Microbial-Derived Butyrate Promotes Epithelial Barrier Function through IL-10 Receptor–Dependent Repression of Claudin-2

The mammalian gastrointestinal tract is home to trillions of bacteria that are separated from the host immune system by a single layer of intestinal epithelial cells (IECs). A finely regulated commensal relationship exists within the intestinal mucosa, where microbes, essential for host health, can also initiate and perpetuate mucosal disease (1). These microbes reside in the distal gut where they metabolize undigested materials and benefit the host through local synthesis of short-chain fatty acids (SCFAs). The three primary SCFAs are acetate, propionate, and butyrate. Butyrate, although the least abundant of the three, is the preferred metabolic substrate for IECs and can reach luminal concentrations of 30 mM in the colon (2). Dysbiosis, in part characterized by a decrease of butyrate-producing microbial species, has been strongly associated with colonic disease, including inflammatory bowel disease (IBD) (3–5). Yet, mechanisms to explain the inverse association between butyrate and disease are not well understood.

In a healthy colon, there exists a fine balance between pro- and anti-inflammatory mediators. Proinflammatory cytokines, including TNF-α and IL-1β, have been implicated in the pathogenesis of IBD (6, 7). However, this inflammation is normally limited by anti-inflammatory cytokines, including IL-10. IL-10 is produced by IECs and leukocytes, and inhibits the secretion of TNF-α and IFN-γ (8). The IL-10 receptor α subunit (IL-10RA) is expressed on numerous cell types, including IECs (9, 10). The functional IL-10 receptor is comprised of the ligand-binding α subunit (IL-10RA) and a β subunit (IL-10RB), which is shared with other IL-10 receptor family members (11). The binding of IL-10 to IL-10R results in activation of the JAK-STAT signaling pathway, and transactivation of latent transcription factors. Numerous studies have shown that STAT3 activation is critical to the ligand’s anti-inflammatory activity (12, 13). Mice deficient in IL-10 or IL-10RA develop severe colitis and mice conditionally lacking intestinal epithelial IL-10RA show increased susceptibility to colitis (14–16).

Although the host has developed defense mechanisms to counter threats from prokaryotes in the gut, bacterial pathogens and commensals have also evolved countermeasures to modulate the host intestinal epithelium. Recently, we have shown that microbial-derived butyrate increases mitochondrial-dependent oxygen consumption in IECs, stabilizes hypoxia-inducible factor (HIF), and induces expression of HIF-target genes that augment barrier function (17). Further, butyrate is a well-known histone deacetylase inhibitor (HDACi) and, through this mechanism, downregulates proinflammatory mediator expression by macrophages and increases regulatory T cell differentiation (18, 19).

Given the role of barrier dysfunction in colonic disease, we sought to define the mechanisms of SCFA regulation of barrier. Our studies reveal that butyrate promotes epithelial barrier formation through IL-10RA–mediated repression of permeability-promoting claudin-2 (Cldn2). Cldn2, one of 27 mammalian claudins that regulate barrier, forms a paracellular channel for small cations and water, and is upregulated in IBD, contributing to diarrhea via a leak-flux mechanism (20). These findings provide a mechanistic link between host-microbial cross-talk within the mucosa and describe a mutualism between butyrate and intestinal homeostasis.

Both Caco2 [#HTB-37; American Type Culture Collection (ATCC), Manassas, VA] and T84 (#CCL-248; ATCC) human epithelial cell lines were obtained from ATCC and maintained in 95% air with 5% CO₂ at 37°C according to instructions provided by ATCC. Where indicated, cells were cultured on polyester transwell inserts (CoStar, Cambridge, MA). Acetic acid, propionic acid, and butyric acid from Sigma-Aldrich (St. Louis, MO) were added to sterile filtered HBSS (Sigma-Aldrich) with the addition of 10 mM HEPES, and pH adjusted with sodium hydroxide to pH 7.4. Transepithelial electrical resistances (TEER) were monitored using an EVOM2 Voltohmmeter (World Precision Instruments). Cytokines were purchased from R&D Systems (Minneapolis, MN) and used at indicated concentrations. To generate short hairpin RNA (shRNA)–targeting IL-10RA and oeIL10-RA cell lines, lentiviruses encoding shRNA–targeting IL-10RA or IL-10RA open reading frames (ORFs) were used (MISSION TRC shRNA or CCSB-Broad, University of Colorado Functional Genomics Facility). We transduced T84 cells using previously described protocols (21). Stable integration was maintained by puromycin selection (3 μg/ml). Knockdown and overexpression were confirmed by PCR analysis.

TRIzol from Invitrogen (Grand Island, NY) was used to isolate RNA from cultured cells and tissues exposed to SCFA, trichostatin A (TSA), and stattic (Sigma). iScript cDNA Synthesis Kit from Bio-Rad Laboratories (Hercules, CA) and SYBR Green from Applied Biosystems (Warrington, U.K.) were used for cDNA synthesis and real-time PCR analysis (7900HT; Applied Biosystems) using the following mouse and human forward (Fwd) and reverse (Rev) primer sequences:

• hIL-10RA Fwd: 5′-CCCTGTCCTATGACCTTACCG-3′, Rev: 5′-CACACTGCCAACTGTCAGAGT-3′; hSOCS3 Fwd: 5′-GGCCACTCTTCAGCATCTC-3′, Rev: 5′-ATCGTACTGGTCCAGGAACTC-3′; hCLDN1 Fwd: 5′-CCAGTCAATGCCAGGTACGAAT-3′, Rev: 5′-TTGGTGTTGGGTAAGAGGTTGTT-3′; hCLDN2 Fwd: 5′-CTCCTGGGATTCATTCCTGTT-3′, Rev: 5′-TCAGGCACCAGTGGTGAGTAGA-3′; hCLDN3 Fwd: 5′-CCACGCGAGAAGAAGTACA-3′, Rev: 5′-GTAGTCCTTGCGGTCGTAGC-3′; hCLDN7 Fwd: 5′-AATGTACGACTCGGTGCTCG-3′, Rev: 5′-AATCTGATGGCCATACCAGG-3′; hOCLN Fwd: 5′-GCTACGGAAGTGGCTATGG-3′, Rev: 5′-GCGGCAATGAAACAAAAG-3′; hECAD Fwd: 5′-GCCCATTTCCTAAAAACCTG-3′, Rev: 5′-CTCTGTCACCTTCAGCCATC-3′; hJAMA Fwd: 5′-CCTGGGAATCTTGGTTTTTG-3′, Rev: 5′-GGAATGACGAGGTCTGTTTG-3′; hTJP1 Fwd: 5′-TGGTGTCCTACCTAATTCAACTCA-3′, Rev: 5′-CGCCAGCTACAAATATTCCAACA-3′; hIL-10 Fwd: 5′-AATAAGGTTTCTCAAGGG-3′, Rev: 5′-AGAACCAAGACCCAGACA-3′; hACTB Fwd: 5′-CCTGGCACCCAGCACAAT-3′, Rev: 5′-GCCGATCCACACGGAGTACT-3′; mIl10ra Fwd: 5′-CCCATTCCTCGTCACGATCTC-3′, Rev: 5′-TCAGACTGGTTTGGGATAGGTTT-3′; mActb Fwd: 5′-TACGGATGTCAACGTCACAC-3′, Rev: 5′-AAGAGCTATGAGCTGCCTGA-3′.

Cells and whole tissue were extracted into Tris lysis buffer with protease inhibitor (Roche), disrupted by sonication, and quantified for normalization using Pierce BCA protein assay kit (Thermo Fisher Scientific). Abs were used at manufacturer-recommended concentrations and included: anti–IL-10RA (rabbit polyclonal), anti-Cldn2 (rabbit polyclonal), and anti–β-actin from Abcam; IL-10 (human), anti-acH3K9 (C5B11, rabbit monoclonal), anti-Stat3 (79D7, rabbit monoclonal), and anti-pStat3 (Y705, rabbit polyclonal) from Cell Signaling; and anti–IL-10RA (rabbit polyclonal) from Thermo Fisher Scientific. The Western blotting Abs, peroxidase goat anti-rabbit IgG and peroxidase goat anti-mouse IgG, were purchased from the Jackson Laboratory (Bar Harbor, ME). Western blotting substrates, Pierce ECL and SuperSignal West Femto, were purchased from Thermo Fisher Scientific.

To localize IL-10RA, T84 cells were exposed to butyrate or buffer control, fixed, and processed for microscopy as described (16). Cells were localized with anti–IL-10RA followed by Alexa Fluor 488 secondary Ab and counter-stained with Alexa Fluor 546 (Invitrogen). Fluorescence images were obtained using an AxioCam MRc5 attached to an AxioImager A1 microscope (Zeiss, Oberkochen, Germany).

An IL-10RA–luciferase reporter plasmid and empty vector control (Switchgear Genomics) were transfected into Caco-2 cells using Lipofectamine 3000 transfection reagent (Invitrogen) using the manufacturer’s recommended protocol. The day following transfection, cells were treated with the designated dose of butyrate. Cells were lysed the day after treatment and luciferase (Promega) was measured and normalized to protein by BCA. Lightswitch (Active Motif) was used to quantify reporters from Switchgear Genomics.

Wild-type C57/B6 mice were purchased from the Jackson Laboratory. Mice were housed in accordance with guidelines from the American Association for Laboratory Animal Care and Research Protocols, and experiments were approved by the Institutional Animal Care and Use Committee of the University of Colorado.

Antibiotic (Abx) treatment consisted of 3 d of 200 μl per mouse oral gavage of Abx mixture ampicillin (1 mg/ml), gentamicin (1 mg/ml), metronidazole (1 mg/ml), neomycin (1 mg/ml), and vancomycin (0.5 mg/ml) from Sigma-Aldrich.

Original germ-free (GF) C57Bl/6 breeding stocks were obtained from the National Gnotobiotic Rodent Resource Center at the University of North Carolina. Mice were bred and maintained in flexible vinyl positive pressure isolators within the Gnotobiotic Facility at the University of Colorado on a 12-hour light cycle, and given free access to autoclaved water and autoclaved chow (Teklad Global Soy Protein-Free Extruded Rodent Diet, 2020sx, Harlan Laboratories, Indianapolis, IN). Animal tissues were extracted into Tris lysis buffer followed by sonication, and protein homogenates were stored at −80°C until use. RNA extraction was performed as described above.

GraphPad Prism 5 software produced by GraphPad (La Jolla, CA) was used to create figures and perform statistical analyses, including one-way ANOVA, Tukey post hoc test, and Student t test. Error bars represent SEM unless otherwise stated, and significance was determined at p < 0.05.

Guided by our recent findings that butyrate decreases the flux of labeled tracers across IEC monolayers (17), we profiled the influence of SCFAs on intestinal barrier formation. As shown in Fig. 1, T84 IECs were plated on transwell inserts and exposed to physiologically relevant concentrations of three major SCFAs: butyrate (5 mM), propionate (20 mM), and acetate (20 mM). We monitored barrier formation, as measured by TEER, over 72 h. We found that butyrate, but neither acetate nor propionate, significantly enhanced both the kinetics and magnitude of barrier formation over 72 h (Fig. 1, p < 0.001), suggesting that butyrate selectively promotes intestinal barrier formation.

Our previous work demonstrated that IL-10RA is central to the formation of the epithelial barrier and that knockdown of IL-10RA results in a loss of barrier formation (16). Based on these findings, we hypothesized that butyrate-enhanced barrier formation is mediated by induction of functional IL-10RA. To test this hypothesis, human IECs (Caco-2 and T84 cells) were exposed to acetate, propionate, and butyrate for up to 24 h and IL-10RA transcript was measured by quantitative PCR (qPCR). IL-10RA transcript increased 140.8 ± 39.3-fold in Caco-2 cells (Fig. 2A, p < 0.05) and 97.0 ± 21.5-fold in T84 cells (Fig. 2B, p < 0.01) after 24 h of 5 mM butyrate treatment. Propionate also increased IL-10RA expression by 40.0 ± 14.1-fold in T84 cells (Fig. 2B, p < 0.05). Further, we observed 118.6 ± 6.2-fold IL-10RA mRNA induction Caco-2 cells treated with 5 mM butyrate for 18 h (Fig. 2C, p < 0.01). We saw a 28.3 ± 7.3-fold dose-response in T84 cells (Fig. 2D, p < 0.01) with the same exposure, and chose to focus on butyrate for ongoing studies.

To further characterize the relationship between IL-10RA and butyrate, we exposed Caco-2 and T84 cells to increasing concentrations of butyrate and examined IL-10RA protein by immunoblot (Fig. 2E, 2F). Our results revealed a dose-dependent induction of IL10-RA by butyrate. As shown in Fig. 2G, analysis by immunofluorescence microscopy also revealed increased IL-10RA protein following exposure to 10 mM butyrate treatment in T84 cells. We next evaluated whether butyrate could transactivate the IL-10RA promoter. We transfected an IL-10RA–luciferase reporter construct in Caco-2 cells, treated the cells with increasing concentrations of butyrate, and discovered 1.6-fold increased IL-10RA–luciferase reporter activity with 2.5 mM butyrate (Fig. 2H, p < 0.001). Additionally, butyrate-elicited IL-10RA was shown to be functional based on IL-10–mediated 3.0-fold induction of suppressor of cytokine signaling 3, a well-documented downstream target gene in intestinal epithelia (Fig. 2I, p < 0.001) (16, 22). Finally, we determined whether exogenous IL-10 would further influence barrier formation in butyrate-treated monolayers. As shown in Supplemental Fig. 1, pretreatment of T84 cells with butyrate (5 mM, 24 h) followed by IL-10 (10 ng/ml) resulted in a significant increase in barrier formation compared with both media alone (p < 0.001) and butyrate alone (p < 0.05). The addition of anti–IL-10 (p < 0.05), but not nonspecific IgG (p = not significant), decreased the rate of barrier formation compared with butyrate alone. These results identify the induction of functional IL-10RA by butyrate.

We next examined mechanisms of butyrate induction of IL-10RA. In this study, we analyzed if STAT3, a downstream target of IL-10RA, could regulate IL-10RA through feed-forward mechanisms, similar to IL-10 (23). T84 cells were treated with butyrate for 18 h followed by IL-10 for 1 h before harvest. Immunoblot of lysates revealed that butyrate alone and butyrate with IL-10 cytokine both phosphorylate STAT3 (Fig. 3A). Further, pretreatment with stattic, a STAT3 activation inhibitor, inhibited butyrate-induced STAT3 activation. In addition, stattic followed by butyrate abolished butyrate-mediated induction of IL-10RA mRNA (Fig. 3B, p < 0.05) (24). These results indicate that butyrate regulates IL-10RA through STAT3 activation.

Additionally, butyrate is a potent HDACi (19). In this study, we investigated whether butyrate mediates IL-10RA expression through HDAC inhibition. Exposure of T84 cells to TSA, a well-known HDACi, at 500 ng/ml increased IL-10RA mRNA 44.4 ± 4.3-fold by qPCR (Fig. 3C, p < 0.05) (25). The addition of butyrate also led to significantly increased acH3K9, an active chromatin mark, in T84 cells (Fig. 3D) (26). These results are supported by a previous analysis that demonstrated increased IL-10RA expression from an array investigating the role of Warburg effect on butyrate-mediated colonocyte proliferation (Gene Expression Omnibus [GEO] accession number GSE 41113) (27). Together, these findings support the HDAC inhibitory function of butyrate in induction of IL-10RA.

Recently, activation of the butyrate-binding G-protein coupled receptor 109a (GPR109a) in the colon was found to suppress colonic inflammation and promote differentiation of regulatory T cells and IL-10–producing T cells (28). GPR109a is a Gi/Go protein-coupled signaling receptor that is inhibited with pertussis toxin (PTX) (29). To determine if G-protein coupling contributes to the induction of IL-10RA, we pretreated Caco-2 and T84 cells with PTX, followed by butyrate, and measured IL-10RA mRNA by qPCR. As shown in Supplemental Fig. 2, IL-10RA expression was not changed by PTX pretreatment, suggesting that butyrate induction of IL-10RA is independent of GPR109a.

We next investigated how changes in IL-10RA expression, as in the case with butyrate levels, influenced barrier formation. For these purposes, we generated cell lines with stable loss and gain of IL-10RA using lentiviral approaches. T84 cells were transduced with shRNA-encoding lentivirus against IL-10RA (knockdown) or IL-10RA lentiORF (overexpression). The shRNA targeting IL-10RA and lentiORF (oeIL-10RA) resulted in significantly attenuated (0.68 ± 0.11-fold) and enhanced (3703 ± 1101-fold) mRNA relative to shControl, respectively (Fig. 4A, 4B, p < 0.01, p < 0.05). Functionally, IL-10RA knockdown practically abolished barrier formation over 72 h both in the absence and presence of butyrate (Fig. 4C, p < 0.001). Conversely, IL-10RA overexpression significantly enhanced barrier formation over 72 h as measured by TEER, and the kinetics are further enhanced in the presence of butyrate (Fig. 4D, p < 0.001). To define if these results paralleled an increase in IL-10, we examined the induction of IL-10 by butyrate. As shown in Supplemental Fig. 3, analysis of IL-10 expression revealed that butyrate induces epithelial IL-10 mRNA by as much as 30 ± 2.5-fold (p < 0.001). This induction, however, was not reflected as measurable protein in cell culture supernatants from butyrate-treated cells (0.02 ± 0.001 versus 0.01 ± 0.001 pg/ml and 0.03 ± 0.003 versus 0.02 ± 0.003 pg/ml for media versus butyrate in Caco2 and T84 cells, respectively, p = not significant). These results indicate that barrier formation is dependent on epithelial expression of IL-10RA.

We next examined the mechanism of barrier regulation by butyrate alone and butyrate-induced IL-10RA. To this end, we exposed Caco-2 and T84 cells to butyrate and profiled the expression of a panel of junctional proteins known to be important in barrier function: Cldn1, Cldn2, Cldn3, Cldn7, occludin, E-cadherin, JAM-A, and ZO-1 (30). We have used this approach successfully in the past to define targets for barrier regulation by HIF (31). Our studies revealed that butyrate preferentially repressed the expression of Cldn2 relative to control in both Caco2 (0.22 ± 0.14) and T84 cells (0.08 ± 0.07) (Fig. 5A, 5B, p < 0.001). This observation extended to the protein level, where Cldn2 protein decreased >70% by densitometry in response to physiologic concentrations of butyrate in T84 cells (Fig. 5C).

Interestingly, we found that Cldn2 repression is dependent on IL-10RA. As shown in Fig. 5C, analysis of Cldn2 protein in shIL-10RA and oeIL-10RA revealed a reverse correlation between Cldn2 and IL-10RA. Moreover, inhibition of STAT3 with stattic reversed the butyrate-mediated repression of Cldn2, suggesting that IL-10RA signaling through STAT3 mediates this response (Fig. 5D). These results demonstrate that butyrate enhances barrier through IL-10RA–mediated repression of Cldn2 and reveals a novel mechanism of barrier regulation through host-microbe cross-talk.

Finally, we determined the physiologic relevance of these findings in vivo. In this study, we employed Abx-treated and GF mice to determine the impact of the microbiota on Il10ra expression. We have previously demonstrated that this Abx mixture resulted in a 92 ± 4% decrease in fecal butyrate levels (17). Administration of broad-spectrum Abx by oral gavage resulted in a 63 ± 7.2% and 55 ± 8.5 decrease in colonic and ileal Il10ra, respectively (Fig. 6A, 6B, p < 0.05 for each). Examination of Il10ra mRNA in colon scrapings enriched in epithelial cells from GF mice revealed attenuated expression of Il10ra, but did not reach statistical significance (p = 0.2, Supplemental Fig. 4). Likewise, as shown in Supplemental Fig. 4, the administration of a butyrate supplement (tributyrin) to Abx-treated mice, as we have done in the past (17), did not significantly change the expression of either Cldn2 or IL10ra (p = 0.17, p = 0.13 for Cldn2 and Il10ra, respectively).

We also sought to test this hypothesis by mining publicly available data sets. We discovered a time-dependent increase in colonic Il10ra rank and mRNA from array analyses investigating immune responses during conventionalization of GF mice (GEO accession number GDS4319) (Fig. 6C, 6D) (32–34). Taken together, these findings confirm in vitro results and implicate luminal-derived signals, such as butyrate, as a significant regulator of Il10ra expression in vivo.

An understanding of the interactions between the enteric microbiota and distal gut is an area of intense investigation. It is clear that microbial-derived factors contribute significantly to human health and that coevolution has benefited both the host and the microbe (35). It has also become evident that changes to microbial community structures can shift homeostasis and contribute to a broad spectrum of diseases, including IBD (36). Recent studies have revealed that IBD is associated with significant depletion of butyrate-producing species. Given that IBD is also strongly associated with barrier dysfunction (37), we sought to understand how butyrate regulates epithelial barrier function. In this study, we report that physiologic concentrations of butyrate repress Cldn2 expression in an IL-10RA–dependent manner to enhance epithelial barrier function, which may serve as a mechanism to prevent unwanted host immune responses against beneficial butyrate-producing microbes in the gut.

Butyrate, an end product of bacterial metabolism, typically constitutes up to 20% of SCFA in the human colon with absolute concentrations above 10 mM in human feces (2). Butyrate has been demonstrated to benefit the host in a number of ways. First, it is the preferential metabolic substrate in IECs and, through oxidative phosphorylation, depletes local oxygen to the extent that HIF is stabilized and enhances IEC barrier via HIF (2, 17). Second, butyrate activates the GPR109a receptor to suppress colonic inflammation and carcinogenesis (28). GPR109a signaling promotes anti-inflammatory properties in colonic macrophages and dendritic cells, and enables differentiation of regulatory T cells and IL-10–producing T cells. Third, through its actions as an HDACi, butyrate stimulates anti-inflammatory mechanisms that may also promote the restoration of mucosal barrier function through decreased inflammatory cytokine production (38, 39). Thus, a role for butyrate in mucosal homeostasis is well established. However, the molecular details of butyrate actions on epithelial barrier function remain incompletely understood.

We confirmed previously published findings that butyrate promotes barrier formation (17). This action of butyrate on increasing both the rate and the magnitude of barrier formation was reminiscent of recent work implicating epithelial IL-10RA signaling (16). Extensions of these initial findings revealed that butyrate, and to a lesser extent propionate, prominently induces IL-10RA. Mice deficient in IL-10 or IL-10R develop spontaneous severe colitis and mutations in either IL-10 or IL-10R have been shown to result in severe cases of human IBD (40–42). Previous studies by Kominsky et al. (16) looking at this receptor on IECs revealed a number of surprising features. First, IL-10RA expression was discovered based on a screen of IFN-inducible genes. This was unexpected given that IFN is one of the more predominant inducers of barrier dysfunction (43). Second, these studies revealed that IL-10RA expression is polarized to the apical surface of IECs in vitro and in vivo. Third, conditional deletion of IL-10RA in IECs revealed both increased baseline permeability and a high susceptibility to DSS-induced colitis, suggesting an important role for IL-10RA in barrier function and recovery following insult. Along with our findings that decreased IL-10RA correlates with decreased luminal butyrate (via Abx treatment or GF conditions), we have identified an additional role for butyrate in the homeostatic maintenance of IECs via IL-10RA.

Butyrate transmits signals to the mucosa through a number of different mechanisms, including surface G-protein coupled receptor (GPCR) stimulation (28), HDAC inhibition (39), and HIF stabilization (17). The primary butyrate signaling receptor is GPR109A, a Gai-coupled GPCR that also functions as the niacin receptor (28). Based on inhibition of Gαi using cholera toxin, we found no role for GPCR signaling in the induction of IL-10RA. Based on direct evidence of H3K9 acetylation and the use of TSA, we concluded that one signaling mechanism for the induction of IL-10RA was HDACi, a function for which butyrate is well established (19). It is interesting to note that propionate and acetate also increased IL-10RA mRNA, which could be due to their ability to inhibit HDACs to a lesser degree, although they are far less potent than butyrate as HDACis (44, 45). We did not find an induction of IL-10RA with hypoxia treatment nor HIF stabilization (data not shown).

A surprising observation was the absolute dependence of IL-10RA expression for epithelial barrier formation. Knockdown and overexpression of IL-10RA using lentiviral approaches revealed a direct correlation between IL-10RA expression and the magnitude of barrier formation. Although the addition of exogenous IL-10 enhanced the butyrate response, IL-10RA also promoted barrier formation in the absence of added IL-10. For a number of reasons, this observation was perplexing. We were not able to measure appreciable amounts of IL-10 in cell culture supernatants. Butyrate readily induced IL-10 mRNA, but no increase in protein was observed. Although we and others have shown that epithelial cells can make and release IL-10 in response to a number of stimuli (46–48), it is possible that butyrate also induces IL-10 mRNA degradation. IL-10 mRNA, for example, is known to be degraded through processing at the 3′UTR by live probiotic bacteria including Lactobacillus paracasei (49), a known butyrate producer (50). It is also possible that the low concentrations of IL-10 protein measured in tissue culture bulk flow are sufficient at the cell surface to activate butyrate-induced IL-10RA. Additional evidence was provided using the small molecule STAT3 inhibitor, stattic, which decreased STAT3 phosphorylation and attenuated butyrate-induced IL-10RA induction. It is also notable that the addition of butyrate to cells overexpressing IL-10RA further enhanced barrier formation, suggesting that butyrate may act through additional mechanisms to increase IEC barrier formation. Although we do not know the exact nature of this additional signaling, it is possible that HDACi activity of butyrate could be involved. Butyrate was recently shown to promote barrier in T84 cells via HDAC inhibition and induction of TWIK-related potassium channel-1 (51).

As an extension of these results demonstrating IL-10–dependent increases in barrier induction by butyrate, we employed an unbiased screen of epithelial junction proteins to identify mechanisms of barrier development, and identified butyrate-mediated repression of Cldn2 in both Caco2 and T84 cells (52). Among the claudin family of tight junction proteins, Cldn2 is considered “leaky.” An understanding of the regulatory pathways of Cldn2 expression is of significant interest because its expression is increased in IBD, celiac disease, and HIV enteropathy and colonic cancer (20). It is noteworthy that butyrate has been previously mentioned in the regulation of Cldn2 (38, 53), without insight into mechanism. The present studies add significantly to this work and identify a role for butyrate-dependent IL-10R signaling in the repression of Cldn2. Given the reciprocal relationship between loss of Cldn2 and compensation by other tight claudins (54), these studies implicate an important role for butyrate in maintaining the repression of Cldn2 and maintenance of barrier function through this mechanism (20).

In a final set of experiments, we examined the role of butyrate on IL-10RA expression in vivo. To this end, IL-10RA expression was examined in mice exposed to broad-spectrum Abx, conditions that we have shown to result in nearly unmeasurable colonic butyrate (17). These conditions were found to significantly decrease epithelial IL-10RA expression in mice. Parallel studies in GF mice supported the results with Abx. Data mined from GEO Profiles similarly revealed increased colonic IL-10RA mRNA during conventionalization of GF mice. IL-10RA has already been shown to play a central role in control of homeostatic barrier function as IEC-specific IL-10RA^−/− mice have significant increases in baseline permeability and systemic inflammatory responses. These results are further evidence that microbial-derived butyrate enhances IL-10RA to promote barrier and dampen immune responses.

Taken together, our results provide insight into how butyrate regulates IEC barrier and identifies a new mechanism by which microbes promote homeostasis in the distal gut. We demonstrate that butyrate, through Stat3 activation and HDAC inhibition, represses Cldn2 in an IL-10RA–dependent manner. This suggests that certain commensal bacteria may be signaling that they are nonpathogenic by producing butyrate, which leads to increased IL-10RA signaling and dampening of the host immune system. In the absence of butyrate-producing species, as occurs in IBD, barrier may be significantly compromised. These results further suggest that administration of butyrate (e.g., tributyrin) and/or promoting butyrate-producing bacteria may be of therapeutic benefit (55).

The authors acknowledge Yaoxing Li for contributions to performing the research presented here.

The authors have no financial conflicts of interest.