Low Concentration of Interleukin-1β Induces FLICE-Inhibitory Protein–Mediated β-Cell Proliferation in Human Pancreatic Islets

Ethical approval for mouse studies was granted by the Zurich Cantonal Animal Experimentation Committee. C57BL/6j wild-type mice were obtained from The Jackson Laboratory (Bar Harbor, ME). IL-1β knockout (IL-1β^−/−) mice, on a C57BL/6j background, were produced by gene targeting as previously described (37). Animals were housed at 22°C with a 12-h light-dark cycle (lights on at 0700) and allowed free access to water and food.

Mice were fasted 12 h overnight and injected intraperitoneally with 2 mg/g body wt glucose (40% glucose solution; Laboratorium Dr. G. Bichsel AG, Interlaken, Switzerland) or with 0.75 mU/g recombinant human (rh) insulin (Novo Nordisk, Bagsværd, Denmark) for the glucose or insulin tolerance tests, respectively. Blood samples were obtained from tail-tip bleedings, and blood glucose concentration was measured with a Glucometer (Freestyle, Disetronic Medical Systems, Burgdorf, Switzerland).

Human islets were isolated from pancreata of nine organ donors at the University of Geneva Medical Center and at the University of Illinois at Chicago. Mouse islets were isolated as previously described (38). The islets were cultured on extracellular matrix–coated plates derived from bovine corneal endothelial cells (Novamed, Jerusalem, Israel), allowing the cells to attach to the dishes and spread. Human islets were cultured in CMRL 1066 medium containing 5.5 mmol/l glucose and mouse islets in RPMI-1640 medium containing 11.1 mmol/l glucose, both supplemented with 100 units/ml penicillin, 100 μg/ml streptomycin, and 10% FCS (Invitrogen, Carlsbad, CA), hereafter referred to as culture medium. Two days after plating, when most islets were attached and began to flatten, the medium was changed to culture medium containing 5.5, 11.1, or 33.3 mmol/l glucose with or without 0.01–5 ng/ml rhIL-1β, 500 ng/ml rhIL-1Ra (R&D Systems, Minneapolis, MN), 500 ng/ml antagonistic Fas antibody (ZB4; MBL, Nogoya, Japan), or transfected as described below.

RNAs of 21 nucleotides, designed to target human IL-1Ra (5′AUCUGCAGAGGCCUCCGCAtt3′/5′UGCGGAGGCCUCUGCAGAUtt3′), human FLIP_long (silencer predesigned small interfering RNA [siRNA]), and scrambled siRNA were synthesized by Ambion (Austin, TX). siRNA was transfected using SiPortAmine and transfection efficiency estimated with cy3-labeled siRNA using a Silence siRNA Labeling Kit (Ambion), as previously described (39).

For β-cell proliferation studies, a monoclonal antibody against the human (Zymed Laboratories, San Francisco, CA) or mouse (Santa Cruz Biotechnology, Santa Cruz, CA) Ki-67 antigen was used. The free 3′-OH strand breaks resulting from DNA degradation were detected by the terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) technique according to the manufacturer’s instructions (In Situ Cell Death Detection Kit; Boehringer, Mannheim, Germany) and as previously described in detail (14,18). Thereafter, islets were incubated for 30 min at 37°C with a guinea pig anti-insulin antibody (Dako, Carpinteria, CA), followed by detection using the streptavidin-biotin-peroxidase complex (Zymed) or a fluorescein-conjugated rabbit anti–guinea pig antibody (Dako).

The pancreata were weighed, fixed in formalin, and embedded in paraffin. Ten representative sections from each pancreas (spanning the width of the pancreas) were used in the analysis of β-cell mass. Tissue sections were deparaffinized, rehydrated, and incubated with guinea pig anti-insulin antibody (Dako) followed by detection with a fluorescein-conjugated rabbit anti–guinea pig antibody (Dako). Subsequently, the specimens were labeled for glucagon with mouse anti-glucagon antibody (Dako), followed by detection with donkey anti-mouse Cy3-conjugated antibody (Jackson ImmunoResearch Laboratories, West Grove, PA). β-Cell mass was analyzed using Openlab software. The relative area of β-cells (green fluorescence) was determined by quantification of the cross-sectional β-cell area divided by the cross-sectional area of total tissue. The β-cell mass per pancreas was estimated as the product of the relative cross-sectional area of β-cells per total tissue and the weight of the pancreas. For detection of FLIP expression in β-cells, islets were fixed in 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, incubated with rabbit anti-FLIP (R&D Systems), followed by detection with donkey anti-rabbit Cy3-conjugated antibody (Jackson). Subsequently, the specimens were stained for insulin as described above.

Equivalent amounts of protein from each treatment group were run on 15% SDS polyacrylamide gels. Proteins were electrically transferred to nitrocellulose filters and incubated with rabbit anti–caspase 8 (Stressgen, Victoria, BC, Canada), rabbit anti-human FLIP-long (R&D Systems), rabbit anti–pancreatic duodenal homeobox (PDX)-1 (kindly provided by Stefan Zahn; NovoNordisk, Bagsværd, Denmark), or rabbit anti-actin (Cell Signaling, Beverly, MA) antibodies followed by incubation with horseradish peroxidase–linked anti-rabbit IgG peroxidase-conjugated antibodies (Santa Cruz Biotechnology). Immune complexes were detected by chemiluminescence using LumiGLO (Cell Signaling).

Total RNA was extracted from the cultured islets by using the RNeasy Mini Kit (Qiagen, Basel, Switzerland), and RT-PCR was performed by using the SuperScript Double-Stranded cDNA Synthesis Kit according to the manufacturer’s instructions (Life Technologies, Gaithersburg, MD). For quantitative analysis, we used the LightCycler Quantitative PCR System (Roche, Basel, Switzerland) with a commercial kit (LightCycler-DNA Master SYBR Green I; Roche). Mouse primers used were 5′TACGGGGTTTGTGAAAGGAG3′ and 5′CACATCATTCCCCAGGAAAC3′ (insulin), 5′GAGGACCCGTACAGCCTACA3′ and 5′CGTTGTCCCGCTACTACGTT3′ (PDX-1), 5′CTAAATTTGGTTGCCCCAGA3′ and 5′CTCCCATTATGGAGCCTGAA3′ (FLIP long), 5′GCTCTTTTTGCCTGGGAGATC3′ and 5′CCCGAAGGACTCGATTGATAGA3′ (Pax4), 5′TATCAAGGAGGCCCATTTTG3′ and 5′GGTCAGGGTGCAGTTTGTTT (Fas), and 5′GTGGCAGTGATGGCATGGAC3′ and 5′CAGCACCAGTGGATGCAGGG3′ (glyceraldehyde-3-phosphate dehydrogenase). Human primers used were 5′CCACCTTGGGACCTGTTTAG3′ and 5′TGATGCCAGAGGAAGAGGAG3′ (PDX-1), 5′CCACCGGAATCGGACTATCTT3′ and 5′TACTGCCCACGCTGGAACTC3′ (Pax4), 5′GAGCAAGCCCCTAGGAATCT3′ and 5′GCCCTGAGTGAGTCTGATCC3′ (FLIP long), 5′CTACCTAGTGTGCGGGGAAC3′ and 5′GCTGGTAGAGGGAGCAGATG3′ (insulin), 5′GCATCTGGACCCTCCTACCT3′ and 5′CAGTCTGGTTCATCCCCATT3′ (Fas), 5′AGAGTCGCGCTGTAAGAAGC3′ and 5′TGGTCTTGTCACTTGGCATC3′ (α-tubulin), 5′AACAGCGACACCCACTCCTC3′ and 5′GGAGGGGAGATTCAGTGTGGT3′ (glyceraldehyde-3-phosphate dehydrogenase), and 5′TACGGGTCCTGGCATCTTGT3′ and 5′CCATTTGTGTTGGGTCCAGC3′ (cyclophilin).

For acute insulin release in response to glucose, islets were washed and preincubated (30 min) in Krebs-Ringer bicarbonate buffer (KRBB) containing 2.8 mmol/l glucose and 0.5% BSA. KRBB was then replaced by KRBB 2.8 mmol/l glucose for 1 h (basal), followed by an additional 1 h in KRBB 16.7 mmol/l glucose. Islets were extracted with 0.18 N HCl in 70% ethanol for determination of insulin content. To determine the total insulin content of the pancreas, the tissue was homogenized in 1 ml 0.18 N HCl in 70% ethanol and left overnight at 4°C. Insulin was determined using a human insulin radioimmunoassay kit (CIS Bio International, Gif-Sur-Yvette, France), which has similar affinity for both mouse and human insulin. Serum insulin was measured using Luminex technology according to the manufacturer’s instructions (Linco Research, St. Charles, MO). IL-1Ra release in the islet supernatant was measured by using human anti–IL-1Ra enzyme-linked immunosorbent assay kits (R&D).

Samples were evaluated in a randomized manner by a single investigator (K.M.) who was blinded to the treatment conditions. Data are presented as means ± SE and were analyzed by Student’s t test or by ANOVA with a Bonferroni correction for multiple-group comparisons.

Human islets were cultured on extracellular matrix–coated plates in the presence of increasing IL1-β concentrations for 4 days. Exposure to 0.02 ng/ml IL-1β induced a twofold increase in β-cell proliferation compared with controls, but exposure to higher concentrations of 2 and 5 ng/ml resulted in an ∼1.5-fold decrease (Fig. 1A and B; Table 1). By contrast, 0.01 ng/ml IL-1β reduced baseline β-cell apoptosis twofold, whereas higher doses of 2 and 5 ng/ml IL-1β increased the apoptosis rate by 2.3- and 3.6-fold, respectively (Fig. 1A and C; Table 1). Those changes in β-cell turnover were accompanied by a 1.6-fold increase in glucose-stimulated insulin secretion (GSIS) at 0.02 ng/ml IL-1β and a 2.8-fold reduction by 2 ng/ml IL-1β (Fig. 1D and E). Insulin content of the islets was not significantly affected (data not shown).

We have recently shown expression of IL-1Ra in human β-cells (39). Interestingly, secreted IL-1Ra will antagonize the effect of IL-1β. To highlight the potential cross-talk between these two molecules, we repressed endogenous production IL-1Ra by RNA interference. As previously shown, siRNA directed to IL-1Ra (siIL-1Ra) with a transfection efficiency of ∼70% suppressed IL-1Ra mRNA by 69 ± 6% (39). Blockade of endogenous IL-1Ra by siIL-1Ra did not significantly affect baseline β-cell proliferation (Fig. 2A) but led to a twofold increase in apoptosis, levels similar to the ones obtained with 2 ng/ml IL-1β (Fig. 2B). Addition of 0.02 ng/ml IL-1β failed to increase β-cell proliferation and to further increase β-cell apoptosis (Fig. 2A and B). In contrast, addition of exogenous rhIL-1Ra (500 ng/ml) increased β-cell proliferation and protected from the deleterious effects of a high concentration of 2 ng/ml IL-1β. Furthermore, the compound protected cells against prolonged exposure to 33.3 mmol/l glucose (Fig. 2A and B). Similarly, siIL-1Ra completely blocked GSIS in the absence or presence of 0.02 ng/ml IL-1β, whereas exogenous rhIL-1Ra prevented the impairment of GSIS in the presence of 2 ng/ml IL-1β or 33.3 mmol/l glucose (Fig. 2C and D). Interestingly, the low concentration of 0.02 ng/ml IL-1β induced a 1.8-fold increase in the release of its receptor antagonist IL-1Ra into the culture medium (Fig. 2E).

Next, we investigated the underlying mechanisms of the proliferative effect of IL-1β. We hypothesized that the Fas pathway mediates IL-1β–induced β-cell proliferation. To test this hypothesis, first we investigated whether low concentrations of IL-1β are capable of inducing Fas. Exposure of human islets to 0.02 ng/ml IL-1β for 4 days induced Fas expression to levels similar to those observed with higher concentrations of IL-1β and 33.3 mmol/l glucose (Fig. 3A). However, low concentrations of IL-1β did not activate caspase 8, the most upstream caspase in the Fas apoptotic pathway, while higher concentrations than 0.02 ng/ml IL-1β cleaved procaspase 8, releasing the active form of the protease (Fig. 3B). To examine whether the induction of proliferation by 0.02 ng/ml IL-1β is caused by the interaction of constitutively expressed Fas ligand (16,28) and upregulated Fas, we used the Fas antagonistic antibody ZB4. ZB4 inhibited the proliferative effect of IL-1β (Fig. 3C). Since Fas signaling may induce proliferation in the presence of FLIP (34), we analyzed FLIP protein modulation by increasing concentrations of IL-1β. Similar to IL-1β–mediated changes in β-cell proliferation (Fig. 1B), low concentrations of IL-1β induced FLIP expression, while higher concentrations were inhibitory (Fig. 3D and E). Of note, FLIP mRNA expression remained unchanged (data not shown) in agreement with previous studies showing regulation of FLIP only at the protein level (34). The functional role of FLIP in IL-1β–induced β-cell proliferation was then investigated by RNA interference (siRNA directed to FLIP [siFLIP]). siFLIP suppressed FLIP expression in most β-cells (Fig. 3F) leading to a 5.2-fold decrease in FLIP protein expression (for representative blot, see Fig. 3G). Repression of FLIP resulted in impaired baseline β-cell proliferation and prevented IL-1β–mediated replication, whereas scrambled siRNA had no effect on the number of Ki-67–positive β-cells (Fig. 3H). We next investigated the potential implication of the transcription factors PDX-1 and Pax4, known to promote β-cell replication (40–45). Low concentrations of IL-1β stimulated PDX-1 mRNA and protein expression, which were decreased by higher concentrations of IL-1β (Fig. 3I–K). Similarly, low concentrations of IL-1β stimulated Pax4 and insulin mRNA expression, which were decreased by higher concentrations (Fig. 3L and M). However, the stimulation of insulin mRNA by low IL-1β failed to reach statistical significance. Furthermore, Pax4 was stimulated at 12 h but no longer after 4 days.

To substantiate the role of IL-1β in normal glucose homeostasis in vivo, glucose tolerance tests were performed in IL-1β knockout (IL-1β^−/−) mice. As predicted, we found that IL-1β^−/− mice displayed glucose intolerance (Fig. 4A). Normal sensitivity of the IL-1β^−/− mice to injected insulin (Fig. 4B) ruled out the possibility that insulin resistance was responsible for the impaired glucose tolerance. Food intake and body weight were similar in IL-1β^−/− and wild-type mice (not shown). Analysis of pancreatic tissue of IL-1β^−/− mice revealed a normal structure of the islets with no significant decrease in β-cell mass (Fig. 4C and D). However, insulin mRNA expression was strongly decreased (Fig. 4E), although pancreatic insulin content showed no significant changes (5.05 ± 1.04 vs. 4.0 ± 0.7 mol/g insulin per pancreas in normal versus IL-1β^−/− mice, respectively). Consistent with our in vitro experiments, Fas, FLIP, PDX-1, and Pax4 were decreased in IL-1β^−/− mice (Fig. 4E). Islets of IL-1β^−/− mice exhibited a strong decrease in baseline β-cell proliferation and GSIS compared with wild-type mice, while apoptosis and insulin content were not significantly different (Fig. 5).

The resistance of IL-1β^−/− islets to β-cell glucotoxicity was then studied. Chronic exposure of wild-type mouse islets to 33.3 mmol/l glucose for 4 days impaired β-cell proliferation, induced apoptosis, impaired GSIS, and decreased insulin content (Fig. 5). In contrast, islets of IL-1β^−/− mice were not further altered by high glucose concentrations (Fig. 5), supporting the concept of IL-1β–mediated glucotoxicity (18). Finally, we have confirmed that glucose regulates IL-1β secretion in mouse islets (2.85 ± 0.8 and 5.68 ± 1.6 pg/ml of IL-1β in islets cultured for 4 days in 11.1 and 33.3 mmol/l glucose, respectively, n = 5) as described in human islets (18).

The role of IL-1β and other cytokines in the pathogenesis of type 1 diabetes is well established (25,46). More recently, the concept emerged that cytokines may also mediate nutrient-induced β-cell dysfunction during the development of type 2 diabetes (5,6,47,48). Intriguingly, some of these cytokines can be produced by β-cells, including IL-6, IL-1β, IL-1Ra, and pancreatic-derived factor (18,19,39,49–53). The fact that the nonimmune β-cell synthesizes cytokines suggests their implication in islet physiology. Accordingly, we observed beneficial effects of low concentrations of IL-1β on β-cell proliferation, apoptosis, and secretory function.

A low concentration of IL-1β stimulated IL-1Ra, which in turn stimulated β-cell proliferation. Likewise, adenoviral expression of IL-1Ra increases β-cell replication in rat islets (54). At first sight, these results suggest a signaling function for IL-1Ra. However, this is unlikely since IL-1Ra only has a single IL-1 receptor–binding domain and is therefore unable to recruit the IL-1 receptor accessory protein, the second chain of the receptor complex, which is believed to be necessary for signaling (55). Therefore, the proliferative effects of exogenous IL-1Ra probably result from a restoration of a beneficial ratio of IL-1 to IL-1Ra and not from direct effects of the latter. Corroborating this hypothesis, repression of endogenous IL-1Ra was deleterious, probably due to unbalanced actions of IL-1.

Fas expression on the surface of pancreatic β-cells contributes to cytokine-induced apoptosis (30,56). However, when FLIP is activated, Fas becomes mitogenic (34). We propose that the Fas-FLIP pathway is modulated by IL-1β. This is supported by a simultaneous elevation of Fas and FLIP proteins at low IL-1β concentrations leading to β-cell proliferation. In contrast, higher concentrations of IL-1β decreased FLIP while Fas remained elevated, leading to decreased proliferation and induction of apoptosis. Furthermore, the Fas antagonistic antibody ZB4 and siRNA to FLIP both prevented IL-1β–induced β-cell proliferation.

The physiological importance of lL-1β was also apparent in IL-1β knockout mice, showing impaired glucose tolerance. Islets of these mice displayed decreased mRNA expression levels of PDX-1, Pax4, insulin, Fas, and FLIP. Of note, these transcripts were enhanced by 0.02 ng/ml IL-1β. At higher concentrations, IL-1β suppressed these factors uncovering its toxic effects. IL-1β may participate in glucotoxicity (18), a concept substantiated by the protection from harmful actions of high glucose of IL-1β knockout mouse islets.

Most factors playing a role in β-cell turnover are also involved in the regulation of β-cell secretory function (57). This is also true for IL-1β, which will have distinct effects depending on the duration of exposure. On the short-term, IL-1β will predominantly influence β-cell secretory function independently of changes in β-cell mass. Indeed, in vitro, a 4-day exposure of human islets to different concentrations of IL-1β led to changes in cell turnover in <1% of β-cells, although β-cell secretory function was almost completely blocked at high concentrations. Nevertheless, this does not mean that prolonged exposure to IL-1β will not affect β-cell mass. Indeed, β-cells remain positive for Ki-67 for ∼12 h. In absolute value, 0.02 ng/ml IL-1β increased the number of proliferating β-cells by 0.2%. The human pancreas has between 500,000 and 1,000,000 islets, comprising 500–2,000 β-cells each. Therefore, it can be estimated that low concentrations of IL-1β will increase the number of β-cells by ∼4,000,000 per day, leading to a doubling of β-cell mass within 250 days. Such an impressive increase in β-cell mass may occur in vivo over a similar time period (e.g., during obesity [4]). Therefore, the observed differences are within an expected range and support the relevance of the finding.

A key signaling molecule regulating β-cell turnover and function is insulin receptor substrate (IRS)-2 (58). Interestingly, it has been proposed that IL-1β may mediate its deleterious effects via IRS-2 degradation (8). Conversely, it is conceivable that at low concentrations IL-1β also signals via IRS-2. Indeed, while IRS-2 degradation leads to apoptosis, IRS-2 activation enhances β-cell proliferation and function. However, further investigations are required to support this hypothesis.

Cytokines are central in the development of diabetes. However, at low concentrations, IL-1β promotes β-cell function and survival. Therefore, glucose-induced IL-1β may have an important role in the long-term adaptation of the β-cells to hyperglycemia. Short-term exposure to hyperglycemia will induce low levels of IL-1β, inducing IL-1Ra, Fas, and FLIP, leading to decreased apoptosis and enhanced proliferation and function. However, prolonged hyperglycemia will increase the ratio of IL-1β to IL-1Ra, decreasing FLIP and directing Fas to signals deleterious to the cell.

This work was supported by the Swiss National Science Foundation (grant no. PP00B-68874/1 to M.Y.D., no. 32-66907.01 to C.B.W., and no. 3100AO-107682/1 to B.R.G.), by the Juvenile Diabetes Research Foundation (no. 3-2004-212 to K.M. and no. 1-2005-826 to M.Y.D.), by the Larry L. Hillblom Research Foundation (to K.M.), by a European Foundation for the Study of Diabetes/MSD Research Award (to M.Y.D.), and by a National Institutes of Health-Seeding Collaborative Research in Beta Cell Biology Award (to C.W. and M.Y.D.).

The authors are greatly indebted to Dr. Peter C. Butler for critical review of the manuscript. We thank I. Dannenmann, G. Siegfried-Kellenberger, and S. Abdullamin for technical assistance.