Endocrinology and Metabolism

Exenatide does not evoke pancreatitis and attenuates chemically induced pancreatitis in normal and diabetic rodents

Krystyna Tatarkiewicz, Pamela A. Smith, Emmanuel J. Sablan, Clara J. Polizzi, Donald E. Aumann, Christiane Villescaz, Diane M. Hargrove, Bronislava R. Gedulin, Melissa G. W. Lu, Lisa Adams, Tina Whisenant, Denis Roy, David G. Parkes


The risk of developing pancreatitis is elevated in type 2 diabetes and obesity. Cases of pancreatitis have been reported in type 2 diabetes patients treated with GLP-1 (GLP-1R) receptor agonists. To examine whether the GLP-1R agonist exenatide potentially induces or modulates pancreatitis, the effect of exenatide was evaluated in normal or diabetic rodents. Normal and diabetic rats received a single exenatide dose (0.072, 0.24, and 0.72 nmol/kg) or vehicle. Diabetic ob/ob or HF-STZ mice were infused with exenatide (1.2 and 7.2 nmol·kg−1·day−1) or vehicle for 4 wk. Post-exenatide treatment, pancreatitis was induced with caerulein (CRN) or sodium taurocholate (ST), and changes in plasma amylase and lipase were measured. In ob/ob mice, plasma cytokines (IL-1β, IL-2, IL-6, MCP-1, IFNγ, and TNFα) and pancreatitis-associated genes were assessed. Pancreata were weighed and examined histologically. Exenatide treatment alone did not modify plasma amylase or lipase in any models tested. Exenatide attenuated CRN-induced release of amylase and lipase in normal rats and ob/ob mice but did not modify the response to ST infusion. Plasma cytokines and pancreatic weight were unaffected by exenatide. Exenatide upregulated Reg3b but not Il6, Ccl2, Nfkb1, or Vamp8 expression. Histological analysis revealed that the highest doses of exenatide decreased CRN- or ST-induced acute inflammation, vacuolation, and acinar single cell necrosis in mice and rats, respectively. Ductal cell proliferation rates were low and similar across all groups of ob/ob mice. In conclusion, exenatide did not modify plasma amylase and lipase concentrations in rodents without pancreatitis and improved chemically induced pancreatitis in normal and diabetic rodents.

  • diabetes
  • mouse
  • rat
  • caerulein
  • sodium taurocholate
  • pancreatic duct

glucagon-like peptide-1 (GLP-1) exerts multiple glucoregulatory actions by enhancing glucose-dependent insulin secretion, regulating gastric emptying, decreasing postprandial glucagon secretion, and decreasing food intake. Moreover, GLP-1 has been observed to improve β-cell mass by augmenting β-cell survival and proliferation in rodents (7, 19, 42). In recent years, the therapeutic potential of GLP-1 receptor (GLP-1R) agonists or dipeptidyl peptidase IV (DPP IV) inhibitors for treatment of type 2 diabetes gained widespread attention, and several drugs affecting the GLP-1 pathway were approved to control hyperglycemia, including the GLP-1R agonists exenatide and liraglutide (12, 18) and the DPP IV inhibitors sitagliptin and saxagliptin.

Cases of pancreatitis have been observed in patients treated with GLP-1 receptor agonists and DPP IV inhibitors (1, 9, 10, 13, 38). Acute pancreatitis is a complex clinical condition that ranges in severity from mild to life-threatening. Abdominal pain, ultrasound-confirmed pancreatic pathological changes and increased plasma amylase and lipase concentrations are the most common markers of acute pancreatitis in the clinic (21). Type 2 diabetes and/or obesity are risk factors for the development of pancreatitis, and obesity increases its severity and mortality rate (5, 28, 34). The cellular functions and molecular mechanisms responsible for initiating and modifying the severity of pancreatitis have not been fully elucidated. In general, acinar cells, which secrete digestive enzymes into pancreatic ducts, play an important role in the development of pancreatitis. Pancreatic injury likely occurs by autodigestion of the pancreas via retention of activated digestive enzymes accompanied by a highly amplified inflammatory response, edema, cellular damage, and necrosis.

Multiple drugs have been suggested to stimulate or worsen pancreatitis (2). It remains unclear whether the GLP-1R is expressed and functional in the exocrine pancreas; therefore, a modulatory role for GLP-1R agonists in pancreatitis cannot be excluded on the basis of mechanism of action (29, 41, 43). Two recent epidemiological studies in drug surveillance data bases showed no increased risk of acute pancreatitis in patients with type 2 diabetes treated with exenatide BID compared with those treated with metformin, glyburide (17), sulfonylurea, biguanidine, or thiazolidinedione (22). However, a very limited number of cases of recurrent pancreatitis have been reported in patients treated with exenatide; therefore, statistical evidence for a causal relationship between exenatide administration and pancreatitis remains to be demonstrated.

Recent animal studies with GLP-1R agonists provided contradictory data in relation to pancreatitis. In one study, mild signs of pancreatitis were observed in normal rats treated subchronically with exenatide (33). In another, exenatide was shown to induce expression of mRNA transcripts encoding anti-inflammatory proteins and had no effect on the severity of caerulein (CRN)-induced pancreatitis in mice (23). The aim of the present series of pharmacology studies was to better understand the role of exenatide in potentially inducing or modulating pancreatitis in animal models. Therefore, we evaluated the acute and subchronic effects of exenatide in normal or diabetic rodents. Additionally, we assessed the actions of exenatide under more challenging conditions, i.e., with experimentally induced pancreatitis. To examine a potential impact of exenatide on different severity of the disease, we applied in normal rats two models of chemically induced pancreatitis; a mild form was induced with CRN and a severe form with sodium taurocholate. In all studies, we measured concentrations of amylase and lipase in plasma at various time points along with the terminal pancreatic weight. Plasma concentrations of inflammatory mediators were assessed in diabetic ob/ob mice after a single dose or subchronic administration of exenatide followed by induction of the experimental pancreatitis. We also present in this article the histopatholgical assessment of pancreata from normal rats with severe sodium taurocholate (ST)-induced pancreatitis and from ob/ob mice, which were treated with exenatide for 4 wk with the subsequent CRN-induced pancreatic injury.



All procedures were performed in accordance with the Institutional Animal Care and Use Committee at Amylin Pharmaceuticals in an Association for Assessment and Accreditation of Laboratory Animal Care-approved facility. Male mice and rats used in these studies are described in Table 1. All wild-type control animals were strain-matched by the vendor. Animals were acclimated for ≥6 days before use and were housed individually (21–24°C, 12:12-h light-dark cycle) with ad libitum access to food and water. Body weights were measured at the termination of each study and weekly in chronic mouse studies.

View this table:
Table 1.

Animal models used in the study

Generation of high-fat-fed and streptozotocin-induced diabetic mouse model.

C57BL/6 mice, maintained on a high-fat (HF) diet at Jackson Laboratories starting at 4 wk of age, were kept on the HF diet throughout the study. To induce experimental diabetes, animals were dosed intraperitoneally (ip) with streptozotocin (STZ; 100 mg/kg) once weekly for 2 consecutive weeks. Controls (nondiabetic) received 0.1 M citrate buffer as vehicle at the same intervals as the STZ groups. At 14 wk of age, diabetic animals were randomized into treatment groups based on Hb A1c.

Induction of experimental pancreatitis.

The CCK receptor agonist CRN (C9026; Sigma-Aldrich, St. Louis, MO) was used to induce acute pancreatitis in fasted animals treated with vehicle or exenatide (acutely or subchronically), as described below. CRN was reconstituted in commercially available 0.9% saline and administered by five consecutive hourly ip injections (10 μg/kg in mice) or a by single ip injection (10 μg/kg in rats) (15). Control animals received saline injections of the same volume. Doses of CRN were selected to induce low levels of pancreatic tissue damage and inflammation. A more severe acute insult to the pancreas was made in anesthetized Wistar rats through transduodenal cannulation of the biliopancreatic duct followed by retrograde infusion of 5% ST (CAS 345909-26-4; Acros Organics, Morris Plains, NJ), a bile salt, at a rate of 0.05 ml/min using an infusion pump (model no. 2400-006; Harvard Apparatus, South Natick, MA). Animals received a total volume of 1 ml/kg body wt. Sham surgery consisted of laparotomy, duodenal incision, and closure (39).

Administration of exenatide.

To assess the effects of acute pretreatment with exenatide on CRN-induced pancreatitis, the compound was administered as a single ip dose (1.2 and 7.2 nmol/kg) in conscious mice or a single subcutaneous (sc) dose (0.072, 0.24, and 0.72 nmol/kg) in anesthetized rats 15 min prior to CRN administration. The same doses of exenatide were administered twice a day for 48 h in rats with ST-induced pancreatitis, with the first injection applied 15 min before the ST infusion. In the subchronic studies in mice, exenatide was administered at 1.2 and 7.2 nmol·kg−1·day−1 via continuous sc infusion for 4 wk. The peptide was reconstituted in 50% DMSO in sterile water supplemented with 0.1% bovine serum albumin and loaded into osmotic 2-wk minipumps (model 2002; Alzet, Cupertino, CA) according to the manufacturer's protocol. After 14 days, mice were reimplanted with new minipumps, with the same treatments administered for the next 2 wk. Control groups received minipumps loaded with vehicle, as described above.

Biochemical analysis.

At various time points, blood samples were collected into heparinized capillary tubes via the retroorbital sinus from nonanesthetized mice. Terminal samples from isofluorane-anesthetized mice were collected by cardiac puncture. In the CRN-induced pancreatitis model in anesthetized rats, blood samples from the femoral artery were collected into heparinized Natelson tubes. In the ST-induced pancreatitis model in nonanesthetized rats, samples from the tail vein were collected into heparinized tubes. Terminal samples were collected from anesthetized animals via cardiac puncture. All samples were processed for plasma and stored at −80°C for later analysis. Whole blood Hb A1c, plasma amylase, lipase, and glucose concentrations were measured using either the Olympus AU400e or AU680 clinical analyzer (Olympus America, Irving, TX) in accordance with the manufacturer's protocol. Plasma samples for interleukin (IL)-1β, -2, and 6, monocyte chemoattractant protein-1 (MCP-1), IFNγ, and TNFα were collected from ob/ob mice at 6 h post-first CRN injection. Cytokine concentrations were measured using the Milliplex Mouse Cytokine/Chemokine multiplexing immunoassay kit according to the manufacturer'a protocol (Millipore, Billerica, MA). Median fluorescent intensity was determined using the Luminex instrument (Luminex, Austin, TX). Cytokine concentrations were calculated using the four-parameter logistic standard curve fit method (SoftMax Pro; Molecular Devices, Sunnyvale, CA). Plasma exenatide concentrations were measured using an immunoenzymetric assay developed at Amylin Pharmaceuticals, as described previously (20).

Histology and gene expression.

Mouse and rat pancreata were dissected, weighed, and immediately fixed in 10% neutral-buffered formalin and embedded in paraffin. Five-micron sections from each animal were stained with hematoxylin and eosin. Histopathological scoring analyses were performed by board-certified independent veterinary histopathologists. Slides were scored for specific characteristics associated with pancreatitis on a score of 1 to 4 (1 = minimal, 2 = mild, 3 = moderate, and 4 = marked). The percentage of animals with findings and the average severity of findings were calculated.

To determine ductal cell replication, Ki-67 and pan-cytokeratin were used as markers of proliferation and ductal cells, respectively. For immunofluorescence, pancreas sections were deparaffinized in SafeClear (23-314-629; Fisher, Waltham, MA) and rehydrated. Antigen retrieval was performed via microwave heating in citrate buffer (S2369; Dako, Carpinteria, CA) for 32 min. Nonspecific binding was reduced by incubation in the avidin/biotin blocking kit (SP2001; Vector Laboratories, Burlingame, CA) and then in protein block serum-free solution (X0909; Dako). The following primary antibodies were used: Ki-67 (rat anti-mouse, 1:25, M7249; Dako) and cytokeratin (rabbit anti-mouse anti-pancytokeratin, 1:50, NC9726269; AbCam, Cambridge, MA). Secondary antibody labeled with FITC (1:50 for 1-h incubation) was used to visualize cytokeratin (AP187FMI; Millipore, Temecula, CA). The secondary antibody for Ki-67 (biotinylated anti-rat, 1:300, A-2002; Vector Laboratories) was applied for 1 h, followed by the fluorophore Rhodamine Avidin D for 10 min (1:200, BA-4000; Vector Laboratories). For quantification of ductal cell replication, cytokeratin-positive cells (∼1,000/section) were counted. The frequency of ductal cell replication in each animal was calculated as the total number of Ki-67-positive ductal cells per the total number of cytokeratin-positive cells.

Expression of genes encoding inflammatory mediators [IL-6, MCP-1, NF-κB, and myeloperoxidase (MPO)], tissue regeneration proteins [Reg3b, Egr-1 (early growth response gene), intracellular adhesion molecule-1 (ICAM-1)], or proteins involved in exocytosis in acinar cells (VAMP8) was assessed in formalin-fixed, paraffin-embedded mouse samples using the QuantiGene Plex 2.0 multiplexing assay (Affymetrix, Fremont, CA) according to the manufacturer's protocol. Expression of individual genes was normalized to GAPDH expression.

Statistical analyses.

Results are graphed using Prism 4 (GraphPad, San Diego, CA), and data are presented as means ± SE. The areas under the curve (AUC) for amylase and lipase were calculated using the trapezoid method. Statistical analyses were performed with Prism 4 or SAS 8.2 (SAS Institute, Cary, NC). Statistical differences between treatment groups (P < 0.05) and appropriate controls were identified with one-way ANOVA followed by Dunnett's or Bonferroni's multiple comparison test. For end points indicated in the results section, measurements were log10-transformed to normalize the data and assessed by ANOVA.


Effect of single-dose or subchronic administration of exenatide on baseline plasma amylase and lipase in normal or diabetic rats or mice.

Exenatide administered as a single injection to normal Harlan Sprague Dawley (HSD) or Zucker diabetic fstty (ZDF) rats at doses ranging from 0.072 to 0.72 nmol/kg had no effect on plasma amylase (Fig. 1 A and B, respectively) or lipase (Fig. 1, D and E, respectively) over 6 h compared with the vehicle control. Accordingly, calculated AUC for 0- to 6-h excursions of amylase (Fig. 1C) or lipase (Fig. 1F) did not differ from the respective vehicle controls in either strains (P > 0.05). Amylase (Fig. 1B) or lipase (Fig. 1E) concentrations in the ZDF animals were ∼1.5- and 2.5-fold higher, respectively, than in normal HSD rats.

Fig. 1.

Plasma amylase (A–C) and lipase (D–F) in rats dosed with single subcutaneous (sc) injections of different doses of exenatide (Ex; 0.072, 0.24, or 0.72 nmol/kg) or vehicle. A and B: time course of amylase in normal Harlan Sprague Dawley (HSD) and Zucker diabetic fatty (ZDF) rats, respectively. C: calculated area under the curve (AUC)0–6 h for amylase. D and E: time course of lipase in HSD and ZDF rats, respectively. F: calculated AUC0–6 h for lipase; n = 4–6/group. No statistically significant differences between Ex- and vehicle-treated groups were found for amylase or lipase.

Similarly, exenatide had no effect on basal plasma amylase or lipase concentrations during subchronic (4 wk) continuous sc dosing at 7.2 nmol·kg−1·day−1 in diabetic ob/ob (Fig. 2, A and C) or HF-STZ (Fig. 2, B and D) mice. This dose of exenatide was selected on the basis of Hb A1c and body weight-lowering efficacy in a parallel study performed in ob/ob mice (Supplemental Fig. S1, A and B, respectively; Supplemental Material for this article can be found on the AJP-Endocrinology and Metabolism web site). The mean exenatide plasma concentration at the termination of this study was 166 ± 24 pg/ml and was in the range of concentrations found in patients dosed with exenatide (20, 25). Additionally, 7.2 nmol·kg−1·day−1 of exenatide significantly decreased Hb A1c, plasma glucose, and body weight in HF-STZ mice compared with corresponding vehicle-treated HF-STZ controls (P < 0.05; Supplemental Fig. S2, A, B, and C, respectively).

Fig. 2.

Plasma amylase (A and B) and lipase (C and D) over time in diabetic mice dosed with Ex via continuous sc infusions for 4 wk. A and C: diabetic ob/ob mice were treated with Ex (1.2 or 7.2 nmol·kg−1·day−1) or with vehicle. Wild-type (WT) control mice were infused with vehicle. B and D: diabetic high-fat (HF)-streptozotocin (STZ) mice were treated with Ex (7.2 nmol·kg−1·day−1) or vehicle. Nondiabetic controls, which did not receive STZ, were infused with vehicle; n = 8–10/group.

Acute effects of exenatide on amylase and lipase during chemically induced pancreatitis in normal or diabetic rats or diabetic mice.

A single ip dose of 10 μg/kg CRN induced dramatic increases in plasma amylase and lipase in normal HSD (Fig. 3, A and D) and diabetic ZDF rats (Fig. 3, B and E). Intraductal infusion of ST resulted in an approximately threefold higher amylase and lipase release than was observed with CRN stimulation, indicating that the ST model induced experimental pancreatitis of greater severity than the CRN model (Fig. 3, C and F).

Fig. 3.

Time course of plasma amylase (A–C) and lipase (D–F) in rats dosed with Ex followed by chemically induced acute pancreatitis. Normal HSD, diabetic ZDF, or normal Wistar rats received a single sc dose of Ex (0.072, 0.24, or 0.72 nmol/kg) or vehicle. Sham-operated Wistar rats received vehicle. Ex, caerulein (CRN), or sodium taurocholate (ST) dosing is indicated by arrows. A, B, and C: plasma amylase for HSD, ZDF, and Wistar rats, respectively. D, E, and F: plasma lipase for HSD, ZDF, and Wistar rats, respectively; n = 6–7/group.

The effects of acute exenatide administration on amylase and lipase release were studied in the CRN and ST models of pancreatitis. Single injections of exenatide prior to induction of pancreatitis with CRN decreased plasma amylase (Fig. 3A) and lipase (Fig. 3D) in HSD rats. Exenatide did not affect amylase and lipase release in CRN-treated ZDF rats (Fig. 3, B and E). In Wistar rats with ST-induced pancreatitis, twice daily dosing of exenatide for 48 h did not change amylase (Fig. 3C) or lipase (Fig. 3F) concentrations. The corresponding AUC0–6 h calculated for both pancreatic enzymes in HSD rats treated with the highest dose of exenatide (0.72 nmol·kg−1·day−1) had significantly lower values for amylase (7,865 ± 670 U·l−1·h) or lipase (1,158 ± 164 U·l−1·h) than amylase (13,281 ± 1,445 U·l−1·h) or lipase (2,648 ± 481 U·l−1·h) in the vehicle-treated group (P < 0.05). The effects of lower doses of exenatide did not differ from those of vehicle controls; amylase AUC0–6 h was 11,273 ± 776 and 10,245 ± 1,211 U·l−1·h for the 0.072 and 0.24 nmol/kg exenatide doses, respectively. Lipase AUC0–6 h values in this study for these doses were 1,861 ± 248 and 1,859 ± 436 U·l−1·h, respectively. No significant differences in amylase or lipase AUCs were found for any dose of exenatide in CRN-treated ZDF or ST-treated Wistar rats (P > 0.05).

In a subsequent acute study in diabetic ob/ob mice, five hourly ip injections of CRN (10 μg/kg) induced dramatic increases of plasma amylase and lipase similarly to rat studies. A single injection of exenatide (5 nmol/kg) given 15 min prior to induction of pancreatitis had no effect on pancreatic enzyme secretion (data not shown).

Effect of 4-wk dosing of exenatide on amylase and lipase in ob/ob mice with CRN-induced pancreatitis.

To examine whether a longer duration of exenatide exposure affected pancreatitis-related end points in rodent models of diabetes, CRN was administered following a 4-wk sc infusion of exenatide (1.2 and 7.2 nmol·kg−1·day−1) in diabetic ob/ob mice. The time courses for plasma amylase and lipase concentrations and the corresponding AUC0–6 h are shown in Fig. 4. Compared with vehicle treatment, exenatide at 7.2 nmol·kg−1·day−1 significantly reduced plasma amylase (Fig. 4, A and B) and lipase (Fig. 4, C and D) responses to CRN. Pancreatic enzyme excursions were similar in magnitude to those observed in normal wild-type mice without pancreatitis. Exenatide at a dose of 1.2 nmol·kg−1·day−1 did not modify amylase or lipase concentrations.

Fig. 4.

Time course of plasma amylase (A) and lipase (C) during CRN-induced pancreatitis in diabetic ob/ob mice pretreated with continuous sc infusion of different doses of exenatide (1.2 or 7.2 nmol·kg−1·day−1) or with vehicle for 4 wk. Strain-matched WT mice were pretreated with vehicle. CRN dosing is indicated by arrows. B and D: calculated 6-h plasma amylase (B) and lipase (D) AUC; n = 8–10/group. *P < 0.05, Ex (7.2 nmol·kg−1·day−1) or vehicle-dosed WT vs. vehicle-dosed ob/ob mice.

Effects of exenatide on pancreatic weight in rats and mice with or without chemically induced pancreatitis.

Pancreatic weight is used as an indicator of fluid retention and edema in experimental pancreatitis. Terminal pancreatic weights normalized to body weight in all rat and mouse studies are summarized in Table 2. In general, exenatide had no effect on pancreatic weight when administered acutely to rats or mice and did not affect pancreatic weight in subchronic studies in two strains of diabetic mice. Similarly, single doses of CRN did not change pancreatic weight in normal HSD or diabetic ZDF rats, although five hourly injections of CRN significantly increased relative pancreas weight in ob/ob mice (P < 0.01). When exenatide was administered either prior to (CRN-treated HSD rats, ZDF rats, or ob/ob mice) or during the induction of pancreatitis (ST-treated Wistar rat, CRN-treated ob/ob mice), exenatide did not modify pancreatic weight in any of the rodent models studied.

View this table:
Table 2.

Effects of exenatide treatment on terminal pancreatic weight normalized to body weight in different rat and mouse models of chemically induced pancreatitis

Effects of subchronic administration of exenatide on inflammatory cytokines in diabetic ob/ob mice with CRN-induced pancreatitis.

To determine whether exenatide influences inflammatory mediators during experimental pancreatitis, we measured IL-1β, IL-2, IL-6, MCP-1, IFNγ, and TNFα in the plasma from ob/ob mice at the study's end. The summary of results from acute and subchronic exenatide administration presented in Table 3 demonstrates that CRN markedly increased plasma concentrations of IL-6 and MCP, whereas exenatide significantly decreased TNFα at a dose of 1.2 but not 7.2 nmol·kg−1·day−1 after 4 wk of treatment. Trends toward decreases in IL-2, IFNγ, and MCP-1 were also observed with continuous infusion of exenatide prior to CRN injury.

View this table:
Table 3.

Effects of exenatide pretreatment on plasma cytokines at 6 h post-CRN-induced acute pancreatitis in diabetic ob/ob mice

Effects of acute or subchronic administration of exenatide on pancreas histology in normal or diabetic rats or mice after induction of experimental acute pancreatitis.

Histological analyses revealed that CRN administration did not adversely affect the pancreatic tissue structure of HSD or ZDF rats, whereas ST administration induced degenerative, inflammatory, and compensatory changes in the pancreas, including cellular infiltration, vacuolation, edema, necrosis of acinar tissue and cells, hemorrhage, and ductal enlargement (Table 4). Exenatide did not affect the normal histology of CRN-treated pancreata from HSD or ZDF rats. However, after ST injury, exenatide at the highest dose (0.72 nmol·kg−1·day−1) was associated with a lower severity of acinar cell vacuolation, interstitial edema, and interstitial adipose tissue necrosis in treated pancreata compared with vehicle-treated pancreata, and interlobular ducts were more likely to be dilated after exposure to a high dose of exenatide (Table 4).

View this table:
Table 4.

Effects of twice daily injections of exenatide on %incidence and mean severity of morphological changes in pancreas of Wistar rats at 48 h post-ST-induced pancreatitis

In ob/ob mice, CRN produced degenerative and inflammatory pathological changes in pancreata (Fig. 5A and Table 5). Markedly fewer histological changes were observed after exposure to CRN in nondiabetic (wild-type) mice than in diabetic ob/ob mice, specifically with respect to acinar cell vacuolation, single cell degeneration in islets, and acute inflammation or necrosis of adipose tissue.

Fig. 5.

Pancreatic histology. A: representative sections of pancreata from diabetic ob/ob mice treated with continuous sc infusion of Ex (7.2 nmol·kg−1·day−1) for 4 wk, followed by CRN-induced pancreatitis. White bar represents 100 μm. B: cells in the cell cycle stained positive for Ki-67 (red) and are indicated by white arrows. Ductal epithelial cells stained positive for pan-cytokeratin (pan-CK; green) and nuclei were counterstained with 4,6-diamidino-2-phenylindole (blue). C: quantification of ductal cell proliferation rates presented as the percentage of Ki-67-positive cells within pan-CK-positive cells; n = 10/group. No statistically significant differences between groups were found (P = 0.11).

View this table:
Table 5.

Effects of 4-wk pretreatment with exenatide or vehicle on %incidence and mean severity of morphological changes in pancreas after CRN-induced pancreatitis in diabetic ob/ob mice or strain-matched WT mice (n = 10/group)

Four weeks of treatment with exenatide (7.2 nmol·kg−1·day−1) had subtle beneficial effects on pancreatic histology after exposure to CRN, decreasing the incidence of single-cell acinar cell necrosis and single-cell degeneration in islets compared with vehicle controls (Table 5 and Fig. 5A). Vascular ectasia (dilation) was also increased in exenatide-treated animals (Table 5). The incidence and severities of focal ductal proliferation and chronic periductal inflammation in exenatide-treated animals more closely resembled those of nondiabetic controls than those of vehicle-treated ob/ob mice.

To determine whether exenatide affected general (nonfocal) pancreatic duct cell proliferation, we examined pancreata from ob/ob mice treated with exenatide for 4 wk, followed by CRN-induced pancreatitis. Representative immunofluorescence images of ductal structures and replicating cells labeled with Ki-67 from wild-type controls, exenatide, or vehicle-treated ob/ob mice are shown in Fig. 5B. The rate of ductal cell proliferation was low (<1%) and similar in all groups (P < 0.11) (Fig. 5C).

Effect of exenatide on expression of genes associated with pancreatitis.

The levels of mRNA for genes known to be associated with pancreatitis [Reg3b, Egr1, Icam1, Il6, Ccl2 (encoding MCP-1), Nfkb1, Mpo, and Vamp8] are shown in Fig. 6 (results for ICAM-1 and MPO are not presented because levels were below the level of detection). Consistent with biochemical studies of pancreatitis-associated inflammation, expression of the genes encoding MCP-1 and IL-6 was dramatically induced after exposure to CRN in the pancreata of diabetic mice compared with those of wild-type controls. Exenatide did not modify expression of these two genes. Similarly, exenatide did not modify expression of Egr1, Nfkb1, or Vamp8. However, an exenatide dose-related induction of Reg3b expression was observed, which reached statistical difference for the high dose of exenatide (7.2 nmol·kg−1·day−1) compared with the vehicle-treated diabetic control (P < 0.05).

Fig. 6.

Effects of Ex on gene expression in the pancreas at 6 h post-CRN-induced acute pancreatitis in diabetic ob/ob mice treated with continuous sc infusion of different doses of Ex or vehicle for 4 wk. Strain-matched WT mice were treated with vehicle infusions and were subject to CRN-induced pancreatitis as well; n = 9–10/group. *P < 0.05, Ex (7.2 nmol·kg−1·day−1) or WT vs. vehicle-dosed ob/ob mice.


The results of this study demonstrate that acute or subchronic pharmacological dosing of exenatide did not induce acute pancreatitis in normal or diabetic/obese rodents and did not increase the severity of two types of chemically induced pancreatitis in rats or mice. In general, the histological results for the pancreata of wild-type and ob/ob mice treated with exenatide were consistent with measurements of plasma amylase, lipase, and inflammatory cytokines. Furthermore, at the highest dose of exenatide tested in mice, which provided plasma exposures similar to those observed in patients with type 2 diabetes (20, 25), exenatide reduced amylase and lipase release in response to CRN and histological evidence of pancreatic injury.

The results of the present studies are consistent with previous extensive nonclinical toxicology assessments, which were undertaken to support the development and marketing of exenatide twice daily. Single- and repeated-dose toxicology studies in mice and studies ≤2 yr in rats and ≤9 mo in monkeys at doses that provided plasma exposures substantially in excess of human exposures showed no drug-related effects on the exocrine pancreas (37). Two-year carcinogenicity studies were also conducted in mice and rats, and no exenatide-related preneoplastic changes or increased incidences of tumors in the exocrine pancreas were observed. Exenatide was also found to be devoid of mutagenic or clastogenic potential. Overall, the nonclinical safety data indicate that acute or chronic administration of exenatide is not associated with any adverse effects on the pancreas (including the exocrine pancreas). Since toxicology studies are performed predominantly in normal animals, additional toxicology studies will be needed in animal models of type 2 diabetes with or without experimentally induced pancreatitis to evaluate the effects of exenatide at doses higher than the pharmacology doses reported in this article.

Others have also investigated the effects of exenatide on the pancreas. In a recent study, Nachnani et al. (33) reported that exenatide (10 μg or 2.4 nmol·kg−1·day−1, ∼10 times the clinical dose) did not affect amylase concentrations when dosed for 75 days in normal Sprague-Dawley rats, which is consistent with our findings in mice. On the other hand, exenatide showed a slight but significant effect on serum lipase concentrations, which we did not observe. Because the concentration of pancreatic enzymes in the systemic circulation increases by orders of magnitude during acute pancreatitis, a slight increase in lipase is unlikely to be clinically important. Nachnani et al. (33) also observed subtle increases in inflammation and pyknotic nuclei (3 to 4 on a scale of 40) with exenatide treatment, which were not observed in our studies. The physiological importance of these findings and their relevance to the effects of clinical doses of exenatide in humans is unknown.

Another study concluded that exenatide did not modify susceptibility to or severity of CRN-induced pancreatitis in mice (23). This previous report observed small increases in pancreas weight upon exposure to exenatide or liraglutide, which differs from our observations. The previous report also observed that exenatide increased expression of immediate early genes such as Egr, which was not observed here. Neither this nor the previous study observed increases in proinflammatory mediators such as MCP-1. However, our present study is the third to observe exenatide-associated increases in the anti-inflammatory gene PAP (Reg IIIβ), which is believed to be protective against cellular injury (11, 23).

Recently, increased ductal cell turnover and metaplasia were reported in diabetic, transgenic, overexpressing human islet amyloid polypeptide rats treated with the DPP IV inhibitor sitagliptin, with one rat exhibiting histological features of acute pancreatitis (29). In this report, a DPP IV inhibitor rather than a GLP-1R agonist was used. DPP IV inhibitors prevent the degradation of multiple signaling peptides (31), including substance P, which is known to be involved in pancreatitis (14, 16). Therefore, potential risks associated with GLP-1R agonism-based therapy cannot be extrapolated from these data and applied to exenatide. It is also feasible that sitagliptin itself may produce direct effects on the pancreas independent of circulating peptides that may be elevated through DPP IV inhibition.

The baseline findings observed in the animal models of pancreatitis appear to be consistent with those of human pancreatitis in patients with type 2 diabetes or obesity (6, 8, 34). Diabetic ZDF rats had elevated plasma amylase and lipase concentrations compared with their nondiabetic counterparts, which is suggestive of a pancreatic, hyperactive state. Likewise, ob/ob diabetic and obese mice exhibited higher amylase concentrations than normal nondiabetic animals prior to chemical injury. Additionally, amylase and lipase responses to chemically induced pancreatitis were greater in these mice than in normal wild-type controls. For our studies, the CRN and sodium taurocholate animal models of pancreatitis were selected on the basis of these being the best-characterized and most commonly reported models in literature (35). These models mimic clinical symptoms of pancreatitis via multiple-fold increases in amylase/lipase concentrations in plasma, pancreatic edema, and histopathological changes in the exocrine pancreas. In general, the exact mechanisms triggering even non-drug-related acute pancreatitis attack in humans are still unclear, and therefore, the full relevance of any animal models of pancreatitis to the clinical condition remains somewhat limited.

Since multiple measurements in several model systems demonstrated improvements on manifestations of pancreatitis with exenatide treatment, it is important to consider the mechanisms that may be responsible for the observed beneficial effects. First, cytoprotection of pancreatic β-cells via downregulation of apoptosis, upregulation of proliferation, or neogenesis by exenatide has been very well documented in multiple studies (7, 19, 42), as have effects on Reg3b in the pancreas (11, 23). Therefore, it cannot be ruled out that exenatide may potentially exert protective effects on pancreatic acinar or ductal cells.

Second, exenatide slows gastric emptying (24, 27), affects antro-pyloro-duodenal motility (36), and inhibits pancreatic bicarbonate and protein secretion (40). The pancreas is densely innervated and vascularized with neuronal and endothelial cells, respectively. These cells can express the GLP-1R (3, 32). Therefore, central or peripheral direct or indirect neural regulation of exocrine secretion, vascular tone, or permeability could contribute to the beneficial effects of exenatide on digestive enzyme secretion.

Finally, pancreatitis is associated with inflammation, which is manifested by elevated proinflammatory cytokines in plasma or pancreatic tissue (4, 30). Exploratory data from our studies in ob/ob mice subchronically dosed with exenatide did not show any increase of several inflammatory mediators (IL-1β, IL-2, IFNγ), as measured in plasma or at gene levels in the pancreas (Il-6, MCP-1, NF-κB). Moreover, a trend toward downregulation was observed for IL-2, IFNγ, MCP-1, and TNFα. TNFα is known to induce apoptosis in pancreatic β-cells, and exenatide directly antagonized this effect in vitro (26). It remains to be demonstrated whether these preliminary rodent results will translate to clinical observations.

In conclusion, the data presented do not indicate that exenatide has detrimental effects on the exocrine pancreas in rodents. To the contrary, the present results from animal models of diabetes, obesity, and/or chemically induced pancreatitis suggest that exenatide might reduce the severity of acute pancreatitis in rodents.


At the time when the work for this article was planned, all authors were employed by Amylin Pharmaceuticals. D. M. Hargrove and B. R. Gedulin were partially involved in the generation and analysis of the data and in preparation of the manuscript. All authors hold stock in Amylin Pharmaceuticals.


We thank Moncerrat Melgoza for technical assistance. We are grateful to Mary Beth DeYoung for assistance with the manuscript writing and editing and to Joseph S. Heilig for the discussion of our data. We thank Tod Coffey and Chi-Hse Teng for their assistance in statistical analysis. We appreciate the contribution made by the contract research organization Charles River Laboratories in preparation of histological specimens and histopathological evaluation, Zyagen in preparation of histological specimens, and Affymetrix in gene expression analysis.

Present address for D. M. Hargrove: Ferring Research Institute, Inc., San Diego, CA 92121.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
View Abstract