p38 MAPK α and β isoforms differentially regulate plasma membrane localization of MRP2

Abstract

In hepatocytes, cAMP both activates p38 mitogen-activated protein kinase (MAPK) and increases the amount of multidrug resistance-associated protein-2 (MRP2) in the plasma membrane (PM-MRP2). Paradoxically, taurolithocholate (TLC) activates p38 MAPK but decreases PM-MRP2 in hepatocytes. These opposing effects of cAMP and TLC could be mediated via different p38 MAPK isoforms (α and β) that are activated differentially by upstream kinases (MKK3, MKK4, and MKK6). Thus we tested the hypothesis that p38α MAPK and p38β MAPK mediate increases and decreases in PM-MRP2 by cAMP and TLC, respectively. Studies were conducted in hepatocytes isolated from C57BL/6 wild-type (WT) and MKK3-knockout (MKK3^−/−) mice and in a hepatoma cell line (HuH7) that overexpresses sodium-taurocholate cotransporting polypeptide (NTCP) (HuH-NTCP). Cyclic AMP activated MKK3, p38 MAPK, and p38α MAPK and increased PM-MRP2 in WT hepatocytes, but failed to activate p38α MAPK or increase PM-MRP2 in MKK3^−/− hepatocytes. In contrast to cAMP, TLC activated total p38 MAPK but decreased PM-MRP2, and did not activate MKK3 or p38α MAPK in WT hepatocytes. In MKK3^−/− hepatocytes, TLC still decreased PM-MRP2 and activated p38 MAPK, indicating that these effects are not MKK3-dependent. Additionally, TLC activated MKK6 in MKK3^−/− hepatocytes, and small interfering RNA knockdown of p38β MAPK abrogated TLC-mediated decreases in PM-MRP2 in HuH-NTCP cells. Taken together, these results suggest that p38α MAPK facilitates plasma membrane insertion of MRP2 by cAMP, whereas p38β MAPK mediates retrieval of PM-MRP2 by TLC.

the multidrug resistance-associated protein-2 (MRP2) is located at the canalicular membrane of hepatocytes and is involved in biliary secretion of conjugated endogenous and exogenous organic anions (17, 33). Choleretic and cholestatic agents affect bile formation in part by regulating the plasma membrane level of MRP2 (PM-MRP2). For example, choleretic agents such as tauroursodeoxycholate (TUDC) and cAMP increase PM-MRP2 levels (5, 16, 38), and cholestatic agents such as taurolithocholate (TLC) and ethinylestradiol-17β-glucuronide (E-17G) decrease PM-MRP2 (32, 43). Although cAMP and TLC have opposing effects on PM-MRP2, both cAMP and TLC have been shown to activate p38 MAPK (18, 42, 45). Because p38 MAPK has been shown to increase plasma membrane localization of canalicular transporters including MRP2 (23, 24, 42), it is unclear why TLC activates p38 MAPK but decreases PM-MRP2. A likely possibility may be that the opposing effects of cAMP and TLC on PM-MRP2 are mediated via different isoforms of p38 MAPK.

There are four known isoforms (α, β, γ, and δ) of p38 MAPK with only α and β isoforms expressed in livers (12, 20, 44). Studies in nonhepatic cells show isoform-specific effects of p38 MAPK. For example, activation of p38α MAPK is crucial to cytokine production and signaling in most inflammatory cells (21). p38α MAPK and p38β MAPK have reciprocal effects on the expression of inducible nitric oxide synthase in renal mesangial cells (31). Activation of p38α, but not p38β, is involved in cAMP-mediated thermogenic gene expression in adipocytes (37). Although p38 MAPK is involved in growth regulation, toxicity, gluconeogenesis, bile acid synthesis, and transporter trafficking in livers (3, 9, 10, 23, 39, 47), it is unknown whether these effects are mediated via specific isoforms of p38 MAPK. Because cAMP increases PM-MRP2 by activating p38α MAPK (42), we hypothesize that TLC decreases PM-MRP2 by activating p38β MAPK.

Three upstream MAPK kinases (MKKs), namely MKK3, MKK6, and MKK4, control activation of the p38 MAPK isoforms. MKK3 and MKK6 are considered the major activators of p38 MAPK (14, 22, 36). MKK4, which functions primarily as an upstream activator of c-Jun N-terminal kinase, can also under certain circumstances activate p38 MAPK (8). Whereas MKK6 preferentially activates p38β MAPK, MKK3 activates p38α and p38γ MAPKs in nonhepatic cells (20). Activation of the MKK3/p38α MAPK has been shown to be essential for cAMP-dependent thermogenic gene expression in adipocytes (37). However, it is unknown whether MKK3 is the major activator of p38α MAPK in hepatocytes.

The aim of the present study was to test the hypotheses that cAMP increases and TLC decreases PM-MRP2 via p38α and p38β MAPK, respectively, and that MKK3 is the major activator of p38α MAPK in hepatocytes. Studies in hepatocytes from MKK3-knockout animals permitted us to test the role of p38α and p38β MAPK in PM-MRP2 localization, and the results are consistent with the proposed hypothesis.

MATERIALS AND METHODS

Materials.

8-(4-Chlorophenylthio)-cyclic adenosine monophosphate (CPT-cAMP) and TLC were purchased from Sigma-Aldrich (St. Louis, MO). Commercially available antibodies used in the present study were total and pan-phospho-p38 MAPK (which detects phosphorylation of all isoforms); total and phospho-MKK4, phospho-MKK3/6, total MKK3 (Cell Signaling, Technology Danvers, MA); total p38α MAPK and actin (Santa Cruz Biotechnology, Dallas, TX); phospho-p38α MAPK (SignalChem, Richmond, BC, Canada); phospho-MKK3 (GeneTex, Irvine, CA); mouse MRP2 (Abcam, Cambridge, MA); and E-cadherin (BD Biosciences, San Jose, CA). Sulfosuccinimidyl-6-(biotin-amido) hexanoate and streptavidin beads were purchased from Pierce (Rockford, IL) and Novagen (Madison, WI), respectively. A small interfering RNA (siRNA) against p38β MAPK (L-003972-00) and nontargeting pool control siRNA (D-001810-10-05) were purchased from Dharmacon (Lafayette, CO). Lipofectamine RNAiMAX was obtained from Invitrogen (Carlsbad, CA).

Animals and cell lines.

C57BL/6 wild-type (WT) mice were purchased from The Jackson Laboratories (Bar Harbor, ME) and MKK3^−/− mice were obtained from Dr. R.J. Davis (8) and bred at the Cummings Veterinary Medical Center (Tufts University, North Grafton, MA). Liver function tests and histological analysis performed on 6- to 8-wk-old MKK3^−/− mice were normal (data not shown). A hepatoma cell line (HuH7) that overexpresses sodium-taurocholate cotransporting polypeptide (NTCP) (HuH-7 cells stably transfected with human NTCP) were generously provided by Dr. Gores (19).

Primary hepatocyte cultures.

Hepatocytes were isolated and cultured according to a previously published procedure (25). Briefly, mice (6–8 wk old, male) were anesthetized with an intraperitoneal injection of ketamine (100 mg/kg body wt) plus xylazine (10 mg/kg body wt), and the liver was perfused for 5 min with liver perfusion media (Invitrogen, Carlsbad, CA), followed by a 6-min perfusion with liver-digestion media (Invitrogen) at 37°C. The digested liver was diced in cold hepatocyte wash media (Invitrogen), passed through an 80-μm nylon mesh followed by a Percoll gradient isolation, and washed three times before cells were resuspended in cold Williams E medium containing 10% FBS, 10 units/ml penicillin, 0.1 mg/ml streptomycin, and 10 μg/ml insulin. Viability was estimated by trypan blue exclusion, and only preparations with viability in excess of 80% were used for experiments. Cells were plated overnight in 60-mm collagen-coated dishes at a density of 1.6 × 10⁶ cells per dish. The following day, hepatocytes were treated with desired reagents and analyzed as described in figure legends. The protocol for isolating hepatocytes was approved by the Tufts University Institutional Animal Care and Use Committee. The Lowry method was used to determine cell protein concentrations (29).

HuH-NTCP cells and siRNA transfection.

HuH-NTCP cells were grown in Eagle's minimum essential medium (MEM) supplemented with 10% FCS, 1.2 g/l G418, 100,000 units/l penicillin, 100 mg/l streptomycin, and 25 μg/ml amphotericin B at 37°C with 5% CO₂.

HuH-NTCP cells were transfected with scrambled siRNA and siRNA against p38β MAPK by using Lipofectamine RNAiMAX according to the manufacturer's instructions as previously reported by us (35). Briefly, after cell culturing, the medium was changed to antibiotics-free regular DMEM media, followed by addition of Opti-MEM containing Lipofectamine RNAiMAX and scrambled siRNA or siRNA p38β MAPK mixture and incubation at 37°C for 48 h. Following transfection, cells were treated with various agents as described in the figure legends.

Plasma membrane MRP2.

A cell surface protein biotinylation method was used to assess PM-MRP2 levels (34, 41, 43). Briefly, after various treatments, cell surface proteins were biotinylated by exposing hepatocytes to sulfosuccinimidyl-6-(biotin-amido) hexanoate, and a whole cell lysate was prepared. Biotinylated proteins were isolated with streptavidin-agarose beads and then subjected to immunoblot analysis to determine PM-MRP2 and E-cadherin (loading control). The level of PM-MRP2 was expressed as a relative value compared with E-cadherin.

Kinase activation assays.

Standard immunoblot analysis was used to assess the activated phosphorylation state of various kinases, and the results were expressed as the ratio of phosphorylated (active form) to total kinase (or actin), where total kinase or actin served as the loading control. Actin was used as the loading control only when immunoblot with total kinase antibodies did not work well on a stripped blot. The blots were scanned and the relative band densities were quantitated with ImageJ software from the National Institutes of Health.

Statistical analysis.

All values are expressed as means ± SE. Differences among multiple means were analyzed by one-way ANOVA, and between two means by Student's t-test with P < 0.05 considered significant.

RESULTS

Upregulation of MKK6 in MKK3^−/− mice hepatocytes.

Previous studies showed that knockout of MKK3 had no effect on the expression of MKK4 and MKK6 in fibroblasts (46) and peritoneal macrophages (30), but increased the expression of MKK6 in kidneys (28). We compared the expression levels of MKK3, MKK4, and MKK6 in WT and MKK3^−/− hepatocytes to determine whether expressions of MKK4 and MKK6 were affected. As expected, MKK3 expression was absent in MKK3^−/− mice (Fig. 1). MKK4 expression was comparable between WT and MKK3^−/− mice, but MKK6 expression increased threefold in MKK3^−/− mice compared with WT hepatocytes. Thus MKK6 but not MKK4 undergoes compensatory upregulation in hepatocytes in MKK3^−/− mice. Because the increased expression of MKK6 and the absence of MKK3 may alter the activation of p38 MAPK and hence PM-MRP2 in different ways, we determined PM-MRP2 levels and p38 MAPK activation by either cAMP or TLC in WT and MKK3^−/− hepatocytes.

Fig. 1.

MKK6 but not MKK4 is upregulated in MKK3^−/− hepatocytes. Immunoblots of lysates prepared from cultured wild-type (WT) and MKK3^−/− hepatocytes were probed for total MKK3, MKK4, and MKK6. Top: typical immunoblots; bottom: bar graph with results of densitometric analysis expressed relative to actin used as the loading control (means ± SE, n = 4). *Significantly different (P < 0.05) from respective WT values.

Effect of MKK3 on PM MRP2 and p38 MAPK.

The known effects of cAMP and TLC on PM-MRP2 and p38 MAPK are derived from studies in rat hepatocytes and human hepatic cell lines (1, 2). To confirm similar effects in mouse hepatocytes, we determined the effect of cAMP and TLC on PM-MRP2 and p38 MAPK. As in rat hepatocytes, cAMP increased, whereas TLC decreased PM-MRP2 in WT mouse hepatocytes (Fig. 2). Despite the differing results with PM-MRP2, both cAMP and TLC increased phosphorylation and thus activation of p38 MAPK in WT mouse hepatocytes (Fig. 3). In contrast to results in WT hepatocytes, cAMP did not increase PM-MRP2 or activate p38 MAPK in MKK3^−/− hepatocytes (Figs. 2 and 3), suggesting that an MKK3/p38 MAPK pathway is necessary for cAMP-induced increases in PM-MRP2. In contrast, TLC still decreased PM-MRP2 and activated p38 MAPK in MKK3^−/− hepatocytes (Fig. 2), indicating that MKK3 is not necessary for TLC-induced decreases in PM-MRP2 or p38 MAPK phosphorylation. A possible explanation for the varying effects of p38 MAPK on PM-MRP2 could be a result of selective activation of p38 MAPK isoforms via different MKKs. Hence we determined the effects of cAMP and TLC on MKK3 in WT hepatocytes and p38α MAPK in WT and MKK3^−/− hepatocytes.

Fig. 2.

Cyclic AMP-induced increases in multidrug resistance-associated protein-2 in the plasma membrane (PM-MRP2), but not taurolithocholate (TLC)-induced decreases in PM-MRP2, are MKK3 dependent. WT and MKK3^−/− hepatocytes were isolated and placed in culture overnight and then treated with CPT-cAMP (100 μM, 15 min) or TLC (10 μM, 10 min) followed by determination of PM-MRP2 by biotinylation. Top: representative immunoblots; bottom: bar graph with results of densitometric analysis (means ± SE, n = 4–7). *Significantly different (P < 0.05) from respective controls.

Fig. 3.

Cyclic AMP but not TLC-induced activation of p38 MAPK is MKK3 dependent. WT and MKK3^−/− hepatocytes were treated with CPT-cAMP (100 μM, 15 min) or TLC (10 μM, 10 min) followed by determination of phosphorylated (p-p38 MAPK) and total p38 (t-p38 MAPK) by immunoblotting. Top: representative immunoblots; bottom: results of densitometric analysis (means ± SE, n = 4–7). *Significantly different (P < 0.05) from respective controls. #Significantly different from respective WT values.

Cyclic AMP but not TLC activates MKK3 and p38α MAPK.

Cyclic AMP but not TLC activated MKK3 (i.e., increased phospho-MKK3) in WT hepatocytes (Fig. 4). To determine whether MKK3 primarily activates p38α MAPK in hepatocytes, we determined the effects of cAMP and TLC on p38α MAPK phosphorylation in WT and MKK3^−/− hepatocytes. In WT hepatocytes, cAMP, but not TLC, activated p38α MAPK (Fig. 5A). In MKK3^−/− hepatocytes, total levels of p38α were the same as in WT hepatocytes (Fig. 5B), but the basal levels of activated p38α MAPK were significantly lower than those in WT hepatocytes (Fig. 5A). These results suggest that MKK3 is likely the major activator of p38α MAPK in mouse hepatocytes (Fig. 5). Neither cAMP nor TLC activated p38α MAPK in MKK3^−/− hepatocytes (Fig. 5A). Taken together, these results suggest that cAMP but not TLC stimulates an MKK3/p38α MAPK pathway. Activation of p38 MAPK by TLC is most likely due to activation of MKKs other than MKK3. Thus we determined whether TLC or cAMP activates MKK4, or MKK6, or both.

Fig. 4.

Cyclic AMP but not TLC activates MKK3. WT hepatocytes were treated with CPT-cAMP (100 μM, 15 min) or TLC (10 μM, 10 min) followed by determination of phosphorylated MKK3 (pMKK3) and total MKK3 by immunoblotting. Top: a representative immunoblot; bottom: bar graph with results of densitometric analysis (means ± SE, n = 3). *Significantly different (P < 0.05) from controls (Con).

Fig. 5.

Cyclic AMP induced activation of p38α MAPK is MKK3-dependent. WT and MKK3^−/− hepatocytes were treated with cAMP (100 μM, 15 min) and TLC (10 μM, 10 min) followed by determination of phosphorylated p38α (p-p38α) and total p38α by immunoblotting. Top: representative immunoblots (X represents another treatment run on the same gel). Bottom: bar graph with results of densitometric analysis (means ± SE, n = 6) for p-p38α (A) and total p38α (B). *Significantly different (P < 0.05) from respective controls. #Significantly different (P < 0.05) from respective WT values.

TLC activates MKK6 in MKK3^−/− hepatocytes.

In addition to MKK3, other MKKs (MKK4 and MKK6) can also activate p38 MAPK (14, 22, 36). Thus we tested the hypothesis that cAMP or TLC (or both) activates p38 MAPK by activating MKK4, or MKK6, or both. Studies in WT hepatocytes showed that MKK4 was not activated by either cAMP or TLC (Fig. 6), indicating that MKK4 is unlikely to be involved in the activation of p38 MAPK. Also, TLC did not activate MKK3 in WT hepatocytes (Fig. 4). Thus phosphorylation of p38 MAPK in MKK3^−/− hepatocytes is most likely due to activation of MKK6. It was not possible to directly determine the effect on MKK6 activation in WT hepatocytes because the available antibody against phospho (active)-MKK6 failed to produce consistent results in mouse hepatocytes. However, an antibody against the phosphorylated (active) form of both MKK3 and MKK6 (MKK3/6) combined is available and produced consistent results with lysates from mouse hepatocytes. Because activation of MKK3/6 in the absence of MKK3 (MKK3^−/− hepatocytes) represents activation of MKK6, we used the phospho-MKK3/6 antibody to determine whether cAMP or TLC activates MKK6 in MKK3^−/− hepatocytes. Our results show that TLC, but not cAMP, activated MKK6 in MKK3^−/− hepatocytes (Fig. 7). Because TLC did not activate MKK3 or MKK4 in WT hepatocytes, it would appear that TLC activates p38 MAPK through MKK6 activation in WT and MKK3^−/− hepatocytes. Because MKK6 exerts many of its effects by activating p38β MAPK (20), we tested whether TLC-mediated decreases in PM-MRP2 are mediated by p38β MAPK.

Fig. 6.

Cyclic AMP and TLC do not activate MKK4. WT hepatocytes were treated with cAMP (100 μM, 15 min) or TLC (10 μM, 10 min) followed by determination of phosphorylated MKK4 (pMKK4) and total MKK4 by immunoblotting. Top: a representative immunoblot; bottom: bar graph with results of densitometric analysis (means ± SE, n = 4).

Fig. 7.

TLC but not cAMP activates MKK6 in MKK3^−/− hepatocytes. MKK3^−/− hepatocytes were treated with cAMP (100 μM, 15 min) or TLC (10 μM, 10 min) followed by determination of phosphorylated MKK3 and MKK6 (MKK3+6) by immunoblotting. Top: a representative immunoblot; bottom: bar graph with results of densitometric analysis (means ± SE, n = 4). *Significantly different (P < 0.05) from respective controls.

TLC-induced decreases in PM-MRP2 are mediated by p38β MAPK.

We used the HuH-NTCP cell line to determine the role of p38β MAPK in TLC-induced decreases in PM-MRP2 for two reasons. First, a chemical inhibitor for p38β MAPK is not available for studies in mouse hepatocytes, and second, TLC has been shown to decrease PM-MRP2 in this cell line (43). HuH-NTCP cells were transfected with an siRNA directed against p38β MAPK, and this resulted in a decrease in the expression of p38β MAPK to 25% of the control with scrambled siRNA. As in WT mouse hepatocytes, TLC decreased PM-MRP2 by 30% in HuH-NTCP cells. However, knockdown of p38β MAPK abrogated the effect of TLC on PM-MRP2 (Fig. 8B). These results suggested that p38β MAPK is necessary for TLC-induced decreases in PM-MRP2.

Fig. 8.

Small interfering RNA (siRNA) knockdown of p38β MAPK abrogates TLC-induced decreases in PM-MRP2. HuH-NTCP cells were transfected with scrambled (scr.) siRNA or siRNA directed toward p38β MAPK, and knockdown of p38β MAPK was confirmed by immunoblot as shown in the representative immunoblot (A). Values in A represent relative values of p38β MAPK/actin compared with untreated cells transfected with scr. siRNA. Forty-eight hours following transfection, cells were treated with TLC (10 μM for 25 min) followed by determination of PM-MRP2. B: representative immunoblot for PM-MRP2 and results of densitometric analysis (means ± SE, n = 4). *Significantly different (P < 0.05) from control values in the absence of TLC. #Significantly different (P < 0.05) from control values in the presence of TLC. Data were analyzed by one-way ANOVA.

DISCUSSION

The aim of the present study was to test the hypotheses that isoform-specific activation of p38 MAPK is responsible for the opposing effects of cAMP and TLC on PM-MRP2, and MKK3 is the major activator of p38α MAPK in hepatocytes. These hypotheses are supported by our studies as discussed below.

The present study provides an explanation for the opposing effects of p38 MAPK activation on PM-MRP2 by cAMP and TLC. Our results suggest that this is a result of activation of p38α MAPK by cAMP and p38β MAPK by TLC. Of the four p38 MAPK isoforms, only α and β isoforms are present in human and mouse liver (20, 44). Thus activation of p38 MAPK in hepatocytes can result from either p38α or p38β, or both MAPKs. Our study shows that activation of p38α MAPK by MKK3 is necessary for cAMP-induced MRP2 translocation to the plasma membrane. This conclusion is supported by results that cAMP increases PM-MRP2; activates MKK3, p38 MAPK, and p38α MAPK in WT hepatocytes; and cAMP does not increase PM-MRP2 or activate p38 MAPK or p38α MAPK in MKK3^−/− hepatocytes. In contrast, TLC decreases PM-MRP2 and activates p38 MAPK, but does not activate MKK3 or p38α MAPK in WT hepatocytes. In addition, TLC still decreases PM-MRP2 in MKK3^−/− hepatocytes, but failed to decrease PM-MRP2 when p38β is knocked down. These results suggest that TLC-induced decreases in PM-MRP2 are not mediated via MKK3 and p38α MAPK, but is instead mediated via p38β MAPK. Because TLC activates MKK6 but not MKK3 or MKK4, and MKK6 preferentially activates p38β MAPK (20), it is highly likely that TLC induces retrieval of PM-MRP2 via an MKK6/p38β MAPK pathway.

Results of the present study provide further insights into the activation of p38 MAPK isoforms by upstream kinases in hepatocytes. Our studies of MKK3^−/− hepatocytes show that expression of p38α MAPK is not affected, but the basal level of phosphorylated (active) p38α MAPK is decreased to 37% of the value in WT hepatocytes (Fig. 5B). This decrease was observed despite a threefold increase in MKK6 expression in MKK3^−/− hepatocytes (Fig. 1), suggesting that MKK3 is the major activator of p38α MAPK in hepatocytes. This conclusion is consistent with the findings that MKK3 does not activate p38β MAPK in COS-7 cells (15, 20) and brown preadipocytes (37). The residual p38α MAPK activity in the absence of MKK3 may be maintained by MKK4 and MKK6, because MKK4 and MKK6 have been shown to activate p38α MAPK in other cell types (15, 20).

Because both p38 MAPK and PKC have been implicated in membrane trafficking in hepatocytes, the localization of PM-MRP2 may involve cross-talk between isoforms of PKC and p38 MAPK. Several studies have shown that isoform-specific activation of PKCs modulates MPR2 trafficking. PKCε is involved in the retrieval of MRP2 by TLC (6, 43), whereas PKCδ promotes the insertion of MRP2 to the plasma membrane by cAMP (34, 40). Conventional PKCs acting upstream of p38 MAPK are involved in E17G-induced retrieval of bile salt export pump (BSEP) and MRP2 from the plasma membrane (7). Furthermore, PKCδ and PKCε activate p38 MAPK in ischemic preconditioning in rat hepatocytes (4), and PKC inhibitors block hypoxia-induced activation of p38 MAPK in chicken hepatocytes (27). These studies suggest that interactions between p38 MAPK and PKCs occur in hepatocytes. However, whether p38α and p38β MAPKs are differentially affected by PKC isoforms has not been reported. Thus further studies will be required to define whether PM-MRP2 localization involves isoform-specific effects of PKCs on p38α and p38β MAPKs in hepatocytes.

Results of the present study may be of clinical significance in view of the reported hepatotoxicity of p38 MAPK inhibitors in human clinical trials. These inhibitors, which were developed to treat inflammatory diseases, inhibit the α and β, but not the γ or δ isoforms of p38 MAPK (26). Although the inhibitors were effective in preclinical models of inflammatory disorders, hepatotoxicity interfered with their clinical development (13, 26). Considering the important role of transporter function in drug-induced liver disease, the hepatotoxicity of p38 inhibitors may be related to inhibition of p38 MAPK-mediated plasma membrane localizations of MRP2 and BSEP (11). A better knowledge of the role of MKKs and p38 MAPK isoforms in hepatocytes may allow for development of less toxic kinase inhibitors for the treatment of inflammatory diseases.

In summary, this study shows that plasma membrane localization of MRP2 is differentially regulated by p38α and p38β MAPK in hepatocytes.

GRANTS

This study was supported in part by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-90010 to M.S. Anwer and DK-65975 to C.R. Webster.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.S.A. conception and design of research; S.W.P. performed experiments; C.M.S., S.W.P., and M.S.A. analyzed data; C.M.S., C.R.W., and M.S.A. interpreted results of experiments; C.M.S., S.W.P., and M.S.A. prepared figures; C.M.S. and M.S.A. drafted manuscript; C.M.S., S.W.P., C.R.W., and M.S.A. edited and revised manuscript; C.M.S. and M.S.A. approved final version of manuscript.