Cholecystokinin Attenuates β-Cell Apoptosis in both Mouse and Human Islets

Abstract

Loss of functional pancreatic β-cell mass and increased β-cell apoptosis are fundamental to the pathophysiology of type 1 and type 2 diabetes. Pancreatic islet transplantation has the potential to cure type 1 diabetes but is often ineffective due to the death of the islet graft within the first few years after transplant. Therapeutic strategies to directly target pancreatic β-cell survival are needed to prevent and treat diabetes and to improve islet transplant outcomes. Reducing β-cell apoptosis is also a therapeutic strategy for type 2 diabetes. Cholecystokinin (CCK) is a peptide hormone typically produced in the gut after food intake, with positive effects on obesity and glucose metabolism in mouse models and human subjects. We have previously shown that pancreatic islets also produce CCK. The production of CCK within the islet promotes β-cell survival in rodent models of diabetes and aging. We demonstrate a direct effect of CCK to reduce cytokine-mediated apoptosis in a β-cell line and in isolated mouse islets in a receptor-dependent manner. However, whether CCK can protect human β-cells was previously unknown. Here, we report that CCK can also reduce cytokine-mediated apoptosis in isolated human islets and CCK treatment in vivo decreases β-cell apoptosis in human islets transplanted into the kidney capsule of diabetic NOD/SCID mice. Collectively, these data identify CCK as a novel therapy that can directly promote β-cell survival in human islets and has therapeutic potential to preserve β-cell mass in diabetes and as an adjunct therapy after transplant.

INTRODUCTION

Diabetes mellitus (DM) is a multifactorial disease that results from impairment of β-cell function and loss of β-cell mass. Type 1 DM (T1D) develops from autoimmune destruction of the pancreatic β-cells, resulting in insulin dependence ¹. Type 2 DM (T2D) results from both increased insulin resistance and decreased insulin production². Failure of adaptive pancreatic β-cell proliferation and increased apoptosis under diabetic stress leads to β-cell mass reduction in T2D ^2,3.There has been an alarming rate of growth in the prevalence of both types of diabetes during the past three decades, consequently driving increases in diabetes-related complications and mortality ⁴. In 2020, the American Diabetes Association reported that 34.2 million Americans are diagnosed with diabetes, while 88 million more Americans have prediabetes ⁵. Yet, no treatment strategies are specifically designed to protect against a decrease in pancreatic islet mass and prevent the loss of insulin-producing cells ^6,7. Many patients with T2D eventually require insulin therapy due to a decline in functional β-cell mass, yet exogenous insulin delivery cannot replicate the dynamic insulin production of a functional islet and dysglycemia often persists.

One possible curative therapy for diabetes is the transplantation of purified human pancreatic islets. The first human islet cell transplantation (ICT) took place in Minneapolis in 1974 ⁸ and with the landmark clinical trial of the Edmonton protocol in 1999, allogenic-ICT became a potential solution to restoring glucose homeostasis for type 1 diabetes (T1D), while auto-ICT became an option of preserving endocrine function for patients requiring total pancreatectomy for chronic pancreatitis ⁹. Considerable improvements in transplant methods were made over the past three decades. However, the major problem with any ICT is the significant loss of islet viability both early and late after transplant ¹⁰. Allogenic ICT only results in 25-50% insulin independence after five years. New therapeutic approaches to protect and preserve β-cell mass in transplanted islets are necessary to improve outcomes.

Cholecystokinin (CCK) is a peptide hormone produced in the small intestine and brain, stimulating the release of digestive enzymes from the exocrine pancreas and inducing satiety ¹¹. A large body of literature from different fields, including neuroscience, cardiology, immunology, and endocrinology, documents that CCK has anti-inflammatory ^12-14, proliferative ^15-17, and anti-apoptotic ^18,19 effects against various insults in a wide range of pathophysiologic models. Combined, there is accumulating evidence that CCK receptor signaling can attenuate the onset and progression of several modes of cell death in various mammalian cell types and different organs. CCK receptor agonists were once extensively studied as a possible anti-obesity agents due to their potential modulation of leptin signaling ^20-22 and their role in central nervous system-mediated appetite suppression ^23,24. In the context of diabetes, systemic injection of CCK analogs in humans and rodents ^22,25-29 has been shown to improve glucose tolerance. While this has previously been thought to be due to an ability for CCK to stimulate insulin secretion, we have recently demonstrated that CCK does not directly enhance glucose-stimulated insulin secretion in mouse or human islets³⁰. In addition to its role as a circulating endocrine hormone secreted from the neuroendocrine cells of the intestinal epithelium, CCK production is highly upregulated in mouse pancreatic β-cells under conditions of obesity and insulin resistance and is also produced and secreted from human islets ^31-35. Mice with β-cell-specific CCK overexpression are protected against β-cell death after injection of the β-cell toxin streptozotocin and have preserved β-cell mass with aging ³². Conversely, CCK null obese mice have decreased islet mass and increased β-cell apoptosis ³¹. In mice and murine Min6 cells ³⁶, as well as in 1.1B4 human β-cell lines ³⁷, exogenous cholecystokinin treatment had been demonstrated to have anti-apoptotic effects.

Despite the accumulated observations of a pro-survival effect of CCK in various cell types and its overall potential benefits in weight regulation and glucose homeostasis the ability of CCK to directly protect human pancreatic β-cells has not been previously tested. Here we show that CCK treatment results in protection against β-cell apoptosis in various models ranging from in vitro insulinoma cell lines and ex vivo intact mouse and human islet. We demonstrate that this is a direct effect of CCK on its receptors in the islet. Additionally, we show that CCK-8 can reduce β-cell apoptosis in human islet grafts following transplant. The translational impact of this study lies in the identification of CCK-based therapeutics as a viable target for preventing β-cell apoptosis in humans. This has relevance for T1D and T2D, including in the setting of transplant or β-cell replacement therapy.

MATERIALS AND METHODS

Viability Assay

INS1E rat insulinoma cells were cultured in RMPI 1640 (Thermo Fisher Scientific, #11875093) with 2.05 mM Glutamax (Cellgro, 35050079) supplemented with 5% (v/v) heat-inactivated FBS, 1% P/S, 10 mM HEPES, 1 mM sodium pyruvate, and 50 μM freshly added beta-mercaptoethanol (Sigma, M7522) at 37 °C and 5% CO2. INS1E cells were seeded in 12 well plates at a density of 0.1 x 10⁶ cells/well and were incubated for 24 hours before any treatment. Cells were then incubated in fresh media containing either 100 nM CCK or 100 nM saline vehicle control up to 72 hours with concurrent treatment with a mouse cytokine cocktail containing 50-ng/mL TNFα (Miltenyi Biotec, #139-101-687), 10-ng/mL IL-1β (Miltenyi Biotec, #130-101-680), and 50 ng/mL IFN-γ (Miltenyi Biotec, #130-105-785) as described ³⁸. Cell media containing the treatment was refreshed at 36 hours of incubation for the time course study groups that were treated for more than 24 hours. At each time point, the growth media containing the suspended cells was collected and centrifuged and combined with the adherent cells that were released with 0.25% trypsin (Sigma, 59428C) and then resuspended in 100 ul of growth media containing FBS. 10 ul of suspended cells were then added to a 10ul solution of 0.4% trypan blue in a buffered isotonic salt solution (Bio-Rad, #1450021), pH 7.3, and measured for viability using TC10 automated cell counter (Bio-Rad, #145-0010).

Apoptosis Assay – Imaging Flow cytometry

Apoptosis was measured in INS1E cells and islets from wild-type (WT) and CCK receptor KO mice treated with cytokine cocktail and CCK for 24 hours. Cultured islets or cells were transferred to a 15 mL conical tube with the incubated media. The culture plate was rinsed with 2 mL PBS and this was added to the tube to ensure complete transfer. Islets and cells were pelleted at 800xg for 3 min at 4°C followed by 1 minute wash using cold PBS. Islets and cells were resuspended in 1 ml Sigma Dissociation Solution (Sigma), then shaken horizontally in a 37°C water bath for 8 min at 100 RPM followed by placement of the tube on ice and addition of 2 mL of media to stop the dissociation process. Cells were then gently disrupted and incubated on ice for another 5 minutes before pelleted again at low speed for 5 min at 4°C. Supernatant was removed and the pellet was resuspended in 1x Annexin Binding Buffer containing 0.5% BSA. 5x Annexin binding buffer was made with 50 mM HEPES, 700 mM NaCl, 12.5 mM CaCl2, in water filled to 50 ml. The cells were then pipetted up and down for 30x, then 1 μl of this cell suspension was placed on a microscope slide to check for dissociation. Depending on the level of clumping, the cells were pipetted up and down for another 20-30x more times or until the majority of the cells were single cells. Cells were then counted on a hemocytometer and were then resuspended in 1x Annexin Binding Buffer containing 0.5% BSA to a final concentration of 1x10⁶ cells/ml. An aliquot of 100 μl of cells (1x10⁵ cells) was mixed with Propidium Iodide (50 μg/mL) and Annexin V-AF488 (Invitrogen #A13201). Unstained and individually stained control samples were prepared in separate tubes. All tubes were incubated at room temp in the dark for 15 min. Samples were diluted immediately before reading with the addition of 400 ul 1x Annexin Binding Buffer containing 0.5% BSA. Imaging Flow data was acquired by imaging flow cytometer (Amnis EMD Millipore, Image Stream Mark II). During the analysis, 10,000 live cell images were captured at 40x. Data were analyzed using ImageStream (Image Stream®X, Amaris) for the quantification of propidium iodide and annexin V stains that indicate the different stages of apoptosis.

Western blot

INS-1E cells cultured as described above were incubated in fresh media containing 100 nM CCK (sulfated-(pGlu-Gln)-CCK-8, American Peptide Company) for 5, 15, 30, 45, and 60 minutes. Protein from the cells was harvested by lysing cells in cold RIPA buffer (ThermoFisher, 89900) containing 1% Nonidet P-40 (Sigma, 21-3277) and a protease inhibitor cocktail (Roche, #04693116001) and phosphatase inhibitor cocktail (Roche, 4906845001). The lysates were centrifuged in QIAshredder columns, and total protein was quantified via BCA assay (Sigma, BCA1). 20 μg of total protein homogenate was loaded per lane onto a 10% polyacrylamide gel (Biorad, 4561036) under constant voltage of 140v for 1 hour using a vertical electrophoresis cell (Biorad, 1658025FC) and were dry transferred onto a 0.2 μm PVDF membrane (Biorad, 1704156) using a Trans-Blot Turbo dry transfer system (Biorad, 1704150). Membranes were then blocked with 5% milk/TBST for 1hr , and incubated overnight with primary antibodies 1:2,000 Rabbit Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) Antibody (CST, 9101), 1:2,000 Rabbit p44/42 MAPK (Erk1/2) (137F5), 1:10,000 mouse GAPDH (Cell Signaling Technologies, 97166) in 1% BSA/TBST solution. Membranes were then washed and incubated with 1:10,000 of Goat Anti-Rabbit IgG H&L (HRP) (Abcam, ab205718), or Goat Anti-Mouse IgG H&L (HRP) (Abcam, ab205719) for 1 hour at room temperature before imaging using chemiluminescent substrate (Thermo, 34579) in an image analyzer (GE Healthcare, Imagequant LAS 4000). Band densitometry was quantified using Image Studio (Li-Cor. Ver.5.2).

Animals, Islet Isolation and Culture

Animal care and experimental procedures were performed with approval from the University of Wisconsin Animal Care and Use Committee to meet acceptable standards of humane animal care. Male C57BL/6J (~10 to 15-week-old) (JAX Stock Number #000664) and NOD/SCID (9-week-old) mice (JAX Stock Number #001303) were purchased from The Jackson Laboratory for islet isolation and transplant studies, respectively. CCK receptor double knockout mice, 129-Cckar^tm1Kpn Cckbr^tm1Kpn/J (JAX Stock Number #006365) ^39,40 age 14-20 weeks were used in the studies. These mice were obtained from Jackson Laboratories after cryorecovery. Mice were housed in facilities with a standard light-dark cycle and fed ad libitum. Mouse pancreatic islets were isolated using collagenase digestion and hand-picked as previously described ⁴¹. Isolated islets were cultured at 37°C and 5% CO₂ in RPMI 1640 media (Thermo Fisher Scientific, #11879020) containing 8 mM glucose, supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 units/ml penicillin and 100 μg/ml streptomycin (1% P/S) (Thermo Fisher Scientific).

Human islet culture and mRNA Levels

Human islets were obtained through the Integrated Islet Distribution Program (IIDP). Upon arrival, islets were handpicked and then cultured in RPMI 1640 media as described above but without antibiotics and antimycotics. Islets were cultured overnight to confirm viability and sterility before treatments. Islets were cultured up to 7 days, and the media was renewed every other day. After culture, RNA was isolated, and gene expression was quantified via quantitative PCR with SYBR and primers for human CCK (hsCCK forward TGA GGG TAT CGC AGA GAA CGG ATG, hsCCK reverse TGT AGT CCC GGT CAC TTA TCC TGT), CCKAR (hsCCKAR forward TGG AAG CAA CAT CAC TCC TC, hsCCKAR reverse CAC GCT GAG CAG GAA TAT CA), and CCKBR (hsCCKBR forward GAT GTG GTT GAC AGC CTT CT, hsCCKBR reverse GGG CTG ATC CAA GCA GAA A) normalized to Beta-Actin (forward TCA AGA TCA TTG CTC CTG AGC, reverse TCA AGA TCA TTG CTC CTG AGC). Human VEGFA ⁴² (hs VEGFA forward TTG CCT TGC TGC TCT ACC TCC A, hs VEGFA reverse GAT GGC AGT AGC TGC GCT GAT A) were also normalized to Beta-Actin.

Ex vivo Dispersed Islet TUNEL and Caspase 3/7 Activity Assays

Using serum-free RPMI 1640 media supplemented with 5 g/l BSA fraction V (Roche, #107351080001), islets were pretreated with 100 nM CCK or saline vehicle control for 24 hours, before additional 24 hours of cytokine or 10 μM thapsigargin treatment to induce apoptosis. Human islets were treated with a cytokine cocktail containing human 1,000 U/ml TNFα, 75 U/ml IL-1β, and 750 U/ml IFN-γ (PeproTech, #200-01B, #300-02, #300-01A). After treatments, islets were dispersed at 37°C using 0.25% Trypsin-EDTA (Thermo Fisher #25200056), plated on Poly-L-lysine pre-coated glass coverslips and fixed with 10% formalin, β-cell apoptosis was measured using DeadEnd Fluorometric terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL) (Promega, #G3250) followed by insulin staining (Dako, A0564). Imaging was performed using an EVOS FL Autofluorescence microscope (Life Technologies). ImageJ (NIH) or Adobe Photoshop (Adobe) was used to count DAPI (Vector Labs, H-1200-10), insulin, and TUNEL in at least 9 randomly chosen fields per treatment group for each replicate. For Figure 2A-D, a minimum of 1,000 β-cells and 100 non-β-cells were counted per biological preparation made from 50 islets. For Figure 3B-C, a minimum of 2,500 β-cells and 300 non-β-cells were counted per biological preparation made from 50 islets.For caspase assay, dispersed human islets were plated in a 96 well plate at 20,000 cells per well and incubated for 24 hours prior to treatment. For Figure 3C, dispersed islets were pretreated with 100 nM CCK or vehicle control for 1-hour prior to treatment with human cytokines as above. CCK or vehicle treatment was continued with or without cytokines for 24 hours. Caspase 3/7 activity was measured 24 hours after cytokine treatment using the Caspase-Glo® 3/7 Assay System (Promega, #G8090).

Fig 2.

CCK protects mouse islets from apoptosis. The percentage of TUNEL-positive insulin-positive cells of mouse islets treated with cytokine cocktail (A) or thapsigargin (C) is reduced when treated with 100 nM CCK (n = 3-4). The percentage of TUNEL-positive insulin-negative mouse islet cells is unchanged after cytokine cocktail (B) and reduced after thapsigargin (D) by CCK treatment (n = 3-4). (E) Percentage of Annexin V and propidium iodide (PI) double-positive cells (Late apoptosis – black bars) and Annexin V positive mouse islet cells (Early Apoptosis – gray bars) co-treated with cytokines and CCK is reduced in wild type controls but not in double CCK receptor KO islets (ABKO) (n = 6) Data are means ± SEM; *P<0.05, **P < 0.01; n.s, non-significant.

Fig 3.

CCK protects human islets from cytokine-induced apoptosis. Human islets express CCK, CCKAR, and CCKBR mRNA at variable levels (n = 25) (A). CCK treatment reduces the TUNEL-positive/insulin-positive cells (B) but does not significantly impact TUNEL positivity in non-insulin-positive islet cells (C). (D) Caspase 3/7 activity of human islets treated with cytokine cocktail is reduced when co-treated with CCK. CCK treatment alone has no effect on caspase activity. Data are means ± SEM (n = 5); *P < 0.05.

Human Islet Transplantation and Graft Harvest

A modified version of human islet transplantation described by Montanya et al. was used for the study ⁴³. A single high dose (200 mg/Kg) of streptozotocin (Sigma, #S0130) was administered via intraperitoneal injection two days before transplant to induce hyperglycemia. Hyperglycemia (glucose >300 mg/dL) was confirmed with tail vein blood sample using a glucometer prior to proceeding with transplant. Human islets were cultured in media supplemented with 100 nM CCK (sulfated-(pGlu-Gln)-CCK-8, American Peptide Company) or saline vehicle control for 24 hours before transplant. Confirmed hyperglycemic mice were anesthetized using isoflurane, and approximately 1,000 islet equivalents were placed under the kidney capsule. While our initial study design was to have a paired sample of both CCK-treated and saline-treated from each human islet donor, in some cases we had death of a mouse during or post-surgery or technical difficulties during the transplant that led to insufficient islet delivery in one animal. Therefore, the final data do not always represent paired treatment groups. An infusion pump (Azlet 1004) containing saline or CCK was placed subcutaneously in the back of the mice at the same time as the islet transplant to have continuous CCK infusion (52 pmol/hr) for 3 weeks. We calculated the concentration of infused CCK to recapitulate the theoretical amount of CCK delivered through twice-daily intraperitoneal injection dosing done in studies by Irwin et al. that led to beneficial effects on glucose and weight in obese mice ^25,26. Mice were monitored during recovery and checked for surgical complications in the post-operative period. Bodyweight and randomly fed blood glucose were measured every 2-3 days following the transplant for a total of three weeks. Blood glucose was measured using a tail nick blood sample and glucometer (Bayer Contour Next EZ). Three weeks post-transplant mice were anesthetized using Avertin (2,2,2-Tribromoethanol, 97% T48402, 500mg/kg) and terminal serum was collected by cardiac puncture and stored for further assays. Kidneys containing islet grafts were harvested and fixed in 10% formalin (Fisher SF100) for 48 hours. 10% Formalin-fixed kidneys were paraffin-embedded and sectioned for immunofluorescence staining.

CCK and Human Insulin Measurement

Serum CCK levels were measured three weeks post-transplant using a CCK radioimmunoassay (Alpco Diagnostics - now discontinued) as described by Rehfeld ⁴⁴. Serum samples were collected through cardiac puncture and plasma extracted by mixing with 96% ethanol and evaporating to dryness using a vacuum centrifuge. The dry extracts were dissolved in Alpco diluent (reagent D) and stored at −20 °C until assayed. The procedure provided by Alpco was followed for the radioimmunoassay. Briefly, samples were mixed with anti-CCK-8 (reagent A) and incubated for 2 days at 2-8 °C followed by the addition of I¹²⁵-CCK-8 (reagent B) and another 4-day incubation at 2-8 °C. Finally, the double antibody solid phase (reagent C) was added, incubated 60 minutes at 2-8 °C, centrifuged, and the supernatant discarded. A gamma counter was used to measure radioactivity with a counting time of 2-4 minutes. Insulin secreted from the islet grafts was measured in the mouse serum three weeks post-transplant using a specific human insulin ELISA (Millipore, EZHI-14K).

Immunofluorescence Staining

Paraffin-embedded islet grafts within the kidneys or intact mouse or human islets were stained for insulin using polyclonal guinea pig anti-insulin antibody (Dako, A0564), and apoptosis was measured using the TUNEL system (Promega, #G3250). Images from each islet graft were obtained using an EVOS FL Autofluorescence microscope (Life Technologies). ImageJ software was used to quantify total insulin-positive cells and TUNEL-insulin co-staining cells. Image quantification was done blinded to experimental group. A total of 7,471 β-cells were counted from control mouse grafts and 8,837 β-cells were counted from CCK treated mouse grafts.

Statistics

Assessment of statistical significance between groups was determined using GraphPad Prism by 2-tailed Student’s t or ANOVA tests as the non-parametric equivalent. Bonferroni posthoc test was performed to correct for multiple comparisons where appropriate. A paired t-test was used when samples were from the same islet prep to account for differences in baseline viability (Figures 2A-C, 3B-C). A probability of error less than 5% was considered significant (i.e., P < 0.05). Statistical information for experiments (data representation, P values, and n numbers) can be found in the figure legends. In all panels, data are represented as mean ± SEM.

RESULTS

CCK Treatment Protects INS1E Cells Against Cytokine-Induced Apoptosis.

We have previously shown that the localized production of CCK in pancreatic β-cells leads to improved β-cell survival in mouse models of obesity, aging, and diabetes ^31-33. However, in these experiments, locally produced intra-islet CCK could indirectly affect other cell types to mediate its pro-survival effects. To look more specifically at the direct pro-survival effects of CCK peptide treatment on the β-cell in vitro, we first turned to the rat insulinoma cell line, INS1E. In all experiments described, we used a stable analog of sulfated CCK-8, s-Glu-Gln-CCK-8, that has high biologic activity in vitro and in vivo ^25,26,45. We will refer to this peptide simply as CCK or CCK-8 throughout the manuscript. A time course of INS1E cell viability over 72 hours of exposure to mouse pro-inflammatory cytokine cocktail (10 ng/μl IL-1β, 50 ng/μl IFN-γ, and 50 ng/μl TNFα) demonstrates that CCK-8 (100 nM) treatment mitigates cytokine-induced cell death, as measured by trypan blue exclusion, (n=7, p <0.05) up to 48 hours (Fig.1A). The viability of INS1E cells treated with CCK-8 was superior to those treated with GLP-1 (100nM), another peptide hormone with known pro-survival effects in β-cells in vitro and a widely used therapeutic target for type 2 diabetes (Fig.1A). We further demonstrate that CCK-8 also protected INS1E cells from cytokine-induced apoptosis in a concentration-dependent manner, using image flow cytometry with annexin V and propidium iodide staining to identify cells in various stages of apoptosis (Fig.1B, n=3-6, p <0.001). Both CCK receptors can activate G_q proteins that can signal through various pathways to activate mitogen-activated protein kinases (MAPK), including the extracellular-regulated kinases (ERK 1/2)^46,47. ERK 1/2 signaling is well known to reduce apoptosis in response to multiple stressors through both inhibition of apoptosis mediators and activation of pro-survival factors⁴⁸. 100 nM of CCK-8 leads to robust phosphorylation of ERK1/2 in INS-1E cells (Fig1.C), suggesting that the protective effect seen in Fig.1A&B involves ERK 1/2 signaling.

Fig 1.

CCK protects INS1E cells from cytokine-induced apoptosis. (A) Time-course of INS1E cells treated with cytokine cocktail and co-treated with 100 nM CCK or 100 nM GLP-1 (n = 6-19). (B) GLP-1 and CCK dose-response effects on the percentage of Annexin V and propidium iodide (PI) double-positive cells (Late apoptosis – black bars) and Annexin V positive cells (Early Apoptosis - gray bars) in INS-1E cells treated with cytokines (n = 3-6). (C) ERK 1/2 phosphorylation induced by 100 nM of CCK-8 in INS1E cells as a function of time (n = 4). Western blot images are representative from single replicate. Data are means ± SEM; *P < 0.05, **P < 0.01, ***P < 0.001, ***P < 0.0001.

CCK Treatment Protects Mouse Pancreatic β-Cells from Apoptosis.

To determine whether CCK also directly affects survival in primary mouse β-cells, we measured β-cell apoptosis of primary mouse islet cells by TUNEL assay. Intact mouse islets were preincubated with CCK-8 or vehicle for 24 hours. Mouse islets were then co-incubated with CCK-8 or vehicle and mouse cytokine cocktail (10 ng/μl IL-1β, 50 ng/μl IFN-γ, and 50 ng/μl TNFα) for an additional 24 hours. Mouse islets treated with CCK-8 (100 nM) had a 27% reduction in β-cell apoptosis after cytokine treatment compared to vehicle control, as measured by TUNEL staining (3.18% ± 0.27 CCK vs. 4.38% ± 0.29 vehicle vs. 0.36% ± 0.15 non-treated) (Fig.2A) while no protection was observed in non-β-cell islet cells (Fig.2B). In addition to the ability to protect from cytokine-mediated apoptosis, which is due to a combination of cell stressors, we wanted to test whether CCK also reduced β-cell apoptosis specifically due to endoplasmic reticulum (ER) stress. ER stress is elevated in β-cells in models of both type 1 and type 2 diabetes ^49,50. Mouse islets that were treated with CCK-8 (100 nM, including pre-treatment for 24 hours) and thapsigargin (10 μM) to induce ER stress have significantly fewer TUNEL positive cells than cells treated with thapsigargin alone (Fig.2C&D). This effect was notably observed in both β-cells and non-β-cells. Finally, to demonstrate that the reduction in apoptosis was a direct, CCK-receptor mediated effect, we tested whether CCK-8 could still protect from apoptosis in islets from mice with knockout of both the CCKA and CCKB receptors (ABKO). Using propidium iodide and annexin V staining via imaging flow cytometry, we demonstrate that the protective effect of CCK-8 (100 nM) seen in the WT islets disappears in islets from ABKOs, indicating that protection of mouse islets against cytokines by CCK is CCK receptor-dependent (Fig.2E). Taken together, these findings suggest that activation of CCK receptors protects rodent β-cells from apoptosis due to multiple stressors directly through the CCKAR, CCKBR, or a combination of both receptors.

CCK Protects Human Pancreatic β-cells from Apoptosis In Vitro.

Human islets express CCK and both CCK receptors, although at highly variable levels (Fig.3A) ^30,33 suggesting that they may also be amenable to protection from CCK receptor-mediated treatments. Intact human islets from deceased organ donors were pretreated with 100 nM CCK-8 or vehicle for 24 hours and then treated with a pro-inflammatory human cytokine cocktail. Human islet donor information in Appendix Table 1. β-cells from dispersed human islets treated with CCK-8 before cytokine cocktail exposure had significantly reduced TUNEL staining in comparison to those treated with vehicle before cytokines (1.81-fold ± 0.21, p<0.05) (Fig.3B). Non-β-cells did not show significant changes in TUNEL staining with CCK treatment (Fig.3C). CCK-8 treatment also decreased apoptosis in dispersed human islet cells, as measured by cytokine-induced caspase 3/7 activity (Fig. 3D). CCK treatment by itself, in the absence of cytokines, did not impact caspase activity (Fig 3D). Together, these results demonstrate the efficacy of CCK treatment in promoting β-cell survival in human islets.

Systemic CCK-8 Does Not Alter Body Weight or Blood Glucose Levels in Mouse.

To provide evidence that CCK treatment could protect human islets in vivo, we transplanted a sub-therapeutic number of human islets (Appendix Table 1) under the kidney capsule of immunodeficient (NOD/SCID) mice with streptozotocin (STZ)-induced diabetes. We treated these mice with CCK or saline via an osmotic pump for three weeks. Human islets were pretreated with CCK or vehicle control for 24 hours before transplant. In our xenograft model, we transplanted less than 1,000 islet equivalents (IEQs) of human islets per mouse, which we predicted would be insufficient to restore euglycemia. Our goal was to keep the islets in a metabolically unfavorable environment of persistent hyperglycemia, similar to that seen in human diabetes. The amount of CCK delivered by an osmotic pump was calculated to approximate the daily dose previously administered to obese mice by twice-daily injection²⁶. We confirmed that the animals that received the CCK treatment had increased levels of circulating serum CCK and achieved absolute serum concentrations of approximately 20 pM, which is only modestly higher than the physiologic postprandial levels of CCK in humans^51,52 (Fig.4A). However, blood glucose levels (Fig.4B) and body weight (Fig.4D) of the human islet transplant recipient animals receiving CCK-8 infusion were unaltered by circulating CCK. Finally, circulating human insulin was measured and was not different in the two treatment groups (Fig.4C). Since the systemic circulation of CCK-8 did not alter the body weight or blood glucose levels in comparison to the control group, any changes in the viability or apoptotic events of transplanted human islets would not have been the result of differences in insulin demand or glucotoxicity among different transplant groups, but most likely a direct effect of CCK-8 on the human islet cells.

Fig 4.

Treatment with CCK does not alter the body weight or blood glucose of diabetic mice. (A) Circulating levels of CCK are increased in CCK-treated mice 21 days after transplant. (B) Non-fasted blood glucose levels measured weekly during 21 days post-transplant do not differ between treatment groups. (C) Human insulin is detected in the serum of mice 21 days after transplant and does not differ between control and CCK-treated mice. (D) Bodyweight was measured weekly and used to calculate percent change during 21 days post-transplant, no significant differences were detected. Data are means ± SEM (n = 4-6). ND, not detectable; n.s, non-significant.

CCK Protects Human Pancreatic β-Cells from Apoptosis Following Transplant.

We measured β-cell apoptosis in the islet grafts three weeks after transplantation. We found reduced TUNEL-positive β-cells in transplanted human islet grafts from mice with CCK treatment (0.35%) in comparison to islet grafts from saline-treated mice (1.53%). This suggests that exogenous CCK treatment protects human β-cells from apoptosis (Fig.5A&B). We did not directly examine vasculature in the islet grafts, but considered that an impact of CCK on revascularization of the transplanted islets may explain our observation of reduced apoptosis. Vascular endothelial growth factor (VEGF)-A is known to be produced by pancreatic islets and is necessary for vascularization of islet grafts⁵³. However, we see no evidence that CCK directly increases VEGF-A expression in human islets (Fig 5C), suggesting that this is less likely to be the mechanism of improved survival post-transplant and further supporting a direct effect of CCK on intrinsic β-cell survival. Taken together, our study indicates that in vivo, CCK protects human β-cells from apoptosis under diabetic conditions and in the transplant setting.

Fig 5.

Human islet grafts in mice receiving CCK treatment contain fewer apoptotic β-cells. (A) Quantitative analysis reveals a reduced percentage of TUNEL-positive insulin-positive cells in human islet grafts from CCK-treated mice 21 days after transplant. (B) Representative images of islet grafts from control (left) and CCK-treated (right) transplant mice. DAPI (blue), insulin (red), and TUNEL (green). (C) VEGF-A expression in human islets is not changed by CCK treatment. Data are means ± SEM (n = 4-6); **P < 0.01. n.s. = not significant.

DISCUSSION

Pancreatic islet transplantation represents a potential cure for diabetes. The fulfillment of this goal is ultimately contingent on achieving the long-term survival of the islet graft. Allogenic-ICT can be a curative therapy for T1D, while auto-ICT can preserve the endocrine function of patients undergoing pancreatectomy. However, the major problem with any islet transplantation, particularly for T1D, is a significant loss of islet viability both early and late after the transplant¹⁰. In T1D patients, only ~50% of successful islet transplant recipients maintain insulin independence after two years. The loss of islets in the early stages after a transplant requires a high number of islets to restore glucose homeostasis in the patient, often requiring more than one donor per recipient. Inflammation and autoimmunity are causes for the initiation of β-cell destruction during the development of T1D. Transplanted pancreatic islets themselves can also release pro-inflammatory cytokines⁵⁴, which play a significant role in cell failure and death, during and post-islet transplantation. Thus, identifying factors that can help improve the preservation of pancreatic β-cell mass is in demand to improve islet transplant outcomes. In rodents, administration of the stable CCK analog (pGlu-Gln-CCK-8) protects mice from obesity-induced diabetes in both high fat-fed and leptin-deficient models ^25,26. Notably, a study by Irwin et al.²⁶ shows that twice-daily injection of pGlu-Gln-CCK-8 (25 nmol/kg body weight) in high fat-fed mice and ob/ob mice reduced body weight, improved glucose tolerance, improved insulin sensitivity and lowered non-fasting glucose, demonstrating the potential of a CCK receptor agonist as an anti-obesity/anti-diabetic agent. However, this study has inherent limitations in isolating the direct beneficial effect of CCK receptor agonism on β-cell survival. Since CCK-treated animals had improved glucose tolerance and insulin sensitivity, this can indirectly reduce the diabetogenic stress in the islets. Additionally, the high-fat diet feeding in their model did not induce the typical compensatory increase in β-cell mass. While these authors reported increased β-cell turnover with CCK treatment, a direct effect of CCK on β-cell survival could not be assessed in this model. In our current study, we demonstrate direct ex vivo effects of CCK to reduce β-cell apoptosis and also show reduced apoptosis of human β-cells in vivo despite persistent hyperglycemia that is comparable to controls.

We did not measure proliferation of β-cells in any of the studies in this paper. We have previously shown that there is no β-cell proliferative response to CCK^32,55 and therefore this was not the focus of the current study. In the future, examination of β-cell proliferation will be an important consideration in long-term treatment studies of human islets with CCK receptor agonists.

This current study demonstrates the relevance of islet CCK signaling in humans by showing that human pancreatic islets express CCK, CCKAR, and CCKBR (Fig.3A), albeit at highly variable expression levels. This is distinct from mouse islets, which primarily express CCKAR with very low CCKBR expression ^30,32. Therefore, this difference in CCK receptor expression is important in assessing species-specific effects of CCK-based therapies on the islet. CCK-8 peptide can activate both CCKAR and CCKBR. We demonstrate that the CCK-8 analog directly protects human islets from cytokine-induced apoptosis (Fig.3B-D). While these in vitro studies were encouraging, it was essential to study this effect under in vivo settings. We find that continuous CCK-8 infusion for three weeks successfully reduced human β-cell apoptosis by over four-fold compared to the vehicle-treated group following the transplant (Fig.5). Not only does this support the promise of CCK-based therapy to promote β-cell survival in human islets, but it also demonstrates that CCK is effective in protecting from the pathophysiologic stressors relevant in diabetes and transplantation such as hyperglycemia, hypoxia, and ER and oxidative stress. We do not see a direct effect of CCK treatment on VEGF-A expression (Figure 5C), but additional studies will be needed to determine if CCK improves islet graft vascularization as a potential mechanism.

While our ex vivo experiments with cytokine treatment do not directly replicate the stressors experienced by islets transplanted into immunodeficient mice, both experiments induce multiple different types of cellular stress and therefore may have overlapping intracellular stress responses. Although they lack functional T-cells and B-cells, NOD-SCID mice still have macrophages and neutrophils and may still exhibit cytokine responses to injury or infection ⁵⁶. Additionally, local production of pro-inflammatory cytokines from the islet cells themselves may occur post-transplant or in response to hyperglycemia ⁵⁷. Therefore, while the exact mechanism of β-cell death in transplanted islets is complex, we and others have shown that CCK is protective against many of these pathways, including hyperglycemia, ER stress, and cytokine exposure^32,36. Additional studies will be needed to determine whether CCK has any impact on β-cell survival in direct response to hypoxia, which likely has shared pathways of apoptosis that include the unfolded protein response and CHOP activation⁵⁸.

The results we present from the transplant study are encouraging, but there are some intrinsic limitations. We acknowledge that, due to study design, including pre-treatment of islets with drug, having a single drug dose, and a single late endpoint for islet graft harvesting, it is impossible to determine whether the protection from apoptosis we see in our transplant model is due to the CCK treatment of the islets before the transplant, the systemic CCK infusion, or a combination of the two. We also do not know if the protective effects of CCK treatment are most pronounced in the early post-transplant period or later in response to persistent hyperglycemic stress. To fully harness the protective effects of CCK and reduce unwanted side effects, it will be essential to determine the appropriate timing and dose of CCK treatment necessary to provide adequate β-cell protection.

It is also important to note that in this study, we show that CCK-mediated protection is occurring in a pharmacologically validated mechanism via receptor-mediated effects. We achieve this by first demonstrating a dose-response relationship of the protective effects of CCK in INS1E cells across a pM to μM range (Fig.1B). The wide range of effective concentrations hints that CCK treatment to target β-cell survival has a large therapeutic index. Using CCK at lower doses will likely minimize the gastrointestinal side effects that impeded progress in prior clinical trials. Notably, the protection of human islets in vivo was achieved with circulating levels of CCK only in the 20 pM range. Additionally, we show that this protection disappears in islets from mice devoid of both CCK receptors (Fig.2E), confirming a direct, receptor-mediated effect.

All experiments in the current study were done using a non-biased CCK analog that presumably results in pharmacological activation of both CCK receptors. CCK signals through the G-protein coupled receptors, the CCK-A receptor (CCKAR or CCK1R) or CCK-B receptor (CCKBR or CCK2R), to activate many different signal transduction pathways⁴⁷. Previously, others have demonstrated the direct effect of CCK on the survival of Min6 cells against apoptosis in response to hyperglycemia and suggested that this effect depended on β-arrestin-mediated signaling ³⁶. However, the CCK receptor signaling pathways employed by β-cells are not fully understood. We show in Figure 1C that CCK can activate ERK 1/2 in INS-1 cells, which predominantly express CCKAR³². CCKBR signaling may also lead to ERK 1/2 activation via G_q protein activation of phospholipase C. Since human islet cells express both the CCKA receptor and the CCKB receptors (Figure 3A), future studies will be needed to address specific CCK signaling pathways in the human islet. Identifying the specific CCK receptor subtype activation that affords protection against β-cell apoptosis in human islets in our transplant model will allow further development of CCKR agonists to improve transplant outcomes with minimal adverse effects.

In conclusion, we demonstrate here, for the first time, that CCK protects human β-cells from apoptosis. We provide evidence through in vitro, ex vivo, and in vivo studies using isolated human islets to suggest that this protection occurs through direct effects on the β-cell and in response to several forms of apoptotic stress. CCK has been studied as a potential therapeutic against obesity and diabetes due to its impact outside the islet on body weight and glucose regulation. Our work demonstrates that CCK also has direct positive effects on protecting human β-cells, adding to its benefits for utilization in diabetes therapy. The endogenous function of CCK in human pancreatic islet remains an open question. CCK has a clear role in intra-islet autocrine and paracrine signaling to promote β-cell survival in mouse islet ^32,33,59 and it also has a potential role in exocrine-endocrine crosstalk for islet mass expansion and regulation in diabetic mouse models ³⁵. Future studies will be essential to discover the CCK signaling pathways activated in β-cells and explore how these pathways can be targeted to develop an effective therapy with few adverse effects.

Brief Commentary.

Background:

Diabetes is a result of decreased functional beta-cell mass. Cholecystokinin (CCK) is a peptide hormone that signals through G-protein coupled receptors and we have previously shown that islet-derived CCK can protect from beta-cell apoptosis in mouse models of obesity, diabetes, and aging.

Translational Significance

In this manuscript we show that CCK treatment can not only directly protect beta-cells in mouse islets from apoptotic cell death, but is also able to reduce apoptosis in human islets both ex vivo and in a xenograft transplant model. These findings support CCK as a viable therapeutic target for islet survival in diabetes.