Previous studies reported that administration of somatostatin (SST) to human patients mitigated their diarrheal symptoms. Octreotide (an analog of SST) treatment in animals resulted in upregulation of sodium/hydrogen exchanger 8 (NHE8). NHE8 is important for water/sodium absorption in the intestine, and loss of NHE8 function results in mucosal injury. Thus we hypothesized that NHE8 expression is inhibited during colitis and that SST treatment during pathological conditions can restore NHE8 expression. Our data showed for the first time that NHE8 is expressed in the human colonic tissue and that NHE8 expression is decreased in ulcerative colitis (UC) patients. We also found that octreotide could stimulate colonic NHE8 expression in colitic mice. Furthermore, the somatostatin receptor 2 (SSTR2) agonist seglitide and the somatostatin receptor 5 (SSTR5) agonist L-817,818 could restore NHE8 expression via its role in suppressing ERK1/2 phosphorylation. Our study uncovered a novel mechanism of SST stimulation of NHE8 expression in colitis.
- inflammatory bowel disease
inflammatory bowel disease (IBD) is a chronic condition characterized by colitis in ulcerative colitis (UC) and in Crohn's disease. Genetic and environmental factors are implicated in the pathogenesis of IBD (8, 12). Diarrhea and abdominal pain are the common symptoms in IBD patients, and these symptoms are attributed to mucosal injury and impaired intestinal homeostasis. Abnormal water/sodium absorption has been established as having a vital role in the pathogenesis of inflammatory diarrhea (27). In the intestinal tract, Na+/H+ exchanger isoform 3 (NHE3) and NHE8 have been identified as important players in water/electrolyte transport. NHE3 knockout mice developed diarrhea, acid-base imbalance, and alteration in Na+ homeostasis (26). Loss of NHE8 resulted in altered mucosal protection and colonic inflammation (19, 32). In TNBS- and DSS-induced models of colitis, NHE8 expression was decreased in the intestine (32, 34). Thus improvement of NHE8 expression may have a role in restoration of mucosal function and also improving diarrhea in inflammatory colitis.
Somatostatin (SST), a neuropeptide produced by D cells, has multifunctions in protecting mucosal epithelial cells, regulating immune function, and inhibiting intestinal motility (9, 31). Dharmsathaphorn et al. (10, 11) found in the 1980s that SST could stimulate sodium and chloride absorption and inhibit diarrhea, and this discovery was confirmed later by others (2, 24). In clinical practice, SST analogs such as octreotide, are also used to treat intractable diarrhea, intestinal fistula, short-bowel syndrome, and carcinoid syndrome (6a, 13). We have previously reported that SST treatment could stimulate NHE8 expression in the intestine in normal mice and that SST could regulate tight junctions in mice with colitis (18, 33). However, whether the SST-mediated antidiarrheal effect is the result of stimulating NHE8 and/or NHE3 expression in colitis remains to be defined.
The current study demonstrated that NHE8 is expressed in the human intestine and that NHE8 expression is decreased in ulcerative colitis (UC) patients. NHE8 expression is also decreased in mice with DSS colitis, as well as in TNF-α-treated Caco-2 cells. SST treatment could ameliorate diarrhea by stimulating NHE8 but not NHE3 expression. Furthermore, SST stimulates NHE8 expression during inflammation through the SSTR2/5-ERK1/2 signaling pathway.
MATERIALS AND METHODS
Human colorectal tissues from age- and gender-matched control individuals and UC patients were obtained during colonoscopy at West China Hospital of Sichuan University. Control individuals were patients with colorectal cancer or polyps, and the colonic tissue biopsies were obtained at least ten centimeters away from the lesion. UC patients were all diagnosed with active disease by endoscopy and pathology tests. Ulcerative colitis is a distal disease and large number of patients has left sided disease. Therefore, majority of the biopsies in this study were collected from rectum and the left colon (Table 1). Tissue biopsies for protein extraction were immediately stored in a liquid nitrogen container at the time they were collected. Tissue biopsies for the immunofluorescence (IF) study were frozen and embedded in OCT compound. This study was approved by the Ethics Committee of West China Hospital of Sichuan University (Chengdu, China).
Female C57BL/6 mice (6–8 wk old) were purchased from Experimental Animal Center of West China Hospital (Sichuan University). Mice were fed with 3% DSS water for 7 days to induce colitis. From the 8th day, the treatment group received octreotide administration at doses of 50 μg/kg body wt three times a day for 3 days, while the control group received PBS injection. On the 11th day, all mice were euthanized and colonic tissues were collected. During the experiment, body weight, stool consistency and occult blood were assessed daily. Diarrheal scores were recorded following the previously described method (7). Scores higher than 2.0 were considered to have diarrheal symptoms. Disease activity index (DAI) was determined from combined scores of body weight loss, stool consistency, and fecal hemoccult following the method described previously (25). All experiments were approved by the Institutional Animal Care and Use Committee of Sichuan University.
Caco-2 cells were purchased from American Type Culture Collection (Manassas, VA). Cells were cultured in a MEM-NEAA medium containing 50 U/ml penicillin, 50 μg/ml streptomycin, and 20% fetal bovine serum in a 5% CO2 incubator. For the TNF-α study, cells were pretreated with 100 ng/ml TNF-α for 18 h, followed by exposure to 1 μM SST or SSTR agonist (1 μM seglitide or 500 nM L-818,817) according to the previously described method (18, 23, 33). For the signaling pathway study, TNF-α-pretreated cells were exposed to 25 μM SB202190 or 25 μM PD98059 for 1 h, respectively (33).
RNA purification and RT-PCR analysis.
To identify the expression of SSTRs in Caco-2 cells, specific primers of SSTRs (Table 2) were used to amplify by RT-PCR. Briefly, total RNA was purified from Caco-2 cells using Trizol reagent (Takara, Japan). Then, 100 ng total RNA were reverse transcribed to cDNA by Moloney murine leukemia virus reverse transcriptase (Thermo). RT-PCR was conducted under the following conditions: initial denaturation at 94°C for 10 min, followed by 35 cycles of 1 min at 94°C, 1 min at 58°C, and 1 min at 72°C. PCR products were then detected by electrophoresis in 1.5% agarose gel.
Protein isolation and Western blot detection.
Total protein from colonic tissues and Caco-2 cells was prepared in a Cell Lyses Buffer for Western and IP kit (Beyotime, Beijing, China) following manufacturer's instructions. Tissue and cell lysates were used for Western blot detection. Briefly, proteins were separated by 10% SDS-PAGE gel and transferred to a polyvinylidene fluoride membrane. Then the membranes were incubated with NHE8 antibody (35), NHE3 antibody (Santa Cruz Biotechnology), ERK1/2 antibody (R&D System, America), phosphorylated ERK1/2 (pERK1/2) antibody (R&D System, America), and GAPDH antibody (Good Here, Hangzhou, China). The membranes were then incubated with appropriate secondary antibody for 2 h. The target bands were detected using a BM Chemiluminescence Western Blotting Kit (Roche, Germany). Targeting protein expression was normalized by GAPDH. The primary antibody dilution used in this study is: 1:2,000 for NHE8 antibody, 1:2,000 for NHE3 antibody, 1:1,000 for ERK1/2 or pERK1/2 antibody, and 1:2,000 for GAPDH antibody.
Histological assessment of colitis and NHE8 IF.
Colonic tissues were embedded in paraffin after being fixed in 10% formalin and were stained with hematoxylin and eosin (H&E) for tissue histological examination. To detect NHE8 expression, an immunofluorescence assay was conducted. Briefly, frozen colorectal tissues were embedded in OCT compound and sliced into 5-μm sections. The sections were then fixed with acetone for 10 min followed by incubation with 5% goat serum. Sections were subsequently incubated with NHE8 antibody (1:100) overnight at 4°C. The specimens were washed three times with PBST before incubation with Alexa Fluor-594 conjugated goat anti-rabbit IgG secondary antibodies (Santa Cruz Biotechnology) for 1 h at room temperature. After incubation with the nuclear-staining agent DAPI (Sigma), sections were mounted on glass slides. Images were examined and analyzed with a Zeiss Imager Z2 inverted fluorescence microscope (Carl Zeiss, Jena, Germany).
Colonic tissues were collected from mice and used for SST radioimmunoassay. SST radioimmunoassay was done by Beijing DORUN International Technology (Beijing, China).
Data were analyzed with SPSS 16.0 software using ANOVA post hoc test for comparison of three and more groups. Student's t-test was used for comparison of two groups. P < 0.05 was considered significant.
NHE8 is expressed in human colorectal tissue and its expression is reduced in UC patients.
Tissue biopsies were obtained from left colon and rectum from 18 controls and 26 newly diagnosed UC patients. Immunofluorescence stain was used to locate NHE8 protein in the tissue. Total proteins were extracted and used for subsequent NHE8 detection by Western blot. As shown in Fig. 1A, fluorescence microscopy detected apically expressed NHE8 protein in human colorectal epithelia. The signal intensity of NHE8 in non-UC tissues was stronger than that in UC tissues. Since our preliminary observation showed a similar expression level of NHE8 protein in the right colon, the left colon, and the rectum, we compared NHE8 protein abundance in all biopsies obtained from left colon and rectum from control and UC patients. As indicated in Fig. 1B, Western blot confirmed the dramatic decrease in NHE8 protein abundance in UC patients compared with controls (0.67 ± 0.08 in UC vs. 1.49 ± 0.18 in control, n = 44, P < 0.05).
Octreotide treatment ameliorates clinical symptoms and histological scores in DSS colitis mice.
DSS induced diarrhea occurred at the fifth day with a diarrhea score of 2.50 ± 0.03. The symptom was exacerbated through the rest of the experiments. However, octreotide administration (from days 8 to 10) significantly reduced diarrhea scores (from 4.00 ± 0.23 in nontreated DSS mice to 2.50 ± 0.21 in octreotide-treated DSS mice, n = 18, P < 0.05; Fig. 2A). DAI scores were also decreased in octreotide-treated DSS mice on the 10th day (4.0 ± 0.0 in nontreated DSS mice vs. 3.33 ± 0.09 in octreotide-treated DSS mice, n = 23, P < 0.05; Fig. 2B). Furthermore, a reduced SST expression was detected in colitis tissues by SST radioimmunoassay. In the proximal colon, SST expression was significantly reduced from 3.72 ± 0.75 pg/mg tissue in control mice to 2.59 ± 0.3 pg/mg tissue in DSS mice (n = 12 mice, P < 0.05). In the distal colon, SST expression was decreased from 3.48 ± 0.62 pg/mg tissue in control mice to 2.63 ± 0.44 pg/mg tissue in DSS mice (n = 12 mice, P < 0.05). Octreotide treatment also restored SST expression to levels similar to control mice (Fig. 2C). H&E staining indicated that DSS disrupted the colonic epithelial layer, and octreotide treatment partially restored the integrity of the epithelial layer. IF detection on NHE8 protein showed that a weaker NHE8 staining was seen in DSS mice, and octreotide treatment increased NHE8 protein intensity (Fig. 3A). Using a colon injury assessment method (20), we found that the histological score in DSS mice was dramatically increased compared with control mice (5.5 ± 0.19 in DSS mice vs. 0.57 ± 0.2 in control mice, n = 18, P < 0.05). Octreotide treatment partially restored the mucosal injury score (3.67 ± 0.3 in treated DSS mice vs. 5.5 ± 0.19 in DSS mice, n = 18, P < 0.05; Fig. 3B).
Octreotide stimulates NHE8 but not NHE3 expression in DSS colitis mice.
DSS treatment significantly reduced NHE8 protein abundance in mice. In the proximal colon, NHE8 protein abundance was significantly decreased in DSS mice compared with control mice (0.66 ± 0.08 in control mice vs. 0.31 ± 0.04 in DSS mice, n = 18, P < 0.05). In the distal colon, NHE8 protein abundance was also significantly reduced from 0.79 ± 0.15 in control mice to 0.36 ± 0.06 in DSS colitis mice (n = 18, P < 0.05). After octreotide treatment, NHE8 expression in the proximal colon was increased from 0.31 ± 0.04 in untreated DSS mice to 0.61 ± 0.08 in treated DSS mice (n = 18, P < 0.05); NHE8 expression in the distal colon was increased from 0.36 ± 0.06 in untreated DSS mice to 0.74 ± 0.1 in treated DSS mice (n = 18, P < 0.05; Fig. 4A). At the meantime, NHE3 expression in DSS mice was also significantly decreased from 2.28 ± 0.29 in control mice to 1.38 ± 0.2 in DSS mice in the proximal colon (n = 18, P < 0.05) and from 2.19 ± 0.4 in control mice to 1.08 ± 0.14 in DSS mice in the distal colon (n = 18, P < 0.05). However, the expression of NHE3 was not changed by octreotide treatment (Fig. 4B).
SST restores NHE8 protein levels in TNF-α-treated Caco-2 cells.
To confirm the stimulatory effect of SST on NHE8 in Caco-2 cells, we treated Caco-2 cells with 1 μM SST for 18 h. As shown in Fig. 5A, SST stimulated NHE8 expression in Caco-2 cells (0.32 ± 0.02 in control cells vs. 0.56 ± 0.1 in SST-treated cells, n = 3, P < 0.05), which was similar to our previous study (33). When Caco-2 cells were treated with 100 ng/ml TNF-α, NHE8 expression was significantly decreased from 0.49 ± 0.1 in control cells to 0.28 ± 0.01 in TNF-α-treated cells (n = 3, P < 0.05). In TNF-α-pretreated cells, SST treatment (1 μM for 1 h) increased NHE8 expression from 0.28 ± 0.01 in cells without SST treatment to 0.56 ± 0.01 in cells with SST treatment (n = 3, P < 0.05; Fig. 5B).
SST regulates NHE8 expression through SSTR2 and SSTR5.
To detect the subtype of SSTR expression in Caco-2 cells, conventional RT-PCR was used with SSTR-specific primers (Table 2). As shown in Fig. 6, SSTR1, SSTR2, SSTR4, SSTR5, but not SSTR3 were amplified from Caco-2 cells. Previous studies have shown that SST is involved in NHE8 regulation and anti-inflammation in Caco-2 cells through SSTR2 and SSTR5 (17, 33). To explore whether SSTR2 or SSTR5 is involved in SST-modulated NHE8 expression in TNF-α-treated Caco-2 cells, selective SSTR2 agonist (seglitide) and SSTR5 agonist (L-817,818) were used. As shown in Fig. 7A, seglitide could stimulate NHE8 expression in TNF-α-treated cells (0.37 ± 0.02 in seglitide/TNF-α-treated cells vs. 0.24 ± 0.02 in TNF-α-treated cells, n = 3, P < 0.05). Similarly, L-817,818 at concentrations of 50 nM and 500 nM could also increase NHE8 expression from 0.26 ± 0.02 in TNF-α-treated cells to 0.43 ± 0.05 in 50 nM L-817,818-treated cells and 0.47 ± 0.02 in 500 nM L-817,818-treated cells (n = 3, P < 0.05; Fig. 7B). To determine the effect of exposure time of L-817,818 on NHE8 expression in TNF-α-treated cells, 500 nM L-817,818 were used to treat cells and NHE8 expression was determined at different time points. As shown in Fig. 7C, L-817,818 exposure for 1 h or 18 hrs stimulated NHE8 expression in TNF-α-treated cells with comparable levels.
SST regulates NHE8 expression through suppressing ERK1/2 phosphorylation.
Our previous study has shown SST simulating NHE8 via activation of p38 phosphorylation (pp38) (33). SST has also been shown to be involved in regulating phosphorylation of ERK1/2 (pERK1/2) in Caco-2 cells (17). To explore the signaling pathway involved in SST treatment, the level of pp38 and pERK1/2 in TNF-α and SST-treated cells were compared by Western blot. As shown in Fig. 8A, TNF-α treatment increased p38 phosphorylation in Caco-2 cells (1.45 ± 0.18 in TNF-α-treated cells vs. 0.98 ± 0.14 in control cells, n = 3, P < 0.05). SST treatment brought p38 phosphorylation down to basal level. Similarly, TNF-α treatment increased ERK1/2 phosphorylation in Caco-2 cells, and SST treatment restored ERK1/2 phosphorylation to basal level (1.22 ± 0.05 in TNF-α-treated cells vs. 0.58 ± 0.03 in SST-treated cells, n = 3, P < 0.05; Fig. 8B). To assess whether SST stimulates NHE8 expression via suppressing phosphorylation of p38 and/or ERK1/2, SB202190 (p38-specific inhibitor), and PD98059 (ERK1/2-specific inhibitor) were used. As shown in Fig. 9, SB202190 treatment failed to restore NHE8 expression in TNF-α-treated cells (0.32 ± 0.10 in SB202190-treated cells vs. 0.23 ± 0.10 in TNF-α-treated cells, n = 3; Fig. 9A). However, when TNF-α-treated cells were exposed to PD98059, TNF-α mediated NHE8 expression inhibition was restored (0.42 ± 0.05 in PD98059-treated cells vs. 0.23 ± 0.10 in TNF-α-treated cells, n = 3, P < 0.05; Fig. 9B). Furthermore, cells treated with seglitide or L-817,818 showed reduced ERK1/2 phosphorylation in TNF-α-treated cells (Fig. 10A). To further compare the effect of SST and PD98059 on NHE8 expression, TNF-α-treated Caco-2 cells were exposed to SST and/or PD98059. As shown in Fig. 10B, all treatment conditions (SST, PD98059, SST/PD98059) restored NHE8 expression in TNF-α-treated cells (0.23 ± 0.02 in TNF-α-treated cells vs. 0.46 ± 0.06 in SST-treated cells, 0.4 ± 0.05 in PD98059-treated cells, 0.46 ± 0.03 in SST-PD98059-treated cells, n = 3, P < 0.05).
The intestinal epithelium plays a key role in mucosal protection and ion movement (4, 6). NHEs expressed at the apical membrane are vital electroneutral sodium transporters. NHE3 and NHE8, which are expressed at the apical membrane of the intestinal epithelial cells, are involved in intestinal ion transport and in diarrheal illnesses (16, 26, 33). Moreover, studies have demonstrated that NHE8 plays an important role in protecting the mucosa of the gastrointestinal tract (19, 36, 37). Thus stimulating NHE expression may have a role in intestinal mucosal protection. In a previous study, we have reported that SST could modulate NHE8 but not NHE3 expression in mice under normal physiological conditions (33). However, it remains unclear if such stimulation also occurs in inflammatory states, such as IBD, and whether this effect would reduce the diarrhea seen in these disorders. The intestinal inflammation has been shown to inhibit NHE3 expression in colitis patients and in animal models of colitis (29). Although studies have also shown that NHE8 expression could be inhibited in chemically induced colitis in animals and by TNF-α in cultured intestinal epithelial cells (34), the effect of inflammation on NHE8 expression in human colitis has not been explored. In the present study, we noted that NHE8 expression is expressed in the colorectal tissue in humans and its expression is significantly reduced in active UC patients.
Somatostatin is a polypeptide that is widely distributed in the nervous system, gastrointestinal tract and endocrine cells. SST could modulate various physiological processes such as proabsorption, antisecretion, and antiproliferation (33). Thus, a decrease in SST expression would impair intestinal function. Previous studies reported that SST levels were reduced in IBD patients and in mice with colitis (15, 18). We observed similar findings in DSS colitis mice. The decreased SST levels in DSS colitis mice could be partially restored after octreotide treatment. The diarrheal scores, DAI scores, and colonic mucosa histology were also improved in octreotide-treated DSS colitis mice. Thus SST treatment could improve diarrheal symptom and restore intestinal mucosa.
Somatostatin has been shown to stimulate NHE8 expression under normal physiological conditions in mice, however, the role of SST as an antidiarrheal and anti-inflammatory agent and whether it involves NHE8 are not clear. Consistent with previous studies (29, 34), the expression of both NHE8 and NHE3 was decreased in DSS colitis mice. Octreotide treatment restored NHE8 expression in the colon to normal levels in DSS colitis mice. However, octreotide did not increase NHE3 expression in the colon in DSS colitis mice. Thus, these observations suggested that the role of SST in ameliorating diarrhea is most likely involved in regulating NHE8 expression in the colon.
To study the mechanism(s) of somatostatin modulating NHE8 expression in inflammatory conditions, we tested the effect of TNF-α on NHE8 expression in Caco-2 cells (34). In the current study, we observed a stimulation of NHE8 expression by SST and a reduction of NHE8 expression by TNF-α in Caco-2 cells. These changes were in agreement with previous studies (33, 34). Moreover, SST treatment restored NHE8 expression levels in TNF-α-treated Caco-2 cells. This observation suggested that SST could upregulate NHE8 expression under normal physiological and pathophysiological conditions.
Somatostatin triggers its downstream biological responses via interaction with its receptors (SSTRs). Multiple SSTRs are expressed heterogeneously in various cells, and different receptor activation has different roles (22). To determine the subtype of SSTRs involving in SST-mediated NHE8 expression regulation during inflammation, we detected the SSTRs expression in Caco-2 cells by RT-PCR. Our result indicated the presence of SSTR1, 2, 4, and 5 in Caco-2 cells. Interestingly, we reported previously that SSTR1, 2, 3, and 5 were expressed in Caco-2 cells (33). The difference on SSTR isoform expression between these Caco-2 cells may be secondary to the heterogeneity of Caco-2 cell lines. Since previous studies showed that SST stimulated NHE8 expression through SSTR2 activation in Caco-2 cells (33) and SST regulated tight junction expression in LPS-treated Caco-2 cells by activating SSTR5 (17), we wanted to know if SSTR2 and/or SSTR5 participate in SST-mediated NHE8 expression regulation during inflammation. Using the SSTR2-specific agonist seglitide and SSTR5-specific agonist L-817,818, we found that both seglitide and L-817,818 could restore NHE8 expression in TNF-α-treated Caco-2 cells in a similar pattern in cells-treated with SST. These observations suggested that SST could restore NHE8 expression via activating SSTR2/5 in the intestinal epithelial cells during inflammatory settings.
The MAPK signaling pathway is closely modulated by SST (21). p38 phosphorylation is involved in SST-mediated NHE8 expression in the intestinal epithelial cells (33). ERK1/2 phosphorylation is associated with inflammation and cytokine expression (1, 14). SST has been shown to inhibit ERK1/2 phosphorylation in endothelial cells (3, 5). To elucidate the signaling transduction pathways involved in SST-mediated NHE8 activation during inflammation, we focused our study on p38 and ERK1/2 signaling pathways. Our results showed that the phosphorylation level of p38 and ERK1/2 was upregulated in TNF-α-treated cells, and this increase was significantly inhibited by SST treatment. By using the p38 MAPK-specific inhibitor SB202190 and the ERK1/2 MAPK-specific inhibitor PD98059, we found that SB202190 could not restore NHE8 expression in TNF-α-treated cells, while PD98059 treatment could restore NHE8 expression in TNF-α-treated cells. Therefore, the ERK1/2 pathway is most likely involved in TNF-α-mediated NHE8 expression inhibition during inflammation. To determine if SSTR2 and SSTR5 modulate NHE8 expression through the ERK1/2 pathway during inflammation, we detected the phosphorylation level of ERK1/2 in the presence of seglitide and L-817,818 in TNF-α-treated Caco-2 cells. As expected, both seglitide and L-817,818 supressed ERK1/2 phosphorylation in TNF-α-treated Caco-2 cells. In the meantime, the expression of NHE8 was restored by seglitide and L-817,818. These observations suggested that SST-mediated restoration of NHE8 expression during inflammation was mainly involved in SSTR2 and SSTR5 activation. Furthermore, we confirmed that SST, PD98059, and SST/PD98059 could restore NHE8 expression at a comparable level. Taken together, these results suggest that ERK1/2 phosphorylation induced by inflammation could be inhibited by SST via SSTR2/5 in the intestinal epithelial cells. Since SST reduces intracellular cAMP concentration (38), we used H-89, a potent inhibitor of cAMP-dependent protein kinase (mainly PKA), to treat Caco-2 cells (20 μM 1 h). Interestingly, NHE8 expression was not affected by H-89 treatment in Caco-2 cells (data not shown). Thus the cAMP-mediated PKA pathway is unlikely involved in NHE8 regulation in the intestinal epithelial cells.
In summary, we have shown for the first time that NHE8 is expressed in the human intestinal tissues and its expression is decreased in UC patients. We also identified that somatostatin could restore NHE8 expression in colitis, which results in improving diarrheal symptoms and mucosal structure. Furthermore, SST exerts its effect by restoring NHE8 expression during intestinal inflammation via the suppression of ERK1/2 phosphorylation through activation of SSTR2/5.
This investigation was funded by National Natural Science Foundation of China Grant 81070294 and National Institute of Diabetes and Digestive and Kidney Diseases Grant R01-DK-073638.
No conflicts of interest, financial or otherwise, are declared by the author(s).
X.L., L.C., C.G., J.L., L.T., and D.S. performed experiments; X.L., L.C., H.X., C.G., J.L., L.T., D.S., and C.W. analyzed data; X.L., L.C., and C.W. interpreted results of experiments; X.L. and L.C. prepared figures; X.L. and L.C. drafted manuscript; H.X. and C.W. conception and design of research; H.X., F.K.G., and C.W. edited and revised manuscript; H.X., F.K.G., and C.W. approved final version of manuscript.
- Copyright © 2016 the American Physiological Society