GLP-1R Agonists Modulate Enteric Immune Responses Through the Intestinal Intraepithelial Lymphocyte GLP-1R

The enteroendocrine system comprises a complex network of specialized endocrine cells situated within the epithelial lining of the stomach and small and large intestine. Gut endocrine cells synthesize and secrete dozens of peptide hormones that function as local neurotransmitters, propagating signals through engagement of the enteric nervous system. Enteroendocrine-derived hormones also act as classical hormones controlling metabolism through regulation of energy intake, digestion, absorption, and the disposal and storage of digested nutrients (1). The majority of gut hormones are secreted at low basal levels in the fasting state, and feeding produces a rapid rise in circulating levels, enabling activation of gut hormone action in response to energy intake. Plasma levels of some gut hormones, including the glucagon-like peptides, also increase rapidly in the presence of intestinal injury or mucosal inflammation (2).

Proglucagon-producing L cells are among the most intensively studied enteroendocrine cells predominantly localized to the distal small bowel and colon (3). Proglucagon liberates multiple proglucagon-derived peptides with unique biological activities, including glicentin, oxyntomodulin, and GLP-1 and GLP-2 (4). Although the actions of glicentin and oxyntomodulin remain incompletely understood, both GLP-1 and GLP-2 exert well-defined actions through distinct G-protein–coupled receptors. GLP-1 controls gut motility and food intake and stimulates insulin while simultaneously inhibiting glucagon secretion (4,5). In contrast, GLP-2 acts more locally within the gut to increase nutrient absorption and maintain the integrity of the gut epithelium (6).

Considerable evidence supports the importance of the GLP-1 system as an integral component of the incretin axis. Indeed, transient blockade of GLP-1 receptor (GLP-1R) signaling reduces insulin, increases glucagon, and impairs glucose tolerance in multiple species (4). Similarly, genetic interruption of the murine Glp1r impairs glucose control, establishing basal GLP-1R signaling as essential for regulation of β-cell function and glucose homeostasis. Nevertheless, the low levels of circulating active GLP-1 together with the distal location of L cells have led some to question whether the major role of GLP-1 is to function as a circulating incretin hormone (7).

More recent studies have expanded concepts of L-cell biology to encompass control of GLP-1 secretion by cytokines and bacterial products. Interleukin (IL)-6 enhances secretion of GLP-1, actions mediated by direct interaction of IL-6 with gut L cells (8). GLP-1 secretion is also induced by bacterial endotoxin, enabling lipopolysaccharide (LPS)-mediated glucoregulation through the GLP-1R (9,10). Furthermore, plasma GLP-1 levels are significantly increased in hospitalized critically ill patients (10), supporting a correlation between systemic inflammation and L-cell secretion. Nevertheless, the importance of GLP-1 in the setting of localized or systemic inflammation remains uncertain. We describe a new axis linking GLP-1R signaling to the control of immune responses in intestinal intraepithelial lymphocytes (IELs), a specialized gut immune cell population that modulates innate immunity and controls gut barrier function, epithelial turnover, and the host response to enteric pathogens (11). These findings extend our understanding of enteroendocrine-immune interactions beyond traditional glucoregulatory concepts of GLP-1 action, with implications for use of GLP-1R agonists in the treatment of metabolic disorders characterized by enhanced activity of the innate immune system.

Male or female C57BL/6 wild-type (WT) mice were from The Jackson Laboratory (Bar Harbor, ME) and acclimated to the animal facility for a minimum of 1 week before analysis. Whole-body Glp1r^−/− mice on the C57BL/6 genetic background (12) and Glp1r^+/+ control mice were generated by crossing Glp1r^+/− mice; 9–18-week-old animals from the same litter or family were studied and housed under specific pathogen-free conditions in microisolator cages and maintained on a 12-h light/dark cycle with free access to standard rodent diet and water unless otherwise noted. All experiments were carried out in accordance with protocols and guidelines approved by the Animal Care Committee at the Toronto Centre for Phenogenomics.

Murine intestinal IELs were obtained following established protocols (13). After flushing luminal contents and removing Peyer’s patches, intestines were opened longitudinally and cut in 5–10-mm pieces. Tissue fragments were incubated twice in 10 mmol/L calcium- and magnesium-free Hanks’ balanced salt solution containing 1 mmol/L dithiothreitol and 5% FBS at 37°C followed by vortexing. Supernatants were combined, filtered through a 40-μm strainer, and pelleted (crude IEL preparations). IELs were further purified by centrifugation in a discontinuous 44/67% Percoll gradient. IELs were collected at the interface (Percoll-purified IEL preparations). At this stage, cells were routinely 50–75% lymphocytes, with epithelial elements accounting for the remainder. Mouse splenocytes and thymocytes were isolated by mechanical dissociation of the spleen and thymus, respectively, followed by red blood cell lysis.

Sorted IELs, splenocytes, or thymocytes were allowed to recover for 1 h and then challenged for 15 min at 37°C with exendin-4 (Ex-4) (CHI Scientific, Maynard, MA), prostaglandin E₂ (5 μmol/L; Sigma-Aldrich Canada), or the adenylyl cyclase activator forskolin (Fk) in Hanks’ balanced salt solution buffer containing 0.1% BSA and 300 μmol/L 3-isobutyl-1-methylxanthine. cAMP concentration was measured using a LANCE cAMP Detection Kit (PerkinElmer, Waltham, MA) and an EnVision Multilabel Plate Reader (PerkinElmer).

Freshly isolated splenocytes and sorted untouched IELs were stimulated in vitro with plate-bound anti-CD3 (clone 145-2C11) and soluble anti-CD28 (clone 37.51) monoclonal antibodies (mAbs) in culture medium (RPMI + 10% FBS + 50 μmol/L β-mercaptoethanol) in the presence or absence of Ex-4, exendin (9-39) (Bachem, Torrance, CA), or Fk. After incubation for 5 h at 37°C, cells were harvested for quantitative RNA and cAMP analyses.

Glp1r^+/+ or Glp1r^−/− female mice were maintained on drinking water containing 3% (weight for volume) dextran sodium sulfate (DSS) (molecular weight 40,000–50,000; USB, Cleveland, OH) or regular drinking water ad libitum for 7 days. Water intake and body weights were determined daily. Animals were euthanized on day 8, and for each mouse, a disease activity index score was calculated (14). Colon damage scores were determined as described (15). In acute studies, C57BL/6 male mice were maintained on drinking water containing 3% (weight for volume) DSS ad libitum for 4 days. On the fourth day, mice were switched to regular drinking water and given two subcutaneous injections of 10 nmol/kg Ex-4 or PBS separated by 12 h. Mice were killed the following morning for analysis of the small bowel and colon.

Intestinal tissue segments were fixed in 10% neutral buffered formalin and paraffin embedded. Digital image acquisition and morphometry were performed on 5-μm histological sections stained with hematoxylin-eosin as described (16). To quantify IELs, sections were stained with a rabbit polyclonal anti-CD3 antibody (Dako Canada, Burlington, ON, Canada) and counterstained with hematoxylin. A minimum of 20 villi/section from three to four sections per tissue sample were scored.

Bone marrow chimeras were generated by lethally irradiating Glp1r^+/+ or Glp1r^−/− female mice (1,100 cGy split into two equal doses 4 h apart) followed by reconstitution with 5 × 10⁶ bone marrow cells from donor C57BL/6 male mice as described (17). Mice were euthanized 12–14 weeks after bone marrow transplantation. Chimerism was assessed by quantitative PCR (qPCR) analysis for the sex-determinant Y chromosome (Sry) using spleen DNA from recipient mice. The efficiency of reconstitution of WT Glp1r expression within the intestinal compartment of recipient mice was assessed by qPCR.

Total RNA was extracted using guanidinium isothiocyanate (18). cDNA synthesis and assessment of full-length Glp1r mRNA expression were performed as described (16,19). Real-time qPCR was performed on an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA) with TaqMan Universal PCR Master Mix and TaqMan Gene Expression Assays (Applied Biosystems). Quantification of transcript levels was performed by the 2^-ΔCt method using 18S rRNA, cyclophilin, Tbp, or β-actin for normalization.

Fecal DNA extraction and 16S rRNA gene amplification were carried out as previously described (20). Microbial population similarities within and between groups of Glp1r^+/+ and Glp1r^−/− mice were determined using metagenomic analysis as described (21). A cluster analysis tree was generated using UPGMA (unweighted pair group method with arithmetic mean) cluster analysis of Bray-Curtis dissimilarity coefficients.

Except where indicated, data are presented as mean ± SD or SE. Statistical significance was determined by unpaired two-tailed Student t test or one- or two-way ANOVA with Bonferroni post hoc analysis using GraphPad Prism version 5.02 software (San Diego, CA). P < 0.05 was considered statistically significant.

Recent studies highlighted limitations of commercially available antisera for detection of immunoreactive GLP-1R expression (16,22). To localize cellular sites of intestinal Glp1r expression beyond the enteric nervous system (23), we fractionated the murine small intestine by sequential collagenase digestion followed by analysis of Glp1r expression and simultaneous assessment of cell lineage–specific RNA transcripts. Quantitatively, Glp1r expression segregated with markers of the villus epithelial compartment (Supplementary Fig. 1A, fraction 1) rather than the crypt compartment, mesenchyme, or smooth muscle layer (Supplementary Fig. 1A, fraction 2). Unexpectedly, a marked increase in relative Glp1r expression was observed in RNA obtained from intestinal cell preparations enriched in IELs (Fig. 1A and B). Consistent with IEL Glp1r expression, we detected abundant expression of IEL markers such as integrin α E (Itgae) and the T-cell surface glycoprotein CD3 γ (CD3g) in RNA from purified small bowel IELs (Fig. 1A). Moreover, the level of Glp1r expression in the intestinal IEL compartment was considerably higher than in other mouse primary and secondary lymphoid organs (Fig. 1B). A 1.4-kb product corresponding to the entire coding region of the canonical mouse Glp1r was amplified from RNA obtained from Percoll-purified IELs at levels approximating those detected in RNA from mouse lung (24) (Fig. 1C). Cloning and sequencing of the 1.4-kb IEL Glp1r mRNA transcript demonstrated that it corresponded to the sequence of the canonical murine Glp1r (data not shown).

Because Tαβ and Tγδ IELs represent the two major subsets of intestinal IELs, we assessed Glp1r expression in RNA from Percoll-purified and FACS-sorted IEL subsets isolated from the small intestine, revealing abundant Glp1r expression in both IEL subsets (Fig. 1D). Furthermore, we detected a strong correlation between levels of Glp1r mRNA and the relative abundance of transcripts corresponding to the IEL markers Itgae and CD3g along the entire murine small intestine and colon (Supplementary Fig. 1B). Taken together, these findings identify the IEL as a previously unrecognized site of Glp1r expression in the small and large bowel.

To determine whether the presence or absence of GLP-1R signaling alters IEL homeostasis, we compared the frequency of IELs in the small intestine and colon of Glp1r^−/− versus Glp1r^+/+ mice and in C57BL/6 mice treated with liraglutide or saline for 1 week (Fig. 2). No significant differences in the phenotypic composition of the IEL population were detected in the small intestine or colon of Glp1r^−/− versus Glp1r^+/+ littermate control mice (Fig. 2A). Furthermore, loss of GLP-1R signaling did not affect IEL density in the jejunum (Fig. 2B). Likewise, liraglutide had no impact on IEL density in the small intestine (Fig. 2C). Thus, GLP-1R signaling is not critical for IEL development or recruitment to the intestinal niche.

To assess whether the IEL Glp1r mRNA transcript encodes a functional GLP-1R protein, we measured levels of cAMP in sorted nonactivated IELs in response to Ex-4 (Fig. 3A and B). Ex-4 directly increased cAMP levels in a dose-dependent manner (half-maximal effective concentration 0.35 nmol/L). Furthermore, the relative extent of Ex-4–induced cAMP production was preserved in activated IELs and was comparable in magnitude to the cAMP response triggered by 10 μmol/L Fk (Fig. 3B). Surprisingly, IELs exhibited a modest cAMP response to the adenylyl cyclase activator Fk compared with T cells from the thymus (Supplementary Fig. 1C). Because IELs produce a broad spectrum of cytokines that play key roles in intestinal inflammation, pathogen clearance, and epithelial barrier function (25), we examined whether GLP-1R signaling modified cytokine expression in sorted intestinal IELs in vitro. As a control, we examined splenocytes that express low levels of Glp1r mRNA. Following activation by immobilized anti-CD3 and soluble anti-CD28 antibodies, transcript levels of the proinflammatory cytokines IL-2, IL-17a, interferon γ, and tumor necrosis factor-α robustly increased in isolated IELs and splenocytes (Fig. 3C). Ex-4 significantly attenuated the induction of both mRNA expression and protein production of proinflammatory cytokines in IELs but not in splenocytes (Fig. 3C and Supplementary Fig. 1E). Notably, the magnitude of reduction in cytokine expression in IELs after Ex-4 treatment paralleled reductions observed with Fk (Fig. 3C). The actions of Ex-4 to suppress cytokine induction required the canonical GLP-1R because they were not detected in IELs from Glp1r^−/− mice (Fig. 3D). Moreover, the GLP-1R antagonist exendin (9-39) attenuated the Ex-4–dependent suppression of cytokine mRNA expression (Supplementary Fig. 1D). In contrast, Fk robustly inhibited cytokine upregulation in IELs from Glp1r^−/− mice (Fig. 3D). These findings link direct activation of GLP-1R signaling to attenuation of proinflammatory cytokine expression in IELs.

IELs play an important role in maintaining the integrity of the epithelial barrier (15). Baseline small intestinal permeability assessed following ¹⁴C-mannitol oral gavage was not different in Glp1r^+/+ versus Glp1r^−/− mice (Supplementary Fig. 3A). To evaluate the importance of Glp1r in the setting of mucosal injury, we compared the severity of DSS-induced colitis in Glp1r^−/− versus Glp1r^+/+ mice. Glp1r^−/− mice lost significantly more weight despite similar intake of DSS (Fig. 4A and B); colon lengths were shorter at baseline yet reduced to a comparable extent after DSS (Fig. 4C). Furthermore, Glp1r^−/− mice exhibited significantly increased disease activity scores (Fig. 4D) and greater epithelial damage (Fig. 4E) reflected by a greater colon damage score (Fig. 4F).

To understand how GLP-1R signaling controls mucosal integrity, we examined expression of genes important for epithelial repair, barrier function, and immune regulation in the colon of Glp1r^−/− and Glp1r^+/+ mice in the presence or absence of DSS-induced colitis (Fig. 5 and Supplementary Fig. 2). Remarkably, even in the absence of experimental gut injury, Glp1r^−/− mice exhibited significant reductions in expression of genes that contribute to epithelial protection and repair (trefoil factor [Tff]-1 and -2, transforming growth factor [Tgf]-b1 and -3, epidermal growth factor receptor [Egfr], keratinocyte growth factor [Fgf7], hepatocyte growth factor [Hgf]), the innate immune response (Il6, Il1b), and inflammation (Il12b). Furthermore, induction of gut injury with DSS produced differential colonic gene expression (Tff3, Tgfb2, Ifng, Hgf) in Glp1r^+/+ versus Glp1r^−/− mice. These findings illustrate that GLP-1R signaling regulates genes important for intestinal homeostasis in the presence and absence of gut inflammation.

Because dysregulated intestinal gene expression in Glp1r^−/− mice potentially reflects loss of GLP-1R signaling in intestinal and extraintestinal cell types, we reassessed colonic gene expression after transplantation of bone marrow from C57BL/6 male donor mice into Glp1r^−/− or Glp1r^+/+ recipient females. In such a setting, Glp1r^−/− mice are reconstituted with Glp1r^+/+ hematopoietic stem cells, including IELs. Chimerism, assessed in spleen DNA from the recipient mice by qPCR, was 84.3 ± 6.4%. Analysis of Glp1r expression in jejunum and sorted IELs from Glp1r^−/− and Glp1r^+/+ recipient females demonstrated that IELs derived from transplanted marrow repopulated the intestinal compartment (Supplementary Fig. 3B). The reconstitution efficiency of Glp1r expression in the jejunum and IELs of Glp1r^−/− recipients 12 weeks after bone marrow transplantation was 64% and 73%, respectively. We focused our analysis on genes shown in Fig. 5 to be dysregulated in the colon of Glp1r^−/− mice. Selective re-establishment of WT IELs in the intestinal mucosa of Glp1r^−/− mice normalized gene expression profiles for the majority of genes examined (Fig. 6).

IELs control intestinal barrier function and repair and protect the intestinal mucosal epithelium by promoting pathogen clearance and lysing pathogen-infected cells (26). Accordingly, we examined expression of immunomodulatory genes in the jejunum of C57BL/6 mice in response to Ex-4 administration (Fig. 7A). Ex-4 markedly upregulated levels of Il1b, Il6, Il22, Il12b, Tnfa, Ccl2, Cxcl1, and Cxcl2 mRNA transcripts. The augmentation of gene expression by Ex-4 was rapid, detectable by 4 h, and transient, with levels of most mRNA transcripts returning to normal by 24 h. Increases in IL-6 and tumor necrosis factor-α protein levels were also observed 4 h after Ex-4 administration (Supplementary Fig. 3C). Ex-4 also significantly increased expression of genes that 1) encode antimicrobial proteins (regenerating islet-derived protein 3 [RegIII]γ and RegIIIβ) or 2) play a role in pathogen clearance (IL-5, IL-13). For some genes, the Ex-4–dependent induction of this group of genes, particularly RegIIIg and RegIIIb, was sustained with continued Ex-4 treatment (Fig. 7A). These data suggest that activation of GLP-1R signaling establishes a generalized cytoprotective response in the murine intestine.

Assessing the effects of Ex-4 or liraglutide in mice with DSS-induced colitis was complicated by marked GLP-1R agonist–mediated reductions in water and, hence, DSS intake, which attenuated intestinal injury. Consequently, we first induced intestinal injury with DSS for 4 days and then treated mice with two injections of Ex-4 12 h apart. Acute Ex-4 had no effect on body weight, colon length, or colon damage score (Fig. 7B and Supplementary Fig. 3D) but significantly reduced colon weight (Fig. 7B) consistent with a reduction of DSS-induced colitis-associated edema. Ex-4 significantly upregulated expression of inflammatory, antimicrobial, and growth factor genes in the ileum (Fig. 7C). Because IELs play important roles in host mucosal defense and response to microbial pathogens, we examined the population of major microbial species by analysis of 16S bacterial rRNA genes in the feces of Glp1r^−/− versus Glp1r^+/+ mice (Fig. 8A). Remarkably, highly significant differences in the relative abundance of Gammaproteobacteria, Bacteroidetes, Firmicutes, and Actinobacteria were detected. Bacteroidetes and Firmicutes represented the majority of the bacterial population in both Glp1r^−/− and Glp1r^+/+ mice (Supplementary Fig. 4A). Less similarity of microbial profiles was seen between groups than within groups. Taken together, these findings suggest that GLP-1R signaling influences the establishment of the gut microbiome and the normal host immune response to intestinal injury.

GLP-1R is expressed in the small intestine and colon (23,27); however, previous efforts to identify cellular sites of GLP-1R expression within the gut yielded conflicting results. Using a GLP-1R antibody and immunohistochemistry, Kedees et al. (28) reported GLP-1R expression in the intestinal mucosa, enteric nervous system, and Paneth cells, whereas studies using the GLP-1R promoter to direct expression of a fluorescent reporter protein localized intestinal GLP-1R expression to enteric neurons (23). GLP-1Rs have also been localized to gastric parietal cells in rats (29) and in the Brunner’s gland of the duodenum, parietal cells and smooth muscle cells in the gastric muscularis externa, and myenteric plexus neurons throughout the primate small and large bowel (30). Given the lack of specificity of the majority of commercially available GLP-1R antibodies and existing discrepancies in studies reporting sites of intestinal GLP-1R expression, we initiated studies to localize the sites of endogenous GLP-1R mRNA expression in the murine gut.

Although mRNA transcripts encoding a functional GLP-1R were detected in rodent immune cells isolated from spleen, lymph nodes, and bone marrow (19), the current experiments reveal that IEL Glp1r expression is considerably more abundant relative to Glp1r mRNA levels in other mouse lymphoid organs. Several lines of evidence support the functional importance of the newly described IEL GLP-1R. First, Ex-4 directly increased cAMP accumulation in isolated IELs. Second, Ex-4 robustly suppressed expression of cytokines following activation of IELs from WT mice but had no effect on IELs from Glp1r^−/− mice. Furthermore, reconstitution of Glp1r^−/− mice with bone marrow from C57BL/6 donors normalized the expression of immune-related genes in the intestine. Taken together, these findings establish the functional importance of the intestinal IEL GLP-1R.

The delineation of an enteroendocrine-immune axis is not without precedent. Infection of severe combined immunodeficient mice with Trichuris muris reduced the number of enteroendocrine cells compared with WT controls, whereas the number of gut endocrine cells increased significantly following reconstitution of immunodeficient mice with CD4⁺ T cells from infected WT mice (31). Furthermore, bacterial-derived products such as LPS and flagellin independently modulated expression of genes encoding peptide hormones and multiple ligands and receptors important for the immune response in LCC-18 enteroendocrine cells (32). Conversely, receptors for vasoactive intestinal peptide and pituitary adenylate cyclase–activating peptide have been localized to intestinal IELs (33), whereas Toll-like receptors are expressed on enteroendocrine cells, and Toll-like receptor activation by bacterial products promotes secretion of cholecystokinin from gut endocrine cells (34).

Considerable evidence attests to the importance of inflammation and innate immunity in the control of metabolism (35,36); however, few studies have examined whether IELs play a role in the development and/or treatment of experimental models of type 2 diabetes or obesity. Emerging data link cytokine and chemokine activity with control of enteroendocrine cell responses. Direct administration of RANTES (regulated on activation, normal T cell expressed and secreted) to NCIH716 cells rapidly reduces GLP-1 secretion and cAMP accumulation, and RANTES reduces plasma levels of GLP-1 and blunts the rise in insulin levels after glucose administration in vivo (37). In contrast, IL-6, IL-1, and LPS all increase GLP-1 levels in mice (8–10), with IL-6 acting directly on gut endocrine cells to promote GLP-1 secretion. Of note, central blockade of GLP-1R signaling enhanced the pyrogenic effect of LPS (38), implying that endogenous GLP-1 exerts anti-inflammatory actions. Indeed, pharmacological administration of GLP-1R agonists has exerted anti-inflammatory actions in multiple studies through incompletely understood mechanisms (39). The current data extend previous observations by demonstrating that GLP-1 rapidly suppresses a proinflammatory cytokine program from activated intestinal IELs ex vivo, actions requiring a functional GLP-1R. Hence, induction of GLP-1 secretion following administration of cytokines or endotoxin has relevance beyond glucose control and leads to attenuation of the inflammatory response (Fig. 8B).

Previous studies examining the relationship between proglucagon-derived peptides and control of gut inflammation have focused on the related L-cell peptide GLP-2, which potently suppresses inflammation in experimental models of gut injury (40–42). Conversely, disruption of GLP-2R signaling was associated with increased bacterial colonization of the small bowel, reduced bactericidal activity, impaired Paneth cell responses (43), and enhanced susceptibility to gut injury and inflammation (43). The anti-inflammatory actions of GLP-2 are indirect, mediated through reduction of gut permeability (44,45). Indeed, modulation (both gain and loss) of GLP-2R signaling robustly modifies gut barrier function and the extent of inflammation in murine models of diabetes characterized by gut injury and endotoxemia (46). The current findings demonstrate that loss of GLP-1R signaling also enhances the susceptibility to gut injury, albeit through different mechanisms, involving the IEL-GLP-1R signaling system.

Germ-free mice exhibit increased levels of circulating GLP-1 and enhanced proglucagon expression in the murine brainstem (47) and distal gut (48), whereas microbial metabolites such as indole directly activate GLP-1 secretion from L cells (49). Collectively, these emerging findings link the gut microbiome and its metabolic by-products with the control of central and peripheral GLP-1 production. Reduction of energy intake and malnutrition has been associated with defective immune responses, and accumulating evidence supports an intimate link among energy availability, nutrient sensing, and the regulation of the innate immune response (50). Of note, Glp1r^−/− mice exhibit resistance to diet-induced obesity (51), increased relative abundance of Bacteroidetes, and reduced abundance of Firmicutes in their microbiota, whereas obese mice show a decrease in Bacteroidetes and an increase in Firmicutes (52). Collectively, the data establish a previously unrecognized local GLP-1-IEL-GLP-1R axis that may influence microbiome composition and the response to intestinal inflammation (Fig. 8B). Because considerable evidence supports a role for gut-derived inflammation in the pathophysiology of insulin resistance arising in subjects with obesity and diabetes (35), we hypothesize that GLP-1R agonists exert favorable nonglycemic actions in part through modulation of IEL-GLP-1R signaling. This hypothesis requires further validation through experimental models of metabolic dysfunction and inflammation.

Funding. P.M.S. holds operating grants for the Canadian Institutes of Health Research (MOP-89894 and IOP-92890) and is the recipient of a Canada Research Chair in Gastrointestinal Disease. D.J.D. was supported in part by a Canada Research Chair in Regulatory Peptides, a Banting & Best Diabetes Centre-Novo Nordisk Chair in Incretin Biology, and operating grant 123391 from the Canadian Institutes of Health Research.

Duality of Interest. D.J.D. received partial grant support from Novo Nordisk. D.J.D. has served as an advisor or consultant or speaker within the past 12 months for Arisaph Pharmaceuticals, Inc., Intarcia Therapeutics, Merck Research Laboratories, MedImmune, Novo Nordisk, NPS Pharmaceuticals, Inc., Receptos, Inc., and Sanofi. Mount Sinai Hospital receives partial operating grant support for other studies in the Drucker laboratory from GlaxoSmithKline, Merck, Novo Nordisk, and Sanofi. Neither D.J.D. nor his family members hold stock directly or indirectly in any of these companies. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. B.Y., L.L.B., J.K., D.H., X.C., L.J.P., K.C.J.-H., W.Y., and K.W.A.B. contributed to the planning and carrying out of experiments, analysis of results, and writing and review of the manuscript. M.G.S., P.M.S., and D.J.D. contributed to the planning of experiments, analysis of results, and writing and review of the manuscript. D.J.D. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.