Glucagon-Like Peptide-1 Receptor Activation Modulates Pancreatitis-Associated Gene Expression But Does Not Modify the Susceptibility to Experimental Pancreatitis in Mice

Experiments were carried out according to protocols approved by the animal care committees of the University Health Network and Mount Sinai Hospital. Mice were housed under a 12-h light/dark cycle with access to standard or high-fat rodent diet (45% kcal from fat; Research Diets, New Brunswick, NJ). Male C57Bl/6J mice (8–10 weeks old) were obtained from Taconic Laboratories (Hudson, NY) and were allowed to acclimatize to the animal facility for 7 days before each experiment. Glp1r^−/− mice in the C57Bl/6J background were generated as described (16).

Ex-4 was from Chi Scientific (Maynard, MA) and dissolved in PBS. Forskolin and caerulein were from Sigma Chemical (St. Louis, MO). Caerulein was dissolved in saline (0.9% NaCl) and forskolin in DMSO. Liraglutide was from Novo Nordisk (Bagsværd, Denmark). Sitagliptin was from Merck (Rahway, NJ).

Pancreatitis was induced in 9- to 11-week-old male C57Bl/6J mice or 3-month-old male Glp1r^−/− and littermate control Glp1r^+/+ mice using the CCK receptor agonist caerulein (17). Caerulein dose-response experiments were performed in male 9- to 11-week-old C57Bl/6J mice to identify a submaximal dose of caerulein that would produce a low level of pancreatic inflammation. Secretagogue-induced pancreatitis was elicited by administration of five sequential hourly intraperitoneal injections of caerulein, at doses of 6, 12, 24, or 50 μg/kg body wt as described (11,17,18); control animals received an equal volume of saline. Mice were killed by CO₂ inhalation 1 h after the final caerulein injection.

To assess whether prior GLP-1R activation exacerbates the subsequent development of pancreatitis, male C57Bl/6J mice were injected (intraperitoneally) twice daily with either 10 nmol/kg Ex-4 or PBS for 7 days. Sixteen hours after the last Ex-4 injection, mice were administered 5 hourly intraperitoneal injections of caerulein at 3 or 6 μg/kg body wt, as indicated, and killed by CO₂ inhalation 1 h after the final caerulein injection.

To assess whether a low level of pancreatitis was worsened by subsequent GLP-1R activation, male C57Bl/6J mice were administered 5 hourly intraperitoneal injections of caerulein (6 μg/kg) followed by twice daily intraperitoneal injections with either 10 nmol/kg Ex-4 or PBS for up to 6 days. Mice were killed 1, 24, 72, or 144 h (day 0, 1, 3, and 6, respectively) after the final caerulein injection such that mice killed on day 1 received two injections of PBS or Ex-4, one in the evening (∼3 h after the last caerulein injection) and one in the morning (∼6 h before being killed). Similarly, the final injection with Ex-4 or PBS was 6 h before being killed for all remaining mice (day 3 and 6). To determine whether loss of GLP-1R signaling protected mice from development of experimental pancreatic inflammation, pancreatitis was induced in 3-month-old male Glp1r^+/+ and Glp1r^−/− mice by administration of 5 hourly intraperitoneal injections of caerulein (6 μg/kg). Mice were killed by CO₂ inhalation 1 h after the final caerulein injection.

To assess whether GLP-1 receptor activation regulates gene expression in the exocrine pancreas of normal mice, male C57Bl/6J mice were injected subcutaneously with a single dose of Ex-4 (1 μg), Ex-4 + 15% glucose (G + X), or PBS, and mice were killed 30, 45, 60, or 90 min later by CO₂ inhalation. To determine whether high-fat feeding and induction of insulin resistance (19) modifies the effects of Ex-4 on the exocrine pancreas, 4-week-old wild-type male mice were fed either standard rodent diet or high-fat diet (45% kcal from fat) for 8 weeks; during the last 4 weeks mice on high-fat diet were given twice daily intraperitoneal injections of PBS or Ex-4 (24 nmol/kg). Separately, to assess the impact of diabetes on the GLP-1R–dependent regulation of exocrine pancreatic gene expression, 13-week-old male C57BL/6 mice were maintained on a high-fat diet for 8 weeks, given one dose of streptozotocin (STZ, 100 mg/kg), and after 4 more weeks were allocated to four treatment groups for an additional 8 weeks of 1) high-fat diet and twice daily intraperitoneal injections of PBS, 2) high-fat diet and metformin (500 mg · kg⁻¹ · day⁻¹) provided in the drinking water, 3) high-fat diet and twice daily intraperitoneal injections of Ex-4 (3 nmol/kg), or 4) high-fat diet and sitagliptin (∼370 mg · kg⁻¹ · day⁻¹) provided in the diet.

Animals were killed by CO₂ inhalation 1, 24, 72, or 144 h after the final caerulein injection. Blood samples were collected by cardiac puncture, placed on ice for 30 min, and centrifuged at 4°C for 10 min at 14,000 rpm, and serum was stored at −80°C. The pancreas was removed, weighed, placed on ice, and cut into sections for RNA, protein, histology, and edema analyses. For RNA isolation, a section of pancreas was homogenized in TRIzol reagent (Invitrogen Life Technologies, San Diego, CA), frozen on dry ice, and stored at −80°C until further analysis. Tissue samples for histology were fixed in 10% neutral-buffered formalin for 24 h and embedded in paraffin. Remaining tissue samples were snap frozen in liquid nitrogen and stored at −80°C until further analysis.

Quantification of serum amylase was performed using the Phadebas amylase test (Magle Life Sciences, Cambridge, MA) using 10 μl serum (20). Pancreatic edema (21) was quantified by measuring tissue water content after desiccation to a constant weight at room temperature for 72 h (dry weight). The results were calculated and expressed as a percentage of wet weight (wet weight−dry weight/wet weight × 100).

RNA from pancreatic tissue was extracted using TRIzol reagent; after first-strand cDNA synthesis, real-time quantitative PCR was carried out as described (22). Quantification of transcript levels was performed with the ABI PRISM SDS 2.1 software. Cyclophilin mRNA or 18S rRNA was used for normalization as expression of both remained unaltered regardless of treatment.

Paraffin-embedded tissues were sectioned (5 μm) and stained with hematoxylin-eosin. Immunohistochemistry was carried out using indirect immunoperoxidase detection with NovaRED substrate (Vector Laboratories, Burlington, ON, Canada) followed by hematoxylin counterstaining. Primary rabbit polyclonal antibodies were used at dilutions recommended by the manufacturer and included: rabbit anti–c-fos (Sigma-Aldrich), rabbit anti–egr-1 (Santa Cruz Biotechnology, Santa Cruz, CA), and rat anti-mouse neutrophils (MCA771GA; AbD Serotec, Kidlington, Oxford, U.K.).

Pancreatic fragments were prepared from C57Bl/J6 mice. Briefly, the pancreas was rapidly cut into small pieces, cultured in DMEM + 1% BSA + 10% trasylol, washed three times with DMEM + 1% BSA + 5% trasylol, resuspended in DMEM + 1% BSA + 10% trasylol, and aliquoted such that each tube contained ∼4% of the entire pancreas. Preparations were then stimulated with 100 nmol/l Ex-4 or 20 μmol/l forskolin for 15 or 30 min at 37°C and frozen on dry ice. Samples were thawed, sonicated with ice-cold ethanol (65% final concentration), and cellular debris removed by centrifugation at 13,000 rpm. cAMP was measured from dried aliquots of ethanol extracts using a cAMP radioimmunoassay kit (Biomedical Technologies, Stoughton, MA).

Statistical significance was assessed by one-way or two-way ANOVA using Bonferroni's multiple comparison post hoc test and, where appropriate, by unpaired Student's t test using GraphPad Prism 4 (GraphPad Software, San Diego, CA). A P value of <0.05 was considered to be statistically significant.

Mice treated with GLP-1R agonists exhibit a significant increase in pancreas weight (23) that cannot be attributed solely to increased β-cell mass. To determine whether GLP-1R activation induces a program of gene expression in the exocrine pancreas associated with development of pancreatic growth or inflammation, we examined whether Ex-4 enhanced the expression of immediate early genes known to play important roles in regulating cell proliferation. Ex-4 rapidly and robustly increased pancreatic levels of egr-1 and c-fos (Fig. 1,A), c-myc, and junB mRNA transcripts (supplementary Fig. 1, available in an online appendix at http://diabetes.diabetesjournals.org/cgi/content/full/db09-0626/DC1). Moreover, the egr-1 protein was not upregulated in islets but was localized exclusively to nuclei of cells within the exocrine pancreas (Fig. 1,B), whereas nuclear c-fos expression was observed in both islets and exocrine tissue (Fig. 1 C). Hence, GLP-1R activation induces a robust induction of gene and protein expression in the exocrine pancreas.

Egr-1 is an early response gene that encodes a transcription factor that regulates cell proliferation, growth, and apoptosis (24). Induction of egr-1 expression occurs early in the course of caerulein-induced pancreatitis, and inflammation-related gene expression is attenuated in egr-1^−/− mice with experimental pancreatitis (25). Accordingly, we examined whether agents that activate the GLP-1Rs upregulate expression of genes associated with pancreatitis. Mice were treated with Ex-4 for 1 week (supplementary Fig. 2), and the expression of pancreatitis-associated genes was assessed by real-time quantitative PCR. Ex-4 (1 or 10 nmol/kg twice daily for 1 week) increased pancreas weight (Fig. 2,A) but had no effect on levels of mRNA transcripts encoding the proinflammatory mediators (monocyte chemotactic protein [Mcp]-1, intracellular adhesion molecule-1, signal transducer and activator of transcription [STAT]-3, ndrg1, or ifitm3) or transcription factors (egr-1, activating transcription factor (ATF)-3, mist1, c-fos, c-myc, and junB) associated with acute pancreatitis (Fig. 2,B and supplementary Table 1). In contrast, Ex-4 markedly increased pancreatic expression of the anti-inflammatory gene PAP (Reg IIIβ) and Reg IIIα (Fig. 2,B). The increase in pancreatic weight and induction of PAP and RegIIIα gene expression was not specific to Ex-4 but was also observed in mice treated with the human GLP-1R agonist liraglutide (Fig. 2,C and D). These actions required a functional GLP-1R as liraglutide had no effect on pancreatic weight or induction of gene expression in Glp1r^−/− mice (Fig. 2,C and D). Moreover, liraglutide reduced the expression of the proinflammatory transcription factor, STAT3, in wild-type mice but not in Glp1r^−/− mice (Fig. 2,D). The expression levels of most pancreatitis-associated mRNA transcripts were comparable in Glp1r^+/+ versus Glp1r^−/− pancreas; however, basal levels of suppressor of cytokine signaling (SOCS)-3 and STAT3 mRNA transcripts were lower in Glp1r^−/− mice (Fig. 2 D). Taken together, these findings demonstrate that structurally distinct GLP-1R agonists produce changes in pancreatic mass and gene expression, actions requiring a functional GLP-1 receptor.

To assess whether antecedent activation of the GLP-1R increases the severity of pancreatitis, mice were treated twice daily with Ex-4 for 1 week before exposure to caerulein (Fig. 3). A low dose of caerulein was employed for these studies (3 or 6 μg/kg) based on preliminary experiments designed to induce a detectable yet submaximal inflammatory response in the pancreas (supplementary Fig. 3). Ex-4 significantly increased pancreas weight (Fig. 3,B) whereas caerulein alone produced a small increase in pancreas weight (Fig. 3,B) and a modest increase in tissue edema (Fig. 3,C). Caerulein consistently produced a low but detectable level of inflammation, as evidenced by increased pancreatic edema, elevated serum amylase, neutrophil infiltration, and upregulated egr-1 and c-fos expression in the exocrine pancreas (supplementary Figs. 3 and 4, and data not shown). Intriguingly, serum amylase levels were lower in mice pretreated with Ex-4 and then treated with 3 μg/kg caerulein (Fig. 3,D). Ex-4 did not significantly modulate levels of most pancreatitis-associated mRNA transcripts regulated by caerulein administration with the exception of Reg IIIα and SOCS3 that were further induced in Ex-4–treated mice exposed to caerulein (Fig. 3,E). Consistent with previous findings (Fig. 2,A), treatment with Ex-4 for 1 week significantly increased the transcript levels of the anti-inflammatory protein PAP (Fig. 3 E), in the presence or absence of caerulein.

To determine whether endogenous GLP-1R signaling influences susceptibility to caerulein-induced pancreatitis, we assessed the effects of a submaximal dose of caerulein on gene expression and severity of pancreatitis in Glp1r^+/+ versus Glp1^−/− mice. Pancreas weight, edema, and serum amylase were not significantly different in caerulein-treated Glp1r^+/+ versus Glp1r^−/− mice (Fig. 4,B–D). Furthermore, GLP-1R genotype had no effect on the expression of pancreatic genes known to be induced by caerulein, including SOCS3, Mcp-1, egr-1, ATF-3, and c-fos (Fig. 4,E). Intriguingly, caerulein-induced levels of PAP, RegIIIα, c-myc, and junB mRNA transcripts were significantly greater in Glp1r ^−/− mice (Fig. 4 E).

To determine whether GLP-1R agonists exacerbate the severity or prolong the recovery after initiation of acute pancreatitis, mice were treated with caerulein followed by 10 nmol/kg of Ex-4 twice daily for up to 6 days (Fig. 5,A). Mice were killed for analysis at 1, 24, 72 h, and 6 days after the final injection of caerulein. Caerulein transiently but significantly increased serum amylase (Fig. 5,B, day 0), pancreatic edema (Fig. 5,C, day 0), and relative pancreas weight (Fig. 5,D, day 1). However, serum amylase and pancreatic edema returned to basal levels after 24 h, independent of Ex-4 treatment (Fig. 5,B and C). Similarly, levels of mRNA transcripts encoding the proinflammatory proteins Mcp-1, interleukin (IL)-6, and ndrg1 were rapidly induced by caerulein but returned to basal levels within 24 h independent of Ex-4 treatment (Fig. 5,E). Levels of STAT3 and ifitm3 transcripts were elevated in Ex-4–treated mice (Fig. 5,E), and mRNA transcripts for the anti-inflammatory proteins PAP and SOCS3 remained elevated in Ex-4–treated mice. In contrast, mRNA transcripts for c-fos, c-myc, and junB were transiently increased in the pancreas of caerulein-treated mice but were not affected by concomitant Ex-4 treatment (Fig. 5 E), whereas levels of pancreatic mRNA transcripts for mist1 and egr-1 were modestly but significantly higher in Ex-4–treated mice.

To determine whether GLP-1R activation modified the expression of pancreatitis-associated genes in the setting of a mild metabolic stress associated with the development of insulin resistance (19), mice were placed on a high-fat diet and treated twice daily with saline or Ex-4 (Fig. 6,A). Consistent with previous observations, Ex-4 prevented weight gain (Fig. 6,C) and significantly increased pancreas weight (Fig. 6,B) but did not affect levels of serum amylase (Fig. 6,D). Moreover, high-fat diet mice treated with saline, but not Ex-4, exhibited a significant increase in the transcript levels of pancreatitis-associated genes including Mcp-1, STAT3, egr-1, and ATF-3 as well as the exocrine-specific transcription factor mist1 (Fig. 6,E). In contrast, Ex-4 (but not saline) administration significantly increased levels of mRNA transcripts for PAP, c-fos, c-myc, and junB (Fig. 6 E).

We next determined whether metformin, sitagliptin, or Ex-4, antidiabetic agents associated with enhanced GLP-1 receptor activation through various mechanisms (3,4,26), modulated pancreatitis-associated gene expression profiles in high-fat–fed diabetic mice. Ex-4, and to a lesser extent the DPP-4 inhibitor sitagliptin, but not metformin increased PAP mRNA transcript levels in diabetic mice (Fig. 7). Neither Ex-4, metformin, nor sitagliptin treatment modulated the expression of other genes associated with pancreatitis including SOCS3, egr-1, STAT3, c-myc, or junB (Fig. 7,B). Finally, as Ex-4 was reported to exhibit high-affinity binding to guinea pig pancreatic acinar membranes and stimulate cyclic AMP formation through a distinct receptor that preferentially recognized Ex-4 relative to GLP-1 (27), we assessed cyclic AMP formation in slices from wild-type mouse pancreas. Forskolin but not Ex-4 rapidly stimulated cyclic AMP accumulation in pancreatic slices ex vivo (Fig. 8).

There are currently limited data regarding putative effects of GLPs on the function of the normal or inflamed exocrine pancreas. Intriguingly, coadministration of glucagon and careulein attenuated the increase in pancreatic weight and amylase expression seen in rats treated with careulein alone (28). Similarly, oxyntomodulin, a peptide structurally related to both glucagon and GLP-1, was shown to be 10-fold more potent than glucagon in the suppression of rat pancreatic exocrine secretion in the basal state or after administration of caerulein (29). Our results extend these findings by demonstrating that although acute and chronic GLP-1R activation modulates a gene expression program in the exocrine murine pancreas, GLP-1 receptor activation does not predispose to or exacerbate experimental pancreatitis in mice.

Several lines of evidence imply that exocrine cells express a functional GLP-1 and/or exendin receptor. Guinea pig pancreatic acini contain high-affinity binding sites for Ex-4, and Ex-4 binding was displaced from acinar cells by coincubation with native GLP-1 (27). Furthermore, both Ex-4– and GLP-1–stimulated cAMP formation but not amylase release in dispersed guinea pig pancreatic acini (27,30), and these stimulatory actions on cAMP were blocked by the GLP-1R antagonist exendin (9–39) (30,31). Similarly, Ex-4 stimulated cyclic AMP formation in rat pancreatic slices, and potentiated calcium ionophore-, neurotransmitter-, or CCK-induced amylase release in vitro (32). In contrast, we found that forskolin, but not Ex-4, rapidly increased cAMP formation in murine pancreatic fragments. Whether these findings reflect species-specific differences in the expression or signaling of functional pancreatic Ex-4/GLP-1 receptors requires further investigation.

We used the CCK agonist caerulein for studies of experimental pancreatitis as previous data demonstrated a functional interaction of CCK and GLP-1 signaling pathways in the endocrine and exocrine pancreas (32,33). Although the expression of most pancreatitis-associated genes was not further modified by administration of GLP-1R agonists, we consistently detected upregulation of PAP after chronic administration of Ex-4 to mice, in the presence or absence of caerulein administration. Similarly, liraglutide also induced PAP expression, in a GLP-1R–dependent manner, and basal levels of pancreatic PAP mRNA transcripts were modestly reduced in Glp1r^−/− mice.

PAP has been shown to be mitogenic, antiapoptotic, and anti-inflammatory and is strongly induced early in the course of inflammatory diseases such as pancreatitis, Crohn's disease, and ulcerative colitis (14). Moreover, inhibition of PAP gene expression in rats augments the severity of acute pancreatitis (34), and PAP knockout mice exhibit more extensive inflammation after induction of caerulein-induced pancreatitis (13), suggesting a protective role for PAP in the inflammatory response to cellular injury. Hence, the induction of PAP after GLP-1R activation may represent a compensatory mechanism that serves to limit damage to the exocrine pancreas.

Our studies of GLP-1R activation on gene expression and the development of and/or recovery from pancreatitis were motivated, in part, by clinical reports of acute pancreatitis in diabetic patients treated with GLP-1R agonists (8,10). Our findings do not support the hypothesis that GLP-1R activation sensitizes the murine pancreas to the development of pancreatic inflammation. Similarly, we did not detect any difference in the extent of pancreatic inflammation in Glp1r^−/− mice. Taken together, the available data demonstrate that GLP-1R activation leads to increases in mass and changes in expression of pancreatitis-associated genes but does not modify pancreatitis susceptibility or severity in the murine pancreas. Whether GLP-1R activation modifies gene expression, enzyme secretion, or pancreatic inflammation in the human pancreas requires further analysis.

These studies were supported in part by a grant from the Canadian Institutes of Health Research and Hoffman LaRoche. D.J.D. is supported by the Canada Research Chairs Program.

The Drucker Lab has received operating grant support from Novo Nordisk in conjunction with the studies of liraglutide. No other potential conflicts of interest relevant to this article were reported.

We thank B. Yusta, A. Maida, and D. Holland for experimental assistance and helpful discussions.