −13, *y* = −2, *z* = −1, *P* \< 0.001 uncorrected, 13 voxels) extending into the thalamus (*x* = 10, *y* = −10, *z* = −1) and the caudate nucleus (*x* = 7, *y* = 6, *z* = −5). In the whole-brain analysis, consistent with the ANOVA effects, the bilateral angular gyri (BA 39; left: *x* = −38, *y* = −72, *z* = 38; right: *x* = 42, *y* = −70, *z* = 41), the bilateral middle occipital gyri (BA 19; left: *x* = −36, *y* = −67, *z* = 24; right: *x* = 37, *y* = −77, *z* = 27), and the bilateral precuneus (BA 7; left: *x* = −6, *y* = 72, *z* = 41; right: *x* = 12, *y* = −57, *z* = 34) also showed significant activity (*P* \< 0.05 cluster FWE-corrected) for the unexpected versus expected familiar stimuli. ![(a) Activity and parameter estimates within SN/VTA (*x* = 7, *y* = −20, *z* = −8) for unexpected novelty and (b) PPI connectivity with the hippocampus (*x* = −36, *y* = −32, *z* = −12 and *x* = −38, *y* = −12, *z* = −26). \[Color figure can be viewed in the online issue, which is available at [wileyonlinelibrary.com](http://wileyonlinelibrary.com).\]](hipo0025-1262-f3){#fig03} ![(a) Activity and parameter estimates within the bilateral ventral striatum/GP (*x* = 17, *y* = 1, *z* = −5 and *x* = −13, *y* = −2, *z* = −1) for unexpected familiarity and (b) PPI connectivity with the perirhinal cortex (x = -31, y = -5, z = -33) and the hippocampus (*x* = 35, *y* = −35, *z* = −5). Parameter estimates are plotted for the right GP. \[Color figure can be viewed in the online issue, which is available at [wileyonlinelibrary.com](http://wileyonlinelibrary.com).\]](hipo0025-1262-f4){#fig04} PPI: Effective Connectivity Analysis ------------------------------------ A PPI analysis with the right SN cluster as the seed area (centered at *x* = 7, *y* = −20, *z* = −8), for unexpected versus expected novelty, showed significantly increased connectivity with the left hippocampus (*x* = −36, *y* = −32, *z* = −12; *P* \< 0.001, 8 voxels and *x* = −38, *y* = −12, *z* = −26; *P* \< 0.001, 16 voxels; [Fig. 3b](#fig03){ref-type="fig"}). The connectivity analysis of the right ventral striatum/GP (centered at *x* = 17, *y* = 1, *z* = −5) showed increased connectivity with the right posterior hippocampus (*x* = 35, *y* = −35, *z* = −5; *P* \< 0.001, 8 voxels) and the left PRC (*x* = −31, *y* = −5, *z* = −33; *P* \< 0.001, 40 voxels; [Fig. 4b](#fig04){ref-type="fig"}) for unexpected versus expected familiarity. Both PPI analyses revealed increased connectivity between the two seed areas and the MTL structures, but no other cortical or subcortical effect was found. Experiment 2: Eye-Tracking Results ---------------------------------- To further investigate the behavior that accompanies the processing of unexpected familiar and novel stimuli, we analyzed the pupil responses and fixation patterns produced while participants processed these stimuli. In the eye-tracking experiment (Experiment 2), familiar and novel stimuli in the two tasks were categorized with a mean accuracy of 0.81 (SD = 0.07) and 0.82 (SD = 0.10). Consistent with the fMRI experiment, the detection of expected stimuli was accompanied by longer RTs than unexpected ones as indicated by the main effect of expectation (*F*~1,36~ = 28.87, *P* \< 0.001; familiar~expected~: *M* = 1,562 ms, SD = 229 ms; familiar~unexpected~: *M* = 1,434 ms, SD = 315 ms; novel~expected~: *M* = 1,591 ms, SD = 261 ms; novel~unexpected~: *M* = 1,394 ms, SD = 228 ms). The stimulus by expectation ANOVA of the peak pupil dilation produced a significant main effect of stimulus (*F*~1,36~ = 4.99, *P* = 0.03), showing that the detection of familiar stimuli produced more elevated pupil dilations than did the detection of novel stimuli. The main effect of expectation was not significant (*F* \< 1); however, a significant interaction between stimulus and expectation (*F*~1,36~ = 11.16, *P* = 0.002) showed that unexpected familiar stimuli produced larger pupil dilations than expected familiar stimuli (*t*~36~ = 2.21, *P* = 0.03). Similarly, unexpected novel produced larger pupil dilations than expected novel stimuli (*t*~36~ = −2.56, *P* = 0.015; [Fig. 5a](#fig05){ref-type="fig"}). The analysis of the number of fixations did not yield any significant main effect (stimulus: *F*~1,36~ = 1.12, *P* = 0.30; expectation: *F* \< 1), but produced a significant interaction between the two factors (*F*~1,36~ = 6.67, *P* = 0.014; [Fig. 5b](#fig05){ref-type="fig"}), revealing that unexpected novel stimuli were accompanied by more fixations than expected novel stimuli (*t*~36~ = 2.80, *P* = 0.008). The same pattern was also found for unexpected versus expected familiar stimuli, but this difference did not reach significance (*t*~36~ = 1.27, *P* = 0.21). ![Pupil dilation (a) and fixation patterns (b) for expected and unexpected familiar and novel stimuli in Experiment 2. Error bars show one standard error of the mean; \**P* \< 0.05; \*\**P* \< 0.01.](hipo0025-1262-f5){#fig05} Experiment 3: Subsequent Memory for Unexpected Stimuli ------------------------------------------------------ Finally, we investigated the extent to which the patterns of midbrain/striatal--MTL connectivity (from Experiment 1) and the eye-tracking patterns that accompanied unexpected stimuli (from Experiment 2) are associated with any memory advantages for unexpected stimuli versus expected ones. In a separate experiment (Experiment 3), participants completed the same experimental procedures as in the previous experiments. Critically, their memory for the originally expected and unexpected familiar and novel stimuli was tested in a surprise recognition memory task at the end of the session (Materials and Methods section). In this task, participants recognized old and new stimuli (from across the original FT and NT conditions) with a mean accuracy of 0.88 (SD = 0.05) and a false alarm (FA) rate of 0.22 (SD = 0.15), giving an overall memory performance (hit rate − FA rate) of 0.66 (SD = 0.19), which was significantly above chance (*t*~26~ = 18.31, *P* \< 0.001). Familiarity *d*′ scores and recollection *d*′ scores for originally expected and unexpected stimuli were entered into a 2 × 2 ANOVA separately for originally familiar and novel stimuli. Both ANOVAs showed a significant expectation by response interaction for novel (*F*~1,26~ = 21.32, *P* \< 0.001) and familiar stimuli (*F*~1,26~ = 12.32, *P* = 0.002). As summarized in Table[1](#tbl1){ref-type="table"}, processing unexpected familiar stimuli produced subsequently higher levels of recollection accuracy than processing expected familiar stimuli (*t*~26~ = −4.15, *P* \< 0.001), whereas there was no difference in the familiarity (*d*′) accuracy for unexpected versus expected familiar stimuli (*t*~26~ \< 1). Similarly, processing unexpected novel stimuli produced higher levels of recollection accuracy than expected novel stimuli (*t*~26~ = −4.03, *P* \< 0.001), whereas familiarity *d*′ scores were higher for expected than unexpected novel stimuli (*t*~26~ = 3.16, *P* = 0.004). ###### Subsequent Familiarity and Recollection Performance (d′) for Unexpected and Expected Novel and Familiar Stimuli Response Unexpected Expected ------------------ -------------- -------------------------------------------------- ------------- Novel stimuli Recollection 0.53 (0.08)[\*\*\*](#tf1-1){ref-type="table-fn"} 0.11 (0.10) Familiarity 1.17 (0.08) 1.30 (0.08)[\*\*](#tf1-2){ref-type="table-fn"} Familiar stimuli Recollection 0.72 (0.11)[\*\*\*](#tf1-1){ref-type="table-fn"} 0.29 (0.16) Familiarity 1.14 (0.12) 1.11 (0.13) *Note*: The categorization of stimuli as novel and familiar refers to their status in the main recognition task. Familiarity and recollection responses are obtained from the surprise subsequent memory test. Numbers in the parentheses are standard errors of the mean. *P*\<0.001. *P*\<0.01. DISCUSSION ========== The findings from this research expand our understanding of the nature of the behavioral conditions that might trigger increased connectivity between striatal/midbrain and MTL regions during the detection of novelty and the effect of this on later memory. In addition, the research illustrates the role of visual exploration (eye movements) in the facilitation of encoding during the detection of contextual novelty. The main findings can be summarized as follows. First, we confirmed that striatal and midbrain structures (along with some cortical areas, [Fig. 2](#fig02){ref-type="fig"}) responded to *contextual* novelty for both familiar and novel stimuli at retrieval. Critically, the detection of unexpected novel versus expected novel stimuli elicited increased activity in the SN/VTA, whereas the detection of unexpected familiar versus expected familiar stimuli was associated with the activity in the striatum (predominantly within the GP and putamen). In addition, these regions displayed increased effective connectivity during the detection of unexpected versus expected stimuli, with important memory formation regions within the MTL; the hippocampus (connecting with both SN/VTA and GP) and the perirhinal cortex (connecting with GP). Finally, a small set of cortical regions (involving frontal and parietal areas) was found to be sensitive to unexpected versus expected familiarity, but no such cortical effect was found for novelty. In Experiment 2, we showed that the detection of unexpected stimuli (familiar or novel) was associated with greater pupillary dilation and a greater number of fixations, denoting enhanced visual exploration and processing of the unexpected information. Both the fMRI connectivity and the eye-tracking findings suggest that enhanced learning during the detection of unexpected familiar and novel stimuli might be taking place. This hypothesis was confirmed in the final experiment (Experiment 3), which revealed that both unexpected familiar and unexpected novel stimuli were later remembered with higher recollection accuracy than expected stimuli. Connectivity Between Midbrain/Striatum and the Hippocampus is Modulated by Contextual Novelty --------------------------------------------------------------------------------------------- According to an influential model proposed by Lisman and Grace ([@b29]), a functional loop between the hippocampus and the SN/VTA underlies long-term memory encoding of new information. According to this view, a novelty signal generated in the hippocampus travels via subcortical connections to the midbrain where it stimulates the VTA neurons. The dopaminergic cells in SN/VTA release dopamine, which converges in the hippocampus resulting in the enhancement of long-term potentiation and leads to the construction of new memories (for an extension of this model, see Lisman et al., [@b30]). Dopaminergic neurons within the SN/VTA and the ventral/dorsal striatal structures have been linked to positive reward prediction errors (D\'Ardenne et al., [@b7]), reward learning (Schultz, [@b44]), the processing of unexpected information (Joshua et al., [@b18]), as well as to the facilitation of the learning of new information (Packard and Knowlton, [@b35]; Lisman and Grace, [@b29]; Lisman et al., [@b30]). Furthermore, SN/VTA has been found to respond to novel stimuli independent of their reward value (Schott et al., [@b43]), whereas coactivation with the hippocampus has been reported as facilitating long-term memory formation for rewarding or highly motivational stimuli (Wittmann et al., [@b54],[@b55]; Adcock et al., [@b1]; Bunzeck and Duzel, [@b3]; Shohamy and Wagner, [@b48]). The findings presented here further extend these observations by showing that the activation of SN/VTA and the striatum is not driven at all by the absolute novelty of a stimulus, but by the contextual novelty of both novel and familiar stimuli. Contextual expectation or prediction of novel stimuli has been reported before as modulating the amplitude and the topography of the P3a novelty ERP component (Cycowicz and Friedman, [@b6]) and as playing a pivotal role in determining the magnitude of novelty-related activity in the MTL (Dudukovic and Wagner, [@b8]; Bunzeck et al., [@b4]). However, this is the first time that such a mechanism is revealed in the SN/VTA. In addition, this study illustrates that contextual expectation modulates the *connectivity* between the midbrain or the striatum, and the hippocampus. Specifically, the *effective* connectivity analysis identified a close coupling between midbrain/striatum and MTL structures when unexpected information is detected. *Intrinsic connectivity* between the SN/VTA and the hippocampus has recently been found in a resting-state fMRI study (Kahn and Shohamy, [@b22]), denoting a functional coupling between the two structures. However, an important question, arising from the current models of midbrain--hippocampal crosstalk (Lisman and Grace, [@b29]; Lisman et al., [@b30]), asks under what conditions such a coupling is utilized to boost memory formation for novel or even familiar information. The results of the current PPI connectivity analysis suggest that the close coupling between the dopaminergic midbrain and the hippocampus is modulated by stimulus expectancy and is especially responsive to *contextual* novelty. This means that the suggested hippocampal--SN/VTA loop (Lisman and Grace, [@b29]) becomes strengthened under conditions when unexpected information is detected and less so when predictable or expected information is detected (whether novel or familiar). Indeed, expected novel stimuli did not yield midbrain activity when contrasted with expected familiar stimuli. Furthermore, a contrast between unexpected familiar stimuli and expected novel stimuli resulted in significant activity in the SN/VTA (*x* = 2, *y* = −27, *z* = −22, cluster FWE-corrected). This indicates that activity in the SN/VTA as a response to the detection of *contextual* novelty even for familiar stimuli exceeds any SN/VTA response to the absolute novelty of a stimulus alone. It should be noted here that while a few previous studies (Schott et al., [@b43]; Adcock et al., [@b1]; Wittmann et al., [@b55]; Shohamy and Wagner, [@b48]) have reported concurrent activation of the hippocampus and the midbrain/striatal system for novel stimuli, especially when these are contrasted with familiar stimuli, coactivation indicates activity synchronization between regions, it does not denote functional coupling (or indeed effective connectivity) between the two structures. In contrast, a PPI (as illustrated here) indicates that the slope of the BOLD response observed in two regions (i.e., SN/VTA and the hippocampus) is modulated by the psychological context (in this case, unexpected vs. expected novelty or familiarity; Materials and Methods section). Although, not precisely identified by the fMRI data, it can be assumed that the character of this midbrain/striatum--hippocampal connectivity is dopaminergic (Eckart and Bunzeck, [@b10]). Therefore, the present findings provide especially strong support for the proposed functional memory circuit linking the dopaminergic midbrain and the hippocampus (Lisman and Grace, [@b29]; O\'Carroll et al., [@b33]; Shohamy and Adcock, [@b47]). Furthermore, this circuit appears to exclusively engage striatal/midbrain and MTL regions without the effect of any other intervening brain structures as revealed by the lack of any significant PPI connectivity of SN/VTA and GP with any other cortical or subcortical region. This makes it highly unlikely that the increased connectivity between the midbrain/striatum and the hippocampus arises from the common influence of a third region. It remains to be established whether there are any functional differences in the contribution made by the ventral striatum (GP) and the SN/VTA to the detection of unexpected incoming information and, critically, whether the nature of their connectivity with the hippocampus is modulated by stimulus type (i.e., nominally novel or familiar). As noted above, SN/VTA and midbrain activity (see also [Fig. 2b](#fig02){ref-type="fig"}) appears to respond to both unexpected novel and unexpected familiar stimuli. In contrast, GP activity selectively signals the detection of unexpected familiar stimuli although this activity most likely also originates from the dopaminergic midbrain. Indeed, the ventral pallidum receives major afferent inputs from the dopaminergic midbrain neurons (Haber et al., [@b14]) and it is very likely that the activity of the dopaminergic neurons in the midbrain results in the modulation of BOLD activity in regions to which these neurons project (Yacubian et al., [@b57]), such as the GP. Therefore, the connectivity patterns between these two regions (GP and SN/VTA) and the hippocampus, during the detection of contextually unexpected stimuli, may suggest two distinct routes within the striatal/VTA--hippocampus loop. More specifically, and based on the connectivity findings from this study, a direct projection between the SN/VTA and the hippocampus appears to be employed during the detection of unexpected novel stimuli. This pathway may be considered a reciprocal one, with the hippocampus signaling the detection of contextual novelty to the VTA and the dopaminergic neurons of the VTA projecting their response back to the hippocampus to trigger encoding. This is consistent with the documented role of the hippocampus in novelty detection (Knight, [@b25]; Köhler et al., [@b26]; Stoppel et al., [@b50]; Kafkas and Montaldi, [@b21]). In contrast, midbrain activity for unexpected familiar stimuli seems to originate from other key memory-related brain regions such as the perirhinal cortex, which showed increased connectivity with the GP for unexpected familiar stimuli, and this signal might project to the hippocampus through the ventral striatum (GP). The selectivity of the deployment of the GP for the processing of unexpected familiar stimuli may reflect the onset of an evaluation process that computes the significance of the incoming familiar information in terms of the need for further encoding by the hippocampus. Therefore, although the hippocampus is not involved in the detection of familiar stimuli (Yonelinas et al., [@b58]; Montaldi et al., [@b32]; Kafkas and Montaldi, [@b20]), it responds to the novelty of the unexpected familiar information through its connectivity with the ventral striatum. Indeed, neurons in the monkey GP have been reported to respond not only to the learning, but critically to the evaluation of the reward value of an action (Pasquereau et al., [@b37]) and are responsible for initiating the transmission of this signal to dopaminergic centers like the SN/VTA (Hong and Hikosaka, [@b16]; see also Tecuapetla et al., [@b51]). Greater Visual Exploration and Learning Facilitation for Unexpected Stimuli --------------------------------------------------------------------------- What is the functional significance of a mechanism whereby midbrain/striatum activation and its connectivity with the hippocampus is selectively strengthened for unexpected stimuli and how is it manifested behaviorally? We hypothesized that midbrain and striatal activations may trigger enhanced visual exploration and processing of unexpected stimuli (see also Düzel et al., [@b9]). It has been shown earlier that less frequent or unanticipated events attract greater pupil dilation even when the cognitive demands of the task are kept minimal (Richer and Beatty, [@b42]; Reinhard and Lachnit, [@b39]). This effect vanishes only when the anticipation of an event plays a less crucial role or when other distinctive features of a stimulus become more salient (Reinhard and Lachnit, [@b40]; Reinhard et al., [@b41]). Indeed, in Experiment 2 we showed that the unexpected familiar and novel stimuli were accompanied by increased pupil responses and that unexpected novel stimuli were characterized additionally by increased numbers of fixations compared to those triggered by expected novel stimuli ([Fig. 2](#fig02){ref-type="fig"}). These findings suggest that the contextually novel stimuli, which are accompanied by striatal/midbrain responses and enhanced connectivity with the hippocampus, are also characterized by enhanced visual exploration and processing. Based on these findings, we hypothesized that the SN/VTA triggered hippocampal activity and the richer visual exploration of the unexpected stimuli would lead to enhanced subsequent memory for those items. Experiment 3 confirmed that unexpected stimuli (whether familiar or novel) were recollected later with higher accuracy than expected familiar and novel stimuli. These findings are consistent with the hypothesis that the role of the enhanced midbrain/striatal--hippocampal connectivity is to facilitate learning for unexpected stimuli (see also Lisman et al., [@b30]) and that this is supported by an enhanced visual exploration of these stimuli (as indicated in Experiment 2). Importantly, as it might be expected from the enhanced connectivity between the midbrain/striatum and the hippocampus, a key region for encoding subsequently recollected events (Ranganath et al., [@b38]; Staresina and Davachi, [@b49]), the memory enhancement for unexpected stimuli was selective to recollection memory with no enhancement of familiarity responses. It should be noted here that the enhanced pupil response and number of fixations characterizing the unexpected stimuli and the ensuing memory benefit cannot be explained by a RT confound. It might have been argued, for example, that the detection of unexpected information is more effortful and thus slower, leading to greater memory for these stimuli owing to the extra time allocated to processing these stimuli. However, in both experiments, the detection of unexpected familiar and novel stimuli was faster than the detection of expected familiar and novel stimuli. Therefore, neither increased effort nor extra processing time can account for the findings in the experiments reported here. The faster RT found for unexpected stimuli relative to expected ones may relate to the differential emphasis put on familiar and novel stimuli within the two retrieval conditions (FT and NT, respectively), or it may result from the engagement in a strength-rating task in the case of target (expected) but not the unexpected stimuli. One limitation of the presented studies, however, is that the fMRI and the eye-tracking data are derived from different experiments. Nevertheless, absolutely matched experimental designs were adopted in both studies with the only difference being that the fMRI data were acquired while participants lay in the MRI scanner, whereas the eye-tracking data were recorded while they were seated in a testing room. We therefore do not believe that this difference could undermine the potential relationship between the results and an intriguing possibility is that visual exploration is the direct product of the enhanced midbrain/striatal activation and connectivity with the hippocampus. Indeed, the activity of SN neurons in the monkey has been found to relate to saccadic eye movements especially when these are directed toward rewarded locations (Handel and Glimcher, [@b15]). More recently, neurons in monkey GP demonstrated synchronized activity with saccadic generation (Shin and Sommer, [@b46]). Therefore, it can perhaps be concluded that the memory facilitation for unexpected stimuli, driven by the hippocampal--midbrain system (Lisman and Grace, [@b29]), is mediated by, or at the very least correlates with, the modulation of eye movement behavior illustrated in Experiment 2. Unexpected Familiarity: a Role for Parietal Cortex -------------------------------------------------- Apart from the striatal/midbrain activations and connectivity patterns with the hippocampus, the analyses also identified areas within the inferior parietal lobe, the occipito-parietal junction, the precuneus, and the middle frontal gyrus ([Fig. 2](#fig02){ref-type="fig"}) that showed a selective response to unexpected versus expected familiar stimuli. The parietal regions, in particular, have been previously found to play a role in identifying unexpected stimuli (O\'Connor et al., [@b34]; Jaeger et al., [@b17]) or stimulus incongruency within a sequence of events (Krebs et al., [@b27]). For example, in a recent study, Jaeger et al. ([@b17]) found that distinct areas within the inferior parietal lobe monitor unexpected novelty and unexpected familiarity and proposed that their function is to orient current attention to the unexpected information. Our findings are consistent with this model although we observed only inferior parietal activations in response to unexpected familiarity, not to unexpected novelty. In addition, our connectivity data did not show any interactions between the parietal regions and the midbrain/striatal--MTL circuitry. Further studies are needed to explore any communication between the midbrain/striatal--hippocampal circuit and the inferior parietal regions. CONCLUSIONS =========== Taken together, these studies show that unexpected familiar and novel stimuli selectively activate striatal and midbrain regions and are also associated with increased visual exploration and enhanced subsequent memory. Consistent with the theoretical model proposed by Lisman and Grace ([@b29]), this memory facilitation is accompanied by increased connectivity between striatal/midbrain and the hippocampus, and the present findings show that this effect is selective to *contextual* novelty (for both familiar and novel stimuli) not related to absolute stimulus novelty. The findings also discriminate between two potential pathways within this system; one for the detection and processing of unexpected familiar stimuli and the other for the detection and processing of unexpected novel stimuli. Overall, the findings support the hypothesis that striatal/midbrain activity and its connectivity with the hippocampus and other medial temporal lobe structures drives and supports exploratory behavior, subsequently leading to the increased recollection of contextually novel events.

{ "pile_set_name": "PubMed Central" }

1. Introduction {#sec1-plants-09-00675} =============== Seed germination is a crucial process, characterized by a series of steps, normally occurring before radicle emergence from the seed coat \[[@B1-plants-09-00675],[@B2-plants-09-00675]\]. Radicle emergence percentage can deeply affect the production and the crop quality since the germination process is a sensitive step for plant growth and it is strictly linked to the seed quality, being the radicle emergence percentage related to seed germination potential \[[@B3-plants-09-00675],[@B4-plants-09-00675],[@B5-plants-09-00675]\]. In fact, during the germination of seeds, different enzymes such as amylases, proteases, and lipases lead to the hydrolysis of reserve substances, producing compounds that are transported to the growing seedlings for their development \[[@B6-plants-09-00675]\]. Sugar beet (*Beta vulgaris* subsp. *vulgaris*) is mainly grown in temperate climates and it is an important industrial biennial root crop, providing about 20% of the world's annual sugar production and representing a source material for bioethanol and animal feed production \[[@B7-plants-09-00675]\]. Sugar beet seed health and quality are crucial for adequate plant growth and they are strictly associated with the productive yield both in terms of quantity and quality of the crop \[[@B8-plants-09-00675]\]. Unfortunately, the productivity of sugar beet is often limited by heterogeneous germination in the field, probably due to the presence of inhibitory substances in the pericarp of the seed as well as pathogen attacks \[[@B9-plants-09-00675],[@B10-plants-09-00675]\]. Moreover, about 35--40% of the sugar beet seeds need to be discarded before sowing in the field because they are defective \[[@B11-plants-09-00675]\]. Finally, after germination, sugar beet seedlings often have to confront different biotic and abiotic stresses, therefore, rapid germination of sugar beet seeds is crucial for plant development and overall yield \[[@B12-plants-09-00675],[@B13-plants-09-00675]\]. For this purpose, different techniques have been adopted in order to increase the physiological potential of sugar beet seeds and their treatment may become an indispensable procedure aiming at the enhancement of the seed vigor and the reduction of the variability in the germination process. Among these, seed priming with water, salicylic acid, or gibberellic acid has been shown to promote sugar beet seed germination as well as enhance sugar beet seedling growth \[[@B10-plants-09-00675]\]. Szajsner et al. \[[@B14-plants-09-00675]\] showed that the seed vigor, germination speed, and the germination time significantly improved by using a magnetic field or a semiconductor laser radiation as pre-sowing treatments of sugar beet seeds. An aqueous extract from microalga *Acutodesmus dimorphus* was successfully used as a seed primer, improving the germination energy of tomato seeds (cv. Roma) and increasing lateral root development and extract concentrations \[[@B15-plants-09-00675]\]. Conversely, other authors found that higher seaweed extract concentrations have caused inhibition of seed germination \[[@B16-plants-09-00675],[@B17-plants-09-00675],[@B18-plants-09-00675]\]. Recently, Barone et al. \[[@B19-plants-09-00675]\] have shown that extracts from microalgae *Chlorella vulgaris* or *Scenedesmus quadricauda* may act as biostimulant in the early stages of sugar beet cultivation, when added after 5 days to the Hoagland growth solution, by improving root and plant growth and modulating gene expression related to the nutrient acquisition in sugar beet. Moreover, these extracts were successfully applied at the root level in the cultivation substrates, showing to exert a biostimulant effect on tomato and lettuce seedlings \[[@B20-plants-09-00675],[@B21-plants-09-00675]\]. Considering that *C. vulgaris* and *S. quadricauda* extract have been successfully applied as a biostimulant to several crops, including sugar beet, the aim of this work was to evaluate whether these microalgae extracts may also be able to positively affect the germination process as a priming treatment for sugar beet seed. Moreover, the concentration-dependent effects of microalgal extracts from *C. vulgaris* or *S. quadricauda* were also evaluated. Therefore, seed germination of *B. vulgaris* was monitored with the aim to calculate several physiological indices, useful to evaluate the effect of pre-soaking seeds with the microalgal extracts. In addition, root morphological traits were also evaluated by using WinRHIZO software in order to evaluate the effectiveness of these priming treatments. 2. Results and Discussions {#sec2-plants-09-00675} ========================== Several germination indices of sugar beet seeds were calculated and monitored (as detailed in the Materials and Methods section) in order to evaluate the effect of the treatments and their concentration dependence, using five amounts of *C. vulgaris* (CVextr) or *S. quadricauda* extracts (SQextr). In all data analysis, a significant effect (*p* \< 0.01) of microalgal extract concentration was observed on all germination variables of *B. vulgaris* seedlings. Similarly, all concentration × microalgal extract interactions were significant (*p* \< 0.01) for all germination parametric indices. Therefore, the experiments were always presented for each type of microalgal extract (*C. vulgaris* or *S. quadricauda*). The percentage of sugar beet seed germination (GP) is reported in [Figure 1](#plants-09-00675-f001){ref-type="fig"}. All the *C. vulgaris* extract-based treatments significantly affected the GP values in comparison to the untreated seeds (C0), 3 days after priming (DAP) ([Figure 1](#plants-09-00675-f001){ref-type="fig"}A). In particular, C2 and C3 *C. vulgaris* extract concentrations increased GP values, showing at 3 DAP a significant increase with respect to the control of 2.8 and 3.8 folds, respectively. At the remaining monitoring times, only C2 and C3 *C. vulgaris* extract concentrations significantly affected GP values in comparison to the untreated seeds, reaching the highest values at 4 DAP, when GP values calculated for C2 and C3 were 1.8 and 2 times higher than those calculated in the control seeds, respectively. The final GP values, calculated 7 DAP, were 1.4 (C2) and 1.5 (C3) times higher than those calculated in the untreated seeds. Conversely, among the different *S. quadricauda* extract concentrations, the effectiveness of the treatment was observed only after 4 DAP ([Figure 1](#plants-09-00675-f001){ref-type="fig"}B). From 4 to 7 DAP, C2 *S. quadricauda* extract concentration positively affected seed germination, by increasing the GP values of around 1.3 times with respect to the untreated seeds. Noteworthy also that C3 concentration significantly increased GP value with respect to the untreated seeds (1.2 folds), limited to 5 DAP ([Figure 1](#plants-09-00675-f001){ref-type="fig"}B). Results reported in [Figure 2](#plants-09-00675-f002){ref-type="fig"} show that at all the monitoring times, mean daily germination (MDG) values calculated for seeds treated with C2 and C3 *C. vulgaris* extract concentrations, were always greater than those calculated for untreated seeds, reaching the highest increase only after 3 DAP (2.8 and 3.8 times higher than those calculated in control seeds for C2 and C3, respectively) ([Figure 2](#plants-09-00675-f002){ref-type="fig"}A). Among *S. quadricauda* extract concentrations, at all monitoring times except for 3 DAP, the highest MDG values, if compared to that calculated in untreated seeds, were recorded in seeds treated with C2, showing always an increase of 1.3 times ([Figure 2](#plants-09-00675-f002){ref-type="fig"}B). These results suggest that treatments exerted different effects on GP and MDG values, strictly related to the microalgae species as well as the extract concentrations. Interestingly, the positive effect on GP values, mainly observed by using C2 and C3 concentrations of CVextr and C2 concentration of SQextr, were higher than those obtained by Szajsner et al. \[[@B14-plants-09-00675]\], who found GP values around 1 fold higher than controls by treating sugar beet seeds with a magnetic field or laser radiation. Therefore, these results indicate that these priming treatments, in particular C2 and C3 CVextr, may be a promising alternative practice aiming to enhance seed germination performance and increase GP and MDG indices, strictly related to the variability of this important physiological process. Interestingly, the lower and higher extract concentrations (C1, C4, and C5) do not significantly affect the GP and MDG germination indices, as these values were always similar to the respective controls. These data are in accordance with that reported by Santos et al. \[[@B22-plants-09-00675]\], who found that the application of algae *Ascophyllum nodosum* extract-based biostimulants in ornamental sunflower requires an optimal concentration to increase germination indices. Conversely, an aqueous extract from microalga *Acutodesmus dimorphus* behaved as an effective primer on tomato seeds, at increasing extract concentrations \[[@B15-plants-09-00675]\]. Indeed, natural biostimulants may contain various biologically active compounds which may determine concentration-dependent effects, making crucial the testing of a broad range of concentrations \[[@B23-plants-09-00675]\]. Moreover, depending on the extract type and concentration applied, natural biostimulants may elicit different responses in treated plants, being sometimes also potentially phytotoxic \[[@B24-plants-09-00675]\]. Our data suggest that the microalgal extracts did not negatively affect the germination process at lower and higher concentrations, showing no phytotoxic effect, as supported by Ronga et al. \[[@B25-plants-09-00675]\], who found no phytotoxicity in an aqueous *C. vulgaris* microalgal extract on cress. Interestingly, the two extracts showed an evident divergent effect on germination indices probably due to the difference in extract compositions \[[@B19-plants-09-00675]\]. Indeed, the two extracts showed a different composition in terms of organic carbon distribution ([Supplementary Table S1](#app1-plants-09-00675){ref-type="app"}) and element composition ([Supplementary Table S2](#app1-plants-09-00675){ref-type="app"}), as reported in Barone et al. \[[@B19-plants-09-00675]\]. In particular, the degree of hydrophobicity for humic substances determined according to Baglieri et al. \[[@B26-plants-09-00675]\], resulted in being much higher for CVextr (6.1) than that calculated for SQextr (3.8), showing CVextr to be more apolar than SQextr ([Supplementary Table S1](#app1-plants-09-00675){ref-type="app"}). These data suggest that the different extract compositions in organic carbon distribution as well as polarity, may be strictly related to the different effect on GP and MDG germination indices, as confirmed by Piccolo et al. \[[@B27-plants-09-00675]\], who found a relationship between the structure of the formulate and its bioactivity in humic substances. Data reported in [Figure 3](#plants-09-00675-f003){ref-type="fig"} show that the two concentrations C3 and C2 of *C. vulgaris* extract significantly increased the germination indices (GI) (1.7 and 1.6 times, respectively), germination energy (GE) (3.8 and 2.8 times, respectively), speed of emergence (SE) (2.4 and 2 times, respectively) and coefficient of the rate of germination (CRG) (around 1.1 times for both concentrations) ([Figure 3](#plants-09-00675-f003){ref-type="fig"}A, B, C and D), if compared to the control; on the other hand, mean germination time (MGT) and T~50~ ([Figure 3](#plants-09-00675-f003){ref-type="fig"}E and F) were significantly reduced (around 1.1 times for both indices and concentrations). As regards to the T~50~ index, all the *C. vulgaris* extract concentrations significantly reduced the time required for 50% germination ([Figure 3](#plants-09-00675-f003){ref-type="fig"}F). These findings support the hypothesis that the treatment with *C. vulgaris* extract at all the concentrations tested may be a very useful priming treatment in order to improve seed germination performance, in terms of reduction of the time required to obtain 50% seed germination. Indeed, it is well known that the higher the GI, GE, SE, and CRG values, the higher the positive effect on seed germination \[[@B28-plants-09-00675]\]. In contrast, the lower the T~50~ and MGT values with respect to the control, the lower the inhibition on seed germination \[[@B28-plants-09-00675]\]. As regards the *S. quadricauda* extract, the concentration effects on germination indices were less evident and differences among concentrations were not always significant ([Figure 3](#plants-09-00675-f003){ref-type="fig"}). Nevertheless, the concentration C2 positively affected GI index (1.4 times) with respect to the control, and the CRG index values were significantly higher in seeds treated with C3 and C5 concentrations than those calculated for untreated seeds. Conversely, MGT values were significantly reduced by the treatments with C3 and C5 concentrations (around 1.1 for both amounts of extract). All other germination indices were not significantly affected by *S. quadricauda* extract at all the tested concentrations ([Figure 3](#plants-09-00675-f003){ref-type="fig"}). These data show that *S. quadricauda* extract also positively affected seed germination, although to a lesser extent than *C. vulgaris* extract. In detail, the calculated GI, GE, and SE values were almost always lower than those relative to the corresponding concentrations of *C. vulgaris*, moreover, T~50~ values were not positively affected by *S. quadricauda* extract, showing values always similar to those calculated for the control ([Figure 3](#plants-09-00675-f003){ref-type="fig"}F). It is noteworthy to underline that these differences between the two algal extracts were significant only for the C2, C3, and C4 concentrations with regard to GI, for the C2 and C3 concentrations with regard to GE and SE indices, for C2 concentration with regard to CRG and MGT, and for C3 concentration with regard to the T~50~ index ([Figure 3](#plants-09-00675-f003){ref-type="fig"}). These results show that CVextr seems to be more effective than SQextr, with C2 and C3 the optimal concentrations, although a higher amount of CVextr did not negatively affect seed germination. These findings are supported by Ronga et al \[[@B25-plants-09-00675]\], who found that a two-fold concentration (around 25 mgC~org~/L) of an aqueous *C. vulgaris* microalgal extract did not show phytotoxicity effect, by measuring the GI index, on a sensitive species to phytotoxic compounds such as cress. Interestingly, from a physiological point of view, several authors \[[@B29-plants-09-00675],[@B30-plants-09-00675],[@B31-plants-09-00675],[@B32-plants-09-00675]\] have reported that the faster emergence in sugar beet showed an enormous influence on plant characteristics, showing a higher dry matter weight compared to the plants emerging later. Among these authors, Podlaski et al. \[[@B32-plants-09-00675]\] demonstrated that the time of emergence was the strongest factor influencing plant weight in sugar beet during harvest season. Analysis of morphological data provided always a significant effect (*p* \< 0.01) of microalgal extract concentrations on all morpho-biometric parameters (length, surface area, mean root diameter, root volume, tips, root 0.000 \< L \< 0.500 and root 0.500 \< L \< 1.000) of *B. vulgaris* seedlings. Since concentration × microalgal extract interactions were significant for all parametric variables, the data were presented for each microalgal extract ([Table 1](#plants-09-00675-t001){ref-type="table"} and [Table 2](#plants-09-00675-t002){ref-type="table"}). The positive effect of C2 and C3 *C. vulgaris* extract concentrations were also confirmed by morphological parameters ([Table 1](#plants-09-00675-t001){ref-type="table"}). In particular, all *C. vulgaris* extract concentrations significantly increased all root morphological parameters, both after 5 and 7 DAP, being the C2 and C3 CVextr the highest values, in comparison to the control ([Table 1](#plants-09-00675-t001){ref-type="table"}). The positive effect of C2 and C3 *C. vulgaris* extract concentrations were also observed at the root volume level, whereas mean root diameter values were unaffected or reduced by the treatments ([Table 1](#plants-09-00675-t001){ref-type="table"}). Similarly, all root morphological traits were positively affected by *S. quadricauda* extracts too ([Table 2](#plants-09-00675-t002){ref-type="table"}), although to a lesser extent than *C. vulgaris*. Noteworthy, the highest values of morpho-biometric parameters were reached using C2 and C3 *S. quadricauda* extract concentrations, whereas each amount of the SQextr did not positively affect the mean root diameter ([Table 2](#plants-09-00675-t002){ref-type="table"}), as it was already observed using *C. vulgaris* extract ([Table 1](#plants-09-00675-t001){ref-type="table"}). These results are in accordance with Barone et al. \[[@B19-plants-09-00675]\], who found that by applying to the hydroponic solution 1 and 2 mg Corg/L of the two microalgal extracts (*C. vulgaris* or *S. quadricauda*), root apparatus of sugar beet seedlings was positively affected by increasing total root length, root surface area, and the number of root tips, whereas the average diameter and the volume of roots were not affected by the treatments. The effect on root morphology may be related to the degree of hydrophobicity of extracts ([Supplementary Table S1](#app1-plants-09-00675){ref-type="app"}), the latter being closely related to the increase of root growth \[[@B27-plants-09-00675]\]. *C. vulgaris* extract showed a degree of hydrophobicity value (6.1) greater than those observed for humic substances of a different origin, ranging between 0.61 and 4.75 \[[@B33-plants-09-00675]\], whereas *S. quadricauda* extract showed an intermediate value (3.8). In particular, methoxilic groups, aryl groups, and carboxylic acids seem to be involved in the bioactivity of natural biostimulant substances, and often related to hormone-like compounds \[[@B27-plants-09-00675]\]. Therefore, both extracts, on the basis of their characterization, seem to be perfectly compatible with the effect observed on the morphological traits in sugar beet roots \[[@B34-plants-09-00675]\]. Moreover, the performances obtained in the present study seem to be higher than those obtained by Szajsner et al. \[[@B14-plants-09-00675]\], who treated sugar beet seeds with a magnetic field or laser radiation. In particular, Szajsner et al. \[[@B14-plants-09-00675]\] achieved an increase of the seedling length of 1.5 times with respect to the control after 4 days from the pre-sowing treatments, whereas the treatment with C3 *C. vulgaris* extract, induced an increase of the seedling length of around 6 times with respect to the untreated seeds at 5 days after priming treatment ([Table 1](#plants-09-00675-t001){ref-type="table"}). Finally, SVI values were also calculated both at 5 and 7 days after priming treatments ([Figure 4](#plants-09-00675-f004){ref-type="fig"}). All CVextr treatments positively affected SVI values, in accordance with other calculated germination indices. In particular, the C3 *C. vulgaris* extract resulted in being the most performant treatment, determining an increase of seedling vigor index (SVI) of around 9 and 3 times higher than the controls after 5 and 7 days, respectively. As regards to *S. quadricauda* extract, C2 induced the greatest increase of 8 and 3 times higher than the controls after 5 and 7 days, respectively ([Figure 4](#plants-09-00675-f004){ref-type="fig"}). These results suggest that the use of these microalgae extract as priming treatment, may be a good alternative to other priming methods adopted for sugar beet, according to Islam et al. \[[@B28-plants-09-00675]\], who observed that the higher the seedling vigor index (SVI) value, the higher the positive effect on seed germination. Moreover, in accordance with Ugena et al. \[[@B35-plants-09-00675]\], pre-sowing treatment with different biostimulant compounds, aiming to increase the vigor of seedlings, represents an innovative alternative to cope with different kinds of stresses. Comprehensively, all these results taken together suggest that best results, in terms of the germination process as well as root morphological traits, were reached by using the concentrations C2 and C3 of *C. vulgaris* extract, showing that this microalgal extract, besides exerting a biostimulant effect when added to the growth medium of sugar beet seedlings \[[@B19-plants-09-00675]\], may also be used as a priming method positively affecting the sugar beet seed germination. 3. Materials and Methods {#sec3-plants-09-00675} ======================== 3.1. Microalgae Cultivation and Extract Preparation {#sec3dot1-plants-09-00675} --------------------------------------------------- The microalgae used in this study were *C. vulgaris* (Beijerinck, CCAP 211/11C) and *S. quadricauda* (isolated from an algal company raceway pond, located in Borculo, Gelderland, the Netherlands in 2011). They were obtained by and maintained in the algal collection of the Department of Agriculture, Food and Environment (Di3A) (University of Catania, Italy) as described in Baglieri et al. \[[@B36-plants-09-00675]\]. Microalgal growth was conducted in 250 mL flask containing 150 mL of sterile BG11 culture medium \[[@B37-plants-09-00675]\] at pH 8.4, incubated on a mechanical shaker (100 rpm) at 25--30 °C, illuminated by a 3500-lx, average photon flux (PPF) 100 μmol photons m^−2^ s^−1^ light source (PHILIPS SON-T AGRO 400) with a 12 h photoperiod for 30 days in a growth chamber and aerated by pumps with 20 L h^--1^ 1.5% CO~2~ \[[@B38-plants-09-00675]\]. Microalgal biomasses were harvested by centrifugation (at 5000 rpm for 15 min), washed with distilled water (up conductivity \< 200 μS cm^−1^), and freeze-dried as described in Puglisi et al. \[[@B39-plants-09-00675]\]. Microalgal extract stock solutions (referred to as CVextr and SQextr) were prepared as described in Barone et al. \[[@B19-plants-09-00675]\]. Briefly, microalgae cells were collected and centrifuged at 5000 rpm for 15 min and the final pellets obtained from each microalgal biomass were added to methanol to lyse the cell wall in order to obtain the intracellular extracts. After centrifugation and evaporation of the organic solvent, the extracts were freeze-dried and collected with distilled water to obtain the microalgal extract stock solution. The complete characterization of extracts was reported in Barone et al. \[[@B19-plants-09-00675]\]. 3.2. Plant Material and Experimental Conditions {#sec3dot2-plants-09-00675} ----------------------------------------------- The sugar beet variety used in this study is the hybrid "Shannon" provided by the Department of Agronomy, Food, Natural Resources, Animals and Environment (DAFNAE) of the University of Padova (Italy) \[[@B40-plants-09-00675]\]. Seeds were soaked in 76% ethanol for 5 min, rinsed with sterilized water, and placed on distilled water moistened filter paper. The treatments were performed by diluting different amounts of microalgal extracts in distilled water, used to moisten the filter paper. For each microalgal extract, five different concentrations were tested and they were calculated on the basis of the extract organic carbon (C~org~) content: C1 = 0,1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Control samples using untreated seeds (C0 = 0 mg C~org~/L) were routinely performed. Germination was carried out in a growth chamber in the dark at 25 °C. Sugar beet seeds were considered germinated when a radicle of at least 2 mm emerged. The germinated seeds were counted and monitored daily for 7 days, after this time no germination, even of those not yet germinated, was detected. The experimental procedure was repeated twice in a complete randomized block design and for each treatment, four replicates consisting of 100 seeds were tested according to the methods of the International Rules for Seed Testing \[[@B41-plants-09-00675],[@B42-plants-09-00675]\]. 3.3. Germination Indices {#sec3dot3-plants-09-00675} ------------------------ In order to evaluate the effect of microalgal extract treatments, several germination indices were calculated as detailed below. The germination percentage (GP) was calculated for each treatment as a percentage of total germinated seeds after 3, 4, 5, and 7 days after priming (DAP):$${GP} = \left( {{{number}~{of}~{germinated}~{seeds}}/{{number}~{of}~{total}~{seeds}~{for}~{bioassay}}} \right) \times 100$$ The mean daily germination (MDG), representing the mean number of seeds germinated per day, was calculated at 3, 4, 5, and 7 DAP \[[@B43-plants-09-00675]\]:$${MDG} = {GP}/t$$ where GP is the germination percentage, and t is the number of DAP. The germination index (GI), also known as mean germination rate or rate of Maguire \[[@B44-plants-09-00675]\], is a measure assigning the maximum arithmetic weight to seeds that germinate at the first days of count and less weight to those germinating later. GI was calculated as follows:$${GI} = \left\lbrack {{{number}~{of}~{germinated}~{seeds}}/{{days}~{of}~{first}~{count}}} \right\rbrack + \ldots + \left\lbrack {{{number}~{of}~{germinated}~{seeds}}/{{days}~{of}~{final}~{count}}} \right\rbrack$$ The mean germination time (MGT) was calculated according to Soltani et al. \[[@B45-plants-09-00675]\] as follows:$${MGT} = \sum\left( {n,t} \right)/\sum n$$ where n is the number of newly germinated seeds at time t. The germination energy (GE) was calculated according to Ruan et al. \[[@B46-plants-09-00675]\] as follows: The speed of emergence (SE), was calculated according to Islam et al. \[[@B28-plants-09-00675]\] as follows: The coefficient of the rate of germination (CRG) was calculated according to Chiapusio et al. \[[@B47-plants-09-00675]\] as follows:$${CRG} = \left\lbrack {\left\lbrack {N_{1} + N_{2} + \ldots + N_{n}} \right\rbrack/\left\lbrack {N_{1} \times T_{1}} \right\rbrack + \left\lbrack {N_{2} \times T_{2}} \right\rbrack + \ldots + \left\lbrack {N_{n} \times T_{n}} \right\rbrack} \right\rbrack \times 100$$ where N~1~, N~2~, ..., N~n~ are the number of germinated seeds on time T~1~, T~2~, ..., T~n~. The time required for 50% germination (T~50~), was calculated according to Coolbear et al. \[[@B48-plants-09-00675]\] as follows:$$T_{50} = t_{i} + \left\lbrack {\left\lbrack {\left\lbrack {N/2} \right\rbrack - n_{i}} \right\rbrack \times \left\lbrack {t_{i} - t_{j}} \right\rbrack} \right\rbrack/(n_{i} - n_{j})$$ where N is the final number of germinated seeds, n~i~ and n~j~ the cumulative numbers of seeds germinated by adjacent counts at times t~i~ and t~j~. Finally, the seedling vigor index (SVI) was calculated at 5 and 7 DAP, according to Noorhossein et al. \[[@B43-plants-09-00675]\] as follows: 3.4. Root Morphological Analysis {#sec3dot4-plants-09-00675} -------------------------------- Root morphological analysis was performed by using WinRHIZO software (Regent Instruments) and total root length, root surface area, mean root diameter, root volume, total number of root tips, lateral root (0.000 \< L \< 0.500 and 0.500 \< L \< 1.000) were determined by computerized scanning (STD 1600, Regent Instruments, Canada) at 5 and 7 days from sowing \[[@B49-plants-09-00675]\]. 3.5. Statistical Analysis {#sec3dot5-plants-09-00675} ------------------------- Data on performances of microalgal extracts from the repeated experiment were analyzed by using the Statistica package software (version 10; Statsoft Inc., Tulsa, OK, USA). The arithmetic means of parameters were calculated, by averaging the values determined for the single replicates of each treatment. Percentage data were transformed into the arcsine (sin^−1^ square root x) prior to the analysis of variance (ANOVA). Initial analyses were performed by calculating associated F and P values to evaluate whether the effects of a single factor (concentration) and microalga × concentration interactions were significant. In the post-hoc analysis, the corresponding mean values of all parameters were subsequently separated by Fisher's least significant difference test (*p* = 0.05). Untransformed arithmetic means of parameters are presented in the figures and tables. 4. Conclusions {#sec4-plants-09-00675} ============== This study leads us to employ microalgae extracts as a priming agent in order to improve the germination process of an important industrial crop such as sugar beet. The novelty of this work consists of the successful application of *C. vulgaris* and *S. quadricauda* as pre-sowing treatments, resulting in the first work in which microalgae extracts are employed as a priming method for sugar beet seed germination. Microalgae extracts, especially *C. vulgaris*, was found to improve all calculated germination indices, the root traits as well as the seedling vigor of sugar beet, putatively promoting the further nutrient acquisition and plant growth. Among the tested extract concentrations, 1 mg C~org~/L and 2 mg C~org~/L of *C. vulgaris* extract revealed to be the best priming treatments, exerting a positive effect both on the germination process and morphological traits of sugar beet seedlings. Although further investigations should be performed, based on our findings, *C. vulgaris* extract may represent a promising practice to increase the physiological potential of sugar beet seeds. The following are available online at <https://www.mdpi.com/2223-7747/9/6/675/s1>, Table S1: Distribution of C intensity of 13C NMR (%) of *Chlorella vulgaris* (CVextr) and *Scenedesmus quadricauda* (SQextr) extracts; Table S2: Element composition (%) of *Chlorella vulgaris* (CVextr) and *Scenedesmus quadricauda* (SQextr) extracts. ###### Click here for additional data file. Conceptualization, A.B.; methodology, A.B. and P.S.; validation, A.B. and I.P.; formal analysis, A.V.; investigation, V.B. and F.F.; resources, A.B.; data curation, I.P. and A.V.; writing---original draft preparation, I.P.; writing---review and editing, A.B., I.P. and A.V.; visualization, I.P. and A.L.; supervision, A.B.; project administration, A.B.; funding acquisition, A.B. All authors have read and agreed to the published version of the manuscript. Authors thank the grants which supported this work: 'Programma ricerca di Ateneo UNICT 2020-22 linea 2′---University of Catania (Italy); "STARTING GRANT 2020"---University of Catania (Italy); Fondi di Ateneo 2020--2022 - University of Catania (Italy), linea Open Access. The authors declare no conflicts of interest. ![Germination percentage (GP) of sugar beet seeds treated with different concentrations of *C. vulgaris* (**A**) or *S. quadricauda* (**B**) extracts, for each day after priming (DAP). C0 = 0 mg C~org~/L; C1= 0.1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Data (± standard error bar) are the means of three replicates (each formed by 100 seeds). Columns within each sampling point followed by the same letters are not significantly different according to Fisher's least significant difference test (α = 0.05). The absence of letters above the columns shows the lack of significant differences.](plants-09-00675-g001){#plants-09-00675-f001} ![Mean daily germination (MDG) of sugar beet seeds treated with different concentrations of *C. vulgaris* (**A**) and *S. quadricauda* (**B**) extracts, for each day after priming (DAP). C0 = 0 mg C~org~/L; C1 = 0.1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Data (± standard error bar) are the means of three replicates (each formed by 100 seeds). Columns within each sampling point followed by the same letters are not significantly different according to Fisher's least significant difference test (α = 0.05). The absence of letters above the columns shows the lack of significant differences.](plants-09-00675-g002){#plants-09-00675-f002} ![Germination index (GI) (**A**), germination energy (GE) (**B**), speed of emergence (SE) (**C**), coefficient of the rate of germination (CRG) (**D**), mean germination time (MGT) (**E**) and time required for 50% germination (T~50~) (**F**) of sugar beet seeds treated with different concentrations of *C. vulgaris* (CV) and *S. quadricauda* (SQ) extracts. C0 = 0 mg C~org~/L; C1 = 0.1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Data (± standard error bar) are the means of three replicates. The same colored columns representing 5 concentrations of each algal extract followed by the same letters are not significantly different according to Fisher's least significant difference test (α = 0.05). The absence of letters above the columns shows the lack of significant differences. The presence of an asterisk (\*) within each concentration show a significant difference between the two algal extracts.](plants-09-00675-g003){#plants-09-00675-f003} ![Seedling vigor index (SVI) after 5 and 7 days of sugar beet treated with different concentrations of *C. vulgaris* (**A**) and *S. quadricauda* (**B**) extracts. C0 = 0 mg C~org~/L; C1 = 0.1 mg C~org~/L; C2 = 1 mg C~org~/L; C3 = 2 mg C~org~/L; C4 = 5 mg C~org~/L; C5 = 10 mg C~org~/L. Data (± standard error bar) are the means of three replicates. The same colored columns representing 5 concentrations of each algal extract followed by the same letters are not significantly different according to Fisher's least significant difference test (α = 0.05). The presence of an asterisk (\*) within each concentration indicates a significant difference between the two algal extracts.](plants-09-00675-g004){#plants-09-00675-f004} plants-09-00675-t001_Table 1 ###### Performance of different dosages of *Chlorella vulgaris* extract on different morpho-biometric parameters on *Beta vulgaris* seedlings. Data derived from two repeated experiments. Standard error of the means = SEM, *n* = 4. Means followed by different letters within the column are significantly different according to Fisher's least significant difference test (α = 0.05). Length (cm) Surface Area (cm^2^) Mean Root Diameter (mm) Root Volume (cm^3^) Tips Number Root Number (0.000 \< L \< 0.500) Root Number (0.500 \< L \< 1.000) -------- -------------- ---------------------- ------------------------- --------------------- --------------- ----------------------------------- ----------------------------------- ----------------- --------------- -------------- -------------- -------------- -------------- ----------------- **C0** 0.8 ± 0.1 d 3.0 ± 0.2 c 0.2 ± 0.01 e 0.3 ± 0.01 c 0.6 ± 0.01 ab 0.6 ± 0.02 a 0.003 ± 0.0 d 0.01 ± 0.0 ab 1.5 ± 0.07 e 2.9 ± 0.1 e 0.2 ± 0.04 b 1.4 ± 0.1 d 0.6 ± 0.02 c 1.4 ± 0.1 a **C1** 2.8 ± 0.2 c 4.5 ± 0.2 b 0.5 ± 0.05 d 0.6 ± 0.05 b 0.6 ± 0.02 b 0.4 ± 0.04 bc 0.01 ± 0.0 c 0.01 ± 0.0 bc 3.0 ± 0.14 d 5.04 ± 0.2 c 3.0 ± 1.1 a 3.2 ± 0.2 bc 1.2 ± 0.2 b 1.2 ± 0.2 a--c **C2** 4.7 ± 0.1 a 6.7 ± 0.1 a 0.8 ± 0.01 b 0.9 ± 0.03 a 0.6 ± 0.01 b 0.4 ± 0.02 c 0.01 ± 0.0 ab 0.01 ± 0.0 a 3.7 ± 0.22 c 8.0 ± 0.2 b 2.3 ± 0.6 a 5.3 ± 0.04 a 2.1 ± 0.2 a 1.3 ± 0.06 a--c **C3** 5.1 ± 0.1 a 6.4 ± 0.2 a 0.9 ± 0.02 a 0.9 ± 0.01 a 0.6 ± 0.02 b 0.4 ± 0.01 c 0.01 ± 0.0 a 0.01 ± 0.0 a--c 4.6 ± 0.18 a 11.2 ± 0.2 a 2.7 ± 0.4 a 5.3 ± 0.2 a 2.3 ± 0.2 a 1.05 ± 0.1 bc **C4** 3.2 ± 0.2 bc 4.2 ± 0.1 b 0.6 ± 0.02 c 0.6 ± 0.01 b 0.6 ± 0.02 a 0.5 ± 0.01 b 0.01 ± 0.0 b 0.01 ± 0.0 a--c 4.3 ± 0.14 ab 4.4 ± 0.1 d 1.7 ± 0.3 ab 2.7 ± 0.2 c 1.2 ± 0.1 b 1.3 ± 0.1 ab **C5** 3.5 ± 0.1 b 4.5 ± 0.2 b 0.6 ± 0.02 c 0.6 ± 0.01 b 0.6 ± 0.02 ab 0.4 ± 0.01 c 0.01 ± 0.0 b 0.01 ± 0.0 c 4.1 ± 0.11 bc 5.2 ± 0.2 c 2.5 ± 0.7 a 3.4 ± 0.2 b 1.2 ± 0.1 b 1.0 ± 0.1 c plants-09-00675-t002_Table 2 ###### Performance of different dosages of *Scenedesmus quadricauda* extract on different morpho-biometric parameters on *Beta vulgaris* seedlings. Data derived from two repeated experiments. Standard error of the means = SEM, *n* = 4. Means followed by different letters within the column are significantly different according to Fisher's least significant difference test (α = 0.05). Length (cm) Surface Area (cm^2^) Mean Root Diameter (mm) Root Volume (cm^3^) Tips Number Root Number (0.000 \< L \< 0.500) Root Number (0.500 \< L \< 1.000) -------- -------------- ---------------------- ------------------------- --------------------- --------------- ----------------------------------- ----------------------------------- --------------- -------------- ------------- -------------- -------------- ------------- ----------------- **C0** 0.8 ± 0.1 d 3.0 ± 0.2 d 0.2 ± 0.01 c 0.3 ± 0.01 d 0.6 ± 0.01 b 0.6 ± 0.02 a 0.003 ± 0.0 c 0.01 ± 0.0 ab 1.5 ± 0.07 d 2.9 ± 0.1 b 0.2 ± 0.04 c 1.4 ± 0.1 e 0.6 ± 0.02c 1.4 ± 0.1 bc **C1** 2.7 ± 0.1 c 3.9 ± 0.1 c 0.5 ± 0.01 b 0.6 ± 0.03 c 0.6 ± 0.01 b 0.5 ± 0.01 c 0.01 ± 0.0 ab 0.01 ± 0.0 b 2.5 ± 0.11 c 3.1 ± 0.1 b 1.1 ± 0.1 a 2.8 ± 0.1 bc 1.9 ± 0.1a 1.3 ± 0.1 c **C2** 5.2 ± 0.3 a 6.2 ± 0.2 a 0.7 ± 0.03 a 0.9 ± 0.03 a 0.7 ± 0.01 ab 0.5 ± 0.01 c 0.01 ± 0.0 bc 0.01 ± 0.0 a 2.8 ± 0.13 c 5.3 ± 0.2 a 0.5 ± 0.2 bc 4.1 ± 0.2 a 1.2 ± 0.4b 1.9 ± 0.1 a **C3** 3.5 ± 0.2 b 5.4 ± 0.3 b 0.7 ± 0.04 a 0.9 ± 0.01 a 0.6 ± 0.04 b 0.5 ± 0.02 c 0.01 ± 0.0 a 0.01 ± 0.0 a 3.4 ± 0.14 b 5.0 ± 0.2 a 1.2 ± 0.3 a 3.3 ± 0.2 b 2.1 ± 0.1a 1.8 ± 0.1 a **C4** 2.9 ± 0.04 c 3.6 ± 0.2 c 0.5 ± 0.02 b 0.7 ± 0.02 b 0.7 ± 0.02 b 0.6 ± 0.04 ab 0.01 ± 0.0 a 0.01 ± 0.0 ab 4.0 ± 0.12 a 4.7 ± 0.2 a 1.0 ± 0.2 ab 1.9 ± 0.2 de 2.1 ± 0.1a 1.6 ± 0.04 a--c **C5** 2.6 ± 0.1 c 3.9 ± 0.1 c 0.5 ± 0.01 b 0.7 ± 0.01 b 0.8 ± 0.06 a 0.5 ± 0.3 bc 0.01 ± 0.0 a 0.01 ± 0.0 ab 2.6 ± 0.3 c 5.1 ± 0.4 a 0.6 ± 0.1 bc 2.5 ± 0.2 cd 1.8 ± 0.1 a 1.8 ± 0.3 ab

{ "pile_set_name": "PubMed Central" }

Dear Editor-in-Chief ==================== Every year in the world there are 600,000 new cases of patients diagnosed with head and neck cancer ([@B1],[@B2]). In Poland head and neck cancer accounts for 4.5% of all cancer cases which in 2008 reached 6046 ([@B3]). The aim of study was to assess the quality of life in patients' treatment due to head and/or neck cancer. Study was conducted of 48 patients (52.08% men), treatment at the Oncology Center, Maria Skłodowska-Curie Institute in Warsaw, Poland, in 2015. The standardized questionnaire EORTC QLQ-C30 and QLQH&N35 module for these patients. The study was approved by the local Ethics Committee and the management of the hospital. All subjects gave their written informed consent to participation in the study. The mean age was 56.29±6.94 yr (range: 43--67 yr). An equal number of respondents had location of tumour: larynx and laryngopharynx and oral cavity (43.75% each) and oropharynx (12.50%). Patients was treatment: radiotherapy (54.17%); chemotherapy (4.17%); surgical treatment (2.08%); radiotherapy and chemotherapy (33.33%); surgery, radiotherapy and chemotherapy (4.17%); surgery and radiotherapy (2.08%). The mean value of subjective assessment of health condition reached 4.04±1.09 and quality of life 3.79±1.17. The statistical analysis did not reveal correlations between subjective assessment of health condition and level of quality of life and sex, age, location of tumour, stage, and type of treatment. Each of aspects of physical function (difficulties in performing wearisome activities, fatigue during long and short walks, help required in the performance of everyday activities, limitations in the performance of everyday activities, limitations in pursuing hobbies, need for rest during the day and the sense of fatigue) was correlated with sex, age, location of tumour, stage and type of treatment. Physical function was affected solely by the stage and type of treatment. Thus, sex, age, location of tumour did not determine the physical ability of patients. The sense of pain (dyspnoea, weakness and pain that made the performance of everyday activities more difficult) was not determined by variables: sex, age, location of tumour, stage and type of treatment. 16.7% of patients\` health condition did not affect their family life, with 29.2% claimed there was no connection between their health condition and social life. Senior patients experienced significant or very significant difficulties (*P*=0.01). The results of QLQ-N&H 35 questionnaire are illustrated in [Table 1](#T1){ref-type="table"}. ###### The results of QLQ-H&N35 questionnaire ***QLQ-H&N35*** ***N (%)*** -------------------------------------------------- ------------- ----------- ----------- ----------- Pain in the mouth 10 (20.8) 19 (39.6) 11 (22.9) 8 (16.7) Pain in the maxilla 13 (27.1) 19 (39.6) 10 (20.8) 6 (12.5) Oral sensitivity 4 (8.4) 22 (45.8) 12 (25.0) 10 (20.8) Sore throat 4 (8.4) 18 (37.5) 16 (33.3) 10 (20.8) Difficulties in swallowing of liquids 7 (14.6) 17 (35.4) 18 (37.5) 6 (12.5) Difficulties in swallowing of mashed foods 8 (16.7) 20 (41.7) 11 (22.9) 9 (18.7) Difficulties in swallowing of solid foods 3 (6.3) 16 (33.3) 18 (37.5) 11 (22.9) Choking when swallowing 13 (27.1) 12 (25.0) 13 (27.1) 10 (20.8) Teeth problems 19 (39.6) 12(25.0) 12 (25.0) 5 (10.4) Difficulties in mouth opening 10 (20.8) 18 (37.5) 17 (35.4) 3 (6.3) Dry mouth 1 (2.1) 19 (39.6) 22 (45.8) 6 (12.5) Gluey saliva 4 (8.3) 14 (29.2) 23 (47.9) 7 (14.6) Impaired smell 7 (14.6) 22 (45.8) 12 (25.0) 7 (14.6) Impaired taste 2 (4.2) 15 (31.3) 18 (37.5) 13 (27.1) Cough 10 (20.8) 18 (37.5) 11 (22.9) 9 (18.8) Hoarse voice 10 (20.8) 17 (35.4) 12 (25.0) 9 (18.8) Feeling sick 6 (12.5) 17 (35.4) 17 (35.4) 8 (16.7) Preoccupied with appearance 11 (22.9) 13 (27.1) 13 (27.1) 11 (22.9) Difficulties in eating 7 (14.6) 17 (35.4) 13 (27.1) 11 (22.9) Difficulties in eating with the family 17 (35.4) 10 (20.8) 10 (20.8) 11 (22.9) Difficulties in eating with other people 17 (35.4) 12 (25.0) 13 (27.1) 6 (12.5) Difficulties in deriving pleasure from eating 6 (12.5) 19 (39.6) 14 (29.2) 9 (18.8) Difficulties in speaking with other people 12 (25.0) 17 (35.4) 11 (22.9) 8 (16.7) Difficulties in phone conversations 8 (16.7) 20 (41.7) 10 (20.8) 10 (20.8) Difficulties in family communication 19 (39.6) 12 (25.0) 12 (25.0) 5 (10.4) Difficulties in social relations 16 (33.3) 16 (33.3) 11 (22.9) 5 (10.4) Difficulties when leaving someone else's house 17 (35.4) 14 (29.2) 9 (18.8) 8 (16.7) Difficulties in contacts with family and friends 15 (31.3) 17 (35.4) 12 (25.0) 4 (8.3) Decreased interest in sex 16 (33.3) 16 (33.3) 13 (27.1) 6 (12.5) Decreased satisfaction with sex 15 (31.3) 14 (29.2) 12 (27.1) 6 (12.5) Oral sensitivity was more common among patients completed treatment compared against those who still received treatment (*P*=0.04). Women reported primarily difficulties in swallowing of mashed foods (*P*=0.03) and impaired smell (*P*=0.02). A statistically significant relation was found between the sense of a dry mouth and the stage of treatment (*P*=0.00). The difficulties associated with phone conversations were dependent on the health condition of patients (*P*=0.04). The difficulties in family communication significantly impaired the quality of life of patients (*P*=0.01). No such relation was found in case of social relations. There was a correlation between satisfaction with sex and health condition of patients (*P*=0.03). The study was not sponsored (own resources). **Conflict of Interests** The authors declare that there is no conflict of interest.

{ "pile_set_name": "PubMed Central" }

Polarized centriole positioning is important for properly oriented cell division, cilia positioning and cell migration. In mammals, planar polarized centriole postioning, as a part of the basal body of cilia, is necessary for proper directional beating of cilia at the apical surface within the node to establish left--right asymmetry, or in ependymal cells to promote cerebrospinal fluid circulation, among many other vital functions[@b1][@b2][@b3][@b4][@b5]. The coordination of cilia/centriole positioning from cell to cell across a tissue has been shown to be dependent on Frizzled/planar cell polarity (Fz/PCP) signalling in vertebrates[@b6][@b7][@b8]. Epithelial cells in *Drosophila*, on the other hand, do not have cilia and so the effect on cilia positioning versus general centriole positioning can be separated. PCP refers to polarization within the epithelial plane and is regulated by two distinct and conserved protein pathways, the Fat/Dachsous pathway (Ft/Ds-PCP) and the core Fz pathway (Fz/PCP)[@b9][@b10]. The conserved mechanism(s) of Fz/PCP signalling are mediated by Wnts and two protein complexes, Fz/Dishevelled/Diego/Flamingo (Fz/Dsh/Dgo/Fmi (a.k.a. Starry night, Stan)) and Van Gogh/Prickle/Flamingo (Vang (a.k.a. Strabismus, Stbm)/Pk/Fmi), which localize to opposite junctional domains within each epithelial cell[@b11][@b12][@b13][@b14][@b15][@b16][@b17], for example, to distal and proximal sides, respectively, in *Drosophila* wing cells. PCP signalling generally coordinates cell polarity across tissues, including ciliary positioning, the latter being reflected in the growing number of human diseases linked to aberrant Wnt-Fz/PCP pathways[@b18]. Ciliopathies, including Bardett--Biedl, Joubert and Meckel--Gruber syndromes, as well as neural tube closure defects in the embryo[@b19][@b20], are linked to vital roles of PCP in cilia positioning and orientation, and directed cell movements during gastrulation[@b21]. There is growing evidence linking core components of the Fz/PCP pathway to ciliary positioning in vertebrates, that is, in the developing mouse embryo the basal body of node cilia shifts from the centre towards the posterior side of the node cells in a PCP-dependent manner[@b22]. In fact, Inversin, a vertebrate homologue of Dgo, localizes to the basal body and axoneme and is part of the NPHP (nephronophthisis) disease module, and its loss-of-function (LOF) alleles affect cilia morphogenesis, convergent extension, and left--right determination[@b23]. Vangl2 can also localize to the basal body and axoneme in ciliated cells[@b24], and it affects the position and tilting of cilia[@b8]. Moreover, a Dishevelled triple knockout (*mDvl1, 2, 3*^*−/−*^) in multi-ciliated ependymal cells causes hydrocephalus and mis-positioning of cilia[@b25], a phenotype similar to LOF of the mammalian Fmi homologues, Celsr2 and Celsr3 (refs [@b26], [@b27]). Fz/PCP signalling controls ciliary positioning in all vertebrates examined, including mice, zebrafish and *Xenopus*[@b6][@b7][@b8][@b28], and it controls intracellular positioning of centrioles during zebrafish gastrulation, biasing it towards the posterior cell region[@b29]. In other contexts, including non-ciliated epithelial cells in *Drosophila*, the effect of Fz/PCP signalling on ciliary components, including acetylated tubulin or centrioles, remains unknown. It is thus an important evolutionary question whether Fz/PCP acts on centriole positioning in general, including in non-ciliated tissues where PCP signalling is 'active\'. We thus decided to explore centriole distribution and positioning in *Drosophila* wing epithelia, where the effects of PCP are well established, but which is a non-ciliated epithelium, as are all *Drosophila* imaginal disc epithelia. The *Drosophila* wing is one of the best-established tissues in which to study PCP pathways[@b11][@b12][@b13][@b14][@b15][@b16][@b17]. Adult wings manifest PCP with a single distally pointing actin-based hair in each cell (a trichome)[@b30]. At the pupal stage, when the wing is formed by two monolayers of non-ciliated epithelial cells juxtaposed at their basal membranes, at around 30--32 h APF, the trichomes start to appear, as actin polymerization becomes activated and focused at the distal apical vertex of each cell. This process depends on Rho family GTPases, which are recruited and activated by Fz-Dsh/PCP complexes[@b31][@b32][@b33]. Microtubules (MTs) also change in arrangement, from a radial to parallel distribution, projecting towards the distal apical portion of the cell with a distal bias of MT plus ends[@b34][@b35][@b36]. Although Fz/PCP signalling induces changes to the cytoskeleton, many unanswered questions remain how PCP regulates cytoskeletal elements, and it is for example unknown what type of MTs are involved in actin-hair formation in pupal wings and if these are linked to actin polymerization or centriole positioning among other options. In this study, we demonstrate that in non-ciliated cells of the *Drosophila* wing the positioning of centrioles is polarized towards the Fz/Dsh side of each cell and, importantly, under the control of the core Fz/PCP system. Our *in vivo* data in *Drosophila* wings argue for and provide evidence that centriole positioning is a conserved PCP readout, likely shared in all epithelial cells. Results ======= Centriole polarization in pupal wing cells ------------------------------------------ Using non-ciliated cells in imaginal discs, we asked whether centriole positioning is linked to Fz/PCP signalling as an evolutionarily conserved readout of Fz/PCP establishment (or if the presence of cilia is a pre-requisite for a Fz/PCP signalling-centriole connection). We first established a quantitative method to assess centriole positioning during pupal wing development, at the time when cytoskeletal rearrangements are being established downstream of Fz/PCP signalling, establishing a distally oriented trichome/actin hair. Two core centriolar components, Sas4 and Asterless (Asl), serve as excellent markers for centrioles; Sas4 and Asl, which is a centriolar scaffold required for centriolar assembly[@b37]. We analysed the localization of centrioles, via Sas4 and Asl staining, in pupal wing epithelial cells relative to other cellular markers, leading to two initial general observations on centriolar positioning: (1) centrioles are localized apically in cells at the level of the adherens junctions ([Fig. 1](#f1){ref-type="fig"}); and (2) Centriole positioning became progressively more polarized and localized to the distal vertex of each cell ([Fig. 1](#f1){ref-type="fig"} and [Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). Centrioles were detected at the adherens junction level, which were labelled with Fmi, and were never detected more basally (for example, at the level of Dlg/Discs large, a marker for baso-lateral membrane[@b38]; [Fig. 1](#f1){ref-type="fig"} and [Supplementary Fig. 1](#S1){ref-type="supplementary-material"}). This apical localization is very similar to that in vertebrate polarized epithelial cells. As the cells matured and started to display polarized, distal actin polymerization (phalloidin staining; [Fig. 1](#f1){ref-type="fig"}), centrioles became also localized to the distal vertex of each cell ([Fig. 1](#f1){ref-type="fig"} and [Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). To characterize the timing of centriole re-positioning in pupal wing epithelial cells, we determined centriole positions before and during hair formation relative to the centre of the cell, measuring distance and angles between the centroid of the cell and the centroid of the centriole in individual cells ([Fig. 1f--w](#f1){ref-type="fig"}; see Methods and [Supplementary Fig. 2](#S1){ref-type="supplementary-material"} for technical details). Before the actin-based hair is formed (28--30 APF), centrioles are positioned near randomly but close to the centroid of the cell ([Fig. 1f--i](#f1){ref-type="fig"}; see quantification in [Fig. 1j--k](#f1){ref-type="fig"}), and can be found at any angle (rosette diagram in [Fig. 1j](#f1){ref-type="fig"}; also [Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). At 31 APF when actin starts to be enriched at distal vertex, centrioles appear biased towards the distal cellular vertex, the distal quadrant between −45° and +45° ([Fig. 1l--q](#f1){ref-type="fig"}, compared with non-polarized localization at 29--30 APF, [Figure 1f--k](#f1){ref-type="fig"}). Once hairs are detected in all cells (32--33 APF), centriole localization is fully polarized to the distal vertex of cells (relative to the wing margin) ([Fig. 1r--w](#f1){ref-type="fig"}). This subcellular distribution of centrioles is very similar to that of the actin-based hair centroids ([Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). Taken together, we conclude that centriole positioning becomes planar polarized in non-ciliated epithelial cells, following largely the same distribution as the trichomes, the actin-based hairs and the best defined 'cellular effect\' of Fz/PCP signalling in *Drosophila* wings. PCP signalling instructs centriole positioning ---------------------------------------------- As PCP signalling regulates wing hair (trichome) formation in *Drosophila* and the position of basal bodies (cilia) in vertebrates, we next investigated whether Fz/PCP regulates the position of centrioles during pupal wing development. We characterized the positioning of the centrioles in loss or gain-of-function core Fz/PCP pathway backgrounds, with the same quantitative approach described above, using *en-Gal4* driven *fmi-IR* (dsRNA knockdown; see Methods) and *fz* null allele (*fz*^*P21*^) wings ([Fig. 2](#f2){ref-type="fig"}, and [Supplementary Fig. 3](#S1){ref-type="supplementary-material"}) and *dpp*-driven Fz overexpression (Fz-OE; GOF) background. In both genetic scenarios, centriole positioning was altered ([Fig. 2](#f2){ref-type="fig"} and [Supplementary Fig. 3](#S1){ref-type="supplementary-material"}). Although apical localization was maintained, centriole positioning remained unpolarized and more centered within mutant cells in both cases (cf. heat maps in [Fig. 2f,l,r,t](#f2){ref-type="fig"}), consistent with randomized PCP in each case. Moreover, angle distribution was spread over a much wider range of angles with WT regions of same wings remaining polarized and serving as controls (compare rosette diagrams in [Fig. 2e,k,q,s](#f2){ref-type="fig"}, respectively; see also [Supplementary Fig. 3](#S1){ref-type="supplementary-material"}). These data are consistent with the notion that in epithelial cells centriole positioning is generally connected to Fz/PCP signalling. Cytoskeleton and centriole localization are linked -------------------------------------------------- To determine how centriole positioning in pupal wing cells relates to the cytoskeleton, we next analysed both MTs and actin in pupal wing cells. MT localization and actin have been previously stained during PCP establishment in pupal wings, using confocal and electron microscopy[@b32][@b34][@b39], but the nature of the MTs remained unexplored. We tested whether they were acetylated, a marker of stable MTs, since acetylated MTs are generally associated with cilia and the basal body in ciliated cells. In pupal wings before hair formation, actin was enriched in the apical plane of the cell and acetylated MTs also formed a web-like structure in the apical plane of each cell ([Fig. 3a--c](#f3){ref-type="fig"}). Once actin-based hairs started to form and actin polymerization was focused at the distal vertex of the apical membrane, acetylated MTs became enriched at the base of the hairs ([Fig. 3d--f](#f3){ref-type="fig"}). Moreover, during hair formation, acetylated MTs started to 'invade\' the hair itself, forming what looks like a scaffold-structure for the trichome ([Fig. 3d--f](#f3){ref-type="fig"}). Of note, the juxtaposition between actin and acetylated tubulin was maintained into deeper areas in the cell, not just in the most apical planes from which the hair projects ([Supplementary Fig. 4](#S1){ref-type="supplementary-material"}). When centrioles were co-labelled with actin in the presence of the hair structure/trichome, they were localized adjacent to the base of the hair (co-labelled with actin; [Supplementary Fig. 5](#S1){ref-type="supplementary-material"}). The close co-localization of centrioles, acetylated tubulin and actin raises the question of whether they remain connected or become disconnected in PCP LOF or GOF backgrounds. Knockdown of *fmi* in the posterior wing compartment causes PCP defects, reflected in aberrant hair orientation and some multiple cellular hair (mch) defects ([Fig. 2g](#f2){ref-type="fig"} and [Supplementary Fig. 3i](#S1){ref-type="supplementary-material"}). Generally in mutant PCP backgrounds, the actin hair in each cell appears unpolarized (within the apical membrane plane), often near the centre of a cell. When acetylated tubulin was examined in *fmi*-IR conditions, it remained associated with the growing actin hair, independently of where the hair growth was positioned within the cell. Furthermore, this link between acetylated tubulin, centrioles and actin remained in GOF scenarios, for example, when Fz was over-expressed (for example in the *dpp* stripe of 8--14 cells close to vein L3; [Supplementary Fig. 6](#S1){ref-type="supplementary-material"}). These results suggested that acetylated tubulin/ MTs are a structural component of the actin-based hair and that the link between actin and acetylated tubulin is independent (or downstream) of Fz/PCP establishment, even though the position of the structure within the cell is regulated by Fz/PCP activity. Next we asked whether the number of actin hairs or hair positioning and thus actin polymerization downstream of Fz/PCP could affect centriole localization. The null allele of multiple wing hair (*mwh*^*1*^) shows mch formation. Mwh, becomes localized in response to Fz/PCP factor interactions and acts downstream of the Fz/PCP effectors Inturned and Fuzzy, and it is thought to directly affect actin polymerization. Moreover, it does not affect localization of the core Fz/PCP factors (for example, Fmi or Fz (refs [@b40], [@b41])). Centriole localization was less polarized in *mwh* mutants ([Fig. 3g--j](#f3){ref-type="fig"}; see heatmap in [Fig. 3e](#f3){ref-type="fig"}, although angular distribution was less affected when compared with *fz* or *fmi* LOF and GOF backgrounds; for example, compare [Figs 3k](#f3){ref-type="fig"} and [2k,s](#f2){ref-type="fig"}). These data suggest that centriole positioning is a downstream event of localized actin polymerization, which in turn is regulated by Fz/PCP. Over-expression of the Sple isoform of *pk* in developing wings has recently been reported to reverse PCP orientation, resulting in actin hair formation being moved to the proximal cellular vertex without affecting Vang or Dsh localization[@b36][@b42]. Importantly, Sple-OE (see Methods) caused not only a reversal of actin hair polymerization but also a reversal of centriole positioning, with the majority of centrioles located to the proximal vertex ([Fig. 3m--r](#f3){ref-type="fig"}). Taken together, these data are consistent with the model that centriole positioning is a downstream readout of core Fz/PCP signalling, similarly to properly localized actin polymerization, and, more importantly (based on the *mwh* LOF effects) that centriole positioning is a downstream effect of localized actin polymerization as regulated by the core Fz/PCP pathway. Loss or gain of centrioles does not affect PCP ---------------------------------------------- We next asked if loss or gain of centrioles could affect core PCP factor localization. Importantly, loss of centrioles, via Sas4 knockdown or the *sas4*^*S2214*^ null allele, did not result in detectable PCP phenotypes in pupal wing cells when Fmi staining/polarization was assessed ([Fig. 4a--e](#f4){ref-type="fig"} and [Supplementary Fig. 8](#S1){ref-type="supplementary-material"}), although growth defects were observed as recently reported ([Supplementary Fig. 8](#S1){ref-type="supplementary-material"}; surprisingly, flies lacking Asl or Sas4, and thus centrioles, survive to adulthood due to compensatory cell proliferation; their adult wings exhibit blisters, overgrowth, and vein mis-patterning[@b43][@b44][@b45]). In contrast, Asl and Sak/Plk4 (Polo-like kinase 4) overexpression caused an increase in centriole number[@b37][@b46], for example under *en-Gal4* conditions (*en\>Asl*; [Fig. 4f,m](#f4){ref-type="fig"}) or ubiquitously (Sak/Plk4, expressed via *ubiquitin* promoter; not shown). In such conditions, we did not detect defects in Fmi localization at the pupal stage ([Fig. 4f--l](#f4){ref-type="fig"}), with respect to either core PCP generated nematic order, actin hair orientation or centriole positioning (see also [Supplementary Fig. 7](#S1){ref-type="supplementary-material"}). Strikingly, however, in the thus generated multi-centriolar cells, when more than two centrioles were present (Asl normally only labels one centriole per cell, the mother centriole; [Supplementary Fig. 1](#S1){ref-type="supplementary-material"}), all centrioles were positioned near the distal vertex ([Fig. 4k--m](#f4){ref-type="fig"}), and moreover were always located near the base of the hair ([Supplementary Fig. 7](#S1){ref-type="supplementary-material"}). These results indicate that the number of centrioles does not impact Fz/PCP signalling and further confirm the notion that centriole positioning is a downstream effect of core Fz/PCP signalling via its effects on actin polymerization. Discussion ========== Taken together with observations that Fz/PCP signalling regulates basal body and cilia positioning in vertebrates[@b6][@b7][@b8][@b28], our data on centriole positioning as a Fz/PCP readout in non-ciliated *Drosophila* wing cells indicate that centriole/MTOC (MT organizing centre)/basal body positioning is an evolutionarily conserved downstream effect of Fz/PCP signalling. Its link with actin polymerization (hair formation in *Drosophila* wing cells) suggests that actin polymerization effectors also affect cilia positioning, possibly through docking of the basal bodies to the apical membranes. Inturned, Fuzzy and Rho GTPases regulate apical actin assembly necessary for the docking of basal bodies to the apical membrane[@b47][@b48] and this apical actin membrane accumulation is lost in Dvl1-3-depleted cells[@b47][@b48][@b49][@b50]. In left--right asymmetry establishment of the *Drosophila* hindgut, which is not a Fz/PCP-dependent process[@b51], asymmetric centriole positioning is observed. During this so-called planar cell shape chirality process, which affects gut-looping and thus embryonic left/right asymmetry, centriole positioning is however still dependent upon actin polymerization downstream of Rho GTPases (Rac and Rho), via MyoD and DE-cadherin control[@b51][@b52]. As Rho GTPases (Rac, Cdc42 and Rho) are downstream effectors of Fz-Dsh/PCP complexes, and their mutants cause PCP-like phenotypes including mchs or loss of hairs in wing cells[@b31][@b32][@b53]. It is thus tempting to infer that both processes, planar cell shape chirality and Fz/PCP, regulate centriole positioning through a common Rho GTPase-mediated actin polymerization pathway, initiated by an upstream cellular communication system, although this assumption will require experimental confirmation. In the mouse, Fz/PCP signalling regulates cilia movement/positioning in cochlear sensory cells via Rho GTPase-mediated processes[@b54], suggesting a similar mechanism in a representative mammalian PCP model system ([Fig. 4n](#f4){ref-type="fig"}). In conclusion, the positioning of centrioles appears to be a key and an evolutionary conserved downstream readout of Fz/PCP signalling, ranging from flies to mammals in both ciliated and non-ciliated cells. Methods ======= Fly strains ----------- Flies were raised on standard medium and maintained at 25 °C, unless otherwise indicated. GAL4/UAS system[@b55] was used for gene expression and RNAi studies. The Gal4 expression drivers were *en-GAL4, dpp-GAL4 and nub-GAL4*. In addition the following lines were used: *fmi* RNAi (Mlodzik lab stock, ML117 *(2)*), Sas4 RNAi (KK106051 from VDRC and BL35049 from Bloomington Stock Centre), Sple-OE (gift from Masakazu Yamazaki; fly ID: TID29239 ref [@b42]), GFP-Asl-OE, Sas4-GFP, *sas4*^*S2214*^ (gift from Jordan Raff), *fz*^*p21*^, *mwh*^*1*^ (described in Flybase), UAS-Fz for Fz-OE experiments[@b56]. The different *Gal4* lines applied in this study were used to direct expression of the UAS-constructs to distinct wing compartments, linked to the localized expression of developmental genes such *engrailed (en)*, restricted to the posterior compartment of the wing, *decapentaplegic (dpp*), expressed in a stripe between L3 and L4, and *nubbin (nub)* expressed in the whole wing. Immunohistochemistry and immunocytochemistry -------------------------------------------- White pupae (prepupae) were collected and staged at 25 °C for indicated time points. Wings were dissected in PBS with 0.1% Triton X-100 (PBT) and fixed with 4% formaldehyde for 45 min at room temperature. Pupae were then washed twice in PBT and blocked in PBT with 2% bovine serum albumin for 30--45 min. Samples were incubated with primary antibodies overnight at 4° in PBT-0.2% bovine serum albumin. Samples were washed five times in PBT and incubated for 1 h in fluorescent secondary antibodies diluted in PBT and fluorescent phalloidin when indicated. Five additional washes in PBT were performed before pupal wings were detached from the pupal cage and mounted on slides with Vectashield (Vector Laboratories). Pupal wing images were acquired using a confocal microscope (× 40∼oil immersion, 1.4 NA; SP5 DM; Leica) with LAS AF (Leica) software. The following antibodies were used: Anti-Asl (gift from Jordan Raff); anti-Fmi (from DSHB); anti-Dlg (gift from Kuyng-Ok Cho); anti-Cnn (gift from Jordan Raff); and acetylated-tubulin (from Sigma). Secondary antibodies used at 1:200 were from Invitrogen with different Alexa fluorophores (Alexa 568 and Alexa 647). FITC-phalloidin, rhodamine-phalloidin and Alexa 647-phalloidin were used at 1:500--1,000 (Invitrogen). Quantitative analyses of centriole positioning ---------------------------------------------- A novel function in MATLAB was developed to assess the relative position of centrioles within cells using immunohistochemical images. As input, this function uses (i) an immunofluorescent image of centrioles marked by Sas4 or Asl and (ii) the same image, marked by Fmi or Dlg and processed with the software 'packing_analyzer_V2\' to obtain the sketches of the cell borders. First, each cell is automatically detected and its centroid (*R*) is calculated as the center of the mass of the cell, whose Cartesian coordinates are where (*r*~ix~, *r*~iy~) are the Cartesian coordinates of the pixel *r*~i~, and n represents the sum of pixels in the cell. Then, the centrioles present within the cell are recognized and, again, the centroid (*C* )representing the center of mass of each one is obtained as where (*p*~*j*x~, *p*~*j*y~) are the coordinates of the pixel *p*~*j*~, and *m* represents the sum of pixels in the centriole. The angle (*α*) formed between the centroid of the cell and the centroid of each centriole is calculated as with a correction in the orientation for angle detection in the 0\<=*α*\<=2*π* radians range. The final results from a full set of cells were plotted as a rosette diagram representing these angles, and χ^2^-tests were used to detect significant differences between angle distributions (see figure legends). Another representation of the relative centriole positions inside the cells was performed as follows: A general cell model for a specific genotype was generated as a 51 × 51 matrix called *M*. Then, each pixel *p* of each centriole was assigned to an element *E* in the matrix following the equation: rounding each element to the nearest integer. *eqRad* represents the equivalent radius of the cell (the radius of a circle with the same area as the cell), and calculated as The final cell model is obtained as the density histogram of *M* using the whole set of pixels of each centriole, and represented as a colour heatmap. This function is available on the website <http://www.neural-circuits.org/>other-software, and can be applied to the relative location of other subcellular structures, like the actin hair. Quantitative analyses of polarized Fmi localization --------------------------------------------------- Polarity as determined from anti-Fmi stained cells was calculated with the software 'packing_analyzer_V2\' as described in ref. [@b57]. The software calculated both angles and strength of polarization (nematic order)[@b57]. Rosette figures were generated to represent the 360° orientation of the nematic order/polarity vector population, with 0° always being oriented as pointing distally using MATLAB. Statistical tests were used to assess differences between cellular orientation distributions (χ^2^-test) or polarity strength/nematic order (*t*-test). Additional information ====================== **How to cite this article:** Carvajal-Gonzalez, J. M. *et al*. Positioning of centrioles is a conserved readout of Frizzled planar cell polarity signalling. *Nat. Commun.* 7:11135 doi: 10.1038/ncomms11135 (2016). Supplementary Material {#S1} ====================== ###### Supplementary Information Supplementary Figures 1-8 We are most grateful to Jordan Raff for many reagents (flies and antibodies), and all members of the Mlodzik and Carlo Iomini labs for helpful suggestions, comments and discussions during the development and execution of the project. We thank the Bloomington and VDRC Stock Centers for fly strains, and the DSHB for antibodies. We are grateful to Giovanna Collu for helping with the manuscript. J.M.C.-G. is a recipient of an Atracción y Retención de talento contract from the GOBEX (Extremadura government). This work was supported by NIGMS and NEI grants from the National Institutes of Health to M.M., and BFU2014-54699-P grant from the Ministry of economy to J.M.C.-G. Confocal microscopy was performed at the Tisch Cancer Institute Microscopy Core, supported by grant P30 CA196521 from the NCI of the National Institutes of Health. **Author contributions** J.M.C.-G. and M.M. designed the experiments, analysed the data and wrote the manuscript. J.M.C.-G. performed all of the experiments. A.-C.R. designed and performed the data analysis. ![Centriole localization and positioning during PCP establishment in pupal wings.\ (**a**) Illustration of pupal wing and its orientation (**b**--**d**) Sas4-labelled centrioles (cyan in **b** and monochrome **d**) are distributed in the junctional planes (*X*--*Z* plane---upper panels) (marked by Fmi staining, red in **b**, monochrome in **c**). Top panels are *x*--*z* sections of respective *x*--*y* views shown below. Scale bar, 10 mm. (**e**) Schematic representation of a pupal wing epithelial cell and the parameters used to study centriole positioning. (**f**--**w**) During pupal wing development, centriole localization changes. (**f**,**l**,**r**) Sas4 (green), Fmi stained in red, and actin (phalloidin) in blue, and the respective monochromes. (**f**--**k**) Before hair formation (29 h APF), centrioles are unpolarized in a central position in the apical portion of each cell (quantified in **f**,**k**). (**l**--**q**) At the onset of hair formation (31 h APF), centrioles begin to localize to the distal portion of each cell (quantified in **p**,**q**). (**r**--**w**) Subsequently, when hairs are fully present in all wing cells (32--34 h APF), centrioles appear to be polarized mostly to the distal sector of each cell (quantified in **v**,**w**). Scale bar, 10 mm. Red sectors in **f**,**p**,**v**---% within distal quadrant. Statistical analyses: rosette diagram distributions panel **g** versus **p**: *P\<*0.0001; **p** versus **v**: *P\<*0.0001 (χ^2^-test).](ncomms11135-f1){#f1} ![Centriole position is affected by PCP signalling.\ (**a**--**l**) Fmi LOF, using *en*-driven *fmi-IR* knockdown affects centriole localization; (**a**,**g**) Dlg: red, cell outline; Sas4: green, centriole; actin (phalloidin): blue; and respective monochromes in **a**--**d**,**h**--**j**. Centrioles within cells in the *en\>fmi-IR* area (**g**) are less polarized and are distributed more centrally, quantified in **e**--**f** and **k**--**l**, respectively. (**m**--**t**) Fz gain-of-function (GOF; see Methods), causes Fmi depolarization (see polarity vectors in **p**), and defects in centriole distribution (Fmi in red, monochrome in **n**; Sas4 in green, monochrome in **o**). Fz overexpressing cells have central distribution of centrioles (quantified in **s**--**t**; see also ROI: Fz-OE in **p**), as compared with *WT* areas of *dpp\>Fz-OE* wings (quantified in **h**,**i**; see ROI: *WT* in **p**). Scale bar, 10 μm. Statistical analyses: centriole rosette diagrams, **e** versus **k**: *P\<*0.0001; **q** versus **s**: *P\<*0.0001 (χ^2^-test).](ncomms11135-f2){#f2} ![Centriole localization relative to trichome position.\ (**a**--**c**) *XZ* and two *XY* optical sections of same wing area, showing polymerized actin (phalloidin, cyan) and acetylated tubulin staining (red, monochrome in **b**). At 29 h APF before actin-based hairs are formed, both actin and acetylated tubulin appear enriched at the apical portion of the pupal wing cells; two planes, apical (*XY1*) and subapical (*XY2*) are shown in the *XY*-axis (indicated in the *XZ* sections as black lines 1 and 2). (**d**--**f**) After hair formation (32 h APF), acetylated tubulin is enriched at the base of each hair and within the polymerized actin structure. *XZ* and two *XY* planes, apical (*XY1*) and subapical (*XY2*) are shown in the *XY*-axis from the same; stainings as in **a**. (**g**--**l**) Centriole localization in the *mwh*^*1*^ mutant is less polarized than *wt* (in 32 h pupal wing cells; compared with [Fig. 1v,w](#f1){ref-type="fig"}). Asl (green), Fmi (red) and phalloidin (blue) stainings and the respective monochromes are shown. **k** and **l** show centriole distribution quantifications in *mwh*^*1*^ mutant, note shift to less polarized, central distribution most evident in heat map (**l**). (**m**--**q**) Overexpression of Sple (Sple-OE) causes reversal of cellular polarity and hair position, and accordingly also centriole positioning is inverted; quantifications in **q**,**r**. Asl staining position was used for quantifications in rosettes (**k**,**q**) and heat maps (**l**,**r**). Scale bar, 10 μm. Statistical analysis: centriole rosette diagrams, **k** versus [Fig. 1w](#f1){ref-type="fig"} (*wt* control): *P\<*0.001; **q** versus *wt* control: *P\<*0.0001 (χ^2^-test).](ncomms11135-f3){#f3} ![Loss or gain of centrioles does not affect core PCP factor localization.\ (**a**--**e**) Imaging and quantifications (sed Methods) of Fmi staining in *Sas4* RNAi-mediated knockdown (in cells marked with GFP) in the posterior compartment in 32--34 h APF pupal wings. *en\>Sas4*^*IR*^ RNAi-mediated KD (**a**--**b**) did not disrupt Fmi localization as compared with the WT anterior compartment from same wings. (**c**,**d**) Quantifications of Fmi polarization using polarity vector angle orientation revealed no significant differences between *en\>Sas4*^*IR*^ (*n*=1,872 cells from five independent wings) and WT cells (*n*=1,513 cells, five independent wings). NS: non-significant (*P=*0.655) (χ^2^-test)). (**e**) Polarity vector lengths (relative to Fmi fluorescence polarization) did not show significant differences between in *Sas4*^*IR*^ cells and adjacent WT cells (five independent wings). NS: non-significant. (**f**--**j**) Gain of centrioles (\>2 per cell) through Asl overexpression (*n*=1,099 cells, three independent wings) in the posterior compartment (*en\>Asl*), did not affect Fmi localization as compared with wild type from the same wings (*n*=1,096 cells, three independent wings). (**h**,**i**) Quantification of angle distribution of polarity vectors, and (**j**) polarity vector length. Note no significant changes between *wt* and *en\>Asl* cells (*P=*0.783) (χ^2^-test). (**k**--**m**) Centriole positioning follows PCP core factor localization even when higher numbers of Asl-positive centrioles are present per cell. Quantifications are depicted in rosette (**k**) and heatmap (**l**) diagrams. (**m**,**n**) Specific examples of centriole localization in multi-centriolar cells, note several examples with 3--4 centrioles per cell (**m**, top), and two Asl-positive centrioles per cell (**m**, bottom); in *WT* only 1 centriole is stained by Asl for comparison (Asl in green; cf. to [Fig. 1a,b](#f1){ref-type="fig"}). In all cases, centrioles remained close to the distal side of each cell (marked with Fmi; magenta). (**n**) Schematic representation of signalling pathways involved in centriole polarization related to planar cell polarity in vertebrates versus *Drosophila*. Scale bar, 10 μm.](ncomms11135-f4){#f4} [^1]: Present address: Departamento de Bioquímica y Biología Molecular, Facultad de Ciencias, Universidad de Extremadura, Badajoz 06071, Spain.

{ "pile_set_name": "PubMed Central" }

![](yjbm00632-0073.tif "scanned-page"){.329} ![](yjbm00632-0074.tif "scanned-page"){.330}

{ "pile_set_name": "PubMed Central" }

Background ========== The notion \'hard to reach\' is a contested and ambiguous term \[[@B1]\] that is commonly used within the spheres of social care and health, especially in discourse around health and social inequalities. Reducing inequalities in health and health provision is a key theme underpinning the NHS Plan and National Service Framework\' \[[@B2]\]. This same report acknowledged that certain groups are marginalized from services and therefore \'harder to reach\' for health services whose goal is to provide appropriate and equitable health care for all populations. The Framework stated that primary care has a significant role to play in providing services that reach those in the greatest need but a Home Office report in 2004 \[[@B3]\] described the \'hard-to-reach\' as \'inaccessible to most traditional and conventional methods for any reason.\'. On the other hand, voluntary and community organisations (VCOs) are considered \'uniquely placed to reach marginalised groups\' \[[@B4]\] and the \'hardest to reach\' \[[@B5]\], and play an important role in the provision of health and social care services for these groups \[[@B6]\]. This paper describes the first stage of a planned study aimed at improving understanding of the \'hard to reach\' and improving access to primary health services for \'hard to reach\' groups. Acknowledgement of the established role of the Voluntary and Community Sector in providing services for these groups informed a decision to begin the study by exploring the perceptions and experiences of organisations from this sector. This first stage of the study had two research objectives - ◦ To describe how service providers from the voluntary and community sector (VCS) conceive of the notion \'hard to reach\'. ◦ To explore perceptions of the barriers and facilitators to accessing services for \'hard to reach\' groups. Using the search terms: \'hard to reach populations\' \'hidden populations\', \'marginalized populations\', a literature search was conducted using the following databases: Books @ Ovid, Journal @ Ovid Full text, British Nursing Index (BNI), CAB Abstracts (1973-2006), CINAHL, EMBASE, Ovid Medline, PSYCINFO, Sociological Abstracts and Web of Science. Searches were restricted to English-language papers and limited from 1997 to present day. A search of the grey literature was also conducted using the Home Office and Department of Health websites. A Google search was undertaken limited to last 12 months using the term \'hard to reach Groups.\' Despite the apparent familiarity of the term and its use in social research and public policy, there would appear to be a lack of consensus about the meaning of the term. Within the literature, \'hard to reach\' is often synonymised with other terms and the sheer multiplicity of alternatives reflects the divergence in the discourse as well as the difficulty in arriving at a definitive description of its meaning. \'Vulnerable\', transient\' \[[@B7]\] \'marginalised\' \'refusers\' \[[@B8]\], \'hidden\' \[[@B9]\], \'forgotten populations\' \[[@B10]\], \'underserved\' \[[@B11]\], \'special populations\' \[[@B12]\] \'disadvantaged populations\' \[[@B13]\] are terms that have all been utilised in the literature pertaining to addressing issues faced by the \'hard to reach\', or indeed those who are trying to reach them. There was an overarching sense that trying to engage the \'hard to reach\' is problematic, as described almost two decades ago by Freimuth & Mettger \[[@B14]\](1990) and this may be due to the inherent ambiguity and lack of clarity of definition. Assignation of the term appears largely dependent upon the context of the organisation doing the reaching. The \'hard to reach\' groups most commonly identified in the literature are sex workers, drug users, people living with HIV and people from lesbian, gay, bisexual, transgender and intersex communities \[[@B15]-[@B17]\], but there are a number of other groups to which the description applies including asylum seekers, refugees, black and minority ethnic communities (BME), children and young people, disabled people, elderly people \[[@B18]\] and traveller families \[[@B19]\]. The \'hard to reach\' may be people who are stigmatised due to the fact they are perceived as being somehow different. If we think in terms of stigma and social exclusion in a healthcare context, the \'hard to reach\' may include people with a variety of conditions or life limiting illnesses. This may include people with congenital abnormalities \[[@B20]\] or genetic conditions \[[@B21]\]; people with hearing loss \[[@B22]\]; people with mental health problems \[[@B23],[@B24]\]. It is important to address also the notion that there are populations who are deemed \'hard to reach\' but who are \'non-associative\'; that is they do not generally associate with other members \[[@B25]\]. These populations are comparatively under-researched and it is acknowledged that trying to gain access to disparate and often isolated individuals, although labour intensive, is an important area of consideration for researchers and indeed service providers working with the \'hard to reach.\' \'Hard to reach\' audiences have been defined as \'inaccessible to most traditional and conventional methods for any reason\' \[[@B3]\], highlighting the difficulty facing statutory health care provision. The voluntary and community sector has arguably been more successful in penetrating some of the barriers for the \'hard to reach\' and has an important role to play in the understanding of service delivery provision. There are around half a million voluntary and community organisations (VCOs) in the UK and there is a long tradition of voluntary action and community service in this country \[[@B4]\]. They have been used increasingly to deliver public services, especially over the last ten years \[[@B26]-[@B28]\]. The VCS is a term used to describe self-governing organisations working in areas of public benefit and they are \'value-driven\' in that they \'exist for the good of the community\' \[[@B29]\]. They can range in size from small community based groups or projects to large national and international organisations. As the term suggests, these services often rely upon volunteers for service provision and delivery and are characterised by their independence from the \'formal structures of government and the profit sector\'. However, in practice it appears that the sector \'defies easy definition\' \[[@B30]\] and it is suggested that the inherent ambiguity and variation in terms of how various organisations define and describe themselves is the very reason they are able to provide such diversity and flexibility. Research Objectives ------------------- The aim of the study was to explore \'hard to reach\'-from the experiences and perspectives of the VCS. The findings would contribute to a larger study considering how these maybe translated to provision of services in primary health care. The specific objectives of the work described in this paper were: ◦ To describe how service providers from the voluntary and community sector (VCS) conceive of the notion \'hard to reach\'. ◦ To explore perceptions of the barriers and facilitators to accessing services for \'hard to reach\' groups. Methods ======= This study utilised a qualitative design, since qualitative methods are well placed to explore sensitive and relatively poorly understood aspects of social life \[[@B31]\]. Semi-structured interviews were conducted with purposefully sampled team leaders or practitioners from VCS organisations across Birmingham. Purposive sampling or criterion based sampling \[[@B32],[@B33]\] was used to enable recruitment of participants who have particular experiences and roles that are pertinent to the study objectives. The Birmingham Index of Voluntary Organisations was employed to access details of organisations working with \'hard to reach groups\'. Letters of invitation were sent to 30 organisations and these were followed up by a telephone call, introducing the researcher and offering further details about the project. The organisations contacted included services working with people with mental health problems; homeless and poorly housed people; problem drug users, people identifying as gay, lesbian, bi-sexual and transgender; parents of young children in disadvantaged areas; people from BME communities; people working as carers; older adults; people with learning difficulties. These organisations were contacted as the groups they worked with are described in the literature pertaining to stigmatised, marginalised or hard to reach groups \[[@B23],[@B34],[@B15],[@B17],[@B34]-[@B38]\]. The initial plan was to interview service leads or managers as it was felt that they may be in the best position to give both an overview of their services, as well as providing accounts of front-line experience. However, it was not always possible to interview team leads, and participation was delegated by the service manager/team leads to other practitioners from within the projects, who in all but one case held a management role of some kind within the project. Eight services agreed to participate and the organisations worked with the following groups - young people, members of the Irish community, homeless people, people from BME communities living with HIV and people from the lesbian, gay and transgender communities, families and carers of drug users, young homeless people and people with mental health problems. The organisations that declined to participate cited time restrictions as their reason for non-participation. The literature review on \'hard to reach groups\' informed the development of the topic guide for the semi structured interviews. The topic guide covered the following areas: • remit of your organisation and the client group you work. • Experience of working with this group/in this field. • Understanding of the term hard to reach. • Who do you regard as \'hard to reach\'? • Why is it difficult to reach these groups? Why do you think they don\'t come? • What measures have been put in place to engage the \'hard to reach\'? • Examples of successes in engaging the \'hard to reach\'. Ethical approval was sought and obtained from the university ethics committee. IRAS approval was not required as no NHS patients or employees were interviewed. All interviews took place at the participants\' place of work. All Participants agreed for the interviews to be audio-recorded and signed a consent form to reflect this. Recorded interviews were transcribed by SF and each transcript was checked for any errors or omissions. Each manuscript was read and the emerging themes and key points were highlighted. Transcripts were analysed by both authors to ensure reliability of interpretation of the results. Labels and categories are used to \'organise and analyse the data\' \[[@B32]\] by assigning key themes and emerging concepts to the raw data. This thematic approach to analysis provides a focus on the identifiable themes and patterns that emerge from the raw data. Three broad areas emerged which include the following ◦ How the term \'hard to reach\' is defined ◦ What barriers the \'hard to reach\' face in accessing services. ◦ What can be done to engage \'the hard to reach\' Each of these broad themes comprised a number of categories which are discussed below. Quotations have been used throughout to add power to participants\' accounts, and to encapsulate the views and experiences of individual participants \[[@B39]\]. Results ======= How the term \'hard to reach\' is defined ----------------------------------------- Two respondents alluded to the apparent familiarity of the term \'hard to reach\', for example, \"one of those terms that people bandy about\" while also indicating their unease with it. As one respondent explained (R5): \"The problem I have with the term is that I don\'t really feel as if groups are actually hard to reach.\" before going on to suggest that it would be better to \"find some other terminology to explain \[it\].\" The term was attributed to describe the service (non) user and attributed to \'services\' themselves. The unengaged - non users of services ------------------------------------- There was an overall sense here that one conception of the \'hard to reach\' refers to the people who are not seen in services. The \"unengaged\" was used to by one respondent to refer to those \"who don\'t come to see us\" (R1). Two other respondents described them as; \"groups and communities who aren\'t accessing a particular service for whatever reason\" (R5) and those who \"traditionally wouldn\'t seek support from the usual avenues\" (R6). It is suggested that there are certain characteristics pertaining to how we may commonly define the notion of the \'hard to reach\': *People that would generally have difficulty accessing services\... younger and more vulnerable people\... you are talking about people who are more deprived, who are less well educated\... may have restrictions placed upon them by society generally or their more narrow social group*. (R1) This respondent further intimated that this does not tell the whole story and acknowledged in the context of providing a service for young people that the \"white middle classes\" may lead \"controlled lives\" that may inhibit them accessing particular services. This alternative explanation of the meaning of the term; the sense that the term \'hard to reach\' can refer to those inhabiting a more privileged section of society \[[@B40],[@B41]\] is an issue that is not raised in the remaining interviews but is worthy of note. More commonly we found that the responses support the sense that the \'hard to reach\' are in some ways more needy; more marginalized; more likely to have experienced poverty, whether that be economic poverty or poverty of opportunity \[[@B7],[@B19],[@B9]\]. Lack of choice is another feature that was used to explain why a group maybe \'hard to reach\' - *You don\'t have those choices\... either you can\'t afford them because you\'ve got limited income and you haven\'t got the information either to be able to make an informed choice*(R7) The transgender community is also particularly stigmatised \[[@B42]\]. The Diagnostic and Statistical Manual of Mental Disorders (DSM) includes \'Gender Identity Disorder\' as a category and this \'continues to raise questions of consistency, validity, and fairness\' \[[@B43]\]. To help overcome this, the project employs a transgender doctor to help engage with members of the transgender community. \'Hard to reach\' understood from a service and societal stance --------------------------------------------------------------- It is difficult to define the \'hard to reach\' without thinking of the wider societal issues that contribute to or create such conditions. Service restrictions and limitations may mean that it is the services themselves that are \'hard to reach.\' One respondent alluded to the sense that it may be the way that an organisation is configured that creates this notion of the \'hard to reach\' and she stated that the term can mean \"people to whom we put up some barriers\" - *it makes it sound like the fault of the non service-user - \"you are hard to reach\" - like they are sat up on a shelf and we have got to lure them down with biscuits, or something\.... actually if you just got a ladder and sat next to them that would be fine*. (R1) Other ways of describing this were given in terms of \'hard to reach\' individuals having \"fallen through the net\" (R7) or being \"left out of the loop.\" This suggests a sense that certain groups or individuals are being excluded from services and this is the fault of the services themselves - *organisations in the mainstream don\'t know \... how to reach them, how to engage them in activities, don\'t know where to find certain people*. (R2) These people have in some way been let down by the system: \"for whatever circumstances \[they\] have been through the system and have either failed the system of the system has failed them.\" (R3) One respondent made explicit the sense that the term is inappropriately attributed to groups and communities \[[@B18]\]. Instead, he stated that \"it is about looking at our own services and looking at why individuals aren\'t engaging. R6 suggested his organisation aims to engage *the hardest to reach people who traditionally wouldn\'t seek support from the usual avenues and traditionally hadn\'t succeeded very well in either society or within the education system*. Two key points are raised here; firstly of somebody being on the margins of what is considered to be the mainstream of society and secondly, a sense that the reason for this maybe due to a failure in the system. This sentiment is developed and made more explicit later when the respondent intimates that the problem may be systemic - *It\'s just about pushing them in that direction and perhaps that\'s where our school education doesn\'t quite serve everybody as well as it should*. This respondent also stated that \"we are very good at stereotyping and boxing people in\" and he uses a quotation from a service user to demonstrate this notion: *I am not \"hard to reach\", generally people don\'t know how to reach me*. Barriers and Facilitators to Services ------------------------------------- In order to understand how appropriate services can be offered to the \'hard to reach\', the experiences of VCS services providers in providing services to these groups were explored. Service providers were asked to describe the kinds of barriers they face in offering services to \'hard to reach\' clients, as well as the factors that helped to facilitate the engagement of these groups. Barriers -------- ### Previous experiences of accessing services A number of participants made explicit reference to the fact that many potential service users may not engage because of their previous experiences of accessing services. In particular, it would seem that statutory services were conceived as being particularly impenetrable \[[@B44]\], thus discouraging individuals to access help. The sense that such experiences may begin in the formative years is indicated by one respondent who suggested that the \"experience of the education system as being quite negative\" (R6) is likely to have an impact upon future expectations. ### Physical Factors There are logistical factors to be considered that may serve as barriers to accessing services, including factors such as location and opening times. One respondent indicated that \"geography and transport\" were barriers to services, but also the fact that young people may be conscious of gang boundaries: quite a lot of young people aren\'t sure if it is ok for them to go to certain places. (R1) A similar point was made by R7 who explained that their particular project moved location from the middle of an estate because: *a lot of young people wouldn\'t use it, because it is all territorial*. Another respondent (R7) suggested that geography may be a barrier to accessing services as potential service users \"don\'t know we are here.\" However, once people become aware of the location and are engaged in the service, limitations can arise in terms of how many people can be accommodated. By becoming victims of their own success, organisations may inadvertently exclude some more needy but less vocal members of their target audience. This idea is encapsulated by this quotation from R1; *the worry is that when you are very busy\... it\'s the more vulnerable that you squeeze out*. ### Funding and Partnership working A common issue raised by a number of respondents concerns the way that VCS organisations are funded \[[@B45]\]. These broad areas of concern were based around the fact the organisations have to vie for different pots of money to provide specific services for the particular remit of client group. A similar issue was raised by one respondent who suggested that there may be a fear of competition amongst partner organisations culminating in the concern that - *they are going to steal my clients. What service am I going to be left to run?*(R4) There was also a concern that excessive attention is given to those things that are easier to measure at the expense of those things that were harder to quantify. Furthermore, some \'causes\' are perceived as more worthy resulting in larger funding pools for certain projects. One respondent describes the inequity in accessing project funding for lesbians: If you work in BME communities, or people with HIV, it can be a lot easier to get funding.(R4) Despite a good deal of praise for the practice of partnership working (see below) one respondent drew attention to its potentially problematic nature, stating that co-working or partnership working may be perceived as synonymous with \"extra work\" and therefore maybe avoided. ### Attitudes and constraints of the statutory services towards the VCS Three of the respondents were explicit in their concerns around the relationship between themselves and statutory sector organisations. One expressed concern about the attitude held by the statutory sector towards the VCS. He suggested that the use of the term \'third sector\' with reference to the VCS reinforces a view that this sector is somehow of less value: *the image the statutory sector has of the voluntary sector is \'there, there nice little organisations. You do your specialist thing and leave us to do the professional work*. (R3) One respondent made explicit their frustrations with the statutory sector suggesting that social services are not equipped to deal with the complexity of problems faced by some young people: *if that young person doesn\'t co-operate, for whatever reason, then they\'re \[social services\] not interested and the young person is left high and dry*. (R7) Problems of access are also faced by the VCS themselves: *as an agency, it\'s also difficult to access these services on behalf of young people, especially when the average user will have more than one, like sort of, special need*. (R7) ### Expectations/limitations of services Expectations or preconceptions of the VCS services may also become a barrier in terms of accessing \'hard to reach\' groups. This is encapsulated by one respondent, who states: *it is about ourselves, the expectations that you must come to us*.(R5) It is suggested that for a service that is dominated by clients from a more narrow demographic this can become a barrier for potential service users who may feel like \"outsiders\" (R3). Even within services for the \'hard to reach\' there is a hard core of individuals who do not access services. In some homeless day centres, for example, where the client group are predominantly white men, then women and people from other ethnic groups may feel more reluctant to utilise the service. Furthermore, the views held by some clients about other service users can contribute to the difficulties faced by service providers in serving diverse groups: *we see individuals coming in with a very blinkered view about some of the service users that we see, maybe refugees - \"why are they over here stealing our jobs?*(R4) Facilitators ------------ As described, there are a number of perceived barriers in terms of providing services to the \'hard to reach.\' Conversely, a number of facilitators in providing accessible services to the difficult to engage were cited. ### Treatment of Clients - trust, respect A perception that service users have been poorly treated by other services, and in particular by statutory service has already been raised. The most commonly cited facilitator for engaging clients was the way VCS services communicate with and treat clients. This key theme is illustrated by the following quotations from R1: *we don\'t interrogate clients when they come here. We ask what questions we need to give them the service they want*. (R1) Reference is also made to the attention paid to the attitude of the staff members: *I think the important thing is to spend an awful lot of time and effort on making sure that staff are genuinely welcoming and non-judgmental about absolutely everyone that walks through the door*. (R1) The quality of the relationship that develops between staff and service user is a major factor in engaging clients; building trust and respect; being non-judgmental and being able to relate to and empower people. Ok, you can trust us, but it is very much about them being able to recognise that they can trust us, or build up that relationship which can take weeks. (R3) ### Flexibility Offering flexible services that respond to the needs of the service users including running outreach services, listening to feedback, offering flexible opening hours and providing service users with the kinds of services they want were all key facilitators cited by the respondents. One respondent made the point that although they are not as \"flexible as we would like,\" many of the people using this service are *very much on the peripheries of normal society\... a lot of people on benefits \...so they can come as and when*. (R4) Furthermore, by offering an outreach service to encourage potential service users to engage, they are able to access people who may not ordinarily feel comfortable to attending a HIV service. Not all the organisations interviewed conducted outreach and this was largely due to resource limitations. There appeared to be a tension between offering a service to all the target audience, including those \'harder to reach\' clients and being able to provide an adequate service to all. Offering services that people want is a key facilitator for engagement. One respondent described how changing their opening times as well as offering service users a hot meal had resulted in a fifty per cent increase in demand for the service (R3). ### Partnership Working As described above there are a number of drawbacks to partnership working that can be perceived as barriers to service participation. However, there are also a number of positive aspects to working in partnership and it seemed to be a priority for a number of the organisations. One respondent makes the point that working in partnership is actively encouraged by the funding bodies who support organisations. Partnerships can be useful in assisting in: *getting information out there, and ensuring that people know who we are*. (R4) Acknowledgement of the limitations of a service was also important: *we don\'t claim to know everything. We don\'t want to do everything. It isn\'t possible to do everything really, really well, but what you can do is build very, very strong partnerships with other organisations that do specialise in particular areas*. (R5) ### User involvement For some of the respondents, emphasis was placed upon the importance of user participation. Service participation and interview panels were conducted at one organisation throughout the year. \"Member helpers\" or trainee volunteers form part of the team with the aim to build the esteem and skills of the service users. Using service users as part of the team encourages other clients to participate and feel more settled: *It\'s about having a different relationship. A service user can better describe it \[how it feels to be a service user\]. They can share that with each other, use that as a bridging gap and a way of opening doors*. (R3) The kinds of relationships the VCS is able to develop with their clients is of the kind that is often found wanting in statutory services. The emphasis they place on user involvement and use of volunteers as demonstrated above can act as an enabler for those they seek to serve. Discussion ========== The VCS a model for the NHS? ---------------------------- As we have seen, tackling health inequalities is a top priority for the Government and part of this includes engaging those \'seldom seen, seldom heard\' in services. Recent Department of Health initiatives have highlighted the need for improved service provision and configuration. The NHS Next Stage Review Interim Report \[[@B46]\] reported that despite \'sustained investment and improvement in the NHS over the past 10 years, access to primary medical care and services and quality of those services continues to vary significantly across the country.\' (pp 3) As noted by a number of the respondents, the importance of listening to the voices of users has been taken onboard by the NHS as a key resource in providing appropriate services \[[@B47]\]. Already, changes in how health services, and in particular primary care services, are delivered seem to bear the hallmarks of VCS practice, but there is a need to go further. Some of the changes in provision of GP services aimed at improving patient access and experience include the introduction of patient feedback and quality assessment (surveying patients about their local GP practice); the extension of opening hours to accommodate the varying lifestyles of patient populations and thus engaging with the \'seldom seen\'. As we have seen the VCS ability to offer flexibility in terms of opening hours and responsive services for service users are cited as facilitators for the more difficult to engage. The Department of Health acknowledges the need to \'establish effective links with frontline services, utilising the potential of VCF sector agencies as valuable catalysts for dialogue, mutual understanding and empowerment \[[@B48]\] (pp4). Partnership working is becoming increasingly important to the health service. From the perspective of primary care provision it is acknowledged that as commissioners of services, Primary Care Trusts (PCTs) working alongside local authorities should \'identify specific communities with particularly poor health, such as travellers, migrant workers, people with learning disabilities, those living in disadvantaged areas or demographic groups\' \[[@B49]\]. It seems that the NHS is moving in the right direction in terms of offering choice and a voice to patients. Before equity in health services can be fully realised that work must continue in engaging the most excluded and \'hard to reach\' in society. Study Limitations ----------------- The study is clearly limited by the small sample size. Most qualitative studies will aim to interview between twenty and thirty people although the aim is to continue data collection until data saturation is met. This is the point were no further themes are being generated. Unfortunately, only eight out of the thirty organisations felt they had the time to participate. It could be postulated that organisations providing services for more marginal groups are resource and time limited and therefore find it more difficult to put time aside, so the views of practitioners from these organisations may have been lost. The participant organisations operate services in a large, diversely populated urban area, and the kind of experiences and views found here are likely to differ from those from organisations operating in rural communities, with a widely dispersed population. There are also likely to regional variations that are not reflected in this study. However as a pilot study, one of the objectives is to provide background to inform a larger scale study proposal which aims to involve the views of a larger number of participants including service users and primary care practitioners. Conclusion ========== The VCS have a long history of working with marginalised, disadvantaged and \'hard to reach\' groups and as discussed their experiences may provide potential lessons for statutory service providers, particularly health services. The VCS is able to provide added value in terms of the resources they offer; the procedures they employ and how they are organised \[[@B30]\]. From the interviews conducted there are four main areas of importance that have arisen that relate to how best to engage \'hard to reach groups\': attitude of staff; service flexibility; working in partnership with other organisations and empowering users involvement. In order to fully engage with the \'hard to reach\' and provide an equitable health service for all, the NHS must embrace some of the philosophies that appear to underpin the VCS. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= SF conducted the interviews and drafted the original manuscript. BH participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1472-6963/10/92/prepub> Acknowledgements ================ This study was funded by The University of Birmingham Research Support Facility 2 year fellowship (May 2006 - September 2008) awarded to SF. The RSF is an academic support unit based in the Department of Primary Care and General Practice. It is funded by the National Co-ordinating Centre for Research Capacity Development (NCCRCD) to contribute to research capacity building by supporting individuals with the potential to develop a career in primary care research.

{ "pile_set_name": "PubMed Central" }

Introduction {#s1} ============ Adaptive phenotypic plasticity across a range of biological traits, from individual cell form and function to advanced learning, enables individuals to cope with fluctuating environments ([@bib64]; [@bib59]). This flexibility is well-recognised in the diploid life stage, where a complex multicellular organism can generate adaptive plasticity in behaviour, morphology and physiology. However, far less is known about what plastic adaptive responses are possible at the seemingly simple unicellular gamete stage, when environmental variation can be profound. We therefore assessed whether males and females can adaptively vary sperm and egg function through gametogenesis in anticipation of impending functional environment. Although sperm and eggs have universal primary roles that are vital for reproductive success, how they achieve these and the environments in which they must succeed can fluctuate considerably, both biotically and abiotically. Sperm almost always operate after ejaculation and release in a non-self and demanding environment, either within the female reproductive tract or through external fertilisation, and many factors that directly influence sperm function can vary profoundly across these environments ([@bib65]; [@bib71]). Ova face a similarly challenging set of biotic and abiotic variables, especially in the many species that are oviparous where zygotes and embryos develop in eggs to hatch in the natural environment outside the mother ([@bib43]; [@bib37]). These intrinsically variable environments for gamete function could lead to selection for adaptive plasticity, allowing males and females to improve their reproductive fitness by matching sperm and egg phenotypes through development in anticipation of different fertilisation and embryogenesis environments. One of the most important abiotic environmental variables is temperature ([@bib12]; [@bib4]), especially in the context of climate change when thermal environments are expected to both warm and become much variable and extreme ([@bib62]; [@bib68]). Thermal variation has profound impacts on living systems, and numerous examples of adaptive plastic responses to temperature variability have been described, from acclimated mitochondrial function ([@bib63]) and sex determination ([@bib95]), up to complex shifts in behaviour ([@bib60]) and phenology ([@bib94]). Temperature also influences gamete function across a number of levels, with sperm production and function being especially sensitive to warming ([@bib82]; [@bib75]) and egg development being directly influenced by thermal regime ([@bib38]; [@bib39]). Here, we therefore use a combination of experimental approaches (overview on [Figure 1](#fig1){ref-type="fig"}) to examine whether males and females adaptively vary sperm and egg biology through gametogenesis at different temperatures, in anticipation of varying thermal environments for fertilisation and reproduction. We test for this potential in the flour beetle *Tribolium castaneum* which, like most sexually reproducing animals, is both ectothermic and oviparous ([@bib84]), and in which spermatozoal sensitivity to temperature is known ([@bib75]). After exposing adult males and females and their gametes to different temperatures, we compare sperm and egg development and reproductive function within thermal regimes that mimic the increasingly variable conditions faced by warmer tropical regions ([@bib62]). We find that thermal regime regulates gamete size development: in the short-term, as the developmental environment warms, eggs get larger and sperm become smaller. Having established these opposing male-versus-female responses, we use reciprocal transplant experiments to measure whether adaptive gamete plasticity exists. We find that warmer reproductive environments present challenges for reproduction. However, gamete plasticity enabled significant improvements in reproductive performance. Our findings reveal a potentially important route by which ectothermic populations can buffer their reproductive output against increasingly stressful and unpredictable temperature fluctuations under climate change. ![Experimental design.\ Overview of experiments investigating thermal adaptation and plasticity in sperm and egg biology and its adaptive significance. (**a**) Set up of long-term thermal selection lines (TSLs) maintained at 30°C or 38°C for 50+ non-overlapping generations with eight replicate populations per regime. (**b**) Experimental design to investigate gamete size divergence in 30°C and 38°C TSLs at their long-term evolving environments, short-term plasticity in gamete size measured after a single-generation of novel temperature exposures, and short-term plasticity in individuals from the ancestral KSS (Krakow Super Strain) population. Symbol fill colour represents long-term background (30°C TSLs in blue, 38°C TSLs in red and KSS in yellow) while outline colour represents short-term exposure temperature (30°C blue, 38°C red). bi) and bii) illustrate experiments on temporal patterns of short-term plasticity in egg size (see main text). (**c**) Experiments investigating the adaptive significance of sperm and egg morphological plasticity in KSS adults. Gametic divergence was achieved by having adults produce gametes at either 30°C or 38°C, whose performance was then tested at 30°C, 38°C or 40°C.](elife-49452-fig1){#fig1} Results {#s2} ======= Sperm morphological divergence and plasticity {#s2-1} --------------------------------------------- Following 54 generations of experimental evolution under increased temperature, we found that at both development temperatures sperm length differed by an average of \~4% between long-term selection regimes, with males from lines evolved at 38°C producing significantly longer sperm than males from lines evolved at 30°C (effect size *β* = 3.4, *t*~1,14~ = 3.4, p=0.004; eight lines per regime, five males per line, and five sperm per male, [Figure 2](#fig2){ref-type="fig"} and [Table 1](#table1){ref-type="table"}). In contrast, a within-generation increase in temperature during pupation and post-eclosion maturation reduced sperm length development irrespective of evolutionary background, with sperm produced at 38°C being \~7% shorter, and indicating developmental plasticity in sperm size (*β* = −6.3, *t*~1,24~ = -11.2, p\<0.001; [Figure 2](#fig2){ref-type="fig"}). The interaction between selection regime and the development environment was not significant (*β* = 0.5, *t*~1,24~ = 0.4, p=0.672). Male body sizes of 30°C and 38°C thermal selection lines (TSLs) did not differ (*F*~1,78~ = 0.38, p=0.53), and similarly we found no evidence for body size divergence between 30°C or 38°C thermal environments from the pupal stage in the ancestral stock population (the Krakow Super Strain, KSS) (*F*~1,46~ = 0.90, p=0.35). ![Long-term divergence and short-term plasticity in sperm size.\ Symbol fill colour represents long-term background (30°C TSLs in blue, 38°C TSLs in red and KSS in yellow) while outline colour represents short-term exposure temperature (30°C blue, 38°C red; [Figure 1b](#fig1){ref-type="fig"}). (**a**) Sperm length in µm ± SEM of the experimentally evolved TSLs, measured either following production from pupation in their long-term thermal environment (highlighted in shaded boxes) or at the reciprocal temperature. (**b**) Sperm length of mature KSS males exposed to either 30°C or 38°C from pupation through eclosion.](elife-49452-fig2){#fig2} 10.7554/eLife.49452.004 ###### Sperm and egg length in relation to long-term selection and short-term exposure. Line Temperature Sperm length N Egg length N ---------- ------------- -------------- ---- -------------- ----- 30°C TSL 30°C 88.4 ± 5.2 40 615.4 ± 51.5 240 30°C TSL 38°C 81.9 ± 7.9 40 685.5 ± 42.2 240 38°C TSL 30°C 91.5 ± 4.9 40 638.4 ± 48.3 240 38°C TSL 38°C 85.5 ± 4.3 40 682.7 ± 42.7 240 KSS 30°C 89.3 ± 4.1 26 662.6 ± 39.4 180 KSS 38°C 84.3 ± 4.8 26 697.9 ± 48.4 180 Shown are mean, standard deviation and sample size (sperm: number of males; eggs: number of eggs) for sperm length and egg length measured in individuals from temperature selection lines (TSL) and the ancestral population (KSS), exposed to different temperatures from pupation onward (see main text and [Figures 1](#fig1){ref-type="fig"}--[3](#fig3){ref-type="fig"}). Egg morphological divergence and plasticity {#s2-2} ------------------------------------------- Egg size also showed divergence and plasticity in relation to thermal regime. In contrast to sperm, egg size showed an increase in response to a hotter proximate temperature (*β* = 70.0, *t*~1,14~ = 13.9, p\<0.001; [Figure 3a](#fig3){ref-type="fig"} and [Table 1](#table1){ref-type="table"}). However, there was a significant interaction between long-term selection regime and short-term temperature exposure (*β* = −25.7, *t*~1,14~ = -3.6, p=0.003; [Figure 3a](#fig3){ref-type="fig"}). Following 58 generations of experimental evolution (logistic contraints prevented simultaneous measurement at generation 54 when sperm lengths were assayed), eggs of females from TSLs evolved at 38°C were larger than those of 30°C TSL females when produced at 30°C (*β* = 23.0, *t*~1,14~ = 3.2, p=0.006), but were very similar in size when produced at 38°C (*β* = −2.8, *t*~1,14~ = -0.4, p=0.677). ![Long-term divergence and short-term plasticity in egg size.\ Symbol fill colour represents long-term background (30°C TSLs in blue, 38°C TSLs in red and KSS in yellow) while outline colour represents short-term exposure temperature (30°C blue, 38°C red; [Figure 1b](#fig1){ref-type="fig"}). (**a**) Egg length in µm ± SEM of the experimentally evolved TSLs, measured either following production in their long-term thermal environment (highlighted in shaded boxes), or at the opposite temperature. (**b**) Egg length produced by mature KSS females exposed to either 30°C or 38°C for mating and oviposition, measured across three experimental replicates (grey lines). (**c**) Egg length of KSS females mated to a standard KSS male ovipositing individually either at 30°C or 38°C for 10 days in two-day intervals. The two bold-face lines indicate the averages across all females within a given treatment, while thin lines show average values for individual females. (**d**) Egg length of groups of 50 females, mated to standard KSS males, and ovipositing alternately at 30°C and 38°C for 12 days in two-day intervals. Egg width did not differ between thermal regimes, demonstrating that oviposition temperature increased egg volumes ([Figure 3---figure supplements 1](#fig3s1){ref-type="fig"} and [2](#fig3s2){ref-type="fig"}).](elife-49452-fig3){#fig3} Egg size was also thermally plastic in standard ancestral stock Krakow Super Strain (KSS) females, and showed significant divergence according to adult rearing and oviposition temperature when assessed in three experimental repeats. KSS females produced larger eggs when ovipositing at 38°C compared to 30°C (*β* = 35.3, *t*~1,356~ = 8.2, p\<0.001; 50 females per group and 60 eggs measured per group at either rearing temperature, [Figure 3b](#fig3){ref-type="fig"}). In addition, this plasticity was shown by individual females ovipositing alone at either 30°C or 38°C (*β* = 81.5, *t*~1,42~ = 11.6, p\<0.001; [Figure 3c](#fig3){ref-type="fig"}), and was reversible when females were alternated between 30°C and 38°C thermal environments (*β* = 58.9, *t*~1,4~ = 7.3, p=0.002; [Figure 3d](#fig3){ref-type="fig"}). Egg width was not different between the oviposition temperatures (*F*~1,198~ = 0.1, p=0.686, [Figure 3---figure supplements 1](#fig3s1){ref-type="fig"} and [2](#fig3s2){ref-type="fig"}) and the interaction between oviposition temperature and egg length was not statistically significant (*t* = −0.77, p=0.44), demonstrating that oviposition temperature increased egg volumes (see [Figure 3---figure supplement 2](#fig3s2){ref-type="fig"} for volume calculations). Adaptive sperm plasticity {#s2-3} ------------------------- As for the selection lines, sperm size was thermally plastic in ancestral stock KSS males, and showed significant divergence according to rearing temperatures (two experimental repeats), with KSS males producing significantly shorter sperm when reared at 38°C compared to 30°C (*β* = −5.1, *t*~1,49~ = -7.0, p\<0.001; [Figure 2b](#fig2){ref-type="fig"} and [Table 1](#table1){ref-type="table"}). To test for the adaptive value of any functional plasticity, the performance of sperm from males of the same ancestral KSS population reared from pupation at either 30°C or 38°C was tracked by comparing the total number of offspring sired across 100 days of oviposition by KSS females at either 30°C or 38°C ([Figure 4a](#fig4){ref-type="fig"} and [Table 2](#table2){ref-type="table"}), by which time all females in the experiment had exhausted their viable sperm stores ([@bib58]) and ceased to produce offspring ([Figure 4b and c](#fig4){ref-type="fig"}). Reproduction was generally sensitive to the proximal thermal environment, with on average 299 (± 20.8, mean ± s.e.m.) offspring eclosing at 30°C, compared with 135 (± 10.3, mean ± s.e.m.) at 38°C (*β* = −5.8, *F*~1,104~ = 69.8, p\<0.001). Despite these overall differences, it was clear that the thermal regime in which sperm production and insemination took place also gave individual males a reproductive advantage when matched to the same thermal environment for fertilisation and offspring development ([Figure 4a](#fig4){ref-type="fig"}). Indeed, the interaction between spermatogenesis temperature (sperm production) and offspring production temperature was highly significant (*β* = 7.5, *F*~1,104~ = 29.6, p\<0.001), while the main effect of male treatment was not significant (*β* = 0.7, *F*~1,104~ = 1.1, p=0.290). Sperm from males exposed as pupae and maturing adults to 30°C achieved greater reproductive success in the 30°C reproductive environment than sperm from males that completed development at 38°C. By contrast, in the 38°C reproductive environment sperm produced by males in the 38°C treatment achieved double the reproductive success compared with sperm from males developed through the 30°C treatment. Across 100 days of oviposition in the 30°C reproductive environment, males reared at 30°C sired an average of 349 (± 29.1, mean ± s.e.m.) offspring, while males reared at 38°C sired 249 (± 27.1, mean ± s.e.m.) offspring. Using the same protocols in the 38°C reproductive environment, males reared at 38°C sired 180 (± 12, mean ± s.e.m. offspring), while males reared at 30°C sired only 91 (± 12, mean ± s.e.m.) offspring ([Figure 4a](#fig4){ref-type="fig"} and [Table 2](#table2){ref-type="table"}). Exploration of the temporal patterns of these effects showed a clear decline in reproduction across consecutive 10-day blocks, and that sperm matched to the fertilisation and development temperature consistently outperformed sperm that were thermally mismatched ([Figure 4b and c](#fig4){ref-type="fig"}). Our model comparison based on AIC values confirmed that the interaction between male rearing temperature and fertilisation temperature was important: the best model included the main effects of male temperature, offspring temperature and time, and the interaction between male and offspring temperature for both parts of the model, and additionally the interaction between rearing temperature and time for the zero-inflation model ([Supplementary files 1](#supp1){ref-type="supplementary-material"} and [2](#supp2){ref-type="supplementary-material"}). ![Adaptive thermal plasticity in sperm.\ Reproductive output of ancestral KSS males following a 24 hr mating bout with a single female. Symbol outline colour represents male short-term exposure temperature (30°C blue, 38°C red) while background colour indicates fertilisation and offspring development temperature (30°C blue, 38°C red; see [Figure 1c](#fig1){ref-type="fig"}). (**a**) Total offspring produced over a 100d period across ten 10 day blocks from sperm produced in either 30°C or 38°C conditions when functioning in either 30°C or 38°C reproductive environments. Temporal patterns in 30°C and 38°C environments are illustrated in (**b and c**), respectively (note the log-scale of the Y-axis). Analyses of individual male reproductive performance and average sperm length across a range of thermal regimes indicate a longer-sperm advantage in this system ([Figure 4---figure supplement 1](#fig4s1){ref-type="fig"}).](elife-49452-fig4){#fig4} 10.7554/eLife.49452.010 ###### Adaptive thermal plasticity in sperm and eggs improves gamete performance. Sperm Eggs ----- ------ ------ --------------- ------ --------- --- KSS 30°C 30°C 349.5 ± 151.5 27 90 ± 2% 8 KSS 38°C 30°C 248.8 ± 140.9 27 80 ± 5% 8 KSS 30°C 38°C 91.0 ± 61.8 27 78 ± 8% 8 KSS 38°C 38°C 179.6 ± 61.8 27 80 ± 3% 8 KSS 30°C 40°C 27 ± 5% 8 KSS 38°C 40°C 36 ± 5% 8 Sperm performance was measured by mating focal males to tester females and counting all offspring produced over a 100d period. Egg performance was measured as the proportion of eggs that developed into adult offspring, tested in groups of 50 eggs (see main text and [Figures 3](#fig3){ref-type="fig"} and [4](#fig4){ref-type="fig"}). Adaptive egg plasticity {#s2-4} ----------------------- The performance of eggs produced by ancestral stock KSS females (mated to 30°C-reared KSS males) ovipositing at either 30°C or 38°C was compared by measuring egg-to-adult offspring viability of pre-counted groups of 50 eggs when incubated and reared at either 30°C, 38°C or 40°C ([Table 2](#table2){ref-type="table"}). There was a significant interaction between oviposition thermal regime and the environmental temperature treatments at which the eggs were incubated, and offspring hatched and developed (*z* = −4.1, p\<0.001; [Figure 5](#fig5){ref-type="fig"}). Thus, there was similar evidence of adaptive egg plasticity as for sperm, although this was only evident in the 30°C and 40°C environmental treatments ([Figure 5](#fig5){ref-type="fig"} and [Table 2](#table2){ref-type="table"}). *Post hoc* testing showed that at 30°C eggs oviposited at 30°C achieved significantly greater egg-to-adult offspring viability rates than eggs oviposited at 38°C (*z* = 4.6, p\<0.0001), whereas the opposite was true at 40°C (*z* = 3.01, p=0.01) where 30°C-oviposited eggs resulted in 25% fewer offspring than 38°C-oviposited eggs. In the 38°C environment, there was no evidence for any adaptive plasticity in egg biology, with the same relative number of eggs from the 30°C and 38°C regime females hatching and producing adult offspring ([Figure 5](#fig5){ref-type="fig"} and [Table 2](#table2){ref-type="table"}). ![Adaptive thermal plasticity in eggs.\ Symbol fill colour represents production and oviposition temperature (30°C blue, 38°C red) for ancestral KSS females while outline and background colour indicate egg incubation, offspring hatching and development temperature (30°C blue, 38°C red, 40°C dark red; see [Figure 1c](#fig1){ref-type="fig"}). Egg-to-adult viability was measured in a fodder medium with 0% yeast, with eight replicate groups of 50 eggs per treatment combination. Point surface area is proportional to the number of observations with identical outcomes.](elife-49452-fig5){#fig5} Discussion {#s3} ========== Our experiments revealed that: (1) gamete function and reproduction is highly sensitive to the local thermal environment; (2) developmental plasticity exists within sperm and eggs in response to temperature; (3) plastic responses in gamete size proceed in different directions in either sex; and (4) gamete plasticity is adaptive, enabling males and females to significantly improve their reproductive success via mechanisms that match sperm and egg development to the imminent thermal environment in which they must function. Sperm production and function is known to be affected by many environmental variables ([@bib65]; [@bib71]), with particular sensitivities to temperature ([@bib82]; [@bib17]), which may be especially important in ectotherms where environmental temperature varies through the reproductive window ([@bib71]; [@bib93]). Such variation will directly influence important elements within the sperm storage and fertilisation environment ([@bib65]; [@bib71]), including biophysical properties of fluids and membranes ([@bib53]), mitochondrial metabolic sensitivity ([@bib85]), flagellar function ([@bib48]) and haplotype integrity ([@bib61]). Our transplant experiments reveal obvious challenges for gamete performance in warmer environments, with an overall halving of reproductive output when sperm and eggs are challenged to function at 38°C versus 30°C ([Figures 4](#fig4){ref-type="fig"} and [5](#fig5){ref-type="fig"}). Interestingly, between-line variance in sperm length also increased when males were forced to develop sperm at 38°C ([Figure 2a and b](#fig2){ref-type="fig"}), possibly indicating a response to thermal stress at the higher temperature. Although we found evidence that laboratory selection across 54 generations at 38°C caused the overall evolution of \~4% longer sperm compared with the thermal line selection at 30°C, our within-generation experiments revealed consistent \~7% reductions in sperm length for all populations when produced through the warmer 38°C regime compared with 30°C ([Figure 2](#fig2){ref-type="fig"}). The different sperm length responses to experimental evolution versus within-generation plasticity under different temperatures is puzzling, and may have arisen through indirect selection acting on the thermal lines for \~five years, given that the 38°C thermal selection lines showed generally longer sperm at both 30°C and 38°C rearing temperatures ([Figure 2a](#fig2){ref-type="fig"}). Recent work examining sperm length evolution in *T. castaneum* revealed that a history of heightened sexual selection led to the evolution of longer sperm ([@bib41]), so one possibility is that increased metabolic and developmental rates in the 38°C regime had elevated mating activity, and therefore promoted male-male competition and selection for increased sperm length. However, our current results ([Figure 4](#fig4){ref-type="fig"}) and past findings ([@bib75]) indicate that warmer conditions generally reduce male reproductive fitness, and we find no evidence that operational sex ratios have deviated between our 30°C or 38°C selection line regimes, with both showing an average adult sex ratio of 50% (54 adults randomly sampled from six lines per regime: 30°C regime male ratios = 0.52, 0.46, 0.50, 0.52, 0.48, 0.52; 38°C regime male ratios = 0.50, 0.59, 0.48, 0.48, 0.52, 0.43); offspring sex ratios are also not changed following male exposure to 42°C heatwaves ([@bib76]). Another possible explanation is that increased developmental rate at 38°C has hastened sperm ageing, leading to correlated changes in length. However, *T. castaneum* is a relatively long-lived insect, with adult males showing no change in fertility even after one year of lifespan ([@bib40]), so reproductive ageing differences over the few days of the sperm length assays conducted here are unlikely to explain the overall sperm length increases in the thermal selection lines at 38°C. Another developmental possibility is that elevated spermatogenic rates at 38°C leads to the production of smaller cells, and longer-term selection has compensated for this reduction in sperm cell size within warmer regimes to evolve longer sperm. There is some evidence for longer-sperm advantage in *T. castaneum* ([@bib41]), and our results for individual male fertilisation success across a range of thermal regimes suggest that males producing longer sperm have improved reproductive fitness ([Figure 4---figure supplement 1](#fig4s1){ref-type="fig"}). Additional possibilities for indirect effects acting on sperm length in the 38°C selection regime include genetic bottlenecking at the start of selection if fertility and offspring production had been compromised, or heightened metabolic rate throughout the entire life cycle allowing improved access to key nutrients which may enable the development of longer sperm ([@bib41]). Whatever the cause behind long-term experimental evolution of sperm length within our thermal lines, it is clear that short-term within-generation impacts of temperature have strong and direct effects on sperm development, with an experimental switch to warmer regimes for either the thermal selection lines or the ancestral stock population resulting in consistently clear reductions in sperm length ([Figure 2a and b](#fig2){ref-type="fig"}). Temperature variation during gametogenesis is known to affect sperm size in ectotherms ([@bib7]; [@bib72]), and cool and warm thermal extremes can reduce fertile function in tandem with reduced sperm length development ([@bib90]). However, no previous study has revealed that the thermal regime during spermatogenesis can shape sperm function to be optimal for the forthcoming thermal reproductive environment. We demonstrate clear evidence for adaptive plasticity in sperm function in anticipation of thermal regime, enabling males (and their mates) to increase their reproductive success by 40% to 100% when males are developmentally 'matched' to the temperature of the subsequent reproductive environment ([Figure 4](#fig4){ref-type="fig"}). Adaptive plasticity in the production of sperm numbers is known in relation to environmental risks of male-male competition ([@bib96]), with the capability for spermatogenesis to increase in response to elevated risks of sperm competition ([@bib70]; [@bib36]). This male plasticity can also change individual sperm cell form and function: domestic cockerels (*Gallus gallus domesticus*) rapidly changed their sperm mobility in relation to their own competitive status ([@bib66]), and male Gouldian finches (*Erythrura gouldiae*) adjusted sperm morphometry in relation to social factors ([@bib49]). Similar changes occured in sperm velocity and density in Arctic charr (*Salvelinus alpinus*) when dominance switched ([@bib74]), and broadcast-spawning ascidians (*Styela plicata*) produced larger, more motile sperm when adults were kept at densities with greater risks of sperm competition ([@bib15]). Far fewer studies have explored adaptive sperm plasticity in relation to abiotic variation, which is a gap because physico-chemical factors can greatly influence sperm function and also vary across reproductive environments ([@bib71]). Acclimation and in vitro fertilisation experiments with the broadcast spawning tubeworm *Hydroides diramphus* revealed adaptive plasticity in sperm and egg function in relation to salinity, with gametes performing best at salinities experienced by their parents prior to spawning ([@bib51]). Likewise, sticklebacks (*Gasterosteus aculeatus*) showed adaptive plasticity in spermatozoal sensitivity to the salinity and osmolarity signal for initiating flagellar motility ([@bib88]). To our knowledge, only two studies have investigated sperm plasticity in relation to temperature: male mosquitofish (*Gambusia holbrooki*) acclimated to cool (18°C) and warm (30°C) regimes for five weeks showed no signs of spermatozoal acclimation or change in thermal limits ([@bib2]). Likewise, although warm conditions reduced sperm motility in brown trout (*Salmo trutta*), warm-acclimated males did not produce sperm with improved relative motility ([@bib26]). Our transplant experiments revealed clear evidence for adaptive developmental plasticity in sperm function according to thermal regime, with males exposed to warmer 38°C temperatures producing sperm that enabled a doubling of reproductive success in warmer 38°C fertilisation and reproductive environments compared with sperm produced at 30°C ([Figure 4a and c](#fig4){ref-type="fig"}). The opposite also applied, with sperm produced in cooler 30°C regimes gaining \~40% greater reproductive success at 30°C compared with sperm produced at 38°C ([Figure 4a and b](#fig4){ref-type="fig"}). This plasticity will confer direct fitness advantages if the thermal regime through development before mating anticipates the temperature for sperm function and reproduction following insemination. Sperm manufacture proceeds rapidly in *T. castaneum*, with production of mature, functional gametes taking around four days at 30°C ([@bib31]). As in many insects, fertilisation and oviposition proceeds within hours of mating in *T. castaneum* ([@bib25]), so sperm production temperature will usually predict insemination, storage and fertilisation temperature. In contrast to males, female *T. castaneum* produced larger gametes in the warmer environment ([Figure 3](#fig3){ref-type="fig"}), which is unusual for arthropods where smaller eggs are usually produced as temperatures increase ([@bib33]); but see [@bib81] and [@bib87]. This egg size plasticity occurred within two days of exposure to the novel thermal regime, and was reversible ([Figure 3d](#fig3){ref-type="fig"}). Egg plasticity was also adaptive when comparing performance between the more extreme 30°C versus 40°C environments ([Figure 5](#fig5){ref-type="fig"}): 36% of eggs produced at 38°C generated viable offspring following development at 40°C compared with only 27% of eggs produced at 30°C. Conversely, 90% of eggs produced at 30°C generated offspring when developed at 30°C, versus 80% of eggs developed at 38°C ([Figure 5](#fig5){ref-type="fig"}). Adaptive plasticity in egg biology is known in relation to a number of environmental factors, but we believe this is the first study to demonstrate it under environmental warming. Adaptive egg plasticity in relation to biotic variation is shown by seed beetles (*Stator limbatus*), where females vary egg size in relation to anticipated host plant quality ([@bib32]). Experiments show that females respond to a switch in host plant in a manner that maximises reproductive fitness by matching egg size (and number) to the host type ([@bib32]; [@bib77]). Likewise, female cowpea weevils (*Callosobruchus maculatus*) are sensitive to levels of larval competition predicted by increased adult density, laying larger eggs to improve larval fitness ([@bib52]). Broadcast-spawning ascidians (*S. plicata*) produce smaller eggs in high density populations, but their embryo-yielding ovicells are larger than eggs from low-density adults ([@bib15]). Similarly, female *Nasonia vitripennis* wasps adjust the sex ratio of their broods depending on whether they are first or second to parasitise a host, improving offspring fitness according to anticipated variation in local mate competition ([@bib97]). Adaptive egg plasticity in relation to some abiotic variables is also recognised. Female stink bugs (*Podisus maculiventris*) detect reflectance at the site where they oviposit, and invest more protective pigment into eggs that will be exposed to stronger ultraviolet solar radiation ([@bib1]). A number of studies have found variation in egg phenotypes according to temperature during development and oviposition (e.g. [@bib33]; [@bib24]; [@bib6]), but few have identified that the changes are adaptive, with thermal variation creating physiological constraints or stress during egg production ([@bib33]). A notable exception is in *Bicyclus anynana* butterflies, where females lay larger (and fewer) eggs when they are exposed to lower temperatures during oviposition ([@bib27]); reciprocal transfer experiments, as we employ here, show that this behaviour is adaptive ([@bib28]). We identify adaptive plasticity in response to the upper ranges of reproductive tolerance by *T. castaneum* females, with adaptive matching through warmer egg development temperatures enabling a 33% improvement in offspring production in the hottest reproductive environment ([Table 2](#table2){ref-type="table"}, [Figure 5](#fig5){ref-type="fig"}). Mechanisms for optimising gamete function in different thermal environments are to be uncovered, but four broad and related possibilities exist through: 1) optimising size, 2) physiological matching, 3) haploid selection, and/or 4) epigenetic modifications for offspring development. Although we observed opposing responses by sperm and egg sizes under temperature variation, it seems unlikely that this phenotypic variation is solely responsible for enabling improvements in reproductive success. Sperm length decreased as temperatures increased, but additional correlative analyses gave no indication that reduced sperm size per se improved reproductive performance in hotter thermal regimes, and the reverse was more evident ([Figure 4---figure supplement 1](#fig4s1){ref-type="fig"}), consistent with previous work showing that sperm elongation is costly and positively selected by competition ([@bib41]). It therefore seems more likely that changes in sperm physiology, rather than morphometry, allow plasticity in sperm function to match thermal reproductive environments. Thermal adaptations influencing cell physiology and biochemistry are known to exist within mitochondrial metabolism ([@bib23]; [@bib42]), essential for sperm function ([@bib69]), and membrane properties influencing cell structure and physiology ([@bib13]) for sperm and flagellum function ([@bib9]). In addition, Heat Shock Proteins, expressed throughout spermatogenesis in the testis, play key roles in sperm development ([@bib22]). These adaptations for different thermal and hydrodynamic environments could be adaptively varied through spermiogenesis so that sperm function is matched to challenges facing sperm migration, storage and fertilisation in warmer environments. Within-ejaculate haplotype selection provides another mechanism for improving offspring performance ([@bib3]). Adaptive thermal plasticity may be possible via changes in egg size, because larger-volume eggs produced at higher temperatures will contain more fluids, and possibly a greater nutrient load, both of which could improve offspring viability under heat stress and desiccation ([@bib29]). However, within the 30°C environment, larger eggs produced at 38°C were outperformed by smaller eggs produced at 30°C, suggesting additional mechanisms beyond size benefits. As for sperm, thermally plastic traits that are essential for egg fertilisation and subsequent function could include differential mitochondrial activity ([@bib21]) and egg plasma membrane properties ([@bib86]). Heat Shock Proteins could also play vital roles in protecting egg development from thermal stress; recent work demonstrates that increased loading of HSP23 into *D. melanogaster* eggs improves embryo thermal tolerance ([@bib55]). In addition to the potential for adaptive plasticity within gamete function, reproductive fitness could be improved if short-term epigenetic modification through gamete development can pass adaptive information to the zygote, embryo and offspring ([@bib35]). The potential for adaptive transgenerational plasticity via the matriline and through transcription factors within eggs is increasingly recognised ([@bib44]), potentially enabling rapid responses to climate change via adaptive plasticity ([@bib18]). There is growing evidence that sperm also have the potential to be transcriptionally labile ([@bib50]), passing environmentally-driven epigenetic information to offspring through histone or protamine modifications, haplotype DNA methylation remodelling, and/or small RNAs ([@bib20]). Sperm carry complex payloads of coding and non-coding small RNAs which can be transcribed into the oocyte and embryo ([@bib16]; [@bib46]) with conserved functions across mammalian and insect models ([@bib30]); the potential importance of these RNA transcripts remains largely unexplored ([@bib11]; [@bib10]; [@bib83]). However, the possibility for transgenerational information transmission via sperm or egg epigenomes in response to environmental variation during gametogenesis is an obvious mechanism to enable adaptive thermal plasticity for populations facing the challenge of reproducing under climate change where increases in both thermal averages, maxima and variation are expected ([@bib62]; [@bib68]). Materials and methods {#s4} ===================== -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Reagent\ Designation Source or reference Identifiers Additional\ type (species)\ information or resource --------------------------------------------------- ---------------------------------------------- ----------------------------------------------------- ---------------------------------------------- ------------------------------------------------------------------ Strain, strain background (*Tribolium castaneum*) Krakow\ KSS: [@bib19] [@bib76]\ KSS\ Live beetles Super Strain ancestral stock\ TSL: this paper and [@bib19] [@bib76] &\ and Thermal Selection Lines at 30°C and 38°C TSL30 or TSL38 Biological sample (*Tribolium castaneum*) Spermatozoa and ova This paper Sperm and eggs Sperm recovered from sacrificed live males, eggs recovered from\ oviposition food medium Software, algorithm R Studio R Studio ([@bib73]) in R ([@bib67]) (version 3.4.1) R Studio version 1.1.463 and R version 3.4.1 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Model system, selection lines and thermal exposure {#s4-1} -------------------------------------------------- *Tribolium castaneum* flour beetles were maintained as previously described (e.g. [@bib56]; [@bib41]) in a 16 hr: 8 hr light: dark photoperiod at 60 ± 10% relative humidity in small stock populations in *ad libitum* fodder comprising 90% organic white flour, 10% brewer's yeast and a sprinkling of rolled oats to aid traction. Approval for the research was granted by UEA\'s Animal Welfare and Ethical Review Board. To identify males from females in mating pairs, individuals were marked with a small dot on the dorsal thorax with correction fluid (Tippex, France) ([@bib92]). We used beetles from the Krakow Super Strain (KSS) in this study, a wild-type outbred lab strain created in 2008 by breeding together eleven different strains, and maintained since at 30°C (see [@bib19]). The thermal selection lines were established in 2010 and maintained since as eight independent replicates for either regime with synchronous generations at either 30 ± 1°C or 38 ± 1°C. Each line was reproduced through every adult generation using 100 haphazardly-selected, sexually mature adults (\>ten days since pupal emergence). The mixed adults were free to mate and oviposit in 150 ml of fodder in 300 ml jars with mesh lids for one week, after which they were removed and the resulting eggs and larvae allowed to develop to the next generation at their respective temperatures. After 50+ generations of experimental evolution, gametes from males and females were assayed from eight independent lines in either thermal regime. To measure the impacts of thermal regime on gamete biology, beetles were exposed to either 30°C, 38°C or 40°C temperatures in a controlled environment facility. Thermal exposure was regulated within a controlled environment facility held at 30°C, or in A.B. Newlife 75 Mk4 forced air egg incubators (A.B. Incubators, Suffolk, UK) at 38°C within the facility. 30°C is our standard rearing temperature, and the optimum laboratory temperature for population productivity in *T. castaneum* is 35°C ([@bib84]; [@bib76]). 38°C approaches the upper thermal limit across more sensitive juvenile stages ([@bib19]; [@bib76]), and this temperature has been recorded in 150 countries ([@bib57]). Our experimental exposure to this thermal regime will have relevance to more extreme thermal conditions and heatwaves, predicted to increase in frequency, severity and duration under climate change (e.g. [@bib62]), and which we know specifically constrain reproduction and sperm function in this system ([@bib75]). Gamete measurements {#s4-2} ------------------- Measures of mature sperm were performed following microdissection (see [@bib41]) and eggs were sieved from fodder following oviposition ([Figure 6](#fig6){ref-type="fig"}). Flour beetle eggs are oblong in shape, and size was quantified by measuring length across the long axis of rotational symmetry of the egg. Initial measures of length and width revealed a positive correlation between the two measures across 100 individual eggs, so a single length measure was used to quantify egg size. Eggs were sieved from the fodder using mechanical sieves (pore size: 300 µm, Endecotts Ltd, London, UK), and placed on a dark tile using a fine paintbrush. Eggs from the fodder were coated with a single layer of flour, and were measured at x30 magnification using a Zeiss Discovery V12 binocular microscope, AxioCam MRc5 camera and AxioVision V5.1 imaging software. Mature sperm were recovered from the base of the testicular follicles dissected out of males frozen at --6°C. Follicles were isolated on microscope slides in 30 µl drops of insect saline (0.9% NaCl), and then ruptured using fine forceps under an Olympus SZX9 binocular microscope. Once ruptured, sperm were dispersed by spreading out the area of the saline drop across the microscope slide using fine forceps, and the smears left to dry at room temperature so that the sperm cells lay in flat two-dimensional planes adhering to the glass. Images of intact spermatozoa were captured at 600X magnification under dark-field phase contrast using an Olympus BX41 microscope connected to a GT Vision GX CAM digital camera and GXCapture 8.2 software (GT Vision Ltd, UK). Sperm length was measured by creating a segmented line that traced the entire length of the cell using the 'ImageJ' analysis package and segmented line tool ([@bib80]). Previous work shows this approach has high repeatability ([@bib41]). ![Gamete measurements were performed on mature sperm dissected from males and eggs following oviposition.\ Sperm length was measured at 600X magnification in Image J by drawing segmented lines along the backbone of the cell (**a and b**). Egg lengths were measured along the long axis of each ovoid at 30X magnification (**c and d**). Further details in [@bib41].](elife-49452-fig6){#fig6} Long-term and short-term gamete divergence ([Figure 1a and b](#fig1){ref-type="fig"}) {#s4-3} ------------------------------------------------------------------------------------- We assessed the impact of temperature on development of gamete sizes following both long-term and short-term variation in thermal regime. Responses to long-term variation were measured following five years of selection (50+ generations) within replicate lines maintained in either 'warm' (30 ± 1°C) or 'hot' (38 ± 1°C) conditions. To measure impacts of short-term thermal variation we created duplicates of each of the eight replicate 30°C and 38°C lines at generation 54 to measure sperm effects and 58 for egg effects, and reared these at both 30°C and 38°C ([Figure 1b](#fig1){ref-type="fig"}) using a balanced, factorial design. Logistic contraints prevented simultaneous measurement at generation 54. To measure effects on sperm, males were exposed to either temperature from the pupal stage in 6 cm Petri dishes containing 15 ml of standard fodder, with adults allowed to emerge in groups of 20 per dish at their treatment temperature ([Figure 1b](#fig1){ref-type="fig"}). Ten days later, when reproductively mature, they were frozen for dissection and sperm measurement. We also measured full body size of males from both regimes (*N* = 80; five males from each of eight lines across two regimes). Sperm length variation was measured in five males per replicate line, recovering sperm from frozen males and measuring the total length of five sperm per male (=200 sperm measures from 40 males across eight lines in each of four thermal x selection treatment combinations). To measure temperature effects on eggs, mature adult females from eight replicate 30°C and 38°C lines at generation 58 were allocated to oviposit for two days at the two temperatures in a fully factorial design. Two days after the introduction of adults, we sieved oviposited eggs from the line's fodder, isolating them for measurement. Egg length measurement followed a balanced design across temperature regimes and lines, with 30 eggs measured from each replicate line at either temperature regime (=960 egg measures across eight replicate lines of either selection regime and two short-term temperature exposures). We also measured gamete size plasticity in the ancestral KSS stock population maintained at 30 ± 1°C, following methods for testing gamete size plasticity as above. Males were developed from pupae at either 30°C or 38°C and, ten days following emergence, five sperm and body length were measured from each of 26 males per treatment. Eggs were measured in three experimental blocks, within which two groups of 50 KSS females (eclosed and mated at 30°C with standard KSS males also developed at 30°C) oviposited at either 30°C or 38°C for two days in 100 ml jars with perforated lids and 80 ml of standard fodder. We measured 60 eggs per oviposition group (total *N* = 3600 eggs). Two additional tests measured short-term thermal impacts on egg size: the first measured egg size plasticity within individual females (*N* = 20 at 30°C and *N* = 20 at 38°C), with 40 KSS females mated to individual 30°C-developed KSS males for 24 hr at 30°C, after which 20 each were randomly allocated to oviposit at either 30°C or 38°C in 4 ml vials containing 0.5 g of standard fodder, transferring females to new vials every two days for a total of ten days, and eggs measured (length and width, µm) from vials immediately after females had been transferred on ([Figure 1bi](#fig1){ref-type="fig"}). The second test examined short-term reversibility of egg size in a group of 50 KSS mated females which were alternated for oviposition between 30°C and 38°C thermal regimes every two days for a total of 12 days, starting at 30°C ([Figure 1bii](#fig1){ref-type="fig"}). Females had been mated to standard 30°C-developed KSS males, and kept in 100 ml jars containing 80 ml of standard fodder, and 30 eggs per jar were measured immediately after transfers (total *N* = 180 eggs). Assessing the adaptive significance of gamete plasticity ([Figure 1c](#fig1){ref-type="fig"}) {#s4-4} --------------------------------------------------------------------------------------------- To measure adaptive plasticity in sperm, we reared KSS male pupae through either 30 ± 1°C or 38 ± 1°C temperatures as described above, and then tested the relative performance of sperm from eclosed males within KSS females and ova that were maintained in either 30 ± 1°C or 38 ± 1°C fertilisation and oviposition regimes for 100 days. Male pupae were isolated from the KSS stock population and completed development to eclosion in groups of 20 in 6 cm plastic Petri dishes with *ad libitum* fodder at either 30 ± 1°C or 38 ± 1°C. Three of these male groups (*N* = 60 pupae) were reared and maintained at 30°C, and three (*N* = 60 pupae) at 38°C. When sexually mature at 10 days post eclosion, individual males from either the 30°C or 38°C eclosion regime were paired with similar-aged, marked virgin females from the KSS stock population (reared at 30°C) in 7 ml vials containing 0.5 g of fodder. Pairs were allowed to mate at the male's eclosion temperature. After 24 hr, pairs were separated, and females isolated in individual 6 cm Petri dishes containing 10 g of fodder. Half the mated females within either male thermal treatment group were allowed to oviposit at 30 ± 1°C, and the other half at 38 ± 1°C, with *N* = 27 females in each of the four oviposition groups. Thus, we executed a fully factorial and balanced experiment in which females developed, fertilised and oviposited eggs at either 30 or 38°C, using sperm that had been produced at either 30°C or 38°C. Every ten days, females were transferred to new Petri dishes containing fresh fodder, for a maximum of 100 days (ten x 10 day blocks), by which time females had ceased to produce fertile eggs (following a single mating period, female *T. castaneum* typically use up all viable sperm within 100 days, after which a new mating allows resumption of fertility and offspring production \[[@bib58]\]). The number of adult offspring emerging from each Petri dish across up to 100 days of oviposition at either 30°C or 38°C quantified the reproductive success of each pair, comparing performance of sperm developed at either 30°C or 38°C when challenged with functioning at either 30°C or 38°C. To measure the adaptive significance of egg plasticity, we isolated eggs that had been developed and laid at either 30 ± 1°C or 38 ± 1°C temperatures from KSS adults, and then tested their egg-to-adult viability through either 30 ± 1°C, 38 ± 1°C or 40 ± 1°C thermal regimes. To generate phenotypic divergence in egg biology, groups of 300 females previously mated to KSS males at 30°C oviposited for 2 days at either 30 ± 1°C or 38 ± 1°C (two groups at either temperature) in 1200 ml tubs containing 600 ml of standard fodder. 600 eggs per group were sieved and isolated from the flour, counted, and transferred in clutches of 50 to develop in 100 ml jars containing yeast-free fodder (applying stronger environmental selection on offspring development). Egg clutches produced at either 30°C or 38°C were transferred to hatch and develop at either 30°C, 38°C or 40°C, with eight groups assayed in each of these three temperature treatments (=2400 eggs assayed across a total of 48 treatment groups). Statistical analysis {#s4-5} -------------------- All analyses were carried out using R Studio ([@bib73]) (v 1.1.463) in R ([@bib67]) (version 3.4.1) with *Rmisc* ([@bib45]), *multcomp* ([@bib47]), *car* ([@bib34]), *MASS* ([@bib91]), *glmmTMB* ([@bib8]) and *lmerTest* ([@bib54]) packages for data exploration and analysis. Graphical figures were plotted using *ggplot2* ([@bib98]). Unless otherwise specified, all data were analysed using Linear Mixed Models (LMM) and Generalised Linear Mixed models (GLMM) in *lme4* ([@bib5]), with the specific approach for each set of results described below. All data generated from the experiments described above were included for analysis, and all replication is biological. To determine the appropriate error distributions the relationship between the variance and the mean of the response variable and the assumptions for data distribution were checked ([@bib14]). Models were fitted using Restricted Maximum Likelihood (REML) methods to enable refinement and validation ([@bib89]). Residuals from linear models were checked for violations of the assumptions of normality and homoscedasticity. Significance of fixed effects in LMMs were obtained using t-tests with Satterthwaite\'s approximation for degrees of freedom implemented in *lmerTest* ([@bib54]). To facilitate the interpretation of main effects in the presence of interactions, we centred the contrasts between factors with two levels by coding them as minus 0.5 and 0.5, respectively ([@bib78]). *Total sperm length divergence* following long-term evolution and short-term temperature exposure in the thermal selection lines (TSL) was analysed using an LMM, with the selection regime (30°C or 38°C), the exposure temperature and their interaction entered as fixed factors, and the replicate male (1--5) as a random factor, nested within each of the eight replicate lines. To account for our fully factorial design, we additionally included random slopes for our replicate lines. Divergence in KSS sperm length was analysed using an LMM with thermal environment (30°C or 38°C) as a fixed effect, and male (five sperm from each of 26 males per thermal environment) nested within the two experimental runs as random effects. *Adaptiveness of sperm length plasticity* was assessed using a General Linear Model (LM) on the total number of offspring per mating pair (after square root transformation), with male temperature treatment (30°C or 38°C), fertilisation temperature (regime; 30°C or 38°C) and their interaction as explanatory variables. To additionally explore temporal variation in offspring production, we additionally ran zero-inflated models with gaussian distribution on offspring counts, implemented in *glmmTMB* ([@bib8]). Almost all females ceased to produce offspring in the last few 10-day blocks which has previously been shown to be due to sperm limitation ([@bib58]). We ran a selection of models, and selected the model with the lowest AIC value as our best model ([Supplementary files 1](#supp1){ref-type="supplementary-material"} and [2](#supp2){ref-type="supplementary-material"}). In our conditional full model, we included KSS male treatment, fertilization temperature, time (using blocks as a continuous variable), and two- and three-way interactions as fixed effects. We included random intercepts for 10-day blocks and random slopes for individual pairs to account for repeated measures across time ([@bib79]). Our zero-inflated full model included male treatment, fertilization temperature (regime), time, and two- and three-way interactions as fixed effects. *Egg length divergence* in groups of TSL females was analysed analogous to sperm length, using an LMM with selection regime (30°C or 38°C), short-term exposure temperature and their interaction as fixed effects, and replicate line (1--8) as a random factor, including random intercepts and slopes. Egg length plasticity in the KSS stock following oviposition at either 30°C or 38°C was analysed in an LMM with thermal regime as a fixed effect and experimental block included as a random effect. Plasticity in egg morphology in individual females was analysed in an LMM with thermal regime as a fixed effect and random intercepts for female ID as well as for two-day blocks. We additionally modelled temporal trends by including two-day blocks as a continuous fixed effect, and random slopes for individual KSS females, but found no evidence for temporal trends on egg morphology (p=0.9). To analyse reversibility of egg size plasticity we used an LMM with thermal regime as a fixed effect and two-day blocks as a random effect. Additionally, on a subset of the eggs, we measured egg width to quantify subsequent changes in morphology and overall volume with oviposition temperature (*N* = 20 KSS females per thermal exposure and five eggs per female measured). To assess the correlation between egg length and egg width (egg morphology) at the two oviposition temperatures, a simple LM was fitted to the data with egg width as a response variable, and egg length, oviposition temperature and their interaction as predictor variables. *Adaptive developmental plasticity in egg function* was assessed by testing the reproductive performance of replicate groups of 50 eggs produced by females at either 30°C or 38°C when hatching and developing in 30°C, 38°C or 40°C thermal environments. Egg performance was analysed using a Generalised Linear Model (GLM) with binomial error structure (logit link) in which egg production temperature (30°C or 38°C), developmental thermal environment (30°C, 38°C or 40°C), and their interaction, were entered as fixed effects. We included the number of successes (developed offspring) and failures (eggs that failed to hatch/develop) using the *cbind* function, and confirmed that our model was not overdispersed. Finally, temperature impacts on male body size were analysed using a simple LM. Female body sizes were not assessed because egg size plasticity tests under thermal variation were conducted on already-emerged mature adult females. An overview of sample sizes is given in [Tables 1](#table1){ref-type="table"} and [2](#table2){ref-type="table"}. Box plots indicate the median and interquartile ranges (IQR), with whiskers indicating data within 1.5 IQR. A central filled marker indicates the mean of the sample. Data accessibility statement {#s4-6} ---------------------------- All data generated and analysed in this study, together with R codes, are openly provided as an associated source file in our Dryad Data Repository with the identifier: <https://doi.org/10.5061/dryad.83bg17q>. Funding Information =================== This paper was supported by the following grants: - http://dx.doi.org/10.13039/501100000270Natural Environment Research Council NE/K013041/1 to Matthew Gage. - http://dx.doi.org/10.13039/501100000867Commonwealth Scholarship Commission Commonwealth Rutherford Fellowship to Ramakrishnan Vasudeva. - http://dx.doi.org/10.13039/501100001711Swiss National Science Foundation P300PA_177906 to Andreas Sutter. This work was supported by NERC grant NE/K013041/1, a Commonwealth Rutherford Fellowship (to RV) by the Commonwealth Scholarship Commission, UK, and a Swiss National Science Foundation fellowship (to AS; grant number P300PA_177906). We thank Maya Krishnan Kumar for help with *ggplot2*, and the eLIFE editorial team and three expert referees for reviewing the work in detail and substantially improving the manuscript. Additional information {#s5} ====================== No competing interests declared. Conceptualization, Data curation, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing---original draft, Project administration, Writing---review and editing. Conceptualization, Data curation, Formal analysis, Supervision, Validation, Visualization, Writing---original draft, Project administration, Writing---review and editing. Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing---review and editing. Data curation, Investigation, Methodology, Writing---review and editing. Data curation, Supervision, Investigation, Methodology, Project administration, Writing---review and editing. Conceptualization, Resources, Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing---original draft, Project administration, Writing---review and editing. Animal experimentation: This study was approved by, and followed strict guidelines to, the University of East Anglia\'s Animal Welfare and Ethical Review Board. Additional files {#s6} ================ 10.7554/eLife.49452.013 ###### Table model summary of the best *glmmTMB* model (lowest AIC value; see [Supplementary file 1](#supp1){ref-type="supplementary-material"} Table 2) for reproductive output of males reared at 30°C or 38°C (Treatment) with offspring developing at 30°C or 38°C (Regime). 10.7554/eLife.49452.014 ###### Table overview of models for adaptive significance of sperm plasticity. Models were run using *glmmTMB* and were sorted along ascending AIC values. All conditional models additionally included random intercepts for ten-day blocks as well as random intercepts and random slopes for individual males. 10.7554/eLife.49452.015 Data availability {#s7} ================= Data accessibility statement: All data generated and analysed in this study, together with R codes, are openly provided as an associated source file in our Dryad Data Repository with the identifier: <https://doi.org/10.5061/dryad.83bg17q>. The following dataset was generated: RamakrishnanVasudevaAndreasSutterKrisSalesMatthewJ. G. Gage2019Data from: Adaptive thermal plasticity enhances sperm and egg performance in a model insectDryad10.5061/dryad.83bg17q 10.7554/eLife.49452.019 Decision letter Donoso David Reviewing Editor Escuela Politécnica Nacional Ecuador Gemmell Neil J Reviewer University of Otago New Zealand Pizzari Tom Reviewer University of Oxford United Kingdom In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included. Thank you for submitting your article \"Adaptive thermal plasticity enhances sperm and egg performance in a model insect\" for consideration by *eLife*. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Patricia Wittkopp as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Neil J Gemmell (Reviewer \#2); Tom Pizzari (Reviewer \#3). The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission. Summary: The authors make use of the Tribolium flour beetle model to assess the potential for thermal adaptation in sperm and egg function. Exposing flour beetles to increased temperature under both long-term experimental evolution and short term acclimation, the authors found that thermal regime regulates gamete size development in both males and females. They find that long-term selection regimes at high temperatures (38°C) led to males producing sperm 4% longer than those evolved at lower temperatures (30°C). However, short term changes in temperature generated the opposite pattern: sperm production in males subjected to 38°C within a single generation produced sperm 7% shorter than those exposed to 30°C. Females also showed a response to both the long and short term selection regimes, producing longer eggs at higher temperatures. Using a series of well-conducted thermal translocation experiments, the authors show that this gametic phenotypic plasticity was adaptive. Both sperm and eggs produced in warmer conditions had superior reproductive performance in warmer environments, and vice versa for colder conditions and reproductive environments. Further, in warmer environments, the gamete plasticity observed enabled males to double their reproductive success, and females could increase offspring production by one-third. Collectively the authors\' work identifies a new dimension in gametic plasticity important for understanding how organisms may respond to the increasing thermal variation in the natural environment. This is an excellent study, presenting a uniquely exhaustive experimental investigation into the consequences of environmental temperature for gametic phenotype and related fitness consequences in an ectotherm invertebrate model species. While the biological ramifications of climatic change, including environmental temperature, have been the focus of intense and increasing research interest in recent years, surprisingly little remains known about the impact on gametic performance and associated implications for reproductive fitness. The study therefore makes an influential contribution to a topical and unresolved question. Essential revisions: Perhaps the most intriguing result is the strong contrast in phenotypic response in sperm length under long-term and short-term regimes. Under long term selection to higher temperatures sperm get longer, while under the short term regime, they got shorter. I find this a puzzling result, which I cannot easily reconcile. The authors have put forward some plausible explanations, which I think are sensible, but I did wonder if other factors might contribute to or explain this result. First, was there potential for the sex ratios in the selection lines to have altered as a consequence of thermal exposure, which may have affected the level of sperm competition across some treatments? I wondered if this might be a particular factor in explaining the difference in sperm length response between short-term and long-term selection experiments? I know the authors are expert in this realm, so assume they controlled for this possibility, but some details on whether sex ratios differed and how these were handled might be useful to include. Similarly, I wondered if there were any data on sperm volumes or concentrations. While sperm length clearly responded to the selection regime, I wondered if the authors might also know how sperm number, ejaculate volume and motility altered as there are often strong interactions among these traits? Probably these traits are not easily measured in flour beetles, but if there are data, they might further strengthen this already impressive piece of work. Last, I wondered if there might be an ageing effect that was driving some of the findings observed. In particular, I was curious as to whether males age faster and/or sperm mature faster at higher temperatures. Might part of the effect observed be driven by male age and ejaculate age? Perhaps there is a way to factor in degree days to control for any variation that might exist among experiments? The authors need to justify the experimental temperatures the beetles were exposed to. In special, 38 C is not a yearly average temperature that exists anywhere in the world. Even the lower temperature (30 C) is not a common average temperature anywhere in the world. And very unlikely it will exist, even under global warming scenarios. So, the paper can be better framed in the context of seasonal temperature variability (growing seasons, summer, etc.) and not necessarily in a global warming context. Comparing only two temperatures 30 vs. 38 C is not ideal, besides being temperatures that are very far apart from each other, it is impossible to draw a general trend when you only have one comparison. It would have been great had the manuscript had an intermediate temperature treatment, at 34 C (or so). 10.7554/eLife.49452.020 Author response > Essential revisions: > > Perhaps the most intriguing result is the strong contrast in phenotypic response in sperm length under long-term and short-term regimes. Under long term selection to higher temperatures sperm get longer, while under the short term regime, they got shorter. I find this a puzzling result, which I cannot easily reconcile. The authors have put forward some plausible explanations, which I think are sensible, but I did wonder if other factors might contribute to or explain this result. First, was there potential for the sex ratios in the selection lines to have altered as a consequence of thermal exposure, which may have affected the level of sperm competition across some treatments? I wondered if this might be a particular factor in explaining the difference in sperm length response between short-term and long-term selection experiments? I know the authors are expert in this realm, so assume they controlled for this possibility, but some details on whether sex ratios differed and how these were handled might be useful to include. Similarly, I wondered if there were any data on sperm volumes or concentrations. While sperm length clearly responded to the selection regime, I wondered if the authors might also know how sperm number, ejaculate volume and motility altered as there are often strong interactions among these traits? Probably these traits are not easily measured in flour beetles, but if there are data, they might further strengthen this already impressive piece of work. Last, I wondered if there might be an ageing effect that was driving some of the findings observed. In particular, I was curious as to whether males age faster and/or sperm mature faster at higher temperatures. Might part of the effect observed be driven by male age and ejaculate age? Perhaps there is a way to factor in degree days to control for any variation that might exist among experiments? We have added a major section to the Discussion where we offer additional explanations for the general evolution of longer sperm in the warmer regime, despite short-term responses leading to shorter sperm. We include the reviewer suggestions for factors arising from sexual selection and/or developmental ageing. We also suggest nutrient access through the whole life cycle for spermatogenesis as a potential factor, since nutrient limited males develop shorter sperm in T. castaneum (Godwin et al., 2017). Finally, we consider the possibility that longer-term evolution in a warmer 38^o^C environment, where sperm are pushed by short-term effects (perhaps via faster development rate) to be shorter, has led to compensatory selection for sperm elongation. We do not have data on sperm number/volume/concentration or motility in our thermal lines, because achieving meaningful measures of these pose problems in this system. Individual male T. castaneum can mate with multiple females throughout their long life, successfully inseminating an average of 50 females within 7 days if given the opportunity (Lumley et al., 2015), so individual ejaculate sperm number/volume/concentration will not reveal the capacity for overall sperm number production. (If the reviewer refers to sperm cell volume for consideration, T. castaneum sperm cells are also small and extremely thin, and may vary in diameter along their long axis, so volume will also be a challenge to measure accurately.) Finally, and as in most insects, motility is also a problem to measure meaningfully in T. castaneum because sperm evidently function through the female tract by moving within narrow tubules or within dense sperm masses, so the motility of individual isolated sperm (even if the tract biochemistry could be replicated) on a microscope slide will likely not reflect function in the natural fertilisation environment. We therefore now suggest the following additional interpretations for why the 38^o^C thermal selection lines have evolved slightly longer sperm, whereas short-term temperature increases through pupal and adult development cause a reduction in sperm length: **"**Although we found evidence that laboratory selection across 54 generations at 38°C caused the overall evolution of \~4% longer sperm compared with the thermal line selection at 30°C, our within-generation experiments revealed consistent \~7% reductions in sperm length for all populations when produced through the warmer 38°C regime compared with 30°C (Figure 2). \[...\] Whatever the cause behind long-term experimental evolution of sperm length within our thermal lines, it is clear that short-term within-generation impacts of temperature have strong and direct effects on sperm development, with an experimental switch to warmer regimes for either the thermal selection lines or the ancestral stock population resulting in consistently clear reductions in sperm length (Figure 2). Temperature variation during gametogenesis is known to affect sperm size in ectotherms..." > The authors need to justify the experimental temperatures the beetles were exposed to. In special, 38 C is not a yearly average temperature that exists anywhere in the world. Even the lower temperature (30 C) is not a common average temperature anywhere in the world. And very unlikely it will exist, even under global warming scenarios. So, the paper can be better framed in the context of seasonal temperature variability (growing seasons, summer, etc.) and not necessarily in a global warming context. Comparing only two temperatures 30 vs. 38 C is not ideal, besides being temperatures that are very far apart from each other, it is impossible to draw a general trend when you only have one comparison. It would have been great had the manuscript had an intermediate temperature treatment, at 34 C (or so). Indeed, we agree that our experimental thermal exposures are not relevant to annual average temperature effects, but seasonal variation, which is becoming more extreme as climate change drives up atmospheric volatility. We now frame the work more in this context, and make it clearer throughout the manuscript that our study is relevant to thermal variability, and not annual averages (including three new references to this). Although we cite climate change as being relevant to our study, we believe that our findings have primary impact in revealing a novel example of basic biological function for dealing with inherent thermal variation in the natural environment. Temperatures exceeding 38°C have been recorded in 150 countries (Mherrera, 2019), so are very frequently encountered in tropical latitudes. It is also relevant to note that the laboratory 'optimum' for population productivity in our T. castaneum model is a warm 35°C (stated in the manuscript Materials and methods), so 38°C is not too far apart from that. We also agree that additional thermal regimes in the study would have been great, but the existing workload to complete this work through a number of years at control and extreme (but relevant) experimental temperatures was itself more than a stretch. Changes to framing the work: "One of the most important abiotic environmental variables is temperature (Cossins and Bowler 1987;Angilletta Jr., 2009), especially in the context of climate change when thermal environments are expected to both warm and become much variable and extreme (Perkins et al., 2012, Raftery et al., 2017)." **"**After exposing adult males and females and their gametes to different temperatures, we compare sperm and egg development and reproductive function within thermal regimes that mimic the increasingly variable conditions faced by warmer tropical regions (Perkins et al., 2012)." "...an obvious mechanism to enable adaptive thermal plasticity for populations facing the challenge of reproducing under climate change where increases in both thermal averages, maxima and variation are expected (Perkins et al., 2012, Raftery et al., 2017)." **"**30°C is our standard rearing temperature, and the optimum laboratory temperature for population productivity in *T. castaneum* is 35°C (Sokoloff, 1974, Sales, 2019). \[...\] Our experimental exposure to this thermal regime will have relevance to more extreme thermal conditions and heatwaves, predicted to increase in frequency, severity and duration under climate change (e.g. Perkins et al., 2012), and which we know specifically constrain reproduction and sperm function in this system (Sales et al., 2018)."

{ "pile_set_name": "PubMed Central" }

Introduction ============ Breast cancer is one of the most common female malignant tumors with high morbidity and mortality rate caused by its strong metastatic ability ([@b1-ol-0-0-6571],[@b2-ol-0-0-6571]). There are no consensus biomarkers for early diagnosis and prognosis assessment of breast cancer with applications in clinical practice. Therefore, the development of breast cancer biomarkers has attracted increased attention recently. Special AT-rich sequence-binding protein 1 (SATB1) binds to T-rich sequences in chromosomes to regulate the expression of downstream genes ([@b3-ol-0-0-6571],[@b4-ol-0-0-6571]). The expression level of SATB1 is low in normal tissue, but is elevated in a variety of tumors ([@b5-ol-0-0-6571]--[@b7-ol-0-0-6571]). Toll-like receptor 4 (TLR4) is mainly expressed in immune cells, but can also be expressed in tumor cells ([@b8-ol-0-0-6571]--[@b10-ol-0-0-6571]). In this study, we detected the expression of SATB1 and RLR4 in 120 cases of cancer and 53 cases of adjacent non-cancerous tissue by immunohistochemistry. The correlation between the expression of these two proteins and the clinical characteristics of patients were analyzed. Patients and methods ==================== ### Patient information A total of 120 patients diagnosed with breast cancer in Yuhuangding Hospital of Yantai from October 2014 to October 2016 was enrolled in the study. Cancer tissue was collected after surgical resection. At the same time, adjacent non-cancerous tissue was collected from 53 patients. All the patients were females, and their ages ranged from 28 to 65 years, with a mean age of 46.5±11.7 years. No patient had been treated with chemotherapy before the study. Specimens were collected from necrotic cancer tissue and the adjacent non-cancerous tissue within 3 cm, fixed, and embedded in paraffin. Cancerous samples were diagnosed as breast cancer by pathological examination. The study was approved by the Ethics Committee of Yuhuangding Hospital of Yantai. All the patients signed an informed consent before being enrolled in the study. ### Reagents and methods Anti-human SATB1 monoclonal antibody and rabbit anti-human TLR4 monoclonal antibody were purchased from Abcam (Cambridge, UK). The DAB kit and hematoxylin were purchased from ZSBG-Bio (Beijing, China). Horseradish peroxidase-conjugated secondary antibody was purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Paraffin-embedded breast cancer samples were cut into 4 µm sections and transferred onto glass slides. After baking for 2 h at 90°C, tissue sections were dewaxed and rehydrated. After that, antigen retrieval was performed by incubating with 0.01 M sodium citrate buffer for 15 min. Endogenous peroxidase blocker was then added and incubated at 37°C for 10 min. After blocking with goat serum at room temperature for 20 min, the primary antibodies of SATB1 (1:300) and TLR4 (1:300) were incubated with the slides at 4°C overnight. After washing, secondary antibody was added and incubated at 37°C for 1 h. DAB staining was then performed and tissue sections were examined under the microscope to observe the staining. After hematoxylin staining, the slides were dehydrated, cleared, and sealed. All the operations were performed in accordance with the manufacturer\'s instructions. ### Determination of experimental results The brown or yellow granules on the slides showed the positive expression of SATB1 and TLR4. SATB1 mainly accumulated in the nucleus and TLR4 mainly accumulated in the cytoplasm. Using ×400 magnification in a bright field microscope (Leica, Wetzlar, Germany), 10 distinct visual fields were selected to count the positive cells and record the degree of staining. We also calculated the percentage of positive cells. Scoring was performed according to the degree of staining: no staining, 0 points; light yellow, 1 point; yellowish-brown, 2 points; chocolate brown, 3 points. Scoring was also performed according to the percentage of positive cells: 0--25%, 1 point; 25--65%, 2 points; 65--100%, 3 points. The product of the 2 scores greater than 3 was taken as positive expression; values below 3 were considered negative expression ([@b11-ol-0-0-6571]). ### Statistical analysis SPSS 19.0 statistical software (IBM SPSS, Armonk, NY, USA) was used to analyze the data. The count data were analyzed by Chi-square test. Correlation analysis was performed by Spearman\'s rank correlation analysis. P\<0.05 was considered to be statistically significant. Results ======= ### Expression of SATB1 and TLR4 in breast cancer and adjacent non-cancerous tissue SATB1 expression was observed in the nucleus under a microscope. SATB1 was positively expressed in 89 cases of breast cancer, and the positive expression rate of SATB1 was 74.1% ([Fig. 1](#f1-ol-0-0-6571){ref-type="fig"}). Positive expression of SATB1 was only detected in 7 cases of adjacent non-cancerous tissue, and the positive expression rate of SATB1 was 13.21%. A statistically significant difference in the expression of SATB1 was found between breast cancer and adjacent non-cancerous tissue (P\<0.05; [Table I](#tI-ol-0-0-6571){ref-type="table"}). TLR4 expression was detected in the cytoplasm. TLR4 was positively expressed in 70 cases of breast cancer, and the positive rate was 58.33% ([Fig. 1](#f1-ol-0-0-6571){ref-type="fig"}). Positive expression of SATB1 was detected in 48 cases of adjacent non-cancerous tissues, and the positive rate was 90.57%. A statistically significant difference in the expression of TLR4 was found between breast cancer and adjacent non-cancerous tissues (P\<0.05; [Table I](#tI-ol-0-0-6571){ref-type="table"}). ### Correlation between expression of SATB1 and TLR4 Following the immunohistochemistry results, we next analyzed the correlation between SATB1 and TLR4. As shown in [Table II](#tII-ol-0-0-6571){ref-type="table"}, the expression of SATB1 was negatively correlated with the expression of TLR4 (r=−0.624, P\<0.05). ### Correlation of SATB1 and TLR4 with the clinical and pathological features of patients We last examined the correlation of the expression of SATB1 and TLR4, and the clinical and pathological features of patients. We found that the expression levels of SATB1 and TLR4 were not significantly correlated with the age, menopause, PR protein, and HER-2 protein expression (P\>0.05). However, the expression levels of SATB1 and TLR4 were significantly correlated with tumor size, local lymphatic metastasis, histopathological grade, tumor stage, and the expression of ER protein (P\<0.05; [Table III](#tIII-ol-0-0-6571){ref-type="table"}). Discussion ========== SATB1 is a nuclear matrix attachment-binding protein with tissue-specific expression. The *SATB1* gene is located on chromosome 3p23 and encodes for a 763-amino acid protein ([@b12-ol-0-0-6571]). SATB1 is highly expressed in thymus where it regulates the development and maturation of T cells ([@b13-ol-0-0-6571],[@b14-ol-0-0-6571]). Previous studies showed that *SATB1* gene knockout in mice can inhibit the production of CD4^+^ and CD8^+^ double positive T cells, leading to disorders of thymus cell maturation ([@b15-ol-0-0-6571]). SATB1 plays a role as a 'gene organizer' in the genome. SATB1 can interact with more than 1,000 proteins to specifically regulate the expression of its target genes by chromatin remodeling and protein modification ([@b16-ol-0-0-6571]). SATB1 can bind the BUR region of target genes and anchor the BUR region on the nuclear matrix to alter the higher-order structure of the chromatin, and regulate gene expression ([@b17-ol-0-0-6571]). SATB1 can also regulate DNA binding capacity and the subcellular localization of proteins through phosphorylation, ubiquitination, and acetylation ([@b18-ol-0-0-6571]). Liu *et al* found that high expression of SATB1 in breast cancer cells significantly increased cell invasion ability ([@b19-ol-0-0-6571]). Clinical data from 1,318 breast cancer patients showed that the expression level of SATB1 was negatively correlated with survival time ([@b20-ol-0-0-6571]). Our study shows that SATB1 is strongly expressed in breast cancer and weakly expressed in adjacent non-cancerous tissue. The expression of SATB1 was not significantly correlated with age, menopause, and the expression of the PR and HER-2 proteins, but was significantly correlated with tumor size, local lymphatic metastasis, histopathological grade, tumor stage, and ER protein expression. Our findings are consistent with previous studies ([@b21-ol-0-0-6571],[@b22-ol-0-0-6571]). Toll-like receptors were first found in *Drosophila*. In 1997, TLR4 homologue was identified in humans, and so far, there are 12 members of the TLRs family ([@b23-ol-0-0-6571]). The *TLR4* gene is located on chromosome 9q32-q33, and encodes for a 224-amino acid protein. TLR4 is widely distributed on the cell surface to sense pathogens ([@b24-ol-0-0-6571]). TLR4 is widely distributed in human monocytes ([@b25-ol-0-0-6571]), neutrophils ([@b26-ol-0-0-6571]), and epithelial cells ([@b27-ol-0-0-6571]). TLR4 can recognize a variety of pathogen-associated molecular patterns (e.g., LPS of Gram-negative bacteria) to induce different immune responses ([@b28-ol-0-0-6571]). Through binding to the corresponding ligands and mediating intracellular signal transduction, TLR4 plays a role as a transcription factor to activate the expression of a variety of cell growth and apoptosis-related factors ([@b29-ol-0-0-6571],[@b30-ol-0-0-6571]). TLR4 can mediate MyD88-dependent pathways through the interaction with a series of cytokines to promote tumorigenesis ([@b31-ol-0-0-6571]). Clinical studies also found that TLR4 was correlated with the growth and metastasis of gastric cancer ([@b32-ol-0-0-6571]), ovarian cancer ([@b33-ol-0-0-6571],[@b34-ol-0-0-6571]), cervical cancer ([@b35-ol-0-0-6571]), and other types of tumor cells. In our study, we found that TLR4 was positively expressed in 58.33% cases of breast cancer tissues and in 90.57% cases of adjacent non-cancerous tissue. The positive expression rate of TLR4 in this study is consistent with previous studies ([@b35-ol-0-0-6571],[@b36-ol-0-0-6571]). We also found that TLR4 expression was not significantly correlated with age, menopause, or the expression of the PR and HER-2 proteins. However, TLR4 was significantly correlated with tumor size, local lymphatic metastasis, histopathological grade, tumor stage, and ER protein expression. In addition, correlation analysis indicated that the expression level of SATB1 was negatively correlated with the expression level of TLR4. In conclusion, SATB1 and TLR4 are involved in the development of breast cancer, which is of great significance for the identification of potential therapeutic targets and prognosis of breast cancer. ![Expression of SATB1 and TLR4 in breast cancer and adjacent non-cancerous tissue. (A) Representative result of SATB1 expression in adjacent non-cancerous tissue (×400). (B) Representative result of SATB1 expression in breast cancer (×400). (C) Representative result of TLR4 expression in adjacent non-cancerous tissue (×400). (D) Representative result of TLR4 expression in breast cancer (×400). The positive expression rate of SATB1 in breast cancer tissues was significantly higher than that in adjacent non-cancerous tissue. The positive expression rate of TLR4 in adjacent non-cancerous tissues was significantly higher than that in breast cancer. SATB1, special AT-rich sequence-binding protein 1; TLR4, toll-like receptor 4.](ol-14-03-3611-g00){#f1-ol-0-0-6571} ###### Expression of SATB1 and TLR4 in breast cancer and adjacent non-cancerous tissue. SATB1 TLR4 ------------------------------- ------ --------- ------ ----------- ----- ---- ------- Breast cancer 120 31 89 74.17 50 70 58.33 Adjacent non-cancerous tissue 53 46 7 13.21 5 48 90.57 χ^2^ value 37.413 26.481 P-value 0.006 0.011 SATB1, special AT-rich sequence-binding protein 1; TLR4, toll-like receptor 4. ###### Correlation between the expression of SATB4 and TLR4 in breast cancer tissues. SATB1 ---- ------- ----- -------- ------- \+ 43 27 −0.624 0.003 − 46 4 SATB1, special AT-rich sequence-binding protein 1; TLR4, toll-like receptor 4. ###### Correlation of the expression of SATB1 and TLR4 with the clinical and pathological features of patients. SATB1 TLR4 ------------------------- ---- ------- ------- ------- ---- ------ ------- Age (years) \<50 46 35 0.25 0.416 28 0.74 0.352 ≥50 74 54 42 Menopause Yes 52 38 1.36 0.129 30 0.62 0.391 No 68 51 40 Tumor size \<2 cm 44 29 5.24 0.017 23 8.49 0.013 ≥2 cm 76 60 47 Lymph node metastasis Yes 56 31 9.15 0.013 26 6.37 0.015 No 64 57 44 Histopathological grade I 38 19 11.48 0.007 18 7.14 0.015 II 53 44 36 III 29 26 16 Tumor stage I 32 18 12.53 0.007 16 9.61 0.011 II 57 43 33 III, IV 31 28 21 PR (−) 48 35 0.94 0.172 28 1.74 0.114 (+) 72 54 42 ER (−) 53 29 6.74 0.015 24 7.93 0.013 (+) 67 60 46 HER-2 (−) 52 38 0.87 0.181 30 0.96 0.172 (+) 68 51 40 SATB1, special AT-rich sequence-binding protein 1; TLR4, toll-like receptor 4. [^1]: Contributed equally

{ "pile_set_name": "PubMed Central" }

The prevalence of diabetes continues to escalate around the world, with an estimated 24 million people affected in the U.S. ([@B1]). Its prevalence has more than doubled in Tennessee the last decade from 5.0% in 1997 to 11.0% in 2007 ([@B2]), with estimated diabetes related mortality of 31 per 100,000 population ([@B2]). Rising prevalence and shortage of physicians, especially in rural areas, contribute to poor outcome in diabetic patients ([@B3]). Telemedicine, the transfer of electronic medical data to a remote location utilizing telecommunications technology, may be beneficial in improving access, care, and outcome in diabetic patients. Therefore, we investigated the impact of telemedicine on the quality of care in diabetic patients in five health professional shortage areas with diabetes related mortality rates of 41.5--84.7 per 100,000. ADT was a prospective interventional study, which recruited diabetic patients aged ≥18 years, with A1C ≥8.0%. Diabetes self-management education (DSME) was delivered by a certified diabetes educator via videoconference every 3 months. The DSME classes addressed pathogenesis of diabetes, nutritional education, physical activity, self blood glucose monitoring, effects of insulin and other diabetes medications, sick day management, and complications of diabetes. DSME classes were delivered from the telemedicine studio of the University of Tennessee Health Science Center in Memphis, which was connected to the remote locations by videoconference using Polycom VSX 7000 video cameras (Pleasanton, CA), television monitors, and the Internet. Vital signs and anthropometry were obtained at each visit. Laboratory data such as A1C and lipid levels were obtained from the patients' primary physician. Statistical analysis was done using Student *t* test or ANOVA and χ^2^ test. Thirty-six patients aged 55.6 ± 10.0 years, with mean duration of diabetes of 12.0 ± 10.0 years completed the study; 55% of the subjects had concomitant dyslipidemia and hypertension. Subjects showed significant improvement in A1C after 3 months (9.8 vs. 8.4%, *P* \< 0.001). Reduction in A1C at 3 months was sustained over 12 months of the study (8.4 vs. 8.3%, *P* \> 0.5). The proportion of subjects achieving target A1C of \<7% increased to 40% after intervention (*P* \< 0.001). There was also a significant rise in the proportion of patients who achieved target blood pressure (61 vs. 78%), HDL (11 vs. 22%), and triglyceride (3 vs. 30%) (*P* \< 0.001); and 97% of the participants were satisfied with the program, and 90% acknowledged it saved time and money. Given the average duration of uncontrolled diabetes of 12 years, this degree of improvement in 3 months is quite encouraging. The improvement in A1C, which was maintained throughout the study, may suggest that a 3-month program may be more cost-effective than a longer one. This prospective study is a useful addition to the evidence for the efficacy of a well-designed telemedicine program in diabetic patients ([@B4],[@B5]). However, it is limited by the small number of subjects. A prospective randomized study would be required to verify the observations in this study. In conclusion, a telemedicine-based management program that delivers diabetes education and consultative service through videoconference was effective in improving diabetes outcome measures. It proved to be an acceptable means of providing care in diabetic patients at high risk for cardiovascular disease. ADT was supported by a grant (ED-08-22875-00) from the Tennessee Department of Health. No potential conflicts of interest relevant to this article were reported. E.A.N. contributed to data research, discussion, and manuscript writing and review. S.A. contributed to data research and manuscript review. J.T. contributed to data research and manuscript review. S.S.N. contributed to data research and manuscript review. A.E.K. contributed to discussion and manuscript writing and review.

{ "pile_set_name": "PubMed Central" }

Background {#Sec1} ========== Natural enemies are important in controlling the population of insect pests in agricultural ecosystems \[[@CR1]--[@CR6]\]. The presence of natural enemies is detected through chemical cues emitted by prey and host plants, alone or in association, and received by the antennal sensory system in the multi-trophic environment to result in various behavioral choices including prey forage, localization, oviposition and escape \[[@CR7]--[@CR17]\]. Recently the study of insect peripheral sensory system and chemosignal transduction has experienced considerable progress due to the development of bioinformatics-based approaches and protein function prediction methods \[[@CR18]\]. In particular, a large amount of information is provided by antennal transcriptome projects from various insect taxa \[[@CR19]--[@CR28]\]. In general, the process of chemoreception, including olfaction and taste, involves several families of genes, including odorant receptors (ORs), ionotropic receptors (IRs), and gustatory receptors (GRs) \[[@CR20], [@CR29]--[@CR31]\]. In addition, odorant binding proteins (OBPs), chemosensory proteins (CSPs) and sensory neuron membrane proteins (SNMPs) also play crucial roles in chemoreception \[[@CR32]--[@CR36]\]. The insect chemoreceptor superfamily including ORs and GRs was first identified in the *Drosophila melanogaster* genome \[[@CR18], [@CR21]\]. Insect odorant and gustatory receptors were once thought to be G-protein-coupled receptors just like ORs in worms and vertebrates, but subsequent studies have shown a lack of homology to vertebrate ORs \[[@CR37]\]. One such superfamily encoding ORs is highly divergent across insect taxa with sequences and frequencies varying to a large extent \[[@CR18], [@CR38], [@CR39]\]. ORs are broadly tuned to alcohols, ketones, and esters generally present in the environment \[[@CR40], [@CR41]\]. Another family encoding GRs, or receptors for taste or contact stimuli, is also very divergent across insect taxa \[[@CR31]\]. On the contrary, one example of exceptionally conserved GRs are GR21a and GR63a, which work together as a CO~2~ receptor in *Drosophila* \[[@CR42]\]. Such chemoreceptors play an important role in host seeking behaviors in many insects but, especially seen in mosquitoes \[[@CR43], [@CR44]\]. A new insect chemosensory family was identified recently and given the name ionotropic receptors (IRs). These IRs belong to the ionotropic glutamate receptor superfamily (iGluRs) and were identified in both the olfactory and gustatory systems \[[@CR30], [@CR45]\]. IRs are more greatly conserved than ORs and GRs but considerable variations can be observed in ligand-binding domains. They are mainly tuned to acids, amines and other odorants that are not detected by ORs \[[@CR30], [@CR45], [@CR46]\]. Since chemosensory gene families were characterized in two important model species, *D. melanogaster* and *Anopheles gambiae* \[[@CR47], [@CR48]\], a growing number of chemosensory genes have been identified from many Dipteran species, such as *Musca domestica* \[[@CR49]\], *Bactrocera dorsalis* (genome: assembly ASM78921v2), *Calliphora stygia* \[[@CR50]\], *Glossina morsitans morsitans* \[[@CR51]\], and *Mayetiola destructor* Say \[[@CR20]\]. Protein prediction methods have been the first step for functional identification of chemosensory genes. All information regarding insect chemosensory was obtained bioinformatically and has been beneficial in understanding insect processing of diverse volatile compounds and cross-species differences in chemical communication. The syrphids belong within the Diptera order and their larvae are aphid-specific natural enemies \[[@CR1], [@CR2], [@CR4], [@CR5], [@CR8]\]. Due to the larvae's agricultural importance via potential applications, several reports have been published on the chemical ecology of these insects. In *Episyrphus balteatus* DeGeer, larvae may use a sesquiterpene as a kairomone \[[@CR17], [@CR52]\] and other potential semiochemicals to locate their prey \[[@CR8], [@CR53]\]. In *Sphaerophoria rueppellii*, adult females are strongly attracted to odors from aphid colonies showing that specific volatile compounds are important to detect their prey \[[@CR1]\]. Some studies on the relationships between aphid or host plant volatile emissions and aphid localization and foraging behavior have shown strong associations with syrphid recognition. One striking finding has shown that volatiles from plants attacked by aphids produce strong electrophysiological responses from the antennae of syrphids \[[@CR7], [@CR8]\]. These studies indicate that detecting prey-derived volatiles ((E)-β-farnesene), herbivore-induced plant volatiles (monoterpenes and sesquiterpenes), or naturally occurring general leaf volatiles (GLVs; alcohols, aldehydes and esters) help natural enemies to select oviposition sites and locate their prey \[[@CR8], [@CR12], [@CR13], [@CR17], [@CR52]--[@CR55]\]. Despite these reports on chemosensory behavior, little is known on the molecular basis of syrphid olfaction. Therefore, the identification of predatory syrphid chemosensory gene families will help reveal how syrphids forage on their prey and choose oviposition sites. In this study we selected two syrphid species, *E. balteatus* and *Eupeodes corollae* Fabricius, active [in](http://cn.bing.com/dict/search?q=in&FORM=BDVSP6&mkt=zh-cn) northern China cotton fields to perform antennal transcriptome sequencing in order to explore and compare chemosensory genes in the two species. A total of 154 and 134 chemosensory candidate genes were identified in *E. balteatus* and *E. corollae* transcriptomes, respectively, including ORs, IRs, GRs, OBPs, CSPs and SNMPs. Furthermore, we report the expression profile of the OR families found in each insect transcriptome. A comparison between these two syrphids and other insect species revealed candidate chemosensory genes that could be involved in prey selection and plant volatile recognition. The discovery of putative chemosensory genes gives way for further exploration into functional assessments regarding chemoreception association. Results {#Sec2} ======= Antennal transcriptome sequencing and sequence assembly {#Sec3} ------------------------------------------------------- *E. balteatus* and *E. corollae* antennal transcriptomes were sequenced using the Illumina HiSeq 2000 platform combined with Trinity assembly. Approximately 68.71 million and 77.28 million raw-reads were obtained from *E. balteatus* male and female antenna, respectively, reduced after filtering, to 65.69 and 74.25 million clean-reads. These were assembled into 57,950 unigenes for male and 68,165 for female. A final transcript dataset with 53,575 unigenes was obtained, consisting of 17,407 distinct clusters and 36,168 distinct singletons. The dataset was 47.61 megabase (Mb) in size with a mean length of 889 bp and N50 of 1724 bp (Additional file [1](#MOESM1){ref-type="media"}: Table S1). Parallel experiments generated 80.15 million and 77.38 million raw-reads in *E. corollae* male and female, respectively, and 65.69 and 74.25 million clean-reads. From these datas 54,116 and 61,220 unigenes were obtained for male and female, respectively. The final transcript dataset of *E. corollae* contained 50,942 unigenes with a mean length of 1039 bp and N50 length of 2104 bp, consisting of 18,054 distinct clusters and 32,888 distinct singletons (Additional file [1](#MOESM1){ref-type="media"}: Table S1). In addition, unigenes with a sequence length \> 500 bp accounted for 47.64% and 43.59% of the *E. corollae* and *E. balteatus* transcriptome assembly, respectively. Homology analysis and gene ontology (GO) annotation {#Sec4} --------------------------------------------------- A BLASTX homology search against the NCBI non-redundant protein database indicated 23,680 (44.2%) and 25,606 (50.3%) unigenes from *E. balteatus* and *E. corollae*, respectively, with sequence similarities to known proteins using a cut-off E-value of 10^−5^. For *E. balteatus*, the larger number of similar genes (27.4%) belonged to *Ceratitis capitata* followed by *M. domestica* (20.4%), *D. melanogaster* (7.4%), *Drosophila willistoni* (3.1%), *Drosophila virilis* (3.0%) and *Drosophila mojavensis* (2.5%). For *E. corollae*, again *C. capitata* was best represented (25.1%), followed by *M. domestica* (18.8%), *D. melanogaster* (7.1%), *D. willistoni* (2.9%), *Acyrthosiphon pisum* (2.8%), and *D. virilis* (2.5%) (Additional file [2](#MOESM2){ref-type="media"}:Fig. S1A). Gene ontology (GO) annotations were used to classify the transcripts into functional groups in accordance with specific GO categories. A total of 12,441 (23.22%) of all predicted proteins from *E. balteatus* and 12,425 (24.39%) predicted proteins from *E. corollae* were assigned to at least one GO term (Additional file [2](#MOESM2){ref-type="media"}: Fig. S1B). The GO terms distribution in the three categories were similar in the two species. In the "molecular function" category, the most abundant GO terms were "binding" and "catalytic activity". In the "biological process" category, "cellular process", "single-organism process" and "metabolic process" were the most represented. Finally, "cell", "cell part", and "organelle" were the most abundant GO terms in "cellular component" category (Additional file [2](#MOESM2){ref-type="media"}: Fig. S1B). GO terms associated with chemosensory genes were distributed in the "biological process" category (e.g. "cellular process", "developmental process", "response to stimulus", "establishment of localization", and "biological regulation", etc.), "molecular function" category (e.g. "molecular transducer activity", etc.) and "cellular component" category (e.g. "extracellular region", "membrane part", "membrane", etc.). Candidate ORs in *E. balteatus* and *E. corollae* {#Sec5} ------------------------------------------------- Based on our analysis of the antennal transcriptomes in the two species, 51 and 42 transcripts for candidate ORs were identified in the combined male and female data sets from *E. balteatus* and *E. corollae*, respectively (Additional file [3](#MOESM3){ref-type="media"}: Table S2). A total of 21 *E. balteatus* ORs (EbalORs) and 29 *E. corollae* ORs (EcorORs) contained full-length open reading frames (ORFs), whose translation products are predicted to possess 2--8 transmembrane domains (TMDs). Other partial length transcripts encoded proteins exhibiting overlapping regions with low identity and were classified as unique genes. After a more exhaustive comparison with OR genes from other insect species, we found that all putative EbalORs shared between 22% and 86% amino acid identity with other ORs, with almost identical values (22% to 87%) for EcorORs. Detailed information is reported in Additional file [3](#MOESM3){ref-type="media"} and Table S2. We next performed a phylogenetic analysis using our candidate ORs and the ORs from four other Diptera species including *B. dorsalis*, *C. stygia*, *D. melanogaster* and *M. domestica* (Fig. [1](#Fig1){ref-type="fig"}). Clustered with DmOR83b, the highly conserved co-receptor Orco, orthologous genes were identified in the antennal transcriptomes of both syrphid species, and named EbalOrco and EcorOrco. As expected, sequence identity between EbalOrco and EcorOrco is very high (97.27%). Among the other ORs, five EbalORs (EbalOR9, 16, 18, 22 and 37) and three EcorORs (EcorOR8, 24 and 29) clustered with DmelOR67d, the pheromone receptor from *D. melanogaster*. This OR67d specific clade also included the OR67d orthologues from *M. domestica* and *B. dorsalis*. Two of these genes, EbalOR16 and EcorOR24, are full-length transcripts with 71.65% amino acid identity. The remaining ORs in this group were highly divergent among different species. Within the Dipteran OR sequences, we found a species-specific clade including eight members from *E. balteatus* (EbalOR7, 10, 30, 32, 41, 46, 47 and 48) and seven from *E. corollae* (EcorOR9, 10, 11, 15, 16, 34 and 38) that shared low identities with other Dipteran ORs (Fig. [1](#Fig1){ref-type="fig"}).Fig. 1Phylogenetic tree of candidate *E. balteatus* and *E. corollae* ORs and other Dipteran ORs. The distance tree was rooted by the conservative Orco gene orthologous. Bootstrap values are shown. The Orco clade, OR67d clade and specific EbalORs and EcorORs clade are shown. Species in this phylogeny include *E. balteatus* (Ebal, *red*), *E. corollae* (Ecor, *blue*), *Drosophila melanogaster* (Dm, *dark*), *Bactrocera dorsalis* (Bdor, *green*), *Calliphora stygia* (Csty, *magenta*), and *Musca domestica* (Mdom, *gray*) Amino acid identity between gene products correlates with similarity between genes. An amino acid sequence comparison between EbalORs and EcorORs revealed 33 pairs of orthologous ORs (including Orco) with 73.86% identity and the amino acid identity of complete OR ORFs was 84.00%. Sequence similarity percentages of the 28 pairs of homologous ORs are greater than 60% (Additional file [4](#MOESM4){ref-type="media"}: Table S3). In addition to the Orco family, orthologous groups with identities higher than 90% include the EbalOR29/EcorOR36, EbalOR3/EcorOR5, EbalOR13/EcorOR19 EcorOR22 with/EcorOR22 with 94.88%, 92.20%, 91.46% and 91.30% sequence identities, respectively. In addition, 10 EbalOR sequences (EbalOR2, 5, 7, 8, 11, 14, 21, 23, 25 and 33) are closely related to EcorOR homologues (EcorOR3, 4, 7, 9, 13, 14, 23, 28, 31 and 35) with identities greater than 80% (Additional file [4](#MOESM4){ref-type="media"}: Table S3; Fig. [1](#Fig1){ref-type="fig"}). All of these highly homologous proteins may play important roles in olfactory recognition. Candidate GRs in *E. balteatus* and *E. corollae* {#Sec6} ------------------------------------------------- We have identified 14 and 16 candidate GR genes from *E. balteatus* and *E. corollae* transcriptomes, respectively (Additional file [3](#MOESM3){ref-type="media"}: Table S2). The majority of candidate EbalGRs and EcorGRs were partial fragments, with only three from *E. balteatus* and six from *E. corollae* encoding full-length proteins. These complete sequences all show six or seven TMDs with an intracellular N-terminus and extracellular C-terminus. Phylogenetic analysis with GRs from six Dipteran species suggest that *Drosophila* GR21a and GR63a, reported as carbon dioxide sensors \[[@CR42], [@CR56]\], clustered first with EbalGR2 and EcorGR2 and second with EbalGR1 and EcorGR1. In addition, EbalGR4, EbalGR13 and EcorGR4 showed high identities to thermoreceptor DmelGR28b responsible for rapid warmth avoidance \[[@CR57]\]. Several other GRs clustered with members of candidate sugar detection GR (GR5a, GR61a and GR64a-f) sub-family (Fig. [2](#Fig2){ref-type="fig"}) \[[@CR58]--[@CR62]\].Fig. 2Phylogenetic tree of candidate *E. balteatus* and *E. corollae* GRs and other Dipteran GRs. The distance tree was rooted by the conservative carbon dioxide GRs gene orthologous. Bootstrap values are shown. The carbon dioxide GRs clade, thermos-sensed GRs clade and sugar GRs clade are shown. This tree was constructed using the species *E. balteatus* (Ebal, *red*), *E. corollae* (Ecor, *blue*), *D. melanogaster* (Dm, *dark*), *B. dorsalis* (Bdor, *green*), *C. stygia* (Csty, *magenta*), and *M. domestica* (Mdom, *gray*) Candidate IRs in *E. balteatus* and *E. corollae* {#Sec7} ------------------------------------------------- We identified 32 transcripts for putative ionotropic receptors in *E. balteatus* and 23 in *E. corollae*. Of these, seven EbalIRs and 14 EcorIRs contained full-length ORFs, with two to five TMDs (Additional file [3](#MOESM3){ref-type="media"}: Table S2). Among these we found the common conserved co-receptors IR8a (EbalIR8a and EcorIR8a) and IR25a (EbalIR25a and EcorIR25a) in both species. Other candidate IRs were found as partial sequences (Fig. [3](#Fig3){ref-type="fig"}).Fig. 3Phylogenetic tree of candidate *E. balteatus* and *E. corollae* IRs and other Dipteran IRs. The distance tree was rooted by the conservative IR25a/IR8a gene orthologues. Bootstrap values are shown. The IR25a/IR8a clade, iGluRs clade and some antennal-associated orthologue clade are shown. This tree was constructed using the following species: *E. balteatus* (Ebal, *red*), *E. corollae* (Ecor, *blue*), *D. melanogaster* (Dmel, *dark*), *Anopheles gambiae* (Agam, *green*), and *C. stygia* (Csty, *magenta*) In order to further distinguish putative IRs from iGluRs, all EbalIRs and EcorIRs were aligned with IRs from *A. gambiae*, *C. stygia* and *D. melanogaster*, as well as some AgamiGluRs and DmeliGluRs for phylogenetic analysis. The results showed that the candidate EbalIRs and EcorIRs clustered with presumed "antennal" orthologues IR76b, IR93a, IR21a, IR68a, IR40a, IR75l, IR75d, IR64a, IR84a, IR31a and IR92a, and were well separated from the AgamiGluRs and DmeliGluRs clade (Fig. [3](#Fig3){ref-type="fig"}) \[[@CR63]\]. Interestingly, the conserved "antennal" orthologues, IR60a, was lacking from *E. balteatus* and *E. corollae* transcriptome assemblies, while IR68a was only absent from *E. corollae.* The sequences of *E. balteatus* clustering with DmelIR94d and DmelIR94e were quite divergent (Fig. [3](#Fig3){ref-type="fig"}; Additional file [5](#MOESM5){ref-type="media"}: Fig. S2). When compared to the orthologues within other species, these IRs may play different roles in olfaction. Candidate OBPs in *E. balteatus* an*d E. corollae* {#Sec8} -------------------------------------------------- We identified 49 different transcripts encoding candidate OBPs in *E. balteatus* and 44 in *E. corollae*, numbers similar to the 52 OBPs of *D. melanogaster* \[[@CR64]\]. Of these, 38 transcripts of EbalOBPs and 31 EcorOBPs contained full-length ORFs with predicted signal peptide sequences (with EbalOBP31 as the only exception) (Additional file [3](#MOESM3){ref-type="media"}: Table S2). A phylogenetic tree was built with these sequences and those of orthologous from *B. dorsalis*, *C. stygia*, *D. melanogaster* and *M. domestica*. Among EbalOBPs, thirty-one showed the classic motif of six conserved cysteines, three were Plus-C (EbalOBP2, 3, 4) and fifteen were Minus-C (EbalOBP5, 7, 8, 9, 11, 12, 14, 15, 16, 18, 21, 25, 26, 44 and 45) (Fig. [4](#Fig4){ref-type="fig"}). For EcorOBPs, we found 30 classic, 5 Plus-C (EcorOBP1, 3, 4, 5, 41) and 9 Minus-C (EcorOBP6, 7, 8, 10, 11, 13, 18, 36 and 37) (Fig. [4](#Fig4){ref-type="fig"}) \[[@CR64], [@CR65]\]. One large group of classic OBPs, including 15 EbalOBPs and 14 EcorOBPs showed large differences compared to sequences of other species and could represent OBPs specific of syrphids (Fig. [4](#Fig4){ref-type="fig"}; Additional file [6](#MOESM6){ref-type="media"}: Fig. S3). We found the orthologue of DmelOBP-lush in both species, EbalOBP17 and EcorOBP14. These two proteins are 98.64% identical at the amino acid level between each other and 39.87% and 40.52% identity to DmelOBP-lush.Fig. 4Phylogenetic tree of candidate *E. balteatus* and *E. corollae* OBPs and other Dipteran OBPs. The distance tree was rooted by lush gene orthologous. Bootstrap values are shown. The classic OBPs clade, Plus-C OBPs clade and Minus-C OBPs clade are shown. The species used to construct tree include *E. balteatus* (Ebal, *red*), *E. corollae* (Ecor, *blue*), *D. melanogaster* (Dmel, *dark*), *B. dorsalis* (Bdor, *green*), *C. stygia* (Csty, *magenta*), and *M. domestica* (Mdom, *gray*) Candidate CSPs in *E. balteatus* and *E. corollae* {#Sec9} -------------------------------------------------- Through bioinformatic analysis, six and seven different transcripts encoding candidate CSPs were identified from *E. balteatus* and *E. corollae* transcriptomes, respectively. Five EbalCSPs and six EcorCSPs represented full-length proteins and only EbalCSP6 lacked a signal peptide (Additional file [3](#MOESM3){ref-type="media"}: Table S2). All of the identified amino acid sequences possessed the highly conserved four-cysteine profile. A phylogenetic tree was built with all the syrphid CSPs and those of *A. gambiae*, *C. stygia*, *D. melanogaster* (Fig. [5](#Fig5){ref-type="fig"}).Fig. 5Phylogenetic tree of candidate *E. balteatus* and *E. corollae* CSPs and other Dipteran CSPs. The distance tree was rooted by AgamCSP1/2/3 genes. Bootstrap values are shown. The species used to construct tree include *E. balteatus* (Ebal, *red*), *E. corollae* (Ecor, *blue*), *D. melanogaster* (Dmel, *dark*), *A. gambiae* (Agam, *green*), and *C. stygia* (Csty, *magenta*) Candidate SNMPs in *E. balteatus* an*d E. corollae* {#Sec10} --------------------------------------------------- In both species, two SNMPs with full-length ORFs were identified possessing two TMDs (with EbalSNMP1 having a single TMD as an exception) (Additional file [3](#MOESM3){ref-type="media"}: Table S2). EbalSNMP1 and EcorSNMP1 are very similar to DmelSNMP1, a protein shown to be required for correct pheromone detection \[[@CR50], [@CR66]--[@CR68]\]. EbalSNMP2 and EcorSNMP2 are similar to DmelSNMP2, reported to be expressed in supporting cells (Fig. [6](#Fig6){ref-type="fig"}) \[[@CR27], [@CR69], [@CR70]\].Fig. 6Phylogenetic tree of candidate *E. balteatus* and *E. corollae* SNMP and other Dipteran SNMP. Bootstrap values are shown. The species used to construct tree including *E. balteatus* (Ebal, *red*), *E. corollae* (Ecor, *blue*), *D. melanogaster* (Dmel, *dark*) and *A. gambiae* (Agam, *green*) Differentially expressed genes (DEGs) analysis {#Sec11} ---------------------------------------------- Gene expression levels of all male and female antennae-associated chemosensory genes in both *E. balteatus* and *E. corollae* were assessed using fragments per kilobase per million fragments (FPKM) values, represented in a heatmap (Fig. [7](#Fig7){ref-type="fig"}). Normalised antennal expression levels of candidate *E. balteatus* and *E. corollae* ORs are shown in Additional file [7](#MOESM7){ref-type="media"}. Of all ORs, Orco had the highest expression level of transcripts in both sexes of each species. There were no significant differences of OR transcript abundances (FPKM value) in the respective male and female antenna, except for EcorOR14 (Additional file [7](#MOESM7){ref-type="media"}). A combined analysis of false discovery rate (FDR) ≤0.001 and \|log2 Ratio\| ≥ 1 showed that EcorOBP shared highest number of differentially expressed genes (DEGs), including eleven high-expression in male and six high-expression in female syrphids (Additional file [7](#MOESM7){ref-type="media"}). In addition, candidate carbon dioxide receptor GR1 and GR2, and SNMP1 in both sexes showed a high expression level (Fig. [7](#Fig7){ref-type="fig"}).Fig. 7Expression profiles of chemosensory genes in *E. balteatus* and *E. corollae*. **a**: CSPs; **b**: GRs; **c**: IRs; **d**: ORs; **e**: OBPs; and **f**: SNMPs Tissue- and sex- specific expression of candidate *E. balteatus* an*d E. corollae* OR genes {#Sec12} ------------------------------------------------------------------------------------------- The expression of the candidate ORs in *E. balteatus* an*d E. corollae* male and female antennae and legs (control sample) was analyzed using semi-quantitative reverse transcription PCR (RT-PCR). All 51 EbalORs and 42 EcorORs were detected in the antennae at high expressing level. Only EbalOR49 was found to be mainly expressed in legs. There were no significant differences of transcript abundances in the respective male and female antenna (Fig. [8](#Fig8){ref-type="fig"}). The Orco co-receptor gene also showed a high expression level in both syrphid species. This is consistent with DEGs analysis of OR transcript abundances using the FPKM value.Fig. 8Tissue- and sex- specific expression of candidate *E. balteatus* and *E. corollae* OR genes. M: male antennae, F: female antennae, L: legs (both sexes mixed) Discussion {#Sec13} ========== The syrphids *E. balteatus* and *E. corollae* are aphid-specific predators and predominately inhabit northern China wheat and cotton fields. Typical of most insects, chemical cues drive several aspects of their behavior, such as foraging on prey and choosing oviposition sites \[[@CR7], [@CR8], [@CR10]\]. Chemosensory proteins play an important role in this process. We analyzed antennal transcriptomes of *E. balteatus* and *E. corollae* and searched for chemosensory genes with the purpose of understanding chemical communication of tritrophic interactions among plants, herbivorous insects, and natural enemies. In our study, we sequenced *E. balteatus* and *E. corollae* antennal transcriptomes using next generation sequencing technology on the Illumina HiSeq 2000 platform. The total RNA was converted into a template library for high throughput DNA sequencing, allowing us to obtain all expressed transcripts. De novo assembly of transcripts using the Trinity method gives high-efficiency and reliable full-length transcripts across extensive expression levels, even without genome information \[[@CR71]\]. Our sequence assembly yielded a final transcript dataset of 50,942 unigenes from *E. corollae* and 53,575 *E. balteatus* unigenes. Total unigenes counts resulted in 44.2% unigenes from *E. balteatus* and 50.3% unigenes from *E. corollae* shared sequence similarities to known proteins using the BLASTX homology search from the NCBI non-redundant protein database. These percentages are very similar to other Dipteran species \[[@CR50], [@CR72]\]. Remaining transcripts without associated GO terms may represent species-specific genes. The antennal transcriptome analysis proved to be a powerful tool to identify chemosensory genes in insects without genome information. It has been successfully employed in many insect orders including Lepidoptera, Coleoptera, Hymenoptera and Hemiptera. In Diptera, the chemosensory genes were successfully identified in *C. stygia*, *B. dorsalis* and *Scaeva pyrastri* antennal transcriptomes \[[@CR50], [@CR72], [@CR73]\]. Here, we identified 154 and 134 candidate chemosensory genes in *E. balteatus* and *E. corollae*, respectively, a number similar to other Diptera antennal transcriptomes (e.g. 128 in *C. stygia*) \[[@CR50]\] but less than the chemosensory genes identified in *D. melanogaster* (254), *M. domestica* (386) and *A. gambiae* (292) genome \[[@CR30], [@CR33], [@CR47], [@CR49], [@CR63], [@CR65], [@CR74]\]. This could be the result of differential expression based on developmental stages of the insect larva or adult olfactory organ development such as maxillary palp and proboscis. All data shows that the chemosensory genes identified by antennal transcriptome sequencing are accurate and reliable. We identified 154 candidate chemosensory genes (51 ORs, 32 IRs, 14 GRs, 49 OBPs, 6 CSPs and 2 SNMPs) in *E. balteatus* and 134 (42 ORs, 23 IRs, 16 GRs, 44 OBPs, 7 CSPs and 2 SNMPs) were identified in *E. corollae*, numbers slightly different compared with those of other Dipteran species \[[@CR50], [@CR73], [@CR75], [@CR76]\]. Such differences could be due to sequencing methods, coverage and/or depth. The number of chemosensory genes is higher in *E. balteatus* than in *E. corollae.* However, assembling and splicing quality (unigene number and N50 length) in *E. corollae* is better than in *E. balteatus.* The differences in the number and quality of transcripts identified could arise from variations in sample preparation or could be due to evolution \[[@CR77]\] and adaptation to the environment (tritrophic interactions). A total of 49 and 44 OBPs were identified in *E. balteatus* and *E. corollae* transcriptomes, respectively. The number of OBPs is variable across species, with 52 members in *D. melanogaster*, 66 in *A. gambiae* (Diptera), 21 in *Apis mellifera* (Hymemoptera), 34 in *Helicoverpa armigera*, 29 in *Helicoverpa assulta* (Lepidoptera), 26 in *Colaphellus bowringi*, 46 in *Tribolium castaneum* (Coleoptera) and 15 in *A. pisum* (Hemiptera) \[[@CR27], [@CR74], [@CR77]--[@CR83]\]. Meanwhile, the OBPs of these two syrphid species are highly divergent with those of other insects. These evolutionary differences may result from different physiological functions or ecological niches. Compared with OBPs, only a small amount of CSPs were detected in Diptera. They are only 4 CSPs in *D. melanogaster*, 4 in *C. stygia* and 7 in *A. gambiae* (Diptera). These numbers are much lower than in other insect orders, such as 18 CSPs found in *H. armigera* and 17 in *H. assulta* (Lepidoptera) \[[@CR27], [@CR50], [@CR80], [@CR81]\]. In our study, six EbalCSPs and seven EcorCSPs are identified in transcriptome sequencing, revealing that the numbers of CSP gene family differ among species. CSPs show a high evolutionary diversity in insecta, probably related to different physiological functions. We identified 51 ORs from *E. balteatus* and 42 ORs from *E. corollae*, respectively. Compared with other Dipteran species, these numbers are similar to those identified in *C. stygia* (50) \[[@CR50]\] and *G. morsitans morsitans* (46) \[[@CR51]\] but lower than those of *D. melanogaster* (62), *M. domestica* (86), *A. gambiae* (79) \[[@CR47], [@CR49], [@CR84]--[@CR86]\], suggesting that sequencing method/depth may be different between studies yielding less genes that may be difficult to detect because of low expression \[[@CR77]\]. Here, we were able to detect species-specific OR transcripts in *E. balteatus* and *E. corollae*. This clade of ORs may have a greater impact on recognizing specific odors, particularly perception of aphids-derived volatiles and herbivore-induced plant volatiles granting syrphid localization access of its prey. The tissue- and sex-specific expression analysis showed no differences between male and female, which is consistent with DEGs analysis of OR transcript abundances using FPKM values. Lepidopteran ORs have shown male-specific expression that is usually involved in the detection of the sex pheromone \[[@CR19], [@CR26], [@CR27]\], but this does not seem to be the case in syrphids. Additional real time quantitative PCR, in situ hybridization and single-sensilla recordings would be required to validate OR expressions and functions. In *D. melanogaster*, males release the volatile sex pheromone *cis*-vaccenyl acetate (cVA) \[[@CR87]--[@CR89]\]. The perception of sex pheromone cVA is mediated by OR67d \[[@CR88]\], OR65a \[[@CR90], [@CR91]\], LUSH \[[@CR92]\], and SNMP1 \[[@CR34]\]. In our two syrphid species, EbalOBP17 and EcorOBP14 are the orthologues of the DmelOBP-lush gene, while EbalOR16 and EcorOR24 are the orthologues of DmelOR67d, and EbalSNMP1 and EcorSNMP1 are very similar to DmelSNMP1, suggesting that these proteins may be involved in detection of their yet unidentified pheromones. Therefore, further functional characterization of these candidate proteins will help reveal any mechanism associated with pheromone reception in *E. balteatus* and *E. corollae*. The *E. balteatus* and *E. corollae* IR family is relatively conserved, especially with respect to common receptors IR8a and IR25a, which are expressed in both olfactory and gustatory systems \[[@CR30], [@CR45]\]. The numbers of IRs identified in *E. balteatus* (32) and *E. corollae* (23) are similar to that of *C. stygia* (22) \[[@CR50]\], but lower than those of *D. melanogaster* (66) and *A. gambiae* (46) \[[@CR63]\]. It is possible that some IRs do not express in antennae tissues or perhaps the number of IRs varies between species and is dependent on natural habitats. A large number of EbalIRs and EcorIRs are clustered with "antennal" orthologues in *Drosophila*, indicating that IRs are highly conserved in Diptera. Furthermore, the IRs identified in these two species may be activated by acids, amines and other odorants that are not sensed by ORs \[[@CR30], [@CR45], [@CR46]\]. In the antennae of *E. balteatus* and *E. corollae*, we identified 14 and 16 candidate GRs, respectively. The total number of GRs in these two species may be much larger, because some members could be exclusively expressed in other gustatory organs, such as maxillary palps, proboscises and legs. However, the numbers are still lower than those reported in other Dipteran antennal transcriptomes \[[@CR50]\]. The conserved receptors identified in the two syrphid species may be involved in CO~2~ perception. However, we infer that the mechanism of CO~2~ perception is different from mosquitoes which concerns host-seeking \[[@CR43], [@CR44], [@CR48], [@CR93]\]. Some GRs may function as taste or contact receptors \[[@CR31]\], particularly with reference to their specific pollination behavior \[[@CR94], [@CR95]\]. Some GRs from these two species are clustered with thermos-sensing GRs and sugar-detecting GRs from *Drosophila*, indicating that they may perform similar functions. Functional analysis of the candidate *E. balteatus* and *E. corollae* chemosensory proteins is required to identify their physiological roles. Conclusions {#Sec14} =========== We have identified and annotated 154 transcripts encoding putative chemosensory proteins in antennal transcriptome of *E. balteatus* and 134 in *E. corollae*. Comparisons between the two syrphid species and among other Dipteran species were deduced using sequence information. This work gives a foundation for future studies aimed at understanding chemical communication in syrphids and tritrophic interactions between plants, herbivorous insects, and natural enemies in agricultural ecosystems. Methods {#Sec15} ======= Insect rearing and tissue collection {#Sec16} ------------------------------------ *E balteatus* and *E. corollae* larvae were fed with aphids (*Aphis gossypii* Glover) and maintained at 22 ± 1 °C with a 12 h light: 12 h dark photo-period at the Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing, China. Following eclosion, adult males and females were separated and provided with pollen and 10% honey solution. Antennae were excised from 2- to 5-day-old adult males and females respectively, and legs were collected together, then immediately frozen and stored in liquid nitrogen. cDNA library construction and Illumina sequencing {#Sec17} ------------------------------------------------- Total RNA of male and female antennae was extracted from *E. balteatus* and *E. corollae* using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). The method for RNA extraction followed in the manufacturer's instruction. Total RNA was dissolved in RNase-free water and RNA integrity was verified by gel electrophoresis. RNA concentration and purity were measured on a Nanodrop ND-2000 spectrophotometer (NanoDrop products, Wilmington, DE, USA). Ten micrograms total RNA of each sample was used to construct the cDNA library. The cDNA library construction and Illumina HiSeq 2000 (Illumina, San Diego, CA, USA) sequencing of the samples was performed at Beijing Genomics Institute (BGI, Shenzhen, China). The insert sequence length was around 200 bp and these libraries were pair-end sequenced using PE100 strategy \[[@CR22], [@CR27]\]. Assembly and function annotation {#Sec18} -------------------------------- Raw reads were pre-processed by filtering low quality reads, trimming low quality nucleotides at each ends and removing 3′ adaptors and poly-A/T tails. Each clean-read dataset of male and female antenna was fed to Trinity \[[@CR71]\]. The Trinity assembly procedure, including Inchworm, Chrysalis and Butterfly were followed using Grabherr *et al.*, 2011 as a reference \[[@CR71]\]. In the first step of Trinity, Inchworm assembles reads into the unique sequences of transcripts using the default parameters (default *k*-mers = 25). Next, Chrysalis clusters related contigs that correspond to portions of alternatively spliced transcripts or otherwise unique portions of paralogous genes. Finally, Butterfly uses read sequences, read-pairings and Chrysalis' read mappings to select the paths that are best supported by read sequences \[[@CR71]\]. The Trinity outputs were clustered by TGICL \[[@CR96]\]. The consensus cluster sequences and singletons make up the unigenes dataset \[[@CR22]\]. The unigenes annotation was performed by NCBI BLASTX against a pooled database of non-redundant and SwissProt protein sequences with e-value \<1e-5. The BLASTX results were then imported into Blast2GO pipeline for GO annotation \[[@CR97]\]. Identification of chemosensory genes {#Sec19} ------------------------------------ Candidate unigenes encoding putative ORs, IRs, OBPs, CSPs, SNMPs and GRs were found by running Perl scripts against transcriptome assembly and annotation in the remote sever. Perl scripts were written to extract sequence from functional annotation results using olfaction keywords. Subsequently, all candidate chemosensory genes were manually checked by BLASTX against local non-redundant database with e-value \<1e-5. Using the BLASTX NCBI database, we manually performed alignments comparing transcripts against all known proteins to examine full-length coverage. The full-length transcripts contain start and termination codons. The ORFs of all putative chemosensory genes were predicted by using ExPASy (Expert Protein Analysis System) server version ([http://web.expasy.org/translate/](http://web.expasy.org/translate)) according to the BLASTX best hit result \[[@CR98]\]. Putative N-terminal signal peptide of OBPs and CSPs were predicted by SignalP 4.0 server version with default parameters \[[@CR99]\]. The TMDs of ORs, IRs and GRs were predicted using TMHMM server version 2.0 \[[@CR100]\]. Sequence and phylogenetic analysis {#Sec20} ---------------------------------- After removing redundancy, alignments of amino acid sequences were performed by MAFFT ([https://www.ebi.ac.uk/Tools/msa/mafft/](https://www.ebi.ac.uk/Tools/msa/mafft)). The phylogenetic trees of *E. balteatus and E. corollae* chemosensory genes were constructed by RaxML version 8 with Jones-Taylor-Thornton amino acid substitution model (JTT) \[[@CR101]\] with the putative chemosensory genes in other Dipteran species (Additional file [8](#MOESM8){ref-type="media"}: Table S4). Node support was assessed using a bootstrap method based on 1000 replicates. The OR data set contained OR sequences identified in Dipteran (51 from *E balteatus*, 42 from *E. corollae*, 62 from *D. melanogaster* \[[@CR85], [@CR86]\], 61 from *B. dorsalis* \[genome: assembly ASM78921v2\], 50 from *C. stygia* \[[@CR50]\] and 81 from *M. domestica* \[[@CR49]\]). The GR data set contained GR sequences identified in Dipteran (14 from *E balteatus*, 16 from *E. corollae*, 68 from *D. melanogaster* \[[@CR47]\], 40 from *B. dorsalis* \[genome: assembly ASM78921v2\], 21 from *C. stygia* \[[@CR50]\] and 43 from *M. domestica* \[[@CR49]\]). The IR data set contained IR sequences identified in Dipteran (32 from *E balteatus*, 23 from *E. corollae*, 76 from *D. melanogaster* \[[@CR30], [@CR63]\], 22 from *C. stygia* \[[@CR50]\] and 54 from *A. gambiae* \[[@CR63], [@CR84]\]). The OBP data set contained OBP sequences identified in Dipteran (49 from *E balteatus*, 44 from *E. corollae*, 71 from *D. melanogaster* \[[@CR64]\], 40 from *B. dorsalis* \[genome: assembly ASM78921v2\], 28 from *C. stygia* \[[@CR50]\] and 52 from *M. domestica* \[[@CR49]\]). The CSP data set contained CSP sequences identified in Dipteran (7 from *E balteatus*, 9 from *E. corollae*, 4 from *D. melanogaster* \[[@CR74]\], 4 from *C. stygia* \[[@CR50]\] and 8 from *A. gambiae* \[[@CR74]\]). The SNMP data set contained SNMP sequences identified in Dipteran (2 from *E balteatus*, 2 from *E. corollae*, 2 from *D. melanogaster* \[[@CR34], [@CR35]\] and 2 from *A. gambiae* \[[@CR35]\]). DEGs analysis {#Sec21} ------------- A mapping-based expression profiling analysis of the chemosensory genes was conducted to compare gene expression between male and female antennae. All of the clean reads were remapped onto the transcripts using SOAPaligner (<http://soap.genomics.org.cn> /soapaligner.html), allowing up to three base mismatches and a minimum length of 40 bp. The FPKM method was used for calculating unigene expression levels \[[@CR20], [@CR50], [@CR102], [@CR103]\]. The suitable *P*-values were calculated to identify differentially expressed genes according to the hypergeometric test \[[@CR103]\]. The FDR was a statistical method used in multiple hypothesis testing to correct for *P*-value. Criteria for estimating significant differential expression was set at FDR ≤ 0.001 and \|log2 Ratio\| ≥ 1. Heatmaps of differential gene expression between male antennae and female antennae in both species were generated by Heml 1.0 software \[[@CR104]\]. Expression analysis by semi-quantitative RT-PCR {#Sec22} ----------------------------------------------- Semi-quantitative RT-PCR was performed to verify the expression of candidate chemosensory genes. Male and female antennae and legs were collected from adult *E. balteatus* and *E. corollae* after eclosion. The extraction of total RNA followed the manufacturer's instruction \[[@CR27]\]. The cDNA was synthesized from total RNA using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Waltham, MA, USA). Gene specific primers were designed using PrimerQuest Tool (<http://sg.idtdna.com/Primerquest/Home/Index>) (Additional file [9](#MOESM9){ref-type="media"}: Table S5) and synthesized by Sangon Biotech Co., Ltd. (Shanghai, China). A Taq MasterMix (CWBIO, Beijing, China) was used for PCR reactions under the general three-step amplification of 94 °C for 30s, 55 °C for 30s, 72 °C for 30s. RT-PCR products were separated on 2% agarose gels, stained by ethidium bromide (EB), and photographed under UV light in Gel Doc XR+ Gel Documentation System with Image Lab Software (Bio-Rad, Hercules, CA, USA). Additional files ================ {#Sec23} Additional file 1: Table S1.Assembly summary of *E. balteatus* and *E. corollae* antennal transcriptome. (DOCX 16 kb) Additional file 2: Fig. S1.(A) Species distribution and annotation summaries in the *E. balteatus* (Ebal) and *E. corollae* (Ecor) antennal transcriptome assembly. (B) Gene ontology classifications of the *E. balteatus* and *E. corollae* unigenes with Blast2GO program, including categories with biological process, molecular function and cellular component. (TIFF 3397 kb) Additional file 3: Table S2.Candidate *E. balteatus* and *E. corollae* antennal chemosensory genes. Unigenes of candidate odorant receptors (2--1), gustatory receptors (2--2), ionotropic receptors (2--3), odorant binding proteins (2--4), chemosensory proteins (2--5) and sensory neuron membrane proteins (2--6) with gene name, length, ORF, best BLASTX hit and identity. (DOCX 112 kb) Additional file 4: Table S3.Comparison of homologous ORs in *E. balteatus* and *E. corollae*. (DOCX 61 kb) Additional file 5: Fig. S2.Protein domain analysis of the species-specific IR clade with *Drosophila* iGluRs and DmelIR94d /e. Amino acid alignments shows the ligand binding domains (S1 and S2), the ion channel pore (P), and TMD (M1, M2 and M3) of ionotropic receptors. The key ligand binding residues are marked in red box. (JPEG 4154 kb) Additional file 6: Fig. S3.Amino acid alignments of the species-specific OBPs clade in the *E. balteatus* and *E. corollae*. The motif of six conserved cysteines are marked with asterisks at the top. (JPEG 3067 kb) Additional file 7:Antennal expression levels of candidate *E. balteatus* and *E. corollae* odorant receptors. (XLSX 99 kb) Additional file 8: Table S4.GenBank accession numbers of chemosensory genes used in phylogenetic analyses. (XLSX 34 kb) Additional file 9: Table S5.Primers of candidate ORs in *E. balteatus* and *E. corollae* used for RT-PCR. (DOCX 20 kb) cDNA : Complementary DNA CSP : Chemosensory protein DEGs : Differentially expressed genes FDR : False discovery rate FPKM : Fragments per kilobase per million fragments GLV : Green leaf volatiles GO : Gene ontology GR : Gustatory receptor IR : Ionotropic receptor OBP : Odorant-binding protein OR : Odorant receptor ORF : Open reading frame PCR : Polymerase chain reaction RT-PCR : Semi-quantitative reverse transcription PCR SNMP : Sensory neuron membrane protein TMD : Transmembrane domain **Electronic supplementary material** The online version of this article (doi:10.1186/s12864-017-3939-4) contains supplementary material, which is available to authorized users. We thank Ph.D. students Kang He, Song Cao, Rui-Bin Zhang and M.S. student Yan-Wei Wang (Institute of Plant Protection, Chinese Academy of Agricultural Sciences) for help collecting insects. We thank Ms. Chun-Yan Wang and Ms. Li-Yan Yang (Institute of Plant Protection, Chinese Academy of Agricultural Sciences) for rearing insects. We are grateful to thank M.S. Xiang-Zhi Liang, Dr. Yong-Lei Liu and Dr. Ya-Nan Zhang for technical assistance. Thanks to Prof. Paolo Pelosi for editorial assistance and comments on the manuscript. Funding {#FPar1} ======= This project is supported by National Natural Science Foundation of China (31,621,064; 31,230,062&31,321,004 to GW, 31,402,023 to BW, 31,471,833 to YL), China Postdoctoral Science Foundation (2014 M550905 to BW), China Postdoctoral Science Foundation 2017T100121 to BW. Availability of data and materials {#FPar2} ================================== The clean reads of the four antennal transcriptomes in this study have been stored in the NCBI SRA database, under the accession number of SRX1812754 (*E. balteatus* male), SRX1812787 (*E. balteatus* female), SRX1829150 (*E. corollae* male), and SRX1829282 (*E. corollae* female). All authors contributed to research design and manuscript preparation. Conceived and designed the experiments: BW, YL, GW. Performed the experiments: BW. Analyzed the data: BW, YL. Contributed reagents/materials/analysis tools: BW, YL, GW. Wrote the paper: BW, YL, GW. All authors read and approved the final manuscript. Competing interests {#FPar3} =================== The authors declare that they have no competing interests. Publisher's Note {#FPar4} ================ Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

{ "pile_set_name": "PubMed Central" }

INTRODUCTION ============ Psoriasis is a chronic immune-mediated inflammatory skin disease that is not explained simply by Mendelian inheritance. Genetic factors play a role in psoriasis development, although the exact causal mechanism of psoriasis is still unknown.^[@r1]^ Psoriasis is seen in nearly 1-3% of the human population all over the world and at every age. Its adverse effect on health is seriously aggravated by physical disruption of daily activities, shame, anxiety, guilt, social withdrawal, social exclusion, depression, and even suicidal tendencies. Psoriasis patients can also experience heart disease, diabetes mellitus, cancer, and depression. A chronic course involving unexplained remissions and exacerbations requires patients to undergo a treatment and follow-up process that extends throughout their lives.^[@r2]-[@r4]^ Psoriasis usually appears in the form of well-demarcated, scaly plaques and has been classified in the group of immune-mediated inflammatory diseases, alongside ulcerative colitis, Crohn\'s disease, rheumatoid arthritis, ankylosing spondylitis, and Behçet\'s disease.^[@r5],[@r6]^ It is known that antimicrobial peptides (AMPs), which provide high resistance to skin infection, show increased synthesis in psoriatic lesions.^[@r7]^ AMP production in the skin is the primary mechanism that protects skin against infections, and AMP levels are increased as a result of microbial infections.^[@r8]^ Granulysin is a member of the AMP family of cytolytic and proinflammatory peptides. The peptide exists in granules of T cells and natural killer (NK) cells and is secreted from these cells together with granzyme and perforin.^[@r9]^ Based on amino acid sequence homology, the 9-kDa granulysin protein belongs to the saposin-like protein (SAPLIP) family, which consists of activators of sphingolipid hydrolase in the central nervous system. Granulysin is tumoricidal and broadly antimicrobial, killing gram-positive and gram-negative bacteria, yeast, fungi, and parasites. Granulysin binds to the cell surface based on charge and appears to make a way through the cell membrane, causing ion fluxes.^[@r10]^ A recent immunohistochemistry study showed that granulysin expression is increased in psoriasis lesions.^[@r7]^ Granulysin encoding by the GNLY gene is located on 2 p11.2, and several polymorphisms have been identified in GNLY that contain 5 exons. C to T substitution in GNLY (rs10180391 (C/T) is located on 2: 85.699.222, and C to G substitution on GNLY (rs 7908 (C/G) is located on 2:85.698.633.^[@r11]^ It has also been shown that GNLY gene polymorphism is associated with infectious diseases such as chronic hepatitis B, but to our knowledge there has been no study on the relationship between GNLY gene polymorphisms and psoriasis.^[@r12]^ Our study investigated the role of GNLY gene polymorphisms in the etiopathogenesis of psoriasis and can help establish a psoriasis susceptibility profile. METHODS ======= Study subjects -------------- The study was conducted in 100 patients diagnosed with psoriasis in the Polyclinic for Skin and Venereal Diseases of Bulent Ecevit University Application and Research Center. A group of 100 individuals with no autoimmune or autoinflammatory disorders was used as the control group. Study participants were matched for gender and age distribution. The age and gender of all the subjects included in the study were recorded, and patients were divided into groups according to the Psoriasis Area and Severity Index (PASI) score. The study was approved by the Clinical Research Ethics Committee, Bulent Ecevit University (Date 28.05.2015 approval number 2015-24-26/05) and performed in accordance with the ethical standards established in the 1964 Declaration of Helsinki and later amendments. All participants were informed and signed consent forms before being included. DNA extraction and genotype analyses ------------------------------------ In order to study the GNLY rs7908 and rs10180391 gene polymorphisms, 2ml blood samples were collected from patients and healthy controls during routine follow-up. Genomic DNA was extracted from 250µL of peripheral blood with the E.Z.N.A.^®^ Blood DNA Mini Kit (Omega Bio-tek Inc., Atlanta GA), according to the manufacturer\'s protocol. Polymerase chain reaction (PCR) restriction fragment length polymorphism (RFLP)-based analysis was used to genotype GNLY rs7908 (C/G) and GNLY rs10180391(C/T) polymorphisms using PCR primers ([Tables 1](#t1){ref-type="table"} and [2](#t2){ref-type="table"}). ###### SNP ID, primer sequences, and annealing temperature ----------------------------------------------------------------------------- SNP ID Primer sequencing Annealing temp. (^0^C) ------------------------- -------------------------- ------------------------ **rs7908** **F:** 5\'- TGT TCA GTA\ 56 0C GGG TCA GGT GG-3\' **R:** 5\'-GAT TCT GGA\ TCG AGG AAG CG -3\' **rs10180391** **F:** 5\'- AAG CAA CAG\ 55 0C AAG TCT CAG CC-3\' **R:** 5\'-CTG CCA AGA\ GAG AAG CGA AC -3\' ----------------------------------------------------------------------------- ###### SNP ID, restriction enzymes for detecting each single-nucleotide polymorphism (SNP), cutting sequencing, working temperature, incubation time and allele size (bp) SNP ID Restriction enzymes Cutting sequencing Working temperature Incubation time Allele size (bp) ---------------- --------------------- -------------------- --------------------- ----------------- --------------------- 5\'-CC\^TCAGC-3\' GG: 225 bp **rs7908** BbvC1 37°C 16 hours GC: 225, 183, 42 bp 3\'-GGAGT\^CG-5\' CC: 183, 42 bp 5\'-AT\^TAAT-3\' CC: 209 bp **rs10180391** Ase1 37°C 16 hours CT: 209, 145, 64 bp 3\'-TAAT\^TA-5\' TT: 145,64 bp PCR was performed in 25µL containing 20 to 100ng genomic DNA, 1.5mM MgCl~2~, 1x PCR buffer with (NH~4~)~2~SO~4~, 0.25mM deoxynucleotide triphosphates (dNTPs), 10% dimethylsulphoxide, 0.5 units of Taq polymerase, and 20pmol of each primer. Amplification conditions for the GNLY rs7908(C/G) polymorphism were initial denaturation at 95°C for 3 minutes, then 35 amplification cycles of 60s at 95°C, 90s at 56°C, and 60s at 72°C, and a final 7-minute extension step at 72°C. The PCR products were checked on 1.5% agarose gel for assay completion, and then the PCR products with 225 base pairs (bp) were digested overnight with BbvCI restriction enzyme (New England *Biolabs*) ([Figure 1](#f1){ref-type="fig"}). Figure 1*GNLY* rs7908 gene poly morphism gel electrophoresis image. DNA marker is 100 bp. Samples with 3,4 are GC; samples with 1, 6 are CC and 2, 5 with GG genotype Amplification conditions for the GNLY rs10180391(C/T) polymorphism were initial denaturation at 95°C for 3 minutes, then 35 amplification cycles of 95°C for 60s, 55°C for 90s, and 72°C for 60s, and a final 7-minute extension step at 72°C. The PCR products were checked on 1.5% agarose gel for assay completion, and then the PCR products with 209 base pairs (bp) were digested overnight with AseI restriction enzyme (Thermo Scientific) ([Figure 2](#f2){ref-type="fig"}). Figure 2*GNLY* rs10180391 gene poly morphism gel electrophoresis image. DNA marker is100 bp. Samples 2,4,5 CT; samples 1 and 6 are in CC and 3 are in TT genotype All digestion products were electrophoresed on 3% agarose gel and viewed by staining with ethidium bromide and evaluated using a gel documentation system (Vilber Lourmat Bio-Print ST4). Statistical analyses -------------------- A case-control study was performed, and allelic frequencies of the polymorphisms were calculated both for cases and controls. The χ^2^ test was used to compare the genotype frequency of each gene polymorphism in psoriasis patients and controls. Odds ratios (OR) and 95% confidence intervals (CI) were calculated to compare the psoriasis risk for the alleles. Distribution of data was determined by the Shapiro-Wilks test. Continuous variables are expressed as mean and standard deviation (SD) or median, and categorical variables as frequency and percent. Continuous variables were compared with the independent-sample t-test or the Mann-Whitney U-test. P values \<0.05 were considered statistically significant. SPSS version 18.0 was used (SPSS Inc., Chicago, IL). RESULTS ======= The study enrolled 100 psoriasis patients and 100 healthy controls. There was no statistically significant difference in age distribution between patients and controls (p = 0.315). Mean age (SD) was 41.31 years (13.46) in patients and 39.49 years (12.06) in controls ([Table 3](#t3){ref-type="table"}). ###### Sex and age distribution of psoriasis patients and controls Healthy controls n (%) Psoriasis patients n (%) *P* value ------------ ------------------------ -------------------------- ----------- **Male** 43(43%) 44(44%) **Female** 57(57%) 5656(%) **Age** 39,49±12.06 41,31±13.46 0.315 As seen in [table 4](#t4){ref-type="table"}, the study showed a significant association between the GNLY rs7908 (C/G) polymorphism and psoriasis versus normal controls (OR=0.305 95% CI=0.121 - 0.773; p=0.033). The GNLY rs7908 CC genotype was significantly more frequent when compared to the control group (GG genotype) (p=0.033, OR=0.305 CI=0.121 - 0.773). Distributions of genotypes of the GNLY rs10180391 (C/T) polymorphism were similar in psoriasis patients and controls, and no relationship was detected between these groups (p \> 0.05). No statistically significant difference was found between patients and controls for GNLY rs10180391 allele frequencies (p = 0.604, OR = 1.114, CI = 0.741-1.627). Neither was there any statistically significant difference in GNLY rs7908 (C/G) and GNLY rs10180391 (C/T) gene polymorphisms in terms of psoriasis severity (PASI) ([Graph 1](#f3){ref-type="fig"}). In addition, there was no statistically significant association between GNLY rs7908 (C/G) and rs10180391 (C/T) gene polymorphisms in terms of haplotype ([Table 5](#t5){ref-type="table"}). ###### GNLY genotypes and alleles and the risk of developing psoriasis (n=100) All PASI\<3 3≤PASI\<1 PASI≥10 OR (95% CI) *P* ------------------------- ------------ ------------ ------------ ------------ ----------- -------------------- ------- **Rs7908/Genotype** **GG** 37 (37%) 51 (51%) 28(46.7%) 17 (53.1%) 6 (75%) Reference 0.033 GC 44 (44%) 41 (41%) 25(41.7%) 14 (43.8%) 2 (25%) 0.676(0.371-1.232) CC 19 (19%) 8 (8%) 7(11.7%) 1 (3.1%) 0 (0.0%) 0.305(0.121-0.773) **Rs7908/Allele** G 118 (59%) 143(71.5%) 81 (67.5%) 48 (75%) 14(87.5%) Reference 0.009 C 82 (41%) 57 (28.5%) 39(32.5%) 16 (25%) 2 (12.5%) 0.574(0.378-0.870) **rs10180391/Genotype** CC 44(44%) 41 (41%) 28(46.7%) 11 (34.4%) 2 (25%) Reference 0.886 CT 41 (41%) 42 (42%) 22(36.7%) 15 (46.9%) 5 (62.5%) 1.099(0.600-2.013) TT 15 (15%) 17 (17%) 10(16.7%) 6 (18.8%) 1 (12.5%) 1.216(0.539-2.746) **rs10180391/Allele** C 129(64.5%) 124 (62%) 78(65%) 37(57.8%) 9 (56.3%) Reference 0.604 T 71 (35.5%) 76 (38%) 42(35%) 27(42.2%) 7 (43.7%) 1.114(0.741-1.672) Graph 1The relationship between PASI score and GNLY genotypes ###### Haplotype distribution of GNLY rs7908 - rs10180391 gene polymorphisms ------------------------------------------------------------------------------- Haplotype Healthy controls Psoriasis patients *P* value OR (95% CI) ----------- ------------------ -------------------- ----------- --------------- **GC** 39 % 42.5 % 0.036 Reference **GT** 20 % 29 % 0.269 1.331\ (0.802-2.208) **CC** 25.5 % 19.5 % 0.180 0.702\ (0.418-1.178) **CT** 15.5 % 9 % 0.060 0.533\ (0.276-1.028) ------------------------------------------------------------------------------- DISCUSSION ========== Psoriasis is a dermatosis characterized by sharply demarcated erythematous, pearly-white scales whose etiology is not clearly known.^[@r5]^ According to current knowledge, genetic factors contribute to age at onset, clinical manifestations, and type and severity of the disease. It has been hypothesized that genetic differences may lead to the development of the disease (PSORS2) and/or affect the initial antigen presentation (PSORS1 associated with the HLA-Cw6 locus).^[@r13],[@r14]^ The most widely accepted mechanism in the etiopathogenesis of psoriasis involves chemokines, which play a key role in cytokine and lymphocyte activation as the fundamental step in this process, followed by keratinocyte hyperproliferation and inflammation.^[@r15],[@r16]^ Granulysin, a positively charged cytolytic antimicrobial peptide (AMP) class found in the granules of cytotoxic T lymphocytes (CTL) and natural killer cells (NK), was first described during the investigation of genes expressed in effector T lymphocytes.^[@r17]^ Granulysin has been associated with a number of diseases, including various infections, cancer, transplantation, skin disorders, and reproductive complications.^[@r18]^ We investigated the relationship between psoriasis, an autoinflammatory disease, and rs7908 (G/C) and rs10180391 (C/T) gene polymorphisms in the GNLY gene. We found no study in the literature on the relationship between psoriasis and GNLY gene polymorphisms. The role of GNLY gene polymorphisms has been investigated in chronic liver disease and HIV infection.^[@r12],[@r19]^ The difference between the two groups in the genotype distribution of GNLY rs7908 (C/G) polymorphism was statistically significant. CC genotype was significantly more frequent than GG genotype in the control group. In terms of allelic frequencies, the C allele has been found to have a protective effect against psoriasis. Due to the alternative splicing mechanism from the GNLY gene, four different protein variants are synthesized in the same cell. In GNLY rs7908 (C/G) variants 1, 2, and 3, this polymorphism occurs in the 3\'UTR region.^[@r20]^ Stabilization of the mRNA, exocytosis, and subcellular localization are affected by 3\'UTR. Therefore, the gene expression is post-transcriptionally regulated.^[@r21]^ Nucleotide changes in the 3\'UTR region lead to serious disruption.^[@r22]^ In the variant GNLY rs 7908, C/G nucleotide exchange is located in the protein-coding region, causing the amino acid transformation of leucine → valine (Leu150Val). Therefore, it is thought that GNLY rs7908 C allele reduces the amount or activity of granulysin protein, leading to the complete elimination of trigger factors causing psoriasis. The amino acid change imposed by the rs7908 SNP can be considered as a reason of activity decrease of the enzyme. Although both leucine and valine are nonpolar and hydrophobic, valine differs structurally from leucine, containing an extra methylene (-CH~2~) group. Most likely, the absence of the methyl group causes distortions in intramolecular interactions and decreases in enzyme activity. When the GNLY rs10180391 gene polymorphism genotype distribution and allele frequencies were analyzed, no statistically significant difference was found between the patient and control groups. In contrast, Hou *et al.* (2015) found a statistically significant association between GNLY rs1866139 and GNLY rs11127 and HBV infection.^[@r12]^ This may be due to the fact that the GNLY rs1866139 and GNLY rs11127 polymorphisms are not effective in protein activity or amount. When the patient group was classified as PASI\<3, 3 ≤ PASI\<10, PASI≥10 according to the PASI severity score, there was no statistically significant change in terms of the two polymorphisms. However, CC ratio was decreased in GNLY rs7908 polymorphism as the PASI score increased with GG and GC. In other words, CC genotype frequency decreased as disease severity increased. This decrease was not statistically significant because of the low number of cases. The study showed that GNLY rs7908 G allele is associated with disease severity and that the C allele protects against psoriasis. The increase in psoriasis severity associated with the GNLY rs7908 G allele in our study suggests that this allele may have an enhancing effect on granulase expression. Similarly, Elgarhy *et al.* (2015) found that the severity of psoriasis increased with the increase in granulysin expression. They corroborated our findings, and their immunohistochemical analyses showed that granulysin expression is increased in psoriasis lesions.^[@r7]^ Raychaudhuri *et al.* (2004) found increased granulysin levels in lesional T cells and dermal dendrocytes in psoriasis plaques.^[@r23]^ Pivarcsi *et al.* (2004) showed that lichen planus (LP) granulysin was expressed 100-200 times above the normal level.^[@r24]^ Granulysin has been identified as an important effector in both Stevens-Johnson syndrome (SJS) and diffuse keratinocyte death in toxic epidermal necrolysis (TEN).^[@r25]^ Tewary *et al.* (2010) demonstrated that granulysin promotes CTL proliferation by acting as an alarmin with chemoattractant and stimulatory effects on dendritic cells and macrophages.^[@r26]^ Our study suggests that the rs7908 C/G polymorphism in the GNLY gene may also reduce the alarmin function of granulysin, thus allowing the cells of the immunopathogenesis of psoriasis to continue to function normally. Our study analyzed GNLY rs7908 and GNLY rs10180391 SNPs and found that the C allele of rs7908 is protective against psoriasis and even inversely proportional to psoriasis severity. There are very few studies in the literature that examine the interactions between GNLY gene polymorphisms and immunological diseases. There is thus serious need for further research to identify the effect of granulysin in psoriasis. Psoriasis prevalence is reported at 1.5-2.8% in the United Kingdom, 2.5-3.5% in Germany, 3.15% in the United States, and 0.05-1.23% in China, and the highest prevalence in Asia was reported at 11.8%.^[@r27]^ Since prevalence of the disease differs between populations, our research in the Turkish population can be useful for elucidating the role of GNLY in psoriasis. CONCLUSION ========== We found that genotype GNLY rs7908 CC and C allele had a protective effect against psoriasis and reduced the severity of the disease. The rs10180391 SNP did not play a significant role in the pathogenesis of psoriasis. Investigation of the effect of GNLY rs7908 and GNLY rs10180391 gene polymorphisms on psoriasis risk and disease progression can lead to the development of new treatment modalities for psoriasis. Acknowledgments: The research was supported by the Unit of Scientific Research Projects of Bulent Ecevit University (BAP-2015-50737594-04), Turkey. Work conducted at the Bulent Ecevit University, Zonguldak, Turkey. Financial Support: The research was supported by the Unit of Scientific Research Projects of Bulent Ecevit University (BAP-2015-50737594-04), Turkey. Conflict of Interests: None. AUTHORS\'CONTRIBUTIONS Esra Ermis 0000-0001-6233-2420 Conception and planning of the study; Obtaining, analyzing, and interpreting the data; Elaboration and writing of the manuscript; Effective participation in research orientation; Critical review of the literature Sevim Karakas Celik 0000-0003-0505-7850 Statistical analysis; Approval of the final version of the manuscript; Conception and planning of the study; Critical review of the literature; Critical review of the manuscript Nilgun Solak 0000-0002-6572-9615 Conception and planning of the study Gunes Cakmak Genc 0000-0001-7222-0377 Conception and planning of the study; Obtaining, analyzing, and interpreting the data Ahmet Dursun 0000-0002-7625-837X Approval of the final version of the manuscript; Critical review of the literature

{ "pile_set_name": "PubMed Central" }

![](hosplond73813-0005){#sp1 .137} ![](hosplond73813-0006){#sp2 .138}

{ "pile_set_name": "PubMed Central" }

Related literature {#sec1} ================== For the synthesis and properties of quinolone derivatives, see: Anderson & Osheroff (2001[@bb1]); Ball *et al.* (1998[@bb2]); Hong *et al.* (1997[@bb5]); Ray *et al.* (2005[@bb6]); Wang *et al.* (2008[@bb9]). Experimental {#sec2} ============ {#sec2.1} ### Crystal data {#sec2.1.1} C~18~H~33~N~3~O~5~*M* *~r~* = 371.47Monoclinic,*a* = 28.867 (3) Å*b* = 6.1887 (13) Å*c* = 25.379 (3) Åβ = 112.769 (2)°*V* = 4180.6 (11) Å^3^*Z* = 8Mo *K*α radiationμ = 0.09 mm^−1^*T* = 298 (2) K0.40 × 0.20 × 0.11 mm ### Data collection {#sec2.1.2} Bruker SMART CCD area-detector diffractometerAbsorption correction: multi-scan (*SADABS*; Sheldrick, 1996[@bb7]) *T* ~min~ = 0.963, *T* ~max~ = 0.99110032 measured reflections3699 independent reflections1915 reflections with *I* \> 2σ(*I*)*R* ~int~ = 0.060 ### Refinement {#sec2.1.3} *R*\[*F* ^2^ \> 2σ(*F* ^2^)\] = 0.050*wR*(*F* ^2^) = 0.144*S* = 1.023699 reflections244 parametersH-atom parameters constrainedΔρ~max~ = 0.23 e Å^−3^Δρ~min~ = −0.22 e Å^−3^ {#d5e367} Data collection: *SMART* (Bruker, 1998[@bb3]); cell refinement: *SAINT* (Bruker, 1999[@bb4]); data reduction: *SAINT*; program(s) used to solve structure: *SHELXS97* (Sheldrick, 2008[@bb8]); program(s) used to refine structure: *SHELXL97* (Sheldrick, 2008[@bb8]); molecular graphics: *SHELXTL* (Sheldrick, 2008[@bb8]); software used to prepare material for publication: *SHELXTL*. Supplementary Material ====================== Crystal structure: contains datablocks global, I. DOI: [10.1107/S1600536808044255/rk2125sup1.cif](http://dx.doi.org/10.1107/S1600536808044255/rk2125sup1.cif) Structure factors: contains datablocks I. DOI: [10.1107/S1600536808044255/rk2125Isup2.hkl](http://dx.doi.org/10.1107/S1600536808044255/rk2125Isup2.hkl) Additional supplementary materials: [crystallographic information](http://scripts.iucr.org/cgi-bin/sendsupfiles?rk2125&file=rk2125sup0.html&mime=text/html); [3D view](http://scripts.iucr.org/cgi-bin/sendcif?rk2125sup1&Qmime=cif); [checkCIF report](http://scripts.iucr.org/cgi-bin/paper?rk2125&checkcif=yes) Supplementary data and figures for this paper are available from the IUCr electronic archives (Reference: [RK2125](http://scripts.iucr.org/cgi-bin/sendsup?rk2125)). This work was supported by the IMB Research Foundation. Comment ======= Quinolones, a class of synthetic antibacterial compounds based on a 4--quinolone skeleton, have been the landmark discovery in the treatment of bacterial infections in both community and hospital setting (Ray *et al.*, 2005; Ball *et al.*, 1998;). The most intensive structural variations have been carried out on the basic group at the C--7 position, partially due to the ease of their introduction through a nucleophilic aromatic substitution reaction on the corresponding halide. Piperazine, piperidine, pyrrolidine and their derivatives have been the most successfully employed side chains, as evidenced by the compounds currently on the market (Anderson & Osheroff, 2001; Hong *et al.*, 1997). Recently, as part of an ongoing program to find potent new quinolones displaying strong Gram--positive activity, we have focused our attention on introducing new functional groups to the piperidine ring (Wang *et al.*, 2008). We report here the crystal structure of the title compound, which is a key intermediate of 4--methoxyimino--3--methylamino--3--methylpiperidine, a novel C--7 substituent of the quinolones. The oxime geometry of the title compound was confirmed to have the *E*--configuration. In the molecule of the compound (Fig. 1), the N1---C6 (1.352 (3) Å) and N2---C12 (1.359 (3) Å) bond lengths are significantly shorter than the normal C---N bond (1.47 Å), indicating some conjugation with the C6═O2 and C12═O4 carbonyl groups, respectively. The six--membered piperidine ring adopts a chair conformation with displacing N1 and C3 atoms (0.593 (3) Å and -0.654 (3) Å respectively) from the mean--plane (C1, C2, C4 and C5). Experimental {#experimental} ============ To a stirring solution of 1-*N*-*tert*-Butoxycarbonyl-3-(*N*-*tert*-butoxycarbonyl) amino-4-methoxyimino-3-methylpiperidine(2.4 g, 6.7 mmol) in dry tetrahydrofuran (40 ml) was added 70% sodium hydride (0.46 g, 13.4 mmol) at 273 K using an ice bath, and then stirred for 0.5 h at the room temperature. After addition of methyl iodide (0.84 ml, 13.4 mol), the reaction mixture was stirred at 313 K for 5 h and cooled to room temperature, adjusted to pH 7 with 1*N* HCl and then concentrated under reduced pressure. The residue was diluted with ethyl acetate (50 ml), washed with distilled water, dried over anhydrous sodium sulfate, and filtered. The filtrate was concentrated under reduced pressure, dried *in vacuo* to give the title compound as a white solid (2.37 g, 95.0%; mp: 380--382 K). Single crystals suitable for X--ray analysis were obtained by slow evaporation of a methanol/water solution (5:1 *v*/*v*). ^1^H NMR (CDCl~3~, δ): 1.34 (s, 3H, CH~3~), 1.41 (s, 9H, *BOC*), 1.46 (s, 9H, *BOC*), 2.24--2.25 (m, 1H, piperidine), 2.87--2.88 (m, 1H, piperidine), 2.91 (s, 3H, NCH~3~), 2.95--3.08 (m, 2H, piperidine), 3.82 (s, 3H, OCH~3~), 3.84--3.86 (m, 1H, piperidine), 4.30--4.31 (m, 1H, piperidine); MS (ESI, m/*z*): 372.2 m/*z* (*M*+1)^+^. Refinement {#refinement} ========== All H atoms were placed at calculated positions, with C---H = 0.96--0.97 Å, and included in the final cycles of refinement using a riding model, with *U*~iso~(H) = 1.2*U*~eq~(C) for methylene or 1.5*U*~eq~(C) for methyl H atoms. Figures ======= ![The molecular structure of the title compound with the atom--numbering scheme. Displacement ellipsoids are drawn at 40% probability level. H atoms are presented as a small spheres of arbitrary radius.](e-65-0o256-fig1){#Fap1} Crystal data {#tablewrapcrystaldatalong} ============ ------------------------ --------------------------------------- C~18~H~33~N~3~O~5~ *F*(000) = 1616 *M~r~* = 371.47 *D*~x~ = 1.180 Mg m^−3^ Monoclinic, *C*2/*c* Mo *K*α radiation, λ = 0.71073 Å Hall symbol: -C 2yc Cell parameters from 1707 reflections *a* = 28.867 (3) Å θ = 2.7--21.1° *b* = 6.1887 (13) Å µ = 0.09 mm^−1^ *c* = 25.379 (3) Å *T* = 298 K β = 112.769 (2)° Prism, colourless *V* = 4180.6 (11) Å^3^ 0.40 × 0.20 × 0.11 mm *Z* = 8 ------------------------ --------------------------------------- Data collection {#tablewrapdatacollectionlong} =============== --------------------------------------------------------------- -------------------------------------- Bruker SMART CCD area-detector diffractometer 3699 independent reflections Radiation source: Fine--focus sealed tube 1915 reflections with *I* \> 2σ(*I*) Graphite *R*~int~ = 0.060 φ and ω scans θ~max~ = 25.0°, θ~min~ = 1.5° Absorption correction: multi-scan (*SADABS*; Sheldrick, 1996) *h* = −33→34 *T*~min~ = 0.963, *T*~max~ = 0.991 *k* = −7→6 10032 measured reflections *l* = −30→23 --------------------------------------------------------------- -------------------------------------- Refinement {#tablewraprefinementdatalong} ========== ------------------------------------- ------------------------------------------------------------------------------------------------ Refinement on *F*^2^ Primary atom site location: structure-invariant direct methods Least-squares matrix: full Secondary atom site location: difference Fourier map *R*\[*F*^2^ \> 2σ(*F*^2^)\] = 0.050 Hydrogen site location: inferred from neighbouring sites *wR*(*F*^2^) = 0.144 H-atom parameters constrained *S* = 1.02 *w* = 1/\[σ^2^(*F*~o~^2^) + (0.058*P*)^2^ + 0.3657*P*\] where *P* = (*F*~o~^2^ + 2*F*~c~^2^)/3 3699 reflections (Δ/σ)~max~ = 0.001 244 parameters Δρ~max~ = 0.23 e Å^−3^ 0 restraints Δρ~min~ = −0.22 e Å^−3^ ------------------------------------- ------------------------------------------------------------------------------------------------ Special details {#specialdetails} =============== --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Geometry. All s.u.\'s (except the s.u. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell s.u.\'s are taken into account individually in the estimation of s.u.\'s in distances, angles and torsion angles; correlations between s.u.\'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell s.u.\'s is used for estimating s.u.\'s involving l.s. planes. Refinement. Refinement of *F*^2^ against ALL reflections. The weighted *R*--factor *wR* and goodness of fit *S* are based on *F*^2^, conventional *R*--factors *R* are based on *F*, with *F* set to zero for negative *F*^2^. The threshold expression of *F*^2^ \> σ(*F*^2^) is used only for calculating *R*--factors(gt) *etc*. and is not relevant to the choice of reflections for refinement. *R*--factors based on *F*^2^ are statistically about twice as large as those based on *F*, and *R*--factors based on ALL data will be even larger. --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å^2^) {#tablewrapcoords} ================================================================================================== ------ --------------- ------------- -------------- -------------------- -- *x* *y* *z* *U*~iso~\*/*U*~eq~ N1 0.18007 (8) 0.2543 (3) 0.60235 (9) 0.0389 (6) N2 0.07900 (7) 0.1687 (4) 0.51326 (9) 0.0367 (6) N3 0.16321 (7) −0.1321 (4) 0.47432 (9) 0.0382 (6) O1 0.16324 (7) 0.3113 (3) 0.67977 (7) 0.0496 (5) O2 0.21149 (7) 0.5507 (3) 0.65653 (8) 0.0554 (6) O3 0.07112 (6) 0.1423 (3) 0.42229 (7) 0.0525 (6) O4 0.01093 (7) 0.3134 (3) 0.44214 (8) 0.0607 (6) O5 0.19807 (6) −0.0836 (3) 0.44880 (7) 0.0452 (5) C1 0.15544 (9) 0.0442 (4) 0.59370 (10) 0.0374 (7) H1A 0.1361 0.0344 0.6175 0.045\* H1B 0.1809 −0.0681 0.6057 0.045\* C2 0.12010 (9) 0.0043 (4) 0.53070 (10) 0.0335 (6) C3 0.15423 (9) 0.0354 (4) 0.49846 (10) 0.0325 (6) C4 0.17891 (10) 0.2510 (4) 0.50509 (11) 0.0387 (7) H4A 0.1536 0.3625 0.4899 0.046\* H4B 0.2006 0.2542 0.4839 0.046\* C5 0.20975 (10) 0.2938 (5) 0.56825 (11) 0.0459 (8) H5A 0.2390 0.2005 0.5813 0.055\* H5B 0.2213 0.4425 0.5732 0.055\* C6 0.18713 (10) 0.3856 (5) 0.64739 (11) 0.0399 (7) C7 0.15940 (11) 0.4430 (5) 0.72637 (12) 0.0495 (8) C8 0.21026 (13) 0.4764 (6) 0.77309 (13) 0.0738 (11) H8A 0.2261 0.3388 0.7856 0.111\* H8B 0.2066 0.5494 0.8046 0.111\* H8C 0.2306 0.5623 0.7589 0.111\* C9 0.13357 (14) 0.6544 (6) 0.70277 (15) 0.0871 (13) H9A 0.1561 0.7461 0.6934 0.131\* H9B 0.1241 0.7241 0.7309 0.131\* H9C 0.1041 0.6272 0.6690 0.131\* C10 0.12728 (14) 0.3024 (6) 0.74676 (14) 0.0903 (13) H10A 0.0957 0.2762 0.7158 0.135\* H10B 0.1216 0.3738 0.7773 0.135\* H10C 0.1441 0.1674 0.7603 0.135\* C11 0.09808 (10) −0.2224 (4) 0.52538 (12) 0.0432 (7) H11A 0.0769 −0.2306 0.5466 0.065\* H11B 0.1248 −0.3256 0.5403 0.065\* H11C 0.0786 −0.2537 0.4859 0.065\* C12 0.05026 (10) 0.2153 (5) 0.45784 (12) 0.0417 (7) C13 0.05602 (10) 0.2229 (5) 0.55381 (11) 0.0492 (8) H13A 0.0295 0.3257 0.5367 0.074\* H13B 0.0810 0.2841 0.5877 0.074\* H13C 0.0425 0.0944 0.5636 0.074\* C14 0.04913 (11) 0.1885 (6) 0.36115 (12) 0.0584 (9) C15 0.04440 (16) 0.4286 (7) 0.35091 (17) 0.1048 (14) H15A 0.0185 0.4844 0.3620 0.157\* H15B 0.0360 0.4573 0.3111 0.157\* H15C 0.0757 0.4972 0.3731 0.157\* C17 0.08735 (13) 0.0915 (7) 0.34093 (13) 0.0907 (13) H17A 0.1198 0.1530 0.3622 0.136\* H17B 0.0779 0.1217 0.3010 0.136\* H17C 0.0887 −0.0621 0.3467 0.136\* C16 −0.00059 (12) 0.0715 (7) 0.33491 (14) 0.0906 (13) H16A 0.0039 −0.0778 0.3460 0.136\* H16B −0.0129 0.0822 0.2940 0.136\* H16C −0.0244 0.1360 0.3481 0.136\* C18 0.20473 (11) −0.2741 (5) 0.42114 (13) 0.0557 (9) H18A 0.2217 −0.3815 0.4493 0.083\* H18B 0.2245 −0.2411 0.3992 0.083\* H18C 0.1725 −0.3283 0.3962 0.083\* ------ --------------- ------------- -------------- -------------------- -- Atomic displacement parameters (Å^2^) {#tablewrapadps} ===================================== ----- ------------- ------------- ------------- -------------- ------------- -------------- *U*^11^ *U*^22^ *U*^33^ *U*^12^ *U*^13^ *U*^23^ N1 0.0452 (13) 0.0367 (15) 0.0395 (13) −0.0085 (11) 0.0215 (11) −0.0060 (12) N2 0.0359 (12) 0.0426 (15) 0.0368 (13) 0.0076 (11) 0.0198 (11) 0.0002 (11) N3 0.0368 (12) 0.0427 (15) 0.0420 (13) 0.0011 (11) 0.0226 (11) −0.0024 (12) O1 0.0665 (13) 0.0501 (13) 0.0394 (11) −0.0098 (11) 0.0284 (10) −0.0081 (10) O2 0.0687 (14) 0.0486 (14) 0.0528 (13) −0.0193 (12) 0.0277 (11) −0.0143 (11) O3 0.0471 (11) 0.0771 (16) 0.0335 (11) 0.0188 (11) 0.0157 (9) 0.0062 (11) O4 0.0469 (12) 0.0716 (16) 0.0619 (14) 0.0252 (12) 0.0193 (10) 0.0080 (12) O5 0.0509 (11) 0.0453 (13) 0.0538 (12) −0.0003 (10) 0.0360 (10) −0.0070 (10) C1 0.0438 (16) 0.0357 (18) 0.0371 (16) −0.0012 (14) 0.0203 (13) −0.0006 (13) C2 0.0361 (15) 0.0331 (17) 0.0349 (15) 0.0017 (13) 0.0178 (12) 0.0011 (13) C3 0.0314 (14) 0.0359 (17) 0.0331 (15) 0.0034 (13) 0.0155 (12) 0.0006 (13) C4 0.0451 (16) 0.0350 (17) 0.0442 (17) −0.0010 (14) 0.0264 (13) −0.0012 (14) C5 0.0468 (17) 0.0457 (19) 0.0510 (18) −0.0088 (15) 0.0254 (15) −0.0077 (15) C6 0.0454 (17) 0.0398 (19) 0.0361 (16) 0.0002 (15) 0.0174 (14) −0.0027 (15) C7 0.066 (2) 0.050 (2) 0.0392 (17) 0.0003 (17) 0.0280 (16) −0.0052 (16) C8 0.090 (3) 0.084 (3) 0.0401 (18) 0.001 (2) 0.0172 (19) −0.0105 (19) C9 0.110 (3) 0.089 (3) 0.074 (3) 0.044 (3) 0.049 (2) 0.007 (2) C10 0.127 (3) 0.099 (3) 0.071 (2) −0.034 (3) 0.066 (2) −0.023 (2) C11 0.0486 (16) 0.0375 (18) 0.0494 (17) −0.0058 (14) 0.0252 (14) −0.0037 (15) C12 0.0406 (17) 0.0440 (19) 0.0439 (18) 0.0051 (15) 0.0201 (14) 0.0005 (15) C13 0.0461 (17) 0.058 (2) 0.0512 (18) 0.0062 (15) 0.0276 (15) −0.0066 (16) C14 0.054 (2) 0.082 (3) 0.0373 (17) 0.0154 (19) 0.0156 (15) 0.0125 (18) C15 0.120 (3) 0.109 (4) 0.092 (3) 0.012 (3) 0.048 (3) 0.044 (3) C17 0.079 (2) 0.151 (4) 0.046 (2) 0.030 (3) 0.0297 (19) 0.010 (2) C16 0.072 (3) 0.130 (4) 0.057 (2) 0.002 (3) 0.0116 (19) −0.011 (2) C18 0.065 (2) 0.052 (2) 0.065 (2) −0.0031 (17) 0.0409 (17) −0.0190 (17) ----- ------------- ------------- ------------- -------------- ------------- -------------- Geometric parameters (Å, °) {#tablewrapgeomlong} =========================== --------------------- -------------- ---------------------- ------------ N1---C6 1.352 (3) C8---H8B 0.9600 N1---C5 1.455 (3) C8---H8C 0.9600 N1---C1 1.457 (3) C9---H9A 0.9600 N2---C12 1.359 (3) C9---H9B 0.9600 N2---C13 1.463 (3) C9---H9C 0.9600 N2---C2 1.494 (3) C10---H10A 0.9600 N3---C3 1.280 (3) C10---H10B 0.9600 N3---O5 1.423 (2) C10---H10C 0.9600 O1---C6 1.342 (3) C11---H11A 0.9600 O1---C7 1.476 (3) C11---H11B 0.9600 O2---C6 1.210 (3) C11---H11C 0.9600 O3---C12 1.342 (3) C13---H13A 0.9600 O3---C14 1.459 (3) C13---H13B 0.9600 O4---C12 1.211 (3) C13---H13C 0.9600 O5---C18 1.423 (3) C14---C15 1.505 (5) C1---C2 1.548 (3) C14---C17 1.510 (4) C1---H1A 0.9700 C14---C16 1.513 (4) C1---H1B 0.9700 C15---H15A 0.9600 C2---C3 1.517 (3) C15---H15B 0.9600 C2---C11 1.525 (3) C15---H15C 0.9600 C3---C4 1.491 (3) C17---H17A 0.9600 C4---C5 1.525 (3) C17---H17B 0.9600 C4---H4A 0.9700 C17---H17C 0.9600 C4---H4B 0.9700 C16---H16A 0.9600 C5---H5A 0.9700 C16---H16B 0.9600 C5---H5B 0.9700 C16---H16C 0.9600 C7---C8 1.502 (4) C18---H18A 0.9600 C7---C10 1.502 (4) C18---H18B 0.9600 C7---C9 1.510 (4) C18---H18C 0.9600 C8---H8A 0.9600 C6---N1---C5 118.1 (2) C7---C9---H9C 109.5 C6---N1---C1 124.7 (2) H9A---C9---H9C 109.5 C5---N1---C1 115.2 (2) H9B---C9---H9C 109.5 C12---N2---C13 114.7 (2) C7---C10---H10A 109.5 C12---N2---C2 123.2 (2) C7---C10---H10B 109.5 C13---N2---C2 118.2 (2) H10A---C10---H10B 109.5 C3---N3---O5 110.9 (2) C7---C10---H10C 109.5 C6---O1---C7 121.1 (2) H10A---C10---H10C 109.5 C12---O3---C14 121.7 (2) H10B---C10---H10C 109.5 C18---O5---N3 107.7 (2) C2---C11---H11A 109.5 N1---C1---C2 112.7 (2) C2---C11---H11B 109.5 N1---C1---H1A 109.1 H11A---C11---H11B 109.5 C2---C1---H1A 109.1 C2---C11---H11C 109.5 N1---C1---H1B 109.1 H11A---C11---H11C 109.5 C2---C1---H1B 109.1 H11B---C11---H11C 109.5 H1A---C1---H1B 107.8 O4---C12---O3 123.7 (3) N2---C2---C3 111.2 (2) O4---C12---N2 124.5 (3) N2---C2---C11 110.1 (2) O3---C12---N2 111.8 (2) C3---C2---C11 113.9 (2) N2---C13---H13A 109.5 N2---C2---C1 109.1 (2) N2---C13---H13B 109.5 C3---C2---C1 103.36 (19) H13A---C13---H13B 109.5 C11---C2---C1 108.9 (2) N2---C13---H13C 109.5 N3---C3---C4 127.0 (2) H13A---C13---H13C 109.5 N3---C3---C2 116.7 (2) H13B---C13---H13C 109.5 C4---C3---C2 115.7 (2) O3---C14---C15 110.5 (3) C3---C4---C5 109.4 (2) O3---C14---C17 102.1 (2) C3---C4---H4A 109.8 C15---C14---C17 111.3 (3) C5---C4---H4A 109.8 O3---C14---C16 108.8 (3) C3---C4---H4B 109.8 C15---C14---C16 112.9 (3) C5---C4---H4B 109.8 C17---C14---C16 110.7 (3) H4A---C4---H4B 108.2 C14---C15---H15A 109.5 N1---C5---C4 110.9 (2) C14---C15---H15B 109.5 N1---C5---H5A 109.5 H15A---C15---H15B 109.5 C4---C5---H5A 109.5 C14---C15---H15C 109.5 N1---C5---H5B 109.5 H15A---C15---H15C 109.5 C4---C5---H5B 109.5 H15B---C15---H15C 109.5 H5A---C5---H5B 108.0 C14---C17---H17A 109.5 O2---C6---O1 124.7 (3) C14---C17---H17B 109.5 O2---C6---N1 123.8 (3) H17A---C17---H17B 109.5 O1---C6---N1 111.5 (3) C14---C17---H17C 109.5 O1---C7---C8 110.8 (2) H17A---C17---H17C 109.5 O1---C7---C10 101.8 (2) H17B---C17---H17C 109.5 C8---C7---C10 110.7 (3) C14---C16---H16A 109.5 O1---C7---C9 109.8 (2) C14---C16---H16B 109.5 C8---C7---C9 112.1 (3) H16A---C16---H16B 109.5 C10---C7---C9 111.3 (3) C14---C16---H16C 109.5 C7---C8---H8A 109.5 H16A---C16---H16C 109.5 C7---C8---H8B 109.5 H16B---C16---H16C 109.5 H8A---C8---H8B 109.5 O5---C18---H18A 109.5 C7---C8---H8C 109.5 O5---C18---H18B 109.5 H8A---C8---H8C 109.5 H18A---C18---H18B 109.5 H8B---C8---H8C 109.5 O5---C18---H18C 109.5 C7---C9---H9A 109.5 H18A---C18---H18C 109.5 C7---C9---H9B 109.5 H18B---C18---H18C 109.5 H9A---C9---H9B 109.5 C3---N3---O5---C18 −177.9 (2) C6---N1---C5---C4 −143.2 (2) C6---N1---C1---C2 139.0 (2) C1---N1---C5---C4 52.3 (3) C5---N1---C1---C2 −57.6 (3) C3---C4---C5---N1 −49.9 (3) C12---N2---C2---C3 48.3 (3) C7---O1---C6---O2 8.0 (4) C13---N2---C2---C3 −155.2 (2) C7---O1---C6---N1 −171.1 (2) C12---N2---C2---C11 −78.9 (3) C5---N1---C6---O2 10.3 (4) C13---N2---C2---C11 77.7 (3) C1---N1---C6---O2 173.3 (3) C12---N2---C2---C1 161.6 (2) C5---N1---C6---O1 −170.6 (2) C13---N2---C2---C1 −41.8 (3) C1---N1---C6---O1 −7.6 (4) N1---C1---C2---N2 −62.4 (3) C6---O1---C7---C8 −66.1 (3) N1---C1---C2---C3 56.0 (3) C6---O1---C7---C10 176.2 (3) N1---C1---C2---C11 177.4 (2) C6---O1---C7---C9 58.3 (3) O5---N3---C3---C4 −5.2 (3) C14---O3---C12---O4 4.0 (4) O5---N3---C3---C2 −176.31 (19) C14---O3---C12---N2 −175.0 (2) N2---C2---C3---N3 −129.7 (2) C13---N2---C12---O4 7.3 (4) C11---C2---C3---N3 −4.6 (3) C2---N2---C12---O4 164.6 (3) C1---C2---C3---N3 113.4 (2) C13---N2---C12---O3 −173.7 (2) N2---C2---C3---C4 58.2 (3) C2---N2---C12---O3 −16.5 (4) C11---C2---C3---C4 −176.7 (2) C12---O3---C14---C15 57.0 (4) C1---C2---C3---C4 −58.8 (3) C12---O3---C14---C17 175.5 (3) N3---C3---C4---C5 −113.4 (3) C12---O3---C14---C16 −67.5 (4) C2---C3---C4---C5 57.8 (3) --------------------- -------------- ---------------------- ------------

{ "pile_set_name": "PubMed Central" }

![A.C.H. Watson (1936-2018)](IJPS-51-260-g001){#d35e66} A CH Watson or Tony as he was commonly and affectionately called unobtrusively influenced the course of plastic surgery not only in his native Britain but also in India for over four decades. Born in 1936, he did his basic medical training in Edinburgh, Scotland, UK and graduated in 1960. His early surgical training started in The Royal Hospital for Sick Children or the 'sick kids' in Edinburgh and went on to the Edinburgh Royal Infirmary in General Surgery and allied branches until 1965. During this period, he qualified with an FRCS (Ed.) in 1964. Plastic surgery training was at the Bangaur General Hospital in the South East Scotland Regional unit under Mr. AB Wallace (of the rule of 9 fame), Mr. Buchan and Ms. Anne Sutherland from 1966 to 1970. The year 1971--1972 was spent as the Maytag-Macahill Fellow under Dr. Ralph Millard. I suppose this led to a lifelong love affair with cleft surgery. In 1972, he came back and became a consultant in the same unit at Bangaur, where he continued to work until his retirement in 1996. He contributed hugely to plastic surgery in general and cleft work in particular. A search on Google Scholar revealed over four pages of published papers and books. He was one of the earliest to adopt nasal endoscopy for diagnosis of velopharyngeal incompetence. Personally, I saw some beautiful cleft work while training under him in the 'sick kids', but what was more impressive was his dedication to the children, his patient manner in the out-patient clinic with fractious and to my eye, often spoilt kids and his very effective collaboration with the Speech therapist. Records were meticulously kept and proved a treasure trove for writing papers and drawing useful conclusions, which influenced care. The hallmark was, of course, the complete lack of fuss and pomp. He also contributed to breast cancer by being one of the early proponents of immediate reconstruction in Edinburgh. This work was later carried on by his colleague, James Watson. There are many other contributions, but they are too many to list here. Organisationally he played a huge role in the Royal College, as well as in the British Association of Plastic Surgeons (BAPS) as an office bearer, examiner and *via* the Specialist Advisory Committee which worked very hard to maintain and raise basic standards in teaching and training. He held multiple positions in BAPS, but the two most important were as Editor of the then BJPS and as President. His influence in India was through his numerous trainees, the author being one of them. Working with him taught us to be sincere, meticulous, balanced and most importantly well behaved. The last being the most challenging. Just as parents live on by spreading their genes to their offspring; I believe firmly that surgeons live through their students who have imbibed their skills and values. In that respect, Tony was a great success in the UK as well as in India and many other parts of the world. In 2010 he was honoured with the Honorary membership of the Association of Plastic Surgeons of India by the author who was then the President of APSI in recognition for his contribution to training Indian Surgeons. He also greatly helped the Indian Journal of Plastic Surgery after I became editor and guided me to make it a somewhat sensible scientific publication. As always this was done silently, with empathy and kindness. Over the last several years he has provided similar editorial input to 'The Short Notes in Plastic Surgery' a blog post run by Dr. RL Thatte with the active help of APSI. Tony lived life with grace and dignity and passed away in the same fashion. His passing is a great loss to plastic surgery in India for we have lost a revered Guru and a genuine friend of this country. Tony is survived by his wife Anne to whom he was married for five decades as well as three daughters and one son. Rest in peace Tony, we hope we can emulate some of your great qualities.

{ "pile_set_name": "PubMed Central" }

Related literature {#sec1} ==================== For the previous determination of the title compound at room temperature (Cambridge Structural Database refcode: JAMFEI), see: FitzRoy *et al.* (1989[@bb5]). For a description of the Cambridge Structural Database, see: Groom & Allen (2014[@bb6]). For previous studies investigating recovered molybdenum(VI) catalysts, see: Amarante *et al.* (2015[@bb1]); Lysenko *et al.* (2015[@bb7]). Experimental {#sec2} ============== Crystal data {#sec2.1} -------------- (C~5~H~9~N~2~)~4~\[Mo~8~O~26~\]·4H~2~O*M* *~r~* = 1644.15Triclinic,*a* = 10.1105 (9) Å*b* = 10.7469 (9) Å*c* = 11.9839 (10) Åα = 64.103 (3)°β = 84.272 (3)°γ = 75.826 (3)°*V* = 1135.67 (17) Å^3^*Z* = 1Mo *K*α radiationμ = 2.24 mm^−1^*T* = 180 K0.20 × 0.14 × 0.01 mm Data collection {#sec2.2} ----------------- Bruker D8 QUEST diffractometerAbsorption correction: multi-scan (*SADABS*; Bruker, 2001[@bb3]) *T* ~min~ = 0.663, *T* ~max~ = 0.74525625 measured reflections4141 independent reflections3175 reflections with *I* \> 2σ(*I*)*R* ~int~ = 0.042 Refinement {#sec2.3} ------------ *R*\[*F* ^2^ \> 2σ(*F* ^2^)\] = 0.026*wR*(*F* ^2^) = 0.056*S* = 1.064141 reflections320 parameters7 restraintsH atoms treated by a mixture of independent and constrained refinementΔρ~max~ = 0.70 e Å^−3^Δρ~min~ = −0.54 e Å^−3^ {#d5e542} Data collection: *APEX2* (Bruker, 2012[@bb4]); cell refinement: *SAINT* (Bruker, 2012[@bb4]); data reduction: *SAINT*; program(s) used to solve structure: *SHELXS97* (Sheldrick, 2008[@bb8]); program(s) used to refine structure: *SHELXL2014* (Sheldrick, 2015[@bb9]); molecular graphics: *DIAMOND* (Brandenburg, 1999[@bb2]); software used to prepare material for publication: *SHELXL2014*. Supplementary Material ====================== Crystal structure: contains datablock(s) I, New_Global_Publ_Block. DOI: [10.1107/S2056989015022823/hb7547sup1.cif](http://dx.doi.org/10.1107/S2056989015022823/hb7547sup1.cif) Structure factors: contains datablock(s) I. DOI: [10.1107/S2056989015022823/hb7547Isup2.hkl](http://dx.doi.org/10.1107/S2056989015022823/hb7547Isup2.hkl) ###### Click here for additional data file. . DOI: [10.1107/S2056989015022823/hb7547fig1.tif](http://dx.doi.org/10.1107/S2056989015022823/hb7547fig1.tif) Schematic representation of the molecular entities composing the asymmetric unit of the title compound. The β-octamolybdate anion has been completed by inversion symmetry for the sake of chemical accuracy. All non-hydrogen atoms are represented as displacement ellipsoids drawn at the 60% probability level and hydrogen atoms as small spheres with arbitrary radii. Non-hydrogen atoms belonging to the asymmetric unit have been labelled for clarity. Dashed green broken lines indicate N---H⋯O and O---H⋯O hydrogen-bonding interactions (see Table for geometrical details). ###### Click here for additional data file. $a$ (b) . DOI: [10.1107/S2056989015022823/hb7547fig2.tif](http://dx.doi.org/10.1107/S2056989015022823/hb7547fig2.tif) Crystal packing of the title compound viewed in perspective along the **(a)** \[100\] **(b)** \[010\] directions of the unit cell emphasising the supramolecular N---H⋯O and O---H⋯O hydrogen-bonding interactions (dashed green lines) interconnecting the three types of chemical species present in the crystal structure of the title compound. ###### Click here for additional data file. . DOI: [10.1107/S2056989015022823/hb7547fig3.tif](http://dx.doi.org/10.1107/S2056989015022823/hb7547fig3.tif) Mixed polyhedral (for the β-octamolybdate anion), ball-and-stick (for the 3,5-dimethylpyrazolium cations) and space filling (for the water molecules of crystallisation) schematic representation of the crystal packing of the title compound viewed in perspective along the \[100\] direction of the unit cell. The Figure illustrates well how the inorganic component of the crystal structure is embedded into an organic matrix, with the entrapped water molecules of crystallization acting as molecular fillers interacting with the hybrid network through hydrogen bonds. CCDC reference: [1439409](http://scripts.iucr.org/cgi-bin/cr.cgi?rm=csd&csdid=1439409) Additional supporting information: [crystallographic information](http://scripts.iucr.org/cgi-bin/sendsupfiles?hb7547&file=hb7547sup0.html&mime=text/html); [3D view](http://scripts.iucr.org/cgi-bin/sendcif?hb7547sup1&Qmime=cif); [checkCIF report](http://scripts.iucr.org/cgi-bin/paper?hb7547&checkcif=yes) Supporting information for this paper is available from the IUCr electronic archives (Reference: [HB7547](http://scripts.iucr.org/cgi-bin/sendsup?hb7547)). **Funding Sources and Entities** Fundação para a Ciência e a Tecnologia (FCT, Portugal), the European Union, QREN, FEDER through Programa Operacional Factores de Competitividade (COMPETE), CICECO--Aveiro Institute of Materials (Ref. FCT UID/CTM/50011/2013) financed by national funds through the FCT/MEC and when applicable co-financed by FEDER under the PT2020 Partnership Agreement.**Projects and Individual grants** We wish to thank FCT for funding the R&D project FCOMP-01--0124-FEDER-041282 (Ref. FCT EXPL/CTM-NAN/0013/2013), and also CICECO for specific funding towards the purchase of the single-crystal diffractometer. FCT is gratefully acknowledged for the post-doctoral research grant No. SFRH/BPD/97660/2013 (to TRA). S1. Context and introduction {#d1e22} ============================ Research efforts in the development and application of new hybrid molybdenum(vi) heterogeneous catalysts require on a daily basis the recovery of used materials and their characterization in the solid state to check for structural modifications of the employed compound (Amarante *et al.*, 2015; Lysenko *et al.*, 2015). Often, the use of drastic experimental conditions leads to the formation of secondary products, which crystallize in the medium as trace amounts of impurity compounds. It is, thus, imperative that most of these possible products are fully described in the solid state in the most accurate fashion. The title compound, (C~5~H~9~N~2~)~4~\[Mo~8~O~26~\]^.^4(H~2~O), was previously reported by FitzRoy *et al.* (1989). Besides the fact that the authors did not fully elucidate the hydrogen bonding network of this material (hydrogen atoms were placed geometrically) and that only the metallic centers could be refined anisotropically (from a room-temperature determination), the unit-cell parameters reported in the main text and in the abstract do not match (*viz.* the *c* axis length). A search and match at the Cambridge Structural Database (Groom *et al.*, 2014) also seems to ignore the presence of Mo(VI) metal centers. In this context, we decided to recollect the crystal structure of the title compound at low temperature to fully elucidate its finer structural details. S2. Structure description {#d1e82} ========================= The asymmetric unit of the title compound is composed of two 3,5-dimethylpyrazolium cations, (C~5~H~9~N~2~)^+^, one half of the β-octamolybdate anion, β-\[Mo~8~O~26~\]^4−^, and two water molecules of crystallization. Noteworthy, while one water molecule was fully located in its crystallographic position and even the associated hydrogen atoms found from difference Fourier maps, the other was found to be disordered over two distinct locations, O2W and O3W (Fig. 1). This feature was not disclosed in the previous structural determination by FitzRoy *et al.* (1989). The molecular geometrical parameters for the β-octamolybdate anion are typical, exhibiting the usual four families of Mo---O bonds: Mo---O*t* to terminal oxido groups \[bond distances in the 1.691 (3)--1.715 (3) Å range\]; Mo---O*b* to µ~2~-bridging oxido groups \[bond distances in the 1.754 (3)--2.268 (3) Å range\]; Mo---O*c* to µ~3~-bridging oxido groups \[bond distances in the 1.953 (3)--2.351 (3) Å range\]; Mo---O*c* to µ~5~-bridging oxido groups \[bond distances in the 2.146 (3)--2.435 (3) Å range\]. The four crystallographically independent Mo(VI) metal centers are thus hexacoordinated in a typical {MoO~6~} fashion resembling highly distorted octahedra: while this *trans* internal O---Mo---O octahedral angles were found in the 145.29 (11)--173.17 (12)° range, the *cis* angles refined instead in the 70.01 (9)--105.85 (14)° interval. We note that this wide dispersion for the internal octahedral angles is a notable and well known consequence of the marked *trans* effect created by the terminal oxido groups which displace the metal centers from the center of the octahedra. The β-octamolybdate anion, located in the center of the unit cell, interacts with the remaining chemical species through a series of both electrostatic interactions and hydrogen bonds (Fig. 2). One crystallographically independent 3,5-dimethylpyrazolium cation donates both the hydrogen atoms bound to nitrogen to form two strong \[D···A distances of 2.801 (4) and 2.869 (5) Å\] and relatively directional \[\<(DHA) angles of 120 and 152°\] charged N---H···O interactions with the polyoxoanion. The other cation has, however, a completely distinct behaviour: the same hydrogen atoms are instead donated in similar (strong and highly directional) interactions with neighbouring water molecules: while the D···A distances are 2.689 (7) and 2.776 (5) Å, the \<(DHA) interaction angles are close to linearity, being 171 and 164°. The water molecules are instead interacting with the β-octamolybdate anion as depicted in both Figs. 1 and 2. Indeed, as depicted in Fig. 3 the water molecules play a decisive role in the overall crystal packing, acting as molecular fillers to effectively occupy the available space left from the arrangement of inorganic anions and organic cations. S3. Synthesis and crystallization {#d1e175} ================================= All chemicals were purchased from commercial sources and used as received without additional purification steps. A Teflon-lined stainless steel vessel was charged with a reaction mixture composed of MoO~3~ (0.34 g, 2.43 mmol), 3,5-dimethylpyrazole (0.11 g, 1.21 mmol) and water (**ca** 25 ml) and heated in an oven at 160 °C for 26 h. The resultant blueish solid was filtered from the aqueous mother liquor and washed with an excess of water and (4×10 ml) diethyl ether, dried at ambient temperature and characterized in the solid state. Colourless plates of the title compound were directly harvested from the walls of the Teflon vessel. Selected FT---IR (KBr, cm^−1^): ñ = 948 (*vs*), 910 (*vs*), 843 (*s*), 733 (*s*), 714 (*s*), 663 (*s*). S4. Refinement details {#d1e222} ====================== Hydrogen atoms bound to carbon atoms were placed at idealized positions with C---H = 0.95 or 0.98 Å (for the aromatic and methyl groups, respectively), and included in the final structural model in riding-motion approximation with the isotropic thermal displacement parameters fixed at 1.2 or 1.5×*U*~eq~, respectively, of the carbon atom to which they are attached. Hydrogen atoms associated with nitrogen atoms have been directly located from difference Fourier maps and were included in the model with the N---H distances restrained to 0.95 (1) Å in order to ensure a chemically reasonable environment for these moieties. These hydrogen atoms were modelled with the isotropic thermal displacement parameters fixed at 1.5×*U*~eq~(N). A total of two water molecules of crystallization were directly located from difference Fourier maps. Though O1W was included in the final structural model by assuming full site occupancy and a typical anisotropic displacement behaviour, the second molecule was found to be disordered over two close crystallographic positions: O2W and O3W. These species were included in the structural model with linked site occupancy \[which ultimately refined to 0.65 (2) and 0.35 (2), respectively\] and by assuming an independent isotropic displacement behaviour. For O1W the two hydrogen atoms were markedly visible in difference Fourier maps and were included in the final model with the O---H and H···H distances restrained to 0.95 (1) and 1.55 (1) Å, respectively, in order to ensure a chemically reasonable geometry for this molecule. These hydrogen atoms were modelled with the isotropic thermal displacement parameters fixed at 1.5×*U*~eq~(O1W). Figures ======= ![Schematic representation of the molecular entities composing the asymmetric unit of the title compound. The β-octamolybdate anion has been completed by inversion symmetry for the sake of chemical accuracy. All non-hydrogen atoms are represented as displacement ellipsoids drawn at the 60% probability level and hydrogen atoms as small spheres with arbitrary radii. Non-hydrogen atoms belonging to the asymmetric unit have been labelled for clarity. Dashed green broken lines indicate N---H···O and O---H···O hydrogen-bonding interactions (see Table for geometrical details).](e-71-0m244-fig1){#Fap1} ![Crystal packing of the title compound viewed in perspective along the (a) \[100\] (b) \[010\] directions of the unit cell emphasizing the supramolecular N---H···O and O---H···O hydrogen-bonding interactions (dashed green lines) interconnecting the three types of chemical species present in the crystal structure of the title compound.](e-71-0m244-fig2){#Fap2} ![Mixed polyhedral (for the β-octamolybdate anion), ball-and-stick (for the 3,5-dimethylpyrazolium cations) and space filling (for the water molecules of crystallization) schematic representation of the crystal packing of the title compound viewed in perspective along the \[100\] direction of the unit cell. The Figure illustrates well how the inorganic component of the crystal structure is embedded into an organic matrix, with the entrapped water molecules of crystallization acting as molecular fillers interacting with the hybrid network through hydrogen bonds.](e-71-0m244-fig3){#Fap3} Crystal data {#tablewrapcrystaldatalong} ============ ---------------------------------------- --------------------------------------- (C~5~H~9~N~2~)~4~\[Mo~8~O~26~\]·4H~2~O *Z* = 1 *M~r~* = 1644.15 *F*(000) = 796 Triclinic, *P*1 *D*~x~ = 2.404 Mg m^−3^ *a* = 10.1105 (9) Å Mo *K*α radiation, λ = 0.71073 Å *b* = 10.7469 (9) Å Cell parameters from 9938 reflections *c* = 11.9839 (10) Å θ = 2.6--25.4° α = 64.103 (3)° µ = 2.24 mm^−1^ β = 84.272 (3)° *T* = 180 K γ = 75.826 (3)° Plate, colourless *V* = 1135.67 (17) Å^3^ 0.20 × 0.14 × 0.01 mm ---------------------------------------- --------------------------------------- Data collection {#tablewrapdatacollectionlong} =============== ------------------------------------------------------------ -------------------------------------- Bruker D8 QUEST diffractometer 4141 independent reflections Radiation source: Sealed tube 3175 reflections with *I* \> 2σ(*I*) Multi-layer X-ray mirror monochromator *R*~int~ = 0.042 Detector resolution: 10.4167 pixels mm^-1^ θ~max~ = 25.4°, θ~min~ = 3.7° ω / φ scans *h* = −12→12 Absorption correction: multi-scan (*SADABS*; Bruker, 2001) *k* = −12→12 *T*~min~ = 0.663, *T*~max~ = 0.745 *l* = −14→14 25625 measured reflections ------------------------------------------------------------ -------------------------------------- Refinement {#tablewraprefinementdatalong} ========== ------------------------------------- ------------------------------------------------------------------------------------------------- Refinement on *F*^2^ 7 restraints Least-squares matrix: full Hydrogen site location: mixed *R*\[*F*^2^ \> 2σ(*F*^2^)\] = 0.026 H atoms treated by a mixture of independent and constrained refinement *wR*(*F*^2^) = 0.056 *w* = 1/\[σ^2^(*F*~o~^2^) + (0.0188*P*)^2^ + 2.2481*P*\] where *P* = (*F*~o~^2^ + 2*F*~c~^2^)/3 *S* = 1.06 (Δ/σ)~max~ = 0.001 4141 reflections Δρ~max~ = 0.70 e Å^−3^ 320 parameters Δρ~min~ = −0.54 e Å^−3^ ------------------------------------- ------------------------------------------------------------------------------------------------- Special details {#specialdetails} =============== ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Geometry. All e.s.d.\'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.\'s are taken into account individually in the estimation of e.s.d.\'s in distances, angles and torsion angles; correlations between e.s.d.\'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.\'s is used for estimating e.s.d.\'s involving l.s. planes. ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å^2^) {#tablewrapcoords} ================================================================================================== ------ ------------- ------------- ------------- -------------------- ------------ *x* *y* *z* *U*~iso~\*/*U*~eq~ Occ. (\<1) Mo1 0.44983 (3) 0.44979 (4) 0.75253 (3) 0.01430 (9) Mo2 0.45658 (3) 0.68126 (4) 0.47103 (3) 0.01229 (9) Mo3 0.23903 (3) 0.56541 (4) 0.38528 (3) 0.01415 (9) Mo4 0.23047 (3) 0.32798 (4) 0.67012 (3) 0.01511 (9) O1 0.3753 (3) 0.4944 (3) 0.5589 (2) 0.0142 (6) O2 0.3635 (3) 0.3005 (3) 0.7899 (2) 0.0177 (6) O3 0.5489 (3) 0.3881 (3) 0.8808 (3) 0.0235 (7) O4 0.5539 (3) 0.5839 (3) 0.6298 (2) 0.0145 (6) O5 0.3166 (3) 0.5748 (3) 0.7683 (3) 0.0208 (7) O6 0.5740 (3) 0.7845 (3) 0.3880 (2) 0.0168 (6) O7 0.3838 (3) 0.6770 (3) 0.3281 (2) 0.0153 (6) O8 0.3261 (3) 0.7924 (3) 0.5040 (3) 0.0190 (6) O9 0.1894 (3) 0.5797 (3) 0.2484 (3) 0.0232 (7) O10 0.1936 (3) 0.3896 (3) 0.4970 (2) 0.0160 (6) O11 0.1191 (3) 0.6842 (3) 0.4197 (3) 0.0220 (7) O12 0.1720 (3) 0.1773 (3) 0.7422 (3) 0.0255 (7) O13 0.1090 (3) 0.4568 (3) 0.6928 (3) 0.0213 (7) N1 0.7819 (4) 0.8121 (5) 0.9458 (3) 0.0294 (9) H1 0.865 (3) 0.747 (4) 0.979 (4) 0.044\* N2 0.6839 (4) 0.7657 (4) 0.9163 (4) 0.0281 (9) H2 0.692 (5) 0.6692 (19) 0.934 (5) 0.042\* N3 −0.0378 (4) 0.2498 (4) 0.3524 (4) 0.0295 (9) H3 −0.116 (3) 0.314 (4) 0.362 (5) 0.044\* N4 0.0827 (4) 0.2507 (4) 0.3913 (4) 0.0289 (9) H4 0.090 (5) 0.314 (4) 0.425 (4) 0.043\* C1 0.8229 (6) 1.0327 (6) 0.9381 (5) 0.0491 (15) H1A 0.8535 0.9870 1.0247 0.074\* H1B 0.7696 1.1293 0.9183 0.074\* H1C 0.9023 1.0365 0.8837 0.074\* C2 0.7364 (5) 0.9494 (5) 0.9191 (4) 0.0319 (11) C3 0.6048 (5) 0.9913 (5) 0.8718 (4) 0.0316 (11) H3A 0.5471 1.0837 0.8443 0.038\* C4 0.5743 (5) 0.8733 (5) 0.8726 (4) 0.0312 (11) C5 0.4486 (5) 0.8533 (6) 0.8341 (5) 0.0420 (14) H5A 0.4728 0.7778 0.8058 0.063\* H5B 0.4035 0.9420 0.7663 0.063\* H5C 0.3866 0.8271 0.9047 0.063\* C6 −0.1363 (5) 0.1101 (5) 0.2791 (5) 0.0329 (12) H6A −0.2222 0.1413 0.3150 0.049\* H6B −0.1209 0.0082 0.3021 0.049\* H6C −0.1412 0.1624 0.1885 0.049\* C7 −0.0218 (4) 0.1381 (5) 0.3270 (4) 0.0208 (10) C8 0.1130 (4) 0.0663 (5) 0.3502 (4) 0.0225 (10) H8 0.1545 −0.0174 0.3399 0.027\* C9 0.1768 (4) 0.1398 (5) 0.3917 (4) 0.0208 (10) C10 0.3202 (4) 0.1135 (5) 0.4290 (5) 0.0303 (11) H10A 0.3497 0.2034 0.3936 0.045\* H10B 0.3787 0.0479 0.3983 0.045\* H10C 0.3268 0.0717 0.5196 0.045\* O1W 0.7299 (3) 0.4764 (4) 0.9840 (3) 0.0276 (7) H1X 0.711 (5) 0.439 (5) 1.0698 (12) 0.041\* H1Y 0.676 (4) 0.449 (5) 0.943 (3) 0.041\* O2W 0.9897 (6) 0.5952 (10) 0.0682 (5) 0.031 (2)\* 0.65 (2) O3W 1.0101 (11) 0.6562 (19) 0.0507 (10) 0.033 (4)\* 0.35 (2) ------ ------------- ------------- ------------- -------------------- ------------ Atomic displacement parameters (Å^2^) {#tablewrapadps} ===================================== ----- -------------- -------------- -------------- --------------- --------------- --------------- *U*^11^ *U*^22^ *U*^33^ *U*^12^ *U*^13^ *U*^23^ Mo1 0.01445 (18) 0.0179 (2) 0.01272 (18) −0.00473 (15) 0.00061 (14) −0.00798 (16) Mo2 0.01153 (18) 0.01158 (19) 0.01458 (18) −0.00190 (14) −0.00116 (14) −0.00643 (15) Mo3 0.01088 (18) 0.0166 (2) 0.01578 (19) −0.00309 (15) −0.00217 (14) −0.00718 (16) Mo4 0.01333 (18) 0.0163 (2) 0.01719 (19) −0.00532 (15) 0.00086 (14) −0.00759 (16) O1 0.0124 (14) 0.0157 (15) 0.0150 (14) −0.0038 (12) 0.0002 (11) −0.0069 (12) O2 0.0179 (15) 0.0185 (16) 0.0157 (14) −0.0065 (12) 0.0021 (12) −0.0055 (13) O3 0.0232 (16) 0.0307 (19) 0.0185 (16) −0.0101 (14) −0.0013 (13) −0.0100 (14) O4 0.0142 (14) 0.0157 (15) 0.0179 (15) −0.0041 (12) −0.0013 (11) −0.0104 (13) O5 0.0194 (15) 0.0228 (17) 0.0223 (16) −0.0053 (13) 0.0038 (13) −0.0122 (14) O6 0.0184 (15) 0.0146 (15) 0.0171 (15) −0.0040 (12) −0.0022 (12) −0.0058 (13) O7 0.0128 (14) 0.0170 (16) 0.0150 (14) −0.0024 (12) −0.0010 (11) −0.0060 (12) O8 0.0186 (15) 0.0155 (16) 0.0227 (16) −0.0007 (12) −0.0028 (12) −0.0091 (13) O9 0.0201 (16) 0.0316 (19) 0.0194 (16) −0.0067 (14) −0.0042 (13) −0.0110 (14) O10 0.0146 (14) 0.0169 (16) 0.0199 (15) −0.0045 (12) −0.0005 (12) −0.0101 (13) O11 0.0159 (15) 0.0214 (17) 0.0273 (17) −0.0002 (13) −0.0014 (13) −0.0110 (14) O12 0.0243 (16) 0.0219 (17) 0.0313 (18) −0.0108 (14) 0.0025 (14) −0.0099 (15) O13 0.0171 (15) 0.0227 (17) 0.0232 (16) −0.0038 (13) 0.0018 (13) −0.0098 (14) N1 0.023 (2) 0.038 (3) 0.022 (2) −0.0038 (19) −0.0028 (17) −0.0087 (19) N2 0.027 (2) 0.029 (2) 0.024 (2) −0.0071 (19) −0.0013 (17) −0.0067 (19) N3 0.020 (2) 0.029 (2) 0.046 (3) 0.0003 (17) 0.0004 (18) −0.026 (2) N4 0.026 (2) 0.031 (2) 0.041 (2) −0.0094 (18) 0.0027 (18) −0.026 (2) C1 0.061 (4) 0.053 (4) 0.039 (3) −0.024 (3) −0.005 (3) −0.017 (3) C2 0.036 (3) 0.035 (3) 0.020 (2) −0.010 (2) 0.004 (2) −0.007 (2) C3 0.032 (3) 0.025 (3) 0.025 (3) −0.002 (2) −0.001 (2) −0.001 (2) C4 0.027 (3) 0.033 (3) 0.022 (2) −0.004 (2) −0.002 (2) −0.001 (2) C5 0.028 (3) 0.046 (4) 0.036 (3) −0.011 (2) −0.006 (2) 0.000 (3) C6 0.024 (3) 0.035 (3) 0.050 (3) −0.003 (2) −0.008 (2) −0.027 (3) C7 0.021 (2) 0.020 (2) 0.026 (2) −0.0049 (19) 0.0010 (18) −0.014 (2) C8 0.023 (2) 0.024 (3) 0.026 (2) −0.0022 (19) −0.0022 (19) −0.017 (2) C9 0.022 (2) 0.024 (2) 0.019 (2) −0.0054 (19) 0.0015 (18) −0.011 (2) C10 0.025 (2) 0.030 (3) 0.043 (3) −0.006 (2) −0.005 (2) −0.021 (2) O1W 0.0298 (18) 0.036 (2) 0.0240 (17) −0.0078 (15) −0.0003 (14) −0.0192 (16) ----- -------------- -------------- -------------- --------------- --------------- --------------- Geometric parameters (Å, º) {#tablewrapgeomlong} =========================== --------------------------- ------------- -------------------- ------------- Mo1---O3 1.702 (3) N2---C4 1.340 (6) Mo1---O5 1.715 (3) N2---H2 0.946 (10) Mo1---O2 1.875 (3) N3---C7 1.332 (5) Mo1---O4 1.982 (3) N3---N4 1.351 (5) Mo1---O7^i^ 2.307 (3) N3---H3 0.947 (10) Mo1---O1 2.320 (3) N4---C9 1.328 (6) Mo1---Mo2 3.2062 (6) N4---H4 0.947 (10) Mo2---O8 1.691 (3) C1---C2 1.488 (7) Mo2---O6 1.754 (3) C1---H1A 0.9800 Mo2---O7 1.953 (3) C1---H1B 0.9800 Mo2---O4 1.953 (3) C1---H1C 0.9800 Mo2---O1 2.146 (3) C2---C3 1.389 (7) Mo2---O1^i^ 2.341 (3) C3---C4 1.373 (7) Mo2---Mo3 3.2178 (5) C3---H3A 0.9500 Mo3---O9 1.695 (3) C4---C5 1.488 (7) Mo3---O11 1.695 (3) C5---H5A 0.9800 Mo3---O10 1.915 (3) C5---H5B 0.9800 Mo3---O7 1.998 (3) C5---H5C 0.9800 Mo3---O1 2.341 (3) C6---C7 1.488 (6) Mo3---O4^i^ 2.351 (3) C6---H6A 0.9800 Mo4---O12 1.692 (3) C6---H6B 0.9800 Mo4---O13 1.711 (3) C6---H6C 0.9800 Mo4---O10 1.928 (3) C7---C8 1.379 (6) Mo4---O2 1.943 (3) C8---C9 1.394 (6) Mo4---O6^i^ 2.268 (3) C8---H8 0.9500 Mo4---O1 2.435 (3) C9---C10 1.483 (6) O1---Mo2^i^ 2.341 (3) C10---H10A 0.9800 O4---Mo3^i^ 2.351 (3) C10---H10B 0.9800 O6---Mo4^i^ 2.268 (3) C10---H10C 0.9800 O7---Mo1^i^ 2.307 (3) O1W---H1X 0.946 (10) N1---C2 1.332 (6) O1W---H1Y 0.944 (10) N1---N2 1.350 (5) O2W---O3W 0.674 (13) N1---H1 0.943 (10) O3---Mo1---O5 104.71 (14) O12---Mo4---O1 162.52 (12) O3---Mo1---O2 102.12 (13) O13---Mo4---O1 91.62 (11) O5---Mo1---O2 100.72 (13) O10---Mo4---O1 74.97 (10) O3---Mo1---O4 99.75 (12) O2---Mo4---O1 73.62 (10) O5---Mo1---O4 97.01 (13) O6^i^---Mo4---O1 70.01 (9) O2---Mo1---O4 147.18 (11) Mo2---O1---Mo1 91.68 (9) O3---Mo1---O7^i^ 89.69 (12) Mo2---O1---Mo2^i^ 104.01 (10) O5---Mo1---O7^i^ 163.56 (12) Mo1---O1---Mo2^i^ 97.04 (9) O2---Mo1---O7^i^ 83.57 (11) Mo2---O1---Mo3 91.54 (10) O4---Mo1---O7^i^ 72.34 (10) Mo1---O1---Mo3 163.07 (12) O3---Mo1---O1 161.88 (12) Mo2^i^---O1---Mo3 98.29 (9) O5---Mo1---O1 93.03 (11) Mo2---O1---Mo4 163.99 (13) O2---Mo1---O1 77.65 (10) Mo1---O1---Mo4 86.35 (9) O4---Mo1---O1 74.01 (10) Mo2^i^---O1---Mo4 91.99 (9) O7^i^---Mo1---O1 72.23 (9) Mo3---O1---Mo4 85.98 (8) O3---Mo1---Mo2 134.82 (10) Mo1---O2---Mo4 116.94 (14) O5---Mo1---Mo2 84.85 (10) Mo2---O4---Mo1 109.13 (12) O2---Mo1---Mo2 119.63 (8) Mo2---O4---Mo3^i^ 110.39 (11) O4---Mo1---Mo2 35.14 (7) Mo1---O4---Mo3^i^ 103.30 (11) O7^i^---Mo1---Mo2 79.31 (7) Mo2---O6---Mo4^i^ 116.79 (14) O1---Mo1---Mo2 41.99 (7) Mo2---O7---Mo3 109.07 (13) O8---Mo2---O6 105.54 (13) Mo2---O7---Mo1^i^ 109.84 (11) O8---Mo2---O7 101.63 (12) Mo3---O7---Mo1^i^ 104.33 (12) O6---Mo2---O7 96.27 (12) Mo3---O10---Mo4 115.92 (13) O8---Mo2---O4 99.93 (12) C2---N1---N2 108.8 (4) O6---Mo2---O4 97.02 (11) C2---N1---H1 133 (3) O7---Mo2---O4 150.55 (11) N2---N1---H1 118 (3) O8---Mo2---O1 97.27 (12) C4---N2---N1 108.9 (4) O6---Mo2---O1 157.19 (11) C4---N2---H2 128 (3) O7---Mo2---O1 78.87 (10) N1---N2---H2 123 (3) O4---Mo2---O1 78.71 (10) C7---N3---N4 109.1 (4) O8---Mo2---O1^i^ 173.17 (12) C7---N3---H3 133 (3) O6---Mo2---O1^i^ 81.21 (11) N4---N3---H3 117 (3) O7---Mo2---O1^i^ 78.31 (10) C9---N4---N3 109.5 (4) O4---Mo2---O1^i^ 77.92 (10) C9---N4---H4 128 (3) O1---Mo2---O1^i^ 75.99 (10) N3---N4---H4 122 (3) O8---Mo2---Mo1 88.75 (10) C2---C1---H1A 109.5 O6---Mo2---Mo1 132.75 (9) C2---C1---H1B 109.5 O7---Mo2---Mo1 125.17 (8) H1A---C1---H1B 109.5 O4---Mo2---Mo1 35.73 (8) C2---C1---H1C 109.5 O1---Mo2---Mo1 46.33 (7) H1A---C1---H1C 109.5 O1^i^---Mo2---Mo1 85.79 (7) H1B---C1---H1C 109.5 O8---Mo2---Mo3 89.31 (9) N1---C2---C3 107.8 (4) O6---Mo2---Mo3 132.20 (9) N1---C2---C1 121.8 (5) O7---Mo2---Mo3 35.93 (8) C3---C2---C1 130.4 (5) O4---Mo2---Mo3 125.36 (8) C4---C3---C2 106.6 (4) O1---Mo2---Mo3 46.65 (7) C4---C3---H3A 126.7 O1^i^---Mo2---Mo3 86.76 (6) C2---C3---H3A 126.7 Mo1---Mo2---Mo3 91.724 (14) N2---C4---C3 107.8 (4) O9---Mo3---O11 105.85 (14) N2---C4---C5 121.2 (5) O9---Mo3---O10 100.45 (13) C3---C4---C5 130.9 (5) O11---Mo3---O10 101.76 (13) C4---C5---H5A 109.5 O9---Mo3---O7 100.71 (12) C4---C5---H5B 109.5 O11---Mo3---O7 97.58 (12) H5A---C5---H5B 109.5 O10---Mo3---O7 146.06 (11) C4---C5---H5C 109.5 O9---Mo3---O1 159.60 (12) H5A---C5---H5C 109.5 O11---Mo3---O1 94.36 (12) H5B---C5---H5C 109.5 O10---Mo3---O1 77.54 (10) C7---C6---H6A 109.5 O7---Mo3---O1 73.40 (10) C7---C6---H6B 109.5 O9---Mo3---O4^i^ 88.84 (12) H6A---C6---H6B 109.5 O11---Mo3---O4^i^ 163.17 (11) C7---C6---H6C 109.5 O10---Mo3---O4^i^ 83.12 (10) H6A---C6---H6C 109.5 O7---Mo3---O4^i^ 71.11 (10) H6B---C6---H6C 109.5 O1---Mo3---O4^i^ 70.77 (9) N3---C7---C8 107.4 (4) O9---Mo3---Mo2 135.69 (10) N3---C7---C6 121.8 (4) O11---Mo3---Mo2 85.49 (10) C8---C7---C6 130.8 (4) O10---Mo3---Mo2 119.34 (8) C7---C8---C9 107.0 (4) O7---Mo3---Mo2 35.00 (7) C7---C8---H8 126.5 O1---Mo3---Mo2 41.81 (7) C9---C8---H8 126.5 O4^i^---Mo3---Mo2 78.16 (6) N4---C9---C8 106.9 (4) O12---Mo4---O13 105.75 (14) N4---C9---C10 121.9 (4) O12---Mo4---O10 103.91 (13) C8---C9---C10 131.2 (4) O13---Mo4---O10 97.77 (13) C9---C10---H10A 109.5 O12---Mo4---O2 101.63 (13) C9---C10---H10B 109.5 O13---Mo4---O2 97.49 (13) H10A---C10---H10B 109.5 O10---Mo4---O2 145.29 (11) C9---C10---H10C 109.5 O12---Mo4---O6^i^ 92.59 (12) H10A---C10---H10C 109.5 O13---Mo4---O6^i^ 161.61 (12) H10B---C10---H10C 109.5 O10---Mo4---O6^i^ 78.80 (10) H1X---O1W---H1Y 109.9 (16) O2---Mo4---O6^i^ 76.95 (11) O3---Mo1---O2---Mo4 177.15 (15) N2---N1---C2---C1 −178.3 (4) O5---Mo1---O2---Mo4 69.40 (17) N1---C2---C3---C4 0.4 (5) O4---Mo1---O2---Mo4 −52.1 (3) C1---C2---C3---C4 179.0 (5) O7^i^---Mo1---O2---Mo4 −94.55 (15) N1---N2---C4---C3 1.4 (5) O1---Mo1---O2---Mo4 −21.39 (13) N1---N2---C4---C5 −179.1 (4) Mo2---Mo1---O2---Mo4 −20.75 (18) C2---C3---C4---N2 −1.1 (5) O8---Mo2---O6---Mo4^i^ 178.38 (13) C2---C3---C4---C5 179.4 (5) O7---Mo2---O6---Mo4^i^ −77.67 (14) N4---N3---C7---C8 −0.1 (5) O4---Mo2---O6---Mo4^i^ 75.99 (14) N4---N3---C7---C6 −178.6 (4) O1---Mo2---O6---Mo4^i^ −1.5 (4) N3---C7---C8---C9 0.4 (5) O1^i^---Mo2---O6---Mo4^i^ −0.57 (12) C6---C7---C8---C9 178.7 (5) Mo1---Mo2---O6---Mo4^i^ 75.31 (16) N3---N4---C9---C8 0.5 (5) Mo3---Mo2---O6---Mo4^i^ −78.01 (15) N3---N4---C9---C10 179.3 (4) C2---N1---N2---C4 −1.1 (5) C7---C8---C9---N4 −0.6 (5) C7---N3---N4---C9 −0.2 (5) C7---C8---C9---C10 −179.2 (5) N2---N1---C2---C3 0.4 (5) --------------------------- ------------- -------------------- ------------- Symmetry code: (i) −*x*+1, −*y*+1, −*z*+1. Hydrogen-bond geometry (Å, º) {#tablewraphbondslong} ============================= ------------------------ --------- --------- ----------- --------------- *D*---H···*A* *D*---H H···*A* *D*···*A* *D*---H···*A* N1---H1···O2*W*^ii^ 0.94 1.77 2.689 (7) 164 N2---H2···O1*W* 0.95 1.84 2.776 (5) 171 N3---H3···O13^iii^ 0.95 2.28 2.869 (5) 120 N4---H4···O10 0.95 1.93 2.801 (4) 152 O1*W*---H1*X*···O5^iv^ 0.95 1.88 2.785 (4) 160 O1*W*---H1*Y*···O3 0.94 1.91 2.848 (4) 172 ------------------------ --------- --------- ----------- --------------- Symmetry codes: (ii) *x*, *y*, *z*+1; (iii) −*x*, −*y*+1, −*z*+1; (iv) −*x*+1, −*y*+1, −*z*+2. ###### Hydrogen-bond geometry (Å, °) *D*---H⋯*A* *D*---H H⋯*A* *D*⋯*A* *D*---H⋯*A* ----------------------- --------- ------- ----------- ------------- N1---H1⋯O2*W* ^i^ 0.94 1.77 2.689 (7) 164 N2---H2⋯O1*W* 0.95 1.84 2.776 (5) 171 N3---H3⋯O13^ii^ 0.95 2.28 2.869 (5) 120 N4---H4⋯O10 0.95 1.93 2.801 (4) 152 O1*W*---H1*X*⋯O5^iii^ 0.95 1.88 2.785 (4) 160 O1*W*---H1*Y*⋯O3 0.94 1.91 2.848 (4) 172 Symmetry codes: (i) ; (ii) ; (iii) .

{ "pile_set_name": "PubMed Central" }

Background ========== MicroRNAs (miRNAs) are short (around 20-22 nucleotides) RNAs that post-transcriptionally regulate gene expression by base-pairing with complementary sequences in the 3\' untranslated regions (UTRs) of protein-coding transcripts and directing translational repression or transcript degradation \[[@B1]-[@B5]\]. There are currently 326 human miRNAs listed in the miRNA registry version 7.1 \[[@B6]\], but the total number of miRNAs encoded in the human genome may be nearer 1,000 \[[@B7],[@B8]\]. The function of most miRNAs is unknown, but many are clearly involved in regulating differentiation \[[@B9]\] and development \[[@B10]\]. It is estimated that up to 30% of human genes may be miRNA targets \[[@B11],[@B12]\]. miRNAs are transcribed by RNA polymerase II into long primary miRNA (pri-miRNA) transcripts which are capped and polyadenylated \[[@B13],[@B14]\]. Genomic analyses indicate that many miRNAs overlap known protein coding genes or non-coding RNAs \[[@B15]\], and that many are in evolutionarily conserved clusters with other miRNAs \[[@B16]\]. Furthermore, intronic miRNAs share expression patterns with adjacent miRNAs and the host gene mRNA indicating that they are coordinately coexpressed \[[@B17]\]. Pri-miRNAs contain a short double-stranded RNA (dsRNA) stem-loop formed between the miRNA sequence and its adjacent complementary sequence. In the nucleus, the ribonuclease III-like enzyme Drosha cleaves at the base of this stem-loop to liberate a miRNA precursor (pre-miRNA) as a 60-70-nucleotide RNA hairpin \[[@B18]\]. The pre-miRNA hairpin is exported to the cytoplasm by exportin-5 \[[@B19]-[@B21]\] where it is further processed into a short dsRNA molecule by a second ribonuclease III-like enzyme, Dicer \[[@B22]-[@B24]\]. A single strand of this short dsRNA, the mature miRNA, is incorporated into a ribonucleoprotein complex. This complex directs transcript cleavage or translational repression depending on the degree of complementarity between the miRNA and its target site. RNA editing is the site-specific modification of an RNA sequence to yield a product differing from that encoded by the DNA template. Most RNA editing in human cells is adenosine to inosine (A-to-I) RNA editing which involves the conversion of A-to-I in dsRNA \[[@B25],[@B26]\]. A-to-I RNA editing is catalyzed by the adenosine deaminases acting on RNA (ADARs). The majority of A-to-I RNA-editing sites are in dsRNA structures formed between inverted repeat sequences in intronic or intergenic RNAs \[[@B25],[@B27]-[@B30]\]. Therefore, the double-stranded precursors of miRNAs may be substrates for A-to-I editing. Indeed, it has recently been shown that the pri-miRNA transcript of human miRNA miR-22 is subject to A-to-I RNA editing in a number of human and mouse tissues \[[@B31]\]. Although the extent of A-to-I editing was low (less than 5% across all adenosines analyzed), targeted adenosines were at positions predicted to influence the biogenesis and function of miR-22. This raises the possibility that RNA editing may be generally important in miRNA gene function \[[@B31]\]. In this study we have systematically investigated the presence of RNA editing in miRNAs. Results ======= To search for RNA-editing sites in human miRNAs, PCR product sequencing was performed from matched total cDNA and genomic DNA isolated from adult human brain, heart, liver, lung, ovary, placenta, skeletal muscle, small intestine, spleen and testis. Primers were designed to amplify pri-miRNA sequences flanking all 231 human miRNAs in miRBase \[[@B6]\]. Of these, 99 miRNA containing sequences were successfully sequenced in both directions and from duplicate PCR products from total cDNA of at least one tissue. Total cDNA sequence traces were compared with genomic DNA sequence traces from the same individual, and A-to-I editing was identified as an A in the genomic DNA sequence compared with a novel G peak at the equivalent position in the total cDNA sequence. In total, 12 of the 99 miRNA-containing sequences (13%) were subject to A-to-I RNA editing according to A-to-G differences between matched genomic DNA and total cDNA sequence traces from at least one tissue (Figure [1](#F1){ref-type="fig"}). These sequences were next oriented with respect to the strand of transcription of the miRNAs. In six cases the A-to-G changes were in the same orientation as the miRNA, and overlap the stem-loop structure of the miRNA, consistent with RNA editing of the pri-miRNA precursor transcript. In an additional case, A-to-I editing was observed in a novel stem-loop structure in sequence adjacent to the unedited miRNA miR-374. This novel stem-loop structure may represent a novel miRNA (Figure [2](#F2){ref-type="fig"}, novel hairpin). In the remaining five cases, the A-to-G changes were from the opposite strand to the miRNA (that is, U-to-C changes in the miRNA sequence). Although U-to-C editing of miRNA sequences cannot be ruled out, no editing of this type has previously been observed and no enzymes capable of catalyzing this conversion are known. The most likely explanation is that these are A-to-I edits in a transcript derived from the DNA strand complementary to the annotated miRNA gene. Consistent with this hypothesis, all of these sequences overlap, or are adjacent to, genes transcribed from the opposite strand to the annotated miRNA gene. To distinguish these sequences from the edited pri-miRNAs, these sequences are referred to here as edited antisense pri-miRNAs (Table [1](#T1){ref-type="table"}). One of the antisense pri-miRNAs contains editing sites overlapping the intended miRNA (miR-144) and miR-451, a recently identified miRNA that was not deliberately included in our list of 231 miRNAs. Collectively the 13 sequences were edited at 18 sites. Ten out of the 13 were edited at a single site. miR-376a and antisense miR-451 were each edited at two sites, and antisense miR-371 was edited at four sites. The extent of editing varied with editing site and with tissue, ranging from around 10% (for example, miR-151 in multiple tissues) to around 70% (antisense miR-371 in placenta). Overall, the levels of RNA editing observed were considerably higher than the approximately 5% editing previously reported for the -1 position of miR-22 \[[@B31]\]. Editing of miR-22 was not detectable by our method, presumably because the low levels of editing of this miRNA fall below our limits of detection. All miRNAs were found to be edited in multiple tissues, with the extent of editing varying from tissue to tissue (Figure [1](#F1){ref-type="fig"}). All novel A-to-I editing sites were found within the dsRNA stems of the predicted stem-loop structures (Figure [2](#F2){ref-type="fig"}). Of the seven editing sites in pri-miRNAs, four were in the 22-nucleotide mature miRNA. Three of these were within nucleotides 2 to 7, which are thought to be important for conferring binding-site specificity between the miRNA and its target sites \[[@B3]\]. Five out of seven editing sites in pri-miRNAs were at single nucleotide A:C mismatches flanked by paired bases. Similarly, five out of seven editing sites were in 5\'-UAG-3\' trinucleotides. These results are consistent with local structural and sequence preferences of RNA editing determined from A-to-I editing sites in inverted repeat sequences \[[@B25]\]. Three of the ten editing sites in antisense pri-miRNAs were in 5\'-UAG-3\' trinucleotides. Six of the ten editing sites were at A:C mismatches. Only one was at a single A:C mismatch, however, with the remainder at extended mismatches involving more than one consecutive nucleotide. Discussion ========== We have identified novel A-to-I editing sites in six out of 99 pri-miRNAs, indicating that at least 6% of all human miRNAs may be targets of RNA editing. We were only able to detect relatively high levels of editing, as illustrated by our failure to detect editing of miR-22, so this estimate is probably a conservative one. Moreover, our method is not strand specific, and cannot distinguish multiple overlapping transcripts from the same genomic locus. Thus, in regions of transcriptional complexity, it is likely that the sensitivity of our assay will be reduced. For example, even miRNAs that are 100% edited would appear to be unedited if transcribed at low levels compared with an unedited overlapping transcript from the opposite strand. We may also be unable to detect RNA editing if it occurs subsequent to the processing of the pri-miRNA (for example, by splicing) such that the binding sites for the PCR primers are removed. In addition to the edited pri-miRNAs, six antisense pri-miRNA transcripts derived from the opposite strand to the annotated miRNA were subject to A-to-I editing. There are many potential explanations for apparent editing on the opposite strand to the annotated miRNA. One possibility is that these sequences are actually due to U-to-C editing of the pri-miRNA. There are, however, no known U-to-C RNA editing enzymes capable of catalyzing such a reaction, and despite extensive searches for RNA editing sites, only a single U-to-C RNA editing site has been reported \[[@B32]\]. It is therefore more likely that these sequences represent an edited transcript from the opposite strand to the annotated miRNA. These transcripts could be another miRNA transcribed and processed from the genomic strand opposite the annotated miRNA, or they could be some other class of transcript, for example the intron of a gene overlapping the annotated miRNA but transcribed from the opposite DNA strand. Alternatively, these may be pri-miRNAs that have been incorrectly annotated to the wrong strand of the genome. To evaluate the possibility that the edited antisense pri-miRNAs are due to incorrect annotation of miRNAs to the wrong genomic strand, we examined previous experimental data obtained for these miRNAs. One of the edited antisense pri-miRNA sequences is derived from the DNA strand opposite the computationally predicted miR-215 \[[@B33]\]. The method used to predict miR-215 successfully predicted 81 out of 109 known miRNAs from a reference set, but around 20% (17/81) were predicted on the wrong strand of the genome \[[@B33]\]. Our data and the direction of overlapping transcripts suggest that miR-215 may have been annotated to the wrong genomic strand. An edited antisense miRNA sequence was also derived from the DNA strand opposite experimentally verified miRNA miR-133a \[[@B34]\]. This miRNA is present in the genome in two copies (miR-133a-1 and miR-133a-2). Copy miR-133a-2 is hosted within a gene transcribed in the same direction as the annotated miRNA gene. In contrast, copy miR-133a-1 overlaps a gene transcribed from the opposite strand. Cloning and expression analysis of miR-133a \[[@B34]\] provides proof that at least one copy of miR-133a is transcribed. As a result of this finding, both copies of miR-133a have been annotated according to the sequence of the cloned copy. Given the direction of overlapping transcripts, however, it remains possible that miR-133a-1 is transcribed from the opposite strand to miR-133a-2, giving rise to a different miRNA. Indeed, our results suggest that miR-133a-1 may have been incorrectly annotated. Similarly, both copies of experimentally verified miR-194 (miR-194-1 and miR-194-2) have been annotated according to the sequence of a cloned copy \[[@B34]\]. Our data and the presence of overlapping transcripts on the opposite strand suggest that miR-194-1 may also have been incorrectly annotated to the wrong genomic strand. In the case of both mir-133a and mir-194, the two copies would generate miRNAs that are perfectly complementary to one another. It has previously been suggested that pairs of complementary miRNAs play a role in miRNA regulation by forming miRNA:miRNA duplexes \[[@B35]\]. Our results suggest that RNA editing may add a further layer of regulation by disrupting complementarity in miRNA:miRNA duplexes. A further two edited antisense miRNA sequences (antisense mir-144 and antisense mir-451) overlap miRNAs that are annotated on the basis of their similarity to mouse miRNAs, and have not been cloned or shown to be expressed by northern blotting in human tissues. The remaining antisense miRNA sequence overlaps mir-371, which has been validated by cloning and northern blotting in human tissues and is therefore correctly annotated. The presence of edited nucleotides in pri-miRNA transcripts indicates that RNA editing occurs early in miRNA biogenesis. Subsequent processes that recognize sequence or structural features of the miRNA precursor could therefore potentially be affected by RNA editing. These include cleavage of the pri-miRNA by Drosha, export of the pre-miRNA to nucleus by exportin-5, cleavage of the pre-miRNA by Dicer, and miRNA strand selection for inclusion in the microprocessor complex. Indeed, it has recently been demonstrated that RNA editing of pri-miRNAs can result in suppression of processing by Drosha, and subsequent degradation of the unprocessed edited pri-miRNA \[[@B36]\]. Although it is unclear whether a miRNA that base-pairs with its target through an I:U wobble would be functional, another possibility is that RNA editing may alter target site complementarity. To investigate the effect of RNA editing of miRNAs on target-site complementarity, we used the miRanda software \[[@B37]\] to predict binding sites of edited miRNAs in 3\' UTRs, and compared these with the predicted binding sites of the equivalent unedited miRNAs. For each of the four pri-miRNAs with an editing site in the mature 22mer, the set of predicted targets of edited miRNAs differs from the predicted targets of edited miRNAs (Table [1](#T1){ref-type="table"}). For the three miRNAs in which the edited adenosine is at a position two to seven bases from the 5\' end of the miRNA (miR-151, miR376a and miR-379) over half of the targets of the edited miRNA are unique to the edited miRNA. In the case of miR-99a the difference is small, with only 5/75 (6%) target predictions differing between edited and unedited miRNAs. In all cases, the top ten predicted targets of the edited miRNA differ from the top ten predicted targets of the unedited miRNA (data not shown). To gain further insight into the potential biological impact of miRNA editing, we identified Gene Ontology (GO) terms in the \'cellular process\' category \[[@B38]\] which were over-represented in the predicted targets of edited and unedited miRNAs compared with all Ensembl genes (Figure [3](#F3){ref-type="fig"}). For the three miRNAs that are edited in the 5\' seed region (miR-151, miR-376a and miR-379), comparison of over-represented GO terms associated with the predicted targets of edited and unedited copies reveals distinct differences (Figure [3](#F3){ref-type="fig"}). Of particular interest are the additional terms that become over-represented; these include regulation of programmed cell death, biosynthesis, RNA metabolism, cell proliferation and transcription (Figure [3](#F3){ref-type="fig"}). RNA editing may therefore contribute to miRNA diversity by generating multiple different miRNAs from an initial pool of identical miRNA transcripts. For example, the total number of predicted targets of Hsa-mir-151 increases from 143 to 229 when taking into consideration both edited and unedited miRNAs. Editing of miRNAs may simultaneously alleviate and augment the gene-regulation effects of miRNAs by changing the concentration of individual miRNAs. Conclusion ========== We have performed the first systematic survey of RNA editing of human miRNAs. We have identified RNA editing sites in at least 6% of human miRNAs that may impact on miRNA processing, including edits that alter miRNA binding sites and contribute to miRNA diversity. Furthermore, our results suggest that some miRNA genes may have been incorrectly annotated to the wrong strand of the genome. This has implications for the interpretation of existing miRNA experiment data and future experimental design. Materials and methods ===================== Total RNA, total cDNA and genomic DNA ------------------------------------- For the initial screen of RNA editing in ten human tissues, total RNA and matching genomic DNA from the same tissue sample was obtained for human brain, heart, liver, lung, ovary, placenta, skeletal muscle, small intestine, spleen and testis from Biochain (Hayward, USA). For each tissue, sequence data was obtained from one individual. The donor was different for each tissue type. Total cDNA synthesis was performed using random nonamers (200 ng per 20 μl reaction) with Superscript III (Invitrogen, Carlsbad, USA) according to the manufacturer\'s instructions. Sequencing of pri-miRNAs ------------------------ Primers were designed to the genomic sequence in the vicinity of all 231 miRNA sequences in the miRNA registry version 7.0 \[[@B6]\], using primer3 \[[@B39]\]. PCR primer design was optimized to give PCR products of approximately 500 bp with at least 75 nucleotides either side of the predicted stem-loop structure. PCR primers were used to sequence PCR products in both directions on ABI3700 DNA sequencers. Sequence traces were quality scored using phred. Sequences with less than 70% of bases having a quality score of 20 or more were rejected. In the first stage of sequencing, duplicate PCR and sequencing was performed for each miRNA from each tissue. A miRNA was considered to be successfully sequenced if the following minimum sequence requirements were met for at least one tissue: good-quality sequence from both strands of one PCR, and good-quality sequence from at least one strand of a second PCR. Successfully sequenced miRNAs that were found to be edited were submitted to a second confirmation stage of sequencing. In the second stage of sequencing, quadruplicate PCR and sequencing was performed for each miRNA from each tissue. For each tissue, a miRNA was considered to be successfully sequenced if the following minimum sequence requirements were obtained: good-quality sequence from both strands of one PCR, and good-quality sequence from at least one strand of a second PCR. See Additional data file 1; primary sequence data is available from \[[@B40]\]. Detection and quantification of RNA editing ------------------------------------------- Sequences were visualized and compared in a gap4 database. A-to-I editing was identified as a novel G peak and a drop in peak height at As in a cDNA sequence relative to the equivalent peak in the matching genomic DNA sequence. The extent of RNA editing was estimated using a modified version of the comparative sequence analysis (CSA) method \[[@B41]\]. Briefly, this program normalizes a cDNA sequence trace to a genomic DNA reference trace by comparison of peak heights at unedited nucleotides. The drop in peak height between the DNA reference trace and the cDNA trace at the edited nucleotide is then reported as a percentage of the peak height in the genomic DNA reference trace. For each edited miRNA, the mean extent of editing for each tissue is calculated from all cDNA sequences obtained for that tissue. Analysis of novel RNA editing sites ----------------------------------- miRNA structures were obtained from the miRBase database \[[@B6]\]. Stem-loop structures of antisense miRNAs were generated by folding the antisense of the miRNA stem-loop sequence obtained from miRBase using MFOLD \[[@B42]\]. To predict edited and unedited miRNA target sites, miRanda (v3.0) \[[@B32]\] was used to scan the edited and unedited miRNA sequences against all human 3\' UTR sequences available from Ensembl v34. The algorithm uses dynamic programming to search for maximal local complementarity alignments, which correspond to a double-stranded antiparallel duplex. The new version of the miRanda algorithm (AJ Enright, personal communication) assigns *P*values to individual miRNA-target binding sites, multiple sites in a single UTR, and sites that appear, from a robust statistical model \[[@B43]\], to be conserved in multiple species. The resulting targets were filtered based on *P*value (*p*\< 0.001) to ensure a high degree of confidence in the predicted target sites. GO analysis ----------- GO terms from level 4 of the \'cellular process\' category were obtained for each human transcript from Ensembl. Over-representation for each term (*O*~term~) in a group of sequences with C terms is calculated as follows: ![](gb-2006-7-4-r27-i1.gif) ![](gb-2006-7-4-r27-i2.gif) where *F*~1~is the frequency of a term in the group being considered, *F*~2~is the frequency of a term in the whole genome and *t*is the term at level *L*. GO terms with low transcript counts (\< 3.0) were excluded from further analysis. Additional data files ===================== The following additional data are available with this paper online. Additional data file [1](#S1){ref-type="supplementary-material"} contains examples of edited sequence traces for each of the edited sites identified in this survey, and the coordinates of edited bases. Additional data file [2](#S2){ref-type="supplementary-material"} contains PCR primer information, details of the initial screen of miRNAs and annotation of edited miRNAs. Supplementary Material ====================== ###### Additional data file 1 Figures containing examples of edited sequence traces for each of the edited sites identified in this survey, and the coordinates of edited bases. ###### Click here for file ###### Additional data file 2 PCR primer information, details of the initial screen of miRNAs and annotation of edited miRNAs. ###### Click here for file Acknowledgements ================ We would like to thank all members of the cancer genome project for technical help sequencing miRNAs and the Wellcome Trust for funding. Figures and Tables ================== ![A-to-I RNA editing of miRNA precursors in human tissues. The extent of A-to-I editing at each editing site is indicated by the color scale. Each colored box represents the average extent of editing calculated from at least two PCR product sequences, at least one of which was sequenced in both directions. Gray boxes indicate miRNAs that could not be amplified. The number in brackets after the miRNA name is the position of the edited adenosine from the 5\' end of the pre-miRNA or equivalent antisense pre-miRNA from the miRNA registry.](gb-2006-7-4-r27-1){#F1} ![Positions of edited adenosines in human pri-miRNAs and antisense pri-miRNAs. Folded pri-miRNA structures were taken from the miRNA registry \[6\]. Antisense pre-miRNA structures were generated from the reverse complement pri-miRNA sequence using MFOLD \[38\]. Mature miRNA sequences of around 22 nucleotides and antisense mature miRNA sequences of around 22 nucleotides are indicated by red letters. Edited adenosines are highlighted in yellow. In antisense Hsa-mir-371, edited adenosines were found to reside in base-paired sequence extending beyond the annotated hairpin. Additional bases are in gray.](gb-2006-7-4-r27-2){#F2} ![GO term comparison of edited and unedited miRNA target predictions. For each edited miRNA, GO terms from level 4 of the \'biological process\' category that are over-represented in predicted targets of the unedited or edited miRNA (indicated by +) compared with all Ensembl genes were identified. All values are normalized and colored in terms of significance, with bright red cells indicating that a miRNA specifically targets genes in that GO functional class.](gb-2006-7-4-r27-3){#F3} ###### Predicted targets of edited and unedited miRNAs MicroRNA Edited only Unedited only Edited and unedited -------------- ------------- --------------- --------------------- Hsa-miR-151 86 84 59 Hsa-miR-376a 74 58 71 Hsa-miR-379 79 70 75 Hsa-miR-99a 5 8 70 Target predictions were performed using the miRanda software using a probability score cut-off of *p*\< 0.001. For each miRNA, the number of targets predicted for both edited and unedited miRNAs is shown against the number of targets predicted exclusively for edited miRNAs, and the number of targets predicted exclusively for unedited miRNAs.

{ "pile_set_name": "PubMed Central" }

Dear Sir, We read with interest the article "Repeat gas insufflation for successful closure of idiopathic macular hole following failed primary surgery" by Rishi *et al*.\[[@ref1]\] We describe two cases with failed primary macular hole (MH) surgery that closed following repeat fluid-gas exchange (FGE) and make some additional observations. Case 1 was a 55-year-old male with a 2-month history of blurred vision in his left eye. Best-corrected visual acuity (BCVA) was 20/120 in this eye, with a stage IV MH. He underwent 23 gauge pars plana vitrectomy with internal limiting membrane (ILM) peeling and 14% C3F8 gas injection, with prone positioning. At 4 weeks postoperatively, the MH was open with a cuff of subretinal fluid (SRF). He underwent repeat FGE with 14% C3F8 gas and prone positioning. At 4 weeks following the second intervention, the MH was found to be closed, and BCVA increased to 20/40 \[[Fig. 1](#F1){ref-type="fig"}\]. Case 2 was a 60-year-old male with a BCVA of 20/80 in his right eye and a stage IV MH of 1-month duration. Primary MH surgery resulted in a smaller but open MH with an elevated inferior edge at 1-month follow-up. Repeat FGE with 14% C3F8 gas and prone positioning was carried out resulting in a closed MH and BCVA of 20/40 1 month later \[[Fig. 2](#F2){ref-type="fig"}\]. ![(a, b) Color fundus photograph and vertical spectral-domain optical coherence tomography scan of case 1 at presentation showing stage IV macular hole (MH). (c, d) One month following primary surgery, the MH was open with a cuff of subretinal fluid. (e, f) Following repeat fluid-gas exchange, type 1 closure occurred at 1-month follow-up](IJO-62-1104-g001){#F1} ![(a, b) Color fundus photograph and vertical spectral-domain optical coherence tomography scan of case 2 at presentation showing stage IV macular hole (MH). (c, d) Following primary MH surgery, the size of the MH decreased but the inferior edge was elevated. (e, f) Repeat fluid-gas exchange was performed resulting in type 1 closure noted 1-month postoperatively](IJO-62-1104-g002){#F2} Nonclosure or reopening of MH after a primary MH surgery can be managed by repeat ILM peeling,\[[@ref2]\] additional FGE done externally (as an outpatient procedure)\[[@ref3]\] or internally (repeat surgery).\[[@ref1]\] Once the gas bubble has absorbed enough to allow adequate examination of the macula, an open MH with elevated edges and a cuff of SRF is unlikely to close or flatten without additional intervention.\[[@ref3]\] The authors describe the MH following primary surgery as "type 2 closure of MH with cuff of SRF" \[[Fig. 2](#F2){ref-type="fig"}\] and mention that the cuff of SRF was a factor favoring re-surgery in their case.\[[@ref1]\] Type 2 closure is used to denote attachment of the entire rim of the MH to the retinal pigment epithelium (RPE) with flattening of the cuff of SRF, and persistence of a foveal defect of the neurosensory retina, while type 1 closure indicates that the foveal defect is closed.\[[@ref4]\] Thus, the appearance of their case following primary surgery is incorrectly labeled as type 2 closure. Postoperatively, careful inspection needs to be done to exclude any epiretinal membrane that may be preventing closure of the MH. If there is no such traction, then an early surgical failure should respond to FGE with a long-acting gas and prone positioning.\[[@ref3]\] A gas tamponade in the vitreous cavity is believed to limit the movement of fluid into the hole, maintain apposition of the edges of the MH to the RPE and facilitate glial proliferation to seal the macular defect. The time taken for this wound healing response that leads to MH closure is probably variable.\[[@ref5]\] So while some MH may close earlier, some may require renewal of the foveal tamponade that provides additional time for the process of MH closure.\[[@ref3][@ref5]\] The use of a long acting tamponade can keep the fovea isolated from the fluid without stringent prone positioning, which may not be possible in a subset of patients. Thus, postoperative FGE with a long-acting gas to provide additional macular tamponade along with repeat patient counseling to ensure compliance with prone positioning appears to be a viable option after failed MH surgery. Successful closure of the MH is associated with improvement in visual acuity in these cases.

{ "pile_set_name": "PubMed Central" }

Background ========== The *Drosophila*adult brain is a highly organized and complex structure that contains thousands of neurons (in the order of 10^5^) \[[@B1]\] exhibiting multiple cell-type identities, as characterized by various morphological, electrophysiological and molecular features. All these neurons arise from the mitotic activity of a small number of progenitor cells (neuroblasts (NBs)), which generate lineages of clonally related neurons via two proliferative phases of neurogenesis \[[@B2],[@B3]\]. The first phase of neurogenesis takes place during embryogenesis and starts with the specification and delamination of the NBs from the procephalic neurogenic region. Between embryonic stages 8 and late 11, around 100 NBs delaminate from this region on either side of the embryo in a reproducible spatiotemporal pattern \[[@B4]\]. Each NB assumes a unique identity, as revealed by the expression of a specific set of marker genes such as the proneural genes, gap genes and segment polarity genes \[[@B5]-[@B7]\], and gives rise to an invariant cell lineage through multiple rounds of asymmetric cell divisions. In each cell division, the NB self-renews and generates a smaller daughter cell (named the ganglion mother cell), which divides only once to give rise to two post-mitotic neurons or glial cells (reviewed in \[[@B8]-[@B10]\]). The first neurogenic process terminates at the end of embryogenesis, when most NBs stop dividing and enter a dormant phase called quiescence \[[@B11]\]. The cells so far generated (primary neurons) wire the larval nervous system and eventually may remodel during metamorphosis to contribute to the adult brain \[[@B12],[@B13]\]. The second phase of neurogenesis starts during the late first (L1) and second (L2) instar larval stage, when the inactive NBs resume mitotic activity and, through rounds of asymmetric cell divisions, generate the population of secondary neurons and glial cells that accounts for more than 90% of the adult brain \[[@B11],[@B14]\]. Hence, a complete NB lineage can be divided into two discrete cell populations, each containing the cells generated by the NB during distinct developmental phases (primary neurons during embryogenesis and secondary neurons during larval development) and each harboring multiple neuronal cell types. It has been proposed that neuron identity within a NB lineage depends on a combination of spatial and temporal cues provided, firstly, by the unique identity the NB acquires during its specification/delamination time \[[@B7],[@B15]\] and, secondly, as a result of a birth time/order-dependent mechanism \[[@B16]\], whereby cell-type specification of the nascent post-mitotic neurons depends on the identity of the progenitor temporal transcription factor expressed by the NB at each particular time during lineage progression \[[@B17]\]. The homeobox gene *orthodenticle*(*otd*), as a cephalic gap gene, is expressed in broad domains of the procephalic ectoderm during early neurogenesis, covering most of the protocerebral anlage and the anterior part of the deutocerebral anlage \[[@B18]\]. Subsequently, its expression is also detected in NBs delaminating from these domains, where it plays instructive roles important for cell viability and spatial identity of the nascent NBs \[[@B5]\]. It has also been proposed that *otd*might control brain NB formation by triggering proneural gene expression \[[@B19]\]. Inactivation of *otd*at this early embryonic phase impairs NB formation and leads to a gap-like phenotype in the anterior head that includes the deletion of the protocerebral anlage and part of the deutocerebral anlage \[[@B20]\]. Later in development, *otd*expression is also detected in post-mitotic neurons of the developing brain and ventral nerve cord, not only in the embryo, but also in the larval and pupal brain and even in mature neurons of the adult brain. In this regard, flies homozygous for an *otd*viable hypomorphic mutation called *ocelliless*(*oc*) show developmental defects that affect the protocerebral bridge, an important neuropile structure in the adult brain that is part of the central complex \[[@B20]\]. However, whether this phenotype is due to the altered expression of *otd*in progenitor cells, post-mitotic cells or both is not known. A comprehensive study of gene function during neuronal cell fate specification requires a previous and thorough cell lineage analysis and demands cell type-specific molecular markers to trace the cells under study. In this paper, we have focused our attention on the array of dopaminergic (DA) neurons that populate the *Drosophila*central brain during larval development. We have used cell lineage tracing genetic techniques together with immunoreactivity against the enzyme tyrosine hydroxylase (TH), the rate-limiting enzyme in dopamine biosynthesis \[[@B21],[@B22]\], to study their development, phenotypic maturation and lineage relationships. Interestingly, one cluster of clonally related DA neurons expresses *otd*from early L1 to adulthood, allowing us to examine the post-mitotic role of *otd*in controlling identity and/or survival of DA neurons in the *Drosophila*larval and adult brain. Results ======= Birth, clustering and phenotypic maturation of dopaminergic neurons in the larval central brain ----------------------------------------------------------------------------------------------- Using immunoreactivity against the enzyme TH as a DA neuron-specific molecular marker, an array of 21 DA neurons could be visualized in the central part of each brain hemisphere during the third instar larval stage (L3; Figure [1A](#F1){ref-type="fig"}). The stereotypical arrangement of their cell bodies in various groups has been used previously to define four clusters of DA neurons that occupy distinct anatomical positions in the central brain \[[@B23]-[@B25]\]: dorso medial 1 cluster (DM1: four cells), dorso medial 2 cluster (DM2: four cells), dorso lateral 1 cluster (DL1: seven cells) and dorso lateral 2 cluster (DL2: six cells). As judged by their projection and innervation patterns, DM2 and DL1 cell clusters contained apparently homogeneous populations of DA neurons. On the contrary, both DM1 and DL2 clusters could be further subdivided into two subclusters according to their differential projection patterns. A single neuron in the DM1 cluster (named DM1a) projected ventrally into the lower part of the ipsilateral brain lobe (Figure [1B](#F1){ref-type="fig"}), whereas the remaining three DM1 DA neurons (named the DM1b cell cluster) innervated more ventrally localized regions of the brain lobe (subesophagial ganglion) and further extended into the thoracic segments of the ventral ganglion (Figure [1C](#F1){ref-type="fig"}). DM2 neurons projected ipsilaterally into the anterior part of the protocerebrum (Figure [1D](#F1){ref-type="fig"}), whereas DL1 neurons were characterized by dorsally projecting neurites that bifurcated into dorsal and ventral branches before crossing the midline to reach the contralateral brain lobe (Figure [1E](#F1){ref-type="fig"}). Although described as homogeneous, DM2 and DL1 cell clusters are possibly heterogeneous at the single-cell level. Indeed, six different cell types with slightly different innervation patterns have been described within the DL1 cell cluster using single-cell labeling techniques \[[@B25]\]. Similarly to the DM1cell cluster, the six DL2 DA neurons displayed two distinct projection patterns: four neurons (named the DL2a cell cluster) projected their neurites dorsally into the anterior part of the ipsilateral brain lobe (Figure [1F](#F1){ref-type="fig"}). The remaining two cells (named the DL2b cell cluster) projected laterally and arborized in the ventral part of the brain lobe before crossing the midline to innervate the contralateral brain lobe (Figure [1G,G\'](#F1){ref-type="fig"}). ![**Birth, clustering and differentiation of larval central brain dopaminergic neurons**. **(A)**Third instar larval stage (L3) brain showing bilaterally symmetrical groups of TH-positive DA neurons clustered according to the position of the cell bodies within the cortex (the number of neurons per cluster is given in parentheses) and the neurite projection patterns. Scale bar: 50 μm. **(B-G\')**Neurite projection patterns. Wild-type cell clones were induced in progenitor cells during early embryogenesis (3 to 7 h after egg laying (AEL)) and analyzed during L3. The clones were labeled with membrane tethered green and red flourescent proteins (both in grey) using the Twin-spot MARCM technique together with the *TH-gal4*driver. When possible, neurite projection patterns of the entire cell cluster are shown (B,E,F); otherwise, just one DA neuron as a representative for the cell cluster is produced (C,D). (G,G\') The neurite projection pattern (red arrowheads) of DL2b DA neurons (yellow arrows) is shown in two consecutive confocal optical sections. **(H)**Early L1 brain (24 to 28 h AEL) showing clusters of TH-positive cells (the number of DA neurons per cluster is given in parentheses) Scale bar: 10 μm. **(I)**Most larval DA neurons in the central brain are born during early neurogenesis between 3 to 13 h AEL. Wild-type cell clones, labeled with the MARCM technique in combination with the *TH-gal4*driver, were induced in progenitor cells at different times during development as indicated and analyzed in L3 brains. Images, except for (G,G\'), represent Z projections of individual confocal optical sections. AEL, after egg laying; DA, dopaminergic; DL, dorso lateral; DM, dorso medial; L, instar larval stage; MARCM, mosaic analysis with a repressible cell marker; ML, midline; TH, tyrosine hydroxylase.](1749-8104-6-34-1){#F1} We next determined the time of birth of the larval central brain DA neurons by using the MARCM (mosaic analysis with a repressible cell marker) technique in combination with the *TH-Gal4*driver. Flippase recognition target (FRT)-mediated mitotic recombination was randomly induced, exposing progenitor cells to a one hour heat-shock treatment (37°C) at different developmental stages and L3 brains were assayed for the presence of green fluorescent protein (GFP)-labeled DA neurons. When mitotic recombination was induced during early L1, a time point commonly used to label neurons born during larval development, GFP-positive DA neurons were not detected in the resulting wild-type cell clones. This result agreed with the observation that *TH*expression was already detectable in early L1 brains (Figure [1H](#F1){ref-type="fig"}). On the contrary, a heat-shock treatment applied during early embryogenesis efficiently labeled DA neurons in the L3 central brain, with the highest labeling efficiencies achieved when cell clones were induced at early embryonic stages (between 3 and 7 and 7 and 11 hours after egg laying (AEL); Figure [1I](#F1){ref-type="fig"}). These results demonstrate that DA neurons present in the L3 central brain are primary neurons that arise during early embryogenesis. Despite their early embryonic origin, the larval central brain DA neurons did not express the cell type-specific marker gene *TH*during embryogenesis, and even at late embryonic stages (stage 17) anti-TH immunoreactivity in the central nervous system was restricted to the ventral nerve cord (data not shown and \[[@B24]\]). However, during early L1 (24 to 28 h AEL) most of the central brain DA neurons already displayed *TH*expression (Figure [1H](#F1){ref-type="fig"}), with two exceptions: the DM1a DA neuron started to show anti-TH labeling during mid-late L1, whereas a DL2a DA neuron only showed anti-TH immunoreactivity at mid-late L3 (data not shown). In summary, DA neurons present in the central brain of the *Drosophila*larva at L3 are generated during early embryogenesis and most of them acquire a mature neurotransmitter phenotype during early L1. Dopaminergic neurons present in the central brain during larval development are generated by seven neuroblast lineages ---------------------------------------------------------------------------------------------------------------------- In *Drosophila*, clonally related neurons typically remain clustered in the mature brain and project their neurites into specific neuropile compartments \[[@B26],[@B27]\]. To address whether the anatomical clustering of DA neurons in the L3 central brain is due to a clonal origin, we analyzed lineage relationships among DA neurons in each cell cluster. For this purpose, we utilized the twin-spot MARCM technique in combination with the *tubulin-Gal4*driver. Wild-type cell clones were induced during early embryogenesis (3 to 7 h AEL) and assayed in early L1 brains (24 to 28 h AEL) for the presence of GFP- and red fluorescent protein (RFP)-labeled NB clones containing TH-positive neurons. Although at this developmental stage two DA neurons still did not express the cell type-specific marker gene *TH*and thus could not be considered in our analysis, two reasons justified our decision. Firstly, the reduced number of total cells present in the larval brain during L1 facilitated an easier lineage analysis. Secondly, we observed that the *tubulin*promoter underwent partial down-regulation in primary neurons during L3 (data not shown), impairing a reliable analysis of the lineage relationships among DA neurons. Our analysis revealed that all the DA neurons assigned to the DM1b (Figure [2A](#F2){ref-type="fig"}), DM2 (Figure [2B](#F2){ref-type="fig"}), DL2a (Figure [2D](#F2){ref-type="fig"}) and DL2b (Figure [2E](#F2){ref-type="fig"}) clusters were contained within individual NB clones and thus were clonally related. By contrast, we only found NB clones containing at most six out of the seven DL1 DA neurons (Figure [2C,C\'](#F2){ref-type="fig"}), indicating that two NBs generate the seven DL1 DA neurons. ![**Twin-spot MARCM lineage analysis of larval central brain dopaminergic neurons at L1 reveals clonal relationship between dopaminergic neurons**. **(A-E)**Membrane-tethered GFP- and RFP-labeled wild-type NB clones containing TH-positive (blue) DA neurons from different DA cell clusters: DM1b cluster (A), DM2 cluster (B), DL1 cluster (C,C\'), DL2a cluster (D) and DL2b cluster (E). All panels represent individual confocal optical sections (0.5 μm thick). The red asterisk in C~2~identifies the DL1 DA neuron that belongs to a different NB lineage (DL1b). The yellow asterisk in C~2~and C\'~2~identifies the same neuron in different optical sections. Scale bars: 10 μm. DA, dopaminergic; DL, dorso lateral; DM, dorso medial; GFP, green fluorescent protein; L, instar larval stage; NB, neuroblast; RFP, red fluorescent protein; TH, tyrosine hydroxylase.](1749-8104-6-34-2){#F2} As mentioned above, due to their delayed *TH*expression, two DA neurons were initially not included in the lineage analysis. In order to validate their allocation to the respective DA cell clusters, we induced MARCM-labeled wild-type NB clones during early embryogenesis (3 to 7 h AEL) and analyzed L3 brains. We tried to circumvent the down-regulation of the *tubulin*promoter by including an additional copy of the *UAS-CD8::GFP*transgene in the genotype of the analyzed larvae. Although this strategy did not always work reliably (for example, we did not observe single NB clones containing the group of six DL1 DA neurons), we were still able to assign the missing DM1 DA neuron to an independent NB lineage (named DM1a), as well as to confirm the allocation of the missing DL2 DA neuron to the DL2a NB lineage (Additional file [1](#S1){ref-type="supplementary-material"}). In summary, we found that seven NB lineages generate the DA neurons present in each hemisphere of the *Drosophila*larval central brain and that their clustered appearance is, at least in part, a consequence of a clonal relationship. Moreover, the assignment of DA neurons to particular NB lineages provides access to study genetic mechanisms of DA neuron cell fate specification. *orthodenticle*is expressed in DL2a dopaminergic neurons -------------------------------------------------------- The homeobox containing gene *otd*is broadly expressed in the anterior ectoderm during the blastoderm stage and is necessary for the specification of most of the NBs that populate the protocerebrum and part of the deutocerebrum \[[@B5],[@B19]\]. Using Otd immunostainings, we found that *otd*was expressed in DL2a DA neurons already at early L1 (Figure [3A](#F3){ref-type="fig"}). At mid-late L3, when the entire complement of DA neurons in the larval central brain was already visible with anti-TH labeling, Otd was specifically detected in the four DL2a DA neurons (Figure [3C](#F3){ref-type="fig"},E~1~-E~3~). ![***otd*is expressed in DL2a dopaminergic neurons in the larval brain**. **(A,C)***otd*(blue) is expressed in a specific cluster of TH-positive (red) DA neurons of the larval central brain during L1 (A) and L3 (C). **(B,D)**In *oc*^*γa1*^hemizygous larvae, two out of three DL2a DA neurons at L1 (B) and three out of four DL2a DA neurons at L3 (D) are not detected using anti-TH immunoreactivity. Scale bars: 10 μm (A,B); 20 μm (C,D). **(E)**The *oc2*enhancer is active in three out of the four DL2a DA neurons at L3. Scale bar: 10 μm. **(E**~**1**~**-E**~**3**~**)***otd*expression and **(E\'**~**1**~**-E\'**~**3**~**)***oc2*enhancer activity in DL2a DA neurons are analyzed in individual optical sections. **(E**~**4**~**,E\'**~**4**~**)**DL2b DA neurons display neither *otd*expression nor *oc2*enhancer activity. Scale bars: 5 μm (E~1~-E~4~). **(F,F**~**1**~**)**In L3 *oc*^*γa1*^hemizygous larvae, *oc2*enhancer activity in DL2a DA neurons is lost. Scale bar: 10 μm. **(G)***otd*^*YH13*^mutant cell clones induced during early embryogenesis impair the development of DL2a DA neurons (white arrow). Yellow arrows point to the DL2b DA neurons in the left hemisphere Scale bar: 50 μm. All the panels, except for (E~1~-E~4~) and (E\'~1~-E\'~4~), represent Z projections of individual confocal optical sections. DA, dopaminergic; DL, dorso lateral; GFP, green fluorescent protein; L, instar larval stage; *oc*, *ocelliless*; *otd*, *orthodenticle*; TH, tyrosine hydroxylase.](1749-8104-6-34-3){#F3} To analyze the significance of *otd*expression in the specification and/or survival of post-mitotic DL2a DA neurons, we made use of the hypomorphic *otd*allele *oc*. *oc*hemizygous flies are viable, but they show a lack of ocelli and associated bristles in the head vertex. These flies also lack the protocerebral bridge, a neuropile structure present in the fly adult brain that is part of the central complex \[[@B20]\]. These phenotypes arise as a consequence of chromosomal rearrangements that remove *cis*-acting regulatory sequences (*oc*enhancer) important for *otd*expression during ocelli and protocerebral bridge development \[[@B18],[@B28]\]. During early L1, *otd*expression in *oc*hemizygous larvae was downregulated and the number of TH-positive DL2a DA neurons was reduced to one or two neurons, as compared to the three DL2a DA neurons present in wild-type brains (Figure [3B](#F3){ref-type="fig"}). In L3 brains, the lack of anti-TH immunoreactivity affected three out of the four DL2a DA neurons (Figure [3D](#F3){ref-type="fig"}). To investigate whether this phenotype is due to the loss of DL2a DA neurons *per se*or to *TH*downregulation in these neurons, we made use of the *oc*enhancer and the *oc2-Gal4*driver \[[@B28]\] as an alternative way of labeling DL2a DA neurons. Detection of *oc*enhancer activity in DL2a DA neurons in *oc*mutant L3 brains would imply that *otd*is involved in the activation and/or maintenance of *TH*expression in these neurons. By contrast, the absence of *oc*enhancer activity in DL2a DA neurons in *oc*mutant L3 brains would indicate *otd*is primarily necessary for survival of DL2a DA neurons. Reporter gene expression under the control of the *oc2-Gal4*driver was detected in three out of the four DL2a DA neurons in wild-type L3 brains (Figure [3](#F3){ref-type="fig"}E\'~1~-E\'~3~). Interestingly, when *oc2-gal4*transcriptional activity was analyzed in *oc*mutant L3 brains, reporter gene expression was not detected in DL2a DA neurons (Figure [3F](#F3){ref-type="fig"},F~1~). We also analyzed the relevance of *otd*expression during embryogenesis for the generation of DL2a DA neurons. We induced MARCM-labeled cell clones mutant for a null *otd*allele during early embryogenesis and assayed L3 brains for the presence of TH-positive cells within these *otd*mutant clones. Very few *otd*- NB clones were recovered in the central brain and none of them affected the DL2a DA neurons. However, L3 brains lacking some of the DL2a DA neurons were observed (white arrow in Figure [3G](#F3){ref-type="fig"}), though the corresponding NB clone was not detected. The simplest interpretation for this result is that Otd depletion during early neurogenesis induces NB cell death and, as a result, loss of DL2a DA neurons. Taken together, the lack of anti-TH labeling observed in three DL2a DA neurons in *oc*mutant L3 brains is not due to *TH*downregulation, but likely reflects the absence of these neurons. Thus, we conclude that *otd*is necessary for survival of the larval DL2a DA neurons. Adult PPL2 dopaminergic neurons derive from the larval DL2a cluster ------------------------------------------------------------------- Similarly to the larval brain, DA neurons also cluster in the adult *Drosophila*brain and these clusters have been annotated according to their anatomical position \[[@B23],[@B24],[@B29]\]. In order to investigate whether adult DA neurons express *otd*, we assayed wild-type young adult brains (3 to 7 days old after eclosion) for the co-expression of *otd*and *TH*. We observed that seven DA neurons, assigned to the protocerebral posterior lateral 2 (PPL2) cluster, expressed *otd*(Figure [4A](#F4){ref-type="fig"},A~1~-A~4~) and five of them also showed *oc2*enhancer transcriptional activity (Figure [4A](#F4){ref-type="fig"},A\'~1~-A\'~4~). Interestingly, the entire PPL2 cluster was not detected in *oc*mutant hemizygous flies using anti-TH immunoreactivity (Figure [4B](#F4){ref-type="fig"}). Also, targeted knockdown of *otd*in DA neurons by RNA interference (RNAi) resulted in a partial phenocopy of this mutant phenotype (*TH*expression was detected in three out of the seven PPL2 DA neurons; Figure [4C](#F4){ref-type="fig"},C~1~-C~3~) and this effect could be rescued by the simultaneous expression of the anti-apoptotic gene *P35*(Figure [4](#F4){ref-type="fig"}D~1~-D~4~). Otd protein levels in the rescued DA neurons (cells 1, 2, 5 and 7) were indistinguishable from background levels or drastically reduced (Figure [4](#F4){ref-type="fig"}D\'~1~-D\'~4~); yet, cell viability and *TH*expression were recovered by *P35*co-expression. These results indicate that *otd*is dispensable for *TH*expression in PPL2 DA neurons and it is mainly required for their survival in the adult brain. ![***otd*is expressed in PPL2 dopaminergic neurons in the adult brain**. **(A)***otd*(blue) expression in the PPL2 cell cluster of TH-positive (red) DA neurons in the adult brain. The PPL2 cell cluster labeled with a yellow asterisk is magnified and *otd*expression **(A**~**1**~**-A**~**4**~**)**and *oc2*enhancer activity **(A\'**~**1**~**-A\'**~**4**~**)**are analyzed in single optical sections. Scale bars: 10 μm. **(B)**In *oc*^*γa1*^hemizygous flies, the seven PPL2 DA neurons are not detected using anti-TH immunoreactivity. **(C)**Targeted expression of an *otd*-specific RNA interference construct in DA neurons (using the *TH-gal4*driver) impairs the viability of four PPL2 DA neurons in the adult brain of *otd*^*YH13*^heterozygous flies (7 days old). **(C,C**~**1**~**-C**~**3**~**)***TH*and *otd*expression in the resulting PPL2 cluster (labeled with a yellow asterisk in (C)) is analyzed at higher magnification. Scale bars: 10 μm. **(D**~**1**~**-D**~**4**~**)**Targeted expression of *P35*in Otd-depleted DA neurons (using the *TH-gal4*driver) rescues cell viability and *TH*expression in PPL2 DA neurons in the adult brain of *otd*^*YH13*^heterozygous flies (7 days old). Pictures represent single optical sections. Scale bars: 10 μm. **(D\'**~**1**~**-D\'**~**4**~**)**Otd protein levels in four PPL2 DA neurons (1, 2, 5 and 7) are not recovered. Panels (A-C,C~1~-C~3~) represent Z projections of individual confocal optical sections. Scale bars in (A-C): 50 μm. DA, dopaminergic; GFP, green fluorescent protein; *oc*, *ocelliless*; *otd*, *orthodenticle*; PPL, protocerebral posterior lateral; RNAi, RNA interference; TH, tyrosine hydroxylase.](1749-8104-6-34-4){#F4} *otd*expression and *oc2*enhancer activity in DA neurons of the larval and adult brain might establish an identity connection between the four larval DL2a DA neurons and the seven PPL2 DA neurons present in the adult brain. However, where did the additional three DA neurons present in the adult PPL2 cluster come from? A closer examination of L3 of the NB lineage that generates the DL2a DA neurons (Figure [5A-C](#F5){ref-type="fig"}) revealed that *otd*was expressed not only in the four TH-positive neurons, but also in three adjacent TH-negative cells (arrows in Figure [5A,B](#F5){ref-type="fig"}). These three cells might represent neurons born during larval development that acquire anti-TH immunoreactivity during pupal stages or, alternatively, they might be undifferentiated embryonic neurons that undergo phenotypic maturation only during pupal development. To distinguish between these two possibilities, twin-spot MARCM-labeled wild-type cell clones were induced in combination with the *TH-gal4*driver during early L1 and adult brains were assayed for the presence of labeled DA neurons. In 13% of the analyzed adult brains (n = 30), one or two PPL2 DA neurons were labeled (Figure [5D-E](#F5){ref-type="fig"}), yet marker gene expression was not detected in three PPL2 DA neurons simultaneously. This result indicates that at least two PPL2 DA neurons in the adult brain are born during larval development. Conversely, the remaining five PPL2 DA neurons are likely to be of embryonic origin. Together, our data support the notion that the adult PPL2 cluster is derived from the larval DL2a cluster. ![**PPL2 dopaminergic neurons in the adult brain are lineage related to the larval DL2a dopaminergic cell cluster**. **(A-C\')**DL2a NB lineage (labeled with membrane tethered GFP) was analyzed during L3 using individual optical sections. Within the NB clone, seven cells are Otd-positive (blue). Three of them (arrows) only express *otd*, whereas the other four express both *otd*and *TH*(red). Scale bars: 10 μm. **(D)**Twin-spot MARCM-labeled wild-type cell clones were induced during early L1 and analyzed in the adult brain. Scale bar: 50 μm. **(E-E\'\')**Magnified view of the PPL2 cell cluster showing two PPL2 DA neurons (arrows) labeled with membrane tethered RFP. Two additional RFP-labeled neurons (arrowheads) do not express *TH*and correspond to neurons in which the *TH-gal4*driver is ectopically activated. Scale bars: 10 μm. DA, dopaminergic; GFP, green fluorescent protein; L, instar larval stage; MARCM, mosaic analysis with a repressible cell marker; NB, neuroblast; *otd*, *orthodenticle*; PPL, protocerebral posterior lateral; RFP, red fluorescent protein; TH, tyrosine hydroxylase.](1749-8104-6-34-5){#F5} Discussion ========== Seven neuroblasts generate the larval central brain dopaminergic neurons ------------------------------------------------------------------------ The *Drosophila*larval central brain contains 21 DA neurons per hemisphere during L3, which express the cell type-specific marker gene *TH*\[[@B21],[@B22]\]. Different methods have been proposed to classify and annotate these neurons according to anatomical criteria (position of the cell bodies within the cortex and/or projection pattern of the axonal tracts) \[[@B23]-[@B25]\]. In this paper, we have analyzed these neurons from a developmental point of view and classified them according to their lineage relationship (Figure [6](#F6){ref-type="fig"}). The MARCM technique is a powerful tool to study lineage progression and cellular pedigrees during *Drosophila*brain development \[[@B30]\]. It allows the labeling of progenitor cells and their offspring at different times during development, depending on the timing of a heat-shock-induced flippase-mediated mitotic recombination event \[[@B31]\]. Using this technique, we have shown that the larval central brain DA neurons are primary neurons born during early embryogenesis. However, when analyzing the lineage relationship among these neurons, two major problems were encountered. Firstly, implicit in the technique is the fact that a labeled NB clone is accompanied by a non-labeled twin clone (two post-mitotic cells derive from the first ganglion mother cell born just after the mitotic recombination event). The exclusion of two cells from the lineage analysis is negligible when larval lineages are analyzed (the average size of a standard larval lineage at L3 is 120 cells \[[@B32]\]). However, embryonic lineages are small (on average between 10 and 20 cells at the end of embryogenesis \[[@B3]\]) and the exclusion of two cells can be significant. Secondly, MARCM-labeled NB clones induced during early embryogenesis can only be visualized with a considerable delay after their generation (from L2 onwards) due to the persistence of the Gal80 repressor protein \[[@B30]\]. These two problems have recently been circumvented by the development of the twin-spot MARCM technique \[[@B33]\]. This technique not only allows the visualization of cell clones earlier in development but also differentially labels the NB clone and the twin clone; thus, the study of the entire NB lineage is now possible. Using this technique, we have analyzed the lineage relationship among the DA neurons present in the *Drosophila*central brain during larval development. We found that seven NB lineages generate the 21 DA neurons present in the larval central brain (Figures [2](#F2){ref-type="fig"} and [6B](#F6){ref-type="fig"}; Additional file [1](#S1){ref-type="supplementary-material"}): DM1a (one DA neuron), DM1b (three DA neurons), DM2 (four DA neurons), DL1a (six DA neurons), DL1b (one DA neuron), DL2a (four DA neurons) and DL2b (two neurons). At large, the lineage analyses agree with the clustering of DA neurons according to anatomical criteria, supporting the general assumption that cell bodies arrangement and axonal projection patterns are reliable ways to classify neurons in *Drosophila*. Just in the case of the DL1 cell cluster was a discrepancy found. The cell bodies of the seven DL1 DA neurons are compactly arranged in a cell cluster that occupies medial-lateral positions in the L3 central brain and their neurites display similar projection patterns (Figure [1E](#F1){ref-type="fig"}). Yet, six DL1 DA neurons are clonally related (DL1a NB lineage; Figure [2C,C\'](#F2){ref-type="fig"}) and the remaining DL1 DA neuron (red asterisk in Figure [2](#F2){ref-type="fig"}C~2~) is generated by an additional NB (DL1b NB lineage). For future functional studies, it would be interesting to find molecular markers differentially labeling these two populations of DL1 DA neurons. ![**Summary of the classification of dopaminergic neurons present in the larval central brain at L3, according to anatomical and lineage relationship criteria**. **(A)**A cartoon depicting a dorsal view of a L3 larva central brain showing cell body locations and neurite projection patterns of the different types of DA neurons. For simplicity, just one neuron per cluster and per brain hemisphere is shown. The cartoon does not pretend to precisely reproduce the innervation pattern of every type of DA neuron, but just gives a general description. **(B)**Table showing the distribution of DA neurons in different NB lineages at different times during development. The lineage harboring Otd+ DA neurons is highlighted in blue. DA, dopaminergic; DL, dorso lateral; DM, dorso medial; L, instar larval stage; NB, neuroblast; PPL, protocerebral posterior lateral.](1749-8104-6-34-6){#F6} Otd acts as a survival factor in DL2a dopaminergic neurons ---------------------------------------------------------- Most studies involving the homeodomain transcription factor Otd in central nervous system development in *Drosophila*have dealt with its role in the specification and proliferation of progenitor cells during early neurogenesis \[[@B5],[@B19]\], whereas a possible function in post-mitotic neurons has been largely overlooked. Our observation that *otd*is expressed in the DL2a DA neurons during larval development prompted us to investigate its role in the specification and/or survival of this DA cell cluster. According to anti-TH labeling, DL2a DA neurons mature mainly during early L1. Thus, null *otd*alleles, which are embryonic lethal, could not be used in our analysis. Therefore, we investigated the hypomorphic *otd*allele *oc*. We found that in *oc*mutant hemizygous larvae, *otd*expression in dorsolateral regions of the central brain was reduced and, as a consequence, only one of the four DL2a DA neurons showed anti-TH labeling during L3. The failure to detect three of the four DA neurons can be due to a defect in the regulation of *TH*expression or to the loss of DA neurons *per se*. Several lines of evidence support the latter hypothesis. Firstly, a general regulator of *TH*expression would be expected to be present in all or most of the central brain DA neurons; yet, *otd*expression during larval development is restricted to the DL2a DA cell cluster. Secondly, misexpression of *otd*in randomly induced cell clones in the central brain during larval development does not result in ectopic *TH*-expressing DA neurons (data not shown). Thirdly, labeling of DL2a DA neurons with the *oc2-gal4*driver shows that reporter gene expression is also abolished in *oc*mutant hemizygous larvae during L3. The *oc2*enhancer has been shown to be positively regulated by *otd*during ocelli development \[[@B28]\] and might not, therefore, be suitable to label DL2a DA neurons in an *otd*-independent way. However, a minimal version of this enhancer harboring the characterized Otd binding site (*oc7*) was active in the ocelli primordium \[[@B28]\], but did not show enhancer activity in DL2a DA neurons during larval development (data not shown). This indicates that the *oc2*enhancer is differentially regulated in the ocelli primordium and in DA neurons during development and, hence, the *oc2-gal4*driver may be used to label DL2a DA neurons in an *otd*-independent fashion. Taken together, our observations support the hypothesis that *otd*expression is required for survival of DL2a DA neurons during larval development. DL2a dopaminergic neurons survive into adulthood and participate in the PPL2 dopaminergic cell cluster ------------------------------------------------------------------------------------------------------ The wild-type *Drosophila*adult brain is populated by about 200 DA neurons distributed in several bilaterally symmetric clusters \[[@B23],[@B24],[@B29]\]. The PPL2 cluster contains seven cells that express *otd*and five of them also show *oc2*enhancer activity in young adult brains (Figure [4](#F4){ref-type="fig"}A~1~-A~4~). Similarly to the larval brain, *otd*expression in PPL2 DA neurons seems to be necessary for their survival, since neither anti-TH immunoreactivity nor transcriptional activity of the *oc2*enhancer is detected in *oc*mutant adult brains (Figure [4B](#F4){ref-type="fig"} and data not shown). Moreover, the effects of targeted depletion of Otd in PPL2 DA neurons (loss of cell viability and/or *TH*expression; Figure [4C](#F4){ref-type="fig"}) can be rescued by the simultaneous expression of the anti-apoptotic gene *P35*, pointing out a role in cell survival as the main function of *otd*in PPL2 DA neurons. Altogether, the simplest interpretation for these results would be that *otd*expression labels homologous DA neuron populations in both the larval (DL2a cell cluster) and adult (PPL2 cell cluster) brains and, hence, both clusters contain the same DA neurons. The discrepancy in cell number between both clusters of DA neurons can be interpreted by analyzing the NB lineage responsible for the generation of the DL2a DA neurons. At L3, this lineage contains seven *otd*expressing cells, four of them are primary neurons that have already undergone maturation and express *TH*. The other three cells might represent immature secondary neurons that differentiate during pupal stages to give rise to the additional three DA neurons present in the adult PPL2 cluster. The distinction between early-differentiating (four cells) and late-differentiating (three cells) PPL2 DA neurons finds support in the targeted depletion of Otd in DA neurons by RNAi. Expression of an *otd*-specific RNAi construct in DA neurons (using the *TH-Gal4*driver) has no effect on the larval brain (data not shown), but impairs the viability of four PPL2 DA neurons in the adult brain. Since these four cells differentiate during larval development, the RNAi machinery would have more time to completely deplete Otd than in the case of the late differentiating DA neurons. Further support for this interpretation also comes from the analysis in the adult brain of wild-type twin-spot MARCM cell clones induced during early L1. According to this analysis, at least two PPL2 DA neurons in the adult brain are secondary neurons, whereas the third DA neuron might represent an undifferentiated primary neuron that only matures during pupal development. Recently, the expression of *Otx2*, an *otd*ortholog, in DA neurons in the mouse adult brain has also been reported \[[@B34]\]. It is selectively expressed in the central DA neurons of the ventral tegmental area, where it is cell autonomously required to antagonize identity features of the dorsal-lateral ventral tegmental area DA neurons \[[@B35]\]. Thus, contrary to *Drosophila*, depletion of Otx2 in these DA neurons does not induce cell death, but it changes neuron subtype identity. Interestingly, *otx2*expression in these DA neurons has been associated with their reduced vulnerability to Parkinsonian neurodegeneration \[[@B35]\]. Finally, in *oc*mutant adult flies most of the protocerebral bridge, a neuropile structure that is part of the central complex, is also missing \[[@B20]\]. In several behavioral paradigms, these mutant flies walk slowly and show altered orientation behavior toward visual objects \[[@B36],[@B37]\]. It has been recently proposed that the protocerebral bridge is an essential part of a visual targeting network that transmits directional clues to the motor output \[[@B37]\]. Thus, with regards to the data presented here, it would be interesting to analyze whether the lack of PPL2 DA neurons in *oc*mutant adult flies contributes to the behavioral phenotypes observed in these mutant flies. Conclusions =========== Using MARCM and twin-spot MARCM techniques together with anti-TH immunoreactivity, we have classified the 21 DA neurons present in the *Drosophila*larval central brain into seven clusters of clonally related DA neurons. The homeobox gene *otd*is specifically expressed in DA neurons belonging to one of these clusters (DL2a cluster); thus, *otd*expression differentially labels a subset of DA neurons. Furthermore, by taking advantage of an *otd*hypomorphic mutation and the *oc2-Gal4*reporter line, we have established a cell lineage relationship between the larval DL2a and the adult PPL2 DA cell clusters. We also studied the role of *otd*in the survival and/or cell fate specification of these post-mitotic neurons. Contrary to mice, where *Otx2*expression in DA neurons of the adult brain is necessary for neuron subtype identity, *otd*is required in the *Drosophila*larval and adult brain for survival of DL2a and PPL2 DA neurons. These findings suggest that *otd*acts as a post-mitotic selector gene whose differential expression among DA neurons might help to establish functional differences. Materials and methods ===================== Fly strains, clonal analysis and RNAi experiments ------------------------------------------------- Flies were reared on standard medium at 25°C. The following transgene and reporter lines were used: UAS-*P35*(Bloomington Drosophila Stock Center, Bloomington, Indiana, USA), UAS-*otd*(J Blanco, unpublished), *oc2-gal4*\[[@B28]\], *TH-gal4*\[[@B24]\]. Mutant alleles used in this study: *oc*^γa1^, *otd*^*YH13*^\[[@B38]\]. Mitotic clones were generated and positively labeled (with membrane tethered CD8::GFP and CD2::RFP) according to the MARCM \[[@B30]\] and twin-spot MARCM \[[@B33]\] techniques. Unless indicated, recombination was induced 3 to 7 hours AEL by a one hour heat shock at 37°C and the larvae were dissected 21 hours later (early L1) or 110 hours later (L3). Genotypes of the analyzed larvae were as follows: *otd*^*YH13*^MARCM clones, *w otd*^*YH13*^*FRT19A*/*w hs-FLP tubP-GAL80*^*LL1*^*FRT19A*; *tubP-GAL4UAS-mCD8::GFP*^*LL5*^/*+*; wild-type MARCM clones, *y w hs-FLP/+*; *FRT82*/*FRT82 tubP-GAL80*^*LL10*^; *tubP-GAL4*^*LL7*^(or *TH-GAL4*) *UAS-mCD8::GFP*^*LL6*^/*UAS-mCD8::GFP*^*LL6*^; wild-type twin-spot MARCM clones, *y w hs-FLP/+*; *FRT40A UAS-mCD8::GFP UAS-CD2-Mir*/*FRT40A UAS-rCD2::RFPUAS-GFP-Mir*; *tubP-GAL4*^*LL7*^(or *TH-GAL4*)/*+*. Depletion of Otd by RNAi was carried out by targeted expression of an *otd*-specific RNAi construct (VDRC-105764) in DA neurons using the *TH-gal4*driver. To increase knockdown efficiency, the experiment was done at 29°C in *otd*^*YH13*^heterozygous flies. Immunohistochemistry -------------------- Antibody staining on brains was performed as previously described \[[@B39]\]. Primary antibodies were as follows: rabbit anti-Otd (1:250) \[[@B40]\], mouse anti-TH (1:100; Chemicon, Millipore AG, Temecula, California, USA), rabbit anti-TH (1:250) \[[@B41]\], rabbit anti-RFP (1:100; Abcam, Cambridge, UK). Secondary antibodies were Alexa488-, Alexa568- and Alexa647-conjugated antibodies generated in goat (1:200; Molecular Probes, Invitrogen, Paisley, Renfrewshire, UK). Fluorescent images were captured with an Olympus FV1000 confocal laser scanning microscope and analyzed in ImageJ \[[@B42]\]. Unless otherwise indicated, pictures correspond to single optical sections (1 μm thick). Figures were assembled using Adobe Illustrator and Photoshop. Abbreviations ============= AEL: after egg laying; DA, dopaminergic; DL: dorso lateral; DM: dorso medial; FRT: flippase recognition target; GFP: green fluorescent protein; L: instar larval stage; MARCM: mosaic analysis with a repressible cell marker; NB: neuroblast; *oc*: *ocelliless*; *otd*: *orthodenticle*; PPL: protocerebral posterior lateral; RFP: red fluorescent protein; RNAi: RNA interference; TH: tyrosine hydroxylase. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= JB and RP carried out all the experiments. JB, MW and GU conceptualized the project. JB and GU wrote the manuscript. All authors read and approved the final manuscript. Supplementary Material ====================== ###### Additional file 1 **Supplementary Figure 1 - MARCM lineage analysis of late differentiating dopaminergic neurons in the larval central brain at L3**. MARCM-labeled wild-type NB clones were induced during early embryogenesis (3 to 7 h AEL) and analyzed at L3. To circumvent the down-regulation of the *tubulin*promoter, an additional copy of the *UAS-CD8::GFP*transgene was included in the genotype of the analyzed larvae. **(A-C)**The DM1a DA neuron belongs to a NB lineage independent of the DM1b DA cell lineage. **(D,E)**The late differentiating DL2 DA neuron belongs to the DL2a DA cell lineage. Scale bars: 10 μm. All panels correspond to Z projections of individual confocal optical sections. AEL, after egg laying; DA, dopaminergic; DL, dorso lateral; DM, dorso medial; GFP, green fluorescent protein; MARCM, mosaic analysis with a repressible cell marker; NB, neuroblast. ###### Click here for file Acknowledgements ================ We acknowledge B Bello, C Desplan, T Lee, B Lu, U Walldorf and the Bloomington Stock Center for kindly providing fly strains and reagents. This work was supported by the Joint Singapore Bioimaging Consortium (SBIC)-Singapore Stem Cell Consortium (SSCC).

{ "pile_set_name": "PubMed Central" }

1. Introduction {#sec1} =============== Management of patients requiring mechanical ventilation is complex and poses a substantial patient safety risk for novice practitioners. Understanding the patient-ventilator interaction is essential to minimize the risk of iatrogenic barotrauma, ventilator-associated pneumonia (VAP), and death \[[@B1]\]. A lack of dedicated formal instruction on ventilator modes and patient-ventilator interaction leads to varying levels of comfort and ability to manage these patients. National work-hour restrictions, medicolegal concerns, and the increasing complexity and volume of critical care patients in the United States significantly limit the time allowed for learning key concepts and developing autonomy with mechanical ventilator management during ICU (Intensive Care Unit) rotations \[[@B2]\]. Simulation is effective as a training methodology within medical education \[[@B3], [@B4]\], including the development and refinement of critical care skills \[[@B5], [@B6]\], invasive procedures \[[@B7]--[@B11]\], and crisis resource management \[[@B12], [@B13]\]. Residents are evaluated on their clinical development by assessing their proficiency in the aforementioned skill sets. The Accreditation Council of Graduate Medical Education (ACGME) has established milestone guidelines to help residency programs standardize performance expectations, facilitate feedback for development, and identify areas needing further guidance. The ACGME mandates the use of simulation in emergency medicine, surgery, and internal medicine training \[[@B14]--[@B16]\]. Singer et al. demonstrated that first-year residents who underwent simulation-based education outperformed traditionally taught third-year residents in critical care medicine topics \[[@B17]\]. Frengley et al. reported that teamwork in critical care teams could be improved after simulated-based training intervention \[[@B18]\]. However, one area lacking sufficient study in the medical education literature is formal training in the management of mechanically ventilated patients. Even in today\'s medicolegal environment where "learning on patients" is no longer acceptable, many intensivists report a lack of standardization to their own mechanical ventilation training. The aim of this study was to develop and pilot test an interactive mechanical ventilation boot camp curriculum for first-year residents in surgery and emergency medicine. The objective of the training was to familiarize learners with common modes of ventilation, common etiologies of ventilator alarms and subsequent management strategies, and high-risk low-frequency presentations. We hypothesize that the use of a standardized curriculum would improve cognitive knowledge, performance, and participant confidence with ventilator management. 2. Materials and Methods {#sec2} ======================== 2.1. Study Location and Equipment {#sec2.1} --------------------------------- The study was performed at a tertiary-care university-affiliated teaching hospital simulation lab during July of 2015. Human-patient simulators in ICU beds were connected to ASL 5000 Breathing Simulators (IngMar Medical). Each of the five human-patient simulators were intubated and connected to a mechanical ventilator (Covidien Puritan Bennett*™*840). ASL 5000 Lung simulators were used to adjust the pulmonary mechanics (lung compliance, airway resistance, respiratory rate, and tidal volume) in real-time based on the actions of the resident. Each bay had a simulated patient monitor that demonstrated the patients\' vital signs, portable chest X-rays, electrocardiograms, and arterial blood gas measurements (if requested), as shown in [Figure 1](#fig1){ref-type="fig"}. This was a quality assurance project that did not meet the definition of human subject research. It was exempt from institutional review board review. 2.2. Curriculum Development and Outline {#sec2.2} --------------------------------------- A three-day pilot boot camp curriculum was developed to educate first-year residents on the management of patients requiring intubation and mechanical ventilatory support. The boot camp program required approximately 65 hours of initial preparation from the simulation and ICU faculty for the development of the curriculum, surveys, questionnaires, and simulation cases with accompanying didactic postsimulation lectures. After initial curriculum development, all participating staff performed a rehearsal of the boot camp to assure that simulation cases were executed without difficulties and all necessary equipment was available and working appropriately. This required an additional 3 hours. The boot camp was scheduled for 12 hours over three days. Overall, total preparation and execution of the curriculum took approximately 80 hours. The curriculum consisted of 4 parts (as shown in*Three-Day Mechanical Ventilation Simulation Curriculum*): preintervention evaluation, independent study, the intervention phase, and postintervention evaluation. Cognitive tests, critical action checklists, and confidence surveys were used to primarily assess the residents. The entire curriculum took participants approximately twelve hours to complete. The cognitive tests, critical action checklists, and confidence surveys were identical for both pre- and postintervention evaluation. *Three-Day Mechanical Ventilation Simulation Curriculum* Day 1 (Pretesting Evaluation): Pretest confidence survey (5 minutes). Pretest cognitive multiple-choice exam (25 minutes). Cases and evaluation by critical actions checklist (10 minutes each for 30 minutes total): ARDS. Complete lung atelectasis secondary to mucus plugging. Pneumothorax in a mechanically ventilated patient. Received supplemental readings. Independent study (estimated 4 hours of reading material provided for asynchronous education). Day 2 (Curriculum and Educational Intervention): Case structure: Two-three residents participated in the case (10 minutes). Evaluation by critical actions checklist (evaluated during case). Bedside debriefing by intensivists (20 minutes). Review of provided didactic PowerPoint (15 minutes). Pathology reviewed: ARDS. Complete lung atelectasis secondary to mucus plugging. Altered mental status secondary to overdose. Pneumothorax in a mechanically ventilated patient. Dynamic hyperinflation. Day 3 (Posttesting Evaluation): Posttest confidence survey (5 minutes). Posttest cognitive multiple-choice exam (25 minutes). Cases and evaluation by critical actions checklist (ten minutes each for 30 minutes total): ARDS. Complete lung atelectasis secondary to mucus plugging. Pneumothorax in a mechanically ventilated patient. 2.3. Participants, Faculty, and Staff {#sec2.3} ------------------------------------- First-year residents from three residency programs were invited to participate in the curriculum, which was conducted in July. Six intensivists were present for debriefing based on availability, as well as one emergency medicine attending physician and two respiratory therapists from a local teaching hospital. When present, the faculty debriefers were assigned to one station for the day and did not rotate through different cases. Each station required a simulation technician for all three days of curriculum. 2.4. Preintervention Evaluation {#sec2.4} ------------------------------- The preintervention stage assessed baseline knowledge and confidence on 3 simulated high-risk mechanical ventilation scenarios, including acute respiratory distress syndrome, complete lung atelectasis secondary to obstruction from a mucus plug (hereby referred to as mucus plugging), and pneumothorax. Participants filled out a pretest 12-question confidence survey (as shown in*Confidence Survey*) and a cognitive 20-question multiple-choice test. Each station accommodated one resident at a time. Residents transitioned through the stations based on availability. One intensivist was based at a station per day. The intensivists rated participants\' performance using a predetermined checklist of critical actions. The critical actions assessed for the mucus plugging case are given herein after. Feedback was not given to participants at any time in the preintervention phase. *Confidence Survey* Four digit identifying code: Surgery or Emergency Medicinevery uncomfortablesomewhat uncomfortableneutralsomewhat comfortablevery comfortable How comfortable do you feel distinguishing between the different modes of ventilation?How comfortable do you feel initially choosing a mode of ventilation?How comfortable do you feel switching between different modes of ventilation?How comfortable do you feel weaning mechanical ventilation?How comfortable do you feel addressing an alarming ventilator?How comfortable do you feel identifying causes of an elevated peak airway pressure?How comfortable do you feel identifying causes of an elevated plateau pressure?How comfortable do you feel managing the mechanical ventilation of a patient with ARDS?How comfortable do you feel managing the mechanical ventilation of a patient with decompensated CHF?How comfortable do you feel managing the mechanical ventilation of a patient with an acute exacerbation of COPD?How comfortable do you feel managing the mechanical ventilation of a patient with an acute exacerbation of asthma?How comfortable do you feel managing the mechanical ventilation of a patient with a traumatic brain injury? *Critical Actions Checklist for Mucus Plugging Case* Identifying mucus plugging□Yes□NoIdentifying increased airway resistance with high peak pressure□Yes□NoActions (suctioning, bronchoscopy, bronchodilators)Must include either suctioning or bronchoscopyYesNoTo increase FiO~2~ to maintain oxygenation□Yes□No (5)To decrease tidal volume while increasing rate to maintain minute ventilation□Yes□No(6)Reevaluate the patient after vent changes□Yes□No(7)Documenting ventilator setting changes□Yes□NoImmediately after initial testing, residents were provided with reading material. This asynchronous material consisted of approximately four hours of reading and included selections taken from book chapters, a review article, and a primer reviewing basic mechanical ventilation principles created by one of the intensivists. 2.5. Educational Intervention {#sec2.5} ----------------------------- The intervention phase was comprised of five 45-minute scenarios and took place six days after pretesting. These scenarios included a review of the basic tenets of mechanical ventilation, acute respiratory distress syndrome (ARDS), pneumothorax, mucus plugging, and dynamic hyperinflation. Residents were divided into groups of 1 to 3 participants for each station. Participants were given a brief clinical description and expected to manage the ventilated patient for the first 10 minutes. Afterwards, intensivists had approximately 35 minutes for bedside debriefing. The debriefing included individualized assessment and feedback, summation of clinical teaching points, and review of a focused didactic PowerPoint presentation, tailored to each case. 2.6. Postintervention Evaluation {#sec2.6} -------------------------------- During the postintervention stage, participants individually underwent the same three scenarios as the preintervention stage, with faculty grading their performance using the same predetermined critical action checklists. This occurred ten days after the initial preintervention evaluation. At the conclusion of the scenarios, participants completed the postintervention cognitive multiple-choice test and confidence survey, as well as a postcurriculum survey. 2.7. Postcurriculum Survey {#sec2.7} -------------------------- A postcurriculum survey was administered soliciting feedback on areas of strength and potential improvement of the curriculum using a 5-point Likert scale. 2.8. Data Analysis {#sec2.8} ------------------ Data were first imported into SPSS v22.0 software for analysis. Test results were summarized using mean (standard deviation), median, and range values for the percentage of correct answers at the pre- and postintervention study time points. The paired change in test results was determined as the postintervention test result minus the preintervention test result and then similarly summarized. The changes in test score were tested for median equality to zero using the nonparametric Wilcoxon signed rank test. Confidence was determined via 12 assessment questions measured on a Likert 1--5 ordinal scale at the pre- and postintervention study time points. Data were summarized using the aforementioned numeric measures of center of spread and the paired change data similarly tested for median equality to zero via Wilcoxon signed rank tests. Due to the potential for type I error rate inflation, *p* values were adjusted using Bonferroni adjustments to the *p* values to protect the overall type I error rate at 5%. Finally, for 3 of the 5 simulations, critical action performance was determined by expert review at the pre- and postintervention study time points. Paired dichotomous data (Yes/No) for each critical action was summarized with two-by-two bivariate frequency tables and tested for improvement in the discordant pairs with McNemar\'s tests. The paired changes, in total number of critical actions performed, were tested for median equality to zero using Wilcoxon signed rank tests for each of the 3 cases. All statistical testing was two-sided with *p* \< 0.05 considered statistically significant. 3. Results {#sec3} ========== 3.1. Demographics {#sec3.1} ----------------- Seventeen first-year residents participated in this study. The majority (64.7%, 11/17) were emergency medicine residents, while 23.5% (4/17) were general surgery residents, and 11.8% (2/17) were urology residents. All of the residents were present for the entirety of the boot camp. 3.2. Cognitive Knowledge Assessment {#sec3.2} ----------------------------------- Cognitive knowledge between the identical pre- and postintervention multiple-choice tests increased significantly ([Figure 2](#fig2){ref-type="fig"}). The mean (SD) preintervention score was 40.3% (9.76%) with a median (range) score of 40% (25%--55%), and the postintervention mean (SD) was 67.1% (9.53%) with a median (range) score of 70% (50%--80%). The mean (SD) change in test score from before to after intervention was +26.8% (11.31%), with a median (range) score increase of 25% (5%--45%). 3.3. Clinical Performance and Critical Actions Assessment {#sec3.3} --------------------------------------------------------- The increase in critical actions performed after intervention was significantly higher in the ARDS and mucus plugging cases but not significant in the pneumothorax case as shown in [Figure 3](#fig3){ref-type="fig"}. The ARDS case had a significant increase in 4 of the 5 critical actions. Participants had a mean of 1.5 critical actions met during preintervention evaluation and 4.1 during postintervention evaluation (*p* \< 0.001). The atelectasis and mucus plugging case had a significant increase in 5 out of 7 critical actions. The mean preintervention score was 1.24 and the mean postintervention score was 4.47 (*p* \< 0.001). The pneumothorax case had no significant increase in critical actions. The mean preintervention score was 3.06 out of 7 critical actions and the mean postintervention score was 3.5 (*p* = 0.123). There was no significant increase in the final critical action of all three cases. The final critical action was documentation of ventilator setting changes. 3.4. Confidence Assessment {#sec3.4} -------------------------- Participants felt more confident with ventilator management based on their pre- and postintervention confidence surveys. This was supported by a statistically significant confidence increase for all questions, with a mean pretest score of 1.56 and a posttest score of 3.64 (*p* = 0.049). Mean confidence gain was 2.1 with a 95% confidence interval (1.6--2.6). 4. Discussion {#sec4} ============= This study demonstrates the feasibility and effectiveness of a boot camp curriculum for residents on the basics of mechanical ventilation. This curriculum resulted in increased competency, knowledge, and confidence of seventeen first-year residents. Using an integration of bedside instruction, supplemental study, and simulation-based training, residents were exposed to several pathologic conditions of mechanically ventilated patients, as well as appropriate assessment and management techniques. A comprehensive ventilator management curriculum is long overdue and serves as a critically important patient safety initiative that may prevent iatrogenic morbidity. Residents often do not have uniform training in critical care before being expected to care for this patient population. Mismanagement of ventilated patients may result in significant morbidity and mortality \[[@B1]\]. It is no longer acceptable for residents and fellows to learn at the expense of the patient\'s health, especially in critical care environments. Medical education has subsequently evolved to include simulation as a way to educate, practice, and reinforce teaching points from cases that may otherwise only be seen rarely or in critically ill patients in the clinical setting. During this curriculum, residents were exposed to several teaching methodologies to accommodate different styles of learning. Not only did residents asynchronously read about critical care topics, but they were able to interactively experience and troubleshoot high-acuity scenarios in a high-fidelity simulated environment. Residents were engaged in a supportive learning environment under direct supervision of an intensivist, who was able to provide personalized feedback and answer any immediate questions. Afterwards, they received a focused didactic PowerPoint presentation on the particular pathology they were managing. Familiarity with the equipment was improved by having the residents physically touch and manipulate the ventilators utilized in the ICU at our institution, helping to ameliorate any anxiety working with unfamiliar equipment. We postulate that these factors contributed to the statistically significant improvement in the various categories evaluated in this study. If this curriculum motivated residents to seek out additional resources for their personal development, this could not be controlled for. Alternatively, we had no objective way of measuring if residents actually completed all provided readings. Our goal was to provide a novel and structured curriculum with dedicated time to improve the foundational knowledge in a skill set that typically did not have formal time set aside upon residents matriculation into residency. We did not intend to prove that any one method, such as simulation, was the sole contributing factor to improved performance but rather that the curriculum, as a whole, provided improved foundational knowledge. The mechanically ventilated pneumothorax case was the sole scenario where there were no statistically significant increases in critical actions when analyzing pre- and postintervention evaluations. This case was unique in that definitive treatment required a procedure and ventilator management was an ancillary measure. All participants performed needle decompression or tube thoracostomy placement in the preintervention sessions, but only five participants increased FiO~2~ to maintain oxygenation, and only one increased tidal volume or rate to maintain adequate ventilation. We postulate that participants were more focused on the definitive management of tube thoracostomy versus adjusting ventilator parameters. We also speculate that medical school curricula more commonly provide instruction on management of pneumothoraces and related cardiorespiratory physiology in comparison to the topics of ARDS, atelectasis, and dynamic hyperinflation. None of the three cases demonstrated a significant improvement in documenting ventilator changes. The importance of documentation was not covered in the supplemental reading material. Any emphasis of its importance would have been dependent on bedside instruction. Formal teaching on the necessity of proper documentation and its role within a patient safety context would likely reinforce appreciation and, therefore, compliance. To date, there are very few studies in the medical education literature addressing mechanical ventilation curriculum for residents. We postulate that the reason that there are so few studies and curricula developed for simulation-based mechanical ventilation is likely due to the significant time and cost commitments required. This curriculum was financially expensive, time-consuming, and human-resource intensive. The significant financial cost arose from renting the lung simulators (approximately \$5,000 USD). Our faculty and technicians needed eight hours of training on the lung simulators and a 6--25 hours\' time commitment from each of the six different intensivists for the boot camp\'s execution. This does not include the time dedicated to the curriculum development, where a multidisciplinary team of specialists developed curricular goals and testing materials. This included determining the selected readings and critical action lists, as well as creating the cognitive multiple-choice questionnaire and confidence surveys. Deadlines for goals and objectives were established as a team and shared throughout the curriculum development over a period of approximately 8 months. This study provides a curricular framework for other educators who desire to provide formal instruction to their residents on mechanical ventilation. Residency programs with an interest to develop and execute such a curriculum will improve their chances of success by pooling time and financial resources. Faculty members should work together to decide on goals and objectives early on in the process. Additionally, identifying how to finance renting necessary simulation equipment as well as how to divide up time for curricular and teaching responsibilities is paramount. Such cooperation will make the curriculum development less tedious and increase the likelihood of successfully executing such a curriculum. This curriculum only addressed pressure and volume control ventilation to focus on modes that residents will most commonly utilize during their training at our institution. Other modes of ventilation were not reviewed in order to prevent inundating residents with information unlikely to be used clinically. The expansion of this curriculum to include additional modes of ventilation and pathologies and the development of a longitudinal curriculum throughout residency training provide further opportunities for curriculum development, patient safety training, and research. This curriculum was well received by participants as supported by feedback from a postcurriculum survey. Participants expressed that hands-on participation and bedside teaching were highly beneficial for their learning. They commented that some additional articles should be included in the reading assignment that provide a basic review of pulmonary physiology and an overview of mechanical ventilation, to establish a stronger foundational knowledge. Additionally, residents did not feel overwhelmed by the amount of material covered, as it was predetermined to only cover the two modes of ventilation. *Limitations*. This was a small prospective pilot study from a community-teaching hospital. A randomized design with a control group of traditionally trained residents compared to residents completing this curriculum would have provided a more powerful study; however, residents and resources were limited. As the boot camp becomes further integrated into the residents\' curriculum, additional medical specialties may be recruited to participate, which may also serve to provide subgroup analyses with greater power. Additionally, this study was done one time and repeating it over several years would both increase the number of residents in the study and check its reproducibility. In regard to data collection, a nonvalidated confidence survey, critical action checklists, and multiple-choice tests were utilized. We also did not survey participants to determine if they completed the provided reading materials or if they used other sources as supplementation. 5. Conclusion {#sec5} ============= A twelve-hour boot camp pilot study demonstrated significantly improved confidence, knowledge, and performance of first-year residents\' ability to manage high-risk, low-frequency scenarios of mechanically ventilated patients. This study may serve as a model for other institutions that wish to implement a mechanical ventilation simulation curriculum within their training programs. Conflict of Interests ===================== There is no conflict of interests regarding the publication of this paper. ![Mechanical ventilation training and testing environment.](CCRP2016-4670672.001){#fig1} ![Pre- and postintervention knowledge assessment.](CCRP2016-4670672.002){#fig2} ![Critical actions assessment.](CCRP2016-4670672.003){#fig3} [^1]: Academic Editor: Djillali Annane

{ "pile_set_name": "PubMed Central" }

Introduction ============ Neuroblastoma is a common extracranial solid tumor that derives from neural crest progenitor cells \[[@r1],[@r2]\]. Neuroblastoma mostly takes place in children younger than 1 year, and the average diagnosis time is about 17 months of age \[[@r3]\]. Neuroblastoma is characterized by a wide range of variable prognosis, spanning from spontaneous regression without chemotherapy to life-threatening tumor progression despite intensive treatment \[[@r4]--[@r7]\]. Approximately 50% of neuroblastomas behave in highly malignant fashion, with distant metastasis at the time of diagnosis \[[@r8],[@r9]\]. Their 5-year survival rates remain less than 40% despite intensive, multi-modal therapy \[[@r10]\]. The affects of environmental factors on the risk of neuroblastoma have been investigated but remains undefined \[[@r11],[@r12]\]. Growing evidence has been directed to the genetic factors predisposing patients to neuroblastoma. Familial neuroblastoma is largely attributed to germline mutations in *PHOX2B* \[[@r13]\] or *ALK* \[[@r14],[@r15]\] gene. In contrast, the etiology of sporadic neuroblastoma, the most common type of neuroblastoma, remains partially unveiled. Several genome-wide association studies (GWASs) and the subsequent replication studies identified a number of neuroblastoma susceptibility alleles, including *BARD1*, *LIN28B*, *HACE1*, *LMO1*, *MMP20* and *CASC15* genes \[[@r16]--[@r23]\]. Moreover, candidate gene approaches also detected the genetic associations of *NEFL* \[[@r24]\] and *CDKN1B* \[[@r25]\] gene polymorphisms with neuroblastoma susceptibility. Ran (Ras-related nuclear protein) is a small Ras-related GTP-binding protein. Ran mainly locates in the nucleus and cycles between the GDP-bound inactive and the GTP-bound active state \[[@r26]\]. It facilitates the movement of molecules in and out of the nuclear-pore complexes \[[@r27]\]. Dysregulated protein level of Ran could cause aberrant nuclear-cytoplasmic transport of tumor suppressors and oncogenes, which might lead to the initiation of cancer \[[@r28]\]. Moreover, Ran also mediates several crucial functions, such as promoting spindle assembly, regulating cell cycle, and facilitating pre-mRNA generation \[[@r29]\]. RanBP2 (Ran-binding protein 2) is the largest protein of the nuclear pore complex (350 kDa). It contains rich FG-repeats, four Ran-binding domains and binds to Ran GTP with high affinity \[[@r30]\]. RanBP2 was initially described to be implicated in regulating nuclear transport due to its linkage with Ran \[[@r31]\]. It was further identified to regulate numerous cellular activities \[[@r32]--[@r34]\]. *RAN/RANBP2* genes are reported to be associated with cancer development. However, the association of polymorphisms in the *RAN/RANBP2* genes and neuroblastoma risk has yet to be elucidated. To address this issue, we conducted a three-center case-control study in a Chinese population. RESULTS ======= Characteristics of study population ----------------------------------- The detailed characteristics of subjects from Guangzhou and Zhengzhou were provided in the previous publications \[[@r35]--[@r37]\]. The detailed demographic characteristics in neuroblastoma patients and controls for Wenzhou, Guangdong and Henan subjects were presented in [Supplementary Table 1](#SD1){ref-type="supplementary-material"}. There were no significant differences between cases and controls from Wenzhou regarding age (20.25 ± 20.73 vs. 23.58 ± 15.36 months old, *P*=0.496) and gender (*P*=1.000). *RAN/RANBP2* polymorphisms and neuroblastoma risk ------------------------------------------------- The genotype frequencies of *RAN/RANBP2* genes polymorphisms ([Supplementary Table 2](#SD1){ref-type="supplementary-material"}) and neuroblastoma susceptibility between all cases and controls were presented in [Table 1](#t1){ref-type="table"} and [Supplementary Table 3](#SD1){ref-type="supplementary-material"}. All genotype frequencies in controls were in Hardy-Weinberg equilibrium (HWE) (rs56109543, *P*=0.587; rs7132224, *P*=0.289; rs14035, *P*=0.800; rs2462788, *P*=0.194). In single locus analysis, no statistically significant association were found regarding all the four SNPs and neuroblastoma risk. We further investigated the combined effect of protective genotypes of *RAN* in neuroblastoma risk. We observed that individuals with 3 protective genotypes were at significantly lower risk of developing neuroblastoma than those without protective genotypes \[adjusted odds ratio (OR)=0.33; 95% confidence interval (CI)=0.12-0.96; *P*=0.042\]. Moreover, subjects with 3 combined risk genotypes of *RAN* have a significant decreased risk of neuroblastoma (adjusted OR=0.33; 95% CI=0.11-0.94; *P*=0.038), compared with those with 0-2 protective genotypes. ###### Association of *RAN* and *RANBP2* polymorphisms with neuroblastoma risk. ------------------------------------------------------- ------------- ------------- ----------- ---------------------- ----------- ---------------------- ----------- Genotype Cases\ Controls\ *P* ^a^ Crude OR\ *P* Adjusted OR\ *P* ^b^ (N=429) (N=884) (95% CI) (95% CI) ^b^ *RAN* rs56109543 (HWE=0.587) CC 304 (70.86) 620 (70.14) 1.00 1.00 CT 118 (27.51) 238 (26.92) 1.01 (0.78-1.31) 0.933 1.01 (0.78-1.31) 0.942 TT 7 (1.63) 26 (2.94) 0.55 (0.24-1.28) 0.165 0.55 (0.24-1.29) 0.168 Additive 0.363 0.93 (0.74-1.16) 0.504 0.93 (0.74-1.16) 0.502 Dominant 125 (29.14) 264 (29.86) 0.787 0.97 (0.75-1.24) 0.787 0.97 (0.75-1.24) 0.781 Recessive 422 (98.37) 858 (97.06) 0.155 0.55 (0.24-1.27) 0.161 0.55 (0.24-1.28) 0.164 *RAN* rs7132224 (HWE=0.289) AA 227 (52.91) 479 (54.19) 1.00 1.00 AG 170 (39.63) 335 (37.90) 1.07 (0.84-1.37) 0.581 1.07 (0.84-1.36) 0.596 GG 32 (7.46) 70 (7.92) 0.97 (0.62-1.51) 0.875 0.96 (0.62-1.51) 0.870 Additive 0.823 1.02 (0.85-1.22) 0.828 1.02 (0.85-1.22) 0.842 Dominant 202 (47.09) 405 (45.81) 0.665 1.05 (0.84-1.33) 0.665 1.05 (0.83-1.32) 0.680 Recessive 397 (92.54) 814 (92.08) 0.771 0.94 (0.61-1.45) 0.771 0.94 (0.61-1.45) 0.770 *RAN* rs14035 (HWE=0.800) CC 285 (66.43) 590 (66.74) 1.00 1.00 CT 135 (31.47) 263 (29.75) 1.06 (0.83-1.37) 0.635 1.06 (0.83-1.37) 0.641 TT 9 (2.10) 31 (3.51) 0.60 (0.28-1.28) 0.187 0.60 (0.28-1.29) 0.191 Additive 0.338 0.96 (0.78-1.19) 0.731 0.96 (0.78-1.19) 0.727 Dominant 144 (33.57) 294 (33.26) 0.912 1.01 (0.79-1.30) 0.911 1.01 (0.79-1.29) 0.918 Recessive 420 (97.90) 853 (96.49) 0.164 0.59 (0.28-1.25) 0.169 0.59 (0.28-1.26) 0.173 *RANBP2* rs2462788 (HWE=0.194) CC 402 (93.71) 810 (91.63) 1.00 1.00 CT 27 (6.29) 74 (8.37) 0.74 (0.47-1.16) 0.187 0.74 (0.47-1.16) 0.188 TT 0 (0.00) 0 (0.00) / / / / Additive 0.185 0.74 (0.47-1.16) 0.187 0.74 (0.47-1.16) 0.188 Dominant 27 (6.29) 74 (8.37) 0.185 0.74 (0.47-1.16) 0.187 0.74 (0.47-1.16) 0.188 Combined effect of protective genotypes for *RAN* ^c^ 0 394 (91.84) 814 (92.08) 0.073 ^d^ 1.00 1.00 1 26 (6.06) 38 (4.30) 1.41 (0.85-2.36) 0.186 1.41 (0.84-2.36) 0.189 2 5 (1.17) 7 (0.79) 1.48 (0.47-4.68) 0.509 1.48 (0.47-4.70) 0.507 3 4 (0.93) 25 (2.83) **0.33 (0.11-0.96)** **0.041** **0.33 (0.12-0.96)** **0.042** 0-2 425 (99.07) 859 (97.17) 1.00 1.00 3 4 (0.93) 25 (2.83) 0.028 **0.32 (0.11-0.94)** **0.037** **0.33 (0.11-0.94)** **0.038** ------------------------------------------------------- ------------- ------------- ----------- ---------------------- ----------- ---------------------- ----------- OR, odds ratio; CI, confidence interval; HWE, Hardy-Weinberg equilibrium. ^a^ *χ^2^* test for genotype distributions between neuroblastoma patients and controls. ^b^ Adjusted for age and gender. ^c^ Protective genotypes were rs56109543 TT, rs7132224 GG and rs14035 TT. ^d^ For additive model. Stratification analysis ----------------------- Stratification analysis was further adopted to assess the effects of the *RAN* polymorphisms on neuroblastoma risk among different strata ([Table 2](#t2){ref-type="table"}). However, we failed to detect significant association for any of the four polymorphisms in single locus analysis. Moreover, the cumulative effects of protective genotypes were also insignificant. ###### Stratification analysis for the association between *RAN* gene genotypes and neuroblastoma susceptibility. ----------------- ---------------- ------------------ ------------------ ---------------- ------------------ --------- ----------------------- ------------------ --------- ------ ------------------ ------- Variables rs56109543\ AOR (95% CI) ^a^ *P* ^a^ rs14035\ AOR (95% CI) ^a^ *P* ^a^ Protective genotypes\ AOR (95% CI) ^a^ *P* ^a^ (case/control) (case/control) (case/control) CC/CT TT CC/CT TT 0-2 3 Age, month ≤18 145/327 1/13 0.17 (0.02-1.32) 0.091 144/326 2/14 0.33 (0.07-1.45) 0.140 145/328 1/12 0.19 (0.02-1.45) 0.109 \>18 277/531 6/13 0.89 (0.33-2.35) 0.806 276/527 7/17 0.79 (0.32-1.92) 0.597 280/531 3/13 0.44 (0.12-1.55) 0.200 Gender Female 181/365 4/11 0.70 (0.22-2.24) 0.550 183/366 2/10 0.38 (0.08-1.77) 0.219 183/366 2/10 0.38 (0.08-1.77) 0.219 Male 241/493 3/15 0.42 (0.12-1.46) 0.170 237/487 7/21 0.68 (0.29-1.63) 0.392 242/493 2/15 0.28 (0.06-1.22) 0.090 Sites of origin Adrenal gland 163/858 1/26 0.22 (0.03-1.61) 0.134 160/853 4/31 0.71 (0.25-2.03) 0.518 164/859 0/25 / / Retroperitoneal 94/858 2/26 0.68 (0.16-2.92) 0.604 93/853 3/31 0.85 (0.25-2.83) 0.785 94/859 2/25 0.71 (0.16-3.03) 0.638 Mediastinum 119/858 4/26 1.08 (0.37-3.16) 0.887 121/853 2/31 0.46 (0.11-1.96) 0.295 121/859 2/25 0.56 (0.13-2.39) 0.432 Others 38/858 0/26 / / 38/853 0/31 / / 38/859 0/25 / / Clinical stage I+II+4s 175/858 4/26 0.73 (0.25-2.13) 0.567 176/853 3/31 0.47 (0.14-1.56) 0.217 177/859 2/25 0.38 (0.09-1.62) 0.190 III+IV 224/858 3/26 0.47 (0.14-1.59) 0.226 221/853 6/31 0.76 (0.31-1.86) 0.550 225/859 2/25 0.33 (0.08-1.39) 0.129 ----------------- ---------------- ------------------ ------------------ ---------------- ------------------ --------- ----------------------- ------------------ --------- ------ ------------------ ------- AOR, adjusted odds ratio; CI, confidence interval. ^a^ Adjusted for age and gender, omitting the corresponding stratification factor. DISCUSSION ========== In the current study, we performed the first investigation into the impact of SNPs in *RAN/RANBP2* genes on the risk of neuroblastoma in Chinese Han children. Our data revealed that the single *RAN* or *RANBP2* gene polymorphism might not be strong enough to confer the neuroblastoma susceptibility in Chinese children. However, three protective *RAN* genotypes were observed to cumulatively reduce the risk of neuroblastoma. Overexpression of Ran has been observed in several human malignancies, including lung, prostate, breast, colon cancer, and neuroblastoma \[[@r38],[@r39]\]. Conditional knockdown of *RAN* gene reduced the viability of activated K-Ras-transformed cells, through inducing S-phase arrest \[[@r40]\]. Barrès et al. found that Ran protein is highly expressed in invasive serous epithelial ovarian cancers and overexpression of Ran is associated with poor patient outcome \[[@r41]\]. They also detected that silencing Ran could impair tumor growth *in vitro* and *in vivo* \[[@r42]\]. Xia et al. showed that RNA interference-mediated knockdown of *RAN* induces aberrant mitotic formation and apoptosis in cancer cells \[[@r38]\]. Silencing *RAN* causes abnormal nucleocytoplasmic transportation of transcription factors in tumor cells \[[@r43]\]. RanBP2 protein also plays critical roles in cellular processes. Knockdown of *RANBP2* results in an aberrant metaphase, mitotic arrest in G~2~/M phase and mitotic cell death \[[@r44]\]. A study by Dawlaty et al. demonstrated that RanBP2 acts as a novel tumor suppressor in lung cancer through regulating TopoII by sumoylation \[[@r45]\]. In addition, RanBP2 hypomorphic mice are more susceptible to spontaneous and carcinogen-induced lung tumors. Consistently, two independent studies also demonstrated that RanBP2 level was downregulated in human lung cancers \[[@r46],[@r47]\]. Herein, for the first time we investigated whether *RAN/RANBP2* SNPs could contribute to the risk of neuroblastoma in Chinese children. However, our findings found no significant relationship between all the analyzed *RAN/RANBP2* polymorphisms and neuroblastoma risk. Such null relationship might be attributed to the relatively small sample size, although we tried to expand the sample by recruiting subjects from three centers. To be highlighted, a study conducted by Luo et al. explored the association between sumoylation-related genes polymorphisms and risk of gastric cancer \[[@r48]\]. They are the first group investigating the role of *RANBP2* gene polymorphism in cancer risk. Their study included 1021 gastric cancer cases and 1304 controls from Chinese population. However, they failed to obtain a significant association between *RANBP2* gene intron variant rs12614691 and gastric cancer risk. In the combined analysis of our study, subjects carrying 3 protective genotypes tend to have decreased neuroblastoma risk in comparison to those without risk genotypes or those with 0-2 protective genotypes. This phenomenon was quite biologically plausible as each single variant in each gene might not be strong enough to influence the risk of cancer. The current study was the first investigation on the association of *RAN/RANBP2* genes SNPs with neuroblastoma risk. Another merit of this study was that this is a three-center case-control study. Several limitations exist in the current study. First, because of the low incidence rate of neuroblastoma, the recruitment of eligible patients was a great challenge for us. Even though we enrolled participants from three hospitals, the sample size is still moderate. This limited sample size inevitably impaired the strength of the statistical power. Second, this study only incorporated four SNPs in the *RAN/RANBP2* genes. Future studies should investigate more potentially functional polymorphisms in *RAN/RANBP2* genes. Third, as all the participants included were of Chinese origin, conclusions should be taken with caution when extrapolated to other populations. Fourth, functional analysis is warranted to justify the described associations, which would illustrate the underlying mechanisms of how theses SNPs modify neuroblastoma susceptibility. Additionally, we only assessed the possible association of the SNPs with neuroblastoma risk. Other environmental factors, such as dietary habit, childhood exposure, and health situation, would help to provide further insight into the influence of *RAN/RANBP2* polymorphisms on neuroblastoma risk. In all, here we demonstrate that common variants at the *RAN/RANBP2* genes are associated with the risk of neuroblastoma in the Chinese children in a low-impact manner. Future larger-sample, functional studies are warranted to address the mechanism by which *RAN/RANBP2* SNPs impacts tumorigenesis of neuroblastoma. MATERIALS AND METHODS ===================== Study populations ----------------- This case-control study was conducted in three centers: Guangzhou Women and Children's Medical Center, The First Affiliated Hospital of Zhengzhou University and The Second Affiliated Hospital and Yuying Children\'s Hospital of Wenzhou Medical University. The study was approved by the Institutional Review Board of the above three hospitals. In total, 429 neuroblastoma cases and 884 controls from three centers were included in this study. To be specific, 275 cases and 531 controls were enrolled from Guangzhou \[[@r35]--[@r37]\], 118 cases and 281 controls were recruited from Zhengzhou \[[@r49],[@r50]\], and 36 cases and 72 controls were enrolled from Wenzhou ([Supplementary Table 1](#SD1){ref-type="supplementary-material"}). The recruitment period lasts from December 2007 to June 2017. All the participants' parents provided signed informed consent before the study. Selection criteria of the included participants were accessible in our previous publication \[[@r51]\]. SNP selection and genotyping ---------------------------- We chose potentially functional polymorphisms in the *RAN/RANBP2* genes from dbSNP database (<http://www.ncbi.nlm.nih.gov/>). An online tool, SNPinfo (<http://snpinfo.niehs.nih.gov/>) was used to predict the functions of SNPs. In brief, we searched the potentially functional candidate SNPs located in the 5'- flanking region, 5' untranslated region, 3' untranslated region, and exon of *RAN/RANBP2* genes. Three potentially functional SNPs in *RAN* gene (rs56109543 C\>T, rs7132224 A\>G, rs14035 C\>T) and one SNP in *RANBP2* (rs2462788 C\>T) were chosen for analysis that captured nine additional SNPs with LD\>0.8 ([Supplementary Table 2](#SD1){ref-type="supplementary-material"}). Three SNPs (rs56109543, rs7132224, rs2462788) are located in transcription factor binding sites (TFBS) and one SNP rs14035 might affect the microRNA binding site activity. As shown in [Supplementary Figure 1](#SD1){ref-type="supplementary-material"}, there was no significant LD (R^2^\<0.8) between each *RAN* SNP pair (R^2^=0.488 between rs56109543 and rs7132224, R^2^=0.582 between rs14035 and rs7132224), except for the rs56109543 and rs14035 (R^2^=0.838). The peripheral blood was used to extract genomic DNA. We genotyped the gene polymorphisms using Taqman real-time PCR \[[@r52]--[@r54]\]. On each 384-well plate, eight negative controls with water were used as quality control samples. The randomized and blinded process method was adopted to genotype all case and control samples. 10% random selection samples were re-genotyped and the genotype concordance rate was 100%. Statistical analysis -------------------- Departures from HWE for the selected SNPs in controls was evaluated using goodness-of-fit χ^2^ test. Allele frequencies and demographic variables between the two groups were assessed by chi-square test. The ORs, 95% CIs, and the corresponding *P* value for each SNP were calculated with adjustment for age and gender. Risk associations between genotypes and neuroblastoma were determined from logistic regression analysis. All calculations were performed using SAS software version 9.4 (SAS Institute, Cary, NC). All statistical tests were two-sided, and significant threshold was set using *P*\< 0.05. Supplementary Material ====================== **CONFLICTS OF INTEREST:** The authors have no competing interests to declare. **FUNDING:** This work was supported by grants from the Pearl River S&T Nova Program of Guangzhou (No: 201710010086), Scientific Research Foundation of Wenzhou (No: 2015Y0492), Zhejiang Provincial Medical and Health Science and Technology plan (No: 2009A148), and Zhejiang Provincial Science and Technology Animal Experimental Platform Project (No: 016C37113).

{ "pile_set_name": "PubMed Central" }

Introduction {#sec1_1} ============ Signet-ring cell carcinoma (SRCC) is a rare adenocarcinoma of epithelial origin, accounting for less than 0.5% of all malignancies \[[@B1]\]. It predominantly affects the stomach, with approximately a fourth of gastric cancers attributed to SRCC. However, SRCC has been known to affect other organs, such as the prostate, ovaries, pancreas, and large intestine to a much lesser extent \[[@B2], [@B3]\]. In addition, only a small number of cases have been reported in patients with diagnosed Crohn\'s disease \[[@B4]\]. It is generally known that inflammatory bowel disease (IBD) increases a patient\'s risk for developing cancer along the gastrointestinal tract (GI), though the exact mechanism is poorly understood \[[@B5]\]. Theories include chronic inflammation causing dysplasia, use of immunosuppressants and biologics, and manipulation of the gut flora \[[@B6]\]. Herein, we present a case of SRCC involving the terminal ileum with concurrent Crohn\'s flare. Case Presentation {#sec1_2} ================= A 44-year-old female with a history of Crohn\'s disease was admitted for progressive fatigue, chronic abdominal pain, and bright red blood per rectum. She was prescribed rivaroxaban approximately 2 weeks prior to this encounter for left calf tenderness determined to be secondary to a deep vein thrombosis. Physical examination on admission was notable for conjunctival pallor and right lower-quadrant tenderness on palpation of the abdomen with no distension; bowel sounds were present and normoactive. There was no appreciable fluctuation in weight in the past 12 months. Initial laboratory findings included a microcytic anemia (hemoglobin 6.6 g/dL), left-shift leukocytosis (14.3 count/μL), elevated sedimentation rate (ESR 45 mm/h), and elevated C-reactive protein (CRP 6.7 mg/L). Rivaroxaban was stopped, and the patient was transfused with packed red blood cells with a goal of hemoglobin \> 7.0. The patient was continued on 20 mg prednisone daily for Crohn\'s disease management with plans to start ustekinumab. Review of medical records showed an MRI enterography performed a month before, which suggested active Crohn\'s disease of the terminal ileum with increased dilatation of portions of the ileum proximal to the area of active disease; no fistula or abscesses were identified. In addition, a colonoscopy performed 1 year prior to this case presentation showed a stenotic but otherwise normally appearing ileocecal valve, preventing passage of the colonoscope and full visualization of the terminal ileum. Given her presentation and previous imaging findings, her symptoms were initially attributed to a lower GI bleed in the setting of rivaroxaban use. The patient had persistent anemia despite red blood cell transfusions prompting repeat colonoscopy. On visualization of the proximal cecum and terminal ileum, an ulcerating, partially obstructing, mass-like process was identified (Fig. [1](#F1){ref-type="fig"}). The mass was biopsied, which revealed SRCC with mucinous features. Additional laboratory tests included a carcinoembryonic antigen, which was within normal limits (CEA 1.8). Computed tomography of the chest, abdomen, and pelvis revealed no evidence of metastatic disease; radiographic measurement of the mass revealed approximately 8 cm in length at the distal ileum. The patient subsequently underwent laparoscopic-assisted right hemicolectomy with ileostomy, with the excised sample sent for gross pathology (Fig. [2a](#F2){ref-type="fig"}). Final pathological diagnosis was poorly-differentiated signet-ring cell adenocarcinoma (T4aN1) of the terminal ileum, ileocecal valve, and cecum (Fig. [2b](#F2){ref-type="fig"}). Molecular diagnostic was negative for *RAS* and *BRAF* genes. The patient was seen by medical oncology as an outpatient and was started on folinic acid, fluorouracil, and oxaliplatin (FOLFOX) adjuvant therapy. Discussion {#sec1_3} ========== Cancer of the small intestine is a rare diagnosis. However, in patients with Crohn\'s disease, the relative risk of having a malignancy of the small intestine increases up to 30 times compared to the general population. Undoubtedly, abnormal imaging and symptom presentations should raise the specter of malignancy in individuals with IBD if the clinical context is ripe. In our patient\'s case, it was difficult to distinguish between malignancy and inflammation. MRI enterography has been shown to be useful in the diagnosis and management of IBD, but its role in deciphering between inflammatory flares in IBD and new malignancies continue to evolve \[[@B7], [@B8]\]. Although it is consensual knowledge that IBD increases the risk of GI malignancies, a proper modality or algorithm of screening or follow-up is not present in standard of care guidelines. In addition, our patient had a previous colonoscopy, which showed abnormalities of the ileocecal valve most consistent with stricturing Crohn\'s disease. As mentioned above, inflammatory changes can be observed, but in our case, there was a quick development of malignancy between colonoscopy procedures. Such a rapidly developing entity is challenging to anticipate even with frequent procedures \[[@B9]\]. Given the various aggressiveness of GI malignancies that may develop in the setting of IBD, further investigation into improving the detection of malignancies early, either through more specific biomarker surveillance or increased sites of biopsy, is warranted. Additionally, our patient was treated with the appropriate medical therapy for her Crohn\'s disease. Our team questioned whether the number of inflammatory flares predisposed her for her diagnosed malignancy \[[@B10]\]. Certainly, SRCC of the small intestine carries a rare diagnosis. In all, the current literature present neither patient studies of long-term medical therapies nor the impact of other comorbidities to prevent the development of malignancies. Though there are studies of chemoprevention of malignancy in Crohn\'s disease, long-term data are limited, and further investigation is needed \[[@B11]\]. Statement of Ethics {#sec1_4} =================== The patient was contacted after her hospital stay. She was told of the team\'s desire to publish her medical case in an academic journal. She was told that all information would be anonymized. She agreed to have her medical case be published. All authors listed were involved in the patient\'s care. The authors have no ethical conflicts to disclose. Disclosure Statement {#sec1_5} ==================== The authors have no conflicts of interest to declare. ![The gross specimen displays an endophytic, ulcerating mass involving the terminal ileum, ileocecal valve, and cecum.](crg-0013-0085-g01){#F1} ![**a** Gross pathology of the resected bowel. The asterisk shows the endophytic mass at the distal ileum. **b** Tumor cells demonstrate signet-ring cell morphology with eccentrically displaced nuclei and abundant vacuolated cytoplasm. HE. ×200.](crg-0013-0085-g02){#F2}

{ "pile_set_name": "PubMed Central" }

Introduction {#s1} ============ Humans are very efficient at processing quantities. Assessing and evaluating the number of desirable or potentially dangerous entities---chunks of food, precious artifacts, or deadly animals---happens in fractions of seconds. Importantly, we are also able to share these rapid impressions with others: Concrete quantities or abstract symbolic representations can be translated into number words (cf. Dehaene et al., [@B4]) or, similarly, other number-describing expressions, so-called quantifiers (e.g., *"many,"* "*at least five*," etc.). Quantifiers can take the form of different semantic classes. On the one hand, there are cardinal quantifiers ("*three* bottles") which contain an explicit numeric expression (in our example: three). There are also other classes of quantifiers that refer to quantity with an implicit notion of quantity that do not refer to an exact number or numeric interval. For instance, Aristotelian quantifiers (comprising existential quantifiers such as "*all* X are Y" and logical quantifiers like e.g., "*some* X are Y") or majority quantifiers ("*at least half* of the X are Y") refer to quantities in the environment in the absence of an explicitly defined number. Items in the environment may be defined by the total size of a set: If there are 12 oranges, for example, "*at least half* of the oranges" means more than 6. However, in other cases, the criterion may be defined more vaguely. For instance, "*much* chocolate" may mean one bar or five bars, depending on the individual\'s preference. Likewise, it may depend on the semantic context: "*many* lions" could be five, but "*many* ants" probably means hundreds or thousands. We focus our study on quantifiers because we can easily and precisely measure the meaning of this class of word, even one that does not explicitly mention a number. This is because the meaning of a quantifier is derived in part from number knowledge---to varying degrees depending on the exact nature of the quantifier. Majority quantifiers refer to a subset of items that require a computation based on an understanding of the entire set of items and thus tend to be more difficult than cardinal quantifiers where an explicit number is mentioned. For instance, children use number expressions much earlier, and far more correctly, than quantifiers (Hurewitz et al., [@B11]; Sullivan and Barner, [@B23]). In particular, the appropriate use of quantifiers is not always easy. Consider the sentence "All shirts are blue." If this sentence is true, then the sentence "Some shirts are blue" is, logically, also true (cf. Geurts, [@B5]). This phenomenon is called scalar implicature, indicating that if the truth value for a set of elements is 1, then it is also 1 for any given subset. However, as Grice ([@B6]) pointed out, this is not how we use quantifier expressions in every-day conversation. Rather, we use "conversational implicatures," trying to be as relevant as possible (for a discussion see Haugh, [@B7]). In the given example, this would mean that we distinguish between *"some"* and *"all"* such that *"some"* means "only *some*, not *all*." This Gricean use of quantifiers can be demonstrated empirically in the laboratory: Different quantifiers are used to refer to different ranges of numerosities (e.g., Oaksford et al., [@B17]). For a given set of objects, "none" refers to the lower end of the scale, followed by *"few," "some," "some not," "most,"* and *"all"* in ascending order. This order was reliably observed with different tasks, indicating that a given quantifier is most appropriate, and thus most informative, for a certain fraction of the overall numerosity. However, the distinction between scopes of quantifiers is not completely clear-cut. Rather, scopes of quantifiers tend to overlap partially (e.g., Holyoak and Glass, [@B10]; Oaksford et al., [@B17]). This may lead to situations in which two quantifiers are (nearly) equally appropriate for the description of a given amount of entities. For instance, in the Oaksford et al. ([@B17]) study, participants were presented with images containing varying amounts of black or white squares, together with a statement like "*Most* of the squares are black." Their task was to decide whether the statement was an appropriate description of the visual scenario. When there were about 18% of black squares on the screen, subjects accepted the statement "*Few* of the squares are black" in about 75% of the cases. Interestingly, when presented with the same scenario but the sentence "Some of the squares are black," they would accept the statement in exactly the same proportion of cases. Compatible results were found in a memory experiment by Holyoak and Glass ([@B10]) who demonstrated that quantifiers like *"many"* and *"a few"* might be confused in memory in some proportion of cases, the chance being higher as the quantities and the referring quantifiers became more similar. To summarize, the evidence suggests that the selection of quantifiers for every-day use seems to follow the Gricean principle of relevance, maximizing the information expressed by this quantifier. Such specific selection is possible because each quantifier has a particular scope, i.e., it refers to a certain interval of probabilities distributed around one criterion value which best represents the semantics of the quantifier. These scopes overlap at least in part. Given this overlap of scopes and the fact that quantifiers may be exchanged under some circumstances, a question arises regarding the invariance of the criteria to which quantifiers refer in an individual. Given the close relationship between processing of numbers and numerosities on the one hand and quantifiers on the other hand, findings from studies of number processing may shed some light on this question. Both behavioral and neurophysiological studies (e.g., Nieder and Miller, [@B16]; Piazza and Izard, [@B18]) indicate that processing of numerosities is not an all-or-nothing phenomenon. Rather, even if a presented number of elements does not exactly match a target numerosity, neural assemblies may also fire to these if they are close enough to the target numerosity, resulting in the subject to show overt behavioral responses to this slightly deviant numerosity. For instance, Piazza and Izard ([@B18]) summarize a series of studies in monkeys and human subjects who performed same-different judgment tasks on stimuli with dot patterns. Intracortical recordings in parietal neurons in monkeys showed the firing pattern described above, which nicely corresponded to the distribution of the percentage of "same" responses in the monkeys. In human subjects, fMRI was used instead of intracortical recordings, yielding again a nice match between button press responses and the shape of the BOLD response in the left parietal cortex around the intraparietal sulcus. Most interestingly, these behavioral responses are not invariant. In his seminal work, Helson ([@B9]) demonstrated that the repeated exposure to a certain stimulus intensity, or numerosity, may result in a change of what he called "adaptation level." Responses to the former target intensities, or numerosities, are thus altered, leading to a so-called peak shift in the response curves. The maximum probability of a YES response is moved toward a new criterion. Following the argument above that the processing of quantifiers and numbers bears some similarities, one might assume that such a peak shift phenomenon can occur not only for number processing, but also for quantifier processing. In other words, some external learning influence is likely to change the internal criterion to which a quantifier refers, resulting in a shift of the scope of this quantifier. It is yet an open question whether such a phenomenon may occur. If so, another question follows: Does the scope of this one quantifier shift in isolation, or does it also affect the scope of other quantifiers? In particular, a change in the criterion for "many" affect the mental representation of the related quantifier "few?" If true, this would provide some support for the claim that quantifiers like "many" and "few" are linked in their representations in semantic memory even though these terms refer to different quantities (cf. Routh, [@B20]). The present study was thus conducted in order to address these two questions. For this purpose, we used the semantic truth value judgment (STVJ) task by Heim et al. ([@B8]) and combined it with explicit feedback trials as e.g., used by McMillan et al. ([@B13]). As a result, we were able to assess the initial criteria to which the quantifiers *"many"* and *"few"* refer, to train subjects to change their internal criteria for "many-ness" (Experiment 1) or "few-ness" (Experiment 2) independently from each other, and to test whether explicitly altering the one criterion (e.g., for *"many"*) would also affect criterion for the other, untrained quantifier (in this case, for *"few"*). We used the quantifiers *"many"* and *"few"* because they are familiar, they do not differ in processing requirements (see above), and at the same time they are sufficiently distant from each other (Routh, [@B20]) to be not mistaken too easily. Experiment 1 ============ The purpose of Experiment 1 was to assess the subjects\' initial criteria for calling a certain amount of circles of a given color *"many"* or *"few"* (Block 1), then to shift the criterion for *"many"* explicitly to a lower proportion of circles (Block 2), and finally to test whether this shift also affected the criterion for *"few"* even though the criterion for this quantifier was not mentioned (Block 3). All subjects participated in an informed consent procedure that was approved by an Institutional Review Board convened at the University of Pennsylvania. Methods ------- ### Participants Twenty-one healthy volunteers (mean age 22.7 years, range 19--29 years; 13 women; average amount of education: 15.8 years) participated in this study. They were recruited at the University of Pennsylvania and received \$10 for their participation. ### Procedure Subjects were presented with pictures showing isoluminant blue and yellow circles on a gray background. The stimuli were taken from the study by Heim et al. ([@B8]). Each stimulus item included a picture containing 50 circles, with proportions of yellow circles ranging from 20/30/40/50/60/70% of the total number of circles. The remaining circles were blue. There were six different stimuli for each proportion. In order to minimize learning effects due to repetition of stimuli, each stimulus was presented in three different rotations (0°, 90°, 180°), a procedure that had previously been shown to reduce repetition memory effects for stimuli (Lassaline and Logan, [@B12]). Each picture was presented together with the sentence "*Many* of the circles are yellow" or "*Few* of the circles are yellow." The subjects performed a truth value judgment task, indicating whether they thought the sentence adequately described the picture or not by pressing the left or right response button[^1^](#fn0001){ref-type="fn"}. The experiment consisted of three blocks (cf. Figure [1](#F1){ref-type="fig"}). The first block served as a baseline block, in which the subjects\' internal criteria for *"many"* and *"few"* were assessed. To this end, subjects performed the STVJ task in a total of 378 trials counterbalanced for "few" and "many" trials along with the proportion of yellow circles. After this baseline block, an adaptation block followed in which subjects received feedback for their response. In this block, we provided feedback and only stimuli with the quantifier *"many"* were presented. Positive feedback was given if they responded YES to pictures containing 40% or more of yellow circles, or NO to pictures with less than 40% of yellow circles. Negative feedback was given if they responded NO to pictures with 40% or more, or YES to less than 40% yellow circles. Based on previous data (Heim et al., [@B8]), this feedback would effectively move the internal criterion from a presumed baseline of 50-60% down to 40%. For this adaptation block for *"many,"* proportions of 20--70% yellow circles were used. ![**Schematic of the trials in blocks 1 and 3 (top) and in the adaptation block 2 (bottom)**.](fpsyg-06-00441-g0001){#F1} The adaptation block consisted of 162 trials, which were organized to form three sub-blocks with equal amounts of trials of each proportion. Sub-dividing the adaptation block into these three homogeneous sub-blocks allowed us to assess the learning curve over the adaptation block with more precision by allowing us to analyze each sub-block independently. Proportions were distributed equally over the first, second, and third part of the adaptation phase, thus creating three parallel sub-blocks of 54 trials each (9 trials for each of the 6 proportions). After the adaptation block, the test block was administered. This was identical to the baseline block, i.e., sentences containing *"many"* or *"few"* were presented, and no feedback was given. The test block was run to assess whether and how the internal criterion for *"many"* had been changed during the explicit training block for *"many,"* and whether the criterion for *"few"* had been altered in parallel even though *"few"* had never been presented in the adaptation block. Prior to the real experiment, all subjects were familiarized with the stimuli and task in two short blocks, one without feedback that resembled the baseline or test phase, and one with feedback as in the adaptation phase. The timing of a baseline/test trial is illustrated in Figure [1A](#F1){ref-type="fig"}. First, a written sentence ("Many of the circles are yellow" or "Few of the circles are yellow") was shown in the upper third of the screen for 3 s. The sentence stayed on for another 1.5 s while the stimulus picture containing yellow and blue circles appeared at the center of the screen for the same amount of time. The words "YES" and "NO" were written in the lower third of the screen to indicate which button to push for which decision. Finally, the screen went blank for another 1.5 s before the next trial began. Responses were recorded from the onset of the picture and through the entire blank period. The trials in the adaptation phase are illustrated in Figure [1B](#F1){ref-type="fig"} and were similar to those in the baseline/test phase, with the exception that feedback was given in these trials. If the subjects responded correctly, they received 10 points which were added to their score. False responses were penalized by deducting 10 points from their score. The feedback screen consisted either of a green arrow pointing upward and the information "+10" in green font, or a downward red arrow and the text "−10" in red font. The initial score was "+100." The actual score was always presented in the middle of the very upper part of the screen in black font. The adaptation phase consisted of 162 trials. Proportions were distributed equally over the first, second, and third portions of the adaptation phase, thus creating three parallel sub-blocks of 54 trials each (9 trials for each of the 6 proportions). These sub-blocks could then be compared in order to test how quickly the subjects learned. ### Data analysis Because of the balanced assignment of YES/NO responses to the left/right response button, the subjects\' button press responses were first recoded into "1" (YES) or "0" (NO) to make data sets comparable. Next, these acceptability ratings were aggregated per subject, experimental block, quantifier, and proportion of circles of the target color. Consequently, a 2 × 2 × 6 ANOVA with factors BLOCK (baseline/test), QUANTIFIER (many/few), and PROPORTION (20/30/40/50/60/70) was run to test for differential learning effects for the trained vs. untrained quantifier at trained vs. untrained proportions. Subsequently, planned contrasts (paired *t*-tests) were computed to compare the change in acceptability ratings at the critical proportion "40%" for the trained and the untrained quantifier. In a second analysis, we also tested how quickly the subjects learned during the adaptation phase. To this end, data of the adaptation block were collapsed over subject, sub-block, quantifier, and proportion. Paired *t*-tests (sub-blocks 1 vs. 2, 2 vs. 3, and 1 vs. 3) were run to compare the acceptability ratings at the critical proportion "40%." We used one-tailed *t*-tests because of our a priori predictions. Results ------- ### Adaptation effects The 2 × 2 × 6 ANOVA yielded significant main effects for BLOCK \[*F*~(1,\ 20)~ = 28.69; *p* \< 0.001\] and PROPORTION \[*F*~(5,\ 16)~ = 5.50; *p* = 0.004\] but not for QUANTIFIER \[*F*~(1,\ 20)~ \< 1\]. Moreover, all two- and three-way interactions were significant \[QUANTIFIER × BLOCK: *F*~(1,\ 20)~ = 34.39; *p* \< 0.001; QUANTIFIER × PROPORTION: *F*~(5,\ 16)~ = 1362.62; *p* \< 0.001; BLOCK × PROPORTION: *F*~(5,\ 16)~ = 4.24; *p* = 0.012; BLOCK × QUANTIFIER × PROPORTION: *F*~(5,\ 16)~ = 9.73; *p* \< 0.001\]. These effects demonstrate a strong adaptation effect for the trained quantifier *"many,"* and most importantly, we also observed a transfer effect for the not trained quantifier *"few"* (Figure [2A](#F2){ref-type="fig"}). ![**(A)**. Average acceptability ratings for a given proportion of circles of the mentioned color, plotted separately for the quantifiers *"many"* (black lines) and *"few"* (gray lines) in the baseline blocks (dashed lines) and the test blocks after adaptation (solid lines). **(B)** Average acceptability ratings for a critical proportion of circles of the mentioned color, plotted separately for *"many"* (black bars) and *"few"* (gray bars) in the baseline blocks (dashed bars) and the test blocks after adaptation (solid bars). ^\*^*p* \< 0.05, ^\*\*\*^*p* \< 0.001.](fpsyg-06-00441-g0002){#F2} Testing the adaptation effect for *"many"* and *"few"* at the critical proportion "40%" illustrates this overall effect (Figure [2B](#F2){ref-type="fig"}). We found a highly significant increase in acceptability for *"many"* \[*t*~(20)~ = −7.79; *p* \< 0.001\], and a decrease for *"few"* \[*t*~(20)~ = 2.58; *p* = 0.018\] (one-tailed *t*-tests, *p*-Bonferroni-corrected). ### Learning during the adaptation phase Testing the learning effect for *"many"* in block 2 yielded on-going adaptation (Figure [3](#F3){ref-type="fig"}). This effect was significant from sub-block 1 to sub-block 3 \[*t*~(20)~ = −7.75; *p* \< 0.001\] as a result of increases from sub-block 1 to sub-block 2 \[*t*~(20)~ = −5.03; *p* \< 0.001\] and from sub-block 2 to sub-block 3 \[*t*~(20)~ = −1.83; *p* = 0.042; all tests one-tailed, uncorrected for multiple comparisons\]. ![**Learning curves for the new meaning of *"many"* during the adaptation phase (block 2) as a function of percent circles of the target color, plotted separately for each of the three sub-blocks**. The learning effect at the trained proportion "40" is visible.](fpsyg-06-00441-g0003){#F3} Discussion ---------- In this study we investigated whether individuals can flexibly adapt their interpretation of quantifier meaning. Moreover, we tested whether such change could affect the criterion of an untrained quantifier as well, thus indicating a shift in the entire semantic reference frame for numerosities. For both research questions we obtained positive evidence. Subjects were able to shift their internal criterion for *"many"* toward the new reference criterion of "40%" following explicit training. A closer examination of the dynamics of this change (Figure [3](#F3){ref-type="fig"}) revealed that learning actually happened quite rapidly, with substantial changes in the acceptability ratings from the first third to the second third of the adaptation phase. Most interestingly, the comparison between baseline and test phase, which were absolutely identical with respect to quantifiers, stimuli, and their pairings, revealed that the internal criterion for the second quantifier (*"few"*) had also shifted although it had not been mentioned in the adaptation phase. These findings have interesting implications. They suggest that contextual influence such as explicit reinforcement can be a driving force to shift the quantifier\'s scope (cf. Price et al., [@B19] for an extensive discussion of feedback mechanisms during the estimation of quantities). Moreover, there is a change in the reference of the related quantifier "few," even though this was not explicitly mentioned during the adaptation phase. This suggests that subjects\' learning was not limited to the superficial mapping of a word to a stimulus, but that the concept underlying this continuum of quantifier word meaning was changed. The data are consistent with the interpretation that the contextual appropriateness of a quantifier (such as the untrained *"few"* in the present study) depends on its relative position in the continuum of reference criteria: If a proportion of 40% circles of a given color is best described as *"many,"* the initial tendency of a subject to call 40% *"few"* cannot be maintained. If it were maintained, the Gricean principle of maximized information would be violated because two different quantifiers that are partly exclusive and that are on different positions on the continuum of reference criteria (Oaksford et al., [@B17]) would describe the same semantic reality. Yet, the interpretation of this data set is limited in two ways. First, a direct shift for the trained quantifier and a transfer for the untrained quantifier were only observed for a decrease of the criterion (from 50 to 40%), i.e., leftwards on the continuum of reference criteria. Second, the trained quantifier *"many"* was positive. From earlier studies (e.g., Heim et al., [@B8]) it is known that positive quantifiers such as *"many,"* "most," or "more than half" are easier to process than their negative counterparts, e.g., *"few"* or "less than half." In line with cognitive load theories (e.g., Bannert, [@B1]) one might assume that the learning transfer observed in Experiment 1 would diminish if the cognitive load associated with the quantifier would be higher, as it would be in the case of the negative quantifier *"few."* This is because there may be reduced potential mental flexibility associated with a word requiring increased cognitive load during processing. Therefore, we conducted Experiment 2 in which the setting was changed accordingly: Here we explicitly trained *"few,"* and then examined whether a transfer effect would be observed for *"many."* Second, since the acceptability ratings for *"few"* in the baseline block of Experiment 1 indicate ceiling effects at low proportions, the criterion shift was now rightwards, i.e., toward the higher reference criteria 50 and 60%. Experiment 2 ============ Thus, the questions in Experiment 2 were the following. (1) Can the criterion of a negative quantifier like *"few"* (the scope of which is on the mental number line below the criterion, e.g., "*less than half*" or "*fewer than five,"* not above as in the case of positive quantifiers such as "*many"* or "*more than half"*) be changed in a manner that is analogous to that of *"many"* in Experiment 1? (2) Is there a generalization of this learning effect to the untrained quantifier *"many"*? Again, all subjects participated in an informed consent procedure that was approved by an Institutional Review Board convened at the University of Pennsylvania. Methods ------- The same subjects as in Experiment 1 completed another experiment that was almost identical to Experiment 1. The only difference was the following: In block 2 of Experiment 2, the subjects were now trained to call proportions of 50% and 60% of circles *"few."* YES responses were encouraged for 60% or less yellow circles, while NO responses were reinforced for proportions above 60% yellow circles. In Experiment 2, the proportion ranged between 30 and 80% of circles of the target color. The order of Experiments 1 and 2 were counterbalanced within a single experimental session such that potential carry-over effects from one experiment to the other could be controlled (10 subjects started with Experiment 1, 11 started with Experiment 2). Analyses were analogous to those in Experiment 1. Results ------- ### Adaptation effect The 2 × 2 × 6 ANOVA yielded significant main effects for BLOCK \[*F*~(1,\ 20)~ = 5.82; *p* = 0.026\] and QUANTIFIER \[*F*~(5,\ 16)~ = 78.67; *p* \< 0.001\] and a trend for PROPORTION \[*F*~(1,\ 20)~ = 2.71; *p* = 0.059\]. Moreover, the following interaction terms were significant (QUANTIFIER × BLOCK: *F*~(1,\ 20)~ = 25.61; *p* \< 0.001; QUANTIFIER × PROPORTION: *F*~(5,\ 16)~ = 763.71; *p* \< 0.001; BLOCK × QUANTIFIER × PROPORTION: *F*~(5,\ 16)~ = 6.92; *p* = 0.001\]. The interaction term for BLOCK × PROPORTION was marginally significant \[*F*~(5,\ 16)~ = 2.43; *p* = 0.081\]. These effects describe a strong adaptation effect for the trained quantifier *"few"* and a weaker transfer effect for the not trained quantifier *"many"* (Figure [4A](#F4){ref-type="fig"}). ![**(A)** Average acceptability ratings for a given proportion of circles of the mentioned color, plotted separately for the quantifiers *"many"* (black lines) and *"few"* (gray lines) in the baseline blocks (dashed lines) and the test blocks after adaptation (solid lines). **(B)** Average acceptability ratings for a critical proportion of circles of the mentioned color, plotted separately for *"many"* (black bars) and *"few"* (gray bars) in the baseline blocks (dashed bars) and the test blocks after adaptation (solid bars). ^\*\*^*p* \< 0.01, ^\*\*\*^*p* \< 0.001.](fpsyg-06-00441-g0004){#F4} Testing the learning/transfer effect for *"many"* and *"few"* at the critical proportion "50%" pinpoints this overall effect (Figure [4B](#F4){ref-type="fig"}), showing a highly significant increase in acceptability for *"few"* \[*t*~(20)~ = −5.49; *p* \< 0.001\] and a significant decrease for *"many"* \[*t*~(20)~ = 3.10; *p* = 0.006\] (one-tailed *t*-tests, *p*-Bonferroni-corrected). At proportion "60%," there was again a significant learning effect for *"few"* \[*t*~(20)~ = −2.90; *p* = 0.009\] but no effect for *"many"* \[*t*~(20)~ = 0.84; *p* = 0.411\]. ### Learning during the adaptation phase Testing the learning effect for *"few"* in the adaptation block yielded on-going learning (Figure [5](#F5){ref-type="fig"}). This effect was significant from sub-block 1 to sub-block 3 \[*t*~(20)~ = −3.57; *p* = 0.001\] as a result of a significant increase from sub-block 1 to sub-block 2 \[*t*~(20)~ = −2.55; *p* = 0.005\] and a strong trend from sub-block 2 to sub-block 3 \[*t*~(20)~ = −1.67; *p* = 0.056; all tests one-tailed, uncorrected for multiple comparisons\]. ![**Learning curves for the new meaning of *"few"* during the adaptation phase (block 2), plotted separately for each of the three sub-blocks**. Learning effects for proportions 50 and 60 are visible.](fpsyg-06-00441-g0005){#F5} ### Stronger adaptation effect for "many" (Experiment 1) than for "few" (Experiment 2) Next, in order to test whether the direct adaptation effect for *"few"* observed in Experiment 2 was equal in size to that for *"many"* in Experiment 1, we tested across experiments for the interaction of BLOCK and trained QUANTIFIER at the respective critical proportions "40%" (for Experiment 1) and "50%" (for Experiment 2). This interaction effect proved to be significant \[*F*~(1,\ 20)~ = 4.44; *p* = 0.048\], indicating that the adaptation effect for *"many"* was larger in size than that for *"few."* ### No effect of the order of Experiments 1 and 2 Finally, in order to test whether the order in which the subjects underwent the two experiments had any systematic influence on the data pattern, we re-ran this analysis adding ORDER as a between-subject factor. Neither the main effect of ORDER \[*F*~(1,\ 19)~ = 1.10; *p* = 0.308) nor any 2-way or 3-way interaction with ORDER was significant (all *p* \> 0.05). Discussion ---------- In Experiment 2, we tested whether the pattern of results from Experiment 1 would be mirrored in adaptation effects for the negative quantifier *"few"* with a generalization to the untrained quantifier *"many."* The findings were straight-forward. There was a clear learning effect for few at both trained proportions "50%" and "60%" as well as a transfer to the criterion for *"many"* at "50%." Moreover, the in-depth analysis of the learning curves in the three parts of the adaptation block revealed again relatively rapid learning of the new criterion from the first-third to the second-third of the adaptation phase. These data corroborate the initial findings and conclusions, i.e., that a change in meaning of a quantifier may be contextually induced and that this change in meaning may also affect related quantifiers even if they are not exact opposites or antonyms. The fact that the same subjects underwent both experiments precluded problems of between-group comparisons; the fact that the order of experiments was counterbalanced over subject ensured that the adaptation and transfer effects observed in Experiment 2 were not just a resetting to the subjects\' original criteria temporally shifted for the duration of Experiment 1. Still, we found that the training for *"many"* elicited a stronger adaptation effect than that for *"few."* The implications of the overall pattern of results will now be discussed in more detail. General discussion ================== In this paper we reported two experiments that tested whether the meaning of a quantifier such as *"many"* or *"few"* can be altered, and whether this would affect the meaning of another, untrained quantifier as well. The findings were straight-forward, with effects for explicit training as well as transfer effects to the untrained quantifier for both quantifiers. These findings emphasize the flexible representations of conceptual knowledge. These data have several implications for the field of quantification and reasoning. For one, they suggest a certain flexibility of our interpretation of quantifiers that depends in part on the context. As in the introductory example, *"many"* can imply different amounts of e.g., animals depending on whether they are rare ("*many* lions" = 10) or not ("*many* ants" = 10,000), i.e., the semantic context in which it is used. As Sanford and Moxey (e.g., in their [@B21] paper) point out, quantifier use may reflect the individual perspective on, or interpretation of, a quantity rather than an inflexible, fixed characterization of its magnitude. This is in line with the finding that quantifier scopes partly overlap, as could be shown by Holyoak and Glass ([@B10]). These authors asked subjects to provide not only their first choice of an appropriate quantifier but also their second choice. They found for this latter task preference for quantifiers with closest distance on the continuum (all---many---some---a few---none). However, the notion of a flexible, partly subjective perspective on quantifier meaning may not be true for quantifiers in general. According to Routh ([@B20]), one way of grouping quantifiers is with respect to the degree of variability of the criterion they refer to. Routh distinguished between "fuzzier quantifiers" such as *"few," "several,"* or *"many"* that are determined in part by context, Aristotelean or logical quantifiers that are "more precise" such as *"all," "none,"* or *"each,"* and parity or majority quantifiers that are "comparative" like *"more"* or *"fewer."* This distinction is based in part on the flexibility of the criteria for the referents of a given quantifier. Whereas "many donuts" may be two for person A and five for person B, comparative quantifiers imply at least the imparity of sets, i.e., ordinal information about which set is bigger. Most objective are quantifiers for which the criterion is explicitly or implicitly named, as in the case of "at least 13" or "more than half" (the latter referring to an amount of more than 50% of the total set)---objective in the sense that they refer to an explicitly stated criterion or degree. Routh ([@B20]) presented evidence from clustering algorithms and multidimensional scaling that this classification of quantifiers has psychological and empirical reality and is thus effective in communication. Given this distinction, one could formulate the hypothesis that the observed changes for quantifier semantics in the present study might generalize to other "fuzzy" quantifiers such as "some," "a few," "several," "quite a few," etc. Interestingly, within the class of "fuzzy quantifiers" like the majority quantifiers *"many"* and *"few"* tested in the present study, there seem to be differences between individual quantifiers. Whereas we did observe adaptation of the meaning of both quantifiers, we also found that this adaptation effect was even more pronounced for *"many"* than for *"few."* This result corresponds to earlier reports in the literature (e.g., Routh, [@B20]; Geurts, [@B5]; Heim et al., [@B8]; Shikhare et al., [@B22]) that quantifiers with negative polarity are more difficult to process. There are various accounts for this effect. One stresses that the positive quantifiers are (unmarked) defaults whereas negative quantifiers are marked (Clark, [@B2]; Clark and Chase, [@B3]). Alternatively, it has been proposed that negative quantifiers refer more or less precisely to the complement of the set denoted by a positive quantifier. Thus, in order to obtain the correct representation of a negative quantifier, the expression about the original set has to be negated, implying extra processing costs[^2^](#fn0002){ref-type="fn"}. Much more work is required here. Still, in the context of the present study, it could well be that negation or markedness pose greater cognitive demands on the comparison of a negative polarity quantifier like *"few,"* resulting in gradually smaller flexibility and thus a smaller adaptation effect. If the meaning of a "fuzzy quantifier" can be changed, the question arises in how far such flexibility can also be found for quantifiers that are less fuzzy, i.e., that have fixed criteria. While the current study suggests that quantity can modulate quantifier interpretation, there is also additional evidence suggesting that object size, or mass, can modulate quantifier interpretation (McMillan et al., [@B15]). The use of quantifiers with different degrees of fuzziness in one experiment may restrict the flexibility of e.g., many and few that was observed here. Alternatively, it could also be the case that shifts along the continuum of criteria might even affect quantifiers with a fixed criterion: For instance, in the case of "at least X," subjects might be more inclined to accept such a statement when the distance of the actual numerosity to the criterion increases (cf. Shikhare et al., [@B22], for first results on distance effects related to "at least" and "at most"). Our findings address the question of the flexibility of word meaning, such as deciding when the diameter of a cup becomes broad enough to be called a bowl. This has been a challenging issue to address because of the difficulty quantifying the associated object concepts. Here we addressed this issue by examining quantifiers, a class of words that is based in part on number knowledge. We found that word meaning is highly flexible. It is trivial to re-label a concept, as in our explicit training paradigm. However, we also found a shift in the meaning of untrained words. This suggests a modification of word meaning at a deeper, conceptual level. The mental flexibility allowing the development of a modified meaning was not transient and determined by the experimental paradigm, but appeared to be maintained throughout the test block. Thus, in contrast to models of word meaning that depend entirely on a set of fixed features (Locke) or reference to an expert (Goodman), our findings are compatible with the notion that a component of word meaning is associated in part with a central tendency (Wittgenstein) that is somewhat flexible in the reference of a word. Another question that needs to be addressed in future research is how long the criterion shift effect induced in the adaptation phase actually lasts, i.e., how long it takes for the learning curve to go back to baseline level. Any considerations of how to make use of the observed effect in therapeutic settings where magnitudes or amounts are being misevaluated by patients (as in additive disorders or body schema disorders) will depend on its natural decay. This decay may in part depend on working memory processes which have been discussed as essential for quantifier processing (McMillan et al., [@B14]; Zajenowski et al., [@B25]). The present findings and the paradigm used to produce them may be helpful in order to resolve a longer-standing debate about the nature of the brain network underlying quantifier processing. McMillan et al. ([@B14]) found increasing involvement of the left inferior frontal cortex in quantifiers with increasing processing demands (*"at least 3," "all," "some"* vs. *"less than half," "odd," "even"*) and related this effect to aspects of working memory (see Wei et al., [@B24], for comparable findings in Chinese). More recently, Heim et al. ([@B8]) reported two fronto-parietal networks for quantifier processing in the left hemisphere that were neighboring but hardly overlapping. One of these networks was related to the initial estimation of numerosities, the other to the comparison with a criterion. Interestingly, only the frontal component also showed systematic effects of processing quantifier meaning, suggesting a functional segregation between the frontal and the parietal components of the two networks. The paradigm presented here may be useful to get deeper insights into the role of the frontal component in order to understand in how far it is involved in the processing of change of meaning during the adaptation phase and/or maintenance of the newly learned meaning in working memory and its decay. Thus, subsequent work both with functional neuroimaging and also with patients with frontal or parietal neurodegeneration is encouraged in order to identify the neural basis of quantifier processing from the initial percept to the final semantic evaluation. Some first steps into that direction have been done by McMillan et al. ([@B15]) and by Morgan et al. ([@B15a]) in studies of patients with fronto-temporal and/or posterior cortical disease. In accordance with Dehaene\'s triple-code model (Dehaene et al., [@B4]) it was again atrophy in the frontal cortex that had a severe effect on the verbal code of numerosities. It is yet an open question whether it is the dysfunction of parts of the frontal cortex itself, or rather dynamic diaschisis (i.e., alterations in the entire network of frontal and parietal areas, as could be supposed based on Troiani et al., [@B24a]) that thus affects the neurobiology of the language-number interface in the human brain. Conclusion {#s2} ========== The present study investigated quantifier processing in a setting in which their meaning could be experimentally altered for the course of the experiment. This novel paradigm yielded results suggesting that, in order to be most highly informative, a change in meaning in one quantifier generalizes also to that of other quantifiers, i.e., to a larger referential frame. Future studies are required to test the usefulness and limits of this paradigm behaviorally and with neuroimaging. Conflict of interest statement ------------------------------ The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The study was funded by the National Institutes of Health (NS044266, AG017586, AG032953, NS053488, and AG043503) and the Wyncote Foundation. SH received financial support for his lab visit at UPenn from the Department of Psychiatry, Psychotherapy and Psychosomatics (Frank Schneider) and the Section for Clinical and Cognitive Neuroscience (Ferdinand Binkofski), Uniklinik RWTH Aachen, as well as from the Institute for Neuroscience and Medicine (INM-1; Katrin Amunts) at the Research Centre Jülich. Thanks go to Yosef Grodzinsky and Klaus Willmes for ongoing discussion of quantifier processing, and especially to Birgit Heim for making this lab visit possible. ^1^The assignment of the YES and NO responses to the one or other button were kept constant for each single subject for the duration of the entire experiment. Across subjects, the assignment of YES or NO to the left or right button was balanced. Thus, 50% of the subjects pressed the left button for YES and the right for NO and vice versa for the other 50%. ^2^Note, however, that "few" is not exactly equivalent to "not many." This is due to the fact that these two quantifiers do not refer to the same reference, or criterion, whatever its value may be in a given context and for a given individual. In the example of peanuts, it is well possible that you neither eat few peanuts (i.e., a handful) nor many (i.e., the whole package), but just some intermediate amount. The situation might be different for quantifiers that are exact opposites not by virtue of explicit negation ("more than half"/"not more than half") but of semantic polarity ("more than half"/"less than half"). [^1]: Edited by: Sonja A. E. Kotz, Max Planck Institute Leipzig, Germany [^2]: Reviewed by: Anna J. Simmonds, Imperial College London, UK; John E. Drury, Stony Brook University, USA [^3]: This article was submitted to Language Sciences, a section of the journal Frontiers in Psychology

{ "pile_set_name": "PubMed Central" }

Introduction {#Sec1} ============ Hair cell function in the cochlea and the vestibular system is calcium (Ca^2+^)-dependent (Tanaka et al. [@CR39]; Ohmori [@CR27]; Kozel et al. [@CR18]) and requires the endolymphatic Ca^2+^ concentration (\[Ca^2+^\]~endolymph~) to be tightly controlled at a level that is unusually low (0.017--0.133 mmol/l (Salt et al. [@CR35])) for an extracellular fluid (e.g., \[Ca^2+^\]~blood\ plasma~ = 1.2 mmol/l (Diem and Lentner [@CR10])). To maintain this physiological \[Ca^2+^\]~endolymph~, numerous epithelial sites within the membranous labyrinth are engaged in Ca^2+^ transport across the endolymph-perilymph barrier, utilizing Ca^2+^-permeable channels (e.g., transient receptor potential cation channel subfamily V members \[TRPV channels\] (Takumida et al. [@CR38]; Wangemann et al. [@CR43]; Nakaya et al. [@CR25]; Ishibashi et al. [@CR17])), ion exchangers (e.g., sodium-calcium exchangers \[NCX\] (Oshima et al. [@CR28]; Yamauchi et al. [@CR44])), and Ca^2+^ ATPases (e.g., plasma membrane calcium ATPase \[PMCA\] (Crouch and Schulte [@CR8]; Ågrup et al. [@CR1])). Many of these transport proteins are directly regulated downstream targets of the calcium sensing receptor (CaSR), a G-coupled cell surface receptor that is prominently expressed in the primary tissues involved in whole-body Ca^2+^ homeostasis, including the kidney and the parathyroid and thyroid glands (Blankenship et al. [@CR4]; Topala et al. [@CR40]; Ranieri et al. [@CR30]). In those tissues, CaSR senses changes in Ca^2+^ levels in urine and plasma and modulates the activity and expression of Ca^2+^ transport proteins to increase renal Ca^2+^ reabsorption (Riccardi et al. [@CR31]) and the Ca^2+^-mediated release of calciotropic hormones (parathyroid hormone, calcitonin) (Brown and MacLeod [@CR5]), respectively. In the inner ear, the endolymphatic sac (ES), among other epithelial tissues, is a site of CaSR gene expression (Beitz et al. [@CR3]; Lin et al. [@CR19]) and critically contributes to endolymphatic Ca^2+^ homeostasis. This function was first indicated by in vivo animal experiments in which surgical separation of the ES resulted in a nonphysiological increase in \[Ca^2+^\]~endolymph~ and endolymphatic hydrops development in the cochlea (Ninoyu and Meyer zum Gottesberge [@CR26]; Salt and DeMott [@CR33]). Structural loss of the distal (extraosseous) ES, as induced in this animal model, is a histopathological hallmark and a presumed etiopathogenic factor in Meniere's disease (MD) (Eckhard et al. [@CR13]). Here, we investigated the immunolocalization patterns of CaSR and Ca^2+^ transport proteins in the mature murine ES epithelium to elucidate the molecular determinants of Ca^2+^ homeostasis in the ES and to better understand the pathophysiological consequences of the structural loss of the ES in MD. Materials and methods {#Sec2} ===================== Animals {#Sec3} ------- All animal experiments were performed according to Swiss animal welfare laws and were approved by the local veterinary authorities (Kantonales Veterinäramt Zürich; protocol no. ZH269/16). Male, 6- to 8-week-old C57BL/J6-Crl1 mice were purchased from Charles River Laboratories, Inc. (Sulzfeld, Germany). A total of 12 animals were used in this study. All animals were kept at an in-house animal facility with free access to a standard chow diet and water under a standard 12 h light/dark cycle. Tissue preparation, fixation, decalcification, embedding, and sectioning {#Sec4} ------------------------------------------------------------------------ Animals were sacrificed by CO~2~ inhalation. For tissue fixation, the thoracic cavity was opened, and a cannula was inserted in the left cardiac ventricle for transcardial perfusion with 10 ml phosphate buffered saline (PBS), followed by 10 ml of a fixative, i.e., (i) 10% neutral buffered formalin (F; Lucerna-Chem AG, Luzern, Switzerland), (ii) F with 1% acetic acid (FA; Sigma-Aldrich, Steinheim, Germany), (iii) F with 0.2% glutaraldehyde (GA; Sigma-Aldrich), or (iv) FA with 0.2% glutaraldehyde (FGA). The respective fixative used in each protocol is given in Table [1](#Tab1){ref-type="table"}. Then, the animals were decapitated, soft tissues were removed from the outer surface of the skull using a sharp scalpel, and the cranial part of the skull was opened to remove most of the cerebrum, while the cerebellum was left in situ. The tympanic bullae were opened to allow faster penetration of the fixative into the inner ear. The specimens were then immersed in the respective fixative used for transcardial perfusion for 6 h on a countertop shaker. Afterwards, the specimens were dehydrated in a graded series of ethanol solutions (50%, 70%, 95%, and 100%), embedded in paraffin and sectioned at 4 μm using an HM 355S Automatic Microtome (Thermo Fisher Scientific, Waltham, MA, USA). Sections were collected on SuperFrost Plus slides (Thermo Fisher Scientific), dried on a heating plate at 37 °C overnight and stored at room temperature.Table 1Primary antibodies used with DAB labeling. Dilutions indicate the highest dilution producing a sufficient signal. All antibodies were diluted in 1% normal horse serum (NHS). Form indicates 10% neutral buffered formalin. *FA* form + 1% acetic acid, *FG* form + 0.2% glutaraldehyde, *FGA* FA + 0.2% glutaraldehyde, *HIER* heat-induced epitope retrievalEpitopeCalbindin D-28 kCaSRNCX2ParvalbuminpPMCAPMCA1PMCA2SERCA1SERCA2TRPV5TRPV6Immunogen (peptide sequence)No infoADDYGRPGIEKFREEAEE RDIRVGDAQGMFEPDGGNo infoExact sequence not available^a^SGVKNSLKEANHDKEIPDPSSINAKTLETPDQVKRHLEKYGCTPNKPSRTSMSKGLNLSEGDGEEVYHFNRGLEDGESWEYQIFixativeFGFAFGFAForm, FGFGFGFAFAForm, FGForm, FGHIERNoYesYesYesNoNoNoNoNoYesYesPrimary antibodyMonoclonal anti-calbindin D-28 k (mouse)Monoclonal anti-CaSR (mouse)Polyclonal anti-NCX-2 (rabbit)Monoclonal antiparvalbumin (mouse)Monoclonal anti-pPMCA (mouse)Polyclonal anti-PMCA1 (rabbit)Polyclonal anti-PMCA2 (rabbit)Polyclonal anti-SERCA1 (rabbit)Polyclonal anti-SERCA2 (rabbit)Polyclonal anti-TRPV5 (rabbit)Polyclonal anti-TRPV6 (rabbit)Dilution1:5001:15′0001:5001:5001:40001:10001:10001:5001:5001:2001:200ManufacturerSwant, Martiny, SwitzerlandNovus Biologicals, Minneapolis, USAAlomone Labs, Jerusalem, IsraelSwant, Martiny, SwitzerlandAbcam, Cambridge, UKAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelAlomone Labs, Jerusalem, IsraelCatalog number300NB120-19347SSANX-012235ab2825ACP-005ACP-002ACP-011ACP-012ACC-035ACC-036Secondary antibody (biotinylated)Anti-mouse IgG (donkey)Anti-mouse IgG (donkey)Anti-rabbit IgG (donkey)Anti-mouse IgG (donkey)Anti-mouse IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Anti-rabbit IgG (donkey)Dilution1:4001:4001:4001:4001:4001:4001:4001:4001:4001:10001:1000ManufacturerMilan Analytica AG, Rheinfelden,Milan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGMilan Analytica AGCatalog number715-065-151715-065-151715-065-152715-065-151715-065-151715-065-152715-065-152715-065-152715-065-152715-065-152715-065-152^a^This antibody recognizes an epitope between amino acids 724--783 of the human erythrocyte calcium pump Immunohistochemical DAB labeling {#Sec5} -------------------------------- Sections were deparaffinized in Histo-Clear (National Diagnostics, Atlanta, GA, USA), rehydrated in a graded series of ethanol solutions (100%, 95%, 70%, and 50%), and rinsed in tap H~2~O. When required, heat-induced antigen retrieval (HIAR) with pressurized coverslipping of the mounted tissue sections was performed (Table [1](#Tab1){ref-type="table"}) according to a previously established protocol (Eckhard et al. [@CR12]). All subsequent steps were performed at room temperature. Nonspecific binding was blocked with 1% normal horse serum (NHS) for 15 min, followed by incubation with primary antibodies overnight. The sections were then incubated with biotinylated secondary antibodies for 1 h, followed by incubation with avidin-biotin-HRP complex (Vectastain ABC HRP Kit, Vector Laboratories, Burlingame, CA, USA). Visualization was performed with 3,3′-diaminobenzidine (DAB) (DAB Peroxidase \[HRP\] Substrate Kit, Vector Laboratories). All incubation steps were performed at room temperature, and each incubation step was followed by rinsing the sections in PBS for 5 min. All primary and secondary antibody combinations utilized in this study are listed in Table [1](#Tab1){ref-type="table"}. Negative controls for all immunolabeling experiments were obtained by omitting the primary antibody or after preabsorption of the primary antibodies with the corresponding commercially available control peptides (for primary antibodies against TRPV5/6, PMCA1/4, NCX2, and SERCA1/2). For nuclear counterstaining, the slices were incubated for 5 s with hematoxylin diluted in distilled water (1:6) and mounted with a permanent mounting medium (VectaMount, Vector Laboratories). Immunofluorescence double labeling {#Sec6} ---------------------------------- Deparaffinization, rehydration of the sections, and incubation with primary antibodies was performed as described for DAB-labeling experiments. All primary antibodies used for immunofluorescence double labeling are listed in Table [2](#Tab2){ref-type="table"}. The sections were then incubated with biotinylated antibodies raised in donkey and directed against mouse (1:400; 715-065-151, Milan Analytica AG, Rheinfelden, Switzerland) and Alexa Fluor 594-conjugated antibodies raised in goat and directed against rabbit (1:400; 111-585-144, Milan Analytica AG) for 1 h, followed by incubation with avidin-biotin-HRP complex (Vectastain ABC HRP Kit, Vector Laboratories). Biotinylated tyramine (10 min) and a subsequent second incubation with avidin-biotin-HRP complex were used for signal amplification. Visualization was performed with Alexa Fluor 488-conjugated streptavidin (1:800; Milan Analytica AG) and Alexa Fluor 594-conjugated anti-goat antibodies raised in donkey (1:400; 705-585-003, Milan Analytica AG). The sections were mounted with an aqueous mounting medium containing DAPI for nuclear counterstaining (H-1200, Vectashield, Vector Laboratories).Table 2Primary antibodies used with fluorescence labeling. Dilutions indicate the highest dilution producing a signal. All antibodies were diluted in 1% normal horse serum (NHS). If not otherwise specified, sodium citrate (10 mM) at pH 6.0 was used as the HIER buffer. Form indicates 10% neutral buffered formalin. *FA* form + 1% acetic acid, *FG* form + 0.2% glutaraldehyde, *FGA* FA + 0.2% glutaraldehyde, *HIER* heat-induced epitope retrieval, *NHS* normal horse serumEpitopeCaSRγENaCTRPV5V-ATPaseImmunogen (peptide sequence)ADDDYGRPGIEKFREEAEERDIAvailable upon request.GLNLSEGDGEEVYHFMAMEIDSRPGGLPGSSCNLGAAREHMQAVTRNYITHPRVTYRTVCSVNGPLVVLDR VKFAQYAEIVFixativeFormFormFGFAHIERYesYesYesYes (Tris-EDTA solution, 10 mM / 1 mM, at pH 9.0)Primary antibodyMonoclonal anti-CaSR (mouse)Monoclonal anti-γENaC (rabbit)Polyclonal anti-TRPV5 (rabbit)Polyclonal anti-Vacuolar-type H^+^ ATPase subunit B, kidney isoform (rabbit)Dilution1:2001:2001:1001:200ManufacturerNovus Biologicals, Minneapolis, USAGift from Prof. Johannes Loffing, University of Zurich, SwitzerlandAlomone Labs, Jerusalem, IsraelNovus BiologicalsCatalog numberNB120-19347SSn. a.ACC-035NBP2-33962 Microscopy analysis {#Sec7} ------------------- Images were acquired using either a Leica DMI6000 microscope (Leica, Wetzlar, Germany) or a Leica SP5 confocal laser scanning microscope (Leica) and processed with Adobe Photoshop CS5 software (Version 12.0, Adobe Systems, San Jose, CA, USA). Quantification of immunolabeled ES epithelial cells {#Sec8} --------------------------------------------------- DAB-labeled (DAB^+^) ES epithelial cells were quantified on light microscopic (× 40 air objective) images of hematoxylin-counterstained tissue sections. Cell counting was performed in at least three non-consecutive sections from two different axial planes of the intraosseous ES (iES) and the extraosseous ES (eES). Each section was divided into ten equally long segments along the longitudinal axis of the ES. The percentage of DAB^+^ cells among the epithelial cells with hematoxylin-stained nuclei was determined in each segment. Single- and double-immunofluorescence-labeled ES epithelial cells in the eES were quantified following double immunofluorescence labeling of CaSR with either V-ATPase, TRPV5, or γENaC. The percentages of fluorescence-labeled cells among the epithelial cells with DAPI-counterstained nuclei were determined. From each double immunofluorescence labeling experiment, at least 100 labeled cells were counted, and sections from at least 3 independent specimens were analyzed. Descriptive statistics {#Sec9} ---------------------- Statistical analyses were performed using Prism software (version 7.0, GraphPad Software). Cell counts are expressed as percentages of the number of epithelial cells in a segment of the ES. Mean values and standard deviations derived from six animals are shown. Results {#Sec10} ======= Based on the previously demonstrated similarities between the ES epithelium and the kidney distal tubular epithelium on the cellular and molecular levels, we mainly focused our immunohistochemical analysis on Ca^2+^ transport-associated proteins that are known to be crucial for the renal tubular handling of calcium. Cellular immunolocalization patterns of the CaSR and Ca^2+^ transport proteins {#Sec11} ------------------------------------------------------------------------------ DAB immunolabeling of the CaSR, as well as of TRPV5, TRPV6, SERCA1, SERCA2, NCX2, PMCA1, and PMCA4 was visible in the iES (proximal to the operculum) and eES (distal to the operculum; Fig. [1](#Fig1){ref-type="fig"} and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}) portions of the ES epithelium. Staining of proteins was absent or very weak in the endolymphatic duct (ED; Fig. [1](#Fig1){ref-type="fig"} and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}). Negative controls (omitted primary antibodies and preabsorption of primary antibodies with control peptides (TRPV5/6, NCX2, PMCA1/4, SERCA1/2) showed no specific staining in the murine ES epithelium.Fig. 1DAB-immunohistochemical staining of the calcium sensing receptor (CaSR) and selected calcium transport proteins along the murine endolymphatic sac. **a**, **b** Overviews of the iES emerging from the ED (**a**) and of the eES (**b**), demonstrating epithelial CaSR immunostaining along the ES. Boxed numbers indicate representative regions shown below in columns. **c**--**f** Representative images corresponding to the regions indicated by the boxed numbers in (**a**) and (**b**), i.e., the ED-iES transition (first column), iES (second column), and eES (third column). Scale bars: 100 μm (**a**), 20 μm (**c**). ED endolymphatic duct, eES extraosseous endolymphatic sac, iES intraosseous endolymphatic sac, sSCC superior semicircular canal CaSR immunolabeling was present in the iES and eES portions (Fig. [1a, b](#Fig1){ref-type="fig"}) and exhibited a scattered labeling pattern within the epithelium. In CaSR-labeled cells, strong cytoplasmic and membranous labeling was found, which in many cells was polarized to the basolateral membranes (Fig. [1c--c](#Fig1){ref-type="fig"}"). Immunolabeling signals for transient receptor potential cation channel subfamily V (TRPV) members TRPV5 and TRPV6 appeared to be cytoplasmic with strong polarization towards the apical cell pole, most clearly in epithelial cells in the eES (Fig. [1d--d](#Fig1){ref-type="fig"}" and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}A-A"). Homogeneous cytoplasmic immunolabeling of the sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA) subtypes SERCA1 and SERCA2 was found throughout the iES and eES epithelium (Fig. [1e--e](#Fig1){ref-type="fig"}" and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}B-B") and was consistent with their known localization in the endoplasmic reticulum (Strehler and Treiman [@CR37]). Labeling for the plasma membrane calcium ATPase (PMCA) subtypes PMCA1 and PMCA4, as well as for the sodium-calcium exchanger NCX2, was polarized in the basolateral membranes, most clearly in the eES (Fig. [1f--f](#Fig1){ref-type="fig"}" and Supplementary Fig. [1](#Fig5){ref-type="supplementary-material"}). No immunolabeling in the ES and ED was found for PMCA2 and NCX1, nor for the Ca^2+^-binding proteins Calbindin D-28k and parvalbumin (Supplementary Fig. [2](#Fig6){ref-type="supplementary-material"}). All immunolabeling patterns and their cellular polarization in the ES epithelium were consistent with those previously described for the distal convoluted tubule and the connecting tubule epithelium (Loffing et al. [@CR21]; Loffing and Kaissling [@CR20]); positive control experiments are shown in Supplementary Fig. [3](#Fig7){ref-type="supplementary-material"}). Spatial (proximal-to-distal) immunolabeling gradients {#Sec12} ----------------------------------------------------- For each protein studied, DAB-labeled cells were quantified along the longitudinal (proximal-to-distal) axis of the ES epithelium, reaching from the proximal transition of the iES in the ED to the distal blind end of the eES. For all of the investigated proteins, a gradual increase in the immunolabeling signal in the proximal-to-distal direction was observed. At the very distal end of the eES portion, on average, more than 50% of the epithelial cells were positively labeled for most of the investigated proteins (Fig. [2](#Fig2){ref-type="fig"}).Fig. 2Quantification of DAB^+^ cells along the ES reveals staining gradients along the ES epithelium increasing from the iES to the eES. The *y*-axis shows the percentage of DAB^+^ cells relative to all ES epithelial cells. For cell counting, both the iES and eES were each divided into ten equal segments. The ticks on the *x*-axis correspond to each of the ten consecutive segments of the iES (left *x*-axis) and eES (right *x*-axis). The first two segments represent the endolymphatic duct epithelium. The values are the percentages of the total number of cells in the respective ES segment and are given as mean values ± SD from six animals. **a** Quantification of DAB^+^ cells for CaSR and the apical Ca^2+^ channels TRPV5 and TRPV6. **b** Quantification of DAB^+^ cells for the cytoplasmatic proteins SERCA1 and SERCA2. **c** Quantification of DAB^+^ cells for proteins located in the basolateral membrane Differential immunolocalization patterns in mitochondria-rich and ribosome-rich cells {#Sec13} ------------------------------------------------------------------------------------- We found scattered DAB-immunolabeling patterns within the ES epithelium for almost all of the investigated proteins (Fig. [1](#Fig1){ref-type="fig"}). We therefore investigated whether CaSR and Ca^2+^ transport proteins are differentially distributed among the two major ES epithelial cell types, i.e., mitochondria-rich cells (MRCs) and ribosome-rich cells (RRCs) (Lundquist et al. [@CR23]; Barbara et al. [@CR2]; Dahlmann and von Düring [@CR9]), using immunofluorescence double labeling. Labeling of the MRC-specific vacuolar-type H^+^-ATPase (V-ATPase; (Stanković et al. [@CR36]; Dou et al. [@CR11])) and CaSR in the eES portion showed a strict cellular colocalization of both proteins (Fig. [3a--a\'\'\'](#Fig3){ref-type="fig"}) in 45.0% of the cells (Fig. [3b](#Fig3){ref-type="fig"}), indicating the localization of CaSR in MRCs (vATPase^+^/CaSR^+^) but not in RRCs (vATPase^−^/CaSR^−^). Colabeling of TRPV5 and CaSR (Fig. [3c-c\'\'\', d](#Fig3){ref-type="fig"}) was found in 31.4% of the cells (TRPV5^+^/CaSR^+^), while the other cells showed exclusive labeling for either TRPV5 (TRPV5^+^/CaSR^−^, 19.3%) or CaSR (TRPV5^−^/CaSR^+^, 22.9%) or were devoid of labeling (TRPV5^−^/CaSR^−^, 26.4%). Colabeling of the gamma subunit of the epithelial sodium channel (γENaC) and CaSR (Fig. [3e-e\'\'\', f](#Fig3){ref-type="fig"}) was present in 28.1% of the cells (γENaC^+^/CaSR^+^), while 20.3% of cells showed labeling for only γENaC (γENaC^+^/CaSR^−^), 29.4% of the cells were labeled for only CaSR (γENaC^−^/CaSR^+^), and 22.2% of cells were devoid of labeling (γENaC^−^/CaSR^−^).Fig. 3Quantitative colocalization of CaSR with V-ATPase, TRPV5, and γENaC in the eES. Representative images and quantification of fluorescent immunohistochemical double labeling of CaSR and V-ATPase (**a**--**a"'**, **b**), TRPV5 (**c**--**c"'**, **d**), and γENaC (**e**--**e"'**, **f**). Scale bar: 10 μm. *N* total number of cells counted CaSR immunolabeling within the murine inner ear is restricted to the ES epithelium {#Sec14} ---------------------------------------------------------------------------------- No specific CaSR immunoreactivity was found in other portions of the membranous labyrinth, including the cochlea, saccule, utricle, and semicircular canal cristae (Fig. [4](#Fig4){ref-type="fig"}).Fig. 4No CaSR immunolabeling was detected within the cochlea or the vestibular end organs. **a** Negative CaSR immunolabeling in the cochlea, in particular the stria vascularis, spiral ligament, Reissner's membrane, limbus, organ of Corti, and spiral ganglion. **b** Ampullary region of the lateral semicircular canal. No CaSR staining was observed in the crista or the SCC epithelium (membranous labyrinth). Adjacent to the crista, several dark cells (melanocytes) are interspersed in the membranous labyrinth. **c**, **d** No CaSR immunolabeling was found in the sensory epithelium or surrounding tissue of the utricle (**c**) and saccule (**d**). Several dark cells are interspersed in the membranous labyrinth covering the utricle (**c**) Discussion {#Sec15} ========== In the present study, the ES epithelium was identified as an exclusive site of CaSR immunolocalization within the mature murine inner ear. Several CaSR-regulated Ca^2+^ transport proteins (TRPV5, PMCA1/4), as well as non-CaSR-regulated Ca^2+^ channels/transporters (TRPV6, NCX2), and intracellular Ca^2+^ sequestering proteins (SERCA1/2) were localized in the ES epithelium, almost all of which exhibited their highest immunolabeling intensities in the distal eES portion. Consistent with the immunolocalization of CaSR in MRCs of the mature murine ES shown here, previous gene expression studies detected CaSR mRNA expression in MRCs of the early postnatal murine ES (Honda et al. [@CR15]), in the adult rat ES (Beitz et al. [@CR3]), and, moreover, in the adult rat organ of Corti, stria vascularis, Reissner's membrane, and vestibulum (Beitz et al. [@CR3]), as well as in neuromast hair cell stereocilia of zebrafish larvae (Lin et al. [@CR19]). The results of the present study indicate that the CaSR protein in the mature murine inner ear is exclusively expressed in the ES or that immunolabeling at other cellular sites was below the detection level. We propose that the CaSR may have a unique role in sensing and regulating \[Ca^2+^\]~endolymph~ in the ES, with significance for the overall endolymphatic Ca^2+^ homeostasis in the inner ear. However, it is unknown whether the CaSR enables ES epithelial cells to sense Ca^2+^ changes in the ES endolymph, the cerebrospinal fluid, or in both extracellular fluid compartments that face the ES epithelium. Moreover, the molecular signaling cascades by which the CaSR may regulate downstream effector proteins, such as TRPV5 and PMCA1/4 (Blankenship et al. [@CR4]; VanHouten et al. [@CR41]; Topala et al. [@CR40]), which are also localized in the ES epithelium (Ågrup et al. [@CR1]; Honda et al. [@CR15]; present study), require further exploration. Although CaSR function in the ES may contribute to maintaining Ca^2+^ within its narrow physiological range in the entire inner ear, its effects on \[Ca^2+^\]~endolymph~ in the vestibular and cochlear portions are probably slow and may have physiological relevance primarily on longer time scales. This assumption is supported by previous observations that (i) longitudinal fluid movements (between endolymph compartments) virtually do not occur, at least under physiological conditions (Salt et al. [@CR34]), and that (ii) surgical separation of the ES only leads to a slow rise of \[Ca^2+^\]~endolymph~ over a time course of several weeks (Ninoyu and Meyer zum Gottesberge [@CR26]; Salt and DeMott [@CR33]). This may be primarily due to local ion homeostatic mechanisms within the cochlea and the vestibular portions (Lundquist [@CR22]; reviewed in Salt [@CR32]), as well as anatomical diffusion barriers, such as the utriculosaccular duct, which (in the adult murine inner ear) largely prevents fluid exchange between the pars inferior (cochlea, saccule) and the pars superior (utricle, semicircular canals) (Cantos et al. [@CR6]). In addition, the CaSR may have local homeostatic functions in the ES, such as maintaining ES epithelial integrity or mediating immune responses (Cheng et al. [@CR7]). Among the two main epithelial cell types in the ES, i.e., MRCs and RRCs, we found that the CaSR was exclusively localized in vATPase^+^ MRCs (Dou et al. [@CR11]; Honda et al. [@CR15]). This cell-type-specific distribution of CaSR resembles that in intercalated cells in the renal cortical collecting duct (CCD; (Everett [@CR14])). On the subcellular level, CaSR was polarized towards both membrane domains (apical and basolateral) with additional strong cytoplasmic labeling. This finding is consistent with the distribution of CaSR in the distal renal tubule epithelium (distal convoluted tubule and collecting duct (Riccardi et al. [@CR31])), where CaSR regulates transepithelial Ca^2+^ reabsorption from the tubular fluid via TRPV5 (Topala et al. [@CR40]) and PMCA (Blankenship et al. [@CR4]). Notably, other shared molecular mechanisms of transepithelial ion (Na^+^) transport between the ES epithelium and the distal renal tubule epithelium have been reported (Mori et al. [@CR24]; Eckhard et al. [@CR13]). In contrast to CaSR, TRPV5 was differentially localized in subsets of MRCs and RRCs, thereby differing from its exclusive localization in principal cells in the renal CCD (Loffing and Kaissling [@CR20]). However, this non-cell-type-specific TRPV5 localization in the ES is consistent with the RNA sequencing analysis of the murine ES (Honda et al. [@CR15]). These and previously noted structural differences between the epithelia of the ES and the renal CCD (Wangemann and Marcus [@CR42]), despite their numerous shared cellular and molecular features, may be due to the different physiological requirements for maintaining extracellular Ca^2+^ homeostasis in the ES endolymph and the blood plasma. In addition to CaSR-regulated Ca^2+^ transporters, we identified TRPV6 and NCX2 in the ES epithelium, both of which are known to enable constitutively active Ca^2+^ transport in the renal tubular epithelium (Loffing and Kaissling [@CR20]). Loss of the proposed Ca^2+^-homeostatic function of the ES may be of significance in various pathological conditions, such as otoconial disorders and MD. The biogenesis of otoconial Ca^2+^ carbonate (CaCO~3~) crystals is believed to occur in the ES luminal microenvironment and is sensitive to changes in \[Ca^2+^\]~endolymph~ (Nakaya et al. [@CR25]). For example, disturbed formation of otoconia was observed in Foxi1 knockout (k.o.) mice and Efnb2 k.o. mice (Hulander [@CR16]; Raft et al. [@CR29]). In these models, the MRC population in the ES is developmentally lacking (Hulander [@CR16]) or mildly decreased and proximally mislocalized in the ES (Raft et al. [@CR29]), respectively. The loss of CaSR-regulated Ca^2+^ homeostasis in the ES is one possible explanation for the compromised otoconia formation in these transgenic mouse models. Another explanation is the loss of (MRC-specific) pendrin- and v-ATPase-mediated anion transport that leads to disturbances of endolymphatic pH, which in turn may inhibit the pH-sensitive TRPV5/6 channels in the ES epithelium and thereby disturb \[Ca^2+^\]~endolymph~ ((Nakaya et al. [@CR25]), present study). In MD, the structural loss of the distal ES epithelium is a consistent histopathological hallmark. Presuming that the human ES harbors Ca^2+^ transport mechanisms in proximal-to-distal expression gradients analogous to the murine ES, structural loss of the distal ES---a pathological hallmark in MD (Eckhard et al. [@CR13])---and in particular its Ca^2+^ transport functions may be crucial in the pathogenesis of endolymphatic hydrops (Ninoyu and Meyer zum Gottesberge [@CR26]; Salt and DeMott [@CR33]). In conclusion, the murine distal ES epithelium presumably has a "calciostatic" function for inner ear fluid homeostasis. Human pathology of the ES likely impairs Ca^2+^ homeostasis of the inner ear fluids and is therefore potentially a significant pathophysiological factor in various inner ear disorders. Electronic supplementary material ================================= {#Sec16} Figure S1DAB-immunohistochemical staining of selected calcium transport proteins along the murine endolymphatic sac. (**A -- E**) Representative images corresponding to the regions indicated by the boxed numbers in Fig. [1a and b](#Fig1){ref-type="fig"}, i.e., the ED-iES transition (first column), iES (second column) and eES (third column). Scale bar: 20 μm. (PNG 165460 kb) High resolution image (TIFF 3415 kb) Figure S2A lack of DAB-immunohistochemical staining in the ES epithelium was observed for calbindin-D28k, parvalbumin, PMCA2 and NCX1. (**A**) Negative DAB immunolabeling of calbindin-D28k in the ES epithelium. Scale bar (for A, D, G and I): 20 μm. (**B -- C**) Positive control DAB staining for calbindin-D28k in the distal nephron (B) and in Purkinje cells of the cerebellar cortex (C). Scale bar in (B): 10 μm (for B, E and J); scale bar in (C): 10 μm (for C, F and H). (**D**) Negative DAB immunolabeling of parvalbumin in the ES epithelium. (**E -- F**) Strong positive control DAB staining in parvalbumin-immunostained cells in the distal nephron (E) and weak positive DAB staining in Purkinje cells of the cerebellar cortex (F). (**G**) Negative DAB immunolabeling of PMCA2 in the ES epithelium. (**H**) Positive control staining of PMCA2, with a predominant apical staining pattern in Purkinje cells of the cerebellar cortex (F). (**I**) Negative DAB immunolabeling of NCX1 in the ES epithelium. (**J**) Positive control staining for NCX1, with a basolateral staining pattern in the distal nephron (J). All scale bars: 20 μm. (PNG 91011 kb) High resolution image (TIFF 1899 kb) Figure S3Positive DAB control staining in the kidney and in the trigeminal ganglion for proteins exhibiting positive immunostaining in the ES epithelium (Fig. [1](#Fig1){ref-type="fig"}, Fig. [S1](#Fig5){ref-type="supplementary-material"}). (**A -- H**) Positive DAB staining in sections along the distal nephron for CaSR, TRPV5, TRPV6, SERCA1, SERCA2, PMCA1, PMCA4 and pPMCA. (**I**) As NCX2 is not expressed in the kidney, somata of trigeminal ganglion neurons served as a positive control for NCX2 staining. Scale bar: 20 μm. (PNG 70669 kb) High resolution image (TIFF 1758 kb) Ca^2+^ : (Ionized) calcium \[Ca^2+^\]~endolymph~ : Endolymphatic Ca^2+^ concentration Calb : Calbindin D-28k CaSR : Calcium sensing receptor CCD : Cortical collecting duct ES : Endolymphatic sac ED : Endolymphatic duct MRC : Mitochondria-rich cell NCX1/2 : Sodium-calcium exchanger type 1/2 Parv : Parvalbumin PMCA1/4 : Plasma membrane calcium ATPase isoform 1/4 pPMCA : Pan-plasma membrane calcium ATPase RRC : Ribosome-rich cell SERCA1/2 : Sarco/endoplasmic reticulum Ca^2+^-ATPase paralog 1/2 TRPV5/6 : Transient receptor potential cation channel subfamily V member 5/6 V-ATPase : Vacuolar-type H^+^ ATPase **Publisher's note** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. The authors thank Sabina Wunderlin for excellent technical assistance. The authors thank Prof. J. Loffing (Institute of Anatomy, University of Zurich) for providing the anti-γENaC antibody. The study was supported by grants from the British Ménière's Society and the Zürcher Stiftung für das Hören. DB is supported by a national MD-PhD scholarship from the Swiss National Science Foundation (SNSF). AHE was supported by a career development (Filling the gap) grant from the University of Zurich. The authors declare that they have no conflicts of interest. The animal research protocol was approved by the local veterinary authorities (permission number ZH269/16, Kantonales Veterinäramt, Zurich, Switzerland). All animal experiments were performed according to Swiss Animal Welfare laws and were carried out according to the approved animal research protocol.

{ "pile_set_name": "PubMed Central" }

![](hosplond70322-0005){#sp1 .185}

{ "pile_set_name": "PubMed Central" }

1. Introduction {#sec1} =============== Despite ongoing efforts to improve maternal and child health in developing countries, mortality rates remain much higher than in developed countries. Women in developing regions face a lifetime risk of maternal death of 1 in 160, as compared with 1 in 3700 for women living in developed regions \[[@B1]\]. These inequalities are driven by many causes, one of which is limited access to preventive services. For example, in low- and middle-income countries, only about 52% of pregnant women receive the World Health Organization- (WHO-) recommended minimum of four antenatal visits \[[@B2]\]. The postnatal period is also critical to the health of a mother and newborn, as the majority of postnatal maternal deaths happen during the first week after birth \[[@B3]\]. However, a recent analysis of Demographic and Health Surveys for 23 African countries found that, of the two-thirds of women giving birth at home, only 13% received a postnatal check-up within two days \[[@B3]\]. Immunization is another critical preventive service that can save the lives of many infants and children. Despite being one of the most cost-effective tools for saving lives, nearly one in five children globally did not receive their full package of immunizations in 2012 and 1.5 million children under the age of 5 died from vaccine-preventable diseases in 2008 \[[@B4], [@B5]\]. Antenatal care (ANC), postnatal care (PNC), and childhood immunization make up an important package of preventive services that can improve maternal and child health. Families tend to use medical services when someone is ill but frequently omit these beneficial preventive services that are essential to improve health. The field of mHealth, or mobile health, has been proposed as a potential solution to many of the challenges that developing countries face, including workforce shortages, lack of health information, limited training for health workers, and difficulty tracking patients. mHealth projects have been implemented all over the world, using mobile phones for record keeping, data collection, or patient communication \[[@B6]\]. Further, mHealth tools have been used to promote behavior change in health workers and/or patients. For example, text message reminders have been shown to increase care-seeking behavior or medication adherence in some patients and mobile data collection and communication tools for health workers have improved follow-up of patients and data reporting \[[@B7]--[@B9]\]. Though there are relatively few thorough evaluations of mHealth programs \[[@B6]\], some published studies do exist. Given that mHealth tools have shown some promise for behavior change more broadly, there is potential for this field to improve essential preventive maternal and child health services as well. Based on the existing evidence in peer-reviewed publications, this literature review aims to determine the effectiveness of mHealth tools to increase the coverage and use of antenatal care, postnatal care, and childhood immunizations through behavior change in low- and middle-income countries. 2. Materials and Methods {#sec2} ======================== 2.1. Information Sources {#sec2.1} ------------------------ This literature review was conducted through a keyword search of the following databases to identify relevant peer-reviewed articles: Google Scholar, PubMed, Embase, PsycINFO, and EBSCO Host. Keywords used in these searches included*mHealth*,*mobile health*,*mobile phone*,*reminder*,*recall*,*mobile medical records*,*antenatal care*,*postnatal care*, and*immunization*. 2.2. Inclusion Criteria {#sec2.2} ----------------------- In order to be included in the review, the article had to meet the following inclusion criteria:study was evaluating an mHealth intervention targeted at increasing antenatal care attendance, postnatal care attendance, or childhood immunization rates through behavior change;study was implemented in a low- or middle-income country;study included measurement of process, behavior change, health, or quality of care outcomes (i.e., studies were excluded that only evaluated willingness of participants to receive an mHealth intervention, without implementing it);study was a peer-reviewed article;study was available in English;study was published between January 1, 2000 and November 20, 2014.These criteria were selected to ensure that the included studies examined outcomes of existing mHealth interventions, not exploratory studies or protocols that have not been implemented yet. Low-, middle-, or high-income status for countries was determined using the World Bank\'s 2014 classification, which is based on estimates of the gross national income per capita for the previous year \[[@B10]\]. In addition, the inclusion of only peer-reviewed articles helped to ensure that higher quality studies were examined. Though there have been well-designed studies using mHealth for behavior change to support maternal and child health in high-income countries, these studies were excluded due to the resource disparities between high-income countries and others. Issues with prevalence of mobile phones and consistency of power and internet access are shared across many low- and middle-income countries and therefore these studies are more comparable than those conducted in high-income countries. No keywords for "low- or middle-income countries" were used in the searches, as these keywords might have excluded relevant results if the study was not specifically labeled as such. Instead, the authors screened manually for this criterion. The review was limited to studies available in English, though this is a limitation of this review, and future reviews should include additional languages, if feasible. Finally, studies only included those that were published after 2000, as mobile technologies were not widely available, especially in low- and middle-income countries, prior to that time. 2.3. Study Selection and Data Collection {#sec2.3} ---------------------------------------- The database searches were undertaken by two researchers (Jessica L. Watterson and Isheeta Madeka) between November 10, 2014 and January 18, 2015. Subsequent review of results was undertaken by one researcher (Jessica L. Watterson). The resulting articles were first screened by title, then by abstract, and finally by full text to progressively eliminate articles not meeting the inclusion criteria. Many systematic reviews of mHealth research were identified in the results (*n* = 26), so the included articles and reference lists of these reviews were all examined to ensure an exhaustive search. Finally, the references of all included articles were reviewed as well. The results of study screening and selection are illustrated in [Figure 1](#fig1){ref-type="fig"}. The database searches identified 1,899 articles initially. After removing duplicates, 508 records remained. Each of these records was screened by title and abstract (if necessary), and 455 records were excluded after this preliminary review. The full text of the remaining 53 articles was reviewed to determine if they met the inclusion criteria. 43 of the articles were excluded and the reasons for exclusion included study being conducted in a high-income country (*n* = 5); not studying antenatal care attendance, postnatal care attendance, or childhood immunization rates (*n* = 7); not studying an mHealth intervention (*n* = 2); or only providing program descriptions or a protocol but no evaluation data (*n* = 4). The other 25 articles that were excluded were mHealth literature reviews that did not identify any new articles for review. One article outlined a protocol for a study that will be very relevant once complete; however it was nevertheless excluded because no evaluation data is published yet \[[@B25]\]. 2.4. Quality Assessment {#sec2.4} ----------------------- Risk of bias was assessed for all included randomized controlled trials (RCTs) (*n* = 2) using the Cochrane Risk of Bias Assessment Tool \[[@B11]\]. This tool was introduced in 2008 by the Cochrane Collaboration and can be used to assess risk of bias in a study by evaluating a study\'s allocation sequence generation (randomization), allocation concealment, blinding, incomplete data, selective reporting, and other potential threats to the study\'s validity. The quality of the observational studies (*n* = 8) was assessed using the Newcastle-Ottawa Quality Assessment Scale \[[@B12]\]. This tool was developed by the universities of Newcastle, Australia, and Ottawa, Canada, and assessed the quality of nonrandomized studies by evaluating potential sources of bias in the selection and comparability of participants, the assessment of outcomes, and the duration and adequacy of follow-up. Scores are awarded out of 9 possible points, with higher scores indicating higher study quality. 2.5. Synthesis of Results {#sec2.5} ------------------------- The primary author (Jessica L. Watterson) extracted information from included articles for tabulation in an Excel spreadsheet. The information extracted included type of study, summary conclusions, methods used, intervention studied, health issue(s) studied, outcomes measured, sample size, intervention frequency, effectiveness of intervention, study quality, study location, clinical characteristics/setting, mHealth tools used, and project name (if any). 3. Results {#sec3} ========== Most articles examined process and behavior change outcomes and made recommendations for future mHealth programs and suggested further research. The study characteristics and key outcomes for each included article are outlined in [Table 1](#tab1){ref-type="table"}. 3.1. Characteristics of Studies {#sec3.1} ------------------------------- In total, ten articles satisfied the inclusion criteria. Of these, two studies were RCTs \[[@B13], [@B14]\] and the other eight were observational studies \[[@B15]--[@B22]\]. Four of the observational studies attempted to limit sources of bias (though not as rigorously as the RCTs) by using a historic control group \[[@B15], [@B16]\] or nonrandomized control group \[[@B18]\] or measuring outcomes before and after implementation of the mHealth intervention \[[@B17]\]. The remaining four observational studies did not use a control group \[[@B17]--[@B20]\], and as such the outcomes of these studies are less reliable. Seven (70%) of the articles studied antenatal care attendance \[[@B13], [@B14], [@B16]--[@B19], [@B21]\]; two (20%) studied postnatal care attendance \[[@B15], [@B19]\]; and four (40%) studied childhood immunization rates \[[@B17], [@B19], [@B20], [@B22]\]. Eight (80%) of the studies used an mHealth intervention that sent reminders to seek care directly to patients \[[@B13]--[@B17], [@B19], [@B20], [@B22]\] and five (50%) sent educational messages to patients \[[@B13], [@B14], [@B16], [@B18], [@B19]\]. Three (30%) studies sent reminders to health workers to follow up with patients \[[@B17], [@B20], [@B21]\] and three (30%) studies used an mHealth tool to improve patient records or identification \[[@B17], [@B20], [@B21]\]. The frequency of these interventions varied widely; educational messages were sent on schedules ranging from daily \[[@B18]\] to twice per month \[[@B14]\]. Some studies specified that appointment reminders were sent a few days in advance of a scheduled appointment \[[@B15], [@B17], [@B22]\] and others did not specify how far in advance patients or health workers were reminded. Eight (80%) of the studies were conducted in Africa \[[@B13]--[@B15], [@B18]--[@B22]\] and two (20%) were conducted in Asia \[[@B16], [@B17]\]. All included studies were published between 2010 and 2014, suggesting that the inclusion criterion of studies published after 2000 was sufficiently conservative and that it is unlikely that any relevant articles were missed from earlier publication dates. One study is taken from a literature review published in English on mHealth tools in China \[[@B16]\]; however, the original study was published in Chinese. Therefore, the information available on this study is less complete than that provided for the studies where the primary publication was included. 3.2. Findings by Intervention {#sec3.2} ----------------------------- All studies showed some evidence that the mHealth intervention implemented had a positive impact on patient or health worker behavior. However, the quality of the studies varied and some of these outcomes cannot be conclusively attributed to the mHealth intervention that was implemented from these studies alone. ### 3.2.1. Antenatal Care Attendance {#sec3.2.1} Two of the seven studies examining antenatal care attendance were RCTs \[[@B13], [@B14]\]. Both studies used text message reminders and education for pregnant women and one also provided the women with mobile-phone vouchers to contact their health worker, if needed \[[@B14]\]. Both studies found a statistically significant increase of over 10% in the proportion of women receiving at least four antenatal care visits between the intervention and control groups. Another study examined antenatal care attendance before and after implementation of an mHealth application for improved patient records and automated appointment reminders; this study similarly found a statistically significant improvement in on-time antenatal care attendance following implementation \[[@B17]\]. A study conducted in China sent text message reminders for antenatal care and health advice to an intervention group and found a statistically significant increase in antenatal care attendance, compared to a historic control group. The remaining studies examining antenatal care attendance found some self-reported behavior change from both patients and health workers \[[@B18], [@B19], [@B21]\]. ### 3.2.2. Postnatal Care Attendance {#sec3.2.2} One study examining postnatal care attendance used a historic control group from the previous 6 months in the same hospital and found that the intervention group, receiving text message appointment reminders, were 50% less likely to fail to attend their appointment (*P* = 0.002) \[[@B15]\]. Another study found that women self-reported intended or actual behavior change, including increased attendance to postnatal care, after receiving voice or SMS messages with education and reminders \[[@B19]\]. ### 3.2.3. Childhood Immunization {#sec3.2.3} A study examining childhood immunization found a statistically significant increase (from 34.5% to 44.2%, *P* \< 0.001) in the proportion of children receiving on-time vaccination after implementation of a mobile application for improved patient records and automated text message appointment reminders \[[@B17]\]. Another study found that mothers reported being influenced by a text message reminder (which were also tied to a conditional cash transfer, if child was vaccinated on time) to bring their child for immunization \[[@B22]\]. In one study, after receiving SMS or voice reminders and education, mothers self-reported intended or actual behavior change, including bringing their child for vaccines \[[@B19]\]. Finally, a study using an mHealth application to improve records and to send reminders during a mass vaccination campaign found that 92% of children visited at home following the campaign had received the measles vaccine \[[@B20]\]. 3.3. Findings on Cost {#sec3.3} --------------------- Two of the studies included information on the cost of their mHealth interventions. Adanikin et al. reported a total cost of only US\$21.12 to send 2252 SMS reminders for postnatal care during the six-month study in Nigeria \[[@B15]\]. Ngabo et al. cited initial investment cost as being considerable, largely due to the fact that they provided all community health workers in Rwanda with a mobile phone to "boost engagement and motivation of CHWs." However, ongoing costs were lowered by Ministry of Health negotiations with the private sector, reducing SMS costs from US\$0.05 to US\$0.005 per message \[[@B21]\]. 4. Discussion {#sec4} ============= Though all included studies showed some evidence that mHealth tools can be effective in changing patient and health worker behavior to increase antenatal care attendance, postnatal care attendance, and childhood immunization rates, the quality of the evidence varied widely. The strongest evidence exists for text message reminders and education delivered to pregnant women\'s mobile phones. The two RCTs that examined this intervention both found evidence of statistically significant increases in antenatal care attendance in their intervention groups, relative to their control groups \[[@B13], [@B14]\]. There is also some suggestion from the results that this intervention may also be effective when applied to the other health issues studied, such as postnatal care attendance and childhood immunization. Though no RCTs studied these health issues, two observational studies of high quality found evidence of effectiveness. Adanikin et al. found that intervention group receiving text message reminders for postnatal care were 50% less likely to fail to attend their appointments than a historic control group from the previous six months (*P* = 0.002) \[[@B15]\]. Kaewkungwal et al. found that after implementation of a smartphone application that supported record keeping and generated text message reminders for health workers and mothers, there was a 15% increase in women attending ANC on time (*P* \< 0.001) and a 10% increase in children receiving on-time immunizations (*P* \< 0.001) \[[@B17]\]. Beyond these findings, much of the evidence is based on self-reported behavior change from health workers and patients, which is not sufficiently reliable to draw any strong conclusions on the effectiveness of mHealth interventions \[[@B19], [@B21], [@B22]\]. Some other observational studies demonstrated good results of their programs, such as 92% confirmed coverage in a measles vaccine campaign \[[@B20]\]; however, it is impossible to determine which factors influenced the campaign\'s success and whether it was due to the use of an mHealth intervention or one of the other program components. In addition, several studies combined multiple mHealth interventions (e.g., text message reminders and conditional cash transfers via mobile phone \[[@B22]\]), making it impossible to determine to what degree each intervention influenced the resulting behavior change. As a result of these methodological limitations and the small number of studies meeting the inclusion criteria, further randomized controlled trials are needed to evaluate the effectiveness of mHealth tools for antenatal care, postnatal care, and childhood immunizations. By employing a multiarm or factorial design, researchers may be able to better ascertain which components of mHealth interventions are most effective. It is also worth noting that many of the mHealth tools studied focused on a single period of time on the maternal, neonatal, and child health (MNCH) continuum. For example, one study focused only on postnatal care, while five others focused only on antenatal care. Given the importance of continued follow-up of families during pregnancy, delivery, postnatal periods, and early childhood, it would be advisable for future mHealth interventions to consider expanding their tools to include more key events along the MNCH continuum \[[@B22]\]. Finally, the majority of these studies were conducted in Africa, suggesting that there is a need for future study of mHealth tools in broader contexts, including Asia and the Pacific, Central and South America, the Caribbean, and other regions. This literature review has provided us with the key knowledge that there is some existing evidence of the effectiveness of text message reminders for antenatal care, postnatal care, and immunizations and it has also helped to identify that this is an area where further research is needed. Given the limited, but largely positive, results of this literature review, researchers and public health practitioners should continue to implement mHealth tools for antenatal care attendance, postnatal care attendance, and childhood immunization. However, careful evaluation and further research are still needed to better determine how effective these tools are and in which settings. 5. Conclusions {#sec5} ============== Based on a systematic review of the literature, there is some evidence that mHealth tools may present an opportunity to influence behavior change and ensure that women and children in low-income countries are accessing prevention services, including antenatal care, postnatal care, and immunizations. Though mHealth programs have been implemented in low- and middle-income countries all over the world \[[@B23]\], there are few peer-reviewed studies and the majority of evaluations relating to maternal and child health have been conducted in Africa. Therefore, greater emphasis needs to be put on the evaluating mHealth tools and disseminating results to inform program design and policy making. In addition, many existing interventions focus on only one component of maternal and child health preventive services, rather than on design of an integrated system that follows women and children through the maternal, neonatal, and child health continuum \[[@B24]\]. The field of mHealth should continue to be supported and studied as it shows promise of improving the lives of women and children in low- and middle-income countries. Conflict of Interests ===================== The authors declare that there is no conflict of interests regarding the publication of this paper. ![PRISMA flow diagram \[[@B26]\].](BMRI2015-153402.001){#fig1} ###### Summary of included articles on mHealth interventions to increase use of antenatal care, postnatal care, and childhood immunization, classified by methods used. -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- First author,\ Title Health issue(s) studied Intervention studied and tools used Intervention frequency Key study outcomes Methods used Sample size Study location Study quality^1^ year ----------------------------------------------------------------- -------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------------------------------------------------- -------------------------------------------------------- ------------------------------------------------------------------------------ Randomized controlled trials (RCTs) Fedha, 2014 \[[@B13]\] "Impact of Mobile Telephone on Maternal Health Service Care: A Case of Njoro Division" Antenatal care attendance Text message reminders and educational messages for mother delivered to mobile phone.\ Appointment reminders every two weeks. Frequency of educational messages not specified 7.4% of women receiving SMS had less than 4 antenatal visits while 18.6% of those not receiving SMS had less than 4 visits (*P* = 0.002) Clinic attendance and antenatal service uptake compared for intervention and control groups Intervention group: 191\ Health facilities in Kenya RCT with low risk of bias No specific mHealth tools mentioned Control group: 206\ Total: 397 Lund, 2014 \[[@B14]\] "Mobile Phones Improve Antenatal Care Attendance in Zanzibar: A Cluster Randomized Controlled Trial" Antenatal care attendance Text message reminders and educational messages for mother delivered to mobile phone and mobile vouchers to contact health workers.\ Two messages per month before gestational week 36 and two messages per week after week 36 44% of women in the intervention group received the recommended four or more antenatal visits, compared with 31% in the control group. The odds for receiving four or more antenatal care visits were 2.39 (1.03--5.55) for women benefitting from the mobile phone intervention. 59% of intervention women stated that received text messages influenced the number of times they attended antenatal care Clinic attendance was compared for cluster randomized intervention and control groups Intervention group: 1311\ Urban and rural healthcare facilities in Zanzibar RCT with low risk of bias Tools used: custom Wired Mothers software Control group: 1239\ Total: 2550 Studies with nonrandomized control group or before/after design Adanikin, 2014\ "Role of Reminder by Text Message in Enhancing Postnatal Clinic Attendance" Postnatal care attendance Text message reminders for mother delivered to mobile phone.\ Two messages sent for each appointment: two weeks prior and 5 days prior Patients who received an SMS reminder were 50% less likely to fail to attend (FTA) their postnatal appointment (relative risk of FTA 0.50; 95% CI, 0.32--0.77; *P* = 0.002) Clinic attendance compared for intervention group and historic control group (from previous 6 months) Intervention group: 1126\ Teaching hospital in Nigeria 7/9 \[[@B15]\] No specific mHealth tools mentioned Control group: 971\ Total: 2097 Fang and Li, 2010 \[[@B27]\] from\ "Mobile Health in China: A Review of Research and Programs in Medical Care, Health Education, and Public Health" Antenatal care attendance Text message appointment reminders and antenatal health advice.\ Four appointment reminders per pregnancy. Frequency of health advice not specified The intervention group received 5.7 ± 1.8 antenatal visits, compared to 3.2 ± 1.1 antenatal visits in the control group (*P* \< 0.01) Clinic attendance compared for intervention group and historic control group (from previous year). Intervention group: 609\ China Unable to determine as not all info. on study design is available in English Corpman, 2013 \[[@B16]\] No specific mHealth tools mentioned Control group: 637\ Total: 1246 Kaewkungwal, 2010 \[[@B17]\] "Application of Smart Phone in "Better Border Healthcare Program": A Module for Mother and Child Care" Antenatal care attendance and childhood immunization (EPI) Smartphone application used by health workers to update antenatal and immunization status when outside clinic and SMS reminders for both health workers and mothers.\ Appointment reminders a few days prior to scheduled appointment 58.68% of pregnant women came to ANC on time after implementation as compared to 43.79% before (*P* \< 0.001). After adjusting for personal characteristics, sending appointment message increased odds of on-time visit by 2.97 (1.60--5.54). 44.22% of children received scheduled vaccines on time after implementation as compared to 34.49% before (*P* \< 0.001). After adjusting for personal characteristics, follow-up cases and updating immunization data on cell phones increased odds of on-time EPI by 2.04 (1.66--2.52). Sending appointment reminder increased odds of on-time EPI by 1.48 (1.09--2.03) Clinic attendance for ANC and EPI were compared before and after MCCM implementation ANC group: 280\ Rural border area in Thailand, near Myanmar 8/9 Tools used: custom Mother and Child Care Module (MCCM) EPI group:\ 544 Lau, 2014 \[[@B18]\] "Antenatal Health Promotion via Short Message Service at a Midwife Obstetrics Unit\ Antenatal care attendance Text messages with antenatal health information.\ Varied from three messages per week to daily messages 92% of participants in the intervention group reported not missing more than two antenatal visits. A focus group of intervention participants reported that they had improved health related behaviors, including attending the clinic regularly, as a result of the text messages. No statistically significant difference in knowledge was seen between the intervention and control groups at the exit interview Baseline questionnaire and exit interview were administered to convenience-sampled intervention and control groups to assess knowledge of antenatal health and clinic procedures. A focus group was conducted with a further convenience sample of the intervention group Intervention group: 102 but 45 were lost to follow-up\ Urban primary care facility in Cape Town, South Africa 4/9 in South Africa: A Mixed Methods Study" No specific mHealth tools mentioned Control group: 104 but 43 were lost to follow-up\ Total: 206 recruited, 118 included in analysis Studies with no control group Crawford, 2014\ "SMS versus Voice Messaging to Deliver MNCH Communication in Rural Malawi: Assessment of Delivery Success and User Experience" Antenatal care attendance, postnatal care attendance, and childhood immunization Text (SMS) or voice message reminders and educational messages for mother delivered to mobile phone or retrieved by calling a toll-free hotline.\ Once (voice) or twice (SMS) per week 91% of SMS enrollees surveyed reported that they had already changed or intended to change their behavior based on the messages, including attending more ANC/PNC or bringing their child for vaccines. SMS enrollees were significantly more likely to report intended or actual behavior change than voice enrollees Phone based surveys of participants. Participants in the pushed SMS and pushed voice groups were randomly sampled but participants in the retrieved voice group were convenience sampled Pushed SMS: 96\ Rural health centers in Malawi 2/9 \[[@B19]\] Tools used: Village-Reach custom application (SMS) and INTELLIVR software (voice messages) Pushed voice: 30\ Retrieved voice: 140\ Total: 266 Mbabazi, 2014\ "Innovations in Communication Technologies for Measles Supplemental Immunization Activities: Lessons from Kenya Measles Vaccination Campaign, November 2012" Childhood immunization Smartphone application used by volunteers to update immunization records when canvassing door-to-door and to provide text message and phone call reminders to caretakers.\ Varied/as needed In precampaign house-to-house visits, 25% of households had no plans to bring their children for the measles supplemental dose if they had not been contacted by the volunteers. Of the children found in the postcampaign house visits, 96% reported to have received a measles supplemental immunization dose, although only 92% had confirmation (finger mark) of vaccination Precampaign household canvassing and data collection for entire target population, followed by postcampaign verification of vaccine coverage Precampaign: 164,643 households with 161,695 children\ Urban areas in Kenya 5/9 \[[@B20]\] Tools used: EpiSurveyor Postcampaign: 17,627 households with 17,993 children Ngabo, 2012 \[[@B21]\] "Designing and Implementing an Innovative SMS-based Alert System (RapidSMS-MCH) to Monitor Pregnancy and Reduce Maternal and Child Deaths in Rwanda" Antenatal care attendance Electronic registration of pregnant women through text messages by community health workers (CHWs) and reminder text messages for antenatal care sent to CHWs\' mobile phones.\ As needed for upcoming antenatal visits and estimated delivery date 81% of the estimated annual pregnancies in the district were registered in the system. Reporting compliance among CHWs was 100%. CHWs reported being more proactive in finding new pregnant women and following up registered pregnant women as a result of reminders forwarded to their mobile phones Reporting compliance, system usage patterns, and error rates were monitored and feedback sessions were held with CHWs CHWs: 432 Rural district of Rwanda N/A, only process outcomes were studied Tools used: customized version of RapidSMS Wakadha, 2013\ "The Feasibility of Using Mobile-Phone Based SMS Reminders and Conditional Cash Transfers to Improve Timely Immunization in Rural Kenya" Childhood immunization Text message reminders for mother delivered to mobile phone and free airtime or mobile cash transfers for mothers that brought child in on time.\ Three days before vaccine due date and on due date 91% of mothers reported that the SMS reminders influenced their decision to come in for vaccination Enrolled mothers were randomized to receive either mMoney or airtime for on-time vaccinations.\ mMoney group: 48\ Rural district of Kenya 4/9 \[[@B22]\] Tools used: customized version of RapidSMS and mPESA Questionnaires were administered in home follow-up visits Airtime group: 24\ Total: 72 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ^1^Quality score assigned using the Cochrane Risk of Bias Assessment Tool (for RCTs) or the Newcastle-Ottawa Quality Assessment Scale (for observational studies). For RCTs, a low risk of bias is the best possible score and for observational studies the highest possible score is 9. Please see [Section 2](#sec2){ref-type="sec"} for more details. [^1]: Academic Editor: Pascale Allotey

{ "pile_set_name": "PubMed Central" }

1. Introduction {#sec1-genes-11-00418} =============== Lilies (*Lilium* spp.) have high commercial and ornamental value due to their various floral colors and coloration patterns. Anthocyanins are one of the major pigments affecting the coloration of the floral tepals in *Lilium* spp. \[[@B1-genes-11-00418],[@B2-genes-11-00418],[@B3-genes-11-00418],[@B4-genes-11-00418]\]. Anthocyanins are synthesized in the cytosol but their transfer to the vacuole is necessary for plant tissues to exhibit brilliant colors \[[@B5-genes-11-00418],[@B6-genes-11-00418]\]. The mechanism of anthocyanin transfer is still unclear, although anthocyanin biosynthesis at the molecular level has been studied extensively leading to the discovery of enzyme-coding structural and regulatory genes related to anthocyanin biosynthesis \[[@B7-genes-11-00418],[@B8-genes-11-00418]\]. To date, glutathione S-transferases (GSTs), multidrug and toxic compound extrusion (MATE) proteins, and the ATP-binding cassette (ABC) transporter have been proposed to be involved in anthocyanin transportation in many species \[[@B9-genes-11-00418],[@B10-genes-11-00418],[@B11-genes-11-00418]\]. MATE transporters perform various transport functions in plants, including the transportation of anthocyanin \[[@B12-genes-11-00418]\]. A MATE transporter related to flavonoid accumulation was isolated for the first time during screening of a *TT12 Arabidopsis* mutant with altered seed coloration \[[@B13-genes-11-00418]\]. Subsequently, orthologs of *AtTT12* were obtained from many other species, such as upland cotton, blueberry, *Medicago truncatula*, apple, and grapevine, and orthologs of the *TT12* gene were verified to be related to flavonoid transportation \[[@B6-genes-11-00418],[@B14-genes-11-00418],[@B15-genes-11-00418],[@B16-genes-11-00418],[@B17-genes-11-00418]\]. To date, the enzyme-coding structural genes and regulatory genes of anthocyanin biosynthesis in *Lilium* have been characterized thoroughly \[[@B2-genes-11-00418],[@B4-genes-11-00418],[@B18-genes-11-00418],[@B19-genes-11-00418],[@B20-genes-11-00418]\]; however, whether MATE transport is involved in anthocyanin transportation in *Lilium* remains unclear. In this study, based on our RNA-seq data published previously \[[@B21-genes-11-00418]\], we isolated a flavonoid transport-related MATE gene, *LhDTX35*, from *Lilium* 'Tiny Padhye'. Then, expression profiling, identification, and functional analysis of this gene in anthocyanin transportation were performed using 'Tiny Padhye' as a subject. This study demonstrates that a gene encoding MATE is critical for anthocyanin accumulation in *Lilium*, and promotes an understanding of the mechanism underlying MATE involvement in anthocyanin transportation in *Lilium*. 2. Materials and Methods {#sec2-genes-11-00418} ======================== 2.1. Plant Material {#sec2dot1-genes-11-00418} ------------------- The Asiatic lily cultivar 'Tiny Padhye' was grown in a greenhouse (25 °C, relative humidity 60%--70%, 16 h/8 h light/dark cycles) at the Chinese Academy of Agricultural Sciences (Beijing, China). The upper parts and bases of the inner tepals were collected at four different developmental stages. The tepal developmental stages were defined as described in Xu LF et al. \[[@B21-genes-11-00418]\]: Stage 1 (S1: no anthocyanin pigment is visible in tepals), stage 2 (S2: anthocyanin pigment becomes visible on tepals), stage 3 (S3: the day before anthesis, lower halves of tepals are fully pigmented), and stage 4 (S4: the first day after anthesis) ([Figure 1](#genes-11-00418-f001){ref-type="fig"}). Samples were obtained from 15 flowers and pooled together as one biological sample; three independent biological replicates were collected for each stage. *Nicotiana tabacum* was also grown in a greenhouse, as described above for environmental conditions. Seedlings at the four-true-leaf stage (6-week-old) were used for the subcellular localization experiment. 2.2. Isolation of a MATE-Like Gene {#sec2dot2-genes-11-00418} ---------------------------------- Six primers (LhMATE-YZ-F1, LhMATE-YZ-R1, LhMATE-3'RACE-GSP, LhMATE-3'RACE-nest, LhMATE-5'RACE, and LhMATE-5'RACE-nest) were designed based on unigenes encoding MATE-like proteins from our previous RNA-seq experiment ([Table 1](#genes-11-00418-t001){ref-type="table"}). First-strand cDNA was cloned using RNA isolated from the colored parts of the tepal at stage 2 as a template, according to the RACE kit manufacturer's instructions (Takara, Dalian China). To obtain the full-length MATE cDNA, the 5'-end and 3'-end sequences of the identified gene were identified using stage 2 total RNA from the basal (colored) tepal regions and a RACE kit, according to the recommendations of the manufacturer (Takara, Dalian, China). Five independent RACE cDNA clones were sequenced by Sangon Biotech (Shanghai, China). Based on the 5'-RACE and 3'-RACE sequencing results, the primers LhDTX35-F and LhDTX35-R ([Table 1](#genes-11-00418-t001){ref-type="table"}) were designed to amplify the entire gene region containing the open reading frame (ORF) from the first-strand cDNA. 2.3. Sequence Bioinformatics Analysis {#sec2dot3-genes-11-00418} ------------------------------------- A BLASTx search of MATE-like gene ORF sequences was conducted to analyze sequence similarity. Protein conserved domain prediction was performed by NCBI conserved domain online software \[[@B22-genes-11-00418]\]. The protein transmembrane domains were analyzed by TMHMM Server V.2.0 \[[@B23-genes-11-00418]\]. 2.4. Phylogenetic Analysis {#sec2dot4-genes-11-00418} -------------------------- The amino acid sequences of genes encoding MATEs from different plants were used for phylogenetic analysis. Sequence alignment was performed using CLUSTAL X. Based on this alignment, a Neighbor-Joining (NJ) tree was constructed using MEGA (version 7.0). Bootstrap values were calculated using 1000 replicate analyses. 2.5. Gene Expression Analysis {#sec2dot5-genes-11-00418} ----------------------------- Quantitative real-time polymerase chain reaction (qRT-PCR) was performed using SYBR Premix Ex Taq (Takara, Dalian, China). The primers used to amplify the MATE-like gene segment are shown in [Table 1](#genes-11-00418-t001){ref-type="table"}. The reaction parameters were as follows: (1) 95 °C for 1 min; (2) 40 cycles of 95 °C for 20 s, 60 °C for 10 s, and 72 °C for 25 s; and (3) a melt curve program (65 °C to 95 °C with an increment in temperature of 0.5 °C every 0.05 s). The signal was monitored using a CFX96 real-time system (Bio-Rad, CA, USA). The average Cq value was calculated from three biological and three technical replicates. To normalize the differences in the amounts of mRNA from other genes, the amount of *Lhactin* mRNA was determined for each sample and the relative expression level of *LhDTX35* was analyzed using the 2^−△△CT^ method \[[@B24-genes-11-00418]\]. The error bars represented the ±SEs from three independent experiments. The data were analyzed by ANOVA using SAS software. 2.6. Subcellular Localization Analysis {#sec2dot6-genes-11-00418} -------------------------------------- Using the transient expression vector pCAMBIA2300-35s-GFP, the subcellular localization of the *LhDTX35* gene was analyzed. To construct the expression vector 35S::*LhDTX35*-GFP, primers containing the XbaI and BamHI sites were designed to amplify the *LhDTX35* gene ORF domain using the primer LhDTX35-DW-F/R ([Table 1](#genes-11-00418-t001){ref-type="table"}). The PCR products were separated on a 1% agarose gel and then purified. The empty vector pCAMBIA2300-35s-GFP was digested with the restriction enzymes XbaI and BamHI according to the manufacturer's instructions (Thermo Scientific FastDigest, America). Subsequently, the purified PCR product and cleaved empty vector were fused using T4 DNA Ligase (TransGen, China). Positive clones were selected and confirmed by sequencing and then transformed into *Escherichia coli* DH5α cells (TransGen, China). The sequencing primers (DW-YZ-F/R) are shown in [Table 1](#genes-11-00418-t001){ref-type="table"}. Finally, the plasmids pCAMBIA2300-GFP and pCAMBIA2300-GFP-*LhDTX35* were each transformed into *Agrobacterium tumefaciens* (LBA4404) using the freeze-thaw method, as previously reported \[[@B25-genes-11-00418]\]. When *N. tabacum* had developed four true leaves, the underside of the leaves was infiltrated with *Agrobacterium* inocula using a 1 mL needleless syringe. The inoculated plants were grown in a phytotron for 24 h in the dark (25 °C, relative humidity 60%--70%). Then, the plants were grown under 16 h/8 h light/dark cycles for 5 days. Subsequently, the LhDTX35-GFP fusion protein in the *N. tabacum* epidermis cells was microscopically detected using a Leica confocal laser scanning microscope. The samples were illuminated with an argon ion laser using 488 nm light for GFP and a green HeNe laser using 543 nm light for chlorophyll autofluorescence. 2.7. Complementation Analysis {#sec2dot7-genes-11-00418} ----------------------------- The 35S::*LhDTX35* vector was constructed by replacing the GUS gene in the pCAMBIA3301 vector with the coding sequence of *LhDTX35*. The *LhDTX35* ORF was amplified by PCR with primers LhDTX35-noci-F and LhDTX35-Bahm-R ([Table 1](#genes-11-00418-t001){ref-type="table"}). The *LhDTX35* ORF fragments were cloned into the linearized vector pCAMBIA3301 and digested with the NocII and HindII restriction enzymes using a one-step seamless cloning kit (Transgen, Beijing, China). The binary vectors were introduced into *A. tumefaciens* GV3101 and subsequently used in *Arabidopsis ttDTX35* mutant transformation with the floral dip method \[[@B26-genes-11-00418]\]. Seeds of the *Arabidopsis LhDTX35* mutant, transgenic lines, and wild-type *Arabidopsis* were germinated and grown on Murashige & Skoog Basic Medium 1/2 Macro (1/2 MS) medium. Transgenic plants were screened on 1/2 MS medium plates that contained 50 mg/L kanamycin. The expression of *LhDTX35* was analyzed by RT-PCR with primers LhDTX35-YZ-F/R, and *Arabidopsis Actin* gene was control with primers Actin-F/R ([Table 1](#genes-11-00418-t001){ref-type="table"}). 3. Results {#sec3-genes-11-00418} ========== 3.1. Cloning and Sequence Analysis of the LhDTX35 Gene {#sec3dot1-genes-11-00418} ------------------------------------------------------ The full-length cDNA sequence of *LhMATE*-like was obtained by RACE and deposited in GenBank with the accession number MT001433. This sequence contained 1948 bp and encoded 507 amino acids, as predicted by the NCBI ORF Finder tool ([Figure 2](#genes-11-00418-f002){ref-type="fig"}). BLASTx results showed high homology to the *DTX35* genes of other species, such as *Dendrobium catenatum*, *Elaeis guineensis*, *Phoenix dactylifera*, and *Oryza sativa* Japonica Group (ca. 71%--76.36% identity at the amino acid level). Therefore, the *LhMATE*-like cDNA isolated in this study was named *LhDTX35*. Conserved domain prediction revealed that *LhDTX35* encoded functional domains typical of the MATE-like superfamily between 258 bp and 1562 bp ([Figure 3](#genes-11-00418-f003){ref-type="fig"}). Transmembrane domain analysis predicted that the protein encoded by *LhDTX35* was a membrane protein with twelve typical transmembrane domains ([Figure 4](#genes-11-00418-f004){ref-type="fig"}). A phylogenetic analysis revealed that *LhDTX35* was closely related to the DTX35 protein in *Dendrobium catenatum*, sharing 76.36% amino acid identity ([Figure 5](#genes-11-00418-f005){ref-type="fig"}). In conclusion, the isolated *LhDTX35* gene was predicted to encode a membrane protein belonging to the MATE-like superfamily. 3.2. Expression Analysis of LhDTX35 {#sec3dot2-genes-11-00418} ----------------------------------- Anthocyanin accumulated in only the basal tepal regions of 'Tiny Padhye'. We previously showed that anthocyanin content increased gradually during tepal development ([Figure 6](#genes-11-00418-f006){ref-type="fig"}) \[[@B21-genes-11-00418]\]. Anthocyanin accumulation was the highest in tepals at stage 3 ([Figure 6](#genes-11-00418-f006){ref-type="fig"}) \[[@B21-genes-11-00418]\]. To exploit whether the *LhDTX35* gene was involved in anthocyanin accumulation in *Lilium*, qRT-PCR was performed using the upper and basal tepal regions at stages 1--4. These results showed that *LhDTX35* gene expression was positively correlated with anthocyanin accumulation, which occurred in the bases of the tepals ([Figure 7](#genes-11-00418-f007){ref-type="fig"}). 3.3. Subcellular Localization of LhDTX35 {#sec3dot3-genes-11-00418} ---------------------------------------- A subcellular localization experiment was performed to further identify whether the protein encoded by the *LhDTX35* gene was a membrane protein. The results showed that GFP fluorescence was distributed in the cytoplasm of *N. tabacum* leaf epidermal cells treated with the empty GFP vector ([Figure 8](#genes-11-00418-f008){ref-type="fig"}a,b). In contrast, GFP fluorescence was located in the membrane of *N. tabacum* leaf epidermis cells treated with the *LhDTX35* and GFP fusion protein vector ([Figure 8](#genes-11-00418-f008){ref-type="fig"}c,d). These results indicated that the *LhDTX35* gene encoded a membrane protein. 3.4. Functional Analysis of the LhDTX35 Gene in the Arabidopsis DTX35 Gene Mutant {#sec3dot4-genes-11-00418} --------------------------------------------------------------------------------- An *Arabidopsis* genetic mutant lacking *DTX35* was selected to investigate the functionality of the *LhDTX35* gene. The coding sequence of the *LhDTX35* gene was transferred into the *Arabidopsis DTX35* mutant under the control of the cauliflower mosaic virus (CaMV) 35S promoter. The hypocotyls of wild-type plants and transgenic lines expressing *LhDTX35* were red. In contrast, the hypocotyls of the *Arabidopsis DTX35* mutant were green, except for the junctions between the cotyledons ([Figure 9](#genes-11-00418-f009){ref-type="fig"}B). The silique fertility of the *Arabidopsis DTX35* mutant was low but normal fertility was restored in the complementation experiment ([Figure 9](#genes-11-00418-f009){ref-type="fig"}C). These results showed that the function of the *LhDTX35* gene was similar to that of the *DTX35* gene in *Arabidopsis*. Furthermore, the *LhDTX35* gene may be involved in anthocyanin accumulation. 4. Discussion {#sec4-genes-11-00418} ============= *Lilium* has high ornamental value due to its wide variety of coloration patterns. Anthocyanins are important pigments responsible for *Lilium* flower pigmentation ranging from pink to purple. Therefore, the characterization of anthocyanins in *Lilium* is of considerable ongoing interest. In *Lilium*, the steps of anthocyanin biosynthesis and regulation are well established; however, the anthocyanin transportation mechanism remains unclear. In recent years, several transport proteins have been reported to be involved in anthocyanin transportation \[[@B9-genes-11-00418],[@B10-genes-11-00418],[@B11-genes-11-00418]\]. Among them, MATE is an important protein that transports anthocyanins. To date, whether the MATE protein is involved in anthocyanin transportation in *Lilium* is unknown. In the present study, a flavonoid transport-related MATE candidate gene, *LhDTX35*, was cloned successfully based on our previous transcriptome data. Phylogenetic analysis revealed that the *LhDTX35* gene clustered with *DTX35* genes from *Dendrobium catenatum, Phoenix dactylifera,* and *Oryza brachyantha* ([Figure 5](#genes-11-00418-f005){ref-type="fig"}). *LhDTX35* is closely related to the *DTX35* gene in *Dendrobium catenatum*, sharing 76.36% amino acid identity. Bioinformatics analysis revealed that the *LhDTX35* gene contained sequences encoding the conserved domains typical of the MATE superfamily ([Figure 3](#genes-11-00418-f003){ref-type="fig"}), suggesting that the protein encoded by *LhDTX35* belonged to the MATE family. In addition, MATEs are important transport proteins that possess 12 typical transmembrane domains, and transmembrane structure prediction showed that *LhDTX35* also has 12 typical transmembrane domains ([Figure 4](#genes-11-00418-f004){ref-type="fig"}). These results indicated that *LhDTX35* is a MATE protein in *Lilium*. The expression profiles of anthocyanin biosynthetic genes and regulatory genes during anthocyanin accumulation at stages 1--4 were previously analyzed in 'Tiny Padhye', and the results showed that anthocyanin accumulation in tepals increased gradually at stages 1--4, reaching its highest level at stage 3 \[[@B21-genes-11-00418]\]. In this study, we revealed that the expression level of *LhDTX35* was consistent with anthocyanin accumulation at stages 1--4 ([Figure 7](#genes-11-00418-f007){ref-type="fig"}). Weak expression of *LhDTX35* was detected in the tepals at stage 1, when the anthocyanin content was lowest. In conjunction with the increasing anthocyanin content in the tepals, the transcript levels of *LhDTX35* were gradually upregulated at stages 1--4, and both anthocyanin content and *LhDTX35* gene transcript levels peaked at stage 3 ([Figure 6](#genes-11-00418-f006){ref-type="fig"} and [Figure 7](#genes-11-00418-f007){ref-type="fig"}). All the results indicated that *LhDTX35* expression was positively correlated with anthocyanin accumulation in *Lilium*, suggesting that the *LhDTX35* gene may be involved in the transport of anthocyanin in *Lilium*. All the plant MATE proteins characterized to date have been localized to membranes \[[@B27-genes-11-00418]\]. To investigate whether *LhDTX35* is involved in anthocyanin accumulation, its subcellular protein localization was examined. The results showed that LhDTX35 was localized to the membrane but not the cytoplasm, further suggesting that the *LhDTX35* gene encoded a protein with the traits of MATE transport proteins. These results were also consistent with the results of the transmembrane structure prediction, that is, the protein encoded by *LhDTX35* was predicted to have 12 typical transmembrane domains ([Figure 4](#genes-11-00418-f004){ref-type="fig"}). Complementation experiment is an effective method to analyze gene function because genetic transformation is a challenge for *Lilium*. Complementation has been used for this purpose in many species \[[@B9-genes-11-00418],[@B11-genes-11-00418],[@B15-genes-11-00418]\]. In this study, the silique-infertility phenotype of the *DTX35* mutant was recovered when *LhDTX35* was overexpressed in the *DTX35* mutant of *Arabidopsis*, suggesting that the *LhDTX35* gene has a similar function to that of *DTX35* in *Arabidopsis*. The *DTX35* gene is also called FFT (flower flavonoid transporter) \[[@B28-genes-11-00418]\]. The anthocyanin content of the *FFT* mutant was slightly less than that of the WT (Wile Type) in seedlings at one week of age \[[@B28-genes-11-00418]\]. Interestingly, the present study showed that the anthocyaninless phenotype of the *DTX35* mutant was rescued by the complementation experiment ([Figure 9](#genes-11-00418-f009){ref-type="fig"}B). These results suggested that the *LhDTX35* gene was required for anthocyanin transportation in *Lilium*. MATE genes are part of a gene superfamily, including 58 MATEs in *Arabidopsis* and 53 MATEs in rice \[[@B29-genes-11-00418]\]. In the present study, only one MATE gene was cloned successfully, and whether any other MATE genes are involved in anthocyanin transportation in *Lilium* should be further studied. 5. Conclusions {#sec5-genes-11-00418} ============== In conclusion, we first identified a flavonoid transport-related MATE candidate gene based on transcriptome data in *Lilium* 'Tiny Padhye' and found that this gene shared high similarity with published *DTX35* genes. Bioinformatics and phylogenetic analysis revealed that *LhDTX35* belongs to the MATE gene family. Expression profiling showed that the expression of *LhDTX35* was positively correlated with anthocyanin accumulation in the tepals of 'Tiny Padhye'. Subcellular localization results showed that the *LhDTX35*-encoded protein was localized to the membrane, suggesting that the protein encoded by the *LhDTX35* gene possessed the traits of a MATE transporter. Silique-infertility and anthocyaninless hypocotyl phenotypes were recovered after the complementation experiment, indicating that *LhDTX35* has a similar function to that of the *DTX35* gene in *Arabidopsis* and is involved in anthocyanin transportation in *Lilium*. Our results provide insight into the mechanism of anthocyanin transportation in *Lilium*. We are grateful to Haiwen Zhang (Beijing Academy of Agricultural and Forestry sciences) for kindly offering *Arabidopsis DTX35* mutant. Conceptualization, J.M. and L.X.; methodology, H.X.; investigation, P.Y., Y.T., Y.C., and G.H.; writing---original draft preparation, H.X.; writing---review and editing, L.X. and J.M.; funding acquisition, J.M. and H.X. All authors have read and agreed to the published version of the manuscript. This research was funded by the National Natural Science Foundation of China, grant number 31672196, 31902048, and this research was also funded by the Fundamental Research Funds for Central Non-profit Scientific Institutions and the Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences. The authors declare no conflict of interest. ![Tepals of the Asiatic lily cultivar 'Tiny Padhye' at four different developmental stages.](genes-11-00418-g001){#genes-11-00418-f001} ![Nucleotide and deduced amino acid sequences of the *LhDTX35* gene. The initiation codon is underlined, while the stop codon is labeled with an asterisk and box.](genes-11-00418-g002){#genes-11-00418-f002} ![Conserved domain prediction for the protein encoded by *LhDTX35.*](genes-11-00418-g003){#genes-11-00418-f003} ![Transmembrane prediction for the protein encoded by *LhDTX35.*](genes-11-00418-g004){#genes-11-00418-f004} ![Phylogenetic tree of the predicted amino acid sequences of the *Lilium* multidrug and toxic compound extrusion (MATE) protein and MATE proteins from other plants. The phylogenetic tree was constructed using MEGA 6.0 software and the NJ method. The red dot indicates the *Lilium* MATE-like protein identified in this study.](genes-11-00418-g005){#genes-11-00418-f005} ![Changes in anthocyanin content at different developmental stages in *Lilium* \[[@B21-genes-11-00418]\].](genes-11-00418-g006){#genes-11-00418-f006} ![Spatiotemporal analysis of *LhDTX35* gene expression at different stages of tepal development in *Lilium* 'Tiny Padhye'. (**a**) Expression profiles of *LhDTX35* at stages 1--4 in the basal tepal region in *Lilium* 'Tiny Padhye'; (**b**) tepal at stage 2; (**c**) expression profiles of *LhDTX35* at stage 2 between the upper and basal tepal regions in *Lilium* 'Tiny Padhye'.](genes-11-00418-g007){#genes-11-00418-f007} ![Subcellular localization of the *LhDTX35*-GFP fusion protein. (**a**,**c**) Bright field images; (**b**,**d**) GFP fluorescence images; (**a**,**b**) cells expressing the *LhDTX35*-GFP fusion protein; (**c**,**d**) cells expressing 35S::GFP as a control.](genes-11-00418-g008){#genes-11-00418-f008} ![Functional complementation of the *Arabidopsis DTX35* mutant with the *LhDTX35* gene driven by the CaMV 35S promoter. (**A**) Expression of *LhDTX35* in wild-type (**a**), mutant (**b**), and transgenic *Arabidopsis* seedlings (**c**--**e**). (**B**) Anthocyanin accumulation in wild-type (**a**), mutant (**b**), and transgenic *Arabidopsis* seedlings (**c**--**e**). (**C**) Silique fertility of wild-type (**a**), mutant (**b**), and transgenic *Arabidopsis* seedlings (**c--e**).](genes-11-00418-g009){#genes-11-00418-f009} genes-11-00418-t001_Table 1 ###### All primers used in this study. ---------------------------------------------------------------------- Primers Primer Sequence (5′-3′) --------------------- ------------------------------------------------ LhMATE-YZ-F1\ GTTGCGGTTATCACATCCCT\ LhMATE-YZ-R1 CTTCTTCCCCTCACAGTCTA LhDTX35-3'RACE-GSP\ CGGTGGGTGGCAAGGTCTGGTAGC\ LhDTX35-3'RACE-nest GGGCTATCCGTTACATTTGGGTGTGC LhDTX35-5'RACE,\ CCAAATCCCCTGCACACCCAAA\ LhDTX35-5'RACE-nest TGGTTCATTCTGTCCACCCCAC LhDTX35-F,\ AGTGGGAGAGAGAGAGAGGCGA\ LhDTX35-R CTTCTTCCCCTCACAGTCTATGC LhDTX35-YG-F\ CCAAGAGGATGACATTGCCGAG\ LhDTX35-YG-R TGGAGGATGAGGGTGAAGAAGC LhDTX35-DW-F\ CCCCCGGGATGGAAGATCCGCTTCTGAGAC\ LhDTX35-DW-R GCTCTAGACACTAACTTGACTTTGTTGGTT DW-YZ-F\ CGGGCTGTTGCGAAAATA\ DW-YZ-R TGCCGTTCTTCTGCTTGTC LhACT-YG-F\ GCATCACACCTTCTACAACG\ LhACT-YG-R GAAGAGCATAACCCTCATAGA Actin-F\ CGTGACCTTACTGATTACCT\ Actin-R AGCGATACCTGAGAACATAG LhDTX35-noci*-F*\ CGGGGGACTCTTGACCATGGAAGATCCGCTTCTGAGACAT\ LhDTX35-Bahm-R GGAAATTCGAGCTGGTCACCTGTAATCTAACTTGACTTTGTTGGTT LhDTX35-YZ-F\ CGGGCTGTTGCGAAAATA\ LhDTX35-YZ-R TGCCGTTCTTCTGCTTGTC ----------------------------------------------------------------------

{ "pile_set_name": "PubMed Central" }

1. Introduction =============== The use of artificial materials in medicine and biology has become very important, especially their application as substitutes for damaged tissues and organs, or as carriers for drug or gene delivery \[[@B1-materials-06-01632],[@B2-materials-06-01632],[@B3-materials-06-01632],[@B4-materials-06-01632]\]. These materials have to be biocompatible, *i.e.*, they have to match the mechanical properties of the replaced tissue, and not act as cytotoxic, mutagenic or immunogenic. For specific applications, biocompatible materials can behave as bioinert and not as promoting cell adhesion and growth. For example, materials of these types have been applied in the construction of artificial eye lenses \[[@B5-materials-06-01632]\] or in the articular surfaces of joint prostheses \[[@B6-materials-06-01632]\], which are implants requiring transparency or smoothness, and are thus completely cell-free surfaces. Bioinert materials have also been used for fabricating polymeric vascular prostheses in order to prevent the adhesion and activation of thrombocytes and immunocompetent cells on the inner surface of these grafts \[[@B7-materials-06-01632]\]. However, adhesion, spreading, growth and phenotypic maturation of cells are required for constructing advanced bioartificial tissue replacements, including replacements for blood vessels or for bone. In these replacements, the artificial materials should act as analogs of the natural extracellular matrix, and thus they should support regeneration of the damaged tissue. For example, in vascular tissue engineering, this means that the material should enable reconstruction of the *tunica intima*, formed by a confluent layer of endothelial cells, and also reconstruction of the *tunica media*, which contain vascular smooth muscle cells. For these purposes, advanced artificial materials should not just be passively tolerated by the cells, but should act as bioactive or biomimetic, which means that they induce the required cell responses in a controllable manner (for a review, see \[[@B8-materials-06-01632],[@B9-materials-06-01632]\]). It is generally known that the cell--material interaction is strongly dependent on the physical and chemical properties of the material surface, such as wettability, roughness and topography, or the presence of various chemical functional groups and biomolecules. From this point of view, the surfaces of most materials currently used for constructing tissue replacements are not optimal for integrating with the surrounding tissues, and need further modification in order to improve their physicochemical surface properties and thus enhance the cell colonization and new tissue formation (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632]\]). A typical example is provided by synthetic polymers (polyethylene, polypropylene, polystyrene, polyethyleneterephthalate or polytetrafluoroethylene), which are biomaterials widely used in various biotechnologies, including cell cultivation and construction of tissue replacements. In their non-treated state, however, some materials usually form inadequate scaffolds for cell colonization, often due to their relatively high hydrophobicity (water drop contact angle in the range of about 90--120°). These materials can be rendered more hydrophilic by several techniques, particularly by irradiation with ultraviolet (UV) light \[[@B10-materials-06-01632]\], an ion beam \[[@B11-materials-06-01632],[@B12-materials-06-01632],[@B13-materials-06-01632]\] or exposure to plasma discharge \[[@B14-materials-06-01632],[@B15-materials-06-01632],[@B16-materials-06-01632]\]. In addition to the changes in surface wettability, these treatments also affect the roughness, morphology, electrical conductivity, stability, mechanical properties and chemical composition of the material surface. A common feature of irradiation with UV light, ion beam or plasma treatment is splitting of the bonds within the polymer molecules, which results in the creation of free radicals, double bonds and new functional groups (especially oxygen-containing groups) on the polymer surface. Oxidized groups increase the wettability of polymers, and this supports the adsorption of cell adhesion-mediating extracellular matrix (ECM) molecules in an appropriate spatial conformation, increasing the accessibility of specific sites in these molecules for cell adhesion receptors. Free radicals and unsaturated bonds can then be used for functionalizing the material surface with various biomolecules, e.g., grafting amino acids, oligopeptides or protein molecules, which can also influence (mediate or attenuate) the cell adhesion and growth (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632]\]). This grafting occurs spontaneously after exposure of the irradiated material to biological environments, including biological fluids such as blood, intercellular liquid or cell culture media. Polyethylene has been used in several tissue engineering applications. For example, composites of HDPE with hydroxyapatite \[[@B17-materials-06-01632]\] or tricalcium phosphate \[[@B18-materials-06-01632]\] have been developed for the construction of bone replacements, composites of HDPE with graphite for the construction of joint replacements \[[@B19-materials-06-01632]\], and copolymers of HDPE with hyaluronan for the reconstruction of osteochondral defects \[[@B20-materials-06-01632]\]. HDPE was also clinically applied for calvarial reconstruction \[[@B21-materials-06-01632]\]. LDPE is used for producing intravascular catheters for arteriography and angioplasty, due to its advantageous mechanical properties, such as elasticity and flexibility \[[@B22-materials-06-01632]\]. In our study, two types of polyethylene, namely high density polyethylene (HDPE) and low density polyethylene (LDPE), were modified by Ar^+^ plasma discharge of various exposure times (50, 100 and 300 s), and were subsequently exposed to solutions of two important components of fetal serum, namely fibronectin (Fn) and bovine serum albumin (BSA). We expected the biomolecules from these fluids to graft spontaneously to the plasma-activated polyethylene and to influence cell adhesion and growth, which were studied using rat vascular smooth muscle cells in cultures on these materials. Polyethylene was chosen as a model material for studies of cell--material interaction, due to its easy availability and particularly its relatively simple molecule, consisting only of carbon and hydrogen, which enables clear and reproducible results to be obtained. A comparison was also made for the cell behavior on HDPE and LDPE. 2. Results and Discussion ========================= 2.1. Physical and Chemical Properties of the Polyethylene Samples ----------------------------------------------------------------- ### 2.1.1. Surface Wettability The water drop contact angle decreased after plasma modification of the polymers, which means that the surface wettability increased ([Figure 1](#materials-06-01632-f001){ref-type="fig"}). The highest decrease in the contact angle was observed on samples irradiated for 50 s, and as the exposure time was prolonged the decrease was less apparent. On HDPE irradiated for 300 s, the contact angle even increased ([Figure 1](#materials-06-01632-f001){ref-type="fig"}A). A similar observation on HDPE was published earlier, and can be explained by the different rearrangement of HDPE after plasma irradiation in comparison with LDPE \[[@B23-materials-06-01632]\]. It is generally known that the increase in wettability is due to the formation of new oxidized groups on the polymer surface after plasma modification \[[@B4-materials-06-01632],[@B24-materials-06-01632]\]. However, surface wettability is also associated with rearrangement of the surface structure, mainly of the oxidized groups, during the aging period of the polymer after plasma exposure. During this period, the oxidized groups are reoriented from the surface to the inside of the polymer. This causes an increase in the contact angle, though in most cases the contact angle of a modified polymer after the aging period still remains lower than the contact angle of the non-treated polymer. The rearrangement of the oxidized groups during aging is influenced by the degree of crosslinking of the polymer structure and the time of plasma modification. A less crosslinked structure and longer modification times lead to greater mobility of these groups. The structure of HDPE is less crosslinked than the structure of LDPE, probably due to the fact that HDPE molecules are more linear and less branched than LDPE molecules, and these differences are further enhanced by plasma irradiation. Thus, the less crosslinked structure of HDPE and relatively long plasma irradiation (300 s) led to higher rotation of the oxidized groups into the polymer, their burrowing into the polymer, their lower exposure on the polymer surface, and thus to lower wettability of the polymer, manifested by a relatively high contact angle (119°), which became even higher than on non-treated HDPE. Moreover, the lesser branching of HDPE creates less opportunity for oxidized groups to occur on this polymer, which might be another important factor influencing the contact angle. ![Water drop contact angle on (**A**) HDPE; and (**B**) LDPE in non-treated form (**N**), treated with plasma for 50, 100 or 300 s (**Plasma**), and subsequently grafted with fibronectin (**Fn**) or bovine serum albumin (**BSA**). Mean ± S.D. from 30 measurements performed on 3 samples (10 measurements on each).](materials-06-01632-g001){#materials-06-01632-f001} Subsequent grafting of Fn or BSA onto the plasma-activated polymer surface usually resulted in a further decrease in the contact angle. This phenomenon was well apparent, particularly on HDPE and LDPE irradiated with plasma for 100 and 300 s. The increase in surface wettability could be explained by the polar groups present in fibronectin and albumin molecules, oxygen-containing and amine groups. However, on the polymers irradiated for 50 s, the presence of BSA and Fn on an HDPE surface and BSA on an LDPE surface did not have significant effects on the surface wettability, as indicated by the contact angles on these polymers, which were comparable with the values on the samples treated only with plasma. Similar results for BSA on HDPE and LDPE have been published earlier \[[@B25-materials-06-01632],[@B26-materials-06-01632]\]. In these studies focused on grafting bioactive substances onto HDPE and LDPE surfaces, only glycin and polyethylene glycol further decreased the contact angle on the plasma-irradiated polymers, while the contact angle remained similar after grafting of BSA, colloidal carbon particles and a combination of BSA and these particles. Similarly, in the present study, grafting the plasma-treated PE with Fn was more efficient in decreasing the contact angle than grafting with BSA, except in the case of HDPE irradiated for 50 s. ### 2.1.2. XPS Analysis and Immunofluorescence Staining of Grafted Biomolecules Fibronectin and albumin are proteins which contain nitrogen and disulfide bonds \[[@B27-materials-06-01632],[@B28-materials-06-01632]\]. The presence of these biomolecules on the PE surface was therefore tested using X-ray photoelectron spectroscopy analysis (XPS) by determining the concentration of nitrogen and sulfur on the polymer surface. XPS also determined the concentration of oxygen as evidence that oxidized polar groups were formed after plasma irradiation. XPS analysis showed that the concentration of nitrogen was significantly higher on HDPE and LDPE activated with plasma and subsequently grafted with Fn and BSA than on samples only treated with plasma ([Table 1](#materials-06-01632-t001){ref-type="table"}). The nitrogen found in a low amount on the plasma-treated samples was probably bound from the ambient atmosphere after plasma exposure. In addition, sulfur was determined on the samples grafted with Fn and BSA. Due to the higher presence of nitrogen and sulfur on the grafted samples than on the only plasma-treated samples, we can conclude that fibronectin and albumin were successfully bound on the plasma-treated polymer surface. XPS also indicates that the concentration of oxygen increased after the polymer was exposed to plasma, and this led to a decline in the contact angle ([Figure 1](#materials-06-01632-f001){ref-type="fig"}). The higher concentration of oxygen on the irradiated samples than on the non-treated samples indicated the formation of newly oxidized polar groups after plasma modification. materials-06-01632-t001_Table 1 ###### Concentration (in at. %) of carbon (C), oxygen (O), nitrogen (N) and sulfur (S), determined by XPS analysis, on HDPE and LDPE treated with plasma for 50 s (**50**), and subsequently grafted with fibronectin (**Fn**) and bovine serum albumin (**BSA**). Mean from three measurements for each experimental group; the measurement error did not exceed 5%. --------------------------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------------------- **Material/Modification** **HDPE** **LDPE** C (at. %) O (at. %) N (at. %) S (at. %) C (at. %) O (at. %) N (at. %) S (at. %) 50 82.2 14.2 2.5 0 83.0 14.8 1.5 0 50 Fn 72.9 20.2 6.7 0.2 82.2 13.6 2.8 at limit of detection 50 BSA 74.4 15.3 9.9 0.4 72.7 15.3 10.9 0.4 --------------------------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------------------- The presence of fibronectin and albumin was also tested by immunofluorescence staining ([Figure 2](#materials-06-01632-f002){ref-type="fig"}). The resultant fluorescence intensity was presented as the difference between the fluorescence intensity measured on samples treated with both primary and secondary antibodies and on samples treated with the secondary antibody only (*i.e.*, the staining control). On HDPE and on LDPE grafted with Fn or BSA, the fluorescence intensity reached high positive values, indicating the presence of Fn and BSA on the polymer surface. By contrast, the fluorescence intensity had negative values on the non-treated polymers \[[Figure 2](#materials-06-01632-f002){ref-type="fig"}(A,B)\]. This was probably due to non-specific binding of the secondary antibody to the polymer surface, because the fluorescence intensity of the staining control, *i.e.*, the sample incubated only in the secondary antibody, was higher than the fluorescence intensity after staining the non-treated polymer with both primary and secondary antibodies ([Figure 2](#materials-06-01632-f002){ref-type="fig"}C). The non-specific binding of the secondary antibody on samples grafted with Fn and BSA was significantly reduced due to the presence of these biomolecules. In this context it should be pointed out that in studies using immunocytochemical methods, BSA \[[@B25-materials-06-01632],[@B26-materials-06-01632],[@B29-materials-06-01632]\], the whole blood serum \[[@B30-materials-06-01632]\] or solutions of other biomolecules, e.g., gelatin \[[@B12-materials-06-01632]\] are commonly used for blocking non-specific binding sites for antibodies. In addition, the fluorescence intensity on HDPE grafted with Fn or BSA was higher than on LDPE, which showed that Fn and BSA were grafted in a higher concentration on the HDPE surface than on LDPE \[[Figure 2](#materials-06-01632-f002){ref-type="fig"}(A,B)\]. ![Immunofluorescence staining of (**A**) fibronectin; and (**B**) albumin and their microphotographs (**C**) on HDPE (**H**) and LDPE (**L**), in non-treated form (**N**), treated with plasma for 50 s and subsequently grafted with fibronectin (**50 Fn**) or bovine serum albumin (**50 BSA**). Olympus IX 51 microscope, obj. 10x, DP 70 digital camera. The measurements were performed at the same exposure time for all experimental groups (1.2 s).](materials-06-01632-g002){#materials-06-01632-f002} ### 2.1.3. Surface Roughness and Morphology Changes in surface roughness and morphology after plasma exposure and grafting with biomolecules were studied by the Atomic Force Microscopy (AFM) method ([Figure 3](#materials-06-01632-f003){ref-type="fig"}, [Table 2](#materials-06-01632-t002){ref-type="table"}). Plasma treatment led to ablation of the surface layer \[[@B23-materials-06-01632]\], which increased the nanoscale surface roughness in proportion to the exposure time. These events were more apparent on HDPE than on LDPE, and could be explained by the different structure of the two polymers. For example, LDPE is more crosslinked than HDPE, and thus it could be more resistant to ablation. In our earlier study, plasma treatment of PE for 400 s resulted in the removal of a surface layer of HDPE that was about 1 μm in thickness, but of a layer of LDPE that was only 0.6 μm in thickness \[[@B23-materials-06-01632]\]. ![AFM images of HDPE (**H**) and LDPE (**L**) in non-treated form (**N**), treated in plasma for 50 s, and subsequently grafted with fibronectin (**Fn**) or bovine serum albumin (**BSA**).](materials-06-01632-g003){#materials-06-01632-f003} materials-06-01632-t002_Table 2 ###### The mean surface roughness values (R~a~) of LDPE and HDPE in non-treated form (**N**), treated in plasma for 50, 100 or 300 s (**50, 100, 300**), and subsequently grafted with fibronectin (**Fn**) or albumin (**BSA**). Sample HDPE LDPE ------------- ------ ------ **N** 6.1 5.0 **50** 7.4 4.7 **50** Fn 10.4 5.5 **50** BSA 6.9 5.4 **100** 8.4 6.7 **100** Fn 11.7 4.6 **100** BSA 9.4 4.7 **300** 11.5 6.6 **300** Fn 12.0 5.6 **300** BSA 7.8 4.6 In addition, the surface morphology of both types of PE was changed after plasma modification. A lamellar structure appeared on the polymer surface, perhaps due to the different ablation rate of the amorphous and crystalline phases. The amorphous phase is ablated preferentially \[[@B23-materials-06-01632],[@B31-materials-06-01632]\]. The newly created lamellar structure was more apparent on the HDPE surface, while the LDPE surface appeared to be ablated more uniformly (homogeneously). The reason could be a higher proportion of the crystalline phase in HDPE, while the amorphous phase prevails in LDPE \[[@B23-materials-06-01632]\]. Subsequent grafting with Fn or BSA further changed the surface roughness (see [Table 2](#materials-06-01632-t002){ref-type="table"}) and the morphology of PE. Grafting with Fn strongly increased the surface roughness of HDPE, while BSA did not significantly increase, or even reduced, the surface roughness of the plasma-treated surface, although this roughness still remained higher than on the non-treated polymer. The lamellar structure was less apparent on the grafted surface than on the only-treated HDPE surface. Grafting LDPE with Fn or BSA produced rather small and non-significant changes in surface roughness and morphology. On LDPE irradiated for 100 or 300 s, both Fn and BSA showed a tendency to smoothen the material surface ([Figure 3](#materials-06-01632-f003){ref-type="fig"}). 2.2. Cell Adhesion and Growth ----------------------------- ### 2.2.1. Cell Adhesion and Growth on LDPE and HDPE Samples in a Standard Serum-supplemented Medium The proliferation of cells in DMEM with fetal bovine serum (FBS) had in general a growing tendency. On day 2 after seeding, there were similar numbers of initially adhered cells on the control polystyrene, on the samples exposed to a plasma discharge, and on all Fn-grafted and BSA-grafted samples ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). This cell behavior was observed on both LDPE and HDPE, and could be explained by a masking effect of the serum in the culture medium, *i.e.*, secondary adsorption of serum components influencing cell adhesion, particularly vitronectin, fibronectin and albumin on the samples, which occurs within a few minutes after these substrates are exposed to the culture medium (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632],[@B9-materials-06-01632]\]). Significant differences were found only between the non-treated samples and some modified samples, usually with the exception of the BSA-grafted samples \[[Figure 4](#materials-06-01632-f004){ref-type="fig"}(A,B,D--F)\], and in two cases also with the exception of the samples only irradiated with plasma \[[Figure 4](#materials-06-01632-f004){ref-type="fig"}(B,D)\]. ![Growth curves of rat aortic smooth muscle cells cultured in a serum-supplemented medium on HDPE (**A--C**) or LDPE (**D--F**) in non-treated form (**N**), treated with plasma for 50, 100 or 300 s (**50, 100, 300**), and subsequently grafted with fibronectin (**Fn**) or bovine serum albumin (**BSA**). A standard cell culture polystyrene dish (**PS**) was used as a reference material. Mean ± S.E.M. from 20 measurements on microphotographs (day 2 and 4) or 200 measurements on a ViCell XR Analyser (Beckman Coulter, Brea, CA, USA) (day 6). ANOVA, Student--Newman--Keuls Method. Statistical significance: \* *p* ≤ 0.05 compared to non-modified HDPE or LDPE.](materials-06-01632-g004){#materials-06-01632-f004} From day 2 to day 4 after seeding, the cells on all modified PE samples showed an intense increase in their growth, while the cell growth on the non-treated PE samples was stagnant. As a result, the cell numbers on all modified samples on day 4 were significantly higher than the values on the non-treated samples. However, the differences among the groups of modified samples still remained non-significant ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). From day 4 to day 6 after seeding, the cells on most of the HDPE samples reached confluence, slowed down their proliferation and entered the stationary growth phase, *i.e.*, their number did not increase as rapidly as between day 2 and day 4, and on some samples the number even decreased. By contrast, the cells on the LDPE samples usually continued their growth, and thus reached on an average higher final cell population densities. On day 6 after seeding, these densities ranged from about 78,000 to 111,000 cells·cm^−2^ on LDPE, while there were only 43,000 to 79,000 cells·cm^−2^ on HDPE ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). Particularly the cells on the BSA-grafted LDPE samples continued their exponential growth, and on day 6 after seeding their population densities were among the highest values reached on these samples, especially on LDPE activated with plasma for 300 s ([Figure 4](#materials-06-01632-f004){ref-type="fig"}F). These results can be considered as surprising, because albumin is known as a cell non-adhesive molecule \[[@B32-materials-06-01632]\], which has been used for creating bioinert cell-free surfaces, applicable e.g., in blood-contacting devices \[[@B33-materials-06-01632],[@B34-materials-06-01632]\]. In our study, the non-adhesive properties of albumin are slightly reflected by the relatively low numbers of initially adhered cells on the BSA-grafted samples, which often did not differ significantly from the values on the non-treated polyethylene ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). Nevertheless, the cell spreading areas were mostly larger than in the cells on non-treated samples, and usually not significantly smaller than on the control polystyrene dishes, except BSA-grafted HDPE treated with plasma for 100 s ([Figure 5](#materials-06-01632-f005){ref-type="fig"}, [Figure 6](#materials-06-01632-f006){ref-type="fig"} and [Figure 7](#materials-06-01632-f007){ref-type="fig"}). Albumin has been reported to promote adsorption of ECM molecules, e.g., vitronectin and fibronectin, in advantageous spatial conformations, supporting the accessibility of these molecules by cell adhesion receptors (for a review, see \[[@B25-materials-06-01632]\]). These molecules cannot only be adsorbed over the albumin from the serum of the culture medium, but are also synthesized and deposited by VSMC themselves \[[@B35-materials-06-01632]\], especially at later culture intervals. This may account for the relatively high proliferation activity of the cells on BSA-grafted PE ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). As for the non-treated HDPE and LDPE samples, the cell number increased from day 4 to 6 on these materials ([Figure 4](#materials-06-01632-f004){ref-type="fig"}). However, the maximum population densities on these polymers still remained significantly lower (from about 24,000 to 29,000 cells·cm^−2^) than on the modified samples (from 43,000 to 111,000 cells·cm^−2^). In accordance with the weak cell proliferation on non-treated PE, the visualization of the cells on day 2 also showed a lower size of the cell spreading area on non-treated PE than on the modified samples ([Figure 5](#materials-06-01632-f005){ref-type="fig"}, [Figure 6](#materials-06-01632-f006){ref-type="fig"} and [Figure 7](#materials-06-01632-f007){ref-type="fig"}). This insufficient cell spreading can be explained by the relatively high hydrophobicity of non-treated HDPE and LDPE. It is known that, on highly hydrophobic surfaces, the cell adhesion-mediating proteins, e.g., fibronectin and vitronectin, present in the serum of the culture medium, are adsorbed in rigid, non-physiological and denatured form, less appropriate for binding the specific sites in the protein molecules by the cell adhesion receptors (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632],[@B9-materials-06-01632]\]). By contrast, the modified polymers were more hydrophilic, with an enhanced nanostructure or grafted with biomolecules. As mentioned above, BSA supports the adsorption of cell adhesion-mediating molecules in a near-physiological conformation, appropriate for binding to the cell adhesion receptors. In addition, the fibronectin grafted on some samples provided the possibility for direct binding of the cells, e.g., through specific RGD-containing amino acid sequences in the Fn molecules \[[@B28-materials-06-01632]\]. ![Morphology of rat aortic smooth muscle cells in two-day-old cultures in a serum-supplemented medium on HDPE in non-treated form (**N**), treated with plasma for 50, 100 or 300 s (**50, 100, 300**), and subsequently grafted with fibronectin (**Fn**) or bovine serum albumin (**BSA**). A standard cell culture polystyrene dish (**PS**) was used as a reference material. Cells stained with Texas Red C~2~-Maleimide and Hoechst \#33342. Olympus IX 51 microscope, obj. 20, DP 70 digital camera, bar 200 μm or 100 μm (Fn-grafted samples).](materials-06-01632-g005){#materials-06-01632-f005} ![Morphology of rat aortic smooth muscle cells in two-day-old cultures in a serum-supplemented medium on LDPE in non-treated form (**N**), treated with plasma for 50, 100 or 300 s (**50, 100, 300**), and subsequently grafted with fibronectin (**Fn**) or bovine serum albumin (**BSA**). A standard cell culture polystyrene dish (**PS**) was used as a reference material. Cells stained with Texas Red C~2~-Maleimide and Hoechst \#33342. Olympus IX 51 microscope, obj. 20, DP 70 digital camera, bar 200 μm or 100 μm (Fn-grafted samples).](materials-06-01632-g006){#materials-06-01632-f006} ![The size of the cell spreading area of rat aortic smooth muscle cells in two-day-old cultures in a serum-supplemented medium on (**A**) HDPE; or (**B**) LDPE in non-treated form (**N**), treated with plasma for 50, 100 or 300 s, and subsequently grafted with fibronectin (**Fn**) or bovine serum albumin (**BSA**). A standard cell culture polystyrene dish (**PS**) was used as a reference material. Mean ± S.E.M. from 32 to 95 measured cells for each experimental group. ANOVA, Student--Newman--Keuls Method. Statistical significance: ▪ *p* ≤ 0.05 compared to other experimental groups, indicated by the colors of these groups above the columns.](materials-06-01632-g007){#materials-06-01632-f007} ### 2.2.2. Cell Adhesion and Growth on LDPE and HDPE in a Serum-Free Medium In the serum-free medium, the differences in the behavior of VSMC on the PE samples with various modifications became more apparent than in the serum-supplemented medium, and the growth dynamics of the cells revealed a different trend. From days 2 to 4 after seeding, the cell number was mostly stagnant, and on some samples it even decreased. There was an increasing tendency only between days 4 and 6 ([Figure 8](#materials-06-01632-f008){ref-type="fig"}). This stagnancy in cell growth was most likely because there was a limited quantity of the ECM proteins normally present in fetal serum, mainly vitronectin and fibronectin, which ensure proper adhesion, spreading and subsequent growth of cells on the material surfaces. The serum-free medium used in this study did not contain these proteins. However, cell adhesion-mediating proteins can be present on cells in small quantities even after trypsinization \[[@B36-materials-06-01632]\], and could be used for cell adhesion after seeding cells on the tested materials, particularly on polystyrene, non-treated PE and plasma-treated PE (on Fn-grafted PE, the Fn molecules could be used directly for cell adhesion). In addition, VSMC are able to synthesize a wide spectrum of cell adhesion-mediating ECM molecules (for a review, see \[[@B35-materials-06-01632]\]). This takes place especially at later culture intervals. By way of these molecules, the cells adhere to the material surface through the bonds between specific amino acid sequences in these molecules and cell adhesion receptors on the cell membrane, e.g., integrins. Another way for cells to attach to the tested materials is by so-called weak chemical bonding, *i.e.*, hydrogen bonding, electrostatic, polar or ionic interactions between various molecules on the cell membrane and functional chemical groups on the polymers, *etc*., but this type of binding does not transfer specific signals into the cells and does not ensure their survival, growth and other functions (for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632],[@B9-materials-06-01632],[@B37-materials-06-01632],[@B38-materials-06-01632]\]). ###### Growth curves of rat aortic smooth muscle cells cultured in a serum-free medium on (**A--C**) HDPE; or (**D--F**) LDPE in non-treated form (**N**), treated with plasma for 50, 100 or 300 s (**50, 100, 300**), and subsequently grafted with fibronectin (**Fn**) and bovine serum albumin (**BSA**). A standard cell culture polystyrene dish (**PS**) was used as a reference material. Mean ± S.E.M. from 20 measurements (day 2 and 4) or 4 measurements (day 6). ANOVA, Student--Newman--Keuls Method. Statistical significance: \* *p* ≤ 0.05 compared to the non-modified LDPE or HDPE. ![](materials-06-01632-g008a) ![](materials-06-01632-g008b) As early as on day 2 after seeding, the cell non-adhesive properties of BSA were completely revealed in the serum-free medium. This effect of albumin was more apparent on HDPE, where the cell densities ranged from 37 to 367 cells·cm^−2^, while on LDPE the values were slightly higher (551 to 1470 cells·cm^−2^; [Figure 8](#materials-06-01632-f008){ref-type="fig"}). Accordingly, the cell spreading was also the lowest on the HDPE samples ([Figures 9](#materials-06-01632-f009){ref-type="fig"}). The cells mainly assumed a spherical shape on these surfaces, while on the other samples, even on the BSA-grafted LDPE samples ([Figure 10](#materials-06-01632-f010){ref-type="fig"}) or on non-treated LDPE and HDPE, the cell morphology was spindle-shaped or polygonal. From day 4 to 6, the cell numbers increased slightly, probably by preferential cell growth on sites with discontinuity of the albumin layer, or secondary deposition of cell adhesion-mediating ECM proteins synthesized by the cells themselves on the albumin layer. On day 6, this increase was slightly more apparent on BSA-grafted LDPE (from 2800 to 4900 cells·cm^−2^) than on BSA-grafted HDPE (from 400 to 1500 cells·cm^−2^). However, these samples still displayed the lowest cell population densities, irrespective of plasma exposure time ([Figure 8](#materials-06-01632-f008){ref-type="fig"}). The effects of the grafted fibronectin were controversial. We expected high adhesion and subsequent growth of cells on these samples, due to specific receptor-mediated adhesion of cells to Fn grafted on the plasma-activated polymer surfaces. However, the cell densities in general on all Fn-grafted samples were similar to the values on the plasma-irradiated surfaces, or on the control polystyrene culture dishes. Among all Fn-grafted samples, the weakest proliferation was observed on LDPE irradiated with plasma for 50 s (L 50Fn; [Figure 8](#materials-06-01632-f008){ref-type="fig"}B), where the cell population density on day 4 was lower than on non-treated LDPE, and as low as on the BSA-grafted samples. On day 6, the number of cells increased, but was still significantly lower than on PS or LDPE irradiated for 50 s ([Figure 8](#materials-06-01632-f008){ref-type="fig"}B). The relatively low cell proliferation activity on Fn-grafted polyethylene could be explained by relatively low quantity of the Fn molecules attached to the polymer surface. As indicated by XPS, the concentrations of nitrogen and sulfur, which are indicators of the presence of protein molecules, were relatively low on the Fn-grafted surfaces in comparison with the BSA-grafted surfaces, particularly on the LDPE samples. On LDPE, the concentration of S was even on the limit of detection ([Table 1](#materials-06-01632-t001){ref-type="table"}). In accordance with this, the intensity of fluorescence of antibody-labeled Fn grafted to plasma-activated surfaces is lower than in the case of albumin \[[Figure 2](#materials-06-01632-f002){ref-type="fig"}(A,B)\] In addition, the pictures indicate less homogeneous distribution of fibronectin on these surfaces compared to albumin ([Figure 2](#materials-06-01632-f002){ref-type="fig"}C). Also the geometrical conformation of Fn, which is more important for cell binding than the absolute amount of Fn immobilized on the material surface (\[[@B39-materials-06-01632]\], for a review, see \[[@B3-materials-06-01632],[@B4-materials-06-01632],[@B9-materials-06-01632]\]) could be less appropriate for binding by the cell adhesion receptors. In spite of this, the cells on the Fn-grafted HDPE and LDPE samples were of similar polygonal morphology as the cells on the control polystyrene and plasma-irradiated samples ([Figure 9](#materials-06-01632-f009){ref-type="fig"} and [Figure 10](#materials-06-01632-f010){ref-type="fig"}), although their spreading area was often smaller ([Figure 11](#materials-06-01632-f011){ref-type="fig"}). ![Morphology of rat aortic smooth muscle cells in two-day-old cultures in a serum-free medium on HDPE in non-treated form (**N**), treated with plasma for 50, 100 or 300 s (**50, 100, 300**), and subsequently grafted with fibronectin (**Fn**) and bovine serum albumin (**BSA**). A standard cell culture polystyrene dish (**PS**) served as a reference material. Cells stained with Texas Red C~2~-Maleimide and Hoechst \#33342. Olympus IX 51 microscope, obj. 20, DP 70 digital camera, bar 200 μm or 100 μm (Fn-grafted samples).](materials-06-01632-g009){#materials-06-01632-f009} ![Morphology of rat aortic smooth muscle cells in two-day-old cultures in a serum-free medium on LDPE in non-treated form (**N**), treated with plasma for 50, 100 or 300 s (**50, 100, 300**), and subsequently grafted with fibronectin (**Fn**) and bovine serum albumin (**BSA**). A standard cell culture polystyrene dish (**PS**) was used as a reference material. Cells stained with Texas Red C~2~-Maleimide and Hoechst \#33342. Olympus IX 51 microscope, obj. 20, DP 70 digital camera, bar 200 μm or 100 μm (Fn-grafted samples).](materials-06-01632-g010){#materials-06-01632-f010} ![The size of the cell spreading area of rat aortic smooth muscle cells in two-day-old cultures in a serum-free medium on (**A**) HDPE; or (**B**) LDPE in non-treated form (**N**), treated with plasma for 50, 100 or 300 s, and subsequently grafted with fibronectin (**Fn**) or bovine serum albumin (**BSA**). A standard cell culture polystyrene dish (**PS**) was used as a reference material. Mean ± S.E.M. from 4 to 269 measured cells. ANOVA, Student--Newman--Keuls Method. Statistical significance: ▪ *p* ≤ 0.05 compared to other experimental groups, indicated by the colors of these groups above the columns.](materials-06-01632-g011){#materials-06-01632-f011} 3. Experimental Section ======================= 3.1. Preparation and Modification of Polymer Samples ---------------------------------------------------- The experiments were carried out on high-density and low-density polyethylene foils (HDPE and LDPE, both purchased from Granitol A. S., Moravsky Beroun, Czech Republic). The HDPE foils were of the Microten M\*S type (thickness 40 μm, density 0.951 g·cm^−3^, melt flow index 0.14 g·10 min^−1^), and the LDPE was of the Granoten S\*H type (thickness 40 μm, specific density 0.922 g·cm^−3^, melt flow index 0.8 g·10 min^−1^). Both types of polyethylene were cut into circular samples (diameter 2 cm) using a metallic perforator (punch). The samples were then treated by an Ar^+^ plasma discharge (gas purity 99.997%) using a Balzers SCD 050 device (BalTec Maschinenbau AG, Pfäffikon, Switzerland). The time of exposure was 50, 100 or 300 s, and the discharge power was 3 W. Immediately after the plasma treatment, the samples were immersed in solutions of two components of fetal bovine serum (FBS), namely 2 wt.% of fibronectin (Fn, Sigma-Aldrich, St. Louis, MI, USA, Cat. No. F2006) or 2 wt.% of bovine serum albumin (BSA, Sigma-Aldrich, Cat. No. A9418), both diluted in phosphate-buffered saline (PBS; Sigma-Aldrich, Cat. N° P4417). The samples were incubated for 24 h at room temperature (RT) in order to allow grafting of biomolecules to the plasma activated polymer surface. The samples were then rinsed in distilled water, air-dried at RT and stored in an air atmosphere for three weeks in order to reorganize and stabilize their surface structure, particularly the orientation of the oxidized structures on the material surface (the so-called aging period) \[[@B23-materials-06-01632]\]. 3.2. Characterization of the Physical and Chemical Properties of the Polymer Surface ------------------------------------------------------------------------------------ ### 3.2.1. Surface Wettability The sessile water drop contact angle was measured by See System reflection goniometry (Masaryk University, Brno, Czech Republic). The volume of the water drop on the polymer surface was 8 μL. The contact angle was measured after an aging period of three weeks. For each experimental group, three samples were used, and on each sample, 10 measurements were performed in different regions homogeneously distributed on the sample surface. The contact angle was presented as mean ± Standard Deviation (S.D.) from 30 measurements. Statistical analyses were performed using SigmaStat (Jandel Corp., Las Vegas, NV, USA). Multiple comparison procedures were performed by the One Way Analysis of Variance (ANOVA), Student--Newman--Keuls method. *p* values equal to or less than 0.05 were considered significant. ### 3.2.2. Chemical Composition of the Polymer Surface The presence of grafted biomolecules, *i.e.*, fibronectin (Fn) and bovine serum albumin (BSA), on the HDPE and LDPE surface was determined by X-ray photoelectron spectroscopy (XPS, Omicron Nanotechnology ESCAProbeP spectrometer) and by immunofluorescence staining. XPS determined the concentration (at.%) of the main elements present in the grafted polymer surface (oxygen, nitrogen, sulfur and carbon). For immunofluorescence staining, specific primary antibodies against Fn (monoclonal anti-human fibronectin, Sigma-Aldrich, Cat. No. F0791) and BSA (monoclonal anti-bovine serum albumin, Sigma--Aldrich, Cat. No. B2901, St. Louis, MI, USA) were used. The primary antibodies were diluted in PBS to a concentration of 1:200 and were applied to HDPE and LDPE samples modified in plasma and subsequently grafted with Fn or BSA, and were also applied to the non-modified control samples. The samples were incubated with primary antibodies overnight at 4 °C. After rinsing with PBS, the secondary antibody, *i.e.*, goat anti-mouse F(ab')2 fragments of IgG (H + L), conjugated with Alexa Fluor^®^ 488 (Molecular Probes, Invitrogen, Cat. No. A11017, Eugene, OR, USA), was diluted in PBS to a ratio of 1:400, and was added to the samples for 1 hour at RT. As a staining control, the secondary antibody was applied to samples which had not been treated with the primary antibodies before. The samples were then rinsed twice in PBS. The fluorescence intensity was evaluated from 10 randomly chosen fields homogeneously distributed on the material surface using an epifluorescence microscope (IX 51, Olympus, Tokyo, Japan, obj. 10×) equipped with a digital camera (DP 70, Olympus, Tokyo, Japan) at the same exposure time for all experimental groups (1.2 s). The data was presented as mean ± Standard Error of Mean (S.E.M.). Statistical analyses were performed using SigmaStat (Jandel Corp., Las Vegas, NV, USA). Multiple comparison procedures were made by the One Way Analysis of Variance (ANOVA), Student--Newman--Keuls method. *p* values equal to or less than 0.05 were considered significant. ### 3.2.3. Surface Morphology and Roughness The changes in surface morphology and roughness were determined by Atomic Force Microscopy (AFM), using a CP II device (VEECO, Santa Barbara, CA, USA) working in tapping mode. We used an RTESPA-CP Si probe, with spring constant 20--80 N/m. By repeated measurements of the same region (1 × 1 µm), it was proven that the surface morphology did not change after three consecutive scans. The mean roughness value (R~a~) represents the arithmetic average of the deviations from the center plane of the sample. 3.3. Cells and Culture Conditions --------------------------------- The samples were sterilized with 70% ethanol for 1 hour, placed into 12-well polystyrene multidishes (TPP, Switzerland, well diameter 2.2 cm) and then air-dried for 12 hours in a sterile environment. The materials were fixed to the bottom of the culture wells by plastic rings in order to prevent them floating in the cell culture media, and were seeded with vascular smooth muscle cells (VSMC), derived from the rat aorta by an explantation method \[[@B40-materials-06-01632]\]. The cells were used in passage 3. Each well contained 50,000 cells (*i.e.*, about 14,000 cells·cm^−2^) and 3 mL of serum-supplemented or serum-free medium. The serum-supplemented medium was Dulbecco's modified Eagle's Medium (DMEM; Sigma, Ronkonkoma, NY, USA, Cat. N° D5648) with 10% of fetal bovine serum (FBS; Sebak GmbH, Aidenbach, Germany) and 40 µg/mL of gentamicin (LEK, Ljubljana, Slovenia), and the serum-free medium was SmGM^®^-2 Smooth Muscle Growth Medium-2 (SmBM; Lonza, Walkersville, MD, USA, Cat. N° CC-3182), supplemented with epithelial growth factor (EGF), fibroblast growth factor-B (FGF-B) and insulin according to the manufacturer's protocol. The serum-free medium was used in order to enhance the effects on cell adhesion and growth of the chemical functional groups and the biomolecules grafted onto the material surface. This influence can be masked, at least partly, by secondary adsorption of various biomolecules on the polymer surface from the serum supplement of the culture medium. The cells were then cultured for 2, 4 and 6 days in a cell incubator at 37 °C and in a humidified atmosphere with 5% of CO~2~ in the air. 3.4. Evaluation of the Cell Number and Cell Spreading Area ---------------------------------------------------------- On days 2 and 4 after seeding, the cells were rinsed with PBS, fixed with cold 70% ethanol (−20 °C, 5 min), and then stained for 2 hours at RT with a combination of the following fluorescent dyes diluted in PBS: Hoechst \#33342 nuclear dye (Sigma-Aldrich, 5 µg·mL^−1^) and Texas Red C~2~-maleimide membrane dye (Molecular Probes, Invitrogen, Carlsbad, CA, USA, Cat. No. T6008, 20 ng·mL^−1^). The cells were counted on microphotographs taken in 10 randomly chosen fields homogeneously distributed on the material surface, using an epifluorescence microscope (Olympus IX 51, Japan, obj. 20×) equipped with a digital camera (DP 70, Japan). This approach was not applicable on day 6, when the cells on some samples reached confluence and started to overlap each another. The cells were therefore detached from the materials by a trypsin-EDTA solution (Sigma-Aldrich, Cat. N° T4174) and were counted using a Vi-CELL XR Analyser (Beckman Coulter, Brea, CA, USA). The pictures taken on day 2 after seeding were also used for measuring the size of the cell spreading area using Atlas software (Tescan Ltd., Brno, Czech Republic). Two independent samples were used for each time interval and experimental group. Non-modified polyethylene (HDPE or LDPE) and standard cell culture polystyrene wells were used as reference materials. The quantitative data was presented as mean ± S.E.M. Statistical analyses were performed using SigmaStat (Jandel Corp., USA). Multiple comparison procedures were made by the One Way Analysis of Variance (ANOVA), Student--Newman--Keuls method. *p* values equal to or less than 0.05 were considered significant. The cell numbers obtained on day 2, 4 and 6 after seeding were expressed as the number of cells per cm^2^ (*i.e.*, the cell population density), and were used for constructing growth curves. 4. Conclusions ============== Modification of polyethylene (HDPE and LDPE) with Ar^+^ plasma and grafting with fibronectin and albumin influenced the cell colonization of both polymers. In the serum-supplemented medium, all modifications improved the adhesion and growth of vascular smooth muscle cells in comparison with the unmodified polymer. The cells reached higher final population densities, on an average, on LDPE than on HDPE. In the serum-free medium, grafting with albumin did not support cell colonization due to its non-adhesive properties, which were not masked by the adsorption of cell adhesion-mediating molecules, normally present in the serum. The beneficial effect of the plasma treatment on cell adhesion and growth could be attributed to the formation of new-oxidized structures on the polymer surface, an increase in surface wettability, changes in surface morphology, and enhanced nanoscale roughness. Changes in the physical and chemical properties of the material surface were more apparent on HDPE than on LDPE. The intensity of the fluorescence indicated that HDPE also contained higher quantities of grafted fibronectin and albumin. This study was supported by the Grant Agency of the Czech Republic (grant No. P108/12/G108 "Center of Excellence"). Robin Healey (Czech Technical University, Prague) is gratefully acknowledged for his language revision of the manuscript. [^1]: These authors contributed equally to this work.

{ "pile_set_name": "PubMed Central" }

Background {#Sec1} ========== Hepatocellular carcinoma (HCC) is one of the most common malignant cancer worldwide and the second leading cause of cancer death for male in China, especially in China's rural areas, and it is also considered as the sixth leading cause of cancer-related mortality for male in developed countries. According to reports, the incidence and mortality of liver cancer in China occupies for nearly 50% of the global liver cancer population \[[@CR1], [@CR2]\]. Unfortunately, the pathogenesis of HCC is concealed, which is difficult to find in early stage. The rapid development and high rate of postsurgical relapse is a major challenge of HCC, owing to the rapid proliferation and differentiation of hepatoma cells \[[@CR3]\]. Thus, new antitumor agents and potential mechanisms are urgently needed to inhibit the proliferation of hepatoma cells. Sorafenib is considered to be the only effective treatment targeted drug for patients with advanced HCC, but it shows poor treatment effect for patients with extrahepatic spread or vascular invasion \[[@CR4]\]. Therefore, screening effective anti-hepatoma drugs from traditional Chinese medicine herbs is a necessary research field. In recent decades, some effective ingredients of Chinese medicinal herbs, such as icariin, psoralen, osthole, matrine, ginsenoside Rg1/Rh2 and celastrol, were reported to inhibit the proliferation of hepatocellular carcinoma cells \[[@CR5]--[@CR11]\]. *Psoralea* species possess unique bioactive compounds with anti-cancer properties \[[@CR12]\], and there are many tumor-suppression patents of traditional Chinese medicine which have psoralea as basic remedy \[[@CR13]--[@CR15]\]. Additionally, psoralen is the main active component of *Psoralea corylifolia* (L.) (Fig. [1](#Fig1){ref-type="fig"}a) and is used as marker to assess its quality \[[@CR16], [@CR17]\], which can inhibit the proliferation of adrenocortical tumors Y1 cells and pituitary tumor AtT20 cells \[[@CR18]\]. Psoralen is a kind of furocoumarin, the molecular formula of psoralen is C~11~H~6~O~3~, with molecular weight of 186.1 and dissolved in DMSO. The chemical structure of psoralen was shown in Fig. [1](#Fig1){ref-type="fig"}b. Furthermore, psoralen plays roles in anti-tumor, anti-oxidation and reversing the multidrug resistance, and it is concerned with cell cycle, apoptosis, calcium antagonism and estrogen-like effects \[[@CR19]--[@CR22]\]. However, the effect of psoralen on HCC is still unclear.Fig. 1Chemical structure of psoralen. **a** The plant diagram of psoralea. **b** The chemical structure of psoralen Endoplasmic reticulum stress (ER-stress) is the basic cells response against various extracellular factors, including UV, anoxia, oxidative stress, toxic substances, nutrient deficiency and drug agonists, which will result in the unfolded proteins accumulation in the endoplasmic reticulum and lead to Ca^2+^ homeostasis imbalance. According to previous studies, three major ER-spanning transmembrane proteins, PERK (protein kinase R-like ER kinase), ATF6 (activating transcription factor 6) and IRE1 (inositol-requiring enzyme 1) subsequently drive mutually reinforcing signaling pathways to correct the protein-misfolding stress \[[@CR23]\]. In this process, glucose-regulated protein of 78 (GRP78), also known as heavy chain-binding protein (Bip), plays an important role in detecting the accumulation of unfolded proteins in ER lumen and releasing the three sensors, so GRP78 is considered to be an ER homeostasis receptor \[[@CR24]\]. In the early stage of ER stress, the unfolded protein response (UPR) acts to help cells to cope with the stress by attenuating protein synthesis, clearing the unfolded/misfolded proteins and increasing the capacity of the ER to fold proteins, which restore the intracellular homeostasis and protect cell functions. However, strong and sustained ER stress induced by the activation of UPR will cause imbalance of ER homeostasis. Prolonged activation of the UPR can induces apoptosis pathway following ER stress, mainly involving in the CHOP/DITT3/GADD153 gene activation pathway, JNK activation pathway, ER-specific cysteine Caspase-12 activation pathway. In this study, we observed the proliferation inhibition effects of psoralen on hepatoma SMMC7721 cells and further investigated its relationship with ER stress. The results suggest that psoralen can induce cell cycle arrest and apoptosis by ER stress to inhibit malignant proliferation of hepatoma cells. Results {#Sec2} ======= Psoralen inhibits the proliferation of SMMC7721 cells {#Sec3} ----------------------------------------------------- We first examined whether psoralen was cytotoxic to hepatoma cell lines. Compared with the control group, the SMMC7721 cells treated with different dose psoralens for 24 h, 48 h and 72 h. We found that 20--80 μM psoralen inhibited cell proliferation and the inhibition rate was 17.50--25.10% in 24 h (Table [1](#Tab1){ref-type="table"}). More than 20 μM psoralen could effectively inhibit cell proliferation in 48 h (P \< 0.001) and 72 h (P \< 0.001) (Table [1](#Tab1){ref-type="table"}). In particular, more than 80 μM psoralen caused cell death (Fig. [2](#Fig2){ref-type="fig"}a, b). Next, we explored the effect of psoralen on L02 hepatocyte cell line and HepG2. We found that 10 μM, 20 μM and 40 μM PSO had little effects on the proliferation of L02, but 80 μM PSO could inhibit L02 proliferation (Fig. [2](#Fig2){ref-type="fig"}c). However, psoralen treatment showed little effect on proliferation of HepG2 cells (Fig. [2](#Fig2){ref-type="fig"}d). It is suggested that psoralen has a significant inhibitory effect on the proliferation of SMMC7721.Table 1Inhibitory effect of psoralen on proliferation of SMMC7721 hepatoma cells ($\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\overline{x}$$\end{document}$ ± s, n = 3)Dose/μMInhibition rate/%24 h48 h72 h0 (Con)0.00 ± 2.700.00 ± 8.900.00 ± 5.70108.40 ± 8.0019.60 ± 6.90\*22.80 ± 10.9\*2017.50 ± 1.9034.80 ± 6.90\*\*44.60 ± 7.80\*\*4022.90 ± 2.9054.80 ± 0.60\*\*\*74.90 ± 2.80\*\*\*8025.10 ± 6.4067.30 ± 3.50\*\*\*96.60 ± 0.50\*\*\*Compared with the control group, \*P \< 0.05, \*\*P \< 0.01, \*\*\*P \< 0.001. The data are from three independent experiments, four wells for each experiment Fig. 2Effects of psoralen on the proliferation of SMMC7721, L02 and HepG2. **a** Different doses of psoralen inhibit the proliferation of SMMC7721 for 24--72 h. The A570 value of SMMC7721 (**b**), L02 (**c**) and HepG2 (**d**) with the treatment of psoralen in 10--80 μM for 72 h. Values are mean ± SD (n = 4) and \* is means compared to the Con group, \*P \< 0.05; \*\*P \< 0.01; \*\*\*P \< 0.001 Psoralen causes shrinkage and endoplasmic reticulum dilation of SMMC7721 {#Sec4} ------------------------------------------------------------------------ The SMMC7721 cells were treated with 40 μM psoralen or 0.1 μM thapsigargin for 12 h, 24 h, and 48 h, respectively. The results showed that SMMC7721 cells appeared mild shrinkage and the cells gap increased. With the treatment prolonged, the SMMC7721 cells shrunk significantly, furthermore the dead and floated cells could be observed. The thapsigargin caused the cells shrinkage as a positive control, which showed significant inhibitory effect than that of psoralen (Fig. [3](#Fig3){ref-type="fig"}a). These results suggested that psoralen can induce the shrinkage and reduce adhesion of SMMC7721 cells. Next, we observed the intracellular structure of SMMC7721, which showed a large nucleus, few cytoplasm, scattered smooth endoplasmic reticulum and some mitochondria in transmission electron microscope. However, endoplasmic reticulum dilation appeared in SMMC7721 by 40 μM psoralen treatment for 24 h, and the cytoplasm presented loose and derangement distribution. Moreover, 0.1 μM thapsigargin can cause the endoplasmic reticulum dilation and cytoplasmic structure disorder (Fig. [3](#Fig3){ref-type="fig"}b). It is suggested that psoralen can cause abnormalities of endoplasmic reticulum and cytoplasmic structure in SMMC7721 hepatoma cells.Fig. 3Effects of psoralen on the morphology of SMMC7721. **a** SMMC7721 cells shrunk in the dose of 40 μM psoralen or 0.1 μM thapsigargin for 12--48 h. **b** Abnormalities of endoplasmic reticulum and cytoplasmic structure in SMMC7721 in the dose of 40 μM psoralen or 0.1 μM thapsigargin for 24 h. Yellow boxes represent the abnormal region Psoralen induces cell cycle arrest at G1 phase in SMMC7721 {#Sec5} ---------------------------------------------------------- We next evaluated whether psoralen modulated cell cycle progression to affect cell viability. PI staining analysis by flow cytometry demonstrated that 40 μM psoralen led to a significant increase in the percentage of cells at the G1 phase and a decrease in cells at the S phase. And 0.1 μM thapsigargin also caused a significant increase in the percentage of cells at the G1 phase (Fig. [4](#Fig4){ref-type="fig"}a). We also determined the effects of psoralen on the expression of G1/S transition-related cell cycle proteins. Compared with the control group, CyclinD1 protein expression was increased significantly by 40 μM psoralen, and CyclinE1 expression was suppressed significantly. However, psoralen showed little influence on the expression of CDK4 and p27Kip1 (CDKN1B). Additionally, the proteins expression of CyclinD1, CyclinE1, CDK4 and p27Kip1 were decreased significantly by 0.1 μM thapsigargin (Fig. [4](#Fig4){ref-type="fig"}b). These results suggested that psoralen could induces cell cycle arrest at the G1 phase by reducing the expression of CyclinE1 in SMMC7721, thus inhibiting the growth of hepatoma cells.Fig. 4Psoralen induces cell cycle arrest at G1 phase in SMMC7721. Flow cytometry (**a**) and Western blot (**b**) assays showing the effects of psoralen and thapsigargin on cell cycle progression and expression of G1/S transition-related proteins, respectively. Relative protein levels are showed on the right. *ns* no significant; \*\*\*P \< 0.001 Psoralen induces apoptosis in SMMC7721 hepatoma cells {#Sec6} ----------------------------------------------------- We next determined the effects of psoralen on cellular apoptosis in SMMC7721 cells. Annexin-V staining revealed that psoralen treatment increased late apoptosis in SMMC7721 hepatoma cells and showed a significant dose-effect relationship (Fig. [5](#Fig5){ref-type="fig"}a). Western blot analysis also demonstrated that treated with 10 μM psoralens for 48 h could increase the level of Bax, and 40 μM psoralen could attenuate the expression of Bcl-2 (Fig. [5](#Fig5){ref-type="fig"}b). These results suggested that psoralen induces cell apoptosis in SMMC7721, which may be related to the inhibition of Bcl-2 protein expression.Fig. 5Psoralen induces apoptosis in SMMC7721 hepatoma cells. AnnexinV-PI staining (**a**) and western blot (**b**) assays showing the effects of different doses of psoralen on cell apoptosis and the expression of apoptosis-related proteins The effects of psoralen on ER-stress related gene expression in SMMC7721 {#Sec7} ------------------------------------------------------------------------ To investigate the effects of psoralen on the key gene expression of ER-stress, we treated SMMC7721 with different doses and time duration. Compared with the control group, GRP78 and SGK1 mRNA expression were increased significantly in the treatment with 20 μM and 40 μM psoralen for 48 h (P \< 0.05), DDIT3 and GDF15 mRNA levels were dose-dependent increased (P \< 0.01), and relative mRNA expression of GADD34 and FOXO3 were also increased by 40 μM psoralen (P \< 0.05). However, different doses of psoralen showed little influence on EIF2A mRNA expression (Fig. [6](#Fig6){ref-type="fig"}a). Next, we detected the mRNA levels with the treatment of 40 μM psoralen for 6 h, 12 h and 24 h. Compared with the control group, DDIT3, XBP1, and ATF4 mRNA expression were time-dependent increased (P \< 0.05), GRP78 and GADD34 mRNA expression were increased significantly with the treatment for 12 h and 24 h (P \< 0.05), and GRP94 and ATF6 mRNA levels were increased significantly after 24 h (P \< 0.05) (Fig. [6](#Fig6){ref-type="fig"}b). We also detected some gene involved in other biological processes on ER. Compared with the control group, GDF15, HERP, WARS, SYVN1, SCAP, ACAT1, DNAJC3, CALR, EDEM1, SEC61A1, and PDIA4 mRNA expression were increased significantly by 40 μM psoralen after treatment for 24 h (P \< 0.05) (Fig. [6](#Fig6){ref-type="fig"}c). However, psoralen treatment showed little effect on the mRNA levels of ER-stress gene in HepG2 (Additional file [1](#MOESM1){ref-type="media"}: Figure S1). These results showed that psoralen can induce unfolded proteins and cause endoplasmic reticulum stress in SMMC7721 hepatoma cells.Fig. 6Effects of psoralen on gene expression of ER-stress. **a** The mRNA levels of key gene in ER-stress by different doses of psoralen treatment for 48 h. **b** The mRNA levels of key gene in ER-stress by 40Μm psoralen for 6 h, 12 h and 24 h. **c** The mRNA levels of ER functional gene by 40Μm psoralen for 6 h, 12 h and 24 h. Values are mean ± SD (n = 3) and \* is means compared to the Con group, \*p \< 0.05; \*\*p \< 0.01; \*\*\*p \< 0.001 The effects of psoralen on ER-stress related protein expression in SMMC7721 {#Sec8} --------------------------------------------------------------------------- 40 μM psoralen was used to explore the effect of psoralen on ER-stress related protein expression. Compared with the control group, GRP78 and XBP-1s protein expression were increased significantly after treatment for 24 h and 48 h (P \< 0.05). And the level of IRE1α was also up-regulated time-dependently (P \< 0.05). However, psoralen had no significant effect on the expression of CHOP protein. As a positive control drug, thapsigargin (TG) promoted the expression of GRP78 and XBP-1s protein only in 4 h (P \< 0.05) (Fig. [7](#Fig7){ref-type="fig"}a). To further evaluate the effects on ER-stress, the ER-stress inhibitor TUDC was added with the treatment of psoralen. Western blot assay showed that TUDC could not inhibit the increased expression of GRP78, XBP-1s, IREα, and CHOP proteins induced by psoralen (Fig. [7](#Fig7){ref-type="fig"}b). Similarly, TUDC had little effects on the psoralen-induced up-regulation of GRP78 and DDIT3 mRNA expression (Fig. [7](#Fig7){ref-type="fig"}c). The results suggested that psoralen can induce the endoplasmic reticulum stress in SMMC7721 hepatoma cells.Fig. 7Effects of psoralen on ER-stress related protein expression in SMMC7721. **a** Western blot assays showing the effects of psoralen and thapsigargin on ER-stress related protein expression. **b** Western blot assays showing the effects of TUDC on the expression of ER-stress related protein induced by PSO. **c** The mRNA levels of GRP78 and DDIT3 with TUDC under PSO and TG treatment. Values are mean ± SD (n = 3) and \* is means compared to the Con group, \*\*P \< 0.01 Discussion {#Sec9} ========== Psoralen is extracted from dried fruits of *P. corylifolia* (L.) from *Leguminosae*, which shows an active effect including anti-tumor, photosensitivity, cardiovascular protection, anti-histamine and estrogen-like pharmacological effects \[[@CR25]--[@CR29]\]. In the traditional Chinese medicine, the property and flavor of *P. corylifolia* (L.) are spicy, bitter and warm, and it has warming kidney and enhancing yang, preventing asthma, to warm the kidney and the spleen and check diarrhea, and removing beverage effect \[[@CR30]\], and *P. corylifolia* can be used as basic remedy to suppress tumor \[[@CR13]--[@CR15], [@CR31], [@CR32]\]. However, the roles of psoralen on the treatment of various tumors are not well understood. In our previous studies, we found that psoralen can inhibit the proliferation of Y1 adrenocortical tumor cells and AtT20 pituitary tumor cells. Based on these results, we explored the inhibition effects of psoralen on SMMC7721 human hepatoma cells for further study. Our results demonstrated that psoralen could significantly inhibit the proliferation of SMMC772 and induce apoptosis. In addition, we found that psoralen could induces cell cycle arrest at G1 phase and induce ER-stress to promote cell apoptosis, thus revealing the molecular mechanism of psoralen in anti-hepatoma effect in vitro. As the malignant proliferation of hepatoma cells is an important characteristic of the biological behavior of tumor cells, the inhibition effect of psoralen on the proliferation of SMMC7721 hepatoma cells has an objective prospect of research and development. Similarly, psoralen can inhibit the proliferation of BGC-823 human gastric cancer cells and the half inhibitory rate (IC50) was 5.82 μg/mL \[[@CR33]\]. Moreover, MG-63 osteogenic sarcoma cells were inhibited with 640 μmol/L psoralen treatment for 72 h, and the cell inhibition rate was 77.74%, and the cells became smaller, rounder, vacuoles and pseudopodia disappeared \[[@CR34]\]. Although there are lots of studies about the pro-apoptotic function of psoralen, however, recent study demonstrated that psoralen could inhibit the apoptosis and increase the proliferation of osteoporotic osteoblasts by modulating IRE1--ASK1--JNK pathway \[[@CR35]\]. Cell cycle disorder is an important feature of tumor cells, especially malignant transformation, so cyclin is an important target for the anti-cancer effect of psoralen. Among them, CyclinD1 and CyclinE1 play vital roles in the regulation of cell cycle, which are positive regulators of cyclin-dependent kinase (CDK). CyclinD1 associates with CDK4/6 and CyclinE1 binds to CDK2, which promote cells enter S phase from G1 phase, causing cell division and proliferation \[[@CR36]\]. Once the proteins are over-expressed, the cell will shorten G1 phase and advance into S phase, causing continued proliferation and increasing the risk of cancer \[[@CR37]\]. It has been reported that psoralen can form a complex with DNA in the form of non-covalent bond that cause DNA damage \[[@CR38]\], arrest cells in G1 phase, and repair damaged DNA. Psoralen can arrest cells in G1 and G2 phases, and significantly reduce the proportion of S phase cells thereby inhibiting cell proliferation of MCF-7 human breast cancer cells \[[@CR27]\]. However, another study reported that 10 μM and 30 μM psoralen can promote the progression of MCF-7 cells from G1 to S phase \[[@CR39]\]. It suggested that the regulation of psoralen on cell cycle in different tumors was not the same. Furtherly, psoralen can induce late apoptosis of SMMC7721 cells and show dose-effect relationship. According to recent studies, psoralen-induced apoptosis of K562 cells was observed under electron microscope \[[@CR40]\]. In gastric cancer, early apoptosis of BGC-803 cells could be obvious after psoralen treatment for 48 h \[[@CR41]\]. In this research, it suggests that psoralen has an exact apoptosis-inducing effect on SMMC7721 hepatoma cells, however, it is mainly not through the molecular mechanisms of the pro-apoptotic protein Bax and the apoptosis inhibitor protein Bcl-2. As well known, apoptosis may be involved in endoplasmic reticulum stress. SMMC7721 cells were observed under transmission electron microscopy after 24 h of psoralen treatment, and we found that the endoplasmic reticulum was dilatated. Then the mRNA and the protein levels of endoplasmic reticulum-associated functional molecules were detected, which also indicated that psoralen could induce ER-stress. The endoplasmic reticulum is a membranous intracellular organelle, including rough endoplasmic reticulum and smooth endoplasmic reticulum. It is mainly involved in the processes like synthesis and secretion of membrane proteins, the correct folding of proteins, the storage of Ca^2+^, the synthesis and the metabolism of lipids and cholesterol \[[@CR23]\]. Due to the influence of extrinsic or intrinsic factors, misfolded and unfolded proteins accumulate in the lumen of the endoplasmic reticulum as well as the disorder of the Ca^2+^ balance are known as ER-stress. The unfolded protein reaction in cells mainly involves three signaling pathways: double-strand RNA-activated protein kinase-like ER kinase (PERK) pathway, inositol-requiring enzyme 1 (IRE1) pathway, activating transcription factor 6 (ATF6) pathway. Under physiological conditions, PERK, IRE1 and ATF6 are all combined with glucose regulatory protein 78 (GRP78/Bip) in an inactive state \[[@CR42]\]. Once ER-stress, a large number of unfolded or misfolded proteins will snatch GRP78 that originally binds to three kinds of responsive proteins, making them exposed and activated. At present, it has been found that GRP78 is closely related to the occurrence and development of liver cancer \[[@CR43]\], breast cancer \[[@CR44]\] and lung cancer \[[@CR45]\]. The expression of GRP78 is increased when ER-stress occurs, which can promote the misfolded or unfolded protein to normal status, thereby reducing the endoplasmic reticulum load and maintaining the stability of the intracellular environment. Glucose regulatory protein 94 (GRP94) is mainly stored in the endoplasmic reticulum and is an important factor in the endoplasmic reticulum stress response. When stress occurs in the endoplasmic reticulum, GRP94 can induce the expression of C/EBP homologous protein, thereby down-regulating Bcl-2 and promoting apoptosis \[[@CR46]\]. This study showed psoralen plays a clear role in promoting the expression of GRP78 and GRP94, which can induce unfolded protein reaction and cause apoptosis. Because PERK is a type I endoplasmic reticulum transmembrane protein, its own activation of phosphorylation also results in phosphorylation of the catalytic substrate eIF2α. Phosphorylated eIF2α can inhibit the synthesis of protein and reduce the endoplasmic reticulum load in a short period of time, but long-term inhibition also leads to apoptosis. Furthermore, it can also promote the translation of activating transcription factor-4 (ATF-4) \[[@CR47]\]. Overexpression of ATF4 can activate the CHOP apoptotic pathway, which can activate GADD34 to cause the damage of oxygen free radicals. ATF6 is a type II transmembrane protein of the endoplasmic reticulum, which is separated from GRP78 and can be activated by S1P and S2P cleavage during ER-stress. Activated ATF6 promotes transcription and expression of CHOP in the nucleus, thereby inhibiting Bcl-2 and causing apoptosis \[[@CR48]\]. This results suggested that psoralen-induced endoplasmic reticulum stress mainly activates the IRE1 pathway and the ATF6 pathway. Once ER-stress occurs, phosphorylated eIF2α or IRE1 protein can activate GADD34 and ATF4. Activated GADD34 dephosphorylates p-eIF2α and restores protein synthesis \[[@CR49]\]. Overexpression of GADD34 and ATF4 induces CHOP expression, inhibits Bcl-2 expression and causes damage of oxygen free radicals which can induce apoptosis. These results showed that psoralen increased GADD34 and ATF4 genes expression, which participating in endoplasmic reticulum stress and promoting apoptosis. C/EBP homologous protein is known as DDIT3, which is in a low expression in physiological state, can activate apoptotic C/EBP homologous protein. During ER-stress, the PERK, IRE1 and ATF6 signaling pathways can promote CHOP expression. However, high expression of CHOP leads to the loss of Ca^2+^ in the endoplasmic reticulum, which increases mitochondrial permeability and induces apoptosis \[[@CR50]\]. On the other hand, it also reduces the expression of anti-apoptotic proteins and increases the expression of pro-apoptotic proteins, finally activating cellular mitochondrial apoptosis pathway \[[@CR51]\]. This study found that psoralen can promote significantly DDIT3 gene expression, but there is no significant effect on the CHOP protein level. Combined with the expression of apoptosis-related proteins Bcl-2 and Bax, it is speculated that psoralen may cause changes in Ca^2+^ concentration in the endoplasmic reticulum through highly expressed DDIT3 which triggers apoptosis. In general, for short-term ER-stress, cells can maintain the stability of the intracellular environment by enhancing the ability of protein folding, inhibiting protein synthesis and translation, and accelerating protein degradation. They are self-protective and belong to normal physiological phenomena. When ER-stress continues, a large amount of unfolded or misfolded protein will accumulate, which will induce unfolded protein response (UPR) and cause apoptosis. This study indicated that psoralen could cause a significant induction of ER-stress in dose- and time-effect relationship, which causing apoptosis of hepatoma cells. In addition, psoralen might regulate some physiological process in HCC by increased relevance mRNA levels on endoplasmic reticulum. Growth differentiation factor 15 (GDF15) is a divergent member of the BMP-subfamily of the TGF-β superfamily, which could inhibit the proliferation, migration and invasion, while promoting apoptosis of A549 cells \[[@CR52]\]. Induction of ER stress leads to upregulation of several genes such as WARS (tryptophanyl-tRNA synthetase), HERP (homocysteine-inducible ER protein with ubiquitin like domain 1), DNAJC3 (also called P58IPK), ER degradation-enhancing alpha-mannosidase-like 1 (EDEM1) and leads to caspase activation, release of mitochondrial intermembrane proteins and dissipation of mitochondrial transmembrane potential (ΔΨm) \[[@CR53]\]. Synovial apoptosis inhibitor 1 (SYVN1), an ER-associated degradation (ERAD) E3 ubiquitin ligase, could inhibit the breast cancer cell growth and metastasis through the miR-96-5p/SYVN1 axis \[[@CR54]\]. Site 1 protease (S1P), SREBP cleavage-activating protein (SCAP) and acetyl-CoA acetyltransferase 1 (ACAT1) were involved in the lipid metabolism \[[@CR55]\]. ER chaperone calreticulin (CALR), protein transport protein Sec61 subunit alpha isoform 1 (SEC61A1), protein disulfide-isomerase A4 (PDIA4) were up-regulated in the ER stress response for binding to misfolded proteins \[[@CR56]--[@CR58]\]. Besides, it is worth noting that TUDC, an ER-stress inhibitor, did not inhibit the expression of GRP78, DDIT3, XBP-1s, IREα, and CHOP, which were increased expression by psoralen. It suggested that psoralen-induced SMMC7721 cells have a strong ER stress response, and the damage to cells is irreversible and is not easily corrected by ER-stress inhibitor. It suggested that the psoralen-induced ER-stress promotes apoptosis of SMMC7721 hepatoma cells with specific characteristics. Conclusion {#Sec10} ========== Psoralen inhibits proliferation of SMMC7721 cells by arresting cells at G1 phase. Moreover, it can also cause the endoplasmic reticulum expansion and dysfunction, so as to induce the ER-stress persistently, leading to the apoptosis of hepatoma cells. The study enriched the pharmacological mechanism of psoralen in the anti-hepatoma effect. Methods {#Sec11} ======= Materials {#Sec12} --------- Psoralen (with the chemical structure illustrated in Fig. [1](#Fig1){ref-type="fig"}) was obtained from the Shanghai Institute for Food and Drug Control (NO. 110739, Shanghai, China; purity: more than 98% HPLC) and dissolved in dimethyl sulfoxide (DMSO, ST038, Biyotime, China). The final concentration of DMSO was 0.1% in all psoralen groups and had no effect on cell viability. Thapsigargin was obtained from Sigma-Aldrich (T9033, St. Louis, MO, USA), TUDC was purchased Selleck Chemicals (S3654, Houston, TX, USA). DMSO and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) were obtained from Sigma-Aldrich (M2128, St. Louis, MO, USA). Fetal bovine serum and RIPM1640 were purchased from Gibco (Grand Island, NY, USA), penicillin and streptomycin and 0.25% Trypsin (0.02% EDTA) were purchased from Hyclone (Marlborough, MA, USA). Trizol was obtained from Invitrogen Life Technologies (Grand Island, NY, USA), PrimeScript^®^RT reagent Kit, SYBR^®^ Premix Ex Taq™ (Tli RNaseH Plus) II were purchased from TaKaRa Bio Inc (Otsu, Japan). RIPA lysis buffer, BCA kit, IgG(H + L) and IgG(H + L) second antibodies were purchased from Beyotime Biotechnology (Shanghai, China), Annexin V-FITC Apoptosis Analysis Kit and PI/RNase Staining Solution were purchased from Sungen Biotech (Tianjing, China). CyclinD1 (ab134175), CyclinE1 (ab33911), CDK4 (ab68266), Bax (ab182733) and Bcl-2 (ab182858) antibodies were purchased from Abcam (Cambridge, MA, USA). GRP78 (3177s), GRP94 (2104s), IRE1α (3294s), XBP-1S (12782s), PERK (3192s) and p-eIF2α (9721s) antibodies were purchased from Cell signaling technology. β-actin (A5441) and p27 (SAB4500067) antibodies were purchased from Sigma-Aldrich (St. Louis, MO, USA). CHOP (NBP2-13172) antibody was purchased from Novus Biologicals (Littleton, CO, USA). Cell culture and treatment {#Sec13} -------------------------- The human hepatoma SMMC7721 cells were cultured in RIPM1640 medium containing 1% penicillin--streptomycin and 10% FBS. According to different experimental purposes, drug treatment was as follows: SMMC7721 cells were plated at a density of 5000 cells/well in 96-well plates with 24 h incubation, and treated with 10--80 μM psoralen for 24 h, 48 h and 72 h, then the proliferation of cells was detected by MTT assay. The cells were plated at a density of 1 × 10^4^ cells/well in 12-well plates with 24 h incubation, and treated with 40 μM psoralen and 0.1 μM thapsigargin for 12 h, 24 h and 48 h, then the cell morphologic change was observed by inverted microscope. The cells were plated into 6-well plates at a density of 1 × 10^5^ cells per well in 6-well plates with 24 h incubation, and treated with 40 μM psoralen and 0.1 μM thapsigargin for 24 h, then the ultrastructure of cells was observed by transmission electron microscope; or the cells were treated with 10--50 μM psoralen for 48 h, then the cell apoptosis and cell cycle were tested by flow cytometry; or the cells were treated with 40 μM psoralen for 6 h, 12 h, 24 h or 48 h, then the expression of mRNA and protein expression were detected by RT-qPCR or Western blotting. Additionally, 0.1% DMSO was as a control group in all experiments. MTT assay {#Sec14} --------- Cell viability was tested using the MTT assay according to the manufacturer's instructions. SMMC7721 cells were plated into 96-well plates at a density of 5000 cells/well for 24 h. Following treatment with psoralen for 24 h to 72 h, the medium was removed and cells were washed with PBS. MTT (0.5 mg/mL) was then added to each well and the mixture was incubated for 4 h at 37 °C. MTT reagent was then replaced with DMSO (150 μL per well) to dissolve formazan crystals. After the mixture was shaken at room temperature for 10 min, absorbance was measured at 490 nm using a microplate reader (Bio-Tek, Winooski, VT, USA). Routine transmission electron microscopy (TEM) of adherent cells {#Sec15} ---------------------------------------------------------------- SMMC7721 cells were plated at a density of 1 × 10^5^ cells/well in 6-well plates and treated with 40 μM psoralen or 0.1 μM Thapsigargin for 24 h. Cells were fixed in situ in 2.5 % glutaraldehyde in 0.05 M phosphate buffer (pH 7.4) for 2 h at 4 °C, then cells were scraped down and to centrifuge for 15 min in 2000 rpm, then discarding the supernatant and resuspending in the 2.5% glutaraldehyde buffer. After fixation, the samples were post-fixed in 1% OsO~4~ for 4 h and dehydrated in increasing concentrations of ethanol and propylene oxide series. Subsequently, the samples were embedded in Epon 618 resin and polymerized for 48 h at 60 °C. Then, the ultrathin sections (90--100 nM) samples were mounted on the copper grids and were stained with 0.5% aqueous uranyl acetate and lead citrate. Subsequently, the grids were air-dried and examined under a transmission electron microscope (Philips Tecnai-12, Netherlands) at 25 kV. Flow cytometry for cell cycle and apoptosis {#Sec16} ------------------------------------------- SMMC7721 cells were plated into 6-well plates at a density of 1 × 10^5^ cells per well in 6-well plates, after 24 h incubation at 37 °C, cells were treated with 40 μM psoralen for 48 h. Cycle arrest and apoptotic cells were detected by flow cytometric analysis. After the treatment, cells were collected by trypsinization and washed twice with PBS. For cell cycle assay, the collected cells were stained with propidium iodide (PI) (Sanjian, Tianjin, China). Cellular apoptosis was determined with Annexin V-FITC and PI using FITC Apoptosis Detection kit (Sanjian, Tianjin, China). Stained cells were assessed by BD FACSalibur flow cytometry (Franklin Lakes, NJ, USA), and the data were analyzed by FlowJo software (TreeStar, Ashland, OR, USA). Quantitative real-time reverse-transcription PCR (qRT-PCR) {#Sec17} ---------------------------------------------------------- Total RNA was extracted using TRIzol (Invitrogen) according to protocols from the manufacturers. The purity and integrity of the RNA were checked spectroscopically using a NanoDrop 2000/c spectrophotometer (Thermo). Then, for each sample, 2 μg RNA was reverse transcribed to obtain the cDNA template using PrimeScript^®^RT reagent Kit (TaKaRa). Each cDNA sample was diluted 5 times for qRT-PCR amplification; qRT-PCR was performed using the fluorescent dye SYBR^®^ Premix Ex Taq™ (Tli RNaseH PlusII, TaKaRa) with a 7500 Fast Real-Time PCR System. Amplification was performed with the following fast time course: 95 °C 30 s, 95 °C 5 s, 65 °C for 30 s for 40 cycles. Relative mRNA expression values were determined by the 2−ΔΔCt method using human β-actin as the normalization control. Western blotting {#Sec18} ---------------- SMMC7721 cells were lysed with RIPA lysis buffer (Beyotime, Haimen, China) supplemented with PMSF. The proteins were separated through 10% SDS-PAGE gels electrophoresis, and transferred to polyvinylidene difluoride membranes, and incubated in block buffer and then incubated with primary anti-bodies, including Bax (1:2000), Bcl-2 (1:2000), CDK4 (1:1000), CyclinD1 (1:30,000), CyclinE1 (1:1000), P27 antibody (1:1000), GRP78 (1:1000), CHOP (1:1000), IRE1α (1:1000), XBP-1S (1:1000), and β-actin antibody (1:20,000), followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibody. Protein expression was visualized by enhanced chemiluminescence assay and signal Signals were detected using a Fluorchem E system (Protein-Simple, USA). Statistical analysis {#Sec19} -------------------- Data were expressed as the mean ± SD. Significant differences were accepted at the 0.05 level of probability and were statistically determined with ANOVA followed by a Newman--Keuls post hoc test using GraphPad Prism6 (San Diego, CA). Additional file =============== {#Sec20} **Additional file 1: Figure S1.** Effects of psoralen on gene expression of ER-stress in HepG2. The mRNA levels of key gene in ER-stress in the dose of 40 μM psoralen or 0.1 μM thapsigargin for 24 h. Values are means ± SD (n = 3) and \* is means compared to the Con group, \*\*, p \< 0.01; \*\*\*, p \< 0.001. PSO : psoralen HCC : hepatocellular carcinoma ER : endoplasmic reticulum TG : thapsigargin PERK : protein kinase R-like ER kinase ATF6 : activating transcription factor 6 IRE1 : inositol-requiring enzyme 1 GRP78 : glucose-regulated protein of 78 UPR : unfolded protein response CHOP : C/EBP homologous protein XBP1 : X-box binding protein 1 TUDC : tauroursodeoxycholic acid qRT-PCR : quantitative real-time PCR **Publisher\'s Note** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Xiaomin Wang and Peike Peng contributed equally to this work Not applicable. All the experiments were carried out by XW, the manuscript was written by PP, all data were analyzed by ZP and ZF. We are thankful to WL and XL for providing her valuable inputs for English. All authors read and approved the final manuscript. This work was supported by the National Natural Science Foundation of China (No. 81473562). All data generated or analysed during this study are included in this published article. This study was approved by the Ethics Committee of Shanghai University of Traditional Chinese Medicine, and the written informed consent was also obtained from each participant in advance. Not applicable. The authors declare that they have no competing interests.

{ "pile_set_name": "PubMed Central" }

Glioblastomas (GBM) are the most aggressive tumors accounting for 45.6% of primary malignant brain tumors[@b1]. The standard therapy for GBM comprises surgery followed by radiation therapy with adjuvant chemotherapy[@b2][@b3][@b4]. Despite advances in multimodal therapies, the median survival of patients with GBM is approximately 15 months with a 5-year survival rate of 5.0% after diagnosis[@b5]. The highly infiltrative, heterogeneous and mutable nature of GBM[@b6] often contributes to tumor recurrence and resistance to therapies. Current cytotoxic chemotherapeutic agents used to treat GBM include carmustine, lomustine, and carboplatin[@b2]. A recent trial of combining bevacizumab with standard chemotherapy and radiation therapy only marginally improved overall survival[@b7]. Therapy regimens are currently being developed to target EGFR, VEGFR, PDGFR, Ras pathway, mTOR, histone acetylation and integrins[@b8], and thus far these molecular-targeted therapies have produced poor-to-modest clinical responses[@b6]. Identification of more effective therapeutic agents that work as a single agent or in combination with existing drugs are clearly needed. The incidence rate of GBM and other glial tumors is higher in males than females[@b1]. Females of reproductive ages demonstrate a survival advantage over both males and post-menopausal women. Usage of exogenous hormones reduces the risk of glioma development[@b9][@b10][@b11][@b12]. Estrogen improves the survival in a glioblastoma orthotopic animal model[@b13]. Estrogen mediates its effects through the estrogen receptor α (ERα) and estrogen receptor β (ERβ). ERβ has quite different function than ERα, and ERβ is considered as a tumor suppressor. Recent studies demonstrated that ERβ reduces proliferation and induces apoptosis in several cancer cells[@b14][@b15][@b16][@b17][@b18][@b19][@b20] and that its expression declines during tumor progression[@b21][@b22][@b23][@b24]. Recent studies including ours demonstrated that GBM cells express ERβ, and high expression of ERβ was an independent favorable prognostic factor[@b25][@b26][@b27]. Collectively, these correlative findings suggest that estrogen and ERβ play a significant role in suppression of GBM; however, the mechanisms are poorly understood. Furthermore, estrogen as potential therapy for GBM has limited therapeutic application due to the risk of breast and uterine cancers in women, prostate cancer and feminization in men, and may increase risk of heart disease in both sexes[@b28][@b29][@b30][@b31]. Even though ERα and ERβ are structurally similar, their ligand-binding domains differ enough to be selective for different ligands[@b32]. Recent studies identified a number of selective synthetic and natural ERβ agonists which are currently being investigated for therapeutic use[@b33]. One synthetic compound LY500307 is a highly potent and selective ERβ agonist; it has a 12-fold higher affinity for ERβ than ERα and exhibits 32-fold more functional potency. Further, preclinical studies showed that LY500307 treatment dose dependently reduced the prostate weight in a mouse model of benign prostatic hyperplasia[@b34]. LY500307 was well tolerated in BPH patients with no side effects[@b35] and more importantly, LY500307 is currently being tested in phase 2 clinical trials for improving negative symptoms and cognitive impairment associated with Schizophrenia. However, it remains unknown whether LY500307 has efficacy in treating GBM. Here, we tested the efficacy of the ERβ agonist LY500307 as a novel therapeutic agent for treating GBM using *in vitro* and *in vivo* preclinical models. Our results demonstrate that LY500307 selectively kills established and patient-derived primary GBM cells with minimal toxicity on normal cells. Mechanistic studies showed that LY500307 modulates cell cycle, apoptosis and DNA damage response pathways and sensitizes GBM cells to current chemotherapeutic agents. Further, LY500307 reduced GBM growth *in vivo* in orthotopic models and prolonged the survival of tumor-bearing mice. This represents the first report demonstrating specificity of ERβ ligand LY500307 on GBM cells and suggests that LY500307-- ERβ-mediated inhibition may be an effective strategy for targeted therapy. Results ======= Selective ERβ agonist LY500307 reduces the cell viability and survival, and induces apoptosis of GBM cells ---------------------------------------------------------------------------------------------------------- To test whether LY500307 reduces cell viability of GBM cells, MTT cell viability assays were performed. Treatment with LY500307 significantly reduced the viability of various GBM cell lines in a dose-dependent manner. In contrast, viability of normal astrocytes was not affected at the tested doses, suggesting that LY500307 has tumor cell--specific activity ([Fig. 1a](#f1){ref-type="fig"}). Further the effect of LY500307 on cell viability of several patient derived GBM cells was examined. As shown in [Supplementary Fig. 1a](#S1){ref-type="supplementary-material"}, all patient derived primary GBM cells tested expressed ERβ but not ERα. Cell Titer-Glo luminescent cell viability assays revealed that LY500307 reduced the viability of various patient-derived GBM cells in a dose dependent manner ([Fig. 1b](#f1){ref-type="fig"}). We next examined the effect of LY500307 on survival of GBM cells using colony formation assays. As shown in [Fig. 1c](#f1){ref-type="fig"}, LY500307 significantly reduced the colony formation of U87 and U251 GBM cells. We then tested whether LY500307 induces apoptosis of GBM cells using the Annexin V assay. LY500307 significantly induced Annexin V-positive apoptotic cells in U87, U251 ([Fig. 1d,e](#f1){ref-type="fig"}) and patient derived primary GBM (GBM10) cells ([Supplementary Fig. 1b](#S1){ref-type="supplementary-material"}). Collectively, these results suggested that LY500307 has potential to selectively reduce cell viability, decrease survival and induce apoptosis of GBM cells. LY500307 enhances ERβ signaling in GBM cells -------------------------------------------- To determine whether LY500307 promotes activation of classical ERβ--ERE signaling, U87 GBM cells were transfected with the ERE-luciferase reporter and treated with LY500307 for 24 h. LY500307 significantly increased the ERE luciferase activity in GBM cells ([Fig. 2a](#f2){ref-type="fig"}, left panel). In addition, LY500307 stimulation enhanced the expression of ERβ and its target gene MDA7/IL-24 in GBM cells ([Fig. 2a](#f2){ref-type="fig"}, middle and right panel). To determine whether LY500307 modulates the non-classical ERβ signaling, GBM cells were transfected with AP-1, SP-1 and NF-κB-Luc luciferase reporter plasmids followed by treatment with vehicle or LY500307. As shown in [Fig. 2b](#f2){ref-type="fig"}, LY500307 significantly increased AP-1 and SP-1 luciferase activities and reduced NF-κB-Luc activity in GBM cells. To determine whether LY500307 has an effect on ERβ-mediated rapid extra-nuclear signaling, GBM cells were treated with LY500307 for short periods of time (5, 15, and 30 min) and the phosphorylation status of p38MAPK, JNK, ERK1/2 and Akt was examined. Western blot analysis showed that LY500307 significantly increased the phosphorylation of proapoptotic stress activated kinases p38MAPK and JNK in GBM cells. However, LY500307 treatment reduced phosphorylation of ERK1/2, and Akt phosphorylation was not affected ([Fig. 2c](#f2){ref-type="fig"}). To examine the ERβ selectivity of LY500307, cell viability assays were performed using ERβ knockdown in GBM cells. LY500307-mediated reduction in cell viability was significantly compromised in ERβ shRNA cells than in control shRNA cells, suggesting the specific requirement of ERβ for LY500307 actions ([Supplementary Fig. 2](#S1){ref-type="supplementary-material"}). We also confirmed the ERβ-mediated GBM suppression using GBM cells that overexpress ERβ. The effect of ERβ expression on GBM cell proliferation was studied using Cell Titer-Glo assay. ERβ overexpression significantly reduced the proliferation of U87 and U251 GBM cells ([Supplementary Fig. 3](#S1){ref-type="supplementary-material"}). Collectively, these results suggest that LY500307 has potential to modulate both classical and non-classical ERβ signaling and that LY500307 can enhance ERβ expression in GBM cells. Analysis of LY500307 induced global transcriptional changes in GBM cells ------------------------------------------------------------------------ To identify the global changes in gene expression following LY500307 treatment in GBM cells, we performed global transcriptome analysis. U87 cells were treated with either vehicle or LY500307 for 48 h, and the isolated RNA was subjected to RNA-seq analysis. Genes that had at least a 1.5-fold change in expression (p \< 0.01) were chosen for analysis. Overall, 3204 genes were differentially expressed in LY500307-treated U87 cells; 1568 genes were upregulated and 1636 genes were downregulated. A representative heat map was shown in [Fig. 3a](#f3){ref-type="fig"}. The complete list of differentially expressed genes is given in [Supplementary Table 1](#S1){ref-type="supplementary-material"}. To further examine the biological significance of the differentially expressed genes, IPA analysis was performed. The top networks that were differentially expressed after LY500307 treatment were related to tissue development, cell cycle, cellular movement, cellular assembly and organization, cellular function and maintenance ([Fig. 3b](#f3){ref-type="fig"}). Analysis of molecular and cellular functions of differentially expressed genes revealed that they are involved in cellular growth, proliferation, cell death and survival ([Supplementary Fig. 4](#S1){ref-type="supplementary-material"}). The canonical pathways that were modulated by LY500307 include cell cycle, DNA damage, p53 signaling, and checkpoint regulation ([Fig. 3c](#f3){ref-type="fig"}). In addition, pathways related to cancer such as molecular mechanism of cancer, GBM signaling, Wnt signaling and glioma invasion were also altered. Ability of LY500307 to modulate these pathways were independently validated by using qRT-PCR on selective genes in U87 ([Fig. 3d,e](#f3){ref-type="fig"}), U251 ([Fig. 3f](#f3){ref-type="fig"}) and patient derived GBM (GBM10) cells ([Supplementary Fig. 5](#S1){ref-type="supplementary-material"}). Collectively, these results suggest that LY500307 modulates the expression of genes involved in cell cycle, cell death, survival, and DNA damage response. LY500307 induces G2/M cell cycle arrest --------------------------------------- Since RNA-seq results demonstrated cell cycle as the top network modulated by LY500307, we examined the effect of LY500307 on cell cycle distribution of GBM cells. Flow cytometry analysis of PI-stained cells revealed that LY500307 treatment significantly increased the percentage of cells in G2/M phase in U87, U251, LN229 and GBM10 cells when compared to vehicle ([Fig. 4](#f4){ref-type="fig"}), further supporting the RNA-seq pathway analysis of cell cycle and G2/M DNA damage checkpoint regulation. LY500307 sensitizes GBM cells to chemotherapeutic agents -------------------------------------------------------- Emerging data provide the evidence that using a multi-targeted approach has an advantage over using a single agent for GBM therapy, and that sensitizer drugs that enhance the utility of chemotherapy are advantageous. We performed an initial *in vitro* screen of 119 FDA-approved drugs in combination with LY500307 on the cell viability of U87 cells. Our results demonstrated that LY500307 sensitized U87 cells to several FDA approved drugs and nine drugs that showed synergism with LY500307 are presented in [Fig. 5a](#f5){ref-type="fig"}. Interestingly, many of the chemotherapeutic agents that LY500307 sensitized cause DNA damage and apoptosis. We independently validated some of the potent compounds from the initial screen including cisplatin, bleomycin. We also tested potent GBM chemotherapeutic agents including lomustine, temozolomide. Our results revealed that LY500307 sensitized GBM cells to cisplatin, bleomycin and lomustine ([Fig. 5b--d](#f5){ref-type="fig"}). Importantly, we also found that LY500307 significantly sensitized temozolomide-resistant cells U138 to temozolomide-mediated reduction in cell viability ([Fig. 5e](#f5){ref-type="fig"}). LY500307 reduces GBM progression in an orthotopic model ------------------------------------------------------- To evaluate the effect of LY500307 on *in vivo* tumor growth, U251 cells that express luciferase reporter were injected orthotopically into the brain of mice. After tumors were established, the mice were randomized into a control group, which received vehicle, and a treatment group, which received LY500307 through oral gavage. GBM progression was measured by monitoring luciferase intensity using Xenogen-IVIS imaging system weekly. Compared to vehicle, LY500037 treatment significantly reduced the GBM progression in the U251 tumor model ([Fig. 6a](#f6){ref-type="fig"}). Immunohistochemical analysis revealed that LY500307 treatment significantly reduced the number of proliferation marker Ki-67-positive cells ([Fig. 6b](#f6){ref-type="fig"}) and increased the number of TUNEL-positive apoptotic cells ([Fig. 6c](#f6){ref-type="fig"}). To further confirm the effect of LY500307 on apoptosis, the expression of Cleaved Caspase-3 and Bcl-2 was examined. As shown in [Supplementary Fig. 6a](#S1){ref-type="supplementary-material"}, LY500307 treatment significantly increased the expression of Cleaved Caspase-3 and reduced the Bcl-2 expression in tumors compared to controls. Collectively, these results suggest that LY500307 inhibits the GBM progression *in vivo* and induces apoptosis. LY500307 increases survival in syngeneic glioma mouse model ----------------------------------------------------------- Immune effects during tumor progression play key roles in GBM progression, and GBM-mediated immunosuppression acts as a barrier to the efficacy of chemotherapeutic agents[@b36][@b37]. To study the anti-tumor activity of LY500307 in the presence of an intact immune system, GL26 cells stably expressing luciferase were implanted into C57BL/6 mice. After establishment of tumors, the mice were randomized into two groups based on luciferase intensity and treated with either vehicle or LY500307 daily by oral gavage. Survival was calculated using Kaplan Meier analysis. Compared to vehicle, LY500307 treatment significantly increased the overall survival of the mice ([Fig. 7a](#f7){ref-type="fig"}). Further, immunohistochemical analysis of tumor sections revealed that LY500307 treatment significantly reduced the expression of the proliferation marker Ki-67 and increased the number of TUNEL-positive apoptotic cells ([Fig. 7b,c](#f7){ref-type="fig"}). Further, the expression of Cleaved Caspase-3 was increased and Bcl-2 expression was decreased significantly in LY500307 treated tumors compared to controls ([Supplementary Fig. 6b](#S1){ref-type="supplementary-material"}). Discussion ========== ERβ was initially discovered as second receptor of estrogen[@b38], and several studies demonstrated that ERβ expression was downregulated during tumor progression. Ligands that increase ERβ expression or activity will have therapeutic utility. Recently, several groups including ours showed that ERβ expression is reduced during the progression of gliomas and that plant-derived ligands of ERβ exhibit anti-tumor activities[@b25][@b39][@b40]. However, plant-derived ligands have low efficacy and are difficult to synthesize in large quantities. Thus, more potent synthetic ERβ ligands that work more effectively are urgently needed for clinical application. Our work provides the evidence that LY500307, a synthetic ERβ agonist has potential to specifically reduce the proliferation of GBM cells with high potency, induce apoptosis, promote G2/M cell cycle arrest and sensitize GBM cells to chemotherapeutic agents. Further, our results also demonstrated that LY500307 has the ability to reduce GBM progression *in vivo* and improve the survival of mice. Selective killing of cancer cells without affecting normal cells is critical for successful chemotherapy. Our results demonstrated that LY500307 selectively reduces the viability of GBM cells with little effect on normal astrocytes. Furthermore, our *in vivo* studies showed no LY500307-related toxicities in various organs (data not shown). Previous studies demonstrated that overexpression of ERβ or activation with ligands results in a decrease in proliferation and induction of apoptosis of cancer cells, and that depending on the cell type, activation of ERβ can promote either G~2~ or G~1~ arrest [@b14][@b41]. Our results suggested that activation of ERβ with LY500307 induces apoptosis of GBM cells. Earlier studies showed that overexpression of ERβ or activation with ligands causes G2/M arrest in several cancer cells, and that ERβ increases the expression of several genes, including p21 and GADD45A[@b14][@b41][@b42][@b43]. Our results demonstrate that LY500307 treatment causes cell cycle arrest of GBM cells in G2/M phase and increases the expression of p21 and GADD45A in GBM cells also confirms the earlier observations in other model cells. ERβ functions as a transcription factor that modulates both classical estrogen response element (ERE)--containing genes and non-ERE genes via non-classical signaling by interacting with AP1, SP1, NF-κB and KLF5 transcription factors[@b44][@b45][@b46]. Our RNA-seq analysis revealed that LY500307 modulates several genes involved in cell cycle regulation, apoptosis and DNA damage response. Further, LY500307 treatment also down regulated several genes involved in tumor development and progression including Wnt signaling, GBM signaling and glioma invasion signaling. Since ERβ regulates multiple pathways directly or indirectly by binding to the key mediators, it is not surprising to observe the modulation of multiple pathways by LY500307 that includes induction of cell cycle arrest and cell death molecules and suppression of oncogenic signaling molecules. In addition, our results showed that LY500307 promotes ERβ extra-nuclear signaling, leading to activation of the proapoptotic p38 and JNK pathways. Our results showing ERβ knockdown reduced the efficacy of LY500307 to inhibit GBM proliferation further confirms that specific role of ERβ in LY500307 actions. The current treatment options for GBM are poor, and the mortality rates are very high. Multiple challenges remain, including difficulty of complete resection of GBM, rapid and aggressive tumor relapse and resistance to external radiation and chemotherapy. Therefore, sensitizer drugs that enhance the utility of chemotherapy are urgently needed. Recent studies showed that selective targeting of ERβ with agonists can sensitize malignant pleural mesothelial cells to cisplatin toxicity[@b19] and that inhibition of ERβ increased DNA repair that contributes to cisplatin resistance in medulloblastoma cells[@b47]. Further, ERβ expression influences the malignant pleural mesothelial cell responsiveness to gefitinib[@b19]. Our results using 119 FDA-approved drugs in combination with a low dose of LY500307 revealed sensitization of GBM cells to several chemotherapeutic agents, including axitinib, doxorubicin, cisplatin, bleomycin and etoposide. Our RNA-seq studies revealed potential of ERβ agonists to down-regulate a number of genes involved in DNA repair and DNA damage response. Most importantly downregulation of DNA repair genes such as DDR1 and DDR2 may provide mechanistic explanation for LY500307-mediated sensitization effects. Further, our studies also revealed that LY500307 sensitizes GBM cells to currently used chemotherapeutic agents temozolomide and lomustine. ERβ agonists' ability to suppress pathways involved in DNA repair can be exploited in future to promote apoptosis of GBM cells. We previously reported that the plant-derived ERβ agonist liquiritigenin reduces the growth of subcutaneous glioma xenograft tumors[@b25]. Recently, salicylketoxime-based estrogen receptor β agonists also reduced the glioma growth in subcutaneous models[@b48]. However, lack of drug testing using an orthotopic model is a limitation of these studies. Further, the tumor microenvironment and the presence of intact immune system must be considered in efficacy testing of chemotherapeutic drug on GBM growth. In this study, we have tested the effect of LY500307 using both orthotopic tumor models and syngeneic model with intact immune system. Our results demonstrate that the selective ERβ agonist LY500307 reduced GBM progression as well as enhanced survival in syngeneic mouse models. Since LY500307 readily crosses the blood--brain barrier, it is currently being tested in clinical trials and it is well tolerated, it can be readily transferred to clinical use with current chemotherapies, thereby providing an additional tool for enhancing survival in GBM patients with limited toxicity. Materials and Methods ===================== Cell culture, reagents and generation of stable ERβ overexpression and ERβ shRNA cells -------------------------------------------------------------------------------------- Human glioblastoma cell lines U87, U251, LN229 and T98G and the mouse glioma cell line GL26 were obtained from the American Type Culture Collection (ATCC). Cell lines were maintained in DMEM supplemented with 10% fetal bovine serum (Sigma Chemical Co, St. Louis, MO). LY500307 was purchased from Selleckchem (Houston, TX) and Apex Biosciences (Durham, NC). ERβ antibodies were obtained from Santa Cruz Biotechnology (Dallas, TX),GeneTex (Irvine, CA) and Millipore (Billerica, MA), Ki-67 was obtained from Abcam (Cambridge, MA) and Bcl-2 was purchased from Dako (Carpinteria, CA) and Santa Cruz Biotechnology (Dallas, TX). The p-ERK1/2, ERK1/2, p-AKT, AKT, p-p38MAPK, p38MAPK, p-JNK,JNK, Cleaved Caspase-3 and GAPDH antibodies were obtained from Cell Signaling Technology (Beverly, MA). ERβ-specific short hairpin RNA (shRNA) lentiviral plasmids, β-actin and all secondary antibodies were purchased Sigma from Chemical Co (St. Louis, MO). Normal human astrocytes were obtained from ScienCell Research Laboratories (Carlsbad, CA) and maintained in astrocyte medium. Stably ERβ-expressing GBM cells were generated by infecting them with pLenti6/V5-D-FLAG ERβ and empty control vectors and positive cells were selected with blasticidine (5 μg/mL). GBM cells stably expressing ERβ-shRNA were generated by infecting cells with human-specific lentiviral ERβ-shRNA particles and selected with puromycin (1 μg/mL), and pooled clones were used for all the studies. Lentiviral particles expressing nontargeted shRNA were used to generate control. Primary GBM cells ----------------- Primary GBM cells were isolated from discarded GBM tumor specimens after approval from Office of the Institutional Review Board (IRB) at UT Health Science Center at San Antonio. All the discarded tumor specimens were obtained from patients undergoing surgery after informed consent, and no clinical linkers or codes for the specimens accessible to any research personnel or the PI. All the methods were carried out in accordance with IRB approved guidelines. Briefly, fresh tumor tissue was collected intraoperatively, dispersed into single cells, and cultured briefly in a neurobasal media for expansion and purification followed by intracranial injection in nude mice. Patient-derived GBM neurosphere lines were maintained in Neurobasal Medium supplemented with B27 serum-free supplement and growth factors EGF (20 ng/mL), bFGF (20 ng/mL) and LIF (10 ng/mL) as described[@b49]. The resulting tumors were characterized for consistency by histology with the primary tumor, growth attributes including time to animal demise, and for gene expression profiling for the molecular subtype. GBM line 101310 (GBM10) used in the assays is grade IV GBM, MGMT hypomethylated and average survival time in mice is 108 days. Cell viability and colony formation assays ------------------------------------------ The effect of LY500307 on cell viability of GBM cells and normal astrocytes was assessed by using MTT assays. For LY500307 treatment, GBM cells were maintained in phenol red--free DMEM containing 5% dextran-charcoal treated FBS. Briefly, GBM cell lines and normal astrocytes were seeded in 96 well plates (2 × 10^3^ cells/well). After overnight incubation, cells were treated with either vehicle or varying concentrations of LY500307 for 72 h. MTT was added to each well and incubated for 4 h. Formazan crystals were solubilized in DMSO and the optical density was measured using micro-plate reader. The effect of LY500307 on the viability of patient-derived primary GBM lines and the cell proliferation rates of control and ERβ-overexpressing GBM cells were measured using CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI) in 96-well, flat, clear-bottom, opaque-wall microplates according to manufacturer's instructions. Total ATP content as an estimate of total number of viable cells was measured on an automatic Fluoroskan Luminometer. For colony formation assays, U87 and U251 cells (500 cells/well) were seeded in 6 well plates. After overnight incubation, the cells were treated with vehicle or LY500307 (5 μM) for 72 h. Then, cells were washed with PBS and the cells were allowed to grow for additional 7 days. The cells were fixed in ice-cold methanol and stained with 0.5% crystal violet solution. Colonies that contained ≥50 cells were counted. Annexin V apoptosis assay ------------------------- The apoptotic effect of LY500307 on GBM cells was analyzed using the Annexin V/PI kit as per the manufacturer's instructions (BioLegend, San Diego, CA). Briefly, U87 and U251 GBM cells were seeded in 60-mm culture plates and treated with either vehicle or LY500307 (5 μM) for 48 h. Cells were harvested at a density of 2.5 × 10^6^ cells/mL in Annexin V binding buffer and 100 μL of cell suspension was incubated with Annexin V FITC and propidium Iodide (PI) for 15 min at room temperature in the dark. Then, 400 μL of Annexin V binding buffer was added to each sample and stained cells were analyzed using flow cytometry. Cell cycle analysis ------------------- U87, U251, LN229 and GBM10 (101310) cells were seeded in 100-mm culture plates, and after overnight incubation, cells were treated with either vehicle (0.1% DMSO) or LY500307 (5 μM) for 24 h. Cells were then trypsinized and harvested in PBS, followed by fixation in ice-cold 70% ethanol for 30 min at 4 °C. Cells were washed again with PBS and stained with a mixture of 50 μg/mL propidium iodide and 50 μg/mL RNase A. The PI-stained cells were subjected to flow cytometry using a FACS Calibur (BD Biosciences). Reporter gene assays -------------------- GBM cells were maintained in phenol red--free DMEM supplemented with 5% dextran-charcoal--stripped serum for 48 h prior to transfection. Cells were transiently transfected with 250 ng of ERE-Luc, AP1-Luc, SP1-Luc, or NF-κB-Luc reporter plasmids using Turbofect transfection reagent (Thermo Scientific, Waltham, MA). After 24 h, cells were treated with either vehicle or LY500307 (1 μM) for additional 24 h. β-galactosidase reporter (50 ng) plasmid was co-transfected and used for data normalization. Cells were lysed in Passive Lysis Buffer, and Luciferase activity was measured using the luciferase assay system (Promega, Madison, WI) in luminometer. Western blotting ---------------- Whole cell lysates were prepared from GBM cells using RIPA buffer (Sigma, St Louis, MO) containing protease and phosphatase inhibitors. Total proteins (50 μg) were mixed with SDS sample buffer and separated on SDS-PAGE gels. Resolved proteins were then transferred onto nitrocellulose membranes, blocked with 5% non-fat dry milk powder for 1 h at room temperature and incubated with respective primary antibodies over night at 4 °C followed by secondary antibody incubation for 1 h at room temperature. Blots were developed using the ECL kit (Thermo Scientific, Waltham, MA). *In vitro* screening of FDA-approved drugs in combination with LY500307 on GBM cells ------------------------------------------------------------------------------------ All the FDA-approved anti-cancerous drugs were obtained from NCI at 10 mM concentration (<https://dtp.cancer.gov/organization/dscb/obtaining/available_plates.htm>). The initial screen of the combination treatment of the 119 FDA-approved drugs and LY500307 was performed on U87 cells using MTT assay. U87 cells (1 × 10^3^/well) were seeded in 96 well plates. After an overnight incubation, the cells were treated with LY500307 (1 μM) and varying doses of each FDA-approved drug at final concentrations of 0.1 μM, 1 μM, and 10 μM for 72 h in phenol red--free DMEM that contains dextran-charcoal treated FBS. Some of the compounds that sensitized GBM cells with LY500307 were further validated using U251 cells. U251 cells were seeded in 96 well plates (1 × 10^3^ cells/well) in phenol red--free DMEM containing dextran-charcoal treated FBS. After an overnight incubation, the cells were pretreated with vehicle or LY500307 (1 μM) for 48 h followed by wash-off and replenished with varying concentrations of each drug alone or in combination with LY500307 for 96 h. Cell viability was then measured using the MTT assay as described above. The temozolomide-resistant cell line U138 was subjected to pretreatment with LY500307 for 72 h followed by varying doses of temozolomide for 5 days, and then cell viability was measured using the MTT assay. RNA sequencing and qRT-PCR -------------------------- U87 GBM cells were treated with either vehicle or LY500307 (5 μM) for 48 h, and total RNA was isolated using RNAesy mini kit (Qiagen) according to the manufacturer's instructions. The purity of prepared RNA was determined using Agilent 2100 BioAnalyzer. Illumina TruSeq RNA Sample preparation was performed following manufacturer's protocol, and the samples were run on an Illumina HiSeq 2000 in duplicates. The combined raw reads were aligned to UCSC hg19, and the genes were annotated by Tophat. Genes were annotated and quantified by HTSeq-DESeq pipeline. Differential expression analysis was performed by DEseq and significant genes with at least 1.5-fold change with p \< 0.01 were chosen for analysis. The interpretation of biological pathways using RNA-seq data was performed with ingenuity pathway analysis (IPA) software using all significant and differentially expressed genes. To validate the selected genes, reverse transcription (RT) reactions were performed by using SuperScript III First Strand kit (Invitrogen, Carlsbad), according to the manufacturer's protocol. Real-time PCR was done using SYBR Green (Thermo Scientific) on an Illumina Real-Time PCR system with the following primers. Results were normalized to β-actin transcript levels, and the difference in fold expression was calculated using delta-delta-CT method. Actin-F-5′- GTGGGCATGGGTCAGAAG-3′; Actin-R-5′- TCCATCACGATGCCAGTG-3′ ESR2-F-5′- GGC AGA GGA CAG TAA AAG CA -3′; ESR2-R-5′- GGA CCA CAC AGC AGA AAG AT -3′ IL24-F-5′- CTTTGTTCTCATCGTGTCACAAC-3′; IL24-R-5′- TCCAACTGTTTGAATGCTCTCC-3′ CDKN1A-F-5′- CTGGAGACTCTCAGGGTCGAAA-3′; CDKN1A-R-5′-GATTAGGGCTTCCTCTTGGAGAA-3′;GADD45A-F-5′- GGTGTACGAAGCGGCCAA-3′; GADD45A-R-5′- GCAGGCACAACACCACGTTA-3′ PUMA-F-5′- ATGCCTGCCTCACCTTCATC-3′; PUMA-R-5′- TCACACGTCGCTCTCTCTAAACC-3′ CTNNB1-F-5′- AAAATGGCAGTGCGTTTAG-3′; CTNNB1-R-5′- TTTGAAGGCAGTCTGTCGTA-3′ SFN-F-5′- TGACGACAAGAAGCGCATCAT-3′; SFN-R-5′- GTAGTGGAAGACGGAAAAGTTCA-3′ AKT3-F-5′-TGTGGATTTACCTTATCCCCTCA-3′; AKT3-R-5′-GTTTGGCTTTGGTCGTTCTGT-3′ PML-F-5′- CGCCCTGGATAACGTCTTTTT-3′; PML-R-5′- CTCGCACTCAAAGCACCAGA-3′ CHEK1-F-5′- ATATGAAGCGTGCCGTAGACT-3′; CHEK1-R-5′- TGCCTATGTCTGGCTCTATTCTG-3′ CHEK2-F-5′- TGAGAACCTTATGTGGAACCCC-3′; CHEK2-R-5′- ACAGCACGGTTATACCCAGC-3′ GADD45B-F-5′- TACGAGTCGGCCAAGTTGATG-3′; GADD45B-R-5′- GGATGAGCGTGAAGTGGATTT-3′ WT1-F-5′- CACAGCACAGGGTACGAGAG-3′; WT1-R-5′- CAAGAGTCGGGGCTACTCCA-3′ WNT5A-F-5′- ATTCTTGGTGGTCGCTAGGTA-3′; WNT5A-R-5′- CGCCTTCTCCGATGTACTGC-3′ SOS2-F-5′- ATGTAGAGGAGCGAGTTCAGAA-3′; SOS2-R-5′- ATGGTAGTCCACTTTGTACCCT-3′ FZD5-F-5′- CATGCCCAACCAGTTCAACC-3′; FZD5-R-5′- CGGCGAGCATTGGATCTCC-3′ FZD7-F-5′- GTGCCAACGGCCTGATGTA-3′; FZD7-R-5′- AGGTGAGAACGGTAAAGAGCG-3′ NTRK3-F-5′- ACGAGAGGGTGACAATGCTG-3′; NTRK3-R-5′- CCAGTGACTATCCAGTCCACA-3′ DDR1-F-5′- AAGGGACATTTTGATCCTGCC-3′; DDR1-R-5′- CCTTGGGAAACACCGACCC-3′ DDR2-F-5′- GCTATATGCCGCTATCCTCTGG-3′; DDR2-R-5′- ACTCTGACCACTGACTGGAAG-3′ MAGI2-F-5′- TCCGGCTCAAGTGTGTCAAG-3′; MAGI2-R-5′- AGGTTGTCACGAATGATTTGCT-3′ MRAS-F-5′ TTCCTCATCGTCTACTCCGTC-3′; MRAS-R-5′ AGGATCATCGGGAATGACTCC-3′ Immunohistochemistry -------------------- Immunohistochemical studies were performed as described previously[@b25]. Coronal brain sections were incubated in xylene and passed through series of graded alcohols and then subjected to antigen retrieval using the antigen retrieval solution (Vector Lab, Inc. CA). Tissue sections were incubated in 3% H~2~O~2~ solution for 20 min and then subjected to blocking using the vector lab blocking kit. Tissue sections were incubated overnight with Ki-67 (1:100), Cleaved Caspase-3 (1:200) and Bcl-2 (1:50) and then with secondary antibodies for 45 min at room temperature. Immunoreactivity was visualized by using the DAB substrate and counterstained with haematoxylin (Vector Lab, Inc. CA). The proliferative index was calculated as percentage of Ki-67-positive cells in five randomly selected microscopic fields at 20X per slide. TUNEL analysis was performed using the *In situ* Cell Death Detection Kit (Roche, Indianapolis, IN) as per the manufacturer's protocol, and five randomly selected microscopic fields in each group were used to calculate the relative ratio of TUNEL-positive cells. DAPI was used to visualize the nuclei. *In Vivo* Orthotopic Tumor Models --------------------------------- All animal experiments were performed after obtaining UTHSCSA IACUC approval, and all the methods were carried out in accordance with the IACUC approved guidelines. U251 and GL26 cells labelled with GFP-Luciferase (1 × 10^6^) were injected orthotopically into the mouse brain at 2 mm right, 2 mm posterior and 3 mm deep from bregma into the striatum. GL26 mice were ovariectomized bilaterally to avoid the effects of circulated estrogen. Five days after injection, the mice were imaged using Xenogen IVIS system and randomized to receive either control or treatment. The control group received vehicle (30% captisol) and the treatment group received LY500307 in 30% captisol (5 mg/Kg body weight/day) orally. Tumor growth was measured weekly using Xenogen IVIS system. After treatment, the mice were euthanized, and their brains were isolated and processed for histological studies. Statistical analyses -------------------- Statistical differences between groups were analyzed with either t-test or ANOVA as appropriate using GraphPad Prism 6 software (GraphPad Software, SanDiego, CA). Student t-test was used to assess the statistical difference between control and LY500307-treated groups. Mice survival was determined using Kaplan-Meier survival curve. All the data represented in bar graphs are shown as means ± SE. A value of p \< 0.05 was considered as statistically significant. RNA-seq data was analyzed using IPA software. Additional Information ====================== **How to cite this article**: Sareddy, G. R. *et al.* Selective Estrogen Receptor β Agonist LY500307 as a Novel Therapeutic Agent for Glioblastoma. *Sci. Rep.* **6**, 24185; doi: 10.1038/srep24185 (2016). Supplementary Material {#S1} ====================== ###### Supplementary Information This study was supported by the NIH/NCI grant NIH-CA178499 (R.K.V; A.B.); ABTA Discovery Grant (GRS), CPRIT training grant (RP140105) (GRS) and the Cancer Therapy and Research Center at the University of Texas Health Science Center at San Antonio through the NCI Cancer Center Support Grant P30CA054174-17. Jan-Ake Gustafsson thanks the Robert A Welch Foundation (grant number E-0004) for support. **Author Contributions** G.R.S., A.B., R.R.T. and R.K.V. designed the experiments; G.R.S., X.L. and S.V. performed *in vitro* and IHC studies; G.R.S., J.L., X.L., L.G., A.G., D.C. and M.G. performed Animal experiments and imaging studies; A.M.S. and J.G. provided reagents and interpreted the data; G.R.S., A.B., R.R.T. and R.K.V. analyzed data; and G.R.S. and R.K.V. wrote the paper. All authors reviewed the manuscript. ![ERβ agonist L500307 reduces proliferation and induces apoptosis in GBM cells.\ (**A**) U87, U251, T98G and normal astrocytes were treated with either vehicle or LY500307 for 72 h and the cell viability was measured using an MTT assay. (**B**) Patient-derived primary GBM cells were treated with either vehicle or LY500307 for 72 h and the cell viability was measured using the CellTiter-Glo luminescent assay. (**C**) U87 and U251 cells were treated with either vehicle or LY500307 for 72 h and then cultured for 7 subsequent days. The number of colonies for each group was counted. (**D--E**) U87 and U251 cells were treated with either vehicle or LY500307 for 48 h followed by Annexin V-FITC and Propidium Iodide (PI) staining for 15 min. The Annexin V--positive apoptotic populations were determined using flow cytometry. Data are represented as mean ± SE. \*p \< 0.05, \*\*p \< 0.01, \*\*\*p \< 0.001.](srep24185-f1){#f1} ![ERβ agonist LY500307 activates ERβ signaling in GBM cells.\ (**A**) U87 GBM cells were transfected with ERE-Luc plasmid. After 24 h, the cells were treated with either vehicle or LY500307 for additional 24 h, and then reporter activity was measured (left panel). U87 cells were treated with vehicle or LY500307 for 24 h, and the mRNA expression of ERβ (ESR2) and MDA7/IL-24 was measured using qRT-PCR (middle and right panels). (**B**) U87 cells were transfected with AP-1 luc, SP-1 luc and NF-kB luc plasmids and 24 h after transfection, the cells were treated with either vehicle or LY500307 for 24 h and reporter activity was then measured. (**C**) U87 cells were treated with either vehicle or LY500307 for 5 min, 15 min and 30 min and lysed in RIPA buffer. The lysates were subjected to Western blotting with the indicated antibodies, and β-actin used as loading control. Data are represented as mean ± SE. \*p \< 0.05, \*\*p \< 0.01, \*\*\*p \< 0.001.](srep24185-f2){#f2} ![Analysis of global transcriptome changes modulated by LY500307.\ Total RNA was isolated from U87 cells that were treated with either vehicle or LY500307 for 48 h using RNeasy mini kit and the quality of RNA was tested using a Bio-analyzer and subjected to RNA sequencing. (**A**) heat map of differentially expressed genes between vehicle and LY500307 is shown. (**B**) Differentially expressed genes were subjected to pathway analysis using IPA software and the top five associated network functions of differentially expressed genes are shown. (**C**) The selected top canonical pathways are shown. (**D**--**E**) U87 cells that were treated with either vehicle or LY500307 for 48 h and differentially expressed genes that were upregulated or downregulated were validated using qRT-PCR. F, U251 cells were treated with either vehicle or LY500307 for 48 h and differentially expressed genes were validated using qRT-PCR.](srep24185-f3){#f3} ![LY500307 promotes G2/M accumulation of GBM cells.\ U87, U251, LN229 and primary GBM10 cells were treated with either vehicle or LY500307 for 24 h, and the cells were fixed in 70% ethanol and subjected to PI staining for 20 min. Cell cycle distribution was analyzed using flow cytometry. Data are represented as mean ± SE. \*\*p \< 0.01, \*\*\*p \< 0.001.](srep24185-f4){#f4} ![LY500307 sensitizes GBM cells to chemotherapeutic agents.\ (**A**) U87 GBM cells were treated with either vehicle or LY500307 for 24 h and then treated with one of 119 different FDA-approved drugs for an additional 72 h. Cell viability was determined using an MTT assay, and the list of the drugs that had synergistic activity in the presence of LY500307 are represented in [Fig. 5A](#f5){ref-type="fig"}. U251 GBM cells were pretreated with LY500307 for 48 h followed by treatment with varying doses of cytotoxic drugs cisplatin (**B**) lomustine (**C**) or bleomycin (**D**) for an additional 96 h. Cell viability was determined using MTT assay. (**E**) U138 cells were pretreated with LY500307 for 72 h followed by treatment with varying doses of temozolomide for additional 5 days. Cell viability was determined using an MTT assay.](srep24185-f5){#f5} ![LY500307 reduces GBM progression and induces apoptosis *in vivo*.\ (**A**) U251 GBM-luc cells were implanted intracranially into the right striatum of nude mice. The mice were then randomized to either the control or the treatment group and received either vehicle or LY500307 (5 mg/kg body weight/day), respectively, for 28 days. Tumor growth in terms of luciferase intensity was measured using Xenogen IVIS imaging (n = 5). (**B**) Mouse brains collected from both the control and LY500307-treated mice were fixed in formalin and subjected to immunohistochemical staining for Ki-67. For quantitation, Ki-67-positive cells from five different fields were counted and plotted as histogram. (**C**) Brain sections were subjected to a TUNEL assay, and the number of TUNEL-positive cells was counted in five different fields and plotted as histogram. DAPI was used to visualize the nuclei. Data are represented as mean ± SE. \*p \< 0.05, \*\*p \< 0.01, \*\*\*p \< 0.001.](srep24185-f6){#f6} ![LY500307 prolongs survival of tumor-bearing mice in a syngeneic glioma model.\ (**A**) GL26 glioma cells were implanted orthotopically into C57BL6 mice that were treated with either vehicle or LY500307 (5 mg/kg body weight/day) orally. Survival of the mice was plotted using Kaplan-Meier curve (n = 4). (**B**) Mouse brains collected from the control and LY500307-treated mice were fixed in formalin and processed for immunohistochemical staining for Ki-67. The number of Ki-67--positive cells from five different fields were counted and plotted as histogram. (**C**) Brain sections were subjected to a TUNEL assay, and number of TUNEL-positive cells was counted for five different fields and plotted as histogram. DAPI was used to visualize the nuclei. Data are represented as mean ± SE. \*\*p \< 0.01, \*\*\*p \< 0.001.](srep24185-f7){#f7}

{ "pile_set_name": "PubMed Central" }

Introduction {#S1} ============ All organisms depend on ring- and spiral-shaped ATPase assemblies to carry out essential processes ranging from proteolysis and membrane trafficking, to signaling events and nucleic acid transactions. DNA replication onset in cells reflects one such process, employing ATP-dependent initiation factors to coordinate replisome assembly^[@R1],\ [@R2]^. Replication initiators of eukaryotes and prokaryotes contain AAA+-family ATPase domains, whose activity is augmented by duplex DNA-binding domains (DBDs) and specialized protein-protein interaction elements that assist with origin recognition and recruit specific replication factors^[@R3],\ [@R4]^. AAA+ enzymes share a common structural core with RecA-type ATPases, together forming the [**A**]{.ul}dditional [**S**]{.ul}trand [**C**]{.ul}atalytic glutamat[**E**]{.ul} (**ASCE**) supergroup of P-loop NTPases^[@R5]^; the molecular logic that allows a common nucleotidyl-hydrolase module to control the disparate activities of replication initiators, and ASCE proteins in general, is not understood. In bacteria, replication initiation relies on the DnaA protein^[@R6]--[@R8]^. In *Escherichia coli*, multiple DnaA molecules bind to the replication origin, *oriC*, through several duplex DNA-binding sites, forming a large nucleoprotein complex in the presence of ATP^[@R9]--[@R11]^. With the aid of of appropriate architectural proteins (such as Integration Host Factor) and negatively-supercoiled DNA, this complex subsequently melts an AT-rich, DNA-unwinding element (DUE) located adjacent to the duplex DnaA binding sites^[@R12],\ [@R13]^. ATP also activates a secondary DNA-binding site within DnaA, postulated to reside within the AAA+ domain, which engages single-stranded regions of the DUE to form a stable open complex^[@R12],\ [@R14]--[@R16]^. DnaA then collaborates with the bacterial helicase loader (DnaC in *E. coli*), to recruit two hexamers of the DnaB helicase to the origin and promote replisome assembly^[@R17]--[@R19]^. Although most AAA+ enzymes form closed-ring assemblies^[@R20],\ [@R21]^, structural studies have indicated that initiators and polymerase clamp-loaders form openring structures^[@R14],\ [@R22]--[@R24]^. Among initiator/loader systems, DnaA is particularly unusual in that it has been seen to oligomerize into a right-handed, spiral filament^[@R14]^. Two models have been proposed to explain how this structure might aid origin melting ([Fig S1](#SD1){ref-type="supplementary-material"}). In one, the wrapping of duplex DNA about a DnaA super-helix would constrain a positive supercoil, generating compensatory negative writhe that could aid opening of the neighboring DUE. In the other, the wrapped DnaA/DNA complex would serve as a nucleation center, allowing DnaA protomers to directly engage and melt the DUE, possibly through the initiator\'s ATPase elements. Thus far, experimental evidence has supported both models^[@R9],\ [@R14]--[@R9],\ [@R25]^, leaving open the question as to how DnaA catalyzes origin melting. The relationship of this mechanism to other initiation systems, or to AAA+/ASCE proteins overall, is also unclear. A DnaA-ssDNA crystal structure {#S2} ============================== To examine these issues, we set out to determine the structure of DnaA bound to single-stranded DNA. Employing a truncation of *Aquifex aeolicus* DnaA consisting of the AAA+ and duplex-DNA-binding domains (which, like its *E. coli* counterpart^[@R16],\ [@R17]^, is active for both ATP-stimulated assembly and single-stranded-DNA (ssDNA) binding^[@R25]^), we first grew DNA-free crystals in the presence of Mg^2+^ and the non-hydrolyzable ATP mimic AMPPCP^[@R14]^. DNA substrates were then soaked into these crystals under low-salt conditions (**Methods**). Data collection and phasing by molecular replacement revealed four DnaA protomers per asymmetric unit, arranged in a spiral configuration that propagates into a continuous protein helix by the action of crystal-symmetry elements ([Fig 1a, 1b](#F1){ref-type="fig"}), along with bound single-stranded DNA. Of the multiple substrates screened (**Methods**), dA~12~ yielded the highest-quality density ([Fig S2a, S2b](#SD1){ref-type="supplementary-material"}), and served as the best target for model building and refinement. The final structure, containing a DnaA:AMPPCP:Mg^2+^:dA~12~ stoichiometry of 4:4:4:1, was refined to an R~work~/R~free~ of 24.9/26.8% at 3.35 Å resolution ([Table S1](#SD1){ref-type="supplementary-material"}). DnaA-ssDNA interactions {#S3} ======================= The overall arrangement of DnaA subunits in the helical assembly is highly similar to a DNA-free form reported previously (0.7 Å r.m.s.d. between all Cα positions)^[@R14]^. AMPPCP•Mg^2+^ binds at the interface between neighboring subunits, with the γ-phosphate of AMPPCP coordinated by catalytic amino acids from pairs of adjoining AAA+ domains. Single-stranded DNA associates exclusively with the AAA+ elements of the initiator, with each protomer binding three nucleotides of the dA~12~ strand ([Fig 1a](#F1){ref-type="fig"}). Almost all contacts are made through the phosphodiester backbone, exposing the DNA bases to solvent. Each trinucleotide segment adopts a B-form DNA conformation ([Fig S3](#SD1){ref-type="supplementary-material"}) with the bases between consecutive segments separated by large (\~10 Å) gaps, extending the substrate by \~50% ([Table S2; Supplemental Material](#SD1){ref-type="supplementary-material"}). DnaA binds single-stranded DNA using just two pairs of helices, α3/α4 and α5/α6, both of which line the central channel of the protein assembly ([Fig 1c](#F1){ref-type="fig"}). The geometry of these two elements creates a single conduit along the length of the DnaA superhelix that allows substrate to traverse consecutive DnaA protomers. Interestingly, helices α3/α4 also comprise the [**I**]{.ul}nitiator [**S**]{.ul}pecific [**M**]{.ul}otif (**ISM**), which both promotes filament formation^[@R14],\ [@R19]^ and distinguishes DnaA as a member of initiator clade of the AAA+ superfamily^[@R26],\ [@R27]^. DnaA uses a simple network of interactions to coordinate ssDNA. The ISM forms a shelf for each trinucleotide, in which a conserved hydrophobic residue, Val156, forms van der Waal contacts with the sugar and base of the first nucleotide in the triplet ([Fig 1d](#F1){ref-type="fig"}). The central phosphate of each trinucleotide is bound by the electropositive, N-terminal helix dipole of α6 and hydrogen bonded by Thr191 ([Fig 1c, 1d](#F1){ref-type="fig"}). These contacts are flanked by two positively charged residues, Arg190 and Lys188, which make salt-bridge interactions with the phosphates of nucleotides 1 and 3, respectively. Importantly, mutant initiators containing substitutions in these observed DNA-binding residues show reduced affinity for ssDNA in solution ([Fig S4](#SD1){ref-type="supplementary-material"}) confirming that the crystals captured a physiologically-meaningful initiator state. Moreover, mutations of the same positions in *E. coli* DnaA (amino acids Arg245, Lys243 and Val211) also disrupt ssDNA binding and origin melting^[@R15]^. Thus, the single-stranded DNA engagement strategy seen here appears conserved across bacterial species. Structural similarities between DnaA and RecA {#S4} ============================================= In considering the assembly patterns of oligomeric ATPases, we were struck by the similarity of DnaA to one system in particular: the homologous recombination protein, RecA. Although the cellular functions of these two proteins are fundamentally different (catalysis of DNA strand-exchange reactions *versus* replication origin melting and coordination of replisome assembly), both RecA and DnaA are predicated upon an ASCE ATPase fold^[@R27],\ [@R28]^. Like DnaA, RecA (and its Rad51/RadA orthologs) forms a helical assembly that engages DNA with its pore regions^[@R28]--[@R32]^. These shared physical properties led us to undertake a more detailed comparison of RecA and DnaA. Of the multiple models available, the structure of a RecA oligomer bound to single-stranded DNA^[@R33]^, representing the presynaptic complex formed during the initial stages of homologous recombination, is globally most similar to the DnaA state we observe ([Fig 2a, 2b](#F2){ref-type="fig"}). As with DnaA, RecA contacts DNA almost exclusively through the phosphodiester backbone, which sits in the interior of a positively-charged filament pore. Each RecA protomer binds three nucleotides in a B-DNA conformation, with the base stacking between each triplet interrupted such that single-stranded DNA is extended \~1.5 fold compared to a B-form duplex ([Fig 2c](#F2){ref-type="fig"}). RecA and DnaA also exhibit some interesting and significant differences. A visual examination of each triplet shows that RecA uses a more extensive network of contacts for engaging single-stranded DNA than does DnaA ([Fig 2d, 2e](#F2){ref-type="fig"}), burying twice as much surface area per triplet (318 Å^2^ and 639 Å^2^ for DnaA and RecA, respectively). This difference derives largely from an additional β-hairpin in RecA that fills the gap between each triplet and reinforces each three-base stack^[@R33]^. Moreover, while two of the three nucleotides within each RecA triplet (positions 1 and 2) align well with those seen in DnaA, position 3 of the DnaA trinucleotide rotates away from the pore axis by \~50° ([Fig 2f](#F2){ref-type="fig"}). This difference skews consecutive DnaA triplets away from one another, disrupting the formation of a smoothly spiral arrangement as seen in RecA ([Fig 2c](#F2){ref-type="fig"}). DNA extension is ATP- and assembly-dependent {#S5} ============================================ The ability of RecA to stretch DNA to the extent observed crystallographically has been amply substantiated by a wealth of methodologies^[@R34]--[@R36]^. Using these efforts as a guide, we set out to determine whether the DNA conformation we observe bound to DnaA accurately represents the state of the substrate in solution. To accomplish this, we employed a bulk-phase Fluorescence Resonance Energy Transfer (FRET)-based ssDNA extension assay analogous to single-molecule approaches applied to RecA^[@R37]^. Using a poly-thymine DNA labeled with Cy3 and Cy5 (FR-dT~21~) ([Table S3](#SD1){ref-type="supplementary-material"}), we monitored changes in the length of single-stranded DNA resulting from DnaA binding ([Fig 3a](#F3){ref-type="fig"}). Analogous studies were performed with RecA as a control. As both RecA and DnaA require ATP for formation of the oligomers observed in the structural models, we expected ATP-dependent extension to lead to a loss of FRET signal. We tested for extension both in the presence of ADP•BeF~3~, to avoid complications that might arise from nucleotide hydrolysis, and in the presence of ADP, which is known to promote DnaA disassembly. Pronounced extension was observed only in the presence of the ATP analog ([Fig 3b and 3c](#F3){ref-type="fig"}), and not in the presence of ADP. The lengths of single-stranded DNA in the ATP-assembled states of both proteins, as calculated from the FRET data, were in close agreement with those observed in the crystal structures ([Table S6](#SD1){ref-type="supplementary-material"}). Likewise, mutations in ssDNA-binding amino acids and residues required for DnaA assembly all significantly reduced ssDNA extension ([Fig S6](#SD1){ref-type="supplementary-material"}), demonstrating that this activity depends on substrate binding to the pore of an initiator oligomer that forms only when activated by ATP. DnaA directly catalyzes duplex melting {#S6} ====================================== How replication origins are opened for replisome assembly is an important, unanswered question. Given the similarities between the ssDNA binding and extension activities of DnaA and RecA, we reasoned that the initiator might directly destabilize and disrupt DNA duplexes. This activity is a known property of RecA^[@R38]^, albeit one that permits the recombinase to actively exchange DNA strands between target substrates^[@R30],\ [@R33]^. To test this idea, we developed a DNA strand-displacement assay for DnaA. First, the initiator was incubated with a short duplex containing one fluorescently-labeled strand. Unlabeled competitor strand was then added to capture any unwound species ([Fig 4a](#F4){ref-type="fig"}). Both ADP and ADP•BeF~3~ were tested to determine whether initiator assembly affected the outcome of the experiment, as were DNAs of different lengths and stabilities. Analysis of the resultant products by gel electrophoresis shows that DnaA readily unwinds a 15mer duplex DNA of moderate stability (T~m~=43°C) in the presence of the ATP mimic ([Fig 4b](#F4){ref-type="fig"}). By contrast, increasing the stability of the DNA substrate by \~30% (using a 20mer, T~m~=55°C) weakens the unwinding activity of DnaA ([Fig 4b](#F4){ref-type="fig"}), whereas increasing DNA stability even further (30mer, T~m~=62°C) abrogates melting completely ([Fig S7a](#SD1){ref-type="supplementary-material"}). Importantly, ADP did not support strand displacement, nor did ssDNA binding and DnaA assembly mutants ([Fig S7b, S7c](#SD1){ref-type="supplementary-material"}). These controls indicate that dsDNA melting is dependent not only upon formation of an assembled DnaA oligomer, but that the initiator is fine-tuned to specifically disrupt DNAs of modest stability. One significant functional difference between RecA and DnaA is that the recombination protein can drive a true strand-exchange reaction; that is, in addition to displacing one strand of a duplex, RecA can also pair homologous single-stranded DNA segments into a double-stranded molecule. By contrast, DnaA\'s function is to separate double-stranded origin regions. Inspection of the RecA and DnaA complexes reveals a physical basis for these differing properties: in DnaA, successive trinucleotide elements are arranged in a state incompatible with the formation of a continuous duplex, whereas ssDNA bound to RecA adopts a smoothly spiraled arrangement permitting the contiguous pairing of a complementary strand ([Fig 4c](#F4){ref-type="fig"}). This distinction arises primarily from the 50° rotation between the nucleotides at the third position of each triplet seen in the RecA and DnaA models ([Fig 2f](#F2){ref-type="fig"}). In DnaA, the orientation of this nucleotide appears to be stabilized by base stacking, whereas in RecA the β-hairpin insertion helps sculpt the configuration of the DNA to create a contiguous base-pairing surface. Implications for origin melting {#S7} =============================== Together, our findings present the strongest evidence yet that DnaA melts replication origins by directly assisting with the separation and sequestration of duplex DNA strands ([Fig S1c](#SD1){ref-type="supplementary-material"}). Notably, this activity does not contradict the demonstrated need for other factors capable of reshaping and/or destabilizing DNA (e.g., IHF and negative supercoiling) during initiation^[@R12],\ [@R13]^. Rather, these elements likely help promote DnaA assembly and prime the origin for melting by what otherwise would be an inefficient unwindase. In this view, the AAA+ domains of DnaA may first engage only one of the two strands of duplex DNA with their ssDNA binding elements (possibly at reported ssDNA or ATP-DnaA binding sites^[@R15],\ [@R16]^). In the presence of ATP, which triggers initiator assembly, subunit/subunit interactions would help restructure the DNA backbone, stretching the contacted strand to facilitate melting. Reannealing would be disfavored by the non-contiguous arrangement of base triplets in the extended state ([Fig 2d](#F2){ref-type="fig"}). Future studies will be needed to define the specific order and effect of these events further. We envision that the propensity of DnaA to open DNA could be adjusted in other bacterial species by strengthening or weakening the association of its ATPase domains with DNA and/or each other. An attractive feature of such a mechanism is that it is amenable to additional layers of control by changes to DUE sequence, superhelical density, and co-resident architectural factors to ensure that a replication origin fires only when DnaA is both present and assembled properly. Such flexibility may have played a role in allowing DnaA to persist as the primary initiator in bacteria that have adapted to markedly different environmental niches. Thematic patterns of substrate recognition in ASCE ATPases {#S8} ========================================================== The mechanism by which DnaA coordinates ssDNA also comports well with findings in other replication initiation systems and with ASCE ATPases in general. For instance, many oligomeric RecA and AAA+ enzymes bind substrate in the interior pore of a closed- or cracked-ring particle^[@R33],\ [@R39]--[@R41]^. DnaA follows this pattern. A comparison of DnaA to other, disparate nucleic acid-dependent AAA+ systems -- e.g., polymerase clamp loaders and processive helicases -- further shows that these factors also associate with client substrates in a remarkably analogous manner, using the same face of the core αβα ATP-binding fold to engage a short backbone stretch of their target DNAs ([Fig 5](#F5){ref-type="fig"}). For AAA+ proteins involved in initiation, these similar contact mechanisms have been differentially co-opted to assist with specific protein functions, ranging from the control of origin recognition (as seen in archaeal Orc1 proteins^[@R26],\ [@R42]^), to mediating processive DNA unwinding (viral superfamily 3 helicases^[@R40],\ [@R43]^). DnaA, with its ability to melt (but not translocate along) DNA, appears to employ an intriguing mix of some of the activities exhibited by related initiation systems. Future efforts will be needed to determine how subtle differences in the position and nature of substrate-binding surfaces, combined with specific alterations in the assembly patterns of central AAA+ domains, endow such molecular motors and switches with their distinct biochemical properties. Methods Summary {#S9} =============== Detailed information regarding experimental methods, substrate sequences, binding constants, and FRET efficiencies and distances can be found in the [Online Methods](#SD2){ref-type="supplementary-material"} section and in the [Supplementary Material](#SD1){ref-type="supplementary-material"}. Supplementary Material {#SM} ====================== We would like to thank Karl Drlica, James Keck, Thomas Murray and the Berger laboratory for helpful comments, and Michael M. Cox for his generous contribution of RecA protein. This work was supported by the NIGMS (GM071747) and the National Institute of Health Molecular Biophysics Training Grant T32 GM008295. **Author Information** Coordinates have been deposited in the RSCB PDB under the accession number 3R8F. **Author Contributions** K.E.D. and J.M.B. designed the experiments, analyzed the data and wrote the paper. Protein purification, crystallization and ssDNA binding assays were performed by K. C. and K.E.D. K.E.D. performed the other experiments. ![The ATPase pore of assembled DnaA binds ssDNA\ **a,** Side view of the asymmetric unit, with DnaA subunits differentially colored. Single-stranded DNA is displayed as red sticks. AMPPCP and Mg^2+^, bound to chain A, are shown as spheres colored by element and in magenta, respectively; AMPPCP•Mg^2+^ bound to chains B-D are occluded in this view. **b,** Side and top views of oligomerized DnaA, reconstructed through crystal packing, showing twelve DnaA subunits and three strands of ssDNA. Coloring as per panel A. **c,** Side view of the DnaA tetramer with helices α3/α4 and α5/α6 highlighted in orange and yellow, respectively ("ISM" -- initiator specific motif). Single-stranded DNA is shown as a transparent stick-and-surface representation colored by element; phosphates are further highlighted as red spheres. **d,** Protein-DNA contacts. Protein chains B (left) and C (right) are displayed with the same coloring as in c. Single-stranded DNA is colored by element.](nihms-318339-f0001){#F1} ![DnaA engages ssDNA in a manner similar to RecA\ **a,** View of a DnaA-AMPPCP-ssDNA pentamer (consisting of one full tetramer, as well as chain A (DnaA~A\'~) and its associated triplet from the adjacent asymmetric unit). AMPPCP•Mg^2+^ is shown as spheres colored by atom, ssDNA as red sticks. **b,** View of a RecA-ADP-AlF~4~-ssDNA pentamer (PDB ID 3CMW)^[@R33]^. ADP•AlF~4~•Mg^2+^ is shown as spheres colored by atom, ssDNA as red sticks. **c,** Comparison of ssDNA bound to DnaA (orange), RecA (green) and a strand of B-DNA (yellow). **d,** Close-up view of triplet bound to DnaA (chain C) with magenta dashed lines indicating key contacts. **e,** Close-up view of triplet bound to RecA (protomer 2) with magenta dashed lines indicating key contacts. **f,** Side (left) and top (right) views of the triplets displayed in **d** and **e** aligned with each other.](nihms-318339-f0002){#F2} ![DnaA extends ssDNA in solution\ **a,** Cartoon of ssDNA extension assay. **b,** Emission scan (donor excitation) of FR-dT~21~ in the presence of 10 μM DnaA with either ADP•BeF~3~ (*top*) or ADP (*bottom*). **c,** Emission scan (donor excitation) of FR-dT~21~ in the presence of 10 μM RecA with either ATPγS (*top*) or ADP (*bottom*). Reported transfer efficiencies and distances were calculated using donor emission as described in the **Methods**.](nihms-318339-f0003){#F3} ![DnaA directly melts duplex DNA\ **a,** Schematic of strand displacement assay. The green circle represents the Cy3 fluorescent end-label used to follow the status of one DNA strand. Complementary strands of duplex substrates are colored grey and black. **b,** Strand displacement assay conducted with 15 and 20mer duplex substrates (C3--15mer and C3--20mer) in the presence and absence of different nucleotides. DnaA concentrations used are indicated above each lane. **c,** (left) Cartoon model showing how complementary base triplets (yellow) would pair (in a B-DNA manner) with ssDNA bound to DnaA (red). The orientation of successive DnaA-bound triplets is such that it prevents the formation of a continuous base-paired strand favoring duplex separation. (right) Same DNA view, but as seen in RecA, where triplets are oriented to allow pairing of an extended complementary strand to promote duplex formation and strand exchange (PDB ID 3CMX)^[@R33]^.](nihms-318339-f0004){#F4} ![Common DNA recognition strategies of AAA+ proteins\ Structures of DNA-bound assemblies (top) and individual domains (bottom) for AAA+ proteins involved in replication. All recognize DNA using the same face of the AAA+ fold (violet) (bottom). **a,** Bacterial clamp-loader (γδδ′) complex (AAA+ domains -- differentially colored) bound to primer-template DNA (PDB ID 3GLF)^[@R41]^. **b,** Archaeal initiators Orc1--1 (gray) and Orc1--3 (AAA+ domain - green) bound to origin DNA (PDB ID 2QBY)^[@R26]^. **c,** Bacterial initiator DnaA (AAA+ domains -- gray/blue) bound to ssDNA. **d,** Viral initiator/helicase E1 (AAA+ domains -- orange/gray) bound to ssDNA (PDB ID 2GXA)^[@R40]^. For all panels, DNA is shown as either red spheres (*top*), or as a red/grey cartoon (*bottom*).\ Nucleotide co-factors bound to AAA+ domains (*bottom*) are represented as spheres colored by atom.](nihms-318339-f0005){#F5}

{ "pile_set_name": "PubMed Central" }

Introduction {#Sec1} ============ Innate immunity is a ubiquitous system that widely protects organisms from infectious pathogens as a front-line host defense mechanism. The immune response is triggered by the recognition of broadly conserved microbial components, known as pathogen-associated molecular patterns, by germline-encoded pattern recognition receptors of the host cells^[@CR1]^. As an early defense system against RNA viruses in mammals, the innate immune response is precisely controlled by two distinct signal transduction pathways mediated by the pattern recognition receptors Toll-like receptor 3 (TLR-3) and retinoic acid-inducible gene I (RIG-I)-like receptors (RLR) that respond to virus-derived RNAs^[@CR2],\ [@CR3]^ (i.e., pathogen-associated molecular patterns). Although the two pathways differ with respect to the initial activation of their downstream effectors, they converge at the point of activation of the transcriptional factors interferon regulatory factor 3 (IRF-3) and nuclear factor κB (NF-κB), which results in the rapid production of type I interferons (IFN-α and -β) and other proinflammatory cytokines to establish adaptive antiviral immunity^[@CR4]^. Mitochondria, eukaryotic cell powerhouses, are crucially involved in numerous cellular processes, including apoptosis^[@CR5]^ and calcium homeostasis^[@CR6]^. Mitochondria also have a unique role in innate immunity against RNA viruses^[@CR7]^. Mitochondrial-mediated antiviral immunity depends on activation of the RLR signaling pathway, and mitochondrial antiviral signaling (MAVS), a downstream adaptor of RLR at the mitochondrial outer membrane (MOM), has a key role in the signal transduction^[@CR8],\ [@CR9]^. Upon viral infection, MAVS recruits various types of effectors at the MOM, and the orchestrated "MAVS signalosome", including the mitochondrial membrane potential (Δψ~m~), is the primary unit governing antiviral innate immunity^[@CR10],\ [@CR11]^. Although the role of the MAVS signalosome in mitochondria with its dynamic morphologic properties^[@CR12]^ to provide a molecular platform that facilitates signal transduction is well characterized, insight into how the organelle functions to facilitate antiviral immunity through the activity of oxidative phosphorylation (OXPHOS) has remained unclear. Results and Discussion {#Sec2} ====================== Cultured cells rely on mitochondrial respiratory activity {#Sec3} --------------------------------------------------------- To evaluate the functional coordination of mitochondrial-mediated antiviral immunity and OXPHOS activity, we first sought to determine optimal cell culture conditions in which the cellular bioenergetics would rely on mitochondrial respiratory activity. We used a fluorescence resonance energy transfer (FRET)-based assay to visualize metabolized intracellular adenosine 5′-triphosphate (ATP) at the single-cell level in human embryonic kidney 293 (HEK293) cells expressing an ATP probe, ATeam1.03^[@CR13]^. The biosensor assay performed with cells cultured under our customized medium containing galactose (10 mM) as the carbon source revealed high FRET signal \[based on an emission ratio of 527/475 nm (denoted YFP/CFP)\] in individual cells \[Fig. [1A](#Fig1){ref-type="fig"}, galactose panels, (−)\], indicating that cells maintained adequate intracellular ATP levels. The intracellular ATP level, however, which affects the YFP/CFP ratio, was dramatically decreased (\~2.5-fold) by the addition of electron transport chain (ETC) inhibitors (rotenone and antimycin A), an ATP synthase inhibitor (oligomycin), or a protonophore \[carbonyl cyanide *m*-chlorophenylhydrazone (CCCP)\] to the media \[Fig. [1A](#Fig1){ref-type="fig"}, galactose panels, (+), and B\], demonstrating that energy production in the cells critically depends on mitochondrial OXPHOS. In contrast, cells maintained in a customized medium containing glucose (10 mM) exhibited no significant decrease in the intracellular ATP level, even after treatment with the pharmacologic drugs (Fig. [1A,B](#Fig1){ref-type="fig"}). These results clearly demonstrated that the cellular bioenergetics are less dependent on OXPHOS activity in the presence of glucose, and that cellular ATP production predominantly relies on glycolysis.Figure 1Bioenergetic profiling of cells under oxidative conditions. (**A**) HEK293 cells stably expressing the ATeam1.03 were cultured in customized media containing either galactose (right) or glucose (left) at 37 °C, and then the indicated mitochondrial inhibitors were added to each medium. The time-course of the fluorescence emission ratio (YFP/CFP) was monitored to visualize cytosolic ATP levels in each living cell, and ratiometric pseudocolor images of cells pre- (−) or post-treated (+) with inhibitors are shown. The images were processed in MetaMorph (Molecular Devices), and the blue color indicates less cytoplasmic ATP. Scale bar, 20 μm. (**B**) Quantification of the YFP/CFP ratio calculated from images in (**A**). Number of cells (*n*) used for the quantification is shown at the top of the graph. Error bars indicate SD (Unpaired *t*-test; \*\*\**P* \< 0.001). Validation of the RLR signaling pathway under oxidative conditions {#Sec4} ------------------------------------------------------------------ We examined RLR-mediated signal transduction under the galactose-containing oxidative condition. Under oxidative conditions, similar to the glucose condition, ectopic expression of MAVS in HEK293 cells potently activated both IFN-β and NF-κB luciferase reporters in a dose-dependent manner (Supplementary Fig. [S1A](#MOESM1){ref-type="media"}). The OXPHOS-dependent RLR-mediated signal transduction was also confirmed by stimulating cells transfected with either a plasmid encoding RIG-I(1--250) (Supplementary Fig. [S1B](#MOESM1){ref-type="media"}) or a synthetic analog of viral double-stranded RNA \[dsRNA; poly(I:C)\] (Supplementary Fig. [S1C and D](#MOESM1){ref-type="media"}), both of which are upstream factors of MAVS^[@CR3]^. The kinetic response of IRF-3 phosphorylation (a hallmark of IRF-3 activation) in cells against infection by Sendai virus (SeV), a negative-stranded RNA virus of the *Paramyxoviridae* family, in the OXPHOS-dependent condition was also similar to that in the glucose condition (Fig. [2A](#Fig2){ref-type="fig"}). The RIG-I-mediated activation of the IFN-β reporter in the OXPHOS-dependent condition was sufficiently impaired by co-expression of a hepatitis C virus serine protease NS3/4A (an inhibitor of the RLR signaling pathway)^[@CR14],\ [@CR15]^ \[Fig. [2B](#Fig2){ref-type="fig"}, wild-type (WT)\], whereas its inactive mutant (S139A) had no functional effect, indicating that the observed signal transduction occurred via RLR-dependent signaling pathway.Figure 2RLR-mediated signal transduction under oxidative conditions. (**A**) The kinetic profile of IRF-3 activation in oxidative or glycolytic medium-cultured HEK293 cells that were challenged with SeV (4 HA units/mL). Each cell lysate was collected at the indicated time-points (3, 6, 9 and 12 h) and analyzed by Western blotting with antibodies against the specific antibody (pIRF-3; phosphorylation of Ser386). Anti-β-actin was used as the loading control and anti-SeV as the infection control. U.I., uninfected. (**B**) Oxidative medium-cultured HEK293 cells were transfected with 50 ng of empty vector (Mock) or expression plasmid for Myc-tagged RIG-I(1--250) (control) together with the IFN-β reporter plasmid. The two right lanes (+NS3/4A) indicate that 100 ng of FLAG-tagged WT or inactive (S139A) NS3/4A serine protease expression plasmids were also co-transfected with the RIG-I(1--250) plasmid. The immunoblot on the top represents an expression profile of Myc-tagged RIG-I(1--250) and FLAG-tagged NS3/4A mutants as well as the loading control of endogenous β-actin. Error bars indicate SD (*n* = 3; Unpaired *t*-test; \*\**P* \< 0.01 and \*\*\**P* \< 0.001, respectively). (**C**) BRET saturation assay of MAVS oligomerization in glycolytic versus oxidative media. HEK293 cells were co-transfected with 5 ng NLuc-MAVS expression plasmid and increasing amounts (0--200 ng) of Venus-tagged MAVS plasmid along with 200 ng of WT (circles) or S139A (squares) FLAG-tagged NS3/4 A plasmids, and analyzed 24 h later using a BRET saturation assay. Closed and open symbols represent glycolytic and oxidative conditions of HEK293 cells, respectively, and inset blots show Western blots from the BRET saturation point of each curve by immunoblotting with the indicated antibodies. Error bars indicate SEM (*n* = 3). (**D**) Heat maps of microarray analysis. Total RNAs were isolated from glycolytic and oxidative cultured conditions of primary MEFs that were unchallenged (−) or challenged (+) with SeV (30 HA units mL^−1^) for 6 h, and microarray analysis was performed. The heat map was generated by MeV software^[@CR42]^, and the color indicates the distance from the median of each row. (**E**) Similar to (**A**), except that J774A.1 macrophages were challenged with SeV (2 HA units/mL). U.I., uninfected. We next analyzed the structural features of MAVS activation accompanied by RLR-mediated signal transduction, homotypic oligomerization at the MOM^[@CR11]^, when cells were dependent on the oxidative condition. The interaction between ectopically expressed Venus- and NanoLuc luciferase (NLuc)-tagged MAVS with co-expression of the S139A mutant was successfully monitored, and a bioluminescence resonance energy transfer (BRET)-based assay^[@CR16]^ revealed a hyperbolic saturation curve characteristic of a specific interaction (Fig. [2C](#Fig2){ref-type="fig"}, squares). Importantly, the specificity of the observed MAVS-MAVS interaction was verified by the disappearance of the saturated curve when the WT NS3/4A protease was co-expressed in cells due to MAVS cleavage^[@CR14]^ (Fig. [2C](#Fig2){ref-type="fig"}, circles, also see immunoblot). We further verified the antiviral immune response of primary mouse embryonic fibroblasts (MEFs) against SeV infection in the OXPHOS-dependent condition. Gene expression profiling of MEFs showing its dependence on the oxidative condition (Supplementary Fig. [S2](#MOESM1){ref-type="media"}, upregulation of OXPHOS genes) revealed the expected induction of numerous IFN-stimulated and antiviral-signaling genes in response to the viral infection, similar to that observed in the glucose condition (Fig. [2D](#Fig2){ref-type="fig"}). Consistent with the aforementioned assays, we also observed that the antiviral immune response of macrophages cultured under the oxidative conditions was intensively upregulated upon SeV infection (Fig. [2E](#Fig2){ref-type="fig"}). RLR-mediated antiviral innate immunity requires OXPHOS activity {#Sec5} --------------------------------------------------------------- Because altering the energy metabolism in cells through the activity of OXPHOS could accomplish the antiviral immune response, we next asked whether arresting OXPHOS activity would disrupt RLR-mediated signal transduction. IRF-3 phosphorylation in cells responding to SeV infection was completely suppressed by the addition of ETC inhibitors to the medium under oxidative conditions (Figs [3A](#Fig3){ref-type="fig"} and [S3](#MOESM1){ref-type="media"}). The functionality of the cells treated with the pharmacologic drugs under these conditions, however, was unaffected when the cells were either stimulated with extracellular poly(I:C) (Supplementary Fig. [S4A](#MOESM1){ref-type="media"}) or infected with a double-stranded DNA virus, Herpes simplex virus 1 (HSV-1) (Supplementary Fig. [S4B](#MOESM1){ref-type="media"}), which activates the TLR-3 or cGAS/STING cytosolic DNA sensing pathways, respectively. Influenza A virus (IAV; PR8 strain), an RNA virus of the *Orthomyxoviridae* family, encodes the viral protein PB1-F2 that translocates into mitochondria and induces mitochondrial depolarization^[@CR17]--[@CR20]^. Using the viral strain, we observed severe PB1-F2--induced inhibition of RLR signaling in cells that depended on OXPHOS (Fig. [3B](#Fig3){ref-type="fig"}, WT); the inhibitory effect was attenuated, however, when we used a recombinant virus with genetic ablation of the PB1-F2 gene (ΔF2)^[@CR21]^. These observations suggest that the severity of the mitochondrial dysfunction correlates with the severity of the defect in the RLR-induced antiviral response.Figure 3OXPHOS activity couples with the RLR pathway to execute antiviral signal transduction. (**A**) HEK293 cells cultured in glycolytic or oxidative media were infected with SeV (4 HAU/mL) for 2.5 h, and the infected cells were further incubated with the indicated mitochondrial inhibitors for 2.5 h (total 5 h infection). The activation of endogenous IRF-3 was analyzed by Western blotting with antibodies against the specific antibody (pIRF-3; Ser386). Anti-β-actin was used as the loading control. U.I., uninfected. (**B**) Comparison of the IRF-3 activation between cells infected with recombinant influenza A/PR8 viruses (*WT* versus Δ*F2*, each used 2 HAU/mL) cultured under oxidative conditions. The ΔF2 strain is a mutant strain with genetic removal of the *PB1*-*F2* gene from the viral genome^[@CR21]^. The graph on the right shows the quantification of pIRF-3 bands analyzed by densitometry. Error bars indicate SD (*n* = 3; Unpaired *t*-test; \**P* \< 0.05). (**C**,**D**) The B82 WT cybrids and ρ~0~ cells were infected with (**C**) SeV (4 HAU/mL) for 18 h or (**D**) HSV-1 (1 × 10^5^ PFU) for 24 h, and the cell-free supernatants were analyzed by ELISA to measure the secreted amounts of IFN-β (left panel) and IL-6 (right panel), respectively. Error bars indicate SD (*n* = 3; Unpaired *t*-test; \*\**P* \< 0.01 and \*\*\**P* \< 0.001, respectively). U.I., uninfected. N.D., not detected. (**E**) Similar to (**C**), except that the mtDNA-less J774A.1 and its parental macrophages were infected with SeV (2 HA units/mL). Inset panel: relative mtDNA copy number was confirmed by qPCR. In (**C**--**E**), cells were maintained in ρ~0~ medium. We next used a genetic approach to examine whether mitochondrial DNA (mtDNA), which is essential for mitochondrial OXPHOS activity^[@CR22]^, modulates RLR-mediated signal transduction. Mouse fibrosarcoma B82 cells devoid of mtDNA (ρ~0~ B82)^[@CR23]^ (Supplementary Fig. [S5A](#MOESM1){ref-type="media"}) produced less IFN-β and IL-6 against SeV infection, although the immune response was fully rescued in a cell cybrid harboring WT mtDNA (Fig. [3C](#Fig3){ref-type="fig"}), despite the fact that both cell types had adequate energy due to the use of a high glucose medium (25 mM) with pyruvate and uridine (Supplementary Fig. [S5B](#MOESM1){ref-type="media"}) or the same infection level (Supplementary Fig. [S5C](#MOESM1){ref-type="media"}). Conversely, ρ~0~ B82 cells had substantial antiviral responses against HSV-1 infection (Fig. [3D](#Fig3){ref-type="fig"}), consistent with a previous observation^[@CR24]^. To confirm that these observations were not due to specific characteristics of the cell type we used, we performed the same experiment in HeLa cells lacking mtDNA^[@CR25]^ (ρ~0~ HeLa), and found that ρ~0~ HeLa cells also exhibited defective antiviral innate immune responses against SeV infection (Supplementary Fig. [S5D and E](#MOESM1){ref-type="media"}). In addition, transient depletion of mtDNA in macrophages (mtDNA-less J774A.1) treated with rhodamine 6G exhibited an impaired immune response against SeV infection (Fig. [3E](#Fig3){ref-type="fig"}). Taken together, these findings indicated that RLR-mediated signal transduction requires OXPHOS activity to execute antiviral immune responses. Pathogenic mtDNA mutations cause defects in antiviral innate immunity {#Sec6} --------------------------------------------------------------------- The importance of mtDNA in the regulation of RLR-mediated signal transduction prompted us to explore the phenotypic effects of pathogenic mtDNA mutations on antiviral immunity. We used human osteosarcoma cybrid 143B cells harboring homoplasmic pathogenic mtDNA mutations: the ND1 cybrid^[@CR26]^ (complex I dysfunction), which is associated with Leber hereditary optic neuropathy, carrying a missense mutation at 3460/ND1 (Supplementary Fig. [S6A](#MOESM1){ref-type="media"}); and the COX cybrid^[@CR27]^ (missing complex IV), which is associated with mitochondrial encephalopathy, carrying a stop codon mutation at 6930/COXI (Supplementary Fig. [S6B](#MOESM1){ref-type="media"}), both of which exhibit severe respiratory defects. As described, ρ~0~ cells from 143B cell lines also exhibited an impaired immune response against SeV infection (Fig. [4A,B](#Fig4){ref-type="fig"}). Remarkably, activation of the RLR-mediated antiviral signaling in the mutant cells (ND1 and COX) was severely disrupted (Fig. [4A,B](#Fig4){ref-type="fig"}) despite the indistinguishable MAVS pattern and energy supply (Supplementary Fig. [S7A, B and C](#MOESM1){ref-type="media"}). Consistent with these results, both pathogenic mutants were more susceptible to viral infection (recombinant IAV-expressing GFP)^[@CR28]^ than WT cells (Fig. [4C](#Fig4){ref-type="fig"}), due to defective mitochondrial-mediated immunity. Most importantly, the immunodeficiency observed in the ND1 cybrid, but not in the COX cybrid, was significantly recovered by the addition of succinate, a complex II substrate, to the media, suggesting that OXPHOS activity in ND1 mutant cells was driven by the substrate through complex II, which consequently restored its phenotype (Fig. [4D](#Fig4){ref-type="fig"}). Together, these results indicated that cells with defective OXPHOS activity due to mtDNA mutations have severely impaired RLR-mediated antiviral signaling.Figure 4Defects in mtDNA cause malfunction in antiviral innate immunity. (**A**,**B**) The 143B cybrids and ρ~0~ cells were infected with SeV (5 HAU/mL) for 18 h, and (**A**) activation of both IRF-3 and IκBα was analyzed by Western blotting with antibodies against its specific phosphorylated-detection antibodies (Ser386 for IRF-3 and Ser32/36 for IκBα) or (**B**) the cell-free supernatants were analyzed by ELISA to measure the secreted amounts of IFN-β (top panel) and IL-6 (bottom panel). U.I., uninfected. Error bars indicate SD (*n* = 3; Unpaired *t*-test; \*\*\**P* \< 0.001). (**C**) Fluorescence microscopy of 143B cybrid cells infected with IAV-GFP (0.6 HAU/mL) for 24 h. In (**A**--**C**), cells were maintained in ρ~0~ medium. (**D**) Similar to (**B**), except that succinate (suc), the complex II substrate with ADP, was added to the oxidative medium as indicated. Error bars indicate SD (*n* = 3; Unpaired *t*-test; \*\*\**P* \< 0.001; N.D., not detected; N.S., not significant). Trans-mitochondrial mice (Mito-miceΔ) are susceptible to IAV infection {#Sec7} ---------------------------------------------------------------------- To evaluate whether our *in vitro* observations were physiologically relevant, we investigated the relationship between OXPHOS and antiviral innate immunity *in vivo*. We used 28-week-old trans-mitochondrial mice carrying mtDNA with a large-scale deletion (Mito-miceΔ)^[@CR29]^ and challenged the mice with the IAV PR8 strain. Mito-miceΔ in a resting state had high blood lactate levels (7.2 ± 2 mmol/L) relative to those in WT mice (3.2 ± 0.4 mmol/L), showing lactic acidosis in the mutant mice. Although the mortality of Mito-miceΔ (*n* = 7) infected with IAV was not much worse than that of WT mice (*n* = 9) (Supplementary Fig. [S8A](#MOESM1){ref-type="media"}), the Mito-miceΔ exhibited significant weight loss during the monitoring period (Fig. [5A](#Fig5){ref-type="fig"}; 17% of Mito-miceΔ versus 4% of WT on day 7). Given the higher pathogenicity in Mito-miceΔ following IAV-infection, we examined the physiologic effects of the infection on the lung tissue. In the lungs of WT mice infected with IAV, mild hyperemia and perivascular inflammatory cells infiltration around the bronchus were minimally observed, but few alveoli were injured by the infection (Fig. [5B](#Fig5){ref-type="fig"}, left panels). On the other hand, Mito-miceΔ infected with the virus had significant inflammation in whole lung tissues with desquamation of the bronchial epithelium (Fig. [5B](#Fig5){ref-type="fig"}, right panels, see arrowheads) and alveolar pneumocyte hyperplasia. Notably, virus-induced expression of endogenous *IFN-*β was impaired in the lungs of the Mito-miceΔ (Supplementary Fig. [S8B](#MOESM1){ref-type="media"}), explaining the histopathologic observation of enhanced pathogenicity. We thus concluded that defective OXPHOS enhances the susceptibility to IAV infection *in vivo*, emphasizing the importance of mitochondrial respiration for modulating innate immunity.Figure 5OXPHOS-defective mice are highly susceptible to viral-infection. (**A**) Mito-miceΔ (*n* = 7) or WT mice (*n* = 9) were challenged with IAV (1 × 10^3^ PFU) and mouse weight loss was monitored for 14 days. In the graph, the percentage change from the initial weight of the mice is shown. Error bars indicate SEM \[Dunnett's test; \*\**P* \< 0.01 and \*\*\*\**P* \< 0.0001 (versus day 0), respectively)\]. ^\#^Dunnett's test could not be applied because dead mice appeared from day 9. (**B**) In a separate experiment, the lungs were obtained from each infected mouse on day 8 post-infection, sectioned, and analyzed for histopathology following staining with hematoxylin and eosin. Enlarged boxes are depicted in lower images (400× magnification). Labels **A** and **B** in the images indicate alveoli and bronchus, respectively, and arrowheads indicate desquamation of the bronchial epithelium. Scale bars, 100 μm (top) and 50 μm (bottom), respectively. OPA1-mediated execution of antiviral innate immunity {#Sec8} ---------------------------------------------------- To elucidate the molecular basis of OXPHOS activity coupled with the RLR signaling pathway, we speculated that optic atrophy 1 (OPA1), a mediator of mitochondrial fusion^[@CR30]^, has a role in antiviral innate immunity^[@CR12]^ because of its stabilizing effects on mtDNA^[@CR31]^ and the cristae structure^[@CR32],\ [@CR33]^, both of which contribute to regulating mitochondrial respiration. As reported previously, we observed that *OPA1* ^−/−^ cells contained less mtDNA (Fig. [6A](#Fig6){ref-type="fig"}), completely lacked mitochondrial fusion (Supplementary Fig. [S9A](#MOESM1){ref-type="media"}), and had a disrupted cristae morphology (Supplementary Fig. [S9B](#MOESM1){ref-type="media"}). These phenotypic defects were restored when a *WT OPA1* isoform gene (L-OPA1; variant 1), but not an inactive mutant *K301A* or short isoform (*S-OPA1*), was re-introduced into null cells. Strikingly, although the *OPA1* ^−/−^ cells exhibited severely impaired SeV-induced production of IFN-β and IL-6 (Fig. [6B](#Fig6){ref-type="fig"}), the immune responses were substantially restored by recovery of the mtDNA level in rescued cells \[Figs [6A,C](#Fig6){ref-type="fig"} and [S9C](#MOESM1){ref-type="media"}; *V1(WT)*\]. Given that OPA1 is important for mtDNA stability and the antiviral immune response, we used histochemical staining for cytochrome *c* oxidase (complex IV) activity to directly assess the OPA1 function involved in mitochondrial respiration. As expected, the COX staining pattern of the *V1(WT)*-rescued *OPA1* ^−/−^ cells showed full recovery of the respiration defect, whereas *K301A* and *S-OPA1* failed to rescue the COX activities (Fig. [6D](#Fig6){ref-type="fig"}). Taken together, these results indicate that functional actions of OPA1 are linked to the improvement of OXPHOS activity and the induction of mitochondrial-mediated antiviral innate immunity.Figure 6OPA1 contributes to mitochondrial-mediated antiviral signaling through stabilizing mtDNA. (**A**) Analysis of mtDNA copy number per nuclear DNA in WT and mutant MEFs. Restoration of mtDNA in *OPA1* null cells was measured in mutant cells infected with a retrovirus-expressing variant 1 WT *OPA1* \[*V1(WT)*\], its K301A mutant \[*V1(K301A)*\], and short isoform *OPA1* (*S-OPA1*). Error bars indicate SEM (*n* = 3; Unpaired *t*-test; \*\**P* \< 0.01). (**B**) SeV-induced antiviral innate immune response in *OPA1*-null MEFs. The *WT* and *OPA1*-null MEFs were either uninfected (U.I.) or infected with SeV (4 HAU/mL) for 18 h, and the cell-free supernatants were analyzed by ELISA to measure the secreted amounts of IFN-β (left panel) and IL-6 (right panel). Error bars indicate SD (*n* = 3; Unpaired *t*-test; \*\**P* \< 0.01 and \*\*\**P* \< 0.001, respectively). N.D., not detected. (**C**) Similar to (**B**), except that the immune response in mutant *OPA1* MEFs was monitored. Error bars indicate SD (*n* = 3; Unpaired *t*-test; \*\*\**P* \< 0.001). (**D**) Cytochemical analysis of COX activity. Cells expressing COX activity were indicated by a brown color. Scale bar, 20 μm. The findings of the present study using both *in vitro* and *in vivo* approaches provided new insight into how mitochondria facilitate antiviral immunity through OXPHOS activity. Mitochondria are believed to have evolved from organisms such as α-proteobacterium, and the discovery of their role in host-cell defense was unexpected. It is not surprising, however, that cells gaining respiratory function through a symbion could greatly advance not only through enhanced energy production, but also through safeguarded host defense against other infectious pathogens, especially in vertebrates. Materials and Methods {#Sec9} ===================== Reagents {#Sec10} -------- CCCP, rotenone, oligomycin, and antimycin A were purchased from Sigma-Aldrich (St. Louis, MO). Cytochrome *c* from horse heart was supplied by Nacalai Tesque Inc. (Kyoto, Japan) and 3,3′-diaminobenzidine was purchased from Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan). Furimazine was supplied by Promega (Madison, WI) and poly(I:C) was purchased from InvivoGen (San Diego, CA). Enzyme-linked immunosorbent assay (ELISA) kits for human and mouse IFN-β were supplied by Kamakura Techno-Science Inc. (Kanagawa, Japan) and PBL Assay Science (Piscataway, NJ), respectively. All other reagents were of biochemical research grade. Antibodies {#Sec11} ---------- A polyclonal antibody against IRF-3 (FL-425; 1:2000) and monoclonal antibody against β-actin (C4; 1:4000) were purchased from Santa Cruz Biotechnology (Dallas, TX), and the rabbit monoclonal antibody against phosphorylated IRF-3 (pIRF-3) Ser^386^ (EPR2346; 1:2000) and polyclonal antibody against MAVS (ab25084; 1:1000) were from Abcam (Cambridge, MA). Monoclonal antibodies against Myc (9E10; 1;1000) and FLAG (M2; 1:1000) were obtained from Covance (Princeton, NJ) and Sigma-Aldrich, respectively. Monoclonal antibodies against phosphorylated IκBα (pIκBα) Ser^32/36^ (5A5; 1:1000), phosphorylated NF-κB p65 (pNF-κB p65) Ser^536^ (93H1; 1:2000) and NF-κB p65 (D14E12; 1:3000) were supplied by Cell Signaling Technology (Danvers, MA), and the monoclonal antibody against OPA1 (1:2000) was from BD Biosciences (San Jose, CA). The anti-human and anti-mouse IL-6 antibodies used for ELISA were obtained from eBioscience, and a polyclonal antibody against SeV (PD029; 1:1000) was purchased from MBL (Nagoya, Japan). Monoclonal antibodies against mtCO1 (1D6E1A8; 1:1000) and mitochondrial heat shock protein 70 (mtHsp70; 1:2000) were purchased from Molecular Probes/Invitrogen (Carlsbad, CA) and Affinity BioReagents (Golden, CO), respectively. The Alexa Fluor 488--conjugated polyclonal antibody against rabbit immunoglobulin G (IgG) and Alexa Fluor 568--conjugated polyclonal antibody against rabbit IgG (1:500) were obtained from Molecular Probes-Invitrogen, and the Cy3-conjugated sheep anti-mouse IgG (1:1000) monoclonal antibody was purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). A monoclonal antibody against PB1-F2 (A/PR8 strain; 1:10) was a kind gift from Viktor Wixler (Münster University Hospital Medical School). Plasmids {#Sec12} -------- Plasmids encoding human MAVS variants, hRIG-I(1--250), NS3/4A variants, and the ATeam1.03 were described previously^[@CR13],\ [@CR16],\ [@CR20],\ [@CR34]^. Plasmid encoding rat OPA1 was a gift from Naotada Ishihara (Kurume University, Japan). To generate the retroviral expression constructs, each cDNA was recloned into the retroviral vector pMXs-Puro (Cell Biolabs, San Diego, CA). The retroviral expression vectors were then transfected into the platinum packaging cell lines (Cell Biolabs), and the retroviral supernatant was harvested 48 h post-transfection and used to infect cells. Cell lines and viruses {#Sec13} ---------------------- HEK293, MEF, and J774A.1 cells were maintained in Dulbecco's modified Eagle medium (high glucose; 4,500 mg/L; GIBCO BRL) supplemented with 1% GlutaMAX, penicillin (100 U/mL)-streptomycin (100 μg/mL), and 10% fetal bovine serum at 5% CO~2~ and 37 °C. The ρ~0~ cells (lacking mtDNA) from HeLa, B82 (mouse fibrosarcoma), and 143B (human osteosarcoma) and its WT or mutant mtDNA cybrids were similarly maintained, except with the addition of sodium pyruvate (110 mg/L) and uridine (50 μg/mL; ρ~0~ medium). The mtDNA-less J774A.1 macrophages were pretreated with rhodamine 6 G (0.06 μg/mL in ρ~0~ medium) for 48 h to transiently eliminate endogenous mtDNA. For glucose-free culture experiments, glycolytic (glucose) and oxidative (galactose) media were customized according to previously published protocols^[@CR35]^ with slight modifications; the glycolytic medium was distinct from the high glucose medium used to maintain the cultures. The customized medium was based on glucose and glutamine-free Dulbecco's modified Eagle medium (GIBCO; A14430--01) externally supplemented with 10% dialyzed fetal bovine serum (GIBCO; 26400--044), 2% GlutaMAX, penicillin (100 U/mL)-streptomycin (100 μg/mL), and either 10 mM glucose (glycolytic) or 10 mM galactose (oxidative), respectively. Sendai virus Cantell strain was purchased from the American Type Culture Collection, and the recombinant influenza viruses A/PR8 (ΔF2)^[@CR21]^ and IAV-GFP^[@CR28]^ were a generous gift from Jonathan A. McCullers (St. Jude Children's Research Hospital) and Adolfo García-Sastre (Icahn School of Medicine at Mount Sinai), respectively. Viral infection *in vivo* {#Sec14} ------------------------- Twenty eight-week-old C57BL/6 mice were anesthetized with isoflurane and inoculated intranasally with 50 μL containing 1 × 10^3^ PFU of A/PR8. One group of mice was mock-infected with sterile phosphate buffered saline (PBS) as a control. Mice were observed daily for weight loss and mortality for 14 days. For histopathologic analysis, dissected lung tissues were fixed in 3.5% formaldehyde and paraffin-embedded. Tissues were sliced into 6-μm-thick paraffin sections and stained with hematoxylin and eosin. To determine the level of *IFN-*β mRNA, mice were euthanized 4 days post-infection, and the total RNAs were purified from 30 mg of lung tissue homogenate using an RNeasy Kit (Qiagen). All animal experiments were carried out in a humane manner after receiving approval from the Institutional Animal Care and Use Committee of the University of Tsukuba, and in accordance with the regulation for Animal Experiments in University of Tsukuba and Fundamental Guideline for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions under the jurisdiction of the Ministry of Education, Culture, Sports, Science, and Technology. Luciferase assays {#Sec15} ----------------- HEK293 cells were plated in 24-well plates (2.5 × 10^5^ cells per well) in the customized media. The following day, the cells were co-transfected with 100 ng luciferase reporter plasmid (p125luc or pELAM), 2.5 ng *Renilla* luciferase internal control vector phRL-TK (Promega), and each of the indicated expression plasmids with Lipofectamine 2000 reagent (Invitrogen). Empty vector \[pcDNA3.1(−)\] was used to maintain equivalent amounts of DNA in each well. Cells were harvested 24 h after transfection and analyzed by a dual-luciferase reporter assay on the GloMax 20/20n luminometer (Promega). Each experiment was repeated at least three times. BRET assay {#Sec16} ---------- All BRET saturation assays were performed as previously described^[@CR16]^ with slight modifications. In brief, HEK293 cells (5 × 10^5^ cells per well) cultured under customized media were co-transfected with a constant amount (5 ng) of NLuc-tagged MAVS plasmid and increasing amounts of Venus-tagged MAVS constructs using Lipofectamine 2000 reagent. Empty vector \[pcDNA3.1(−)\] was used to maintain equivalent amounts of DNA in each well. The cells were harvested at 21 h post-transfection and transferred to each well of white 96-well microplates. NLuc substrate (furimazine; 5 μM) was added, and the plates were analyzed via a BRET saturation assay using a Flexstation 3 Microplate Reader (Molecular Devices, Sunnyvale, CA) at 37 °C. ELISA {#Sec17} ----- Measurements of species-specific IFN-β and IL-6 production were performed as described previously^[@CR20],\ [@CR36]^. Microarray analysis {#Sec18} ------------------- Total RNA was isolated from each cultured condition of primary MEFs using TRIzol Reagent (Invitrogen) and purified using SV Total RNA Isolation System (Promega) according to the manufacturer's protocol. RNA samples were quantified using an ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE) and the quality was confirmed with an Experion System (Bio-Rad Laboratories, Hercules, CA). The cRNA was amplified, labeled (Low Input Quick Amp Labeling Kit), and hybridized to the SurePrint G3 Mouse Gene Expression 8 × 60 K microarray according to the manufacturer's instructions (Agilent Technologies, Santa Clara, CA). All hybridized microarray slides were scanned by an Agilent scanner (Agilent Technologies). Relative hybridization intensities and background hybridization values were calculated using Agilent Feature Extraction Software (9.5.1.1). Raw signal intensities and flags for each probe were calculated from the hybridization intensities (gProcessedSignal), and spot information (gIsSaturated, etc.), according to the procedures recommended by Agilent. The raw signal intensities of four samples were normalized by a quantile algorithm with the 'preprocessCore' library package^[@CR37]^ of the Bioconductor software^[@CR38]^. To identify upregulated or downregulated genes, we calculated the Z-scores^[@CR39]^ and ratios (non-log scaled fold-change) from the normalized signal intensities of each probe to compare between control and experiment samples, and established criteria for regulated genes: Z-score ≥2.0 and ratio ≥1.5-fold for upregulated genes, and Z-score ≤−2.0 and ratio ≤0.66 for downregulated genes, respectively. Visualization of cytoplasmic ATP in living cells {#Sec19} ------------------------------------------------ Imaging of cytoplasmic ATP levels in living cells was performed as previously described^[@CR13],\ [@CR40]^. Briefly, cells expressing ATeam1.03 were plated on collagen-coated 35-mm glass-bottom dishes (MatTek, Ashland, MA). The following day, the cells were washed once with the customized media and the cells were subjected to ATP imaging. Observations of cells that were maintained at 5% CO~2~ and 37 °C using a stage-top incubator (Tokai Hit, Shizuoka, Japan) were performed using a Nikon ECLIPSE Ti-E microscope with an oil-immersion type CFI Plan Apo VC 60x lens (Nikon Instruments Inc, Tokyo, Japan). The filters, 438/24-nm BrightLine single-band bandpass filter (FF01--438/24--25), 458-nm edge BrightLine single-edge dichroic mirror (FF458-Di02), and two emission filters (FF01--483/32 and FF01--542/27) used for dual-emission ratio imaging of ATeam1.03 were obtained from Semrock Inc. (Rochester, NY). Fluorescence emission from the ATeam1.03 was imaged by altering the emission filters with a filter changer and a scientific CMOS camera (Zyla 4.2, Andor Technology), and the imaging analysis was performed using MetaMorph (Molecular Devices). The YFP/CFP emission ratio was calculated by dividing the YFP intensity by the CFP intensity of each cell. Confocal microscopy {#Sec20} ------------------- Cells were plated on coverslips in 12-well plates (1.5 × 10^5^ cells per well). The following day, cells were fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.2% Triton X-100 in PBS (pH 7.4), and blocked with 5% bovine calf serum. Endogenous MAVS was detected with the polyclonal primary and AlexaFluor488 secondary antibodies, and mitochondria were stained with anti-mtHsp70 primary antibody followed by the Cy3-conjugated secondary antibody. Cells were imaged with the C2^+^ confocal microscope (Nikon Instruments Inc). Electron microscopy analysis {#Sec21} ---------------------------- For electron microscopy, cells were fixed in 100 mM cacodylate buffer (pH 7.4) containing 2% paraformaldehyde and 2% glutaraldehyde (GA) for 1 h, washed and fixed again with 100 mM cacodylate buffer containing 2% GA overnight at 4 °C. After fixation, the cells were washed four times with 100 mM cacodylate buffer for 20 min each, and postfixed with 2% osmium tetroxide (OsO~4~) in cacodylate buffer at 4 °C for 1 h. Cells were then dehydrated in a graded series of ethanol (50%, 70%, 90%, and 100%), embedded in resin (Quetol 812, Nisshin EM Co., Japan), and polymerized at 60 °C for 48 h. The polymerized resins were cut in ultra-thin sections at 70 nm with an LEICA UTC ultramicrotome, mounted on copper grids, and stained with 2% uranyl acetate and lead solution (Sigma-Aldrich). Images were collected with a transmission electron microscope (JEM-1400Plus; JEOL Ltd, Japan) operating at 80 kV and equipped with a CCD camera (VELETA; Olympus Soft Imaging Solutions GmbH, Germany). Quantification of mtDNA {#Sec22} ----------------------- Total genomic DNA was extracted and purified from cells, and diluted to 5 ng/μL. To quantify the amount of mtDNA per nuclear DNA, quantitative real-time polymerase chain reaction (qPCR) was performed using FastStart Essential DNA Green Master (Roche). Quantification of relative copy number differences was performed based on the difference in the threshold amplification between mtDNA and nuclear DNA \[ΔΔC(t) method\]. We used the following sets of primers for mtDNA: TK702 (5′-cctatcacccttgccatcat, forward) and TK703 (5′-gaggctgttgcttgtgtgac, reverse); and for nuclear DNA: TK704 (5′-atggaaagcctgccatcatg, forward) and TK705 (5′-tccttgttgttcagcatcac, reverse). Cytochemical analysis of cytochrome *c* oxidase (COX) activity {#Sec23} -------------------------------------------------------------- COX staining was performed as previously described^[@CR41]^ with slight modifications. MEFs plated on 18-mm coverslips were fixed with 2% paraformaldehyde in PBS for 15 min and washed twice with PBS for 5 min each. To visualize COX activity, cells were stained with 100 mM phosphate buffer (pH 7.4) containing 0.6 mg/mL 3′,3′-diaminobenzidine, 0.3 mg/mL cytochrome *c*, and 45 mg/mL sucrose at 37 °C for 2 h. Cells were then washed twice with 0.1 M Tris-HCl buffer (pH 8.0) for 5 min each and analyzed by microscopy. Statistical analysis {#Sec24} -------------------- An analysis of variance test (GraphPad QuickCalcs) was used for the statistical analyses. A *P*-value of less than 0.05 was considered statistically significant. Electronic supplementary material ================================= {#Sec25} Supplementary Information **Electronic supplementary material** **Supplementary information** accompanies this paper at doi:10.1038/s41598-017-05808-w **Accession code:** The GEO accession number for the microarray data reported in this paper is GSE75399. **Publisher\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. We are grateful to Thomas Langer (University of Cologne) for helpful discussions regarding the study. We also thank Katsuyoshi Mihara (Kyushu University), Yuichi Michikawa (National Institute of Radiological Sciences, Japan), and Mitsuyasu Kato (University of Tsukuba) for their valuable comments on the study. We appreciate Yuko Fuchigami and Norio Sasaoka for their technical assistance with DNA cloning, sequencing, establishing the cell line expressing ATeam1.03, and technical comments on the figure preparation. We are also grateful for technical support regarding the fluorescence measurements (BRET assay) from the Research Support Center, Graduate School of Medical Sciences, Kyushu University. The 143B COX and ND1 cybrids were a generous gift from Giovanni Manfredi (Cornell University) and Valerio Carelli (University of Bologna, Italy), respectively, and *OPA1*-null MEFs were obtained from David Chan (HHMI and Caltech). This work was supported by the JSPS KAKENHI Grants (No. 15H01265, 17H03667, 26291032, and 26620135 to T.K.), the Daiichi Sankyo Foundation of Life Science, the Japan Foundation for Applied Enzymology, the Takeda Science Foundation, the Naito Foundation, the NOVARTIS Foundation (Japan) for the Promotion of Science, the Life Science Foundation of Japan, and KANEKA Corporation to T.K, and partly supported by a Grant-in-Aid for JSPS Fellow (T.Y.). T.Y. and T.K. performed most of the *in vitro* experiments. H.I. established the biosensor ATeam1.03 and analyzed the cellular ATP level. K.N. and K.I. established the Mito-miceΔ and ρ~0~ cells from cultured human and murine cells. A.K., T.T., and T. Kuroki performed the mouse infection experiments and K.I. analyzed the histopathologic analysis. T.Y., H.I., A.K., T.T., T. Kuroki, K.I., K.N., and T.K. analyzed the data and interpreted all the experimental results. T.K. designed the research program and contributed to writing the paper. Competing Interests {#FPar1} =================== The authors declare that they have no competing interests.

{ "pile_set_name": "PubMed Central" }

Sir, Central venous pressure (CVP) monitoring is a simple, relatively inexpensive method of assessing a patient's circulating blood volume, cardiac status and vasomotor tone. It is essential to be aware of the inherent fallacies and inadequacies of the information derived. Inaccurate measurements are often obtained by the aberrant lodgement of the central venous catheter (CVC) tip. In continuation of the previously published letter to editor concerning misdirected CVC, we describe an unusual case of CVC coiling in the superior vena cava (SVC) leading to falsely high CVP measurement. A 55 year old male patient was brought to the emergency room (ER) with head injury, blunt trauma abdomen with haemodynamic instability. He was further posted for an emergency laparotomy. In view of the clinical condition of the patient and the need to know intravascular volume status, a 7 French triple lumen CVC was inserted in the right internal jugular vein (IJV) in the operating room (OR). All the three ports were checked for free flow of blood and the CVC was fixed at 11cm at skin level. On connecting the transducer to the monitor, ideal waveform was absent. Intra operatively CVP tracing was suboptimal despite the change of transducer, the cable, flushing the unit and repeated zeroing. Post operatively the patient was shifted to intensive care unit on ventilator support for further management. Chest radiograph revealed coiled CVC in the SVC \[[Figure 1](#F0001){ref-type="fig"}\]. Hence it was removed and right subclavian vein was cannulated. ![Coiled CVC in superior vena cava](IJA-54-351-g001){#F0001} The correct placement of the CVC tip is an important factor in obtaining accurate CVP measurements. Malposition of a CVC may occur at the time of insertion or later as a result of spontaneous migration due to anatomic positioning or pressure changes within the thoracic cavity.\[[@CIT1]\] There has been case report on CVC folding back during guide-wire removal inside IJV.\[[@CIT2]\] In our case CVC coiling inside SVC was unusual as it is a large calibre vessel with high flows. There was no anatomical vascular abnormality and no manufacturing defect in CVC. The abutting of the guide wire against the wall of the SVC probably caused the coiling of CVC in the SVC. The J tip of guide wire probably was unknowingly directed cephalad while insertion, which could have caused CVC to further angulate in the upward direction over the guide wire. In conclusion, inaccurate CVP measurements or inability to obtain an ideal wave from tracing are suggestive of an undesirable location of the catheter tip. Awareness of this possibility and careful review of the CVC tip position on X-ray pictures in suspicious cases are important. Inaccurate CVP readings lead to improper assessment of the intravascular status of the patient. Careful clinical co-relation under such circumstances is essential. Roentgenograms after insertion of CVC are essential to eliminate this problem, which is often encountered in clinical practice.

{ "pile_set_name": "PubMed Central" }

Introduction ============ The effectiveness of drug eluting stents for the prevention of restenosis after percutaneous coronary interventions (PCI) relies on the potent effects of cell cycle inhibitors on vascular smooth muscle cell (VSMC) proliferation ([@B2]). Intimal hyperplasia, however, remains a significant clinical problem in for example coronary artery bypass graft failure ([@B17]). In addition to direct effects on VSMCs, intimal hyperplasia is also driven by inflammation, by means of neutrophil, and macrophage infiltration, as well as cytokine and matrix metalloproteinase (MMP) release ([@B17]). In particular, a failure in the resolution of the acute inflammatory response to vascular injury prevents re-endotheliazation and promotes VSMC proliferation and migration ([@B18]). Omega-3 fatty acids decrease inflammation ([@B15]), and inhibit intimal hyperplasia in mice ([@B7]). Other studies have shown similar effects after administration of the pro-resolving lipid mediators resolvins and maresins, which are enzymatically formed from omega-3 fatty acids ([@B5]; [@B1]; [@B8]; [@B19]). The latter studies provided the initial evidence that stimulating the resolution of inflammation by means of lipid pro-resolving mediators would promote an adequate healing of vascular injury. However, the receptor(s) involved in this protective effects *in vivo* remain unknown. We recently established a protective role for the G protein-coupled receptor ChemR23 for the omega-3-derived pro-resolving lipid mediator resolvin E1 in atherosclerosis ([@B6]), but its implications for intimal hyperplasia have remained hitherto unexplored. The aim of the present study was therefore to establish the role of ChemR23 in the downstream signaling of omega-3 fatty acids in intimal hyperplasia, in a pro-inflammatory vascular injury murine model ([@B20]). Methods ======= Carotid Ligation ---------------- The study was approved by the Regional Ethical Review Board in Stockholm. All animals used were male and on a C57BL/6J background. ChemR23^-/-^ mice were obtained from Deltagen. Mice expressing the *Caenorhabditis elegans* *fat-1*transgene were bred as previously described ([@B9]). The two mice strains were then crossbred to generate four groups; ChemR23^+/+^, ChemR23^-/-^, *fat-1*/ChemR23^+/+^, and *fat-1*/ChemR23^-/-^. 10 weeks old littermates (*n* = 6--8/group) were subjected to a complete left carotid ligation as previously described ([@B12]) and indicated in Figure [1A](#F1){ref-type="fig"}. In brief, the four groups were fed in a 10% v/w Omega-6 (Sigma-aldrich, S8281) enriched diet to increase the *fat-1* desaturase substrate. EPA and DHA were quantified by gas chromatography as a control as previously described ([@B6]), and exhibited the expected increase in *fat-1* transgenic mice independently of ChemR23 expression (data not shown). After 7 days, mice were anesthetized with isoflurane/O~2~ (2:1) followed by 0.1 mg *s.c.* injection of buprenorphine for pain relief. Left common carotid artery was exposed, followed by a complete ligation at the bifurcation level with a 7-0 suture. After 28 days from ligation, mice were euthanized by CO2, PFA fixated, and the ligated carotid was collected in PFA, and paraffin embedded. For carotid intimal hyperplasia evaluation, 8 sections of 10 um each every 100 um, were collected. Next, H&E staining was performed to assess neointimal growth area at the site of ligation. Furthermore, sections were stained with antibodies (Supplementary Table [1](#SM1){ref-type="supplementary-material"}) for rabbit anti α-SMA (Abcam), rat anti CD206 (Serotec), rat anti Mac2 (Cedarlane), and rat anti Ly6G (BD) for macrophage and neutrophil content determination, respectively. Negative and isotype controls are shown in Supplementary Figure [2](#SM1){ref-type="supplementary-material"}. Staining in the neointima was assessed using the automated software Leica QWin Standard Y 2.8 (Leica Microsystems) and normalized to neointima area. Rat and rabbit isotype controls were purchased from R&D, and Abcam, and secondary antibodies from Vector. ![ChemR23 deletion promotes intimal hyperplasia under pro-inflammatory conditions. **(A)** Schematic representation of *in vivo* experimental procedure. **(B)** Mouse intima hyperplasia quantification in: ChemR23^+/+^ *n* = 7, ChemR23^-/-^ *n* = 7, *fat-1*^tg^ ChemR23^+/+^ *n* = 6, *fat-1*^tg^ ChemR23^-/-^ *n* = 8; after left common carotid ligation, and representative H&E stained photomicrographs. **(C)** Ly6G^+^ (neutrophil) ChemR23^+/+^ *n* = 5, ChemR23^-/-^ *n* = 7 and Mac-2 (macrophage) ChemR23^+/+^ *n* = 4, ChemR23^-/-^ *n* = 7 immunohistochemistry quantification, and representative photomicrographs. Data represent mean ± SEM. *P*-values derive from **(B)** 2-way ANOVA, **(C)** Student's *t*-test.](fphar-09-01327-g001){#F1} VSMC Isolation and Evaluation of Proliferation ---------------------------------------------- Abdominal aortas from ChemR23^+/+^ and ChemR23^-/-^ mice (*n* = 3/group) were isolated, fat and adventitia removed, and digested in a sterile mixture of 1 mg/mL collagenase type II (Worthington) and 0.3 mg/mL elastase (Sigma, E0127) in DMEM with 10% FBS for 90 min at 37°C and 5% CO~2~. Cell suspension was spun down, resuspended in complete medium (DMEM, 10% fetal calf serum, 100 units/ml penicillin, 100 μg/ml streptomycin, 1 mM sodium pyruvate, 10 mM HEPES, and 2 mM [L]{.smallcaps}-glutamine) and plated. Cells were passaged using trypsin when they reached 80% confluency. Proliferation was assessed by WST-1 reagent (Roche) according to manufacturer's protocol. Peritoneal Macrophages Conditioned Media Generation --------------------------------------------------- Peritoneal macrophages from ChemR23^+/+^ and ChemR23^-/-^ mice (*n* = 4/group) were obtained as previously described ([@B6]). Macrophages were treated for 24 h with complete medium supplemented with LPS (100 ng/ml), washed and followed by a 24 h incubation in complete medium without LPS. After that, cell supernatant was collected and frozen. After thawing, supernatants were diluted 1:10 in complete medium and transferred to ChemR23^+/+^ VSMCs. RNA Extraction and Real-Time PCR -------------------------------- RNA from ChemR23^+/+^ and ChemR23^-/-^ peritoneal macrophages was isolated after 24 h of LPS (100 ng/ml) stimulation using the RNeasy Mini Kit (Qiagen). RNA concentration was quantified by Nanodrop (Thermo Scientific). Relative gene expression was assessed using Taqman assays from Life Technologies: GAPDH as endogenous control (Mm99999915), TNF-α (Mm00443258), IL-6 (Mm00443258), MMP9 (Mm00442991), IL-1β (Mm00434228), and IL-10 (Mm01288386). Statistics ---------- Results are expressed as mean ± S.E.M. Statistical significance was assigned at *p* \< 0.05 as assessed with Student *t-*test when comparing two groups, and with two-way ANOVA as appropriate followed by recommended *post hoc* tests, for multiple comparisons. All analyses were performed using GraphPad Prism 7 (GraphPad Software Inc., CA, United States). Results ======= We here report for the first time that genetic disruption of ChemR23 significantly increased intimal hyperplasia (Figure [1B](#F1){ref-type="fig"}). Furthermore, and as predicted, mice expressing the *Caenorhabditis elegans* *fat-1* transgene, which enables the endogenous production of omega-3 fatty acids, exhibited reduced intimal hyperplasia (Figure [1B](#F1){ref-type="fig"}). Crossbreeding of the two models also allowed us to determine the interaction between the observed effects attributed to omega-3 fatty acids and ChemR23, respectively. Unexpectedly, the *fat-1* transgene was protective in both ChemR23^+/+^ and ChemR23^-/-^ mice (Figure [1B](#F1){ref-type="fig"}). Immunohistochemistry in intimal lesions revealed an infiltration of neutrophil granulocytes (Ly6G) and macrophages (Mac-2), which were not significantly different between ChemR23^+/+^ and ChemR23^-/-^ mice (Figure [1C](#F1){ref-type="fig"}), whereas no T-lymphocyte infiltration (CD3) was detected (data not shown). Analysis of CD206, as a marker of M2 macrophages, revealed no significant differences between ChemR23^+/+^ and ChemR23^-/-^ mice in the percentage of CD206 positive cells in the intima (1,52 ± 0,47, 1.26 ± 0.38, *p* = 0.682), nor in the ratio between M2 and total macrophages (CD206/Mac-2) (31.17 ± 9.66, 40.98 ± 12.30, *p* = 0.538) (Supplementary Figure [1A](#SM1){ref-type="supplementary-material"}). Finally, intimal lesions stained positive for α-smooth muscle actin, but revealed no differences in the percentage of α-smooth muscle actin in the intima between ChemR23^+/+^ and ChemR23^-/-^ mice (Supplementary Figure [1B](#SM1){ref-type="supplementary-material"}). To decipher the mechanisms involved in the phenotype associated with ChemR23-deficiency, we subsequently isolated abdominal aortic VSMCs from ChemR23^+/+^ and ChemR23^-/-^ mice. ChemR23^-/-^ VSMCs exhibited significantly less proliferation compared with ChemR23^+/+^ VSMCs (Figure [2A](#F2){ref-type="fig"}). These *in vitro* findings were hence in sharp contrast to the increased intimal hyperplasia observed in ChemR23-deficient mice *in vivo*. To further characterize the ChemR23-dependent phenotype, we subsequently studied the inflammatory response in LPS-activated peritoneal macrophages, isolated from ChemR23^+/+^ and ChemR23^-/-^ mice. ChemR23^-/-^ macrophages exhibited a more pro-inflammatory phenotype, with significantly higher mRNA levels of TNFα, and MMP9 (Figure [2B](#F2){ref-type="fig"}). These results indicated that in intimal hyperplasia, ChemR23^-/-^ VSMCs may have a protective phenotype (less proliferation), whereas ChemR23^-/-^ macrophages would be detrimental (increased inflammation). To assess the consequences of the differential macrophage phenotypes on VSMC proliferation, VSMCs were treated with CM derived from LPS-activated ChemR23^+/+^ and ChemR23^-/-^ macrophages, respectively. ChemR23^-/-^ macrophage CM significantly increased VSMC proliferation as compared with VSMC treated with ChemR23^+/+^ media (Figure [2C](#F2){ref-type="fig"}). ![ChemR23 deletion alters the Vascular smooth muscle cells (VSMC) and macrophage phenotype. **(A)** Basal ChemR23^+/+^ and ChemR23^-/-^ VSMC proliferation (48h) *in vitro,* assessed by WST-1. **(B)** mRNA expression of LPS-activated (100 ng/mL, 24 h) peritoneal macrophages *in vitro*. **(C)** VSMC proliferation (48 h) treated with conditioned media (CM) derived from LPS-activated ChemR23^+/+^ and ChemR23^-/-^ macrophages. For VSMC *n* = 3/group, for macrophages *n* = 4/group. Data represent mean ± SEM. *P*-values derive from Student's *t*-test.](fphar-09-01327-g002){#F2} Discussion ========== Three major observations emerge from the present study. First, our results revealed a protective role of ChemR23 in intimal hyperplasia, which was unrelated to the beneficial effects of enriching tissues with omega-3. Secondly, we identify ChemR23 as a transducer of opposing effects in different cell types, in terms of suppressing inflammatory activation of macrophages, but stimulating proliferation of VSMCs. Third, ChemR23^-/-^ macrophages in turn promoted VSMC proliferation. Taken together, these results suggest that ChemR23 signaling toward the resolution of inflammation protects from intimal hyperplasia by means of reducing macrophage activation. The observation that genetic ChemR23 deletion increased the inflammatory response in macrophages supports previous reports ([@B3]; [@B10]) and is consistent with our recent findings that ChemR23 deletion in hyperlipidemic mice accelerated atherosclerosis ([@B6]). The present study extends that observation by showing that a failure in the resolution of macrophage-induced inflammation by means of ChemR23 deficiency increased VSMC proliferation. This effect was independent of an increased total number of macrophages or M2 macrophage subtype. In accordance with this findings, in atherosclerotic lesions ChemR23 expressing macrophages do not correspond to an M2 phenotype ([@B6]). Indeed, previous results indicate that ChemR23 expression in M1 macrophages acts to promote the resolution of inflammation ([@B4]). Altogether the expression of ChemR23 in both macrophages and smooth muscle cells in human atherosclerotic lesions ([@B6]), support an extrapolation of these results to human pathophysiology. Previous reports have shown that pro-resolving lipid mediators decrease VSMC proliferation ([@B19]), and that VSMC lacking the lipoxin, and D-resolvin receptor ALX/FPR2 exhibit decreased migration ([@B12]). In the present study, however, ChemR23 deficient VSMCs proliferated less compared with wild-type VSMCs. Hence, in the absence of a macrophage-derived response, ChemR23 signaling may in contrast promote VSMC proliferation. These observations suggest that while ChemR23 signals to limit macrophage-induced inflammation, ChemR23 may transduce deleterious effects on VSMCs under non-inflammatory conditions. Whether different agonists transduce those differential responses, however, remains to be established. In general, pro-resolving agonists are associated with a beneficial smooth muscle response in vascular injury ([@B5]; [@B19]), abdominal aortic aneurysms ([@B14]; [@B13]) and atherosclerosis ([@B16]; [@B11]). The reduced intimal hyperplasia by ChemR23 appeared independently of the beneficial effects of omega-3 fatty acids, and suggest ChemR23 expression being directly coupled to the VSMC, and macrophage phenotypes. Other studies have indeed implicated signaling through the free fatty acid receptor-4 to mediate beneficial effects of omega-3 fatty acids in and murine models of intimal hyperplasia ([@B7]). In conclusion, the opposing effects of ChemR23 on VSMC and macrophages reported in the present study raise a novel notion for intimal hyperplasia pathophysiology, namely that the same receptor may transduce both protective and deleterious effects, which may vary over time depending on different stages in the resolution of inflammation. Ethics Statement ================ This study was carried out in accordance with the recommendations and guidelines of the Regional Ethical Review Board in Stockholm. The protocol was approved by the "Regional Ethical Review Board in Stockholm." Author Contributions ==================== GA, MC, and AL-F designed and performed the experiments, and analyzed the data. MB, GA, MC, and AL-F conceived the study. GA, MC, and MB wrote the manuscript. All authors participated in the interpretation of the data and provided critical review of the manuscript. Conflict of Interest Statement ============================== The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. **Funding.** This work was the supported by grants from the Swedish Research Council (Grant No. 2014--2312), the Swedish Heart and Lung Foundation (Grant Nos. 20150600 and 20150683), King Gustaf V and Queen Victoria Freemason Foundation, the Stockholm County Council (Grant No. 20170365), and Marianne and Marcus Wallenberg Foundation (2015.0104). Supplementary Material ====================== The Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fphar.2018.01327/full#supplementary-material> ###### Click here for additional data file. [^1]: Edited by: Lucy V. Norling, Queen Mary University of London, United Kingdom [^2]: Reviewed by: Asif Jilani Iqbal, University of Birmingham, United Kingdom; Soon Yew Tang, University of Pennsylvania, United States [^3]: ^†^These authors have contributed equally to this work [^4]: This article was submitted to Inflammation Pharmacology, a section of the journal Frontiers in Pharmacology

{ "pile_set_name": "PubMed Central" }

Background {#S0001} ========== Female lower abdominal pain has long posed diagnostic difficulties for clinicians, especially when they have had little more than the history and physical examination to guide them. Even first-trimester pregnancy diagnoses were expensive and time consuming until disposable qualitative immunological β-HCG assays became widely available in the 1980s. More accurate diagnoses have resulted from using transabdominal and transvaginal ultrasonography, computerized tomography, and diagnostic laparoscopy. These modalities, however, usually are not available in resource-poor areas. When they are lacking, as in the case presented here, clinicians may continue to follow the wrong diagnostic path, leading to unnecessary surgical procedures and potentially devastating treatment delays. Patient and case report {#S0002} ======================= A 19-year-old sexually active high school student with a history of regular menses believed she was pregnant after missing her period for three months. Still in school and not ready to have a child, she wanted to terminate the pregnancy but was reluctant to visit a healthcare professional. She could not pay for such a visit and even though she had government-sponsored medical insurance, it did not cover abortion services. Her girlfriend, the only person in whom she confided, suggested misoprostol for termination; she had never been pregnant or previously used misoprostol. A Licensed Chemical Seller (LCS) sold her misoprostol without a prescription. The fourth of 6 siblings, the patient had been sent to school from a rural farming community and lived near the school with her two younger sisters. Her first coitus was at 16 years, but she had never used birth control or condoms, claiming that she did not believe that she would get pregnant or contract sexually transmitted diseases (STDs). Her religious beliefs did not influence her behaviour. Vaginal bleeding began a few hours after taking 200 mcg misoprostol orally and the same dose vaginally. She bled for four days, accompanied by severe constant lower abdominal pain. Nevertheless, she remained at home. Five days later, the bleeding subsided but right sided pelvic pain persisted. Two days later, she went to a hospital near her home, where she was given the clinical diagnosis of septic incomplete abortion with severe anaemia (Hgb 7.3 g/dl). They had no ultrasound or transfusion capability. Twenty-four hours after performing a manual vacuum aspiration (MVA) with no observable products of conception, they transferred the patient with a diagnosis of pelvic abscess. Despite her level of illness, she came to our district hospital, about 20 minutes away by taxi, the common transport for ambulatory patients. On arrival, the thin and small-for-her-age woman looked pale and obviously ill. Her vital signs were temperature 37.8°C, BP 100/60 mm/Hg, pulse 100/min, regular and good volume. Her abdomen was diffusely tender with peritoneal signs. Since the one surgeon for the facility was operating, she was sent to the maternity ward, where the nurses, immediately recognizing her level of illness, started an IV, drew labs, and notified the surgeon. She was found to have a Hgb 6.1 g/dl, WBC 15.9 x109/L, with 10.8% lymphocytes, 9.0% mononuclear cells, and 80.2% neutrophils. A physician-performed ultrasound transabdominal examination showed an empty but bulky uterus with fluid in the dependant parts of the abdominal cavity. The presumptive diagnosis was pelvic abscess, although other processes were considered that also required surgery. She received two units of whole blood, additional crystalloids, ciprofloxacin and metronidazole. Under general anesthesia, exploration through a Pfannenstiel incision demonstrated that the uterus, fallopian tubes and both ovaries were normal, but that the peritoneal cavity contained a large quantity of serosanguinous fluid. The incision was extended in an inverted "T" to the epigastrium for a more extensive laparotomy. After draining the fluid, an inflamed appendix was found adhering to the cecum and no other pathology was identified. An appendectomy was performed ([Figure 1](#F0001){ref-type="fig"}), the peritoneal cavity was lavaged and the abdomen was closed primarily. ![Inflamed appendix in a girl treated for missed appendicitis after self-induced abortion](PAMJ-10-38-g001){#F0001} Post-operative course {#S20001} --------------------- Post-operative management included prolonged use of parenteral antibiotics. During the post-operative healthcare education, the patient stated that she did not intend to use either contraception or STD protection in the future, relying instead on sexual abstinence. Discussion {#S0003} ========== Presentation {#S20002} ------------ This patient presented to a relatively resource-poor hospital after treating herself for a self-diagnosed pregnancy. Although she never took a pregnancy test, her self-diagnosis was reasonable given her menstrual and sexual history. Doing a pregnancy test at the initial hospital would have been positive if she were still pregnant or had recently aborted. The persistent right lower quadrant abdominal pain should, however, have steered those clinicians in the direction of immediate referral, rather than first subjecting the patient to an MVA and then delaying her transfer by 24 hours. While lower abdominal pain, usually colicky, is often associated with misoprostol use, the cramps would have dissipated long before this patient presented to the hospital. Misoprostol {#S20003} ----------- Misoprostol is a prostaglandin analogue with uterotonic properties. It is 94% to 98% effective in inducing abortions when used orally, vaginally, or both. The failure rate is greater if used after 9 weeks gestation \[[@CIT0001],[@CIT0002]\]. The drug commonly causes vaginal bleeding within 12 hours of administration, with vaginal bleeding and uterine cramps being the most common reasons for women to seek medical care\[[@CIT0003]\]. In this case, misoprostol use may have complicated and delayed appropriate treatment. Under Ghanaian law, however, the patient should never have obtained misoprostol without first seeing an experienced physician. Misoprostol is a class C medication that may only be prescribed by doctors at a District Hospital or higher level institution, and only dispensed by registered pharmacists. In this case, an LCS illegally provided her with misoprostol. LCS\'s usually have only a few weeks of training and staff small retail outlets that sell over-the-counter medication. The LCS could be fined or their outlet closed down for selling misoprostol, although it is unclear whether this occurred. ### Diagnostic Process {#S30001} The differential diagnosis of persistent abdominal pain, especially following menstrual irregularities in a 19-year-old woman, has the potentially life-threatening ectopic pregnancy near the top of the list. It should also include all other causes of abdominal pain in an adult woman, including acute appendicitis. Acute appendicitis is the most common general surgical problem encountered during pregnancy \[[@CIT0004]\]. The incidence averages 1 in 550 pregnancies \[[@CIT0005]\]. With a recently pregnant patient, the potential dangers inherent in missing an ectopic pregnancy diagnosis would have been serious. It\'s unclear why that possibility wasn\'t considered, with the patient transferred immediately for diagnosis and treatment. When the patient arrived at our hospital, her clinical picture plus free fluid on the ultrasound made it clear that she required surgery, probably for an ectopic pregnancy or, given her lack of STD protection, a pelvic abscess. Appendicitis was a lesser consideration but, as Hickam\'s Dictum tells us, patients can have as many diseases as they wish \[[@CIT0006]\]. This contrasts with the more commonly used and cited principle of Ockam\'s Razor, which says that "of two explanations, each capable of explaining a given phenomenon, one should choose the simpler of the two" \[[@CIT0007]\]. Or, in other words, the hoof beats you hear probably mean horses, not unicorns. Even in non-pregnant women, however, it can be very difficult to differentiate appendicitis from pelvic inflammatory, gastrointestinal, and urinary tract disorders. Up to 1/3 of these patients are initially misdiagnosed. Such misdiagnosis is associated with an increased incidence of perforation and abscess formation, as well as prolonged hospitalization \[[@CIT0008]\]. Differentiating appendicitis from gynecological disease often requires adequate ultrasound capability (the hospital had only transabdominal probes on a non-moveable machine) and laparoscopy (not available)\[[@CIT0009]\]. One would imagine that this brush with death would alter the patient\'s view of safe sexual activity. Yet, when receiving education from nurses on these practices, this relatively well-educated patient demonstrated a cavalier attitude toward both birth control and sexually transmitted diseases. She asserted that she would rely on abstinence. Unfortunately, history shows that this will most likely lead to both pregnancy and sexually transmitted diseases. Conclusion {#S0004} ========== System problems in resource-poor areas can limit access to simple laboratory and healthcare services due to their unavailability in healthcare facilities or patients' lack of healthcare education or poverty. One problem that is difficult to remedy is unsafe sexual practices, even among better educated young women. In many countries, those selling medications often subvert official regulations and dispense potentially dangerous medications without a clinician\'s authorization. This leads to many problems, not the least of which is obscuring or delaying serious diagnoses and appropriate treatment. Clinicians must be wary of relying on patients' self-diagnoses while disregarding others that may better explain their symptomatology. The differential diagnosis for a woman\'s lower abdominal pain is extensive; neglecting other potential diagnoses risks increased morbidity (as in this case) or mortality. Sometimes clinicians must make more than one diagnosis to explain temporally and anatomically related symptoms. Authors' contributions {#S0005} ====================== Both authors participated in diagnosing and treating the patient in this case, drafting this article, and approving the final version for publication. Competing interests {#S0006} =================== The authors declare no competing interests.

{ "pile_set_name": "PubMed Central" }

Introduction ============ Elevated blood glucose concentration is an important factor for mortality and morbidity in critically ill patients. Previously, a randomized controlled trial showed that a blood glucose level of about 6 mmol/l is associated with less multiorgan failure and a significantly higher survival rate compared with levels above 8 mmol/l. Platelets play a crucial role in hemostasis and inflammation. However, the effect of short-term elevated blood glucose levels on platelet activation has not yet been evaluated systematically. Objective ========= To evaluate the influence of blood glucose levels on platelets *in vitro*. Methods ======= Citrated blood samples were drawn from healthy blood donors (40% male, age 38 ± 13 years \[mean ± SD\]). Exclusion criteria were smoking and the use of drugs interfering with platelet function. Blood samples were adjusted with glucose (Sigma, Taufkirchen, Germany) to final concentrations of 5 mmol/l (control group), 10 mmol/l (group 1) and 15 mmol/l (group 2), respectively. Samples were incubated for 10 min at 37°C with fluorescence-labeled monoclonal antibodies against CD62P, CD41, CD36, or CD42b (all: Beckman-Coulter, Krefeld, Germany). To evaluate platelet reactivity 2 and 6 μM thrombin-receptor-agonist-peptide-6 (TRAP-6; Bachem, Heidelberg, Germany) or 5 and 10 μM adenosine-di-phosphate (ADP; Sigma, Taufkirchen, Germany) were added. Analyses were performed in a flow-cytometer (EpicsXL; Beckman-Coulter). The mean fluorescence intensity was calculated. Determination of platelet aggregation was performed by the turbidimetric procedure (BCT, Dade Behring, Marburg, Germany). Aggregation was induced with ADP (200 μM/l), collagen (2 mg/l) and epinephrine (100 μM/l; all Dade Behring). Statistics for intergroup differences were performed by one-way ANOVA. Results ======= The initial blood glucose concentration was 5.0 ± 1.1 mmol/l. The blood glucose level had no significant influence on the expression of CD36 and CD62P, with and without stimulation. By contrast we observed a significant decrease in expression of CD42b in group 2 compared with the control group (unstimulated, *P*\< 0.001; TRAP-6, *P*= 0.005; ADP, *P*= 0.012). A similar observation was made for CD41 expression (unstimulated, *P*\< 0.001; TRAP-6, *P*\< 0.05; ADP, *P*\< 0.001). Also in group 1, a significant decrease in CD41 expression was observed after ADP stimulation. No significant differences were seen by aggregometry with either agonist. Conclusion ========== This *in vitro*study demonstrates that elevated blood glucose levels reduce expression of platelet receptors CD41 and CD42b. In contrast, platelet function measured by aggregometry showed no impairment. We conclude that acute hyperglycemia does not lead to lowered platelet function.

{ "pile_set_name": "PubMed Central" }

1. Introduction {#sec1} =============== Pluripotent stem cells are defined by their ability to undergo self-renewal and capability of differentiating into all three germ layers (mesoderm, endoderm, and ectoderm). Even though during development pluripotency is a temporary condition, ad hoc cultures can sustain indefinite self-renewal of cells isolated from the inner cell mass (embryonic stem cells, ESCs) in vitro \[[@B1]\]. Also, it is now well assessed that somatic cells can be reprogrammed back to the pluripotent state, obtaining the so-called induced pluripotent stem cells (iPSCs), thus reverting the physiological differentiation process \[[@B2]\]. Given the potential of pluripotent cells in regenerative medicine applications, in the recent years, a big research effort has been put in understanding the molecular mechanisms behind ESCs decision-making. Outcomes of this research are crucial to define optimal culture conditions to push cells into the desired pluripotent or differentiated state, to optimise somatic cell reprogramming, to better understand in vivo development, and to guide the use of reprogrammed cells for regeneration purposes. High variability in terms of functionality, gene expression, and epigenetic signature has been highlighted as a peculiar feature of both ESCs and iPSCs \[[@B3]\]. Focusing on gene expression variability in undifferentiated, isogenic mouse ESCs (mESCs), a number of pluripotency-related genes have been shown to be expressed heterogeneously and to present temporal fluctuations in mESCs cultured in standard serum/leukemia inhibitory factor (LIF) medium. Nanog, a master regulator of pluripotency and development \[[@B4]--[@B6]\], was the first pluripotency gene for which heterogeneity and temporal fluctuations were observed \[[@B7], [@B8]\]. This was followed by the discovery of heterogeneous expression of other pluripotency factors, such as T-box 3 \[[@B9]\], zinc-finger protein 42 (also known as Rex1) \[[@B10]\], Klf4 \[[@B10]\], Stella \[[@B11]\], Esrrb \[[@B12]\], and *β*-catenin \[[@B13]\]. Importantly, the mosaic and interconvertible distribution of pluripotency genes often correlates with different degrees of potency; at population level, mESCs are fully pluripotent, but subpopulations show a different predisposition towards differentiation. Alternative media to serum/LIF have been proposed \[[@B14]\]. The so-called 3i/LIF medium is serum free and contains 3 small molecules: PD0325901 (hereafter named PD), a MEK inhibitor \[[@B15]\]; SU5402, a FGF receptor inhibitor \[[@B16]\]; and CHIR99021 (Chiron), a glycogen synthase kinase-3 (Gsk3) inhibitor \[[@B17]\]. Intriguingly, most pluripotency genes become overall homogeneous in mESCs cultured in the 3i/LIF medium, even in the absence of FGF inhibitor (i.e., in 2i/LIF medium), and the "ground state" of pluripotency is achieved \[[@B14], [@B18], [@B19]\]. Is heterogeneity of Nanog and other pluripotency factors an inherent property of mESCs, fundamental for their competence to choose different fates \[[@B20]\], or is it an impediment to achieve standardizes cultures \[[@B21]\]? How do heterogeneity and complex temporal patterns of Nanog originate, and how do different culture media regulate them? In the recent years, both experimental and computational studies have tried to address these questions. In vitro cultures are an artificial approximation of in vivo systems and self-renewal is not part of in vivo development; still, the range of pluripotent states that exists in vitro as a response to different culture conditions \[[@B1]\] mimics different stages of development (preimplantation embryo and late blastocyst for 2i/LIF and serum/LIF cultures, resp. \[[@B22], [@B23]\]); thus, experimental and computational studies of gene expression patterns and dynamics in both media are relevant. Notably, many of the metastable mESC genes are, directly or indirectly, regulated by Nanog \[[@B20]\]; thus, in this review, we provide an updated overview of the major recent computational attempts to explain origins and functions of Nanog dynamics in mESCs, alongside with a review of available experimental data. We show that mathematical models can both aid in elucidating the mechanisms behind complex temporal gene expression dynamics and generate testable predictions. We start reviewing works that use small transcriptional gene regulatory networks (GRNs) to explain Nanog dynamics and their role in mESC fate decision; then, we report non-GRN approaches; finally, we discuss open questions in the field and possible future research directions. 2. Dynamics of mESCs Explained through Transcriptional Regulatory Networks: Multistable Models for mESC Pluripotency {#sec2} ==================================================================================================================== Notable is the work of Peterson\'s group in developing computational models to understand the role of transcription factor dynamics in mESC decision-making. We firstly review two early Peterson\'s works which consider the mESC switch between the pluripotent and the differentiated states a bistable system; Nanog heterogeneity is not considered, but the basis for formalisms developed later by the same authors to recapitulate more complex dynamics are given. In 2006 \[[@B24]\], a nonlinear ordinary differential equation (ODE) model, based on the Shea-Ackers formalism \[[@B25]\], was developed to describe mESC decision-making in view of the dynamics of the core pluripotency network, formed by Oct4, Sox2, and Nanog \[[@B26]\]. The topology is composed of nested positive feedback loops, which respond to environmental signals (inputs) and act on target genes (outputs), generically indicated as stemness and differentiation genes. Oct4 and Sox2 form a heterodimer, which positively regulates both genes and Nanog \[[@B27]\]; also, Nanog activates itself, Oct4, and Sox2 ([Figure 1()](#fig1){ref-type="fig"}). The ODE model describing such network predicts correlated dynamics of Nanog and Oct4-Sox2. Particular input signals (such as addition of LIF or Wnt to culture) can switch both on, pushing cells to a pluripotent state, while different inputs (such as p53) can switch them off. The system shows bistability; the pluripotent and differentiated steady states are stable and mutually exclusive, meaning that the system can converge to one of the two, depending on the initial conditions and inputs. Model analysis highlights a central role of Nanog autoregulation for the core pluripotency network maintenance. Also, the model predicts that, if Nanog is high enough, pluripotency is maintained even in the absence of Oct4 and Sox2 induction by external factors. Intriguingly, this prediction has been partly confirmed experimentally. Nanog can compensate for loss of self-renewal genes such as Esrrb, Tbx3 or Tcl1 \[[@B28]\], and *Oct4*^+/−^ mESCs lack Nanog-low, undifferentiated cells \[[@B29]\]. The model presents some limitations: it is unable to describe an effect of Oct4 overexpression in initiating differentiation \[[@B4], [@B30], [@B31]\], Nanog heterogeneity in the pluripotency state is not considered, and noise is neglected. In 2008, the same authors proposed an extension of the above model to better recapitulate mESC differentiation dynamics \[[@B32]\]. While the topology of the core pluripotency network is kept identical to the previous work, two new mutual antagonistic interactions are introduced to describe differentiation: Cdx2/Oct4 (responsible for trophectoderm lineage specification) \[[@B33]\] and Nanog/Gata6 (responsible for endoderm lineage specification) \[[@B34]\]. In addition, Cdx2 and Gata6 activate the orphan nuclear receptor germ cell nuclear receptor (Gcnf), a transcriptional repressor of pluripotency genes \[[@B35]\], which in turn represses Oct4. Finally, in order to account for evidence of cell tendency to specify towards an endoderm state when Oct4 is overexpressed \[[@B36]\], Gata6 activation by Oct4 is considered; this interaction had not been reported in the literature, so it is inserted as a model hypothesis, which is needed to recapitulate dynamics ([Figure 2()](#fig2){ref-type="fig"}). The authors developed an ODE formalism, similar to the one in their previous work \[[@B24]\]; the model reproduces the coexistence of three steady states, corresponding to pluripotent, trophectoderm, and endoderm cellular fates. In the absence of an external factor that forces Oct4 expression, the cells stay in the pluripotent state; the switch to a lineage-specific steady state is possible only by adding such external factor. The main achievement of this model is its ability to explain the biphasic response ("bell/inverse bell shaped") of TFs as functions of Oct4 (i.e., they can be expressed in low/high Oct4 ranges or at intermediate Oct4 levels, only), a mechanism not easy to explain by intuition. Interestingly, Gata6 activation by Oct4 hypothesized by the authors has been later proved, experimentally, in early blastocyst development \[[@B37]\]. As in the 2006 model \[[@B24]\], Nanog is expressed homogeneously in all cells in the pluripotent state. 3. Experimental Evidence of Nanog Heterogeneity in Serum/LIF-Cultured mESCs {#sec3} =========================================================================== In 2005, the first experimental evidence of Nanog multimodal distribution was reported \[[@B7]\]. Immunofluorescence experiments showed in serum/LIF-cultured mESCs the coexistence of two subpopulations, one positive for both Oct4 and Nanog and one positive for Oct4 only. In 2007, two different studies confirmed these results and highlighted new features of Nanog heterogeneity. Singh and colleagues \[[@B38]\], using a cell line previously developed by Mitsui\'s group (*β*-geo cells, in which the beta-galactosidase-neomycin fusion gene is inserted on one allele under the endogenous Nanog promoter \[[@B6]\]), confirmed heterogeneity of Nanog in contrast to homogeneous Oct4 expression. Analyzing by microarray the two Nanog subpopulations, a clear pluripotency signature characterized high Nanog (henceforth called HN) cells, while low Nanog (henceforth called LN) cells showed pronounced expression of mesodermal genes. Also in 2007, Chambers and colleagues \[[@B8]\] generated a mESC reporter cell line for Nanog, named TNGA, obtained by inserting a GFP-IRES-Puro-pA cassette at the Nanog start codon on one allele. This study provided evidence of both multimodal Nanog distribution and its temporal fluctuations between the two states; at steady state, two distinct (GFP negative and GFP positive, corresponding to LN and HN, resp.) subpopulations appeared clearly separated and could interconvert in time. In the following years, a number of additional reporter cell lines have been generated, and different experimental approaches, including flow cytometry and fluorescence-activated cell sorting (FACS), time-lapse microscopy, single-cell sequencing and single-molecule RNA-FISH, have been used to study Nanog distribution and dynamics at the levels of both single cells and cell populations \[[@B20]\]. 4. An Excitable Model for Nanog Heterogeneity in mESCs Grown in Serum/LIF {#sec4} ========================================================================= The first mathematical model accounting for Nanog heterogeneity in mESCs grown in serum/LIF medium was presented in \[[@B39]\]. Kalmar and colleagues combined experiments and mathematical modeling and proposed that Nanog dynamics arise from a noise-driven excitable system. Experiments were performed analyzing Nanog levels with the mentioned TNGA mESC cell line \[[@B8]\] both at population level by FACS and at single-cell level by time-lapse imaging. Nanog was proved to change dynamically, as sorted LN and HN mESCs were able to reproduce the original bimodal distribution, both at population level and at single-cell level; although with different timescales, sorted LN cells tended to rapidly switch to the HN state (within 24 hours), while sorted HN cells showed a slower switch off. At single-cell level, LN mESCs were able to switch Nanog on in 2-day time-lapse experiments (time lapses for the HN to LN transition were not included in the paper). Following these experimental observations, the authors derived a mathematical model showing that the HN state is stable for the system, with noise driving sporadic excursions of cells in the LN state. The system is excitable, thus highly sensible to perturbations (transcriptional noise, in this case) beyond a certain noise threshold. In detail, nonlinear ODEs describe the underlying GNR network, in which Oct4 and Nanog autoactivate their own and also each other expression and Oct4 inhibits Nanog, resulting in an overall negative feedback loop topology ([Figure 3(a)](#fig3){ref-type="fig"}). The authors assumed that Oct4 activates Nanog to saturated levels (thus this interaction is neglected in the equations) and only high levels of Oct4 can inhibit Nanog (a full biological motivation for such negative interaction, crucial for the system dynamics, is left by the authors for further investigation). The equations, based on Hill-like kinetics, consider protein dynamics only (mRNA steady-state assumption); Gaussian noise is added to the Nanog equation, in order to describe the stochastic excursions from the unique steady state of the system (HN). The model predicts that Oct4 is more homogenous in HN than in LN cells; this prediction is validated experimentally by immunofluorescence in FACS-sorted mESCs. While interesting, this result is not entirely surprising, taking into account the network topology. Finally, a stochastic model is derived and simulated using Gillespie\'s first reaction method. Both single-cell dynamics and population steady-state distribution of Nanog are reproduced, showing a good agreement with experimental data, and indicating that the regulation of occupancy of the two Nanog states is noise driven. The authors showed that the excitable regime is robust to parameter variations and concluded that Nanog cooperative autoregulation is crucial for the system\'s dynamics. 5. Multistable Models for Nanog Heterogeneity in mESCs Grown in Serum/LIF {#sec5} ========================================================================= In 2010, Glauche and colleagues proposed an alternative GRN-based formalism to recapitulate again experimental data obtained with the TNGA cell line. The modeled GRN includes Oct4-Sox2 heterodimer autoactivation and its activation on Nanog; also, as in Kalmar\'s model, Nanog autoactivation is considered ([Figure 3(b)](#fig3){ref-type="fig"}). This latter positive feedback loop, for proper parameter choices, can give raise to bistability. The stochastic differential equation (SDE) model, based on Hill kinetics, describes only protein dynamics. At steady state, the Nanog bimodal distribution observed experimentally in \[[@B8], [@B39]\] is reproduced; in time course simulations, fluctuations between the two states randomly happen due to the introduction of Gaussian noise (only in the Nanog equation). The main difference with the Kalmar model is that, in this case, both the LN and HN states are stable, with noise allowing the switch between them. Altering the noise changes the transition probability of the two Nanog states and, consequently, the residence times within each state. Interestingly, the authors simulate also the redistribution of TNGA cells upon sorting, performed experimentally in \[[@B39]\], matching the reestablishment of both populations on a timescale of circa 10 days. The model is finally analyzed to study the effect of Nanog heterogeneity on differentiation; the authors also consider an external differentiation signal, which inhibits Oct4 and is inhibited by Nanog, generating a double negative feedback loop. As a result of such loop, only LN mESCs respond to the differentiation signal with a decrease in Oct4-Sox2 complex level, indicating that Nanog has a "gatekeeper" role in mESC pluripotency maintenance. Of note, the differentiation signal is external; thus, no mESC spontaneous cell differentiation is conjectured. In the same paper, the authors also consider an alternative scenario in which, due to the introduction of a negative feedback loop through a not specified gene "X" that is activated by Nanog and represses it ([Figure 3(c)](#fig3){ref-type="fig"}), Nanog shows sustained oscillations, even in the absence of noise. Notably, such Nanog-mediated autorepression, hypothesized in the paper for modeling aims, has been later experimentally proven \[[@B40]\]. While still being able to recapitulate the experimental steady-state Nanog distribution, the oscillation scenario significantly differs in the sorting simulations: while the system is predicted to continuously shift from one steady state to the other in the fluctuation case, damped oscillations lead to the reestablishment of the steady-state distribution in the oscillation scenario. The same authors, more recently, tested experimentally these two discordant predictions, showing higher plausibility of the noise-induced bistable dynamics \[[@B41]\]. More recently, Lakatos and colleagues proposed a different formalism to describe Nanog dynamics arising from a bistable system \[[@B43]\]. The authors compared dynamics of five GRN topologies compatible with experimental evidence of interactions in the core pluripotency network. In all the analyzed GRNs, the genes considered are Oct4, Sox2, and Nanog; Oct4-Sox2 is a common transcription factor for the three genes and Nanog activates Sox2 ([Figure 4()](#fig4){ref-type="fig"}, blue arrows). In addition, the following interactions are included or not: Nanog activation on Oct4 (as in earlier Chickarmane\'s works \[[@B24], [@B32]\]), repression of Oct4 on Nanog, and autoinhibition of Nanog (experimental evidence in \[[@B44]\] and \[[@B40]\], resp.). Of note, in all topologies, Nanog autoactivation is neglected. Quasi-steady-state approximation is used for mRNA levels, proteins are modeled using ODEs based on the Shea-Ackers representation for transcriptional regulations, and variations to parameters are introduced to account for cell-to-cell diversity. Intriguingly, modeling shows that all the topologies considered can give rise to bistability, due to the Oct4-Sox2 autoregulation and dimerization. Due to the feedback loop between Nanog and Oct4, original parameter values predict the coexistence of two steady states for both Nanog and Oct4, contradicting experimental evidence of Nanog heterogeneity in contrast to Oct4 homogeneity \[[@B45]\]. However, parameter manipulation can shift nullclines and reproduce the mentioned experimental data. The authors also considered an extended topology ([Figure 4](#fig4){ref-type="fig"}, additional interactions indicated with black arrows), which includes the downstream pluripotency genes Klf4 and Esrrb (the latter acting as an amplifier \[[@B46]\]), fibroblast growth factor (FGF) signaling as an autocrine module, and a negative feedback through the MAP kinase/ERK kinase (MEK) cascade \[[@B47]\], which affects the cooperativity of the Oct4/Sox2-Nanog interaction \[[@B48]--[@B50]\]. Of note, FGF and MEK signaling have been proved experimentally to drive differentiation of mESCs, with a potential role in mESC heterogeneity \[[@B47]\]. Simulations of this extend topology show that the system allows the existence of distinct substates of Nanog and downstream transcription factors, still within the HN state; LN mESCs have higher Nanog expression than cells fully committed to differentiation, in agreement with experimental evidence \[[@B45]\]. Noise induces stochastic fluctuations between the multiple steady states: Nanog fluctuates in phase with Esrrb and Klf4, and in antiphase with FGF5, with no significant changes in Oct4 levels. Interestingly, multiple substates of the ON state of Nanog are predicted, suggesting the coexistence of multiple subpopulations of mESCs, responding differently to external signaling and having different pluripotency signature. The authors confirm experimentally that several components of the pluripotency network present a range of expression levels in mESC subpopulations, as also previously reported \[[@B45]\]. 6. Nanog Dynamics in Serum-Free Cultures {#sec6} ======================================== The development of culture conditions that enable mESC ground state pluripotency (2i/LIF medium) prompted the development of new mathematical formalisms to understand the mechanisms behind the overall loss of pluripotency gene heterogeneity. In 2012, the Peterson group proposed a novel model \[[@B51]\] to recapitulate distribution and dynamics of Nanog in different culture conditions (heterogeneity and fluctuations in LIF/bone morphogenetic protein (BMP4), homogeneity in 2i/LIF-3i/LIF) and their impact on somatic cell reprogramming (the BMP4/LIF culture condition was considered equivalent to serum/LIF \[[@B52]\]). In the network topology, the authors included Nanog and Oct4 autoactivations, and the activation of Oct4-Sox2 on Nanog ([Figure 5(a)](#fig5){ref-type="fig"}), as in their aforementioned previous works \[[@B24], [@B32]\]. In order to represent cells with high Oct4 levels while low Nanog \[[@B39]\], the previously considered Nanog activation on Oct4 and Sox2 is now removed. Also, the network includes a differentiation gene (loosely indicated as "G" in [Figure 5(a)](#fig5){ref-type="fig"}), which autoactivates itself, is activated by Oct4-Sox2 and mutually represses Nanog. Importantly, medium-regulated external factors are included: LIF/BMP4, which activates Nanog, the FGF4 receptor signaling, and Gsk3. FGF4 and Gsk3, lumped into a single variable, inhibit Nanog, are activated by the Oct4-Sox2 complex, and are blocked by 2i/3i medium ([Figure 5(a)](#fig5){ref-type="fig"}). Thus, in the overall topology, Nanog is regulated by a feed-forward loop including Oct4-Sox2 direct activation and inhibition via FGF4. An important assumption is that only when the Oct4-Sox2 heterodimer is bound to Nanog promoter additional TFs can bind (experimental evidence in \[[@B53]\]). A nonlinear ODEs model based on the Shea-Ackers formalism \[[@B25]\] is used, assuming steady-state levels for mRNA concentrations. To account for stochasticity originated within the network, the authors also formulated a master equation, simulated using the Gillespie algorithm. In BMP4/LIF medium, the system is bistable, and stochastic Nanog fluctuations arise. When Nanog levels are very low, cells are irreversibly pushed into a differentiation state, indicating again that Nanog has a pluripotency gatekeeper function. Also, the model predicts that suppression of FGF4/Gsk3 by 2i/3i can push cells into the monostable state (HN); in single-cell simulations, Nanog fluctuates around the unique steady state due to noise, but heterogeneity at population level is lost. Notably, the Nanog steady state in 2i/3i is higher than the HN steady state in BMP4/LIF medium. Of note, the noise is kept at the same level, both in BMP4/LIF and 2i/3i simulations; although LIF is not present in the 2i/3i simulations, the stem cell state is initialized with low "G," and high Nanog and Oct4-Sox2 levels. Interestingly, the model is also used to predict somatic cell reprogramming dynamics, performing simulations upon overexpression of Oct4 and Sox2: model predictions match the experimental observation of the importance of reprogramming factor dosage for the efficiency of the process \[[@B54]\]. If cells are cultured in 2i, an increase in reprogramming efficiency is predicted, with synchronous Oct4 and Nanog switch-on dynamics; these predictions are in agreement with experimental evidence of an increased efficiency in the late reprogramming phase if partially reprogrammed cells are cultured in 2i/LIF medium \[[@B55], [@B56]\]. In 2014, Herberg and colleagues proposed another GRN-based modeling framework to describe medium-dependent Nanog dynamics and the resulting mESC pluripotency signature \[[@B57]\]. The authors extended the network topology modeled in 2010 \[[@B42]\] introducing (i) the pluripotency marker Rex1 \[[@B58]\], directly activated by both Nanog and the Oct4-Sox2 heterodimer, in order to reproduce experimental data generated using the Rex1GFPd2 cell line \[[@B59]\]; (ii) FGF/Erk signaling, which inhibits Nanog and is activated by Oct4; and (iii) a phenomenological external differentiation signal, which inhibits the Oct4-Sox2 complex and is inhibited by Nanog, to recapitulate differentiation dynamics ([Figure 5(b)](#fig5){ref-type="fig"}). Of note, Rex1 is only an output of the system, as it is not involved in any feedback. Overall, the topology is comparable to the one modeled by Chickarmane in 2012 \[[@B51]\], but here, spontaneous differentiation arising from Nanog heterogeneity is not considered: only the influence of the cellular environment is investigated. Hill-based ODEs are derived for all genes; Gaussian noise is added to Oct4-Sox2, Nanog, and Rex1 equations; finally, steady state for mRNAs is assumed. As in the Chickarmane et al. 2012 paper, the system presents bistability, abrogated when culture medium is switched to 2i (LIF is implicitly considered here). In particular, in serum/LIF (FGF/Erk active), Nanog shows bistability due to its positive autofeedback loop, which strength is adjusted in order to have a steady-state distribution of circa 20% LN and 80% HN cells, as observed in TNGA experiments \[[@B8], [@B39]\]; the model can also reproduce Rex1 distribution measured culturing Rex1GFP2 mESCs. In time simulations, noise allows the switch between the HN and LN states, which is predicted to happen in the timescale of days. Increasing the level on noise results in more frequent switch and raised proportion of LN cells. Parameters of the Oct4-Sox2 complex are adjusted to keep its level constantly high and homogeneous. When simulating the 2i condition, the block of Erk (due to the presence of PD in the medium) reduces the inhibition of Nanog, pushing the system outside the bistability region into the "ground pluripotency" state, characterized by high Nanog, Oct4-Sox2, and Rex1 and low FGF/Erk. Of note, the effect of Chiron in the 2i medium is not considered. The model is used to predict differentiation dynamics: after the first 24 hours of 2i withdrawal, Nanog decreases with fast dynamics, followed by Rex1 and, ultimately, Oct4-Sox2, matching experimental observations \[[@B60]\]. Experimental data in the paper validate model predictions about the asynchronicity of the differentiation process, possibly due to individual cell variability arising when the inhibitor is removed from the medium. A comparison with differentiation dynamics of cells cultured in serum/LIF (thus, starting from bimodal Nanog and Rex1 distribution) is not reported. 7. A Probabilistic Model for Nanog Multistationary Dynamics {#sec7} =========================================================== Luo and colleagues \[[@B61]\] proposed an alternative approach to describe medium-depended Nanog dynamics in mESCs. Differently from the above formalisms, the authors did not rely on a regulatory network, to avoid introducing assumptions in case of lacking or contrasting data about underlying regulations, and possible overparameterization. Instead, they used a statistical model, based on probability density functions and Gaussian noise. The model has just one variable (representing Nanog), for which the existence of a stationary distribution is assumed (represented as a mixture of Gaussian distributions). Using nested sampling, model parameters are fitted on novel experimental data presented in the paper: the authors identify in TNGA mESCs the distribution of Nanog as a mixture of high, medium, and low (HN, MN, and LN, resp.) subpopulations. Then, they focus on the MN state, which is shown by experiments to have the highest dynamical changes. Nanog dynamics are measured in sorted MN cells replated in different culture conditions for four days (LIF/BMP4, PD only, Chiron only, and 2i/LIF); Nanog distribution shifts towards the HN state in the 2i/LIF condition and these data are used to validate model predictions about sorting dynamics. Intriguingly, Chiron widens the LN and MN states, while PD narrows them. One interesting result of the fitting is a competitive effect of PD and Chiron: while the first tends to add noise to the system, the second can filter noise. The approach used in this paper is definitively interesting: it allows inferring directly the shape of the potential function from the data, without a priori knowledge of the underlying signaling network. The experiments were performed 4 days after the sorting; it could be of interest to perform both experiments and simulations on a longer timescale, to check if the distributions reported in the paper are or not at steady state. 8. Nanog Heterogeneity Arising from Allelic Switching {#sec8} ===================================================== Miyanari and Torres-Padilla \[[@B62]\] recently suggested an alternative mechanism contributing to Nanog heterogeneity based on allelic regulation. The authors firstly analyzed Nanog nascent transcription by RNA fluorescent in situ hybridization (RNA-FISH) in vivo, observing monoallelic and biallelic expression in preimplantation embryo and in late blastocyst, respectively. To study the dynamic switch of Nanog in vitro, the authors generated a novel reporter cell line based on a dual-reporter system (named NGR mESCs), in which two distinct fluorescent reporters (destabilized TurboGFP and mCherry) are inserted downstream of the NANOG-coding region in the two alleles. In NGR mESCs cultured in serum/LIF, Nanog transcriptional firing was observed, with an odd distribution of allelic firing: only a very low percentage of cells showed Nanog biallelic expression, in contrast to biallelic expression of other pluripotency genes such as Oct4. Consequently, in vitro cultures showed distinct populations for Nanog: monoallelic, biallelic, and no expression; similar results were obtained using single-molecule RNA-FISH (smRNA-FISH) \[[@B63]\]. In contrast, biallelic expression was reported in 2i/LIF. These results suggested a crucial role of allelic expression, and the possibility of achieving ground state pluripotency by activating the second Nanog allele. In light of such results, Wu and Tzanakakis proposed a mathematical model to recapitulate medium-dependent Nanog dynamics \[[@B64]\]. The mESC population is divided into 4 groups (biallaleic, monoallelic, and both alleles inactive), with percentages of residence in each group and probabilities of transition among groups extrapolated from RNA-FISH and allele-specific RT-PCR experiments in \[[@B62]\] ([Figure 6](#fig6){ref-type="fig"}). The stochastic allelic switching is modeled using a homogeneous Markov chain, assuming that the next state does solely depend on the current state; asynchronous proliferation rates are considered. The model is simulated using a Monte Carlo algorithm, setting the half-life of endogenous Nanog to 2 hours, as measured in \[[@B65]\]. In the absence of noise, at population level, Nanog distribution shows 3 peaks (low for both alleles inactive, middle for monoallelic, and high for biallelic cells), while addition of noise leads to a bimodal distribution (LN and HN only). In single-cell-simulated dynamics, HN cells show a slow (circa 20 hours) switch off, while LN cells switch on in circa 5 hours. The predicted dynamics are much faster than the ones previously observed with TNGA mESCs; thus, the authors simulate the expected output of the latter cells, in which the reporter is inserted only in one allele and has half-life longer than endogenous Nanog. In this case, even in the absence of transcriptional noise, only 2 peaks in the reporter distribution are predicted, confirming experimental observations in \[[@B8], [@B39]\]. Also, the predicted correlation between endogenous Nanog and the reporter dynamics is very high (Pearson coefficient = 1) if both alleles are tagged and the half-life of the reporter is comparable to the one of Nanog, while correlation is lost (Pearson coefficient = 0.06) using a TNGA-like tagging approach. Finally, the authors use the model to simulate dynamics of Nanog^+/−^ cells, predicting that Nanog deletion on one allele can impact its distribution at population level. The paper does not include results about Nanog dynamics in 2i/LIF, but the authors mention that the model could easily be extended to describe the latter case by properly changing the proportion of cells in each subpopulation. Simulation results about the impact of differences between endogenous and reporter kinetics could explain mismatches in steady-state distributions and dynamics observed using different Nanog-tagged cell lines. Such predictions have been partly corroborated recently: the group of Henrique generated a novel reporter cell line (Nd mESCs), which contains a transgenic bacterial artificial chromosome (BAC) with a destabilized Venus reporter protein (Venus-Nuclear-PEST) under the control of Nanog regulatory regions \[[@B66]\]. Nd mESCs differ from TNGA cells for two main reasons: the two endogenous Nanog alleles are kept intact and the half-lives of the reporter mRNA and protein are comparable to the endogenous Nanog ones (circa 4 and 2 hours, resp.). In Nd mESCs grown in serum/LIF, as compared to TNGA cells, Nanog expression is again mosaic but within an overall narrower distribution; also, fluctuations between the LN and the HN state and restoration of the original distribution from LN- and HN-sorted cells occur in a shorter timescale (circa 4 hours and 2--4 days of culture, resp. \[[@B66]\]). 9. Nanog Dynamics Arising from a Protein Interaction Network {#sec9} ============================================================ Recently, Muñoz-Descalzo and colleagues \[[@B67]\] considered an alternative model to describe Nanog dynamics in mESCs focusing on posttranscriptional interactions. In view of recent data generated by the same research group about the key role of protein balance for mESC pluripotency maintenance \[[@B68]\], the authors reanalyzed the excitable model of Kalmar and colleagues and pointed out its failure in reproducing (i) the increased correlation between Oct4 and Nanog observed by single-cell immunofluorescences in 2i/LIF medium and (ii) a critical region characterized by low levels of both Nanog and Oct4, in which pluripotency is lost. The authors started considering a minimal model, named NOC ([Figure 7(a)](#fig7){ref-type="fig"}), which encompasses only Oct4 and Nanog, and assume that the correlation between the two proteins results from the formation of a protein complex (O:N) previously described experimentally \[[@B69]--[@B71]\]. In the model, Oct4 and Nanog proteins exist either free or bound together in a complex; transcriptional regulations are neglected. Importantly, the model takes into account different stabilities for proteins (high for Oct4 and Nanog/Oct4 complex and low for Nanog), measured experimentally in the paper. To account for transcriptional heterogeneity of Nanog, its transcriptional bursting \[[@B40], [@B62]\] is considered. The resulting discrete stochastic formalism is simulated for both serum/LIF and 2i/LIF conditions by tuning Nanog expression (infrequent transcriptional bursting in serum/LIF and high frequency bursting in 2i/LIF). The model correctly matches Nanog and Oct4 correlation in the two media, and Nanog mRNA distribution (unimodal in both media, but shifted towards high levels in 2i/LIF) observed experimentally by mRNA-FISH. Still, the model has some pitfalls: it predicts that knock-out of Nanog results in increased Oct4 levels, the latter experimentally associated with differentiation, contradicting experimental evidence about mESC ability to maintain pluripotency even in absence of Nanog \[[@B8]\]. Thus, the authors developed a refined model (named TBON), which also includes Tcf3 and *β*-catenin, two Wnt pathway proteins associated with pluripotency \[[@B59], [@B72], [@B73]\] and involved in a protein-complex with Oct4 \[[@B74], [@B75]\]. The new topology ([Figure 7(b)](#fig7){ref-type="fig"}) considers, in addition to the Oct4-Nanog complex (O:N), the inducers (PD and Chiron) and three additional complexes: *β*-catenin with Oct4 (*β*:O) \[[@B74], [@B75]\], *β*-catenin with Tcf3 (*β*:T) \[[@B76]\], and *β*-catenin with Oct4 and Nanog (*β*:O:N). Stochastic simulations of the model match protein distributions and correlations of protein levels in both serum/LIF and 2i/LIF. The model shows that, in 2i/LIF, ground state pluripotency is achieved by attenuation of free Oct4, thus limiting its effect in promoting differentiation. Also, the model predicts that, in absence of Nanog, the *β*:O complex becomes stronger, enabling cells to maintain pluripotency, as confirmed in experiments reported in the paper. In case of lack of *β*-catenin, the model predicts that the correlation between Nanog and Oct4 is unchanged, but their levels are lower, due to Tcf3 increase. These model predictions are in agreement with experimental evidence about the not absolute requirement of *β*-catenin for pluripotency \[[@B77]\], although its lack makes mESCs more prone to differentiate \[[@B59]\]. Conversely, abrogation of Tcf3 sustains pluripotency through increased Nanog, decreased Oct4 and *β*-catenin, and impaired Oct4-Nanog correlation. Finally, the model predicts that the removal of Oct4 from the system causes a drop in Nanog levels (as, if not bound in protein complex, Nanog has faster degradation), in agreement with loss of pluripotency in Oct4^−/−^ cells \[[@B31]\]. 10. Limitations of Existing Formalisms for Ground State Pluripotency Cultures {#sec10} ============================================================================= All the reviewed models but the Muñoz-Descalzo et al. one just consider the molecular effect of the MEK inhibitor (PD), neglecting effects of the Gsk3 inhibitor (Chiron) in ground state pluripotency cultures. Given the failure of PD alone in sustaining mESC clonal propagation \[[@B1], [@B59]\] and the crucial functions of the Wnt/*β*-catenin pathway in pluripotency maintenance and successful reprogramming of somatic cells to pluripotency \[[@B13], [@B78]--[@B82]\], the role of the Wnt pathway in cellular heterogeneity and mESC plasticity should be better characterized. Also, possible effects of nonhomogenous, or cell density-dependent, drug uptake in cellular cultures on the system dynamics might be considered. Finally, no current model accounts for a certain level of Nanog heterogeneity and temporal fluctuations persistent in 2i/LIF recently observed experimentally \[[@B83], [@B84]\]. It remains to be shown whether long-term cultures in 2i/LIF could impair mESC karyotypic and epigenetic stability, given prolonged exposure of cells to Gsk3 inhibition \[[@B85]\]. Recently, R2i, an alternative serum-free, chemical media, has been proposed to sustain mESC ground state pluripotency \[[@B86]\]. Of note, also R2i eliminates mESC heterogeneity, but it acts on the pluripotency circuit through other routes (it contains inhibitors of TGF*β* and FGF signaling pathways); this provides a good indication that mosaic expression of pluripotency genes in serum/LIF might originate from multiple sources, which mathematical formalisms should consider. 11. Crosstalk between Gene Expression Dynamics and the Cell Cycle {#sec11} ================================================================= Mathematical models could be particular useful in elucidating the complex interconnection between cell cycle, the pluripotency network, and cellular fate. In the experimental works mentioned in this review, fluctuations of Nanog reporter genes were observed within a mESC cell cycle \[[@B62], [@B84]\]; however, there is also evidence of coupling between pluripotency network gene dynamics and the cell cycle, as recently reviewed in \[[@B87]\]. In an early work reporting Nanog heterogeneity \[[@B38]\], FACS-sorted HN cells were found to upregulate cell cycle genes characteristic of the S-G2 phases, while LN cells expressed genes characteristic of the G1 phase. Also, MacArthur and colleagues, using an inducible system, found a correlation between Nanog and cell cycle checkpoints genes \[[@B60]\]. Recently, Nanog but not Oct4 expression was proved to oscillate in mESCs synchronized for the cell cycle \[[@B88]\]. In a recent work \[[@B41]\], Herberg and colleagues extended their previous GRN model \[[@B57]\] to include mESC proliferation; the resulting agent-based model predicts an effect of different cell cycle times in the proportion of cells in low-high Nanog subpopulations in serum/LIF; analysis in ground state pluripotency cultures (in which mESC cell cycle is known to differ \[[@B89]\]) was not performed. Multiscale modeling approaches, able to account simultaneously for processes at subcellular, intercellular, and population levels, could be highly informed by quantitative single-cell measurements (such as live imaging and sequencing), and be able not only to reproduce experimental data but also to generate useful predictions, usable for targeted control of mESC fate, both in pluripotency maintenance and in differentiation. 12. Conclusions and Future Perspectives {#sec12} ======================================= Here, we have reviewed recent computational/experimental results about mechanisms and consequences of Nanog dynamics in populations of isogenic mESCs. We reported main assumptions, results and predictions of mathematical models based on regulatory networks (in which Nanog dynamics result from its interactions with other pluripotency genes, signaling pathways, and drugs at transcriptional and posttranslational levels \[[@B39], [@B42], [@B43], [@B51], [@B57], [@B67]\]), statistical models \[[@B61]\], and unbalanced Nanog allelic expression \[[@B64]\]. A common conclusion is that Nanog acts as a molecular gatekeeper, fine-controlling cell fate in response to pluripotency and differentiation genes\' regulations, internal noise, and external stimuli. We showed that different formalisms are able to reproduce Nanog dynamics observed experimentally with reporter mESCs cultured in both serum/LIF and chemical, serum-free culture media, and generate testable predictions. Still, many open questions remain regarding both discordant experimental results and the validity of modeling assumptions. How can models based on different GRN topologies reproduce the same experimental data? One critical step in deriving mathematical models is parameterization. Notably, given the same system of equations, different set of parameters and timescales of the variables involved can result in completely different dynamic scenarios; conversely, different sets of genes can be included in GRNs to reproduce specific dynamics, as far as the topology encompasses key elements (e.g., positive feedback loop(s) for bistability). It is therefore crucial, as far as possible, to use parameters directly measured and rely only on interactions unambiguously identified experimentally. Also, when comparing simulations to experimental data, it is vital to critically account for the experimental settings and the timescales of the considered variables. Among the models reviewed here, only Wu and Tzanakakis explicitly took into account the differences in the degradation rates of endogenous Nanog and its reporter; their results suggest modeling as a powerful tool for experimentalists to infer endogenous dynamics from fluorescent reporter data. The Wu and Tzanakakis paper, as the Luo and colleagues one, is for sure interesting also for being able to describe Nanog dynamics at population level starting from single-cell dynamics, without relying on a GRN model with consequent assumptions on the network topology. However, the Luo et al. model, while capable of reproducing Nanog heterogeneity, cannot explain its source. Focusing on timescales, the commonly used steady-state assumption for mRNA dynamics, while reducing the parameter space, might lead to misrepresentative results. Indeed, in 2014, two independent reports \[[@B90], [@B91]\] focused on Nanog distribution at the protein level and found that mESCs cultured in serum/LIF do not show protein biallelic expression, in contrast with mRNA firing measured in \[[@B62]\]. In particular, in \[[@B91]\], Filipczyk and colleagues generated mESCs encoding dual fusion proteins to the Nanog C-terminus, with green and red fluorescent reporters placed on the two alleles. At cell population level, FACS analysis showed bimodal distribution for Nanog in mESCs cultured in serum/LIF, although with a narrower expression range as compared to TNGA cells (as with Nd cells), but correlated expression of the two fusion proteins, suggesting biallelic expression of Nanog protein. These results are in agreement with examination of mature cytoplasmic RNA transcript, which was found to be biallelic regardless allelic Nanog firing \[[@B63]\]. Given that the fusion fluorescent proteins generated by Filipczyk and colleagues have an half-life comparable to the endogenus Nanog\'s one, protein biallelic expression can be compatible with the mRNA allelic switching reported in \[[@B62]\] only if different timescales for mRNA and protein are considered \[[@B20]\]. Taken together, these results suggest that the quasi-steady-state assumption for Nanog mRNA dynamics might be revised and that GNR feedback mechanisms are still needed to explain heterogeneous Nanog protein dynamics in serum/LIF. Also, in our view, the impact of intrinsic and extrinsic noise on mESC fate should be better addressed: molecular instability is often represented incorporating a stochastic term in a set of differential equations, but whether it is a good representation has not been addressed. It would be extremely interesting to combine quantitative estimate of noise (as in \[[@B92]\]) with GRN-based formalisms. Scaling up the reviewed mathematical models for small GRNs, involving other key genes involved in pluripotency and differentiation identified with bioinformatics and statistical methods \[[@B60], [@B93]--[@B95]\] and unraveling their role in mESC temporal dynamics, remains an open challenge. Also, the descriptive and predictive power of mathematical models could be significantly increased accounting also for the combined role of noncoding RNAs \[[@B96]--[@B99]\], epigenetic mechanisms \[[@B100]\], metabolism \[[@B101]\], and posttranscriptional modifications \[[@B102]\]. From a more broad prospective, we do believe that the questions about the role of heterogeneity in mESC decision-making stated at the beginning of this review have not yet been fully addressed. Xenopoulos et al. \[[@B103]\], using high-resolution live cell imaging, reported "salt and pepper" Nanog pattern in early blastocyst, but lack of temporal fluctuations, and irreversible commitment to epiblast and primitive endoderm. If and how heterogeneity can be advantageous for mESC plasticity is currently not known; the principles behind the use of noise and fluctuations by genetic circuits, also in an evolutionary prospective, are beginning to be elucidated only recently \[[@B104]\]. The hypothesis that pluripotency is an emergent property of a population of cells rather than a characteristic of a single cell \[[@B60]\], with noise and cellular heterogeneity conferring stem cell high entropy, and thus the potential to choose a number of specialized, differentiated fates \[[@B105], [@B106]\], still needs to be more extensively characterized not only for embryonic but also induced and adult stem cells. Also, the "exploratory hypothesis" for pluripotent cells, which conjectures stem cell decision-making as a two-step process in which firstly stochasticity induces a critical state that primes diverse transcriptional programs, and then one particular fate is chosen via interaction with external inputs \[[@B107], [@B108]\], represents, in our view, an interesting research avenue for further investigation. This work was funded by MRC (MR/N021444/1) (Lucia Marucci) and BrisSynBio, a BBSRC/EPSRC Synthetic Biology Research Centre (BB/L01386X/1) (Lucia Marucci). Conflicts of Interest ===================== The author declares that there is no conflict of interests regarding the publication of this paper. ![Core pluripotency regulatory network modeled in \[[@B24]\]. The core network includes the transcription factors Oct4, Sox2, and Nanog and the heterodimer Oct4-Sox2. Arrow- and bar-headed lines represent activation and inhibition, respectively, among genes (green ovals) and inputs/outputs (grey rectangles).](SCI2017-7160419.001){#fig1} ![Extended regulatory network for pluripotency modeled in \[[@B32]\]. The network includes, in addition to the core pluripotency factors Oct4, Sox2, and Nanog, Cdx2 and Gata6 to represent trophectoderm and endoderm commitment through Oct4-Cdx2 and Nanog-Gata6 interactions, respectively, and the nuclear receptor Gcnf. Green and purple ovals represent pluripotency and differentiation genes, respectively.](SCI2017-7160419.002){#fig2} ![Pluripotency regulatory networks to describe excitable, bistable and oscillatory Nanog dynamics in serum/LIF. (a) Regulatory network proposed in \[[@B39]\] to describe Nanog dynamics in serum/LIF as excitable; additive noise is considered for Nanog and allows the system to transiently escape the HN stable steady state. (b) and (c) Networks proposed in \[[@B42]\] to reproduce Nanog dynamics as bistable (b) or oscillatory (c). The noise term, summed to Nanog\'s equation, is indispensable to reproduce Nanog heterogeneity and dynamics in (b) only. In (c), the negative feedback loop between Nanog and a not specified gene ("X") can generate oscillatory dynamics.](SCI2017-7160419.003){#fig3} ![Extended regulatory network to describe Nanog multistable dynamics in serum/LIF \[[@B43]\]. Blue and black arrows indicate interactions of the core and extended network, respectively. The extended network interactions include additional pluripotency genes (Klf4, Esrrb), and differentiation signaling pathway (FGF and MAPK) genes.](SCI2017-7160419.004){#fig4} ![Regulatory network to recapitulate Nanog dynamics in 2i/LIF. Networks proposed and modeled in \[[@B51]\] (a) and \[[@B57]\] (b). Both systems include Nanog, the Oct4-Sox2 heterodimer, a module for FGF signaling, and the effect of the medium (2i/3i) on network regulations. In addition, in \[[@B51]\] (a), the interactions with LIF/BMP4 and a differentiation gene ("G") are included, while Herberg and colleagues (b) include Rex1 as a system output only and a differentiation signal ("Y," similar to "G" in (a)). In both models, the effects of the two inhibitors in 2i are lumped.](SCI2017-7160419.005){#fig5} ![Schematic of Nanog allelic expression, as modeled in \[[@B64]\]. mESCs are grouped into 4 types given different patterns of Nanog allelic expression; the percentage fraction of mESCs residing in each group is indicated as well as the transition probability (in percentage).](SCI2017-7160419.006){#fig6} ![Pluripotency protein interaction networks, as modeled in \[[@B67]\]. Scheme of the minimal (a) and refined (b) protein interaction networks. O:N, *β*:T, *β*:O, and *β*:O:N indicate Oct4-Nanog, *β*-catenin-Tcf3, *β*-catenin-Oct4, and *β*-catenin-Oct4-Nanog complexes, respectively.](SCI2017-7160419.007){#fig7} [^1]: Academic Editor: Gary E. Lyons

{ "pile_set_name": "PubMed Central" }

Background ========== The effectiveness of conservative treatment of scoliosis is controversial. Some studies suggest that brace is effective in stopping curve progression, whilst others did not report such an effect. The purpose of the present study was to effectiveness of PASB in the correction of lumbar curves, in agreement with the SRS Committee on Bracing and Nonoperative Management Standardisation Criteria \[[@B1]\]\[[@B2]\]. Materials and methods ===================== Fourty adolescent females (mean age 12.95 ± 1.72 years) with lumbar curve and a pretreatment Risser score ranging from 0 to 2 have been enrolled. The minimum duration of follow-up was 24 months (mean: 41.75 ± 34.47 months). Antero-posterior radiographs were used to estimate the curve magnitude (CM) and the torsion of the apical vertebra (TA) at 5 time points: beginning of treatment (t1), one year after the beginning of treatment (t2), intermediate time between t1 and t4 (t3), end of weaning (t4), 2-year minimum follow-up from t4 (t5). Three situations were distinguished: curve correction, curve stabilisation and curve progression. Results ======= CM mean value was 26.43 ± 2.77 SD at t1 and 13.80 ± 7.94 SD at t5. TA was 10.83 ± 3.74 SD at t1 and 7.88 ± 4.24 at t5. The variation between measures of Cobb and Perdriolle degrees at t1,2,3,4,5 and between CM t5-t1 and TA t5-t1 were significantly different. Curve correction was accomplished in 82.5% of patients, whereas a curve stabilisation was obtained in 17.5% of patients. Conclusions =========== The PASB, due to its peculiar biomechanical action on vertebral modelling, is highly effective in correcting lumbar curves.

{ "pile_set_name": "PubMed Central" }

Published: March 17, 2020 Introduction {#sec1} ============ The rodent hippocampus coordinates a wide array of behaviors, from spatial navigation ([@bib34]) to decision making under approach-avoidance conflict ([@bib17], [@bib22]) and reward processing ([@bib17]). A central hypothesis of how the hippocampus might participate in such diverse behaviors is the presence of heterogeneous principal neurons that differ widely in their gene expression ([@bib10], [@bib42]), electrophysiological properties ([@bib23]), and behavioral function ([@bib11], [@bib12], [@bib17], [@bib42]). In particular, the main ventral hippocampal output region, the ventral subiculum (vS), is composed of multiple neuronal populations that send parallel, long-range projections to distinct areas, including prefrontal cortex (PFC; vS^PFC^), lateral hypothalamus (LH; vS^LH^), and nucleus accumbens shell (NAc; vS^NAc^) ([@bib32]). These populations are proposed to integrate a myriad of local and long-range inputs ([@bib4], [@bib42], [@bib51]) to perform their unique behavioral functions ([@bib1], [@bib11], [@bib17], [@bib22], [@bib41]). However, to date, knowledge of the input connectivity of vS output neurons is lacking. Further, vS populations are spatially patterned, in particular along the proximal-distal (PD) axis (ranging from the CA1 to the presubiculum borders) ([@bib11]), and synaptic input varies dramatically across different spatial locations in vS ([@bib2], [@bib11], [@bib25], [@bib30], [@bib46]). Based on this, we hypothesized that different vS output populations receive distinct upstream inputs, and we reasoned that these inputs may in turn depend on the spatial location of postsynaptic neurons ([@bib10], [@bib11]), their downstream target ([@bib17], [@bib22], [@bib32]), or a combination of these two factors. To address this hypothesis, we studied the anatomical organization of projection-defined neurons in vS; we then applied rabies tracing across different vS projections and postsynaptic cell locations, and obtained a brain-wide map of inputs to vS subpopulations. We identified quantitative differences in multiple long-range input regions to vS that depended to different extents on the spatial location and projection target of vS neurons. Results {#sec2} ======= Hippocampal Projection Populations Are Topographically Organized in vS {#sec2.1} ---------------------------------------------------------------------- Each vS projection population is thought to occupy a unique spatial distribution in the hippocampus ([@bib23]), and long-range input into hippocampus has been shown to be highly topographical ([@bib2], [@bib11], [@bib25], [@bib30], [@bib46]). Therefore, we first wanted to determine the spatial distribution of the different projection populations within vS, because this spatial distribution could be an important determinant for differential input connectivity. To do this, we stereotaxically injected the retrograde tracer cholera toxin subunit-B (CTXβ) into PFC, LH, or NAc in a pairwise manner ([Figure 1](#fig1){ref-type="fig"}A); collected coronal sections that spanned the hippocampus; conducted whole-hippocampus cell counts; and registered the data to the Allen Brain Atlas (ABA; see [STAR Methods](#sec4){ref-type="sec"}; [Figures 1](#fig1){ref-type="fig"}B--1F). We first confirmed that vS^PFC^, vS^NAc^, and vS^LH^ neurons resided predominantly in vS ([Figure 1](#fig1){ref-type="fig"}C), and that the fraction of colocalized cells (neurons that projected to more than one injection site) ranged from 2% to 6% ([@bib32]). Using this approach, we found that vS projection populations occupied spatially distinct locations within vS ([Figures 1](#fig1){ref-type="fig"}D and 1E). The locations of these neurons varied across all three axes (anterior-posterior \[AP\], medial-lateral \[ML\], and dorsal-ventral \[DV\]), but most notably along the AP axis ([Figures 1](#fig1){ref-type="fig"}E, 1F, 1H--1J), whereas vS^PFC^ neurons were located at more anterior locations, vS^LH^ neurons were located at more posterior locations, and vS^NAc^ neurons were spread across the entire range of vS. Using correlation analysis, we confirmed that projection type covaried across all axes, but most dramatically with AP position ([Figure 1](#fig1){ref-type="fig"}F), and that the spatial position of neurons along AP was most predictive of their output target (see [STAR Methods](#sec4){ref-type="sec"}; [Figure 1](#fig1){ref-type="fig"}F).Figure 1Non-overlapping vS Neurons Occupy Distinct Spatial Sites(A) Schematic of pairwise CTXβ injections.(B) Left: side view of the hippocampus with the plane of coronal sectioning illustrated. Right: example schematic of a coronal section. Boxed area is the approximate location of the example images in (C).(C) Left: examples of stitched coronal sections of vS with retrogradely labeled cells. Scale bar: 500 μm. Middle: zoom-in images of boxed area from left. Scale bar: 50 μm. Right: proportion of single- and dual-labeled hippocampal neurons.(D) 3D whole-brain diagrams with CTXβ-labeled hippocampal neurons (n = 10,8291 hippocampal neurons from 6 brains; see [Video S1](#mmc4){ref-type="supplementary-material"}).(E) Cumulative distributions of CTXβ-labeled cells in ventral hippocampus along AP, ML, and DV axes (vS^LH^: n = 7,717 cells, 8 injections; vS^NAc^: n = 9,989 cells, 6 injections; vS^PFC^: n = 9,287 cells, 6 injections).(F) Left: Pearson correlation coefficient was highest along AP compared with ML and DV (n = 10 injections). Right: logistic regression analysis predicts projection type based on spatial location. Removal of AP as a predictor led to the largest decrease in accuracy (see [STAR Methods](#sec4){ref-type="sec"}). Data are summarized as mean ± SEM. Individual injections are overlaid in gray.(G) Right: side view of the hippocampus, with horizontal cutting plane illustrated. Middle: example stitched horizontal brain section. Boxed area indicates location of the left ventral hippocampus. Scale bar: 500 μm. Right: schematic of ventral hippocampus in the horizontal plane, illustrating PD axis. Boxed area is the approximate location of the example images in (H).(H) Example stitched horizontal sections of retrogradely labeled vS neurons. Scale bar: 100 μm.(I) Cumulative distributions of cell counts along the proximal-distal (PD), deep-superficial (DS), and DV axes (vS^LH^: n = 5,785 cells, 7 injections; vS^NAc^: n = 3,405 cells, 6 injections; vS^PFC^: n = 3,509 cells, 7 injections; see [STAR Methods](#sec4){ref-type="sec"}).(J) Pearson correlation (left) and logistic regression (right) analyses analogous to (F). The PD axis captures the most variation and best predicts the projection type of vS neurons (n = 10 injections).(K) LDA of registered neurons in vS unbiasedly identifies the anatomical plane that best discriminates between projection populations. Left: cell positions projected onto the first and second linear discriminant axes. Cumulative distributions of cells along the first (middle) and second (right) linear discriminant demonstrate overlap of projection populations.(L) Schematic of vS topography, where vS neurons are intermingled, vS^PFC^ neurons are located most proximally, vS^LH^ neurons most distally, and vS^NAc^ neurons span the area between both vS^PFC^ and vS^LH^.Video S1. Rotating 3D Whole-Brain View of CTXβ-Labeled Hippocampal Neurons Registered to the Allen Brain Atlas, Related to Figure 1Data from 108,921 registered, CTXβ-labeled hippocampal neurons are presented in a virtual glass brain (color code: vS^PFC^ in yellow, vS^NAc^ in magenta, vS^LH^ in cyan). Each dot represents one labeled hippocampal neuron. We reasoned that the marked distribution of projection populations across the AP axis might be a reflection of the known PD distribution of projection populations along the pyramidal cell layer of the hippocampus ([@bib11]), which in ventral hippocampus is oriented approximately along the AP axis ([Figures 1](#fig1){ref-type="fig"}G and 1H). We confirmed that this was the case by making horizontal slices of labeled vS ([Figure 1](#fig1){ref-type="fig"}G). This experiment showed that the variance in AP is well explained by a PD gradient of neurons in vS, where vS^PFC^ neurons were located at more proximal locations near the CA1 border, vS^LH^ neurons were located at more distal locations near the presubiculum border, and vS^NAc^ neurons were spread across the entire PD axis ([Figures 1](#fig1){ref-type="fig"}H and 1I). In keeping with previous observations, correlation and regression analyses also indicated that projection type varies most along the PD compared with the deep-superficial (DS) and DV axes ([Figure 1](#fig1){ref-type="fig"}J). Interestingly, the spatial distributions of the different projection populations were highly overlapping across each of the AP, ML, DV, and PD axes. This is notably different from that observed in dorsal subiculum ([@bib11]), where there is a sharp PD border separating distinct projection populations. To control for the possibility that this sharp border was not simply obscured by our coronal or horizontal slicing angles, we carried out linear discriminant analysis (LDA) on the registered whole-hippocampus neuronal distributions ([Figure 1](#fig1){ref-type="fig"}K). This method allowed us to unbiasedly find a virtual plane that best separates the three populations of projection neurons. By examining the distribution of neurons in this plane, we observed that, although there was a clear subregion organization, the spatial distribution of the three populations in vS still remained highly overlapping ([Figure 1](#fig1){ref-type="fig"}K). Thus, in contrast with dorsal subiculum, vS appears to be organized as a gradient, especially with vS^NAc^ neurons being present at almost every point along the AP axis. Overall, our data indicate that vS projection populations are segregated cell types and occupy overlapping yet distinct locations within vS, best separated across the AP axis. Labeling of Hippocampal Input Dependent on Spatial Location and Projection Target {#sec2.2} --------------------------------------------------------------------------------- Next, to directly assess the organization of presynaptic inputs onto these vS projections, we applied *tracing the relationship between input-output* (TRIO) ([@bib6], [@bib8], [@bib37], [@bib39]) to vS projections ([Figure 2](#fig2){ref-type="fig"}A; [Figures S1](#mmc1){ref-type="supplementary-material"}A--S1E). This approach involved injecting *AAV2-retro-Cre* ([@bib45]) into the output target region to retrogradely express Cre recombinase in vS neurons that projected to the target site. In the same surgery, we injected a single Cre-dependent helper construct (*AA2/1-synP-FLEX-split-TVA-2A-B19G* \[TVA-G\]) into vS. After at least 2 weeks of TVA-G expression in projection-specific neurons, G-deleted, pseudotyped rabies virus (*EnvA*-*RVΔG-H2B-mCherry*) harboring nuclear-localized mCherry was injected into vS to infect TVA-G^+^ cells.Figure 2TRIO of vS Output Neurons(A) Schematic of TRIO.(B) Left: example stitched sagittal section of the brain with rabies-labeled cells (black) in the hippocampus. Right: zoom-in of boxed area from left. Starter cells colocalize for TVA-G (cyan) and rabies (red). Scale bars: 1,000 μm (right); 200 μm (left).(C) Overall input from coarse anatomical structures normalized to the total number of inputs counted in a single brain (n = 20 brains). Note the break in the y axis. Data are summarized as mean ± SEM.(D) Representative stitched images of long-range monosynaptic inputs to vS neurons by projection type and the center of mass (COM) of the starter cells along the AP axis. Classification of starter cells into anterior and posterior groups was based on the rank order of COM along the AP axis. Approximate sagittal planes are displayed with red-shaded boxes indicating the estimated locations of the corresponding images. For images of MPO input, three consecutive images (spanning 180 μm) were projected to produce the displayed representative images. 3V, third ventricle; ac, anterior commissure. Scale bars: 200 μm (all input regions except PVT); 100 μm (PVT).See also [Figure S1](#mmc1){ref-type="supplementary-material"}. Importantly, we systematically varied the injection site of TVA-G and rabies within vS ([Figure 2](#fig2){ref-type="fig"}A). This method ensured that for each projection target, we sampled distributions of starter cells (i.e., cells from which rabies virus begins the monosynaptic retrograde tracing) from different locations within vS ([Figures 2](#fig2){ref-type="fig"}A and 2B; [Figure S1](#mmc1){ref-type="supplementary-material"}F). This provided us with experimental control over both starter cell location (by injection location of TVA-G and rabies within vS) and output projection (by injection location of *AAV2-retro-Cre* in target sites). This approach allowed us to assess how variations in spatial position of vS neurons (across the AP, ML, and DV axes), projection target (across PFC, NAc, and LH), and their combination influence input size and identity. In parallel, we performed control experiments where either Cre, TVA-G, or both Cre and TVA-G were excluded, to show that input labeling by rabies infection required the presence of both Cre and TVA-G ([Figures S1](#mmc1){ref-type="supplementary-material"}G--S1K). Brain-wide Rabies Tracing Reveals Biased Connectivity of vS Projection Neurons {#sec2.3} ------------------------------------------------------------------------------ To quantify input to vS, we conducted brain-wide cell counts of rabies-labeled neurons in sagittal sections and registered the data to the ABA ([@bib20], [@bib33]). Consistent with previous studies, the majority (∼90%) of direct inputs to all vS projection types and spatial locations arose locally within the hippocampal formation ([Figure 2](#fig2){ref-type="fig"}C; [Figures S2](#mmc1){ref-type="supplementary-material"}A--S2E). In addition, there were numerous long-range inputs arising from ∼35 brain regions spanning thalamic, striatal, pallidal, cortical, hypothalamic, and amygdalar regions ([Figures 2](#fig2){ref-type="fig"}C and 2D; [Figure S2](#mmc1){ref-type="supplementary-material"}E), including the nucleus of diagonal band, medial septum, nucleus reuniens (RE), posterior amygdala (PA), and preoptic area. Further, we observed that there were variations in labeling from these regions when we compared tracing from different projection targets or starter cell position (center of mass \[COM\]) along the AP axis ([Figure 2](#fig2){ref-type="fig"}D). Next, we wanted to quantitatively investigate whether long-range input onto vS neurons depended on either projection target (PFC, LH, or NAc; [Figure 3](#fig3){ref-type="fig"}A), spatial location (COM in AP, ML, and DV; [Figures 3](#fig3){ref-type="fig"}A and 3B), or their combination (see [STAR Methods](#sec4){ref-type="sec"}; [Figure 3](#fig3){ref-type="fig"}). For each brain region that contributed more than 1% of extrahippocampal input, we built a linear regression model with both COM and projection information as predictors ([Figure 3](#fig3){ref-type="fig"}C) and tested the ability of the model to predict the percentage of total extrahippocampal input from that region. Using this strategy, we found that input from medial preoptic area (MPO), PA, RE, and paraventricular thalamus (PVT) provided quantitatively different input sizes to vS depending on either COM, output projection, or both predictors ([Figure 3](#fig3){ref-type="fig"}D). Importantly, these results were independent of normalization method, because we found similar projection and spatial dependence using alternative normalization procedures ([Figures S3](#mmc1){ref-type="supplementary-material"}A and S3B).Figure 3TRIO of vS Output Neurons Reveals Projection and Spatial Bias of Inputs(A) Schematic of TRIO protocol for assessing projection (via retro-Cre injection in the output site) and spatial dependence (via varying AP injection sites of TVA-G and rabies) of inputs.(B) 3D scatterplots (left) and geometric COM of individual brain samples (right) of starter cells. COM of individual brains are represented as ellipsoids (center defined as the mean position and radii as 1 standard deviation of the cell distribution in AP, ML, and DV). The spatial distribution of starter cells along the three brain axes was comparable (standard deviation of COM: AP = 0.23 mm; ML = 0.19 mm; DV = 0.28 mm).(C) Schematic of linear regression analysis. Linear models were constructed with COM (in all three brain axes), and output type as independent variables and the extrahippocampal input fraction (log-transformed) as the dependent variable (n = 20 brains).(D) Full model fits assessed by computing the F-statistic of full models with COM and projection as predictors. Inputs from PVT, MPO, PA, and RE (red) exceed the threshold of p = 0.05 after correction for multiple comparisons.(E) Likelihood ratio tests (see [STAR Methods](#sec4){ref-type="sec"}) show that MPO and PA inputs are projection dependent, whereas PVT and RE are spatially dependent. p values have been corrected for all 15 input regions tested, but only the significant hits shown in (D) are illustrated. Dashed lines indicate p = 0.05.(F) Quantification of extrahippocampal inputs. The number of inputs from a brain region is expressed as a percentage of the total number of extrahippocampal inputs (vS^NAc^: n = 8 brains; vS^PFC^: n = 5 brains; vS^LH^: n = 7 brains). Shaded boxes indicate projection-dependent input regions with significant model fits, followed up with post hoc pairwise Tukey multiple comparisons (^∗^p \< 0.05). See also [Video S2](#mmc5){ref-type="supplementary-material"}. Data are summarized as mean ± SEM.(G) Same dataset as in (E) where the input fractions are plotted as a function of COM along the AP axis. Shaded continuous line represents smoothed input density. Inputs from RE and PVT (∗) resulted in a significant model fit with a statistically significant COM AP coefficient.See also [Figures S2](#mmc1){ref-type="supplementary-material"} and [S3](#mmc1){ref-type="supplementary-material"} and raw cell count data in [Tables S1](#mmc1){ref-type="supplementary-material"} and [S2](#mmc1){ref-type="supplementary-material"}.Video S2. Rotating 3D Whole-Brain Views of Rabies-Labeled Inputs Registered to the Allen Brain Atlas, Related to Figure 3Example data for each condition (vS^PFC^, vS^NAc^, and vS^LH^ TRIO in posterior and anterior locations) illustrated as rotating virtual glass brains. Each dot represents one rabies-labeled input neuron. We next sought to directly investigate the relative contribution of projection type or COM on the amount of RE, PVT, MPO, and PA inputs. To do this, we compared single-predictor (projection or COM) models with combined (projection and COM) models (see [STAR Methods](#sec4){ref-type="sec"}; [Figure 3](#fig3){ref-type="fig"}E). Using this approach, we found that MPO and PA innervated vS dependent solely on the projection target of the postsynaptic neuron, whereas RE and PVT inputs were dependent on COM ([Figure 3](#fig3){ref-type="fig"}E). Utilizing a similar approach, we next investigated whether AP, ML, or DV information was important for the spatial dependence of PVT and RE input, and found that for both RE and PVT, input was most dependent on starter cell location along the AP axis ([Figure S3](#mmc1){ref-type="supplementary-material"}C), consistent with our predictions from [Figure 1](#fig1){ref-type="fig"}. Post hoc testing revealed that MPO input selectively innervated vS^NAc^ neurons, and PA input selectively targeted vS^LH^ and vS^NAc^ neurons, whereas RE and PVT input targeted vS^LH^ and vS^NAc^ neurons only at posterior locations within vS ([Figures 3](#fig3){ref-type="fig"}F and 3G). Subsequently, we obtained a qualitative description of all extrahippocampal inputs (35 regions) and their relative dependence on COM or projection, which revealed a wide range in the relative dependence of different synaptic input on both COM and projection identity ([Figures S3](#mmc1){ref-type="supplementary-material"}D and S3E). Finally, we confirmed the spatial predictions of the rabies tracing data by analyzing anterograde labeling experiments available publicly from the ABA ([Figures S3](#mmc1){ref-type="supplementary-material"}F and S3G), which showed strong spatial dependence of axon innervation across the AP axis for RE and PVT, but not for MPO or PA input. Overall, our rabies tracing dataset identified brain-wide regions that project to vS, including quantitatively biased inputs from MPO, RE, PVT, and PA that depend differentially on the location and projection identity of the postsynaptic neuron. Biased Nucleus RE Input to Hippocampal Projection Neurons {#sec2.4} --------------------------------------------------------- A surprising finding in our dataset was that RE input was anatomically biased to avoid vS^PFC^ neurons ([Figure 4](#fig4){ref-type="fig"}A). This finding runs counter to classic models of the vS-PFC-RE-vS circuit where RE functions as a relay between PFC and hippocampus via long-range input to vS^PFC^ projections ([@bib19], [@bib47]). We thus sought to confirm these anatomical data using *channelrhodopsin-assisted circuit mapping* (CRACM) to ensure that this result was not due to methodological constraints such as viral tropism ([@bib28]) or activity dependence of viral tracing ([@bib7]) (see [Discussion](#sec3){ref-type="sec"}). From our tracing data, we hypothesized that RE input was spatially biased, i.e., RE input targets posterior areas where vS^PFC^ neurons are not abundant (COM). In addition, we wanted to ask whether RE input does not form synaptic connections even with less abundant vS^PFC^ neurons in the most posterior locations (projection).Figure 4Nucleus Reuniens Inputs to vS Projection Neurons Are Functionally Biased(A) Extrahippocampal fractions of RE inputs as a function of COM and projection population. Non-PFC projectors (i.e., pooled vS^NAc^ and vS^LH^) have overall higher RE inputs than vS^PFC^ (n = 15 non-PFC projectors, n = 5 PFC-projectors; Mann-Whitney U test, U = 12.0, p = 0.013). Data are summarized as median and interquartile range.(B) AAV expressing ChR2 under the *synapsin* promoter was injected bilaterally into RE.(C) Stitched confocal image of a horizontal section of hippocampus. RE axons expressing ChR2 target distal regions of vS. Scale bar: 500 μm.(D and H) In the same surgery after ChR2 injection into RE, red and green retrobeads were injected into either LH and PFC (D) or NAc and PFC (H).(E and I) Light-evoked excitatory postsynaptic currents in pairs of (E) vS^LH^ and vS^PFC^ or (I) vS^NAc^ and vS^PFC^. All data represent responses to 10 ms of blue light. Solid line: mean response; shaded region: 95% confidence interval.(F and J) Scatterplots of light-evoked photocurrents from vS^LH^ and vS^PFC^ cell pairs (F) and vS^NAc^ and vS^PFC^ cell pairs (J).(G and K) Normalized EPSCs from vS^LH^ (G) and vS^NAc^ (K) neurons scaled to the photocurrent elicited in neighboring vS^PFC^ neurons; the relative amplitudes of light-evoked photocurrents are higher in vS^LH^ (Wilcoxon signed-rank test, n = 18 pairs from 3 animals, V = 4.0, p = 3.86 × 10^−4^) and vS^NAc^ (Wilcoxon signed-rank test, n = 11 pairs from 3 animals, V = 6.0, p = 0.016). Data are summarized as median ± 95% confidence intervals.(L--N) Models of vS input connectivity based purely on (L) spatial location or (M) projection identity. (N) Our dataset supports a combined model from (L) and (M), where inputs can be either projection or COM dependent.See also [Figure S4](#mmc1){ref-type="supplementary-material"}. We injected adeno-associated virus (AAV) to express ChR2 in RE ([Figure 4](#fig4){ref-type="fig"}B) and found that ChR2^+^ axons emanating from RE were observed most densely in the distal, posterior region of vS ([Figure 4](#fig4){ref-type="fig"}C), consistent with the spatial dependence of our rabies tracing data. In these slices, we recorded light-evoked postsynaptic currents from pairs of retrogradely labeled neurons within axon-rich distal vS. By recording from pairs of neighboring neurons within the same slice in the presence of tetrodotoxin (TTX) and 4-aminopyridine (4-AP), this approach allowed us to directly compare the relative monosynaptic RE input strength across projection populations while controlling for PD and AP position, ChR2 expression, and light intensity. We observed that excitatory RE input was indeed much weaker onto vS^PFC^ neurons, whereas it strongly targeted neighboring vS^LH^ and vS^NAc^ neurons ([Figures 4](#fig4){ref-type="fig"}D--4K; [Figure S4](#mmc1){ref-type="supplementary-material"}). Overall, these results demonstrate a functional bias of RE inputs away from vS^PFC^ and toward both vS^LH^ and vS^NAc^. Discussion {#sec3} ========== Using a combination of retrograde tracing, conditional rabies tracing, and optogenetics with whole-cell electrophysiology, we demonstrated that the vS output circuitry is composed of projection-specific, topographically organized populations that receive a range of local and long-range inputs. In turn, the targeting of these inputs depends to different degrees on the spatial position and projection target of the postsynaptic vS neuron. Topography of vS Projections {#sec3.1} ---------------------------- We used CTXβ retrograde tracing to reveal the distribution of neurons in vS that project to PFC, NAc, and LH, and confirmed previous findings that suggest that there are unique populations of neurons in vS that project to each downstream region ([@bib32]). However, due to the efficiency of retrograde labeling, significant proportions of neurons projecting to multiple downstream sites cannot be definitively ruled out. Using single-neuron tracing, it was recently shown that many neurons in dorsal subiculum do indeed project to multiple downstream areas ([@bib50]). In the future, it will be interesting to look at individual neurons in vS and directly compare the extent of projection specificity and how this varies with spatial location. Our data, however, show that a large proportion of neurons in vS appear to project to unique downstream areas. Consistent with previous findings, we found that different vS projections are organized topographically. vS^PFC^ neurons were located most anteriorly (along AP), vS^LH^ were found posteriorly, and vS^NAc^ were widely distributed throughout vS ([@bib23], [@bib11]). We demonstrated that this AP gradient is most likely due to the previously described divergence along the PD axis of the pyramidal cell layer. The spatial variation of projection neurons within vS prompted us to experimentally vary our injection coordinates during rabies tracing, which enabled us to sample starter cells from different COM positions ([Figure 3](#fig3){ref-type="fig"}B). Thus, we were able to investigate the relative contribution of both projection identity and spatial location of vS neurons on the type and amount of rabies-labeled inputs. Biased Input to vS Based on Projection Identity and Spatial Location {#sec3.2} -------------------------------------------------------------------- Rabies tracing identified many inputs to vS, including MPO, PA, RE, and PVT, that targeted distinct populations in vS differentially; MPO selectively innervated vS^NAc^, PA selectively innervated vS^NAc^ and vS^LH^, and RE and PVT innervated vS^NAc^ and vS^LH^ only in posterior vS. Notably, the consistent MPO inputs to vS^NAc^ have not, to our knowledge, been previously described in the literature ([@bib51], [@bib9]). The specificity in input labeling across the different vS projections raises interesting questions regarding the function of these upstream neurons. For example, vS is important for social memory, and this behavior may be controlled in part by MPO input to vS^NAc^ ([@bib31], [@bib35]). More generally, the rabies tracing data support a model of combined topographical and output-defined connectivity of vS inputs ([Figures 4](#fig4){ref-type="fig"}L--4N) where, depending on the upstream region, vS inputs are biased according to the location, projection type, or both of these attributes. This variation in inputs across space and projection type is in keeping with the known spatial and projection-specific functions of subiculum; across PD subdivisions, proximal vS is involved in sensory encoding ([@bib25]) and distal vS supports path integration ([@bib11], [@bib25]), whereas across projection populations, vS^PFC^ and vS^LH^ encode innate threat ([@bib1], [@bib17], [@bib22]) and vS^NAc^ encodes social memory ([@bib35]). Crucially, the fact that RE inputs target vS^NAc^ and vS^LH^ neurons in posterior vS also support the existence of spatial- and projection-specific function. For example, goal-directed locomotion has been proposed to be specific to both vS^NAc^ neurons and distal subiculum ([@bib11], [@bib17], [@bib35]). Biased Connectivity of RE Input Away from vS^PFC^ Neurons {#sec3.3} --------------------------------------------------------- A surprising result from our dataset was that the RE input did not innervate vS^PFC^ neurons. RE is essential for bidirectional communication between hippocampus and PFC. This thalamic region is proposed to form an anatomical link between hippocampus and PFC, thereby closing a PFC-RE-vS-PFC loop ([@bib19], [@bib21], [@bib47]). Little information is present about the extent to which the projection-defined vS populations are involved in this circuit loop, and it is generally assumed that RE input is integrated by vS^PFC^ neurons to relay signals from hippocampus to PFC. We found that vS^PFC^ neurons receive few and functionally weak RE inputs. Thus, although vS^PFC^ neurons do indeed receive input from RE, our data suggest that the main effect of this input is by integration upstream of vS^PFC^ neurons (i.e., through dorsal hippocampus, entorhinal cortex, local interneurons, or other pyramidal populations) before being transmitted back to PFC. It will be necessary for future work to investigate in more detail the hippocampal microcircuitry that is involved in integrating RE input, the details of other multi-synaptic routes that may allow reciprocal connectivity between the hippocampus and PFC, and how this circuit organization contributes to behavior. Interestingly, we also show that RE input is anatomically and functionally biased toward vS^NAc^ and vS^LH^ neurons. This observed circuit connectivity complements previous data that demonstrated key roles for RE in goal-directed planning ([@bib21]) and fear generalization ([@bib52]), functions that may be crucial for the proposed roles of vS^NAc^ in reward seeking and vS^LH^ in anxiety ([@bib11], [@bib17], [@bib22]). Therefore, it will be important to elucidate the role of RE input in these vS-related behaviors, especially given the key role of this circuit in preclinical models of disorders such as schizophrenia, depression, and Alzheimer's disease ([@bib19], [@bib21], [@bib47]). Limitations of the Study {#sec3.4} ------------------------ Although conditional rabies tracing is now a widely used technique that enables systematic brain-wide mapping of synaptic input ([@bib6], [@bib8], [@bib18], [@bib37], [@bib3], [@bib44]), it is important to note the many caveats to this technique when considering the biological meaning of such tracing data. For instance, the mechanism of transsynaptic retrograde spread is almost completely unknown ([@bib28]). This raises potential confounds; for example, the efficiency of rabies viral spread may depend on uncontrolled variables such as tropism for certain cell types and synapses ([@bib28]), the level of upstream circuit activity ([@bib7]), or even non-synaptic transfer ([@bib28]). In addition, there is conflicting evidence as to whether the quantity of rabies-labeled input neurons correlates with functional synaptic strength ([@bib28]). Although the number of input neurons has been shown to match the connection probability and synaptic strength in multiple studies across many brain regions ([@bib27], [@bib43], [@bib26]), other studies did not reach similar conclusions and showed marked divergence ([@bib40], [@bib15]). To mitigate the potential shortcomings of the technique, it is therefore important to conduct complementary experiments and analyses. One method is to combine findings from rabies tracing experiments with analysis of publicly available anterograde tracing experiments, such as those from the ABA Mouse Connectivity Atlas ([@bib8]). For example, we used this approach to confirm direct input to vS from MPO, and also the spatial targeting of RE and PVT input ([Figure S3](#mmc1){ref-type="supplementary-material"}). A second, more powerful method is to combine tracing with follow-up experiments that directly address synaptic strength, such as CRACM ([@bib27]). We used this approach to confirm that the RE input is stronger at synapses onto vS^NAc^ and vS^LH^ and weaker onto vS^PFC^ ([Figure 4](#fig4){ref-type="fig"}). In this instance, the functional strength of input correlated with the number of rabies-labeled inputs, but this is not always guaranteed. Finally, there are other technical considerations that are important to highlight with the use of whole-brain mapping. First, the plane of sectioning is an important consideration and should be motivated by the purpose of the experiment. This is because there is potential for reduced sampling resolution across the axis of slicing, although error associated with this was recently shown to be minimal when detailed registration is carried out ([@bib20]). In our study, we used sagittal sectioning in order to allow accurate estimates of AP position within the brain ([@bib48], [@bib38], [@bib5]), but both horizontal and coronal sections have also been repeatedly and successfully used ([@bib9], [@bib11], [@bib43], [@bib6], [@bib49]). In addition, the use of stereotaxic injections means that despite small injection volume and controlled rate of injection, we cannot completely exclude the possibility that small amounts of virus or tracer might have leaked into neighboring regions. Finally, the success of tracing experiments is dependent on the efficiency of the tracing system, and novel variants of the rabies system, such as more efficient glycoproteins ([@bib24]), or more stable rabies virus ([@bib36], [@bib16], [@bib14]) could enhance the power of whole-brain anatomy experiments in the future. Overall, our study has revealed a basis for the selective control of vS projection neurons through the biased organization of brain-wide input connectivity. Further work will be required to comprehensively delineate the functional connectivity and behavioral relevance of these dedicated circuits in the execution of adaptive behavior. STAR★Methods {#sec4} ============ Key Resources Table {#sec4.1} ------------------- REAGENT or RESOURCESOURCEIDENTIFIER**Bacterial and Virus Strains***pAAV2-retro-CAG-Cre*[@bib45], UNC Vector CoreN/A*pENN-AAV-hSyn-Cre-WPRE-hGH*a gift from James M. WilsonAddgene viral prep \# 105553-AAV1; <http://addgene.org/105553>; RRID:Addgene_105553*pAAV-synP-FLEX-splitTVA-EGFP-B19G*a gift from Ian WickershamAddgene viral prep \# 52473-AAV1; <http://addgene.org/52473>; RRID:Addgene_52473*RabiesΔG-EnvA-H2B-mCherry-2A-CLIP*a gift from Marco Tripodi, MRC LMBN/A*pAAV-hSyn-hChR2(H134R)-EYFP*a gift from Karl DeisserothAddgene viral prep \# 26973-AAV1; <http://addgene.org/26973>; RRID:Addgene_26973**Chemicals, Peptides, and Recombinant Proteins**Red retrobeadsLumafluorItem \#: R180Green retrobeadsLumafluorItem \#: G180Cholera Toxin Subunit B (Recombinant), Alexa Fluor 647 ConjugateInvitrogenCat \#C34778Cholera Toxin Subunit B (Recombinant), Alexa Fluor 555 ConjugateInvitrogenCat \#C34776ProLong Gold Antifade Mountant with DAPIInvitrogenCat \#P36930ProLong Glass Antifade Mountant with DAPIInvitrogenCat \#P36984IsofluranePiramal Critical CareCarprofenNorbrookTetrodotoxinHello BioCat \#HB10354-aminopyridineHello BioCat \#HB1073**Experimental Models: Organisms/Strains**Mouse: C57BL/6Charles RiverN/A**Software and Algorithms**ImageJ (Fiji) Software<https://fiji.sc/>N/APython 3.6<https://www.python.org/>N/AJupyter Notebook<https://jupyter.org/>N/AR<https://www.r-project.org/>N/AWholeBrain<http://www.wholebrainsoftware.org/>[@bib20]N/A Lead Contact and Materials Availability {#sec4.2} --------------------------------------- Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Andrew MacAskill (<a.macaskill@ucl.ac.uk>). This study did not generate new unique reagents. Experimental Model and Subject Details {#sec4.3} -------------------------------------- ### Mice {#sec4.3.1} Young adult C57BL/6 male mice (CTXβ and rabies tracing: at least 7 weeks old; physiology: 7 -- 9 weeks old) provided by Charles River were used for all experiments. All animals were housed in cages of 2 to 4 in a temperature- and humidity-controlled environment with a 12 h light-dark cycle (lights on at 7 am to 7 pm). Food and water were provided *ad libitum*. All experiments were approved by the UK Home Office as defined by the Animals (Scientific Procedures) Act, and strictly followed University College London ethical guidelines. Method Details {#sec4.4} -------------- ### Stereotaxic surgery {#sec4.4.1} Stereotaxic surgeries were performed according to previously described protocols ([@bib13]). Mice were anaesthetised with isoflurane (4% induction, 1.5 to 2% maintenance) and secured onto a stereotaxic apparatus (Kopf). A single incision was made along the midline to reveal the skull. AP, ML and DV were measured relative to bregma, and craniotomies were drilled over the injection sites. Long-shaft borosilicate pipettes were pulled and backfilled with mineral oil, and viruses were loaded into the pipettes. Viruses were injected with a Nanoject II (Drummond Scientific) at a rate of 13.8 nL every 10 s. Following infusion of the virus, the pipette was left in place for an additional 10 mins before being slowly retracted. The following coordinates (in mm) were used:Injection siteMLAPDVMedial prefrontal cortex:0.4+2.3−2.4Lateral hypothalamus:0.9−1.3−5.2Nucleus accumbens (medial shell):0.9+1.1−4.6Ventral subiculum (anterior):3.4−3.2−4.3Ventral subiculum (posterior):3.4−3.7−4.3Nucleus reuniens0.25−0.7−4.4 Following injection of substances into the brain, animals were sutured and recovered for 30 mins on a heat pad. Animals received carprofen as a peri-operative s.c. injection (0.5 mg / kg) and in their drinking water (0.05 mg / mL) for 48 hours post-surgery. The titers of viruses used are the following:•*RabiesΔG-EnvA-H2B-mCherry-2A-CLIP,* 1.8 × 10^8^ genome copies (gc) / mL;•pAAV-synP-FLEX-splitTVA-EGFP-B19G, 3.9 × 10^12^ gc / mL;•pENN-AAV-hSyn-Cre-WPRE-hGH, 1.3 × 10^13^ gc / mL;•pAAV-hSyn-hChR2(H134R)-EYFP, 2.5 × 10^13^ gc / mL;•pAAV2-retro-CAG-Cre, 5.3 × 10^12^ gc / mL ### Retrograde tracing {#sec4.4.2} For CTXβ retrograde tracing, 150 nL of Alexa 555- or 647-tagged CTXβ was injected into one of three output regions (PFC, NAc or LH). After at least 14 days post-surgery, animals were sacrificed for histology. For rabies monosynaptic tracing, experiments were done according to previously described protocols ([@bib6]). Adult male mice were injected with 100 nL of *AAV2-retro-CAG-Cre* or *AAV2-retro-synP-Cre* into one of three output regions (PFC, NAc or LH), and in the same surgery, 250 nL of *AA2/1-synP-FLEX-split-TVA-EGFP-B19G* was injected into ventral subiculum (anterior) or ventral subiculum (posterior). In a second surgery after at least 2 weeks, 300 -- 400 nL of *EnvA-RVΔG-H2B-mCherry* was injected into vS. After 7 days of rabies expression, animals were sacrificed for histology. ### Histology and imaging {#sec4.4.3} Animals were deeply anaesthetised with a lethal dose of ketamine and xylazine (100 mg/kg) and perfused transcardially with phosphate-buffered saline (pH = 7.2) followed by 4% paraformaldehyde. Brains were dissected and post-fixed overnight at 4°C prior to sectioning. For CTXβ tracing analysis, 60-μm sections in the horizontal plane were prepared using a vibratome (Campden Instruments). For brain-wide rabies tracing analysis, 60-μm sections in the sagittal plane were prepared with a supporting block of agar, and every 2^nd^ section was mounted and analyzed. Sections were mounted on Superfrost Plus slides with ProLong Gold or ProLong Glass antifade mounting medium (Molecular Probes) and imaged with a 5x objective using a Zeiss Axioscan Z1 or 10x objective using a Zeiss SLM 800, using standard filter sets for excitation/emission. Tiled images were stitched automatically during image acquisition by Zen software to generate the full images displayed. ### Whole-hippocampus cell quantification and analysis {#sec4.4.4} For quantification of CTXβ-labeled hippocampal neurons, consecutive 60-μm coronal sections spanning the hippocampus (approximately AP −4.3 mm to −1.0 mm) were collected. Cell counting of retrogradely labeled neurons was conducted using custom written scripts in R based around the WholeBrain package ([@bib20]), a recently developed automatic segmentation and registration workflow in R. Only sections containing labeled neurons in the hippocampus (∼-4.1 mm to −2.7 mm) were analyzed. Segmentation was performed using wavelet multiresolution decomposition on the WholeBrain platform in R, and the segmentation parameters (pixel threshold, soma size, brain outline etc.) were adjusted slice by slice to achieve accurate segmentation of neurons. For the registration of coronal sections, six to eight random brain sections were sampled and manually annotated with approximate AP coordinates, and the remaining sections were assigned AP coordinates based on interpolation. For a given experiment, two projection-specific hippocampal populations (pairs of NAc, LH or PFC) were labeled with two fluorophores (Alexa Fluor 555 and Alexa Fluor 647) in a counterbalanced order. First, segmentation was conducted for each channel individually. Subsequently, registration was conducted on one image data from one channel in a semi-automated manner; this involved an automatic registration of the tissue section by the WholeBrain software, and then refining the registration through manually adding, subtracting or changing correspondence points at clear anatomical landmarks. The same anatomical registration was applied to register the tissue section image obtained from the other channel. Cell count data were saved as RData files, imported into Python and analyzed using Python 3.6. For spatial distribution, correlation and logistic regression analyses, individual sections were manually assigned AP coordinates by first estimating the posterior-most AP coordinate using the ABA as a reference, and then labeling the next anterior coronal section at 60-μm increments until the last section in the coronal section series. ML and DV positions were calculated through registration of individual neurons to the ABA (see above). Only hippocampal cells in subiculum ('SUB') and CA1 residing in ventral hippocampus (∼DV −3.5 and −4.5 mm) were analyzed. Cell distributions across AP, ML and DV were determined by kernel density estimation using a Gaussian kernel function and bandwidth estimated by Scott's rule. Correlation analysis was conducted to determine which brain axis covaries most with the projection type of hippocampal neurons. For each pair of injections in a given hemisphere, the covariation of spatial position along each axis with projection type was computed using Pearson correlation using the *scipy* function *scipy.stats.pointbiserialr*. The absolute value of the correlation coefficient indicates the degree of covariation, where \|r\| = 1 indicates perfect correlation, and \|r\| = 0 indicates no correlation; the absolute correlation coefficient was then compared across the different brain axes. To identify which axis contributes most information to predict the projection class of hippocampal neurons, we conducted logistic regression analysis using the *sklearn* package in Python. For this analysis, the total ventral hippocampal cell counts from each hemisphere were divided into 80% training dataset to train the multinomial logistic regression model, and 20% test dataset to examine the performance of the model. The train-test-split was crucial to assess how well the linear classifiers generalized to unseen data. The 10-fold cross-validated accuracy of the models was calculated using the *LogisticCV* function, and the classifiers were assessed for how much their performance degraded after removal of each brain axis (AP, ML or DV) as a predictor in the model. This reduction in accuracy after the removal of one predictor (cv. Δ accuracy) indicates the unique information that spatial position along one axis contributes to their performance. Note that 2 hemispheres out of 12 total hemispheres from 6 animals were excluded from the dataset due to poor hippocampal labeling. ### Analysis of spatial positions of vS projections along the PD axis {#sec4.4.5} For a subset of CTXβ injected brains, brain sections were prepared in the horizontal plane to analyze the spatial positions of retrogradely labeled vS projections to PFC, NAc and LH along the PD axis. Six to eight sections spanning vS (approximately DV −3.5 to −4.5 mm) were analyzed per brain hemisphere. Using ImageJ, horizontal sections of the hippocampus were digitally straightened from the dentate gyrus to the end of subiculum using the *Straighten* function on ImageJ to approximate the PD axis. Labeled cells in each slice were manually counted and registered to this axis using the ImageJ CellCounter plugin. The radial (y-coordinate) and PD (x-coordinate) positions of each cell were used for spatial position analysis. The DV positions of each cell was manually annotated for each hippocampal slice based on the Paxinos atlas. The CA1/subiculum border occurs approximately at 0.7 within this normalized PD axis range and was anatomically defined as the disappearance of stratum oriens and the fanning out of the pyramidal cell layer. The spatial positions were analyzed with custom Python routines. ### Linear discriminant analysis (LDA) of registered hippocampal neurons {#sec4.4.6} LDA was used to unbiasedly identify the plane which most optimally separates the clusters of CTXβ-labeled, projection-specific hippocampal neurons. LDA is a dimensionality reduction method that identifies the subspace which maximizes the ratio of the between-class over the within-class variability. The between- and within-class variability were calculated as scatter matrices S~B~ and S~W~, respectively, where S~B~ and S~W~ are 3x3 matrices and the number of rows or columns corresponds to each brain axis (AP, ML and DV).The predictor variables for the LDA analysis were the registered spatial positions in AP, ML and DV, and the target variable was the projection type (encoded labels of NAc, PFC or LH). We focused on hippocampal cells labeled between DV −4.5 mm and −4.0 mm, and this dataset (n = 14,527 ventral hippocampal neurons counted from 10 experiments; 6 animals) was split into an 80% training and 20% held-out test dataset. The projection matrix used to transform the retrogradely labeled neurons to the subspace that best maximizes discriminability was solved by matrix factorisation using singular value decomposition (SVD) based on the *LinearDiscriminantAnalysis* function from the *sklearn* package. The spatial positions of the held-out test dataset were then projected onto the subspace by matrix multiplication of the spatial position and projection matrices:$$X' = X\varphi$$where $\varphi$ is the eigenvector (projection) matrix whose columns correspond to the eigenvectors (linear discriminant vectors), and *X* is the matrix of spatial positions of each registered neuron. The transformed spatial positions were plotted in the LD subspace, and the cell distributions in the first and second linear discriminants were determined by kernel density estimation. ### Mapping and analysis of rabies-labeled inputs {#sec4.4.7} Cell counting of rabies-labeled inputs was conducted using WholeBrain ([@bib20]). After acquiring the imaged sections and exporting them as 16-bit depth image files, images in the rabies mCherry channel were manually assigned a bregma coordinate (ML −4.0 to 0.0 mm) and processed using WholeBrain ([@bib20]) and custom cell counting routines written in R. The workflow comprised of (1) segmentation of cells and brain section, (2) registration of the cells to the ABA and (3) analysis of anatomically registered cells. As tissue section damage impairs the automatic registration implemented on the WholeBrain platform, sections with poor registration were manually registered to the atlas plate using corresponding points to clear anatomical landmarks. Once all cells have been registered, the cell counts were further manually filtered from the dataset to remove false-positive cells (e.g., debris). Virtually all cells were detected in the injected hemisphere, apart from a consistent set of contralateral CA3 inputs ([@bib9]). Therefore, only the injected hemisphere up to the midline was used for cell quantification analysis. Each cell registered to a brain region was classified as belonging to an anatomically defined region as defined by the ABA brain structure ontology. Information on the ABA hierarchical ontology was scraped from the ABA API (link: <http://api.brain-map.org/api/v2/structure_graph_download/1.json>) using custom Python routines. The coarse-level (or parent) structures to which each cell belonged were defined *a priori* and comprised the following: *Hypothalamus, Isocortex, Hippocampal formation, Thalamus, Cortical subplate, Pallidum, Striatum, Midbrain, Pons* and *Medulla.* All cells falling under these parent structures were analyzed. The fine-level (or child) structures represent all brain regions existing as subcategories of the corresponding coarse-level structure, e.g., *nucleus reuniens* and *paraventricular thalamus* are fine-level (child) structures relative to *Thalamus* (parent). For quantification of input fractions, cells residing in different layers within the same structure, e.g., CA1 stratum oriens (CA1so) and stratum lacunosum-moleculare (CA1slm), or subdivisions of nuclei, e.g., basomedial amygdala, posterior division (BMAp) and anterior division (BMAa), were agglomerated across layers and subdivisions and counted as residing in one single region (CA1 and BMA, respectively). Note that lateral and medial entorhinal cortex (LEC and MEC, respectively) were analyzed as separate structures. ### Starter cell center of mass (COM) quantification {#sec4.4.8} To determine the starter cell COM, every 2^nd^ sagittal section from both the rabies mCherry+ and TVA-G GFP+ channels that spanned the extent of the TVA-G GFP+ expression in vS was obtained and analyzed. Images were collected in order from lateral to medial. Colocalized cells representing starter cells were manually counted and registered onto digital plates from the Paxinos atlas using the ImageJ ROI Manager function. Cell positions in AP and DV were calculated by first obtaining the spatial position of cells in pixel space. Then, the location of bregma in pixel space was determined by placing an ROI at the approximate site of bregma in the Paxinos atlas. Finally, a scaling factor was calculated by estimating the change in pixel space with a unit change in ML and DV. The final positions of cells were calculated by offsetting the positions of the cells in pixel space to that of bregma and normalizing the positions by the scaling factor. As the TVA-G construct expressed TVA and G bicistronically (i.e., the exon for TVA and G are linked by a self-cleaving 2A peptide), all cells were assumed to express TVA and G in a 1:1 stoichiometry and unlikely to express one gene without the other. Therefore, all colocalized cells were treated as starter cells. ### Analysis of COM versus projection dependence of inputs {#sec4.4.9} Inputs arising from within the hippocampal formation were analyzed as input fractions of the total inputs counted in the same brain, while inputs arising from outside the hippocampal formation were analyzed as input fractions normalized to the total number of extrahippocampal inputs. The dataset containing cell counts from n = 20 brains were analyzed according to projection or COM. For extrahippocampal inputs, only fine-level structures exceeding 1% of extrahippocampal input were assessed (15 brain regions). Multiple linear regression modeling was performed to compare the relative influence of COM and projection identity on the amount of rabies-labeled inputs in an input region of interest. For each input brain region, a multiple regression model -- in which the predictor variables were the COM and projection identity -- was constructed using the *ols* function from the *statsmodel* package in Python. The overall statistical significance of full models (containing COM and projection as predictors) were assessed by computing the F-statistic values. All *p* values generated from multiple comparisons were corrected using the Benjamini-Hochberg method (false-discovery rate \< 0.05). The models with adjusted p values \< 0.05 were further analyzed for statistically significant coefficients using the Wald test and followed up with post hoc pairwise Tukey multiple comparisons between vS projection populations (see [Figures S3](#mmc1){ref-type="supplementary-material"}C and [3](#fig3){ref-type="fig"}F). To further assess the importance of each predictor to the model, the likelihood ratio (LR) test was used to compare the full model to a reduced model containing only one predictor -- either COM or projection. The reduced models containing either COM or projection identity as a predictor were built using the same *ols* function. The LR test was computed using the following formula from the *statsmodel* package function *compare_lr_test:*$$D = - 2\ \log\left( \frac{\mathcal{L}_{restricted}}{\mathcal{L}_{full}} \right)$$where $\mathcal{L}$ is the likelihood of the model, and D is a test statistic that follows a $\chi^{2}$ distribution with degrees of freedom (df) equal to the difference in the number of predictors between the full and restricted (reduced) model. The p values generated from multiple LR tests were corrected using the Benjamini-Hochberg method (FDR \< 0.05). As complementary analyses, the projection-dependence of extrahippocampal inputs from [Figure 3](#fig3){ref-type="fig"}F was further analyzed using multiple one-way ANOVAs, while the spatial-dependence of these inputs from [Figure 3](#fig3){ref-type="fig"}G was assessed using multiple Spearman rank correlation tests of input fractions against starter cell COM. This analysis revealed similar patterns of biased connectivity to the multiple linear regression analysis ([Table S2](#mmc1){ref-type="supplementary-material"}), where MPO and PA were detected as projection-dependent inputs, but RE was detected as being COM-dependent. For spatial-dependence analysis, the input density for each brain region as a function of COM was visualized by first sorting the input fractions of each brain by the COM in the posterior-to-anterior direction. The array of input fractions was then interpolated to produce 500 data points, smoothed with a Savitzky-Golay filter (window size = 51 points, order = 3), and normalized by dividing each data point by the total area under the curve. To assess the relative goodness-of-fit of non-nested *COM models* or *projection models*, we performed linear regression models with only one predictor by using the *ols* function from the Python package *statsmodels*. Linear models were built for each brain region, where the target variable was the input fraction observed in that brain region normalized to the total number of inputs, and the predictor variable was either the COM or projection. For each of 35 extrahippocampal inputs in total, there were 20 observations (n = 20 brains). The models were fitted, and goodness-of-fit was measured using the Bayesian information criterion (BIC). The BIC was computed using the following formula:$$BIC = k\ \mspace{9mu}\log\ n - 2\ \log\widehat{\mathcal{L}}$$where *k* is the number of model parameters, *n* is the observation number and $\widehat{\mathcal{L}}$ is the maximum likelihood of the model. Finally, to compare model fits, the difference between the BIC values obtained from COM and projection models were computed to obtain ΔBIC. ### ABA Mouse Brain Connectivity analysis of axonal projections {#sec4.4.10} To validate the spatial targeting of rabies-labeled inputs as predicted by the multiple linear regression analysis, analysis of axonal input density arising from the input region 'hits' (i.e., MPO, RE, PA and PVT) was conducted. For MPO, PVT and RE, three separate experiments were analyzed. For PA, the only two experiments available were analyzed. All images were downloaded from the ABA API using the Python package Allen SDK (2015). The following experiments were used:InputExperiment IDTransgenic lineImage sections used for analysisMPO158738180*C57BL/6*158738374, 158738370, 158738366, 158738362, 158738358, 158738354119846838*C57BL/6*119847348, 119847344, 119847340, 119847336, 119847332, 119847328182459635*Gal-Cre_KI87*182460001, 182459997, 182459993, 182459989, 182459985, 182459981RE204832205*Adcyap1-2A-Cre*204832574, 204832570, 204832566, 204832562, 204832558, 204832554175374982*C57BL/6*175375180, 175375176, 175375172, 175375168, 175375164, 175375160174957972*C57BL/6*174958286, 174958282, 174958278, 174958274, 174958270, 174958266PA304721447*Rbp4-Cre_KL100*304721806, 304721802, 304721798, 304721794, 304721790, 304721786545415593*Npr3-IRES2-Cre*545415944, 545415940, 545415936, 545415932, 545415928, 545415924PVT278510903*Ppp1r17-Cre_NL146*278511259, 278511255, 278511251, 278511247, 278511242, 278511238183225830*Grm2-Cre_MR90*183226060, 183226056, 183226052, 183226048, 183226044, 183226040301209502*Efr3a-Cre_NO108*301209844, 301209840, 301209836, 301209832, 301209828, 301209824 For each experiment, six consecutive coronal images (spanning approximately −3.7 to −3.1 mm AP relative to bregma) of ventral hippocampus were downloaded and analyzed using ImageJ. Only one hemisphere per brain was analyzed. The brightness and contrast values were fixed for all sections obtain from a single experiment to allow comparison of the relative fluorescence of axonal projections across slices. Images were downloaded with a downsample factor of 8, ROIs were manually drawn around the vS (defined as the region below the rhinal sulcus, between the alveus and the boundary of the CA1 stratum lacunosum moleculare and adjacent dentate gyrus or CA3 stratum radiatum border, see also [Figures S3](#mmc1){ref-type="supplementary-material"}F and S3G), and the mean pixel intensity within the ROIs were measured. The pixel intensity values were normalized to that from the posterior-most section, and the relative fluorescence intensities were compared across the AP axis. ### CRACM experiments {#sec4.4.11} #### Surgeries {#sec4.4.11.1} CRACM of projection-defined vS neurons was done according to previously described protocols ([@bib29]). 7- to 9-week old animals were injected with 250 nL of red (1:10 dilution in sterile saline) or green (undiluted) retrobeads in a counterbalanced order into two of three output regions (PFC, NAc or LH). In the same surgery, 250 nL of *AAV1-synP-ChR2-YFP* was injected into RE bilaterally. After at least 14 days of ChR2 expression, animals were sacrificed for electrophysiological recording. #### Slice preparation {#sec4.4.11.2} Acute, transverse hippocampal slices were used for all electrophysiological recordings. Mice were deeply anaesthetised with a lethal dose of ketamine and xylazine (100 mg / kg), and perfused transcardially with ice-cold sucrose solution containing (in mM): 190 sucrose, 25 glucose, 25 NaHCO~3~, 1.2 NaH~2~PO~4~, 10 NaCl, 2.5 KCl, 1 Na^+^ ascorbate, 2 Na^+^ pyruvate, 7 MgCl~2~ and 0.5 CaCl~2,~ bubbled continuously with 95% O~2~ / 5% CO~2~. Following perfusion, mice were decapitated, and their brains were rapidly dissected. The dissected brains were then placed in ice-cold sucrose solution and hemisected. The cerebellum was removed, and transverse slices were prepared using a vibratome (VT1200S, Leica), with a ∼10° angle along the ventromedial plane to obtain sections that were orthogonal to the long-axis of the hippocampus. The thickness of hippocampal sections was 300 μm. Slices were transferred to a bath containing artificial cerebrospinal fluid (aCSF) and recovered first for 30 mins at 37°C, and subsequently for 30 mins at room temperature. The aCSF solution contained (in mM): 125 NaCl, 2.5 KCl, 1.25 NaH~2~PO~4~, 22.5 glucose, 1 Na^+^ ascorbate, 3 Na^+^ pyruvate, 1 MgCl~2~, 2 CaCl~2~. All recordings were performed at room temperature (22 -- 24°C). All chemicals were from Sigma or Tocris. #### Whole-cell electrophysiology and optogenetics {#sec4.4.11.3} Whole-cell recordings were performed on retrogradely labeled hippocampal pyramidal neurons with retrobeads visualized by their fluorescent cell bodies and targeted with Dodt contrast microscopy. For sequential paired recordings, neighboring neurons were identified using a 40x objective at the same depth into the slice. The recording order of neuron pairs was alternated to avoid complications due to rundown. Borosilicate recording pipettes (3 -- 5 MΩ) were filled with a Cs-gluconate internal solution containing (in mM): 135 Gluconic acid, 10 HEPES, 10 EGTA, 10 Na-phosphocreatine, 4 MgATP, 0.4 Na~2~GTP, 10 TEA and 2 QX-314. Presynaptic glutamate release was elicited by illuminating ChR2 expressed in the presynaptic terminals of long-range inputs into the slice, as previously described ([@bib29]). Wide-field illumination was achieved through a 40x objective with brief 10 ms pulses of blue light over the recorded neuron from an LED centered at 473 nm (CoolLED pE-4000, with corresponding excitation-emission filters). Light intensity was measured as 4--7 mW at the back aperture of the objective and was constant between all recorded cell pairs. When recording a cell pair, the LED power was adjusted until responses were ∼200 pA in a connected neuron or set to maximum in an unconnected neuron, and fixed at the same level when recording light-evoked responses in the other cell. In all experiments, the aCSF contained 1 μM TTX and 100 μM 4-AP to isolate monosynaptic connectivity and increase presynaptic depolarisation, respectively. Recordings were conducted using a Multiclamp 700B amplifier (Axon Instruments), and signals were low-pass filtered using a Bessel filter at 1 kHz and sampled at 10 kHz. Data were acquired using National Instruments boards and WinWCP (University of Strathclyde) and analyzed using custom routines written in Python 3.6. For synaptic connectivity analysis, six recording sweeps were obtained for each optical pulse duration. The signals were preprocessed by baselining the signals to the first 100 ms, low-pass filtering using a Bessel filter (cutoff = 2 Hz, order = 2), averaging across the six recording sweeps for each optical pulse duration, and decimating the averaged signal to 1 kHz. The peak amplitude response of light-evoked EPSCs was measured as averages over a 2-ms time window around the peak compared to a 2-ms baseline period preceding the optical pulse. Only paired data in which at least one cell received \> 5 pA were included for analysis for [Figures 4](#fig4){ref-type="fig"}D--4K, while all cells irrespective of connectivity were included for the analysis described in [Figure S4](#mmc1){ref-type="supplementary-material"}. We note that due to the jitter in release as a result of the use of presynaptic ChR2, across-cell averages of postsynaptic currents ([Figures 4](#fig4){ref-type="fig"}D--4K) may appear relatively slow. For each recorded neuron, the spatial position was obtained by manually registering the cell to a digital atlas depicting a horizontal section of the vS. The long axis of the CA1 and subiculum area (i.e., the y-coordinates) of each registered cell was used to approximate the anterior-posterior position. Input connectivity was determined by a threshold light-evoked response of \> 5 pA. Quantification and Statistical Analysis {#sec4.5} --------------------------------------- All statistics were calculated using Python's *scipy* and *statsmodels* packages, and R. Summary data are reported throughout the text and figures as mean ± sem unless otherwise stated. For [Figures 1](#fig1){ref-type="fig"} and [4](#fig4){ref-type="fig"}, normality of data distributions was determined by visual inspection of the data points. For [Figures 3](#fig3){ref-type="fig"} and [S2](#mmc1){ref-type="supplementary-material"}, the rabies tracing dataset was assessed for normality with the Jarque-Bera test and equal variance with the Levene's test. The input fractions for multiple regions showed non-normality and heteroscedasticity in the input fractions across projection populations. Therefore, all analysis for [Figure 3](#fig3){ref-type="fig"} were conducted after log-transformation of input fractions. Distributions of the data became more Gaussian-like after log-transformation, as assessed again by the Jarque-Bera test and Levene's test. All data were analyzed using statistical tests described in the figure legends and in [Table S2](#mmc1){ref-type="supplementary-material"}. The alpha level was defined as 0.05. No power analysis was run to determine sample size *a priori*. The sample sizes chosen are similar to those used in previous publications. Data and Code Availability {#sec4.6} -------------------------- Programme code, original data and scripts for statistical tests and analysis are available upon request. Raw data are available at <https://github.com/macaskill-lab>. Supplemental Information {#app2} ======================== Document S1. Figures S1--S4 and Tables S1 and S2Data S1. Raw Cell Counts in Each Condition of the TRIO Experiments (vS^PFC^, vS^NAc^, and vS^LH^ Projections with Associated COM Information), Related to Figure 3Data S2. Abbreviations of the 115 Brain Regions with Raw Cell Counts Shown in Data S1, Related to Figure 3Document S2. Article plus Supplemental Information We thank Marco Tripodi and Fabio Morgese for the unpublished pseudotyped rabies virus, Francesca Cacucci for assistance with surgery, Harsha Gurnani for discussions about LDA, and Rachel Wee for help with illustrations. We are grateful to Larry Swanson, Andy Murray, Tara Keck, Kenneth Harris, and the MacAskill lab for helpful comments on the manuscript. A.F.M. was supported by a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society (grant number 109360/Z/15/Z) and by a UCL Excellence Fellowship. R.W.S.W. was supported by a UCL Graduate Research Scholarship and a UCL Overseas Research Scholarship. Author Contributions {#sec5} ==================== Conceptualization, R.W.S.W. and A.F.M.; Methodology, R.W.S.W. and A.F.M.; Investigation, R.W.S.W. and A.F.M.; Formal Analysis, R.W.S.W. and A.F.M.; Writing -- Original Draft, R.W.S.W. and A.F.M.; Writing -- Review & Editing, R.W.S.W. and A.F.M.; Funding Acquisition, A.F.M.; Supervision, A.F.M. Declaration of Interests {#sec6} ======================== The authors declare no competing interests. Supplemental Information can be found online at <https://doi.org/10.1016/j.celrep.2020.02.093>. [^1]: Lead Contact

{ "pile_set_name": "PubMed Central" }

SpecificationsOrganism/cell line/tissue*Nicotiana tabacum*Sequencer or array typeIllumina HiSeq™ 2000-RNA sequencingData formatRaw and processedExperimental factors*Wo*^v^ transgenic and wild-type tobaccoExperimental featuresRNA-seq dataset for gene expression profiling in young primary leaf in *Wo*^v^ transgenic and wild-type tobaccoSample source locationWuhan, China 1. Direct link to deposited data {#s0005} ================================ Deposited data can be found here: <http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE72310>. 2. Experimental design, materials and methods {#s0010} ============================================= 2.1. Sample collection and RNA isolation {#s0015} ---------------------------------------- *Wo*^v^ transgenic tobacco plants and their WT species (*N. tabacum*) were grown in a naturally illuminated glasshouse in the experimental field of the Huazhong Agricultural University. Environmental control was implemented based on standard protocols [@bb0005]. Two independent *Wo*^v^ transgenic tobacco plants and two negative regenerated plants were used for transcriptomic analysis. All samples were immediately frozen in liquid nitrogen, and total RNA was isolated using TRIzol reagent (Invitrogen). The integrity of all RNA samples was examined by 1.2% agarose gel electrophoresis after treatment with RQ1 DNase (Promega). The quality and quantity of these RNA samples were further determined by measuring the absorbance at 260/280 and 260/230 nm by using SmartSpec Plus spectrophotometer (Bio-Rad). 2.2. RNA-seq and data processing {#s0020} -------------------------------- The materials used for RNA-seq analyses were three young primary leaves from the transgenic and control plants. Up to 10 μg of total RNA was sent to ABlife Inc. (Wuhan, China) where the libraries were produced. The cDNA libraries were then sequenced in BGI Inc. (Shenzhen, China) by using Illumina HiSeq™ 2000 by 100 nt pair-end sequencing. RNA-seq was performed as previously described [@bb0010]. Whole RNA-seq data were submitted to the Gene Expression Omnibus (series accession number [GSE72310](ncbi-geo:GSE72310){#ir0015}). The clean reads were aligned to the *N. tabacum* genome (<ftp://anonymous@ftp.solgenomics.net/genomes/Nicotiana_tabacum/assembly/Ntab-K326_AWOJ-SS.fa.gz>) using TopHat (2.0.12) software [@bb0015]. Based on the length of the gene and read counts uniquely mapped to this gene, gene expression levels were calculated using FPKM method. 2.3. Functional analysis of differentially expressed genes {#s0025} ---------------------------------------------------------- We used the DESeq software to identify the differentially expressed genes between samples, which were specific for differential expression analysis of the RNA-Seq data with biological replicates [@bb0020]. Genes with P-value ≤ 0.05 and fold change ≥ 2 were regarded as differentially expressed. To characterize the putative functions of differentially expressed genes, Gene Ontology (GO) term analysis was also performed using GOseq based Wallenius non-central hyper-geometric distribution [@bb0025]. For further identification of pathways significantly influenced by the *Wo*^v^ gene, KEGG enrichment pathways analysis of DEGs was implemented by the KOBAS (2.0) software by a hypergeometric test and the Benjamini--Hochberg FDR correction (FDR ≤ 0.05) [@bb0030]. [^1]: These two authors contributed equally to this work.

{ "pile_set_name": "PubMed Central" }

INTRODUCTION {#s1} ============ Notch signaling is involved in various cellular processes, including cell fate specification, differentiation, proliferation, and apoptosis. Abnormal Notch signaling is oncogenic in several cancers, including hematologic malignancies \[[@R1]\]. Disregulated Notch signaling has also been associated with chronic lymphocytic leukemia (CLL), a frequent adult leukemia characterized by the accumulation of CD19^+^/CD5^+^ B lymphocytes resistant to apoptosis \[[@R2]\]. The involvement of Notch in CLL has been recently demonstrated by findings that a *NOTCH1* PEST domain mutation, generating a highly active truncated protein, and affecting up to 10-15% of patients, is associated with poor prognosis, disease progression and refractoriness to chemotherapy \[[@R3]-[@R7]\]. Previous evidence that constitutive activation of Notch1 and Notch2 signaling contributes to apoptosis resistance in CLL also underscores the importance of Notch in this leukemia \[[@R8], [@R9]\], encouraging further investigation of its therapeutic potential. Indeed, a better understanding of the mechanisms involved in the anti-apoptotic signaling of Notch in CLL cells may provide insight for designing future Notch-targeted therapies. In cancer, Notch signaling prevents apoptosis through different networks, involving cell cycle and survival pathways, and interactions with mitochondria. Notch suppresses p53 \[[@R10]\] or JNK function \[[@R11]\] as well as the expression of the pro-apoptotic proteins Bax, Bim and Noxa \[[@R12], [@R13]\]. Notch increases the activation of the pro-survival PI3K/AKT \[[@R14]\] and NF-kB pathways \[[@R15]\] and the expression of the anti-apoptotic proteins Bcl-2 and Bcl-xL \[[@R12]\], stabilizes the apoptosis inhibitor protein XIAP \[[@R16]\] and induces mitochondrial integrity and functions \[[@R17]\]. Some of these functions are mediated by the transcriptional activity of Notch-intracellular domain (ICD) \[[@R10], [@R14], [@R15]\], which, after Notch-ligand interactions, is released from the membrane to the nucleus. Other functions involve a non-canonical Notch-ICD-activated signaling which operates in the cytoplasm \[[@R11], [@R16]\] and can also converge on the mitochondria by promoting cell survival \[[@R17]\]. A crucial role in controlling mitochondrial integrity and apoptosis is played by the balance between pro-apoptotic and anti-apoptotic Bcl-2 family members \[[@R18]\]. In CLL as well as in other hematologic malignancies, the over-expression of the anti-apoptotic Mcl-1 and Bcl-2 proteins is one of the major causes of apoptosis resistance \[[@R19], [@R20]\], poor prognosis \[[@R21], [@R22]\] and chemoresistance \[[@R23]-[@R25]\]. Mcl-1 and Bcl-2 bind and sequester the pro-apoptotic proteins Bax and Bak blocking their ability to form pores in the mitochondrial membrane with the consequent release of cytochrome c into the cytoplasm. Degradation of Mcl-1 frees Bax and Bak allowing their polymerization and activating apoptosis \[[@R18]\]. Mcl-1 is a short-lived protein tightly regulated by transcriptional \[[@R26], [@R27]\], translational \[[@R28], [@R29]\], and degradation mechanisms \[[@R30], [@R31]\]. Interestingly, Mcl-1 mRNA translation is highly dependent on the eukaryotic initiation factor 4E (eIF4E), a key component of the mRNA cap-binding complex, which preferentially enhances translation of a subset of mRNAs with complex 5^\'^ untranslated regions, such as those of Mcl-1 and several other transformation-related and survival proteins \[[@R32]-[@R34]\]. eIF4E has been associated with cancer development and progression, and proposed as an important therapeutic target \[[@R35], [@R36]\]. Recently, it has been demonstrated that CLL cells also express high levels of eIF4E, and that its pharmacologic targeting increases *in vitro* fludarabine cytotoxicity, suggesting an involvement of eIF4E in chemoresistance of these cells \[[@R37]\]. Considering the critical role of Mcl-1 and eIF4E in CLL and in other malignancies, in this study, we investigated whether Mcl-1 and eIF4E are targets of the anti-apoptotic Notch signaling in CLL. RESULTS {#s2} ======= Notch1 and Notch2 downregulation decreases viability of CLL cells from different patient subgroups {#s2_1} -------------------------------------------------------------------------------------------------- We used small-interfering RNA (siRNA) and nucleofection to silence expression of Notch1 and Notch2 in all 22 patients included in the study. Table [1](#T1){ref-type="table"} gives clinical and biological characteristics of CLL patients. Downregulation of the expression of each Notch receptor, achieved at different levels in all samples examined (Table [2](#T2){ref-type="table"}), did not affect the levels of the other receptor (Figure [1A](#F1){ref-type="fig"}), suggesting that the expression of each of them is independent of the other. As previously reported \[[@R8], [@R9]\] and shown in Table [2](#T2){ref-type="table"} and Figure [1B](#F1){ref-type="fig"}, silencing of either Notch1 (siNotch1) or Notch2 (siNotch2) decreased, to a similar extent, CLL cell viability compared with cells transfected with control siRNA (siCtrl). This effect was observed in 18 of 22 samples (Table [2](#T2){ref-type="table"}), suggesting that both receptors contribute to CLL cell survival in the vast majority of patients (81.8%). The four patients with Notch-independent viability (CLL4, 17, 20, 22) did not belong to any specific subgroup regarding clinical and biological characteristics (Table [1](#T1){ref-type="table"}). Decrease in cell viability induced by silencing of each receptor varied among the different CLL samples, ranging from 17.7 to 65.2% for Notch1 and from 16.1 to 51.4% for Notch2 (Table [2](#T2){ref-type="table"}). However, similar responses were observed irrespective of Binet stage, previous therapy, *IgVH* mutational status and ZAP70 and CD38 expression (Table [3](#T3){ref-type="table"}), suggesting that Notch targeting is effective in CLL cells despite the presence of adverse prognostic factors. Even in the three samples with *NOTCH1* PEST domain mutation (CLL1, 7, 11; Table [1](#T1){ref-type="table"}), either Notch1 or Notch2 downregulation reduced CLL cell viability at levels similar to those observed in *NOTCH1*-unmutated samples (Table [3](#T3){ref-type="table"}). These results suggest that *NOTCH1* mutation does not influence the sensitivity of CLL cells to Notch targeting, at least when it is harboured by a small fraction of leukemic cells, as indicated by the low *NOTCH1* mutant allele burden detected in all three mutated samples examined (Table [1](#T1){ref-type="table"}). ###### Characteristics of CLL patients ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Patients Binet stage Previous treatment^[a](#tfn_001){ref-type="table-fn"}^ *IgVH* status^[b](#tfn_002){ref-type="table-fn"}^ ZAP70 expression^[c](#tfn_003){ref-type="table-fn"}^ CD38 expression^[d](#tfn_004){ref-type="table-fn"}^ *NOTCH1* status(% mutant allele burden)^[e](#tfn_005){ref-type="table-fn"}^ Cytogenetic alterations^[f](#tfn_006){ref-type="table-fn"}^ ----------- ------------- -------------------------------------------------------- --------------------------------------------------- ------------------------------------------------------ ----------------------------------------------------- ----------------------------------------------------------------------------- ------------------------------------------------------------- **CLL1** B no Unm \+ − Mut\ Normal (1.9) **CLL2** C yes Unm \+ − Unm del 11q22-23\ del 13q14 **CLL3** A no Unm \+ ND Unm ND **CLL4** C yes Unm \+ \+ Unm del 11q22-23\ del 13q14 **CLL5** C yes Unm \+ \+ Unm del 13q14\ del 14q32 **CLL6** A no Mut \+ − Unm Normal **CLL7** B yes Unm \+ \+ Mut\ del 11q22-23 (2.4) **CLL8** B yes Unm \+ \+ Unm ND **CLL9** C yes Unm − − Unm ND **CLL10** C yes Mut − − Unm del 13q14\ del 17p13 **CLL11** B yes Unm − − Mut\ del 11q22-23\ (1.3) del 13q14\ del 14q32 **CLL12** C yes Unm \+ − Unm del 11q22-23\ del 13q14 **CLL13** A yes Unm \+ \+ Unm Normal **CLL14** C yes Unm \+ − Unm del 14q32 **CLL15** B no Mut − − Unm Normal **CLL16** C yes Mut \+ \+ Unm ND **CLL17** A no Mut − − Unm del 13q14 **CLL18** A no Mut − − Unm del 13q14 **CLL19** C yes Mut \+ − Unm ND **CLL20** B yes Mut − \+ Unm Normal **CLL21** B yes Unm \+ − Unm del 13q14\ del 14q32 **CLL22** B no Mut \+ − Unm Normal ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Mut, mutated; Unm, unmutated; ND, not determined. Treated patients had not received treatment for at least 3 months before the study. Mutated was defined as having a frequency of mutations \>2% from germline *VH*. Positivity refers to detection of \>20% ZAP70+/CD19+. Positivity refers to detection of \>20% CD38+/CD19+. c.7544_7545delCT in *NOTCH1* exon 34. Assessed by FISH. ###### Effect of Notch1 and Notch2 silencing on CLL cell viability and Mcl-1 protein expression Patients siRNA Receptor \% reduction in Notch expression \% reduction in CLL cell viability \% reduction in Mcl-1 protein expression ---------- ---------- ---------- ---------------------------------- ------------------------------------ ------------------------------------------ ------ CLL 1 siNotch1 Notch1 55.0 67.0 32.6 39.0 siNotch2 Notch2 51.2 61.1 17.9 38.0 CLL 2 siNotch1 Notch1 76.0 53.8 19.8 none siNotch2 Notch2 30.2 38.1 16.2 none CLL 3 siNotch1 Notch1 98.0 99.3 65.2 70.0 siNotch2 Notch2 71.0 64.0 51.4 29.8 CLL 4 siNotch1 Notch1 87.0 69.5 none none siNotch2 Notch2 84.0 64.4 none none CLL 5 siNotch1 Notch1 57.0 80.1 32.7 36.9 siNotch2 Notch2 43.7 65.3 35.1 51.0 CLL 6 siNotch1 Notch1 61.5 52.8 38.1 44.4 siNotch2 Notch2 84 78 43.7 50.9 CLL 7 siNotch1 Notch1 73.2 89.1 29.7 44.2 siNotch2 Notch2 69.0 56.7 23.1 32.6 CLL 8 siNotch1 Notch1 78.2 72.1 29.3 36.0 siNotch2 Notch2 52.0 47.3 24.8 23.2 CLL 9 siNotch1 Notch1 67.0 73.0 33.8 27.7 siNotch2 Notch2 81.0 72.5 39.5 34.0 CLL 10 siNotch1 Notch1 78.0 58.0 25.0 none siNotch2 Notch2 75.0 51.0 28.4 none CLL 11 siNotch1 Notch1 48.0 59.0 20.8 22.9 siNotch2 Notch2 41.0 54.0 38.2 41.7 CLL 12 siNotch1 Notch1 51.0 65.0 31.5 none siNotch2 Notch2 35.3 40.1 17.1 none CLL 13 siNotch1 Notch1 31.0 35.2 17.7 23.0 siNotch2 Notch2 75.7 61.2 16.1 48.0 CLL 14 siNotch1 Notch1 79.0 72.0 26.1 92.0 siNotch2 Notch2 88.0 98.3 20.7 80.0 CLL 15 siNotch1 Notch1 44.4 52.0 58.4 41.9 siNotch2 Notch2 72.7 59.8 34.9 26.0 CLL 16 siNotch1 Notch1 58.0 69.0 43.9 59.4 siNotch2 Notch2 65.0 71.0 36.1 50.0 CLL 17 siNotch1 Notch1 47.1 53.0 none none siNotch2 Notch2 69.0 71.0 none none CLL 18 siNotch1 Notch1 89.0 74.5 23.9 52.9 siNotch2 Notch2 73.9 52.8 31.3 71.6 CLL 19 siNotch1 Notch1 36.0 31.3 18.4 32.0 siNotch2 Notch2 56.4 45.1 20.5 35.4 CLL 20 siNotch1 Notch1 95.3 71.0 none none siNotch2 Notch2 50.7 45.4 none none CLL 21 siNotch1 Notch1 68.3 75.3 33.6 21.0 siNotch2 Notch2 57.3 61.3 44.3 44.0 CLL 22 siNotch1 Notch1 45.6 61.0 none none siNotch2 Notch2 81.9 82.3 none none Expression of Notch1 and Notch2 and Mcl-1 was analyzed by western blot in CLL cells transfected with Notch1 (siNotch1), Notch2 (siNotch2) or control siRNA (siCtrl). The band intensities of Notch1 and Notch2 (TM and IC) and Mcl-1 were quantified by densitometric analysis and normalized to GAPDH. Values represent the percentage reduction of Notch1 and Mcl-1 expression in siNotch1 cells, and of Notch2 and Mcl-1 expression in siNotch2 cells, compared with siCtrl cells. Viable CLL cells were quantified by flow cytometry using Annexin V/PI staining, and the percentage reduction in viability of siNotch1 and siNotch2 cells was calculated relative to siCtrl. None indicates reductions in cell viability or Mcl-1 expression lower than 5%. ![Notch1 and Notch2 silencing decreases cell viability and expression of Mcl-1 protein in CLL cells\ CLL cells were transfected with control siRNA (siCtrl), Notch1 siRNA (siNotch1), Notch2 siRNA (siNotch2) or combined siNotch1 and siNotch2 (siNotch1/2) as described in "siRNA transfection" and then cultured in complete medium for 72 hours. **A.**, **C.** Western blot analysis of Notch1, Notch2, Mcl-1 and Bcl-2 expression was performed on 15 μg whole-cell lysates separated on a 7.5% SDS-PAGE. The antibodies used for Notch1 and Notch2 recognized the 120-kDa transmembrane/cytoplasmic subunit (TM) and the 100-kDa active intracellular domain (IC). Protein loading was assessed by reprobing the blots with an anti-GAPDH antibody. Vertical line inserted in Notch2 blot of CLL7 indicates a repositioned gel lane. Data of CLL7 and 18 are representative of six samples. **D.** The blots of Mcl-1 were subjected to densitometric analysis and densitometry units (U) were calculated relative to GAPDH. Data are the mean ± SD of six samples. \**P* \< 0.05, \*\**P* \< 0.01 (each siNotch transfection condition *versus* siCtrl) according to Student *t* test. **B.** Cell viability was evaluated by flow cytometric analysis of Annexin V/PI (An V/PI) staining. Viability (An V^−^/PI^−^) of siCtrl cells was set to 100%. Data are the mean ± SD of six samples. \*\**P* \< 0.01 (each siNotch transfection condition *versus* siCtrl) according to Student *t* test.](oncotarget-06-16559-g001){#F1} ###### Notch1 and Notch2 silencing decreases CLL cell viability independently of clinical characteristics and prognostic factors Binet stages and prognostic factors Number of patients \% Viability related to siCtrl (mean ± SD)^[a](#tfn_007){ref-type="table-fn"}^ ---------------------------------------------------------- -------------------- -------------------------------------------------------------------------------- ---------------- ------------- ---------------- *Binet stage* A 5 70.5 ± 23.7 0.778 (A vs B) 71.3 ± 20.4 0.594 (A vs B) B 8 73.8 ± 18.2 0.974 (B vs C) 76.9 ± 16.1 0.924 (B vs C) C 9 74.1 ± 11.8 0.707 (C vs A) 76.2 ± 12.2 0.580 (C vs A) *Previous treatment^[b](#tfn_008){ref-type="table-fn"}^* No 7 68.1 ± 24.6 74.2 ± 20.0 Yes 15 75.5 ± 11.4 0.337 75.9 ± 13.1 0.814 *IgVH status^[c](#tfn_009){ref-type="table-fn"}^* Unm 13 71.2 ± 14.1 73.5 ± 14.2 Mut 9 76.1 ± 20.0 0.505 78.0 ± 16.9 0.501 *ZAP70 expression^[d](#tfn_010){ref-type="table-fn"}^* Negative 7 76.1 ± 19.2 75.0 ± 16.6 Positive 15 71.8 ± 15.6 0.579 75.4 ± 15.0 0.954 *CD38 expressione^[e](#tfn_011){ref-type="table-fn"}^* Negative 14 73.8 ± 14.2 74.7 ± 14.3 Positive 7 77.4 ± 15.8 0.599 80.4 ± 14.4 0.402 *NOTCH1 status^[f](#tfn_012){ref-type="table-fn"}^* Unm 19 73.3 ± 17.7 75.6 ± 15.9 Mut 3 72.3 ± 6.2 0.925 73.6 ± 10.5 0.839 Mut, mutated; Unm, unmutated. Percentage of viable cells determined by Annexin V/PI staining in siNotch1 or siNotch2 transfected CLL cells related to siCtrl cells as 100% viability. Treated patients had not received treatment for at least 3 months before the study. Mutated was defined as having a frequency of mutations \>2% from germline *VH*. Positivity refers to detection of \>20% ZAP70+/CD19+. Positivity refers to detection of \>20% CD38+/CD19+. c.7544_7545delCT in *NOTCH1* exon 34. The evidence that Notch1 and Notch2 exert redundant effects in promoting CLL cell survival prompted us to examine the effect of combined Notch1 and Notch2 silencing. We simultaneously transfected CLL cells with Notch1 and Notch2 siRNA (siNotch1/2), and performed these experiments in six CLL samples (patients 6, 7, 9, 11, 16, 18), selected to include patients with different clinical and biological characteristics. Results showed that although the combined siNotch1/2 transfection efficiently downregulated the expression of both receptors (Figure [1A](#F1){ref-type="fig"}), it did not further decrease CLL cell viability with respect to transfection of each single receptor (Figure [1B](#F1){ref-type="fig"}). Notch1 and Notch2 silencing decreases Mcl-1 but not Bcl-2 protein expression in CLL cells {#s2_2} ----------------------------------------------------------------------------------------- Based on the evidence that Mcl-1 and Bcl-2 proteins are highly expressed in CLL cells and play a crucial role in apoptosis resistance and CLL pathogenesis \[[@R19], [@R23]\], we analyzed the effect of Notch downregulation on the expression of these proteins (*n* = 22). We found that in 15 of the 18 samples where Notch1 and Notch2 downregulation reduced CLL cell viability, there was a decrease in Mcl-1 levels compared with siCtrl cells, whereas in the four samples with Notch-independent viability, Mcl-1 levels remained unchanged (Table [2](#T2){ref-type="table"}, Figures [1C and 1D](#F1){ref-type="fig"}). When we transfected CLL cells with combined siNotch1/2 (*n* = 6), we observed that reduction of Mcl-1 levels was more pronounced (Figures [1C and 1D](#F1){ref-type="fig"}). In contrast, in all 22 patients, Bcl-2 levels were not affected by either each single or combined Notch receptor downregulation (Figure [1C](#F1){ref-type="fig"} and data not shown). Altogether, these data indicate that apoptosis of CLL cells induced by Notch silencing involves downregulation of Mcl-1 but not Bcl-2 expression. Combined Notch1 and Notch2 silencing prevents the increase in Mcl-1 levels and cell viability induced in CLL cells by IL-4 {#s2_3} -------------------------------------------------------------------------------------------------------------------------- Several micro-environmental stimuli have been shown to promote *ex vivo* CLL cell survival by increasing Mcl-1 expression \[[@R38], [@R39]\]. The involvement of Notch signaling in Mcl-1-mediated CLL cell survival induced by the microenvironment has never been explored. We examined whether the T-cell derived cytokine IL-4, known inducer of cell survival and Mcl-1 expression in CLL \[[@R39]\], enhanced Notch expression in promoting these effects, and if so, whether this increase was required for Mcl-1-mediated CLL cell survival induced by IL-4. CLL cells, transfected with siCtrl or combined siNotch1/2, were cultured for 72 hours with or without IL-4, and then examined for Notch1, Notch2 and Mcl-1 expression, and cell viability/apoptosis (*n* = 6). In agreement with previous studies \[[@R39]\], we found that in siCtrl cells, IL-4 increased both Mcl-1 levels (Figures [2A and 2B](#F2){ref-type="fig"}) and cell viability (Figure [2D](#F2){ref-type="fig"}). In siCtrl cells, IL-4 significantly upregulated also Notch1 and Notch2 expression (Figures [2A and 2C](#F2){ref-type="fig"}), and interestingly, combined Notch1/2 silencing partially abrogated the increase in both Mcl-1 expression and cell viability induced by the cytokine (Figures [2A-2D](#F2){ref-type="fig"}). These results indicate that in CLL cells, IL-4 enhances Notch expression and that this event is required to induce the increase in Mcl-1-mediated cell survival. ![Combined Notch1/2 silencing prevents the increase in Mcl-1 levels and cell viability induced by IL-4 in CLL cells\ CLL cells, transfected with control siRNA (siCtrl) or combined Notch1 and Notch2 siRNA (siNotch1/2) as described in "siRNA transfection", were cultured for 72 hours in complete medium with or without 25 ng/ml IL-4 (*n* = 6). **A.** Expression of Notch1, Notch2 and Mcl-1 was analyzed as described in Figure [1A,C](#F1){ref-type="fig"}. Vertical line inserted in Notch1 blot of CLL18 indicates a repositioned gel lane. The blots of Mcl-1 **B.** and those of Notch1 and Notch2 **C.** were subjected to densitometric analysis, and densitometry units (U) were calculated relative to GAPDH. **A.** Data of CLL7 and 18 are representative of six samples. **B.**, **C.** Data are the mean ± SD of six samples. \**P* \<0.05, \*\**P* \< 0.01 calculated by Student *t* test. **D.** Cell viability was evaluated by flow cytometric analysis of Annexin V/PI (An V/PI) staining. Viability (An V^−^/PI^−^) of IL-4-untreated siCtrl cells was set to 100%. Data are the mean ± SD of six samples. \*\**P* \< 0.01 calculated by Student *t* test.](oncotarget-06-16559-g002){#F2} Mcl-1 downregulation by Notch silencing is not due to transcriptional control or degradation by caspases but in part to degradation by proteasome {#s2_4} ------------------------------------------------------------------------------------------------------------------------------------------------- Mcl-1 protein expression is regulated at multiple levels, including transcription, translation and degradation \[[@R26]-[@R31]\]. To define whether the reduced Mcl-1 expression induced by Notch targeting was due to transcriptional inhibition, we analyzed Mcl-1 mRNA expression by real-time PCR (*n* = 6). In siNotch1 as well as in siNotch2 cells, Mcl-1 mRNA levels were similar to those observed in siCtrl cells (Figure [3A](#F3){ref-type="fig"}), suggesting that Notch silencing reduces Mcl-1 expression at posttranscriptional level. ![Mcl-1 downregulation by Notch silencing partially depends on proteasome degradation\ **A.** Mcl-1 downregulation by Notch silencing is independent of reduced transcription. CLL cells were transfected with control siRNA (siCtrl), Notch1 siRNA (siNotch1) or Notch2 siRNA (siNotch2) as described in "siRNA transfection" and then cultured in complete medium for 72 hours (*n* = 6). Mcl-1 mRNA levels were evaluated by real-time PCR, normalized to GAPDH and represented as fold change with respect to siCtrl cells. Data are the mean ± SD of six samples. Differences between each siNotch transfection and siCtrl were not significant. **B.**, **C.** Mcl-1 downregulation by Notch silencing is independent of degradation by caspases. siCtrl, siNotch1 or siNotch2 transfected cells were cultured for 72 hours in complete medium with 50 μM pan-caspase inhibitor z-VAD-fmk or 0.005% DMSO as control (*n* = 6). **B.** PARP cleavage, indicator of caspase activity, and Mcl-1 expression were analyzed by western blot on 15 μg whole-cell lysates. Protein loading was assessed by reprobing the blots with an anti-GAPDH antibody. Data of CLL18 are representative of six samples. **C.** The blots of Mcl-1 were subjected to densitometric analysis, and densitometry units (U) were calculated relative to GAPDH. Data are the mean ± SD of six samples. ^\*^*P* \< 0.05; *ns*, not significant (z-VAD-fmk-treated cells *versus* DMSO-treated cells in each transfection condition) according to Student *t* test. **D.**, **E.** Mcl-1 downregulation by Notch silencing partially depends on degradation by proteasome. siCtrl, siNotch1 or siNotch2 transfected cells were cultured for 72 hours in complete medium additioned, during the last 4 hours, with 2.5 μM proteasome inhibitor MG132 or 0.001% DMSO as control (*n* = 6). **D.** Accumulation of polyubiquitinated proteins, indicator of proteasome inhibition, and Mcl-1 expression were analyzed by western blot on 15 μg whole-cell lysates. Protein loading was assessed by reprobing the blots with an anti-GAPDH antibody. Data of CLL18 are representative of six samples. **E.** The blots of Mcl-1 were subjected to densitometric analysis, and densitometry units (U) were calculated relative to GAPDH. Data are the mean ± SD of six samples. \**P* \< 0.05, \*\**P* \< 0.01 (MG132-treated cells *versus* DMSO-treated cells in each transfection condition) according to Student *t* test.](oncotarget-06-16559-g003){#F3} It has been shown that Mcl-1 can be degraded by caspases during apoptosis \[[@R40]\]. To test whether Notch silencing decreased Mcl-1 levels through degradation by caspases, we cultured siNotch1, siNotch2 and siCtrl cells with or without the pan-caspase inhibitor z-VAD-fmk (*n* = 6). Results showed that whereas in siCtrl cells, z-VAD-fmk increased Mcl-1 levels, in siNotch1 as well as in siNotch2 cells treated with z-VAD-fmk, Mcl-1 levels continued to be downregulated, although PARP cleavage, an indicator of caspase activation, continued to be inhibited as in siCtrl cells (Figures [3B and 3C](#F3){ref-type="fig"}). These results suggest that reduction in Mcl-1 expression induced by Notch targeting was independent of caspase activation. Another mechanism controlling Mcl-1 protein expression is proteasomal degradation \[[@R26], [@R30], [@R31]\]. To determine whether the decrease in Mcl-1 levels induced by Notch targeting was due to proteasomal degradation, we tested the effect of the proteasome inhibitor MG132 (*n* = 6). It was added to siNotch1, siNotch2 and siCtrl cells during the last 4 hours of the 72-hour post transfection culture, and its action was demonstrated by the accumulation of polyubiquitinated proteins, an indicator of proteasome inhibition (Figure [3D](#F3){ref-type="fig"}). In siCtrl cells, MG132 increased Mcl-1 levels, indicating that in CLL cells cultured *ex vivo*, Mcl-1 is degraded by proteasome. In both siNotch1 and siNotch2 cells, MG132 prevented the loss of Mcl-1 protein which returned to levels observed in siCtrl cells (Figures [3D and 3E](#F3){ref-type="fig"}). These results suggest that Mcl-1 downregulation by Notch targeting depends, at least in part, on increased degradation by proteasome. Mcl-1 downregulation by Notch silencing is accompanied by reduced eIF4E phosphorylation {#s2_5} --------------------------------------------------------------------------------------- Mcl-1 mRNA translation is highly dependent on the activity of the eukaryotic initiation factor 4E (eIF4E) \[[@R33]\], a key translation regulator associated with tumorigenesis \[[@R35], [@R36]\]. eIF4E is indeed frequently over-expressed and over-activated in human cancers, and acts at a converging point of relevant oncogenic pathways. One pathway regulating eIF4E activity is mediated by the MAPK-interacting kinases 1 (MNK1) and MNK2, which are targets of the Ras/Raf/MAPK signaling and directly phosphorylate eIF4E at Ser209, a crucial event for its oncogenic activity \[[@R41]\]. Another pathway is mediated by the eIF4E-binding protein 1 (4E-BP1), a target of the PI3K/AKT/mTOR signaling, which, when is hypophosphorylated, prevents eIF4E activity through inhibitory interactions \[[@R42]\]. Based on these observations, we examined whether in CLL cells, Mcl-1 downregulation by Notch silencing was accompanied by effects on the expression and phosphorylation of eIF4E (Ser209), MNK1 (Thr197/202), and 4E-BP1 (Thr37/46 and Ser65). Consistent with previous studies \[[@R37]\], our results showed that all examined CLL cells (*n* = 6) expressed high levels of both total and phosphorylated eIF4E forms. eIF4E phosphorylation, but not expression, was reduced by either Notch1 or Notch2 targeting, and to a greater extent, by combined Notch1/2 targeting (Figures [4A and 4B](#F4){ref-type="fig"}). In contrast, expression and phosphorylation levels of MNK1 and 4E-BP1 were not affected by either each single or combined Notch receptor downregulation (Figure [4A](#F4){ref-type="fig"}). These results indicate that Notch signaling controls eIF4E activity, but in doing this, it does not involve MNK1 and 4E-BP1 regulation. These results, along with recent evidence that eIF4E contributes to CLL cell survival \[[@R37]\], also suggest that eIF4E activity is another target of the anti-apoptotic Notch signaling in CLL, in addition to Mcl-1. ![Mcl-1 downregulation by Notch silencing is associated with reduced phosphorylation of eIF4E\ CLL cells were transfected with control siRNA (siCtrl), Notch1 siRNA (siNotch1), Notch2 siRNA (siNotch2) or combined siNotch1 and siNotch2 (siNotch1/2), as described in "siRNA transfection" and then cultured in complete medium for 72 hours (*n* = 6). **A.** Expression and phosphorylation of eIF4E (Ser209), 4E-BP1 (Thr37/46 and Ser65) and MNK1 (Thr197/202) were analyzed by western blot on 15 μg whole-cell lysates, using antibodies able to detect total and phosphorylated forms. Protein loading was assessed by reprobing the blots with an anti-GAPDH antibody. Data of CLL7 and 18 are representative of six samples. **B.** The blots of phosphorylated eIF4E (Ser209) were subjected to densitometric analysis, and densitometry units (U) were calculated relative to total eIF4E. Data are the mean ± SD of six samples. \**P* \< 0.05, \*\**P* \< 0.01 (each siNotch transfection condition *versus* siCtrl) according to Student *t* test.](oncotarget-06-16559-g004){#F4} DISCUSSION {#s3} ========== In the present study, we have identified in the anti-apoptotic Mcl-1 protein and in the key translational regulator eIF4E, two targets of the pro-survival activity of Notch signaling in CLL cells. These findings are interesting because Mcl-1 and eIF4E are oncogenic in several malignancies \[[@R26], [@R35]\], including CLL. In this leukemia, both proteins are important mediators of cell survival \[[@R19], [@R20], [@R37]\], with Mcl-1 which is closely associated with adverse prognosis \[[@R21], [@R22]\] and chemoresistance \[[@R23], [@R24]\]. Specifically, we have shown that the increase in CLL cell apoptosis induced by either Notch1 or Notch2 genetic inhibition is accompanied by Mcl-1 protein downregulation. In contrast, the levels of Bcl-2 protein remained unaffected indicating that Bcl-2 is not a target of the anti-apoptotic signaling of Notch in CLL cells and that Mcl-1 reduction is not due to a general effect of Notch silencing. This observation along with the evidence that the decrease in Mcl-1 levels induced by Notch silencing was observed in the vast majority of CLL cells with Notch-dependent viability (83.3%), but not in CLL cells whose viability was independent of Notch, suggest that Mcl-1 is important for Notch-mediated CLL cell survival. Thus, one of the mechanisms by which Notch1 and Notch2 sustain CLL cell survival is by maintaining the constitutive high levels of Mcl-1. In CLL cells, several pro-survival signals have been implicated in promoting Mcl-1 expression, including those mediated by STAT3 \[[@R27]\], NF-kB \[[@R43]\], and Syk/AKT pathways \[[@R20], [@R44]\]. This is the first evidence that in CLL cells, Mcl-1 expression is also sustained by Notch signaling. This effect of Notch is not due to transcriptional regulation because reduction in Mcl-1 levels induced by Notch1 and Notch2 targeting is not accompanied by changes in Mcl-1 mRNA expression. Mcl-1 downregulation by Notch targeting is also independent of cleavage by caspases, but, as shown in studies of pharmacologic proteasome inhibition, it is partially due to increased Mcl-1 proteasome-mediated degradation, suggesting that Notch signaling controls this process, and contributes to stabilize Mcl-1 levels by interfering with it. Whether in CLL cells, Notch receptors act directly on Mcl-1 protein to stabilize it, as Notch1 does to stabilize XIAP protein \[[@R16]\], or indirectly, by influencing some of the pathways which regulate Mcl-1 proteasome-mediated degradation in these cells, including the Syk/PKCδ \[[@R45]\] and AKT/GSK3 \[[@R46]\] pathways, remains to be defined. Furthermore, the evidence that Mcl-1 downregulation by Notch targeting is accompanied by a decreased activity of eIF4E, an essential factor for Mcl-1 translation, also suggests that Notch signaling may control Mcl-1 expression by regulating its biosynthesis. Further studies will be needed to define this point and to clarify the mechanisms whereby Notch signaling sustains eIF4E activity in CLL cells, given that Notch targeting does not have any effect on the phosphorylation levels of 4E-BP1 and MNK1, which in other cell types, are two key upstream regulators of eIF4E activity \[[@R41], [@R42]\]. The impact of Notch signaling on Mcl-1 expression in CLL cells is also supported by the evidence that the combined Notch1/2 downregulation induces a higher reduction of Mcl-1 levels than that induced by downregulating each single receptor. However, these results seem discordant with the observation that combined Notch1/2 silencing does not enhance CLL cell apoptosis induced by downregulating each single receptor. A possible explanation is that the anti-leukemic activity of a strong Notch downregulation is limited by compensatory survival mechanisms, suggesting that inhibition of Notch signaling alone is not sufficient to kill all leukemic cells. This is consistent with the evidence that in several malignancies, the best clinical activity of Notch-targeted therapies was observed when the specific Notch inhibitors, including γ-secretase inhibitors (GSI) or monoclonal antibodies to Notch receptors or Notch ligands, were administered in combination with conventional chemotherapy or other targeted agents \[[@R47]\]. In line with this evidence, even in CLL cells, it has been recently demonstrated that the clinically relevant GSI PF-03084014 improves the pro-apoptotic effect of fludarabine \[[@R48]\]. This occurs because Notch inhibition overcomes the resistance mechanisms activated in CLL cells by fludarabine \[[@R23]\], including increased NF-kB activation \[[@R48]\] and Mcl-1 expression \[[@R21]\]. In this context, a critical role of Notch in regulating NF-kB pathway has been previously described in T-cell leukemia \[[@R15], [@R49]\] and a role of Notch in controlling Mcl-1 expression is here demonstrated in CLL. Another important aspect of this study is that Notch signaling also contributes to Mcl-1 accumulation induced by survival micro-environmental stimuli. Specifically, we demonstrate that IL-4, known to induce in CLL cells an increased Mcl-1-mediated cell survival \[[@R39]\], also enhances Notch1 and Notch2 activation. Interestingly, combined Notch1/2 downregulation partially prevents the increase in both CLL cell survival and Mcl-1 expression, suggesting that Notch targeting, in addition to reducing the constitutive Mcl-1 levels, is also able to prevent Mcl-1 accumulation induced by micro-environmental stimuli. This effect of Notch downregulation is important because in CLL, the major resistance mechanisms to current chemotherapy, including Mcl-1 expression, are highly favored by the microenvironment \[[@R38]\]. Overall, Notch signaling sustains CLL cell survival by promoting Mcl-1 expression and eIF4E activity. These findings along with the evidence that both Mcl-1 and eIF4E contribute to survival and chemotherapy resistance of CLL cells highlight the importance to target Notch signaling for CLL treatment, especially in combination with agents whose poor efficacy is mainly due to the elevated Mcl-1 expression and eIF4E activity detected in these leukemic cells. MATERIALS AND METHODS {#s4} ===================== Patients {#s4_1} -------- Twenty-two CLL patients entered this study. Diagnoses of CLL were based on Stanford criteria defined by the National Cancer Institute-sponsored Working Group \[[@R50]\], and clinical staging was based on the Binet classification \[[@R51]\]. This study was approved by the local Ethics Committee, and all patients signed informed consent in accordance with the Declaration of Helsinki. CLL cell isolation {#s4_2} ------------------ Peripheral blood mononuclear cells were isolated from heparinized blood of CLL patients by Ficoll density-gradient centrifugation (Nycomed, Oslo, Norway). Monocytes were removed by plastic adherence, and T cells by sheep erythrocyte rosetting. All CLL samples contained more than 96% CD19^+^/CD5^+^ CLL cells, as assessed by flow cytometry (EPICS-XL-MCL; Beckman Coulter, Fullerton, CA). CLL clinical laboratory characteristics {#s4_3} --------------------------------------- *IgV~H~* mutations, CD38 surface and ZAP70 intracellular expression were analyzed as previously reported \[[@R8]\]. Cytogenetic abnormalities were examined by fluorescent in situ hybridization (FISH) using probes for chromosomes 11, 12, 13, 14 and 17. *NOTCH1* exon 34 mutations were analyzed as previously reported and the percentage of mutant allele burden was determined using a semi-quantitative assay on the basis of Genescan analysis \[[@R4]\]. Table [1](#T1){ref-type="table"} gives clinical and biological characteristics of CLL patients. siRNA transfection {#s4_4} ------------------ CLL cells (12 × 10^6^) were resuspended in 100 μl Cell Line Solution Kit V (Lonza Group Ltd, Basel, Switzerland) with ON-TARGETplus SMARTpool small interfering RNA (siRNA) to Notch1 (0.5 μM), Notch2 (0.5 μM) or ON-TARGETplus siCONTROL nontargeting pool as negative control (all from Dharmacon RNA Technologies, Lafayette, CO). In the case of combined Notch1 and Notch2 silencing, 0.25 μM per siRNA was added. Cells were then transfected with the Amaxa Nucleofector II device (program U-013) and cultured for 72 hours in 12-well plates in complete medium consisting of RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum (Hyclone Laboratories, Logan, UT), 2 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin (all from Invitrogen, Milan, Italy). In some experiments, transfected cells were incubated with 25 ng/ml recombinant human IL-4 (Immunotools, Friesoyte, Germany), 50 μM pan-caspase inhibitor z-VAD-fmk or 2.5 μM proteasome inhibitor MG132 (both from Calbiochem, La Jolla, CA). The inhibitors z-VAD-fmk and MG132 were dissolved in DMSO and diluted in complete medium at the used concentrations. DMSO concentrations, which did not exceed 0.005%, did not affect CLL cell responses. IL-4 and z-VAD-fmk were added at the beginning of the 72-hour post transfection culture and maintained until cell collection, whereas MG132 was added during the last 4-hour culture. Analysis of cell viability and apoptosis {#s4_5} ---------------------------------------- Cell viability and apoptosis were assessed by flow cytometric analysis (EPICS-XL-MCL) after Annexin V-fluorescein isothiocyanate/propidium iodide staining, performed using a commercial kit (Immunotech, Beckman Coulter) according to the manufacturer\'s instructions. Western blot analysis {#s4_6} --------------------- Whole-cell lysates were extracted in RIPA buffer. Equal amounts of proteins were separated by 7.5% SDS-PAGE and transferred to nitrocellulose membranes, which, after blocking, were incubated with primary antibodies to: Notch1 (clone bTAN20) and Notch2 (clone C651.6DbHN), developed by Spyros Artavanis-Tsakonas, obtained from DSHB developed under the auspices of the NICHD, and maintained by Iowa University; Mcl-1 (Santa Cruz Biotechnology, Santa Cruz, CA); Bcl-2 (DakoCytomation, Milan, Italy); phospho-eIF4E (Ser209), total eIF4E, phospho-4E-BP1 (Ser65), phospho-4E-BP1 (Thr37/46), total 4E-BP1, phospho-MNK1 (Thr197/202), total MNK1 and PARP (Cell Signaling Technology, Beverly, MA); ubiquitinated proteins (BIOMOL Research Laboratories, Plymouth Meeting, PA) and GAPDH (Sigma-Aldrich, St. Louis, MO). Signals were detected using appropriate horseradish peroxidase-conjugated secondary antibodies and the ECL system (GE Healthcare, Milan, Italy). Densitometric analysis was performed using Quantity One software (Bio-Rad, Milan, Italy). Real-time quantitative PCR {#s4_7} -------------------------- Total RNA was isolated using Trizol (Invitrogen) and 1 μg reverse-transcribed using RT-kit plus (Nanogen Advanced Diagnostics, Milan, Italy). Real-time quantitative PCR was performed with PCR Master Mix Power SYBR Green (Applied Biosystems, Warrington, UK), using the 7900HT Fast Real-Time PCR System (Applied Biosystems). The primer sequences used for *MCL-1* were forward: 5′-GAG ACC TTA CGA CGG GTT-3′ and reverse: 3′-TTT GAT GTC CAG TTT CCG-3′ (Invitrogen). Relative fold change was normalized to *GAPDH* and calculated using the 2^−ΔΔCt^ method. Statistical analysis {#s4_8} -------------------- Statistical differences between mean values were evaluated using the Student *t* test. The minimal level of significance was *P \<* 0.05. This work was supported by grants from Progetti di Ricerca di Interesse Nazionale (PRIN 2010-2011; 2010MCLPLB_005 to E.R.), and the Associazione Italiana per la Ricerca sul Cancro (IG13314 to I.S.). **CONTRIBUTION** F.D.F. conceived and performed experiments and discussed data; R.S., B.D.P. and S.B. performed experiments and interpreted the data; F.F., M.D.I. and P.S. provided patient samples, patient clinical and laboratory data, and critical suggestions; I.S. and P.M. supervised the study and provided critical suggestions; E.R. planned every aspect of the proposal, conceived and supervised the research, analyzed and interpreted data, and wrote the manuscript. All authors revised the manuscript and approved the final version. **CONFLICTS OF INTERESTS** The authors declare no conflict of interest.

{ "pile_set_name": "PubMed Central" }

Murray SR, Stock SJ, Cowan S, Cooper ES, Norman JE. Spontaneous preterm birth prevention in multiple pregnancy. The Obstetrician & Gynaecologist 2018;20;57--63. DOI: [10.1111/tog.12460](10.1111/tog.12460). The copyright line for this article was changed on 10 May 2018 after original online publication. Introduction {#tog12460-sec-0003} ============ Twin pregnancies are high‐risk and associated with increased perinatal morbidity and mortality.[1](#tog12460-bib-0001){ref-type="ref"} Multiple pregnancy is associated with adverse maternal outcomes including increased rates of pre‐eclampsia, pregnancy‐induced hypertension, maternal anaemia and venous thromboembolism.[1](#tog12460-bib-0001){ref-type="ref"} Compared to singleton newborns, newborn infants of twin pregnancies are also at increased risk of adverse outcomes including congenital anomalies, cerebral palsy, intrauterine growth restriction and stillbirth. Although only 3% of all live births are twin pregnancies,[1](#tog12460-bib-0001){ref-type="ref"} twin babies account for up to 15% of special care unit admissions.[2](#tog12460-bib-0002){ref-type="ref"} A 2009 survey conducted by the UK charity the Twin and Multiple Birth Association (TAMBA) found that in 44% of all twins born to the 1298 mothers interviewed, at least one baby entered special care.[3](#tog12460-bib-0003){ref-type="ref"} According to the 2014 MBRRACE‐UK (Mothers and Babies: Reducing Risk through Audits and Confidential Enquiries across the UK) report,[4](#tog12460-bib-0004){ref-type="ref"} the perinatal mortality rate in twins was three times higher than that of singleton pregnancies.[4](#tog12460-bib-0004){ref-type="ref"} Much of this perinatal mortality is driven by prematurity. In Scotland, 50% of twins are delivered preterm (at fewer than 37 weeks of gestation), with around 20% delivering before 34 weeks of gestation.[5](#tog12460-bib-0005){ref-type="ref"} Figures from the USA are similar: a 12‐fold higher preterm birth (PTB) rate of 56.6% was found in twins compared to 9.7% in singletons (odds ratio \[OR\] 12.8, 95% confidence interval \[CI\] 12.6--12.9).[6](#tog12460-bib-0006){ref-type="ref"} Despite attempts by the UK\'s National Institute for Health and Care Excellence (NICE) to reduce multiple pregnancy with the universal use of single embryo transfer (SET) during assisted reproduction technologies (ARTs),[7](#tog12460-bib-0007){ref-type="ref"} the UK twin pregnancy rate remains high among women undergoing ART, with up to 24% of successful in vitro fertilisation (IVF) procedures resulting in multiple pregnancy.[1](#tog12460-bib-0001){ref-type="ref"} As well as the increased risk of perinatal morbidity and mortality, the economic cost of healthcare provision for a twin baby is significantly higher than for singletons, and is quoted to be twice as high in the first 5 years.[8](#tog12460-bib-0008){ref-type="ref"} The James Lind Alliance (JLA) Preterm Birth Priority Setting Partnership (PSP) has prioritised the following question as the number one uncertainty in preterm birth: which interventions are most effective in predicting or preventing PTB? The JLA brings together clinicians, patients and carers to set research priorities in several aspects of obstetrics and gynaecology, including stillbirth and endometriosis.[9](#tog12460-bib-0009){ref-type="ref"} Aetiology {#tog12460-sec-0004} ========= The aetiology of spontaneous PTB in multiple pregnancies is likely to be multifactorial, different from singletons, and remains largely unknown. Different proposed pathophysiology mechanisms include intrauterine infection, cervical insufficiency and increased uterine stretch/distension. There is also increased secretion of mediators such as corticotrophin‐releasing hormone (CRH) from the larger placental mass, and factors produced by the maturing fetal lung such as surfactant protein‐A, which stimulates myometrial contractility and may contribute to preterm parturition.[10](#tog12460-bib-0010){ref-type="ref"} Twins are associated with a higher risk of obstetric intervention and therefore iatrogenic PTB is higher. Approximately one‐third of all premature deliveries in multiple pregnancies are medically indicated.[11](#tog12460-bib-0011){ref-type="ref"} This review focuses on the prevention of spontaneous PTB in twin pregnancies. Prediction of preterm labour in multiple pregnancy {#tog12460-sec-0005} ================================================== Detailed descriptions of the methods of prediction of PTB in multiple pregnancy are out of the scope of this review, but the evidence surrounding the two most researched methods, cervical length and fetal fibronectin, is summarised below. Cervical length measurement {#tog12460-sec-0006} --------------------------- ### In asymptomatic women with a twin pregnancy {#tog12460-sec-0007} A 2010 meta‐analysis of 16 cohort/cross‐sectional studies (n = 3213)[12](#tog12460-bib-0012){ref-type="ref"} showed that in asymptomatic women with twins, a cervical length of \<25 mm was associated with a 25% risk of delivery before 28 weeks of gestation. This review also found a cervical length of \<20 mm at 20--24 weeks of gestation to be associated with a 42.4% risk of birth before 32 weeks, and a 62% risk of birth before 34 weeks. A subsequent systematic review[13](#tog12460-bib-0013){ref-type="ref"} upheld these findings and concluded that in asymptomatic women with a twin pregnancy, a cervical length measurement at 20--24 weeks of gestation was a good predictor of spontaneous PTB. A recent systematic review of 1024 women with twins[14](#tog12460-bib-0014){ref-type="ref"} assessed repeated measures of a change in cervical length as a predictor of PTB, and found that the shortening of cervical length over time had a low predictive accuracy for preterm birth at fewer than 34 weeks of gestation. The largest and most up‐to‐date individual patient data (IPD) meta‐analysis on the effect of gestational age and cervical length measurements in the prediction of PTB in twin pregnancies was published by Kindinger et al.[15](#tog12460-bib-0015){ref-type="ref"} in 2015 (n = 4409 twin pregnancies). The benefit of an IPD meta‐analysis is that it is a more robust method of combining and easily comparing studies, easily allowing subgroup analysis; for example, a short cervix group. This analysis consisted of 12 twin cohorts and found that when cervical length was \<30 mm at 18 weeks of gestation, it was most predictive of birth at ≤28 weeks. Prediction of later spontaneous PTB (28--34 weeks) improved with cervical length measurements taken at later gestations (≥22 weeks). The authors concluded by recommending cervical length screening in twin pregnancies as a predictor of spontaneous PTB from 18 weeks of gestation. In summary, in asymptomatic women with a twin pregnancy, current evidence supports the use of measuring cervical length from 18 weeks of gestation as a predictor of spontaneous PTB, but does not support repeated measures of cervical length. ### In symptomatic women with a twin pregnancy {#tog12460-sec-0008} In a meta‐analysis of five combined cohort/cross‐sectional studies (n = 310), Conde‐Agudelo et al.[12](#tog12460-bib-0012){ref-type="ref"} showed that, in women with twin pregnancies and symptoms of spontaneous PTB, cervical length measurement had a low predictive accuracy for PTB at \<34 weeks of gestation. However, the review concluded that the sample size was small and there is a paucity of evidence in the area of predicting PTB in twin pregnancies. Fetal fibronectin {#tog12460-sec-0009} ----------------- ### In asymptomatic women with a twin pregnancy {#tog12460-sec-0010} NICE does not recommend the use of fetal fibronectin (fFN) in multiple pregnancies. A 2010 meta‐analysis[16](#tog12460-bib-0016){ref-type="ref"} summarising 11 studies of fFN use in asymptomatic twin pregnancy to predict PTB suggested only limited prediction accuracy, and better negative predictive rates than positive (6% risk of PTB before 34 weeks of gestation with a negative test compared to a 33% risk of PTB before 34 weeks with a positive test). ### In symptomatic women with a twin pregnancy {#tog12460-sec-0011} The meta‐analysis by Conde‐Agudelo et al.[16](#tog12460-bib-0016){ref-type="ref"} synthesised the results of five studies using fFN to predict PTB in symptomatic women with a twin pregnancy. The authors concluded that fFN testing in multiple pregnancy was most accurate in women with symptoms of PTB (positive and negative likelihood ratios 85% and 75%, respectively, within 7 days of testing). However, although fFN testing is recommended for use in singleton pregnancies with symptoms of PTB if cervical length measurement is not available,[17](#tog12460-bib-0017){ref-type="ref"} it is not recommended in the NICE guideline for multiple pregnancy. A combination of cervical length measurements and fFN may be a more accurate predictor of PTB. One small study[18](#tog12460-bib-0018){ref-type="ref"} (n = 155 twin pregnancies) reported that if fFN was positive and cervical length was \<20 mm, then 54.4% of twins would deliver before 34 weeks of gestation, and this was significantly higher than the overall rate of PTB. Methods of preterm birth prevention {#tog12460-sec-0012} =================================== Cervical cerclage {#tog12460-sec-0013} ----------------- Cervical cerclage is a surgical technique to prevent PTB and has been described in the UK literature since 1902.[19](#tog12460-bib-0019){ref-type="ref"} Although cerclage is endorsed by NICE for use in singleton pregnancies, it is not recommended for use in multiple pregnancy as its effectiveness remains controversial.[11](#tog12460-bib-0011){ref-type="ref"} A subgroup of twins (n = 49 women) in a 2005 meta‐analysis[20](#tog12460-bib-0020){ref-type="ref"} investigating the use of cerclage showed an increase in prematurity and a trend towards harm. In multiple pregnancies, cerclage was associated with an increased risk of premature delivery (relative risk \[RR\] 2.15, 95% CI 1.15--4.01) and a trend towards an increased risk of perinatal mortality. However, this finding was not statistically significant, with wide confidence intervals likely reflecting the small sample size (RR 2.66, 95% CI 0.83--8.54). A subsequent Cochrane review published in 2014[21](#tog12460-bib-0021){ref-type="ref"} examined five trials, of which two (n = 73 women) assessed history‐indicated cerclage and three (n = 55 women) assessed ultrasound‐indicated cerclage. This review found no benefit of cervical cerclage in reducing preterm delivery in twin pregnancies at fewer than 34 weeks of gestation (RR 1.16, 95% CI 0.44--3.36, four trials, n = 98 women). It also showed a trend towards harm with an increased risk of perinatal death, although this was not statistically significant (RR 1.74, 95% CI 0.92--3.28). There was no reduction in a composite of adverse neonatal outcome (RR 1.54, 95% CI 0.58--4.11). In the pre‐specified subgroup, ultrasound‐indicated cerclage was associated with an increased risk of low birthweight (RR 1.39, 95% CI 1.06--1.83, three trials, n = 98 women) and respiratory distress syndrome (RR 5.07, 95% CI 1.75--14.70, three trials, n = 98 women). The authors concluded that there was no evidence of the usefulness of cerclage in reducing the risk of PTB in twins, but that more research was needed because of a small number of trials, which each had a small number of patients. It is also important to note that no trials were identified that reported long‐term infant neurodevelopmental outcomes following cervical cerclage. ### Ultrasound‐indicated cerclage in twin pregnancies {#tog12460-sec-0014} Following the Cochrane review, a further IPD meta‐analysis was published looking specifically at ultrasound‐indicated cerclage (cervical length ≤25 mm before 24 weeks of gestation) in twin pregnancies.[22](#tog12460-bib-0022){ref-type="ref"} Three randomised controlled trials (RCTs; n = 49 women) were identified. No significant differences in PTB rates before 34 weeks of gestation were found between cerclage and no cerclage in twin pregnancies with a trend towards harm (RR 2.19, 95% CI 0.72--6.63). Similar to the Cochrane review, the authors concluded that further large trials were necessary to determine the effectiveness of cerclage in twin pregnancies. The latest retrospective cohort study was published in 2015 and involved 140 women with twin pregnancies.[23](#tog12460-bib-0023){ref-type="ref"} This showed that ultrasound‐indicated cerclage with a cervical length of ≤25 mm did not reduce the risk of preterm delivery before 34 weeks of gestation (adjusted OR \[aOR\] 0.37, 95% CI 0.16--1.1) nor was it associated with increased neonatal morbidity (aOR 0.7, 95% CI 0.4--1.2). In the pre‐specified subgroup analysis of women with a cervical length of ≤15 mm (n = 32 women), there was a significant reduction in the rate of PTB before 34 weeks of gestation (OR 0.42, 95% CI 0.24--0.81). In summary, unlike singleton pregnancies, there appears to be no benefit of cervical cerclage in reducing PTB rates in multiple pregnancies. Although the study by Roman et al.[22](#tog12460-bib-0022){ref-type="ref"} reported a benefit in those with a short cervix (≤15 mm), the numbers in the study were small and further large trials are needed to address this question adequately. A search on <https://clinicaltrials.gov/> identified two currently recruiting studies comparing ultrasound and emergency cerclage in twin pregnancies with expectant management.[24](#tog12460-bib-0024){ref-type="ref"}, [25](#tog12460-bib-0025){ref-type="ref"} One trial[24](#tog12460-bib-0024){ref-type="ref"} is a multicentre, international RCT of physical exam‐indicated cerclage in twin gestations with a primary outcome measure of preterm delivery at fewer than 34 weeks of gestation. The other study[25](#tog12460-bib-0025){ref-type="ref"} is a single‐centre RCT of cervical cerclage versus expectant management for women with a twin pregnancy and a short cervix (≤25 mm), with a primary outcome of pregnancy prolongation. The 'C‐STICH' RCT[26](#tog12460-bib-0026){ref-type="ref"} of monofilament versus braided sutures for insufficient cervix is currently recruiting in multiple centres in the UK. Although this study only looks at singleton pregnancies, a similar trial in twin pregnancies would be appropriate after the results are published. Progesterone {#tog12460-sec-0015} ------------ ### Is it biologically plausible? {#tog12460-sec-0016} The use of progesterone in the prevention of PTB in both singleton and multiple pregnancies has been extensively investigated. Arguably, the use of progesterone is biologically plausible given that uterine quiescence is maintained throughout pregnancy by progesterone and progesterone receptor‐mediated inhibition of inflammation, which causes suppression of the contractile genes.[27](#tog12460-bib-0027){ref-type="ref"} Labour is thought to occur as a result of a functional withdrawal of progesterone.[28](#tog12460-bib-0028){ref-type="ref"} Anti‐progesterones such as mifepristone are used to stimulate abortion and induce labour. However, others have argued that progesterone levels are high during pregnancy, and progesterone receptors are fully occupied, hence the therapeutic benefit of adding further progesterone is unclear. Progesterone is available as an intramuscular injection of 17α‐hydroxylase caproate (only licensed in the USA), or vaginal progesterone (the only available progesterone product in the UK, but not licensed for prevention of PTB in the USA or Europe). ### Use of progesterone in unselected twin pregnancies {#tog12460-sec-0017} A 2012 IPD meta‐analysis by Romero et al.,[27](#tog12460-bib-0027){ref-type="ref"} which investigated the use of vaginal progesterone in singletons, also included a subgroup of women with twins (n = 52 women). This revealed no reduction in PTB before 34 weeks of gestation when vaginal progesterone was used in twin pregnancies (RR 0.7, 95% CI 0.34--1.44). A subsequent IPD meta‐analysis by Schuit et al. (2015)[29](#tog12460-bib-0029){ref-type="ref"} upheld these main conclusions. The review included 13 trials (3768 pregnancies) of the use of progestogens in unselected twin pregnancies and the primary outcome was neonatal morbidity. Treatment with vaginal progesterone did not reduce the risk of adverse perinatal outcome (RR 0.97, 95% CI 0.77--1.2), and there were no significant differences for delivery before 32 weeks of gestation between the progesterone and the control groups (RR 0.91, 95% CI 0.68--1.2).[29](#tog12460-bib-0029){ref-type="ref"} A recent systematic review and meta‐analysis comparing progesterone, cerclage and the cervical pessary for prevention of PTB in unselected twin pregnancies found no reduction in PTB rates with any of the interventions.[30](#tog12460-bib-0030){ref-type="ref"} Some secondary outcomes were reduced with vaginal progesterone only (very low birth weight, \[RR 0.71, 95% CI 0.52--0.98\], and need for mechanical ventilation, \[RR 0.61, 95% CI 0.45--0.82\]). One of the larger RCTs in the Schuit et al.[29](#tog12460-bib-0029){ref-type="ref"} IPD meta‐analysis, the 'STOPPIT' study (n = 500 women),[31](#tog12460-bib-0031){ref-type="ref"} performed a baby follow‐up study of the effect of prophylactic progesterone in twin pregnancy on childhood outcome.[32](#tog12460-bib-0032){ref-type="ref"} Performed through record linkage of childhood records, this study found no increased incidence of perinatal death (15 twins in the progesterone group versus 11 in the placebo group), congenital anomalies (OR 1.04, 95% CI 0.49--1.21) or hospitalisation (OR 0.97, 95% CI 0.71--1.33) of those exposed to progesterone versus placebo. The authors concluded there was no evidence of a detrimental or beneficial impact on health and developmental outcomes at 3 and 6 years of exposure to progesterone in utero. ### Use of progesterone in twin pregnancies in women with a short cervix {#tog12460-sec-0018} Debate still exists as to whether progesterone is effective in singleton pregnancies of women with a short cervix. The Schuit et al.[29](#tog12460-bib-0029){ref-type="ref"} IPD meta‐analysis published in 2015 included a subgroup analysis of women with a twin pregnancy and a short cervix of ≤25 mm (n = 116 women). Among this subgroup, vaginal progesterone provided a protective effect over the control with regards to adverse perinatal outcome (RR 0.57, 95% CI 0.47--0.70). However, the number of women in the subgroup was small, thus limited conclusions can be inferred. The primary outcome in this study was neonatal morbidity, therefore differences in PTB in the subgroup of women with a short cervix ≤25 mm were not investigated in this review. In summary, there is no benefit of universal vaginal progesterone to reduce PTB rates in multiple pregnancies. One meta‐analysis showed a benefit in adverse perinatal outcome in a subgroup of women with a short cervix ≤25 mm, suggesting it may be useful in this group, but the numbers in the study were small and further research is needed. There appears to be no long‐term harm caused to infants exposed to progesterone in utero.[32](#tog12460-bib-0032){ref-type="ref"} The NICE guidelines for multiple pregnancy currently followed by UK practitioners do not promote the routine use of cervical cerclage or progesterone for the prevention of PTB in multiple pregnancies (see Box [1](#tog12460-fea-0001){ref-type="boxed-text"}). ###### Preterm birth prevention in multiple pregnancy from the National Institute for Health and Care Excellence guideline 2011[1](#tog12460-bib-0001){ref-type="ref"} #### {#tog12460-sec-0019} Do not use the following interventions (alone or in combination) routinely to prevent spontaneous preterm birth in twin or triplet pregnancies: Bed rest at home or in hospitalIntramuscular or vaginal progesteroneCervical cerclageOral tocolytics Cervical pessary {#tog12460-sec-0020} ---------------- The Arabin pessary is a cervical pessary used to prevent PTB. It is a flexible silicon ring with a smaller inner diameter that encompasses the cervix, aiming to tilt it posteriorly and provide cervical support.[33](#tog12460-bib-0033){ref-type="ref"} It is usually inserted at around 18--22 weeks of gestation, and in twin pregnancies is removed before 36 weeks. The Arabin cervical pessary and its correct positioning are shown in Figure [1](#tog12460-fig-0001){ref-type="fig"}. Available evidence surrounding the use of the cervical pessary in twin pregnancies is conflicting and is discussed here in terms of unselected twin pregnancies and twins in women with a short cervix. ![a) The Arabin cervical pessary; b) the position of the Arabin cervical pessary.](TOG-20-57-g001){#tog12460-fig-0001} ### Use of the cervical pessary in unselected twin pregnancies {#tog12460-sec-0021} Two RCTs looking at the use of the cervical pessary in unselected twin pregnancies have been performed. The largest, by Nicolaides et al.[34](#tog12460-bib-0034){ref-type="ref"}, which involved 1180 women who received a cervical pessary versus expectant management, found no difference in the rate of PTB before 34 weeks of gestation (RR 1.05, 95% CI 0.79--1.41). The ProTWIN study[35](#tog12460-bib-0035){ref-type="ref"} was an RCT of 808 unselected twin pregnancies. This study found no difference in PTB before 32 weeks of gestation (10% versus 12%, RR 0.86, 95% CI 0.65--1.15) and no reduction in adverse perinatal outcome (RR 0.98, 95% CI 0.69--1.39). ### Use of the cervical pessary in twin pregnancies with a short cervix {#tog12460-sec-0022} The PECEP‐Twins trial,[36](#tog12460-bib-0036){ref-type="ref"} published in 2016, is the only RCT to focus on twins in women with a short cervix (defined as ≤25 mm). In this study, 137 women with a short cervix were randomly assigned to cervical pessary or expectant management groups. The rate of PTB before 34 weeks of gestation was reduced in the pessary group compared with the expectant management group (RR 0.41, 95% CI 0.22--0.76). In a post hoc subgroup analysis of the study by Nicolaides et al.,[34](#tog12460-bib-0034){ref-type="ref"} the incidence of PTB before 34 weeks of gestation among 106 women with a cervical length of \<25 mm was not significantly different between the pessary group and the expectant management group (31% versus 26%, RR 1.2, 95% CI 0.8--1.8). In contrast, subgroup analysis of the ProTWIN study,[35](#tog12460-bib-0035){ref-type="ref"} which used a cervical length cut‐off of \<38 mm, reported a reduction in PTB before 32 weeks of gestation (16.2% versus 39.4%; RR 0.41, 95% CI 0.22--0.76). In a systematic review and meta‐analysis, Saccone et al.[37](#tog12460-bib-0037){ref-type="ref"} combined the results of the three trials of cervical pessary use for the prevention of PTB of twins of women with a short cervical length, and a forest plot was produced (Figure [2](#tog12460-fig-0002){ref-type="fig"}). It is important to note here that because the Nicolaides et al. and ProTWIN studies used different cervical length cut‐offs (\<25 mm and \<38 mm, respectively), they were not directly comparable in the meta‐analysis and therefore were not powered to detect a difference in PTB in these subgroups. The review concluded that use of the Arabin pessary in twin pregnancies of women with a short cervix may not prevent PTB or improve perinatal outcome. ![Risk of preterm birth \<34 weeks of gestation in twin pregnancies with a cervical length \<25 mm.[37](#tog12460-bib-0037){ref-type="ref"}](TOG-20-57-g002){#tog12460-fig-0002} In summary, evidence for the use of the cervical pessary for the prevention of PTB in twins is conflicting, though there is some evidence to suggest it may be useful in twin pregnancies of women with a short cervix. The STOPPIT 2 RCT[38](#tog12460-bib-0038){ref-type="ref"} is currently recruiting in the UK to help address this paucity of evidence. This RCT aims to resolve the uncertainty surrounding whether or not the Arabin pessary reduces spontaneous PTB in twins of women with a short cervix (\<30^th^ centile, which equates to 35 mm[15](#tog12460-bib-0015){ref-type="ref"}). The cervical pessary is not currently routinely used in clinical practice outside of the research setting. Other methods of preterm birth prevention {#tog12460-sec-0023} ----------------------------------------- As stated in Box 1, the following methods of PTB prevention in multiple pregnancies are not endorsed by NICE. The evidence is summarised below: ### Bed rest and uterine monitoring {#tog12460-sec-0024} In the past, bed rest was used as a method of PTB prevention for both singleton and twin pregnancies. A 2010 Cochrane review[39](#tog12460-bib-0039){ref-type="ref"} comparing hospitalisation and bed rest against expectant management summarised the results of six trials (n = 600 women). It found no benefit of this intervention, but did find an increased risk of PTB before 34 weeks of gestation (OR 1.84, 95% CI 1.01--3.34). Similarly, a 1995 meta‐analysis[40](#tog12460-bib-0040){ref-type="ref"} of six trials on the use of home uterine monitoring showed no reduction in PTB. Thus, neither of these interventions have been endorsed by NICE. ### Prophylactic tocolytics {#tog12460-sec-0025} For singleton pregnancies, the RCOG\'s Green‐top Guideline[41](#tog12460-bib-0041){ref-type="ref"} (GTG) supports the use of tocolysis for the completion of corticosteroids or an in utero transfer, but not for PTB prevention. Similarly, for multiple pregnancy, the GTG states that there is insufficient evidence for its use in the prevention of preterm labour. A 2005 Cochrane review[42](#tog12460-bib-0042){ref-type="ref"} of five trials (n = 344 women) found no reduction in PTB before 34 weeks of gestation with the use of tocolytics (RR 0.47, 95% CI 0.15--1.50). In summary, the use of prophylactic tocolytics to prevent preterm labour in multiple pregnancy is not recommended. Conclusion {#tog12460-sec-0026} ========== There is a lack of effective, evidence‐based interventions for the prevention of PTB in twin pregnancies. There is limited evidence for the use of vaginal progesterone and cervical cerclage, and the cervical pessary is currently only used within a research setting. There are no reported trials comparing the effectiveness of each of these interventions against each other, whether in isolation or in combination.[43](#tog12460-bib-0043){ref-type="ref"} Likewise, although fFN and cervical length scanning may be beneficial in predicting PTB in twins either alone or in combination, no reliable evidence supports the use of any one predictor. Research is needed to further evaluate the benefit of the cervical pessary and the use of cervical cerclage in twins of women with a short cervix. A recent article by Stock et al.[43](#tog12460-bib-0043){ref-type="ref"} concludes by advising clinicians to share with women the uncertainty of methods to prevent PTB in multiple pregnancy, and offer the opportunity to participate in clinical trials. Disclosure of interests {#tog12460-sec-0027} ----------------------- JEN has research grants from government and charitable organisations (Tommy\'s, the baby charity, the Medical Research Council and National Institute of Health Research \[NIHR\]) for preterm birth prevention (including in twins). The University of Edinburgh receives also funding for her contribution to a data monitoring committee for a preterm birth study run by GlaxoSmithKline. SRM, SJS, SC and ESC have no conflicts of interest. Contributions to authorship {#tog12460-sec-0028} --------------------------- SRM and JEN instigated the article. SRM researched and wrote the article. SJS, ESC, JEN and SC edited the article. All authors approved the final version.

{ "pile_set_name": "PubMed Central" }

Introduction {#sec1-1} ============ Dental caries is still a major concern in public health management.\[[@ref1]\] Establishment of a productive dental health philosophy, especially for the low socioeconomic levels of the society, where the lack of development endangers the dental health, even more, seems to be necessary.\[[@ref2]\] Microbial nature of caries and the contemporary concept of treatment based on the medical model necessitate the consideration of chemical control plans along with the traditional mechanical means of caries control.\[[@ref3]\] Innovations in dental material science are come to the point of developing smart materials.\[[@ref4]\] glass-ionomer cement (GIC), a tooth-colored acid-base material, with the capability of fluoride release in an aqueous environment is the first of this category.\[[@ref3][@ref4]\] It can be used as base, liner, or direct restorative material and is also considered as the material of choice in atraumatic restorative technique (ART). The ART is a minimal intervention approach, particularly beneficial for pediatric and elderly patients as well as those with dental anxiety or learning difficulties.\[[@ref3]\] While the bulk of evidence supports the protective nature of fluoride in public water and oral health products; however, the data available does not endorse the anti-caries ability of GIC or any useful outcome of fluoride-releasing restorative materials.\[[@ref3][@ref5]\] To address the issue, multiple modifications are being under investigation to develop a GIC with direct antimicrobial effect based on its ability to participate in ion-exchange reactions with the oral environment.\[[@ref3][@ref5]\] Such a material would be of tremendous clinical benefits. Providing antibacterial seal leads to suppression of remaining bacteriae under the restoration, reduced postoperative sensitivity, improved pulpal and periodontal health, and finally reduction in recurrent caries. Different studies are designed attempting to develop a modified GIC by incorporation of antimicrobial agents such as chlorhexidine hydrochloride, cetylpyridinium chloride, cetrimide, and benzalkonium chloride.\[[@ref6]\] Hook *et al*. proposed a GIC functionalized with chlorhexidine-hexametaphosphate particles. The suggested material released chlorhexidine in a dose-dependent manner and showed to have antimicrobial properties.\[[@ref3]\] In another study, de Castilho *et al*. added doxycycline hyclate as an antibiotic with local antimicrobial and antimatrix metalloproteinase activities into the GIC. The modified material showed inhibitory effects on caries producing bacteria.\[[@ref7]\] Deepalakshmi *et al*. incorporated chlorhexidine and cetrimide as wide-spectrum antiseptics into GIC and developed an antibacterial material.\[[@ref6]\] Herbal extracts have recently gained even more attention as active agents to be incorporated into oral care products and dental materials. Many of them have potential antimicrobial activities which could lead to the production of safe, economical, and efficient alternative materials for use in caries management.\[[@ref8]\] The Iranian herbal medicine with a long history of effectiveness and safety is a rich source of such agents. *Salvia officinalis*, which is also called Sage, is a perennial evergreen plant native to Mediterranean region\[[@ref8]\] and is also naturalized in Iran.\[[@ref9]\] It has woody stems, grayish leaves with blue-to-purplish flowers. A long ancient and contemporary history of medicinal and pharmacological applications supports its use. *S. officinalis* extract has antimicrobial, analgesic, anti-inflammatory and antioxidant properties.\[[@ref8][@ref10]\] The essential oil of *S. officinalis* consists of alpha- and beta-thujone, borneol, camphor, and cineole. In contemporary European herbal medicine, a type of *S. officinalis* tea is commonly advocated to control a sore throat, inflammatory oral lesions, and gingivitis.\[[@ref8][@ref10]\] Kermanshah *et al*. showed Sage mouthwash to have an inhibitory effect on caries-producing oral microbial flora.\[[@ref9]\] In another study, Beheshti-Rouy *et al*. proved the *S. officinalis* extract to be effective in reducing the number of *Streptococcus mutans* colonies in bacterial plaque.\[[@ref11]\] Different other studies also reported the inhibitory effect of *S. officinalis* extract on collagenolytic activity of *Porphyromonas gingivalis*,\[[@ref12]\] and also on the attachment of *Candida albicans* to denture surfaces.\[[@ref8]\] In an attempt to develop a GIC with direct antimicrobial properties, this study is conducted to investigate any possible inhibitory effects *S. officinalis* modified GIC may have on *S. mutans* and *Lactobacillus casei* as the main bacteria involved in caries initiation and progression process. The experiment is performed for *S. officinalis* weight concentration levels of 0.5%, 0.75%, 1%, and 1.25%. Methods {#sec1-2} ======= Some high-quality dried *S. officinalis* is purchased from a local herbal medicine retail store (Parand herbal products, Isfahan, Iran). Leaves of the plant are chopped and fragmented into small pieces and filtered through a \#40 mesh (Sina Lab. Inst., Tehran, Iran). Each 50 g of leaves are soaked in 1500 ml of solvent (50% water, 50% ethanol \[96%\]) in a shaker apparatus (Heidolph Unmax, Schwabach, Germany) at 90 rpm for 48 h. Thereafter, the solution is passed through a strainer and then transferred to a rotary evaporator apparatus (Heidolph WD2000; Schwabach, Germany) to separate the solvent from the extract.\[[@ref11]\] The purified extract is then dried by applying freeze-drying technique in three stages over a 1-week period. The final extract powder is stored in sealed vial at low temperature to be used in the next steps. To filter the particles, the same size as range of GIC powder which is to be \<50 μ, the powder is grinded and again filtered through a \#270 laboratory mesh (Sina Lab. Inst., Tehran, Iran). The procedure is performed under the supervision of a pharmacology professor at the main laboratory of the School of Pharmacy, Isfahan University of Medical Sciences and Health Services, Isfahan, Iran. A conventional powder and liquid Fuji IX GIC (GC Corporation, Tokyo, Japan) is used as control group named Group I. Experimental GIC samples are prepared by incorporating *S. officinalis* extract powder into the powder component of Fuji IX GIC (GC Corporations, Tokyo, Japan) at 0.5% (Group II), 0.75% (Group III), 1% (Group IV), and 1.25% (Group V) weight concentration levels using a digital weight scale (CAS MWP Micro Weighing Balance, CAS Scales, New Zealand). Thus, four groups of experimental powder formulations are obtained, each sample weighs 1000 mg according to the procedure outlined in.\[[@ref6]\] The concentrations used in this research are based on the minimal inhibitory concentration and minimal bacterial assays suggested in.\[[@ref9]\] The groups are presented in [Table 1](#T1){ref-type="table"}. ###### Test groups assigned to each bacterial category based on the weight concentration of *Salvia officinalis* extract powder Test group number\* Weight percentage concentration of *S. officinalis* extract powder^\^^ --------------------- ------------------------------------------------------------------------ I (control) 0 II 0.5% III 0.75% IV 1% V 1.25% \*These groups are assigned to each bacterial category (*S. mutans* and *L. casei*), ^\^^Weight concentration of *S. officinalis* extract powder added to the powder component of Fuji IX Glass-ionomer cement (GC Corporation, Tokyo, Japan) to form test groups. *S. mutans*=*Streptococcus mutans*, *L. casei*=*Lactobacillus casei*, *S. officinalis*=*Salvia officinalis* Five specimens are prepared from each experimental group by a specialist in restorative dentistry. Powder/liquid is dispensed on a mixing pad according to the manufacturer recommendations and mixed for 30 s with a sterile plastic spatula, and then inserted into the stainless steel molds of 10 mm in diameter and 2-mm thickness (Abzar Iran, Arak, Iran) within 1 min with a sterile dental instrument and allowed to set for 30 min at room temperatures, same as the procedure in.\[[@ref6]\] Stock cultures of *S. mutans* (UA159) and *L. casei* (ATCC \#193) obtained from the Microbiology and Immunology Laboratory of the School of Medicine, Isfahan University of Medical Sciences and Health Services, Isfahan, Iran are used by an expert laboratory technician. For each bacterial category, cells are cultured freshly from frozen stock on brain--heart infusion broth (BHI; DIFCO Laboratories, Detroit, MI, USA) for 24 h at 37°C in a 10% CO~2~ incubator. The viability and absence of contamination is confirmed by plating in a specific medium and using Gram procedures. Cultures are again grown in BHI for 20 h at 37°C and the inoculums for subsequent testing are obtained according to.\[[@ref7]\] In each sterilized Petri dish (15 mm × 90 mm), a base layer containing 15 ml of BHI agar mixed with 300 ml of each inoculums (*S. mutans* and *L. casei*) was prepared according to.\[[@ref7]\] The set disk-shaped specimens are placed on a BHI agar plate inoculated with bacterial strain and left at 37°C for 48 h. Zones of inhibition are then measured in millimeters using a digital caliper at three different points. Size of the inhibition zones is calculated by subtracting the diameter of the specimen from the average of the three measurements of the halo according to.\[[@ref6]\] Statistical analysis {#sec2-1} -------------------- Data are analyzed using Statistical Package for the Social Sciences, SPSS software version 22 for Windows (IBM Inc., NY, USA). SPSS is a software package used in statistical analysis of data. This software package was developed by SPSS incorporation and acquired by IBM in 2009. In 2014, the software was officially renamed IBM SPSS statistics. Kolmogorov--Smirnov test is used to test the normality of the data. The test is rejected at significance level of 5%, and the data follow distributions different than normal. Thus, nonparametric Kruskal--Wallis, Mann--Whitney, and Wilcoxon signed-rank sum tests are performed to compare the test groups and also the test groups with the control group. Data are analyzed using Statistical Package for the Social Sciences (SPSS, version 22) for Windows (IBM Inc., NY, USA). Kolmogoro v--Smirnov test is used to test the normality distribution of the data. The nonparametric Kruskal--Wallis and Mann--Whitney tests are then employed to compare the test groups and also the test groups with the control group. The statistical analyses are all performed at significance level of 5%. Results {#sec1-3} ======= According to Kruskal--Wallis test, there was a significant difference between the mean diameter of the inhibition zone for *Streptococcus mutans* (*P* \< 0.001) and *Lactobacillus casei* (*P* \< 0.001) in different concentrations of salvia extract \[[Table 2](#T2){ref-type="table"}\]. ###### The mean diameter of the inhibition zone for *Streptococcus mutans* and *Lactobacillus casei* in different concentrations of salvia extract Concentration Mean±SD *P* ------------------------ --------- ----------- ----------- ----------- ----------- --------- *Streptococcus mutans* 0 0.26±0.05 0.02±0.34 0.02±0.62 0.07±1.1 \<0.001 *Lactobacillus casei* 0 0.05±0.04 0.02±0.47 0.05±0.56 0.05±1.06 \<0.001 According to Mann-Whitney test, the mean diameter of the inhibition zone for *Streptococcus mutans* was significantly different between all groups with each other and with the control group (*P* \< 0.05) and the mean diameter of the inhibition zone for *Lactobacillus casei* was significantly different between all groups with each other and with the control group (*P* \< 0.05), except between the control group and the group with the concentration of 0.5% (*P* = 0.134). \[Tables [3](#T3){ref-type="table"} and [4](#T4){ref-type="table"}\]. ###### Comparison of the mean diameters of the inhibition zones between the test groups in the *Streptococcus mutans* category\* Test group Group I Group II Group III Group IV Group V ------------ --------- ---------- ----------- ---------- --------- Group I \- Group II 0.005 \- Group III 0.004 0.014 \- Group IV 0.005 0.007 0.006 \- Group V 0.005 0.007 0.006 0.007 \- \*Wilcoxon signed-rank sum test as a nonparametric test was used for comparing the means. The numbers show *P* - value within groups. Significant level was considered at 5% ###### Comparison of the mean diameters of the inhibition zones between the test groups in the *Lactobacillus casei* category\* Test group Group I Group II Group III Group IV Group V ------------ --------- ---------- ----------- ---------- --------- Group I \- Group II 0.134 \- Group III 0.005 0.007 \- Group IV 0.005 0.007 0.02 \- Group V 0.005 0.007 0.007 0.007 \- \*Wilcoxon signed-rank sum test as a nonparametric test was used for comparing the means. The numbers show *P* - value within groups. Significant level was considered at 5% Pairwise comparison of the mean diameter of inhibition zones of samples with similar concentrations in the *Streptococcus mutans* and Lactobacillus acidophilus showed that only the inhibitory effect of the concentration of 0.5% on *Streptococcus mutans* was significantly higher than that of Lactobacillus acidophilus \[[Table 5](#T5){ref-type="table"}\]. ###### Comparison of the mean diameters of the inhibition zones of the same group between two bacterial categories of *Streptococcus mutans* and *Lactobacillus casei\** Test group Mean diameter of the inhibition zones *P* ------------ --------------------------------------- --------- ----------- Group II 0.26 mm 0.04 mm 0.006 Group III 0.34 mm 0.47 mm 0.058^\^^ Group IV 0.62 mm 0.56 mm 0.339^\^^ Group V 1.10 mm 1.06 mm 0.99^\^^ \*Wilcoxon signed-rank sum test as a nonparametric test was used for comparing the means. Significant level was considered at 5%, ^\^^*P*-value indicates no statistically significant difference. *S. mutans*=*Streptococcus mutans*, *L. casei*=*Lactobacillus casei* Discussion {#sec1-4} ========== In this research, the possibility of producing an herbally modified GIC with direct antibacterial properties is investigated. A distinguishing feature of this study is the use of an herbal extract with a long history of safety and effectiveness instead of chemical agents to produce a modified GIC. The results showed that the incorporation of *S. officinalis* extract into the powder component of a routinely used GIC produces a dental material with direct inhibitory properties against the most important bacterial causes of dental caries. It is also shown that this ability has a dose-response nature. The statistically nonsignificant difference between Group I and Group II (control group) in the *L. casei* category shows a lower potential in *S. officinalis* to inhibit the *L. casei* at its lowest concentration. The statistically significant higher value of the mean inhibition zone diameter in Group II of the *S. mutans* category compared to the Group II of the *L. casei* category points also to the same finding which should be considered cautiously. The findings of this study challenge the traditional idea that the dental materials must be passive without any interactions with surrounding oral environment\[[@ref4]\] and also recommends the use of materials with active capabilities as a possible approach in the management of different carious states. It seems that the ability of such a modified GIC to inhibit the bacterial activity depends on its interactions with the surrounding aqueous environment.\[[@ref3][@ref5]\] It is shown that bacteria responsible for caries may survive up to 2 years even under regular GICs, which many believe to have anticarious properties.\[[@ref13]\] This means that any intentional (in case of close proximity to pulp) or unintentional residual caries under the restorations may progress and encroach on pulp and produce discomfort for patient and finally endanger the pulp vitality.\[[@ref14][@ref15][@ref16]\] This also questions the accuracy of ART, in which multiple carious teeth are only excavated with a hand instrument and filled with a regular GIC, especially in low socioeconomic areas of society.\[[@ref2][@ref16][@ref17]\] A modified GIC with direct antimicrobial properties addresses the situation here. Different investigators incorporated different antimicrobial agents in different concentrations into GIC in an attempt to find a novel solution.\[[@ref1][@ref3][@ref6][@ref7][@ref18]\] Wide-spectrum antimicrobials such as chlorhexidine,\[[@ref1][@ref3][@ref18]\] cetrimide,\[[@ref6]\] and antibiotics\[[@ref7][@ref6]\] were at the center of the scope of those studies. The agent(s) were added into the powder or liquid component of the GIC, utilizing different incorporation techniques. However, none of them are incorporated any herbal agent as a safer source with fewer or no possible complications such as mucosal or dental staining seen in chlorhexidine products.\[[@ref19]\] The reasons for selecting the *S. officinalis* in this research are the long history of successes and no history of any negative side effect documented for, in traditional and contemporary medical literature.\[[@ref12]\] It is also reported that *S. officinalis* extract has superior efficiency compared to the antibiotics and has also less chance of drug resistance when used in high concentrations.\[[@ref20]\] In addition, researchers are recently investigating possible applications of *S. officinalis* essential oil extract in oral care products. The studies reported inhibitory effect of the extract on many oral bacteria and fungi responsible for carious and periodontal infections.\[[@ref8][@ref9][@ref11][@ref12]\] They also reported positive effect on reducing the number of plaque bacterial colonies,\[[@ref11]\] impairing plaque formation,\[[@ref21][@ref22]\] and fungi attachment to prosthetic surfaces.\[[@ref8]\] Based on the available evidence, which recognizes the anti-inflammatory properties in *S. officinalis* extract,\[[@ref9][@ref12][@ref20][@ref22][@ref23]\] one may suggest the use of the *S. officinalis* modified GIC may alleviate and control the inflammatory status of the pulp or periodontium. It is recommended to conduct experiments to investigate such abilities. While the presence of *S. officinalis* extract as an antibacterial agent in GIC is effective against oral pathogens, investigations to evaluate other important properties such as mechanical, physical, and biocompatibility properties of the material are highly recommended before any routine clinical applications. Considerations should be taken to improve the color of *S. officinalis* modified GIC since the current product has an intense yellow color. Conclusions {#sec1-5} =========== Within the limitations of this study, it is concluded that a *S. officinalis* modified GIC has a direct antibacterial property against *S. mutans* and *L. casei*. This, in turn, provides many advantages for use as a base, liner, or restorative material in operative procedures and management of the complications of carious teeth. It also emphasizes the importance of the herbal agents in developing new dental materials with astonishing properties. Financial support and sponsorship {#sec2-2} --------------------------------- Nil. Conflicts of interest {#sec2-3} --------------------- There are no conflicts of interest. The authors are thankful to the staff of the Department of Microbiology, School of Medicine, and the Main Laboratory of the School of Pharmacy -- Isfahan University of Medical Sciences and Health Services, Isfahan, Iran, for their support and friendly cooperation during every step of this investigation. We also would like to thank Mr. Tavangar, School of Dentistry -- Isfahan University of Medical Sciences and Health Services, Isfahan, Iran, for his kind help in organizing and analyzing the data.

{ "pile_set_name": "PubMed Central" }

INTRODUCTION {#sec1-1} ============ The development of whole slide imaging (WSI) has made it possible to capture images of an entire histological slide. While different hardware platforms support the image capture process, it is the development of user-friendly software interfaces that has enabled pathologists to begin to seriously contemplate integration of digital pathology activities into their routine clinical practice. The user-friendly interfaces make it possible to navigate to various regions of the scanned slide and change magnification for diagnosis (digital pathology), to view the image remotely by different pathologists in real time (telepathology), to capture static images of the slide for reporting or archiving, and to perform computer-aided analysis (digital image analysis). Roles for WSI in surgical pathology to support off-site frozen section interpretation;\[[@ref1]--[@ref4]\] primary diagnosis, not only as telepathology consultations,\[[@ref5]--[@ref7]\] but also to improve service to underserved areas;\[[@ref8][@ref9]\] educational activities;\[[@ref10]--[@ref14]\] and laboratory quality assurance (QA) activities are all well established.\[[@ref15]--[@ref18]\] While the advantages of WSI for all of these applications are well established, the criteria for evaluating the cost of WSI-based activities have traditionally been based solely on direct costs (hardware and software) and diagnostic accuracy.\[[@ref6][@ref19]--[@ref22]\] However, these types of analyses largely ignore a spectrum of workflow and cost issues that arise when WSI is integrated into routine diagnostic surgical pathology as a result of redundancy (due to the fact that histological sections on glass slides are an intrinsic component of surgical pathology, and the addition of WSI necessarily involves duplication of the primary diagnostic material), the need for additional staffing (so that WSI does not introduce delays in diagnosis), and customized software development (often required to address unique aspects of specific pathology practices). The pre-scan, scan, and post-scan costs; quality control and QA costs; and IT process costs that result when WSI is integrated into routine surgical pathology can be significant.\[[@ref6][@ref21][@ref23][@ref24]\] Consequently, most pathology groups find it difficult to perform a realistic cost--benefit analysis of adding WSI to their practice, or are uncertain as to the process they should use to decide whether WSI will have utility in their spectrum of patient care activities. In this paper, we report a "value added" approach developed to guide our decisions regarding integration of WSI into our surgical pathology practice.\[[@ref25]--[@ref28]\] The approach focuses on specific operational measures (cost, time, and enhanced patient care) and the settings in which they can provide enhancement (patient care, education, and research). Application of this value added approach provided the focus to identify several routine activities in which the addition of WSI could improve our practice, and resulted in expanded and improved surgical pathology operations. SETTING {#sec1-2} ======= Our group is a large surgical pathology practice (33 attending pathologists, 35 residents, and 16 subspecialty fellows from 10 different pathology subspecialty fellowship programs) at an academic tertiary care medical center and affiliated medical school. The practice has a subspecialty emphasis model with 12 subspecialty sections and multiple sign-out areas (individual pathologist\'s offices, common sign-out rooms, two frozen section areas that are in different buildings one city block apart); handles a large volume of high complexity cases (approximately 55,000 cases per year, including in-house cases and consultations); staffs over 40 patient care conferences per month; supports clinical, translational, and basic research activities of the department and medical school; and is involved in medical student teaching. In support of the diagnostic activities of the group, the histology laboratory processes approximately 200,000 blocks per year (approximately 800 blocks per day) and produces about 380,000 slides per year (approximately 1500 slides per day) including over 28,200 immunohistochemical stains per year, over 1600 *in situ* hybridization (ISH) slides per year, and over 3100 immunofluorescence stains per year. Value added in our practice is defined in purely operational terms in three categories \[[Table 1](#T1){ref-type="table"}\], specifically cost savings, time savings, and improvements in patient care.\[[@ref1]--[@ref3][@ref5]--[@ref12][@ref15][@ref16]\] Although added value can be assessed on a number of different axes within each of these three categories, analysis was limited to support of patient care, educational activities, and research. Some components of the analysis are quantifiable and objective (see below), other components of the analysis are subjective. Added value can be achieved overall despite negative impacts in some categories and/or along some axes within each category. Finally, it should be emphasized that some aspects of WSI are specifically not included in a value added approach; for example, the mere capability to produce a digital image that can be used for diagnosis is, in and of itself, not value added simply because it is technically feasible or novel. ###### Operational definition of "Value added" ![](JPI-2-39-g001) VALUE ADDED ANALYSIS {#sec1-3} ==================== [Table 2](#T2){ref-type="table"} presents the operational costs associated with WSI in our practice. Note that the time to scan slides in routine surgical pathology workflows is significantly higher than advertised by vendors, which reflects the operational realities of whole slide scanning as part of daily practice. The capital and personnel costs to incorporate WSI into our practice are summarized in [Table 2](#T2){ref-type="table"}. In the most recently completed calendar year (2010), we scanned 2.7% of the slides from our laboratory, for which we needed one scanner (Aperio ScanScope, Aperio Technologies Inc., Vista, CA, USA) at approximately 30% maximal utilization, one digital imaging technician (at approximately 0.5 FTE) and one information technology (IT) support person (at approximately 0.3 FTE). Were we to scan every slide in our practice, it would require an additional initial capital investment of approximately \$2,000,000 (for hardware and software), and a yearly indirect cost of approximately \$650,000 for support personnel and \$10,000 for data storage. ###### Operational costs for incorporating WSI into routine surgical pathology workflows ![](JPI-2-39-g002) Value added: Enhanced patient care {#sec2-1} ---------------------------------- There are five areas in which WSI provides capabilities which are superior to current options, or for which other options are simply not available \[[Table 3](#T3){ref-type="table"}\]. Although these areas do not improve diagnostic accuracy of the specimen of origin of the scanned slides, they provide operational advantages that improve patient care in our practice in four ways. First, the use of WSI of selected slides from cases sent in consultation (the medicolegal climate in the United States dictates return of all diagnostic materials to the referring institution) provides the opportunity to enhance patient care by providing an immediately available permanent record of the slides to guide frozen section diagnosis at the time of subsequent definitive excision; for comparison at sign-out of subsequent excision or post-therapy specimen; for presentation at patient care conferences; in QA activities; and so on. Second, WSI of selected slides sent to other institutions as requested or required by their policies for patient care, or slides encumbered by medicolegal proceedings, provides us with a permanent record for use in patient care activities even though we lose control of the original glass slides. Third, WSI of slides that will be destroyed as part of ancillary testing makes it possible to retain the diagnostic content of the slides; given the demonstration that molecular tests can be performed on nucleic acids collected from diagnostic areas microdissected from glass slides,\[[@ref29]\] the electronic record of slides produced by WSI will likely become more important. Fourth, WSI of slides for digital image analysis, for example, *HER-2/neu* analysis, supports emerging slide-based diagnostic paradigms.\[[@ref30][@ref31]\] ###### Value added in enhanced patient care ![](JPI-2-39-g003) Value added: Educational programs {#sec2-2} --------------------------------- The use of WSI to support educational activities is well established.\[[@ref10]--[@ref14]\] At our institution, the added value was provided by the opportunity to produce virtual study sets. WSI also made it possible to view selected slides remotely from anywhere in the institution, which can be used to support medical lectures, medical student labs, and both intramural and extramural departmental educational activities. Value added: Research {#sec2-3} --------------------- WSI makes it possible to produce virtual study sets that can be viewed for expert panel review, eliminating the need to produce multiple recuts of the study sets, mailing the various panel numbers, and even travel to a common site for real-time synchronous viewing by panel members. WSI is also used to support various clinical trials (in which digital images can substitute for glass slides for consensus for centralized review, e.g., by the Gynecologic Oncology Group), as well as to create a permanent record of slides that will be destroyed by the ancillary testing that is part of a clinical trial. Not value added {#sec2-4} --------------- In our practice, WSI does not provide added value in many areas classically associated with digital imaging, although it must be emphasized that aspects of WSI that are not value added for our group may well provide a benefit in other practice settings. Since we have a large laboratory, we already have on site the glass slides from both routine histochemical and immunohistochemical stains, and so WSI does not provide the opportunity to save money or time in primary diagnosis. Also, since our practice has a subspecialty emphasis model with experts on site in virtually every organ system (quite literally across the hallway, if not one block away), there is no value added component for WSI to save time or increase diagnostic accuracy. Similarly, since we already have a well-developed "bricks and mortar" consultation service, there is no immediate advantage to be gained in our practice by WSI for telepathology or branding. In fact, several impediments were identified that actually impede the capture of value added components of WSI in our consultation practice. First, among our referral base, the 1-day delay introduced by overnight express shipping of slides for primary or subspecialty review is not viewed as a significant detriment to patient care (in fact, overnight shipping of slides for early morning delivery integrates nicely into the routine workflow of our group). Second, although overnight express shipping is somewhat cumbersome and introduces the risk of slide loss or damage, it is extremely cost-effective (at current rates, the cost of express overnight shipping one 18 × 12.5 × 3 inch box that can hold up to 100 slides and 100 blocks, every day for 1 year, is less than \$4000; in contrast, the current cost of entry level WSI hardware and software is approximately \$135,000--\$160,000 for a 5--10 slide scanner). Third, significant IT resources are required to develop, implement, and support the HIPAA compliant processes necessary to transfer the images; to support automated reporting to numerous outside client lab information systems; and to bill for the patient care activities (to say nothing of navigating the quagmire of state licensure requirements). COSTS ASSOCIATED WITH IMPLEMENTATION {#sec1-4} ==================================== It became clear early in our evaluation of WSI that the faculty and trainees at our institution varied in their experience with the software for viewing digital slide images, their willingness to incorporate WSI into their routine practice, and in the increases in time they were willing to tolerate (even among faculty who were willing to include WSI into their workflow). The faculty and trainees were vocal in their unwillingness to incorporate a WSI process that required them to move back and forth between different software packages (with concomitant multiple log-ons using different usernames and passwords, after time-outs due to inactivity of the different programs, and so on), and most were unenthusiastic about a need to have several computer monitors so that multiple software packages could be open at the same time (notably, faculty that already had two computer screens running various software packages simultaneously were reluctant to add a third computer monitor to accommodate WSI). Space constraints limited the feasibility of multiple computer screens at all work stations, and the department was reluctant to purchase additional monitors. Finally, from a purely operational perspective, we were reluctant to pursue a model incorporating a requirement for multiple monitors that might work well at the medical center, but would not support remote faculty sign-out (either at home or while traveling), remote faculty and trainee research activities, or remote faculty and trainee educational activities. Consequently, in collaboration with several vendors, we pursued a model of "one-stop-shopping" in which there was an automatic in-house interface between our lab information system (Cerner CoPath Plus, Cerner Corporation, Kansas City, MO, USA) and the imaging software (Aperio Spectrum). Development of this new functionality took approximately 11 months overall, including 9 months to design the system architecture and write the associated software code and 2 months to implement within our existing workflow. Implementation of the automatic interface had an overall cost of approximately \$70,000 \[for software development for the Aperio and the CoPath HL7 interfaces, and for purchase of the underlying CoPath Advanced Bar Coding and Tracking (AB and T) module\]. The "one-stop-shopping" solution is illustrated in [Figure 1](#F1){ref-type="fig"}. ![One-stop-shop solution. In CoPath, clicking on the Digital Imaging tab (upper red oval, panel a) immediately opens an Aperio Spectrum WSI window (panel b) that automatically displays the scanned slides from that case. Clicking on the Image Gallery tab (lower red oval, panel a) immediately opens a static image gallery (panel c) that includes the scanned paperwork from the case, gross images, and microscopic images. Secure remote access via the Citrix solution opens a window (panel d) with links (red ovals) to CoPath (for access to clinical WSI, as above) or directly to Spectrum (for access to research and educational WSI)](JPI-2-39-g004){#F1} The need for the automatic interface emphasizes the types of additional costs that are often overlooked in evaluations of the utility of WSI in routine pathology workflow. Specifically, so-called off-the-shelf hardware and software packages, regardless of the vendor, have generic functionality. Integration of WSI into specific pathology practices often requires hardware and software changes to the standard products which can be costly and time consuming to develop and implement. OUTCOMES {#sec1-5} ======== We also sought to evaluate whether the value added approach was successful in guiding integration of WSI into our clinical practice in ways that enhanced our patient care, research, and education activities. Our metrics in this analysis included changes in the number of scans, changes in faculty and trainee acceptance, and utilization \[[Table 4](#T4){ref-type="table"}\]. ###### Outcomes of a value added approach ![](JPI-2-39-g005) Number of scans {#sec2-5} --------------- The number of cases scanned per year has shown consistent growth (at least 33% per year over the last 3 years). Of note, the average number of slides scanned per cases has decreased, which we interpret as evidence that the faculty and trainees are gaining experience with WSI and are learning which slides contribute the most to patient care. This change in the number of slides scanned per case suggests that the economic model presented above may require revision as we gain more experience with WSI. Acceptance {#sec2-6} ---------- Measurement of faculty and trainee acceptance is not a simple endeavor. However, we interpret faculty and trainee demands for remote access to clinical WSI activities via laptop computer, tablet personal computers, and smart phones as evidence that the faculty and trainees are integrating WSI into their routine workflows. In this regard, our pathologists are voicing an expectation in line with use by their clinical colleagues of mobile devices to access lab test results\[[@ref32]\] and imaging studies.\[[@ref33]\] We have implemented the same one-stop approach to support the demands for remote access through development of a secure infrastructure in which all data are housed in an HIPAA approved environment. Access to the environment is via a Citrix application (Citrix Systems, Fort Lauderdale, FL, USA), specifically a Citrix Presentation server 4.5 installed on a Windows 2003 server-based system, which permits delivery of applications as services and thus provides on-demand secure access for users. Although this solution gives us the capability to access patient reports as well as view whole slide images in a HIPAA compliant environment, our use of Citrix\'s Independent Computer Architecture (ICA) protocol has clear limitations when accessing and scrolling through large amounts of graphical data; since ICA protocols were not expressly designed for this functionality, current user experience can be choppy. Expanded utilization {#sec2-7} -------------------- Our initial value added approach identified WSI of slides seen in consultation as an enhancement to patient care. We interpret faculty interest in extending WSI to include select in-house cases as evidence of increasing recognition of a role for WSI in patient care activities. We likewise interpret faculty and trainee interest in extending WSI into a broader range of educational and research roles as evidence that our value added approach was successful in demonstrating key areas in which WSI could provide utility. Available FREE in open access from: <http://www.jpathinformatics.org/text.asp?2011/2/1/39/84232>

{ "pile_set_name": "PubMed Central" }