Acute pancreatitis-induced islet dysfunction in ferrets.

To test whether cerulein induces acute pancreatitis in ferrets, we measured amylase, lipase, cytokine levels, MPO staining, and performed histological scoring of the H&E-stained pancreas sections. Compared to controls, ferrets treated with cerulein exhibited histopathologic changes observed in secretagogue-induced acute pancreatitis including mild interstitial edema and scattered neutrophil infiltration ( Figure 1A – D ). Although the histological scoring of H&E slides did not show any significant difference in edema, inflammation or acinar cell necrosis between the groups ( Fig 1E – G ), there was a mild increase in pancreatic MPO staining within the pancreas of cerulein-treated ferrets, consistent with neutrophilic infiltration ( Fig. 2A – D ). Plasma amylase and lipase were significantly elevated in cerulein-treated animals compared to controls (1.8 and 2.8-fold higher respectively) ( Fig. 3A , B ). In addition, plasma IL-6 and TNF-α, both markers of inflammation were significantly upregulated, 23- and 30-fold higher respectively, in cerulein-injected ferrets compared to controls ( Fig. 3C , D ). Inflammation was confined to pancreas, with no injury to the lungs (data not shown). Taken together, these results indicate that cerulein administration leads to mild pancreatic injury and a profound inflammatory response. We next examined whether acute pancreatitis induced by cerulein impacted fasted blood glucose and plasma endocrine and incretin hormone levels. Twenty hours after cerulein injection, fasting glucose remained significantly elevated ( Fig. 4A ), plasma insulin was significantly lower ( Fig. 4B ), pancreatic polypeptide (PP) and glucagon like peptide-1 (GLP-1) levels were significantly higher ( Fig. 4D , E ) in 3 hr fasted ferrets with acute pancreatitis compared to controls. Glucagon and GIP levels were not significantly different between the groups ( Fig. 4C , F ). Plasma cortisol and growth hormone levels were not different between cerulein and saline-injected ferrets ( Supplemental Fig. S2A , B ). These results suggest that AP induced by cerulein significantly impacts glycemic regulation at 24h post-injury, possibly by inhibiting insulin secretion. To determine whether hyperglycemia occurs during the early stages of acute pancreatitis, we monitored the plasma glucose levels every 4 to 8 hours for 24 hours after cerulein administration. Compared to controls, cerulein-treated animals exhibited a marked increase in plasma glucose at all time points, with a peak at 12 hours ( Fig. 5A ). Area under the curve (AUC) measured during the course of AP was significantly higher in ferrets with acute pancreatitis compared to controls ( Fig. 5B ). To better understand how acute pancreatitis influences the glucose metabolism, we performed OGTT, MMTT, and ITT, 24 h after cerulein or saline injection in 3h-fasted animals. Plasma glucoses and AUC measured during OGTT ( Fig. 5C , D ), MMTT ( Fig. 5E , F ), and ITT ( Fig.5G , H ) demonstrated significantly elevated glucoses in animals that have received cerulein compared to controls. Taken together, these results show that ferrets with acute pancreatitis have impaired glucose tolerance (IGT) and insulin resistance compared to controls. We postulated that hyperglycemia in acute pancreatitis could be related to defects in islet and enteric hormone secretion. We proceeded to answer this question by measuring hormone levels during OGTT and MMTT. Following ingestion of oral glucose, plasma insulin levels rose sharply and peaked at 30 min, then fell back to baseline over the next 120 min in both control and cerulein-treated animals. Compared to controls, plasma insulin levels were markedly lower in ferrets treated with cerulein, with significant difference at 30 min and by AUC ( Fig 6A , B ). Plasma PP levels also rapidly reached the peak at 30 min after glucose ingestion. Compared to controls, PP levels were higher in cerulein-treated ferrets at 30 min, but there was no difference in overall secretion as judged by AUC ( Fig. 6C , D ). There was no difference in the secretion of glucagon, GLP-1 and GIP secretion between the groups ( Fig. 6E – J ). These OGTT results suggest that the hyperglycemic response in animals with acute pancreatitis is, at least in part, due to an insulin secretory defect. We next proceeded with MMTT to better determine islet and incretin hormone responses in animals with AP ( Fig. 7 ). Plasma insulin levels peaked at 30 min and they were significantly lower in ferrets treated with cerulein compared to controls at 30 to 120 min ( Fig 7A ). Likewise, the AUC for insulin was also significantly lower in in ferrets with acute pancreatitis ( Fig 7B ). There was no difference in PP levels between treatment and control groups ( Fig. 7C , D ). Glucagon and GLP-1 levels were significantly elevated in AP ferrets with relative higher levels at 60 and 120 min and 60, 120 and 180 min, respectively ( Fig 7E – H ). In contrast, GIP levels were significantly higher at 30 min in controls ferrets and overall secretion was decreased in ferrets that received cerulein ( Fig. 7I , J ). MMTT results suggest that animals treated with cerulein demonstrate reduced insulin secretion and increase in secretion of glucagon and incretin hormones. To evaluate whether changes observed in cerulein-treated ferrets were due to possible structural changes in the endocrine pancreas, we evaluated islet insulin expression by immunofluorescence. As shown in Fig. S3 , both AP and control ferrets had intact islets with no discernable differences in insulin expression. Thus, we conclude that defective insulin secretion in AP ferrets was not due to the destruction of islets. To examine whether the impact of acute pancreatitis on islet dysfunction was long-lasting, we repeated MMTT and OGTT 3 months after the animals have received cerulein, measuring glucose and hormone levels ( Supplemental Figs. S4 and S5 ). There were no differences between the groups suggesting that the effect of cerulein on the islet dysfunction was transient.

The study was performed on four-week-old ferrets. Animal maintenance conditions and experimental protocols were approved by the University of Iowa Animal Care and Use Committee. Figure S1 summarizes the study protocol. The rationale was to induce acute pancreatitis using cerulein and immediately thereafter conduct studies to measure islet cell dysfunction and insulin resistance. We utilized the cerulein-induced AP protocol described in mice 25 with some modifications to accommodate oral glucose tolerance test (OGTT), mixed meal tolerance test (MMTT), and insulin tolerance test (ITT). Briefly, after 3h fasting, animals were given intraperitoneal (IP) injections of cerulein (50 μg/kg body weight, Sigma-Aldrich, St. Louis, MO) or saline every hour (total 9 injections) to induce acute pancreatitis. They had access to water during cerulein injections and were offered food for 12h after completion of injections. Blood was obtained at 6h after the first cerulein injection for interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-α) measurements in plasma. Twenty-four hours after the first cerulein injection, blood was obtained for amylase, lipase and hormone measurements in plasma; OGTT, MMTT, ITT were then performed, followed by euthanasia and harvesting of the pancreas. At necropsy, pancreas was fixed in 10% neutral buffered formalin, processed and embedded in paraffin, and sectioned at 4 μm. Sections were stained with hematoxylin and eosin (H&E) for light microscopy and scored in a semi-quantitative fashion by a board-certified veterinary pathologist (KNGC) for pancreatic interstitial edema, inflammation and acinar cell necrosis as previously described 26 , 27 : edema 0 = none present; 1 = very mild, rare increases in interstitial clear space (<10% of tissue on slide is affected); 2 = multifocal expansion of the interstitium with clear space (between 10–50% of tissue on the slide is affected); 3 = diffuse, marked expansion of the interstitium with clear space (>50% of tissue on the slide is affected). For inflammation, 0 = none present; 1 = rare, scattered, individualized inflammatory cells are present with the pancreatic parenchyma and/or interstitium; 2 = multifocal to coalescing inflammatory cell infiltrates are present; 3 = diffuse, marked inflammatory cell infiltrates are present. For acinar cell necrosis, 0 = none present; 1 = rare, scattered, individualized acinar cells are necrotic, characterized by hypereosinophilic cytoplasm with pyknotic nuclei and/or nuclear fragmentation; 2 = multifocal, sometimes clustered acinar cell necrosis; 3 = coalescing to large foci of acinar cell necrosis. Sections were also stained for myeloperoxidase (MPO) via immunohistochemistry. Briefly, slides were antigen retrieved with 10 mM Tris/EDTA buffer at a pH of 9.0 in a decloaker at 125°C for 5 minutes, quenched with 3% hydrogen peroxidase for 8 minutes and the MPO antibody (ThermoFisher PA5–16672; Waltham, MA) was applied at a dilution of 1:50 in buffer for 60 minutes at room temperature followed by detection with the DAKO (Santa Clara, CA) Rabbit Envision HRP system reagent for 30 minutes, DAKO DAB plus for 5 minutes and counterstained with hematoxylin. MPO stained slides were evaluated and MPO immunoreactive cells were counted in five separate high-powered fields (40x magnification) per slide and these were averaged for each slide for final quantification. Blood was obtained from the tail vein of ferrets at 0, 4, 8, 12 and 24 hours after first cerulein or saline injection. Glucose levels were measured using a glucometer. Oral glucose tolerance test was performed following 3h fasting on animals treated with saline or cerulein injections, as previously described 23 . An oral glucose solution (TRUTOL 100 Glucose tolerance beverage, Thermo Fisher Scientific, Middletown, VA) of 1g per kg body weight was ingested by ferrets over 2 minutes. Blood samples were collected in chilled heparinized tubes at 0, 30, 60, 120, and 180 min after the ingestion of oral glucose. Blood samples were centrifuged at 4°C applying 3000 xg to separate plasma, which was then stored at −80°C for hormone assays. Blood glucose levels were determined using a glucometer. A mixed meal tolerance test was performed following 3h fasting on animals treated with saline or cerulein injections, as previously described 23 . Elecare (Abbott, Chicago, IL) was ingested by animals over 5 minutes. Elecare dose was normalized to the body surface area estimate (BSAE in cm 2 ) of each animal, as previously described 23 . Typically, 6.45 mg carbohydrate per cm 2 of BSAE was used. Blood samples were collected in chilled heparinized tubes at 0, 30, 60, 120, and 180 min after Elecare ingestion. Blood glucose levels were determined with a glucometer. Insulin tolerance test was performed following 3h fasting on animals treated with saline or cerulein injections. Insulin was given as 1U per kg body weight intraperitoneally. Blood samples were collected in chilled heparinized tubes at 0, 30, 60, 120, and 180 min after insulin injection. Blood samples were centrifuged at 4°C applying 3000 xg to separate plasma, which was then stored at −80°C for hormone assays. Blood glucose levels were determined using a glucometer. Plasma insulin, glucagon, PP, GLP-1, and GIP were measured using a Human Metabolic Hormone Magnetic Bead Panel (Millipore Burlington, MA). The antibody pairs were specific only to the desired analyte and exhibited no or negligible (<2%) cross-reactivity with other analytes in the panel. Plasma IL-6 and TNF-α were measured using a Canine Cytokine Magnetic Bead Panel (Millipore Burlington, MA). There was negligible cross-reactivity among different analytes within the panel. Plasma Growth Hormone was measured using a Rat/Mouse Growth Hormone ELISA kit (Millipore, Burlington, MA). Plasma cortisol was measured using a Cortisol Competitive ELISA kit (Invitrogen/Thermo Fisher Scientific, Waltham MA). Plasma pancreatic lipase was measured using a Human Pancreatic Lipase Assay (Kamiya Biomedical Company, Seattle, WA) following the manufacturer’s protocol. Plasma amylase was measured using an Amylase Assay Kit (Abcam, Cambridge, MA) following the manufacturer’s protocol. At necropsy, pancreata were fixed in 10% neutral buffered formalin, embedded in paraffin, and sectioned at 4 μm. Tissue sections were subsequently deparaffinized, rehydrated, boiled 20 min in 0.1 M sodium citrate (pH 6.0) for antigen retrieval, and blocked with 100 −120 μl per slide of PBS with 10% donkey serum + 0.5% Triton-X-100 for 1 hour at room temperature. Sections were incubated overnight at 4°C with the primary antibody, guinea pig anti-insulin at a 1:500 dilution (MP Biomedicals, Santa Ana, CA). The secondary antibody was rhodamine red-conjugated donkey anti-guinea pig at 1:250 dilution (Jackson ImmunoResearch, West Grove, PA). Following staining, samples were rinsed in PBS and placed in Vectashield Mounting Medium (Vector Laboratories, Burlingame, CA) containing DAPI to stain nuclei prior to imaging with Zeiss 700 confocal microscope. All data were presented as mean ± SD. Area under the curve (AUC) of OGTT and MMTTs was calculated using the trapezoid rule. Statistical comparisons between two groups were performed by two tail student’s t-test analysis and p < 0.05 was considered to indicate significant differences.

In the ferret model, exposure to cerulein induced mild injury to the pancreas but marked increases in proinflammatory cytokines. Animals exhibited profound islet dysfunction characterized by fasting hyperglycemia, impaired insulin secretion and altered pancreatic and incretin hormones. There was no morphological injury to the islets, or insulin-expressing β cells, and the impaired glucose tolerance and hormonal abnormalities were transient. To our knowledge, this is the first report of cerulein-induced acute pancreatitis in ferrets. Similar to mouse and rat models of cerulein-induced AP 28 – 30 , ferrets exhibit an inflammatory reaction to cerulein. The major difference between mouse and rat models of cerulein-induced AP is that mice develop less interstitial edema and intracellular vacuolization, but more acinar cell necrosis compared to rats 31 . As in rats 28 , pancreatic injury is mild in ferrets in response to cerulein, characterized by interstitial edema, neutrophil infiltration, elevations in plasma amylase and lipase, and marked elevation in proinflammatory cytokines, IL-6 and TNF-α. In contrast to rats, which demonstrated no abnormalities in serum glucose and insulin levels after cerulein exposure 32 , ferrets have developed glucose intolerance and significant hormonal dysregulation. The differences in cerulein-response across animal models is unknown but may be due to variations between species. There are differences among nearly all species, including distinct contrasts between rodents and humans. One possibility for variations in hormonal responses across animal models may be due to their islet cryoarchitecture and composition 33 . As emphasized in the review of animal models of pancreatitis by Saloman et al. 29 , various induction protocols exist for acute and chronic pancreatitis, therefore it is important to consider the most appropriate time points to assess the chosen parameters. Since there were no previous published studies of cerulein-induced acute pancreatitis in ferrets, we used published protocols in mice and had to make modifications to accommodate time needed for MMTT, OGTT and ITT. Because cytokines are expected to peak earlier compared to amylase and lipase levels and histopathologic changes 34 , we determined cytokines at an earlier time point. It is possible that modifying the protocol to accommodate the islet cell dysfunction and insulin resistance studies has impacted our ability to detect the histological grade of pancreatitis at earlier time points. Hyperglycemia during AP is long believed to occur as a reaction to “stress” triggered by the acute illness 13 . Stress hyperglycemia is caused by the interplay of counter-regulatory hormones including growth hormone, cortisol and cytokines 14 . This phenomenon is associated with excessive hepatic glucose production through glucagon secretion and gluconeogenesis, and characterized by insulin resistance 35 – 37 . We have indeed observed insulin resistance during ITT, increased glucagon secretion during MMTT, and high levels of proinflammatory cytokines that may all be consistent with stress hyperglycemia during AP. However, the stress hormone cortisol and growth hormone were not elevated in ferrets with AP. One possibility is that cerulein may have direct or indirect effect on the islets. This is unlikely as cerulein has a very short-half life, measured in minutes 38 . Our observations in ferrets are consistent with an acute and transient glucose intolerance. This is in contrast to recent observations that reported an increased risk for prediabetes and/or diabetes risk after an attack of acute pancreatitis in humans 15 – 18 . The mechanisms of increased long-term risk of diabetes mellitus after a single attack of acute pancreatitis remain unknown. There are no longitudinal and prospective studies that followed patients from the initial acute pancreatitis attack to the development of diabetes mellitus to identify the risk factors. Only few studies have reported functional islet studies such as OGTT 39 , 40 . Some patients had abdominal obesity, pancreatic fat deposition and insulin resistance after AP 41 , 42 , concerning that mechanisms typically seen in type 2 diabetes may be contributing factors. Although exocrine and endocrine systems have long been viewed as separate systems, there is an increased understanding of their anatomical, morphological and functional connections and their joined influence on pancreatic diseases 43 . Indeed, early studies in animals have suggested insulin having a role in zymogen release and acinar cell growth, thus the term “insulin-acinar axis” was proposed 44 . For example, streptozotocin-induced islet destruction not only causes islet cell loss, but also functional reduction in the exocrine pancreas 45 . Likewise, exocrine pancreatic insufficiency without evidence of chronic pancreatitis is now well-recognized in patients with type 1 or type 2 diabetes mellitus 46 . Moreover, a new type of diabetes, pancreatogenic diabetes or type 3c diabetes mellitus (T3cDM) has been reported in patients with exocrine pancreatic diseases 47 . The mechanism by which islet dysfunction occurs in exocrine pancreatic diseases and specifically acute pancreatitis are unknown. It is possible that pancreatic islet microenvironment, their rich vascular and neuronal connections and their cell-to cell contact with acinar cells may be playing a role 48 . Although the effects of cerulein-induced acute pancreatitis on the endocrine pancreas was transient in our studies in ferrets, some hormone levels during the acute phase resembled those found in patients with pancreatogenic diabetes or T3cDM including insulin deficiency, reduced glucose-stimulated insulin secretion and hyperglucagonemia 49 – 52 . As seen in T3cDM 53 , 54 , we observed insulin resistance, but no PP deficiency in our model. Our findings pertaining to incretins showed a stimulated GLP-1 and low GIP. In humans with T3cDM, incretin responses have been mixed probably due to low sample size 55 . Overall, there are similarities between the glycemic defects and hormonal changes observed in the ferret AP model and human pancreatogenic diabetes mellitus. These changes may be due to the inhibition of insulin secretion caused by cytokine release during the acute phase 56 . It is possible that repetitive proinflammatory insults in people with chronic pancreatic diseases would impact the islet microenvironment and lead to persistent changes. In summary, acute exposure to cerulein causes acute pancreatitis, hyperglycemia, transient islet dysfunction, but no structural injury to the islets in ferrets through a mechanism that involves exocrine pancreatic inflammation. Future studies in the ferret cerulein model may contribute to the understanding of pancreatogenic diabetes.

Acute pancreatitis (AP) is an inflammatory disease of the pancreas associated with local tissue damage, systemic inflammatory response and organ failure 1 . Acute pancreatitis creates a major burden for children and adults with increased need for emergency room visits, hospitalizations, multiple tests and imaging studies, procedures and surgeries 2 , 3 . Indeed, AP is the leading cause of inpatient admissions among gastrointestinal disorders, causing an estimated $2.6 billion in inpatient costs 4 . Acute pancreatitis is considered a disease specific to the exocrine pancreas with little understanding of the endocrine complications. Exocrine pancreatic function may be impacted during AP, sometimes requiring pancreatic enzyme replacement therapy, depending on the severity of pancreatitis, the amount of necrosis, and underlying risk factors 5 – 7 . It is expected that the exocrine pancreatic function will completely recover after an AP attack, unless there is significant tissue necrosis, chronic inflammation, fibrosis and profound acinar cell loss. It has long been recognized that patients with AP may suffer from varying degrees of hyperglycemia during the attack 8 , 9 . Hyperglycemia during AP has been considered as a poor prognostic sign 10 , 11 and more commonly seen with severe acute pancreatitis or pancreatic necrosis 12 . The mechanisms of hyperglycemia during AP are unknown, but thought to be secondary to the stress of acute illness 13 , called “stress hyperglycemia” 14 . The reason why acute pancreatitis causes hyperglycemia is poorly understood because most patients are not closely monitored or undergo broad functional metabolic testing during or after the event. In general, it is assumed that a single acute pancreatitis event resolves without long term sequelae, and that blood glucoses return to normal as patients recover from AP. However a recent systematic review reported persistent hyperglycemia in people with acute pancreatitis after discharge from the hospital 15 and three recent population-based studies reported prediabetes and/or diabetes risk two-fold or higher after a single attack of acute pancreatitis in humans 16 – 18 . These studies suggest that exocrine pancreatic inflammation may have long-lasting effects on the endocrine pancreas for unknown reasons. The pathogenesis of hyperglycemia induced by AP remains unknown due to the inaccessibility of human tissues and lack of representative animal models. To investigate the mechanisms of glucose homeostasis disturbances during AP, we developed a ferret acute pancreatitis model using cerulein. Ferrets have been broadly used in biochemical studies 19 and they display many features that resemble human physiology. In the cystic fibrosis (CF) ferret model, pancreatic involvement mimics exocrine pancreatic damage and CF-related diabetes (CFRD) typically seen in CF patients 20 – 24 . Thus, to better understand the impact of acute pancreatitis on glycemic control and insulin secretion, we administered cerulein injections to ferrets. Animals developed hyperglycemia and decreased insulin secretion that dissipated after resolution of pancreatic inflammation. These findings suggest that the AP ferret model is a feasible tool to investigate the pathogenesis of glycemic disturbances associated with acute pancreatitis.

Fig. S1. Study Protocol. Schematic presentation of the research protocol. After a 3h fast, ferrets were injected with saline or cerulein (50 μg/kg) every hour x9. After completion of injections, they were offered food for 12h. Twenty-four hours after the first cerulein injection and following a 3h fast, oral glucose tolerance test (OGTT), mixed meal tolerance test (MMTT), insulin tolerance test (ITT) were performed. Blood was obtained for glucose, amylase, lipase, interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-α) at time points indicated. Animals were euthanized and pancreata were harvested upon completion of OGTT, MMTT, ITT. Fig S2. Plasma growth hormone and cortisol levels are not elevated in ferret AP model. Blood was drawn from fasted saline- or cerulein-injected ferrets at 24 hr following the last cerulein injection. (A) Plasma Growth Hormone, (B) plasma cortisol. Results are shown as mean ± SD. P values are shown. Fig. S3. Ferrets with acute pancreatitis have intact β-cells. Insulin immunostaining from saline or cerulein-treated ferrets. Insulin (green), DAPI (blue). Image is representative of three independent experiments. Fig. S4. Islet dysfunction is transient after cerulein-OGTT. Three months after cerulein or saline injections, OGTT was performed. (A) Blood glucose, (B) plasma insulin, (C) glucagon, (D) pancreatic polypeptide (PP), (E) glucagon-like peptide-1 (GLP-1), (F) and gastric inhibitory polypeptide (GIP) were measured. Results are shown as mean ± SD. Two-tailed student’s t-test is performed for AUC and ANOVA test for the progress curves during 0–120 min. Fig. S5. Islet dysfunction is transient after cerulein-MMTT. Three months after cerulein or saline injections, MMTT was performed. (A) Blood glucose, (B) plasma insulin, (C) glucagon, (D) pancreatic polypeptide (PP), (E) glucagon-like peptide-1 (GLP-1), and (F) gastric inhibitory polypeptide (GIP) were measured. Results are shown as mean ± SD. Two-tailed student’s t-test is performed for AUC and ANOVA test for the progress curves during 0–120 min.