RyR2 and calpain-10 delineate a novel apoptosis pathway in pancreatic islets.

Cells are programmed to die when critical signaling and metabolic pathways are disrupted. Inhibiting the type 2 ryanodine receptor (RyR2) in human and mouse pancreatic beta-cells markedly increased apoptosis. This mode of programmed cell death was not associated with robust caspase-3 activation prompting a search for an alternative mechanism. Increased calpain activity and calpain gene expression suggested a role for a calpain-dependent death pathway. Using a combination of pharmacological and genetic approaches, we demonstrated that the calpain-10 isoform mediated ryanodine-induced apoptosis. Apoptosis induced by the fatty acid palmitate and by low glucose also required calpain-10. Ryanodine-induced calpain activation and apoptosis were reversed by glucagon-like peptide or short-term exposure to high glucose. Thus RyR2 activity seems to play an essential role in beta-cell survival in vitro by suppressing a death pathway mediated by calpain-10, a type 2 diabetes susceptibility gene with previously unknown function.

The pancreatic ␤-cell plays a central role in the pathogenesis of diabetes mellitus. A reduction in ␤-cell mass mediated at least in part by an increase in apoptosis is characteristic of the diabetic state (1)(2)(3). It is becoming clear that several pathways can lead to ␤-cell apoptosis, including cytokine signaling, excessive Ca 2ϩ influx during chronic hyperglycemia, high levels of free fatty acids, hypoxia or hypoglycemia, endoplasmic reticulum (ER) 1 stress, and loss of growth factor signaling (1,(3)(4)(5)(6)(7)(8)(9)(10)(11)(12). Whether various inducers of apoptosis employ distinct molecular mechanisms has not been systematically studied. Intracellular Ca 2ϩ stores play an important role in the regulation of apoptosis in many cell types (13,14). The present study was undertaken to test the hypothesis that alterations in specific intracellular Ca 2ϩ stores may induce apoptosis in pan-creatic ␤-cells. There are at least three classes of intracellular Ca 2ϩ stores in ␤-cells, and these are sensitive, respectively, to inositol trisphosphate (IP 3 )/thapsigargin, nicotinic acid adenine dinucleotide phosphate, and cyclic ADP ribose/ryanodine (15)(16)(17)(18). In many cell types, ryanodine receptor Ca 2ϩ channels (RyR) transmit Ca 2ϩ signals directly to closely associated mitochondria (19). In the MIN6 ␤-cell line, RyR were shown to regulate ATP production (20). Because of their role in regulating intracellular Ca 2ϩ and mitochondrial function, we focused specifically on RyR as likely mediators of ␤-cell apoptosis. Of the three RyR subtypes, two have been reported to be present in ␤-cells, RyR1 and RyR2. The latter is more abundant and can be distinguished from the former by its insensitivity to dantrolene (21,22). Ryanodine, a plant alkaloid, is the most specific probe for all RyR subtypes, and its activity is lost in RyR-deficient cells (23,24).
In the present study, we examined the role of RyR in the survival of human and mouse pancreatic islets. We uncovered a novel apoptosis pathway that is initiated when Ca 2ϩ flux through RyR2 is blocked. The mechanism of ryanodine-induced programmed cell death shares important features with palmitate-induced apoptosis, and both require activation of calpain-10, a type 2 diabetes susceptibility gene.
Immunofluorescence Staining-Ryanodine receptors were detected using a monoclonal antibody that recognizes RyR1 and RyR2 (Affinity BioReagents, Golden, CO). For double-labeling experiments, insulin staining was performed using guinea pig anti-insulin antibody (Linco, St. Charles, MO). Secondary antibodies conjugated to either Alexa Fluor 488 or 596 were applied for 2 h. Controls using no primary antibody or no second antibody were negative for each experiment. A FluoView™ laser scanning confocal microscope (Olympus, Melville, NY) was used for studies of RyR localization.
Cell Culture and Ca 2ϩ Imaging-Human islets were obtained from the Washington University Human Islet Isolation Core Lab. Standard culturing methods were employed as described for human islets (16), mouse islets (1), and MIN6 cells (9) using RPMI 1640 media (with 10% fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin). RPMI 1640 medium contained 10 mM glucose unless otherwise indicated. Human and mouse islets were gently dispersed, loaded with 1 M Fura-4F-AM for single-cell Ca 2ϩ imaging studies as described (1,16). Ringer's solutions contained 3 mM glucose, unless otherwise indicated.
Measurement of Apoptosis-We used four independent methods to measure apoptosis in primary islets and MIN6 cells: namely, PCRenhanced DNA laddering, ApoPercentage dye labelling, cell density/ viability, and terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling (TUNEL). The PCR-enhanced DNA ladder analysis was adapted from the ApoAlert Kit (Clontech, Palo Alto, CA) to measure apoptosis in groups of 8 -15 islets. This method is semi-quantitative and extremely sensitive (i.e. requires very little tissue). This approach uses adapter nucleotides and short PCR runs to selectively amplify DNA ladders. Briefly, after lysing islets at 55°C for 15 min, genomic DNA was isolated by using the DNeasy kit (Qiagen, Valencia, CA), taking great care not to break the large strands of DNA; concentration was quantified by UV spectrophotometry. 200 ng of genomic DNA was ligated to adapters using T4 ligase (New England Biolabs, Beverly, MA) at room temperature for 2 h. PCR products were run on a 2% agarose/ ethidium bromide gel in 0.5ϫ TBE (90 mM Tris/64.6 mM boric acid/2.5 mM EDTA, pH 8.3). DNA ladders were quantified as the mean intensity using Adobe Photoshop. This approach is conservative because blank lanes exhibit a small amount of gray-scale density. For each gel, DNA ladder density was expressed as a percentage of untreated/wild-type islets. These percentages were then averaged to give the mean change Ϯ S.E. for each condition.
Apoptosis in single isolated human or mouse islet cells was examined qualitatively using the ApoPercentage dye (Biocolor, Belfast, Northern Ireland), as described previously (1). This dye labels live cells bright pink that are undergoing phosphatidylserine translocation to the outer plasma membrane, a characteristic of apoptosis, but not necrosis. Nonapoptotic cells remain clear.
Cell death was measured quantitatively in cultures of MIN6 cells by measuring viability/cell density and by TUNEL analysis using the DeadEnd Colorimetric kit (Promega, Madison, WI) as described (1). The number of cells in each well and the number of TUNEL-positive cells were quantified using MetaMorph™ image analysis software. To further characterize MIN6 cell apoptosis, cells were stained with antibody recognizing the active, cleaved form of caspase-3 (Trevigen Inc., Gaithersburg, MD).
␤-Cell-specific Calpastatin and Calpain-10 Transgenic Mice-Animal use protocols were approved by Washington University. Male mice, 8 -30 weeks old, and littermate controls were used for all experiments. Transgenic mice were engineered by standard techniques. The rat insulin 2 promoter was used to drive the expression of human calpastatin or human calpain-10a using the strategies outlined (see Figs. 4 and 7). The expression of the human calpastatin transgene was confirmed by immunohistochemistry of pancreatic sections as described previously (1) using an antibody to human calpastatin (Calbiochem). The expression of the human calpain-10a transgene was confirmed by immunoblot using a polyclonal calpain-10 antibody from Biogenesis (Kingston, NH). The function of the transgenes was confirmed using measurements of islet calpain activity.
Calpain-10 Knockout Mice-Calpain-10 knockout mice (Capn10 Ϫ/Ϫ ) were generated by deleting exon 2, as follows. A BAC clone containing Capn10 from a 129/SvJ mouse genomic library (Genome Systems Inc, St. Louis, MO). A BamHI/SpeI (6.9 kb) fragment was used to construct a gene-targeting vector pACN-Capn10. A self-excision ACN cassette was inserted at a BglII site (5Ј of exon 2) and loxP sequence at an SnaBI site (3Ј of exon 2). A targeting vector was designed to delete the predicted active-site Cys 73 in exon 2 of the gene. Linearized pACN-Capn10 plasmid was electroporated into embryonic stem (ES) cells. Genomic DNA from G418-resistant ES cell clones was analyzed for homologous recombination by Southern blotting. Two ES cell clones containing the targeted mutation were injected into C57BL/6B embryos. Chimeric mice were backcrossed to C57BL/6B mice. Germline transmission of the mutant allele was detected by Southern blot and PCR analysis of tail DNA from F1 offspring with Agouti coat color. One ES clone showed germline transmission. Homozygous Capn10 Ϫ/Ϫ mice from the F1 cross were identified by PCR. The recombinant Capn10 ACN allele was also confirmed by Southern blotting (not shown). The loss of calpain-10 protein expression was confirmed by Western blotting of brain lysates using an affinity-purified anti-calpain-10 antibody generated against a 15 amino acid peptide of human calpain-10 ( 194 GGQQDRPGRWE-HRTC 208 ) from Research Genetics (Huntsville, AL).
Statistical Analysis-The unpaired t test was used to test the significance of differences between groups. Differences were considered significant when p Ͻ 0.05. Results are presented as mean Ϯ S.E.

Localization of Active RyR in Human and Mouse ␤-Cells-
Immunofluorescence staining demonstrated the presence of RyR in a majority of human and mouse ␤-cells (Fig. 1A). Confocal studies of RyR localization showed a punctate and vesicular pattern of expression (Fig. 1B). Although previous studies have documented the presence of specific ryanodine-binding sites and mRNA for RyR2 in human and rodent ␤-cells (17,22), whether these Ca 2ϩ channels are functional has not been directly tested. Nanomolar concentrations of ryanodine, which bind to the open RyR channel pore and increase the probability that the channel will open (27,28), evoked an increase in cytosolic Ca 2ϩ in ␤-cells from mouse islets (Fig. 1C) and human islets (18). This finding suggests that ␤-cell RyR are active under basal conditions. Micromolar concentrations of ryanodine, which inhibit the RyR, did not elevate cytosolic Ca 2ϩ in unstimulated cells (Fig. 1D). These findings provide strong evidence that functional RyR are present in both human and mouse ␤-cells.
Role of RyR in ␤-Cell Apoptosis-We employed several approaches to determine whether ryanodine induced apoptosis in cultured islets, including PCR-enhanced DNA ladder analysis, a sensitive technique for detecting DNA fragments generated during apoptosis but not necrosis (1). DNA ladders, reflecting the organized cleavage of DNA, were detected within 36 h after exposure of human islets ( Fig. 2A) or mouse islets (see below) to a blocking concentration (10 M) of ryanodine. In contrast, . A significant increase in red TUNEL-positive cell nuclei is also seen in these representative images. C, normalized to control, the percentage of active caspase-3-positive MIN6 cells (stained red), relative to the total number of cells in each treatment, was 129 Ϯ 7% for ryanodine (p Ͻ 0.05), 1679 Ϯ 15% for thapsigargin (p Ͻ 0.05), and 232 Ϯ 3% for palmitate (p Ͻ 0.05; n ϭ 3). D, in ryanodine-treated mouse islets, the caspase-3 inhibitor DEVD-CHO (10 M) did not block apoptosis (123 Ϯ 22%; n ϭ 3). DEVD-CHO reduced apoptosis 69 Ϯ 11% (p Ͻ 0.05) in mouse islets treated with 1 M thapsigargin) (n ϭ 3). activation of RyR with 1 nM ryanodine or inhibition of RyR1 with dantrolene did not induce significant apoptosis. We confirmed that 10 M ryanodine induced apoptosis by measuring phosphatidylserine translocation in human ␤-cells (not shown), as well as by measuring a decrease in cell number and an increase in TUNEL-positive cells in the MIN6 cell line (Fig.  2B). Together, these results indicate that inhibition of basal Ca 2ϩ flux through RyR2 activates ␤-cell apoptosis.
The Role of Caspase-3 in Ryanodine-induced Apoptosis-The mechanism of ryanodine-induced apoptosis was studied in comparison with two known inducers of ␤-cell apoptosis, thapsigargin and the free fatty acid palmitate (1,8). The mechanism of the potent effect of palmitate on apoptosis is poorly understood. Thapsigargin causes ER stress and induces apoptosis by depleting IP 3 -sensitive Ca 2ϩ stores that are distinct from those targeted by ryanodine in ␤-cells (29). In contrast to thapsigargin-induced apoptosis (9), apoptosis caused by inhibiting RyR2 was associated with relatively little caspase-3 activation (Fig.  2C). Apoptosis induced by palmitate showed an intermediate association with caspase-3 activation, when compared with ryanodine and thapsigargin. To address whether ryanodineinduced cell death was mediated by caspase-3, we treated islets with the membrane permeant caspase-3 inhibitor DEVD-CHO. Ryanodine-induced apoptosis was not inhibited by DEVD-CHO, whereas thapsigargin-induced apoptosis was significantly reduced (Fig. 2D). The implication of a cell death pathway relatively independent of caspase-3 prompted us to test whether an alternate pathway, involving calpain, mediated ␤-cell apoptosis in response to ryanodine.
Role of the Calpain System in ␤-Cell Apoptosis-Ryanodine, but not dantrolene, increased calpain activity in mouse islets (Fig. 3A). The small molecule calpain inhibitor N-acetyl-Leu-Leu-Met-CHO (ALLM), which we have previously shown decreases calpain activity by Ͼ50% in islets (25), blocked ryanodine-induced apoptosis but had no effect on thapsigargininduced apoptosis or basal apoptosis (Fig. 3B). ALLM also blocked apoptosis induced by palmitate. A genetic approach was used to confirm the role of calpain in ␤-cell apoptosis. We generated transgenic mice that express calpastatin, an endogenous calpain inhibitor protein (30), exclusively in ␤-cells using the rat insulin promoter (Cast RIP mice; Fig. 3C). Calpain activity was reduced more than 50% in islets from these mice (Fig.  3E), and this was associated with reductions in apoptosis induced by ryanodine or palmitate (Fig. 3F). The reduction in ␤-cell apoptosis with ALLM and calpastatin, both of which inhibit multiple isoforms of calpain, establishes an important role for a calpain-mediated pathway in this process.
The Role of Calpain-10 in ␤-Cell Apoptosis-Next, we sought to determine the specific isoform of calpain that is involved in ryanodine-and palmitate-induced apoptosis. Ry- anodine increased calpain-10 mRNA in mouse islets by ϳ2.5fold (Fig. 4, confirmed by real-time PCR, not shown), but had no significant effect upon calpain 1 and calpain 2 mRNA (not shown). Importantly, the ryanodine-induced increase in mRNA was an early event, evident by 12 h and lasting at least 5 days. To define the role of calpain-10 in ␤-cell apoptosis, we examined Capn10 Ϫ/Ϫ islets from mice in which the calpain-10 gene had been deleted (Fig. 5A). Ryanodine failed to increase calpain activity in Capn10 Ϫ/Ϫ islets (Fig. 5B). Ryanodine-induced apoptosis and palmitate-induced apoptosis were also prevented in Capn10 Ϫ/Ϫ islets (Fig. 5C). Conversely, ryanodine-induced apoptosis was enhanced in islets from transgenic mice with ␤-cell-specific overexpression of human calpain-10 (Fig. 6, A-C, Capn10 RIP ). Capn10 RIP islets showed enhanced calpain activity (Fig. 6D), suggesting that our assay measures the activity of calpain-10, in addition to other calpains. Together, the results strongly suggest that calpain-10 mediates apoptosis induced by inhibition of RyR2. The observation that apoptosis resulting from exposure to palmitate was inhibited in Capn10 Ϫ/Ϫ islets and enhanced in Capn10 RIP islets suggests that cell death induced by free fatty acids shares common steps with the RyR2 pathway.
Effects of GLP-1 on Calpain Activity and Apoptosis-We next tested whether agents known to promote ␤-cell survival may act by modulating the RyR2/calpain-10 pathway. GLP-1, a potent inhibitor of ␤-cell apoptosis (31), mobilizes intracellular Ca 2ϩ and activates mitochondria via RyR (17,20). GLP-1 decreased basal calpain activity and abolished ryanodine-induced calpain activation (Fig. 7A). GLP-1 also prevented apoptosis and cell loss in MIN6 cells treated with ryanodine (Fig. 7B). Similar results were seen with the GLP-1 receptor agonist, exendin-4 (not shown). Thus, the anti-apoptotic effects of GLP-1 may, in part, be due to activation of RyR2 and subsequent inhibition of calpain activity.
Interaction between Glucose and RyR2-The glucose level to which pancreatic islets are exposed is another important determinant of apoptosis (1, 4 -6, 32). First, we wanted to determine the relationship between glucose signaling and RyR2. Surprisingly, ryanodine had significant effects on Ca 2ϩ homeostasis during stimulation with glucose or KCl, in contrast to the situation in the basal conditions (3 mM glucose) described in Fig. 1. Concentrations of ryanodine that inhibit RyR potentiated the Ca 2ϩ response to elevated glucose in a subpopulation of human or mouse ␤-cells (Fig. 8, A and B) and Ca 2ϩ responses to 30 mM KCl in all cells (Fig. 8C), suggesting that ryanodine-sensitive FIG. 5. Calpain-10 is required for ryanodine-and palmitateinduced apoptosis. A, a fragment of ϳ25 kDa was apparent in immunoblot of brain lysate from wild type but not Capn10 Ϫ/Ϫ mice. B, calpain activity after 36-h culture was measured in the presence or absence of 10 M ryanodine in wild-type or Capn10 Ϫ/Ϫ islets. Asterisks denote significant difference (p Ͻ 0.05) from control. C, apoptosis in islets from Capn10 Ϫ/Ϫ mice was compared after 60-h culture in 10 mM glucose control conditions (100 Ϯ 9% versus wild type), 1 M thapsigargin (72 Ϯ 15% versus wild type), 10 M ryanodine (51 Ϯ 6% versus wild type, p Ͻ 0.05) or 250 M palmitate (56 Ϯ 9% versus wild type, p Ͻ 0.05) (n ϭ 6).
FIG. 6. Transgenic over-expression of calpain 10 enhances ryanodine-and palmitate-induced apoptosis. A, transgene construct for Capn10 RIP mice. B, Western blot of islet lysate using antibodies for the Xpress Tag (right panel) and an antibody to human calpain 10 (Biogenesis, left panel). The band below ϳ94 kDa is abundant in transgenic tissue and is approximately the expected size of calpain-10 (including the Xpress tag). C, apoptosis in islets from Capn10 RIP mice was compared after 60-h culture in 10 mM glucose control conditions (117 Ϯ 22% versus wild type), 1 M thapsigargin (114 Ϯ 16% versus wild type), 10 M ryanodine (138 Ϯ 11% versus wild type, p Ͻ 0.05) or 250 M palmitate (125 Ϯ 10% versus wild type, p Ͻ 0.05) (n ϭ 6). D, calpain activity is increased in islets from Capn10 RIP mice. Asterisks denote significant difference (p Ͻ 0.05) from control. Ca 2ϩ stores are involved in Ca 2ϩ uptake under these conditions. In contrast, dantrolene evoked an immediate decrease in Ca 2ϩ in stimulated ␤-cells, consistent with a role for RyR1 in Ca 2ϩ -induced Ca 2ϩ release (CICR). Neither stimulating RyR with 1 nM ryanodine nor blocking IP 3 R with xestospongin C had any effect upon depolarization-induced Ca 2ϩ responses (18). These findings suggest that RyR2 play a novel non-CICR role during glucoseinduced Ca 2ϩ -influx, whereas RyR1 mediates CICR in ␤-cells.
Having established the link between glucose signaling and RyR2, we examined whether high glucose modulates calpain activity and protects islets from ryanodine-induced apoptosis, because short-term exposure (i.e. 2 days) to high glucose (25 mM) has been shown to inhibit ␤-cell apoptosis (33). This experiment would also shed some light on the possible mechanisms of ryanodine-induced apoptosis. Notably, if 10 M ryanodine were causing apoptosis by augmenting CICR, 25 mM glucose would be expected to potentiate ryanodine-induced cell death. However, a 2-day culture of islets in 25 mM glucose blocked ryanodine-stimulated calpain activity and ryanodine-induced apoptosis (Fig. 9). Thus, the beneficial effects of short-term exposure to high glucose on ␤-cell survival may be mediated by a novel interaction with RyR2. Accordingly, we also examined the role of the RyR2/calpain pathway in cell death caused by a lowered rate of ␤-cell metabolism. Apoptosis induced by in vitro hypoglycemia (2 mM glucose for 60 h) was completely absent in Capn10 Ϫ/Ϫ islets (Fig.  10A). On the other hand, apoptosis was enhanced in Capn10 RIP islets incubated in moderately low glucose (5 mM) but not at 10 mM or 25 mM glucose (Fig. 10B). These results suggest that calpain-10 plays an important role in apoptosis induced by exposure to low glucose concentrations but not apoptosis induced by prolonged exposure to high glucose concentrations. Thus, the RyR2/calpain-10 death pathway is turned on when ␤-cell metabolic activity is low and turned off by stimuli that increase metabolic activity.
Roles of RyR2 and Calpain-10 in Apoptosis Induced by Chronic Hyperglycemia-Unlike the pro-survival effects of short incubations in 25 mM glucose, chronic stimulation (i.e. 7 days) with high glucose induces apoptosis (1,32). Apoptosis induced by chronic (7 day) hyperglycemia was not affected by calpain-10 knockout (Fig. 10C) or calpastatin over-expression (102 Ϯ 15% versus wt, n ϭ 6), suggesting that apoptosis mediated by high glucose does not involve a calpain-10-dependent pathway. Like thapsigargin-induced apoptosis, glucose toxicity is known to involve caspase-3 (32). Therefore, the RyR2/calpain-10 apoptosis pathway is separate from other known ␤-cell apoptosis pathways. Interestingly, inhibiting RyR protected ␤-cells against apoptosis induced by 7-day culture in high glucose (Fig. 10D). Therefore, we tested whether the protective effect of ryanodine on apoptosis induced by chronic stimulation may be mediated by blocking CICR through RyR1. Apoptosis induced by chronic exposure to high glucose was inhibited by dantrolene (Fig.  10E), implicating RyR1 and CICR in this process. A prominent role for dantrolene-sensitive RyR in excitotoxic neuronal cell death has been proposed (34). DISCUSSION The present studies were undertaken to assess the role of RyR in the survival of pancreatic ␤-cells and to determine the mechanism by which these Ca 2ϩ channels regulate apoptosis. Our findings indicate that inhibiting the RyR2 is associated with increased apoptosis, suggesting that maintenance of normal basal Ca 2ϩ flux through this channel is essential for ␤-cell survival. To our knowledge, these are the first results in any cell type to suggest that blocking RyR can induce apoptosis, although excessive RyR activity has been linked to cell death in other tissues, including the brain and heart (30,35). That RyR2 may be a central molecule in the control of programmed cell death is perhaps not surprising. RyR are very large proteins that can directly sense cytosolic Ca 2ϩ , luminal Ca 2ϩ , ATP, redox potential, and nitric oxide (36), thus placing them in a key position to integrate multiple signals known to influence apoptosis. The general importance of RyR2 activity is underscored by the embryonic lethality of RyR2 Ϫ/Ϫ mice after embryonic day 9.5 (37). In contrast, no differences in islet morphology or ␤-cell ultrastructure were observed in mice lacking both RyR1 and RyR3 (38). Previous studies have reported that the expression of RyR2 is reduced in islets from several rodent models of diabetes (22). Together, these results suggest that RyR2 may be important in apoptosis in the ␤-cell and could, therefore, be involved in the impairment of insulin secretion and the pathogenesis of diabetes.
Our results define a number of novel aspects of the mechanisms of apoptosis in the pancreatic ␤-cell. We have clearly demonstrated the existence of multiple pathways leading to ␤-cell apoptosis (Fig. 11). In the pancreatic ␤-cell, as least two major two apoptosis pathways are apparent, based upon their requirement for calpain-10 or caspase-3. Thapsigargin-induced ER stress and chronic hyperglycemia are known to be associated with the classical caspase-3-dependent pathway (5,9). On the other hand, a novel calpain-10-dependent apoptosis pathway mediates cell death induced by ryanodine, hypoglycemia, and palmitate. Whether the pathways associated with these apoptotic stimuli involved additional molecules other than calpain-10 remains to be determined. Although we present evidence that RyR2 participates in glucose-induced Ca 2ϩ signaling, the link between RyR2 and palmitate remains unclear. Although other calpain isoforms have been suggested to regulate cell death in other tissues, our results are the first to demonstrate that any calpain isoform is involved in apoptosis in primary ␤-cells. Previously, we have shown that prolonged FIG. 9. Effects of short-term treatment with high glucose on ryanodine-induced calpain activity and apoptosis. A, ryanodinestimulated calpain activity was significantly lower after a 60-h incubation in 25 mM glucose, compared with 10 mM glucose. *, significant difference from control; **, significant difference from ryanodinetreated islets. B, although ryanodine induced significant apoptosis in mouse islet culture for 48 h in 5 mM glucose (200 Ϯ 14% versus no ryanodine, p Ͻ 0.05), no additional apoptosis was seen in 25  incubation with calpain inhibitors has deleterious effects upon the function of mouse islets (26). However, the links between inhibition of RyR2 activity, increased calpain-10 expression, and apoptosis have not been identified previously. Genetic variation in the calpain-10 gene has been linked to increased susceptibility to diabetes (39), although the function of the calpain-10 protein was not known.
We have shown that the RyR2/calpain-10 pathway can be directly modulated by high glucose and GLP-1. Both treatments would be expected to stimulate mitochondrial activity, and both are known to protect ␤-cells from apoptosis (20,31,33). In addition, GLP-1 signaling in human ␤-cells and MIN6 cells is known to be mediated by RyR-gated intracellular Ca 2ϩ stores (17,20). Our results suggest that the stimulation of Ca 2ϩ release from RyR2, possibly leading to increased mitochondrial ATP production (20), and the suppression of calpain activity may be important mechanisms by which GLP-1 promotes ␤-cell survival. GLP-1, or related agonists that activate RyR2, may be useful, therefore, in the context of clinical islet transplantation, where apoptosis during the procurement, isolation, storage, and engraftment of islets is a significant problem. Because improved hematopoietic stem cell engraftment was seen with exposure to cyclic ADP ribose (40), an endogenous activator of RyR2, the possibility that this pathway could be modulated to improve graft survival in human islet transplantation should be examined. Interestingly, we have documented distinct effects of RyR1 and RyR2 on apoptosis. Under basal conditions, Ca 2ϩ flux through RyR2 is required to prevent apoptosis, whereas active RyR1 are critical for the deleterious effects of prolonged stimulation with high glucose, consistent with a CICR mechanism. These studies, therefore, suggest that both the RyR1 and RyR2 should be explored as novel targets for diabetes drug development. Although multiple studies in other endocrine cell types have made use of the specificity of ryanodine for the RyR (41)(42)(43), and we are not aware of alternate molecular targets of this agent (23,24), the molecular ablation of RyR2 will be required to confirm our findings.
In conclusion, our results are the first to clearly define a physiological role in any tissue for calpain-10, recently identified as a type 2 diabetes susceptibility gene by linkage studies and positional cloning (39). The demonstration that a RyR2/calpain-10 pathway plays a critical role in ␤-cell sur-vival suggests novel mechanisms for the pathophysiology of ␤-cell dysfunction in type 2 diabetes as well as novel targets for therapeutic intervention to preserve ␤-cell function. Our results may also provide a new framework for the investigation of potential mechanisms whereby alterations in Ca 2ϩ handling by RyR may lead to cell death in other pathological states, such as Alzheimer's disease, ischemia/reperfusion injury, and heart failure.