Perspectives in Diabetes Apoptosis and the Pathogenesis of IDDM A Question of Life and Death Didac Mauricio and Thomas Mandrup-Poulsen

In type 1 diabetes, an immune-mediated process leads to the destruction of pancreatic P-cells. In the last decade, considerable progress has been made in under- standing the cellular and biochemical pathogenic processes of the disease. However, more needs t o be learned about the immune mechanisms leading to the development of autoreactive immune cells and the molecular mechanisms of P-cell death. The study of apoptosis of autoreactive lymphocytes as well as apo- ptosis of p-cells may give answers to many still unsolved questions. This review focuses on the possi- ble role of apoptosis both in the regulation of immune mechanisms involved in the pathogenesis of type 1 dia- betes and as a way for p-cells to die. The advancement in the knowledge of the possible role of apoptosis and its regulation in the pathogenesis of type 1 diabetes may provide new therapeutic tools for the prevention of the disease. Diabetes 47:1537-1543, 1998

here are two conceptually different ways through which eukaiyotic cells can die: one is death by

I necrosis caused by ischemic, chemical, or thermal cell iaury; the other is by apoptosis, a

type of programmed cell death, which is required for the normal maintenance of development and homeostasis in a cell system, an organ, or an individual as a whole (1). The machinery required for apoptosis is present in all mammalian cells (2). The interest in apoptosis has accelerated tremen- dously in the past few years (3). Recent evidence indicates that alterations in cell survival contribute to the pathogene- sis of a variety of diseases such as cancer, neurodegenerative disorders, viral infections, and autoimmune diseases, includ- ing IDDhl(4). A direct linkage between genes involved in the control of apoptosis and human autoimmune diseases is yet to be demonstrated. However, this is a field in which inves- tigations are just beginning. Understanding the molecular mechanisms and the genetic control of apoptosis and how the apoptotic process can be modulated have obvious implica-

From the Endocrine Unit (D.M.), Consorci Hospitalari Parc Tauli, Sabadell, Spain; and the Steno Diabetes Center (T.M.-P.), Gentofte, Denmark.

Address correspondence and reprint requests to Thomas Mandrup- Poulsen, MD DMSc, Steno Diabetes Center. Niels Steensens Vej, 2, DK- 2820 Gentofte, Denmark. E-mail: tmpo@novo.dk.

Receiveti for publication 11 May 1998 and accepted in revised form 2 July 1998.

FasL, Fa5 ligand; ICE, interleukin-lpxonverting enzyme; IFN, interferon; IL, interleukin; NO, nitric oxide; NOD, nonobese diabetic; Th, T-helper; TNF, tumor necrosis factor.

tions for the understanding of the pathogenesis of these dis- eases and the future design of new therapeutic strategies.

APOPTOSIS: AN OVERVIEW

Apoptosis plays a central role in the regulation of develop- ment and homeostasis in metazoan animals (2). The term "apoptosis" (Greek: leaf-fall) is often used synonymously with programmed cell death, the latter being a more func- tional definition. Apoptosis refers to the morphological fea- tures of programmed cell death, which is characterized by shrinkage, nuclear condensation, membrane blebbing, and membrane changes that eventually lead to phagocytosis of the affected cell. As opposed to what is seen in necrosis, apop- totic cells do not swell and burst, but undergo phagocytosis before the intracellular content leaks, thereby avoiding to evoke an immune response (5) (Table 1).

An unresolved question is how the apoptotic program is regulated so that only selected cells die. Apoptosis is con- trolled by many different signals that converge toward a common apoptotic program (2). Triggering of apoptosis may take place in different stages of the apoptotic cascade, depending on the initiating stimulus (6). The pathways that lead to activation of apoptosis vary among different cells, but the final pathway leading to cell death seems to be common. Therefore, when studying apoptosis, it is important to dlscern between processes that are cell-specific and those that are death-specific (7). The signals inducing apoptosis are varied and even the same signals can induce differentiation and proliferation, depending on conditions and cell type (3). The following signaling pathways have been implied in apoptosis: the Fas system (also known as APO-1 or CD95), protein tyro- sine kinases, serinelthreonine kinases, the Ras signaling pathway, protein kinase C, calcium signaling pathways, ceramide, CAMP, and free radicals (Fig. 1).

There are two gene families that are particularly important in the control of apoptosis: the genes encoding the inter- leukin-lp-converting enzyme (ICE) family of cysteine pro- teases (now known as caspases), and those related to the pro tooncogene bcl-2 (3). Proteolytic activity is involved in many apoptotic systems, the ICE family being of special impor- tance, since it seems to be central in F'as-mediated and tumor necrosis factor (TNQ-induced apoptosis. Proteolysis is prob- ably a common event in the apoptotic process, with different proteases involved and several proteins having been shown to be the subject of proteolytic activity (Table 2). The bcl-2 family of proteins are important modulators of apoptosis (Table 3). New evidence suggests that Bcl-2 protein has two

DIABETES, VOL. 47, OCTOBER 1998

TABLE 1 Main differential features of apoptosis and necrosis

Apoptosis Necrosis

Spontaneous (programmed cell death)

Chromatin becomes fragmented and condensed along the nuclear envelope. Finally, formation of one large or several (oligonucleosomal) fragments.

Loss of cell junctions and cell detachment from neighboring cells. Cell surface ruffling and blebbing (zeiosis). Cell membrane keeps the capacity to maintain osmotic gradients.

Organelles and cell shrink

Idury-induced (accidental cell death)

Only slight morphological changes in the nucleus.

Major damage in cell membrane with loss of its control of selective permeability.

Early mitochondria1 changes in shape and function, with swelling of cytoplasm and organelles. Organelle dissolution and rupture.

Budding off of membrane-bound apoptotic bodies, containing Cell swelling and rupture, with leakage of cell contents into the organelles and nuclear fragments. Stimulation of phagocytosis surrounding tissue space. This induces an inflammatory by nearby cells (professional or nonprofessional phagocytes). response and the initiation of tissue repair process. No spilling of intracellular contents. Therefore, no inflammatory response is induced.

different functions: 1) as an ion channel protein and 2) as an adaptorldocking protein through its binding to several other proteins. However, the precise way in which these proteins modulate apoptosis is unclear and conflicting theories have been proposed (3). Further, the gene product of bcl-2 does not prevent apoptosis under all circumstances (e.g., does not protect target cells from apoptosis induced by cytotoxic T- cells [8]). Finally, genes involved in cell differentiation and proliferation are also important in modulating the apoptotic process (e.g., the protooncogene c-myc, the tumor sup- pressor p53, and the apoptosis suppressor gene A20). Both c-myc and p53 are implicated in the induction of apoptosis under certain conditions, whereas A20 is a cytokine-induced primary response gene involved in the inhibition of the apoptotic process (3).

APOPTOSIS IN THE REGULATION OF IMMUNE MECHANISMS INVOLVED IN THE PATHOGENESIS OF IDDM Apoptosis and the immune repertoire. Apoptosis plays an important role in determining the quantitative and qualitative composition of the immune system. In IDDM, tolerance to the pancreatic p-cells is lost and immunological effector mech- anisms are then directed against this cell type (9). For the pur- pose of this study, we will focus mainly on the role of pro- grammed cell death in the selection of T- but not of B-cells. During the process of precursor T-cell receptor gene recom- bination, it is estimated that only -lo?? of T-cells vv~ll succeed in expressing both a and P chains of the T-cell receptor (5). The others die by apoptosis. Once a complete T-cell receptor is expressed on the cell surface, there are three alternatives awaiting the T-cell in the thymus: negative selection, positive selection, or no selection. If T-cell receptors bind with high affinity to self-antigens presented by the thymic antigen-pre- senting cells, the cell is eliminated by apoptosis to avoid autoimmunity. Failure or resistance of these cells to undergo

INTRACELLULAR Metabolic imbalance Cell-cycle perturbation DNA - Damage

EXTRACELLULAR Death receptor activation Growth factor withdrawal Hormones,cytokines, etc..

INFORMATION SIGNAL TRANSDUCTION MODES + DEATH SIGNALLING PATHWAYS

- Fas - Calcium - Ceramide - CAMP - Free radicals

- Ras signalling pathway - Protein tyrosine Kinases - Serine 1 threonine Kinases - Protein Kinase C

I

REGULATORY FACTORS -+ FINAL COMMON PATHWAY@) - bcl - 2 - related genes - Proteases (ICE, ICE- related, etc.) - Other oncogenes - Endonucleases

APOPTOSIS

FIG. 1. Schematic representation of the apoptotic cascade. The cell is able to integrate the information received from intracellular as well as extracellular milieu. Different death signals may activate one or more signal transduction pathways, which probably may converge in one or few final death pathways. Components of the different signal- ing pathways may interact with each other producing regulatory steps in which inhibition or potentiation of the process may take place. Further, triggering of apoptosis may take place at different points of the cascade, depending on the initiating stimulus.

DLaETES, VOL. 47, OCTOBER 1998

TABLE 2 Main proteases involved in apoptosis and some proteins shown to be cleaved during the apoptotic process

TABLE 3 The bcl-2-related genes, their protein products, and their effects on apoptosis

Proteases

ICE The ICE family of cysteine proteases prICE ICErel-I1 ICErel-111 CPP32 Mch-2

Serine proteases (e.g., urokinase-like plasminogen activator)

Calpain Cathepsins B and D Collagei~ase

Proteins undergoing cleavage

poly (ADP-ribose) polymerase Lamin B 7OkDa protein Topoisomerases I and I1 Histone H1 Protein kinase Cpl Fodrin

the apoptotic process results in the delivery of autoreactive T-cells t,o the periphery. Cells with no affmity for histocom- patibility antigens are also useless to the immune system and die by apoptosis, in a process that is probably triggered by endogenous glucocorticoids (10). Those T-cells express- ing T-cell receptors that recognize a foreign peptide, cross- reactive with a self-peptide, will undergo further maturation and will finally express either CD4 or CD8 (positive selection).

One of the central signaling pathways of apoptosis in the immune system is the Fas system. Fas is a 45-kDa surface receptor belonging to the TNFInerve growth factor receptor family, which on bin- by Fas ligand (FasL), t r a d u c e s the signal for programmed cell death (10,ll). Fas has been sug- gested t,o play an essential role in tissue development and homeostasis (12). In the immune system the Fas and FasL are essentially involved in downregulation of immune reactions and in T-cell-mediated cytotoxicity (13). The FasL is almost exclusively expressed by activated CD4' and CD8' T-cells (13). Although both T-helper (Th)l and Th2 cells express the Fas ligand, Thl cells are capable of lysing other target cells more readily than Th2 by a Fas-mediated mechanism (14,15). Thus, mature T-cells in the periphery may undergo an additional selection process known as peripheral clonal deletion, in which those T-cells interacting with self-anti- gens expressed in the periphery are eliminated (16). Further, to ensure that the organism can downregulate activated lym- phocytes, there must be a process of elimination of lym- phocytes after they have been activated by antigen (17). Whereas the negative and positive selections, which take place in the thymus, do not seem to involve a Fas-mediated mechanism, the Fas system is implicated in both the clonal deletion of autoreactive T-cells in peripheral lymphoid organs and the elhimtion of activated Tcells (13). Therefore, the Fas system is central in the regulation of peripheral immune responses.

Fas is constitutively expressed by mature T-cells, but expression is enhanced upon activation by antigen, thus ren- dering T-cells more sensitive to FasL-mediated apoptosis (18). If the Fas-FasL system is defective, activated lympho- cytes may accumulate (19). Thus, if these cells are not effi- ciently eliminated, the occurrence of autoimmune disease is

Genes Protein products Effect on apoptosis

Bcl-2a Bcl-20 Bcl-X, Bcl-X, Baxa BaxP Baxy Mcl-1 Bak

+ and - denote a permissive/inducer and a suppressor effect on apoptosis, respectively.

enhanced as has been shown in murine experimental models: the lpr and gld mutations in mice lead to a loss of function of Fas and FasL, respectively, which cause activated lympho- cytes to accumulate leading to the appearance of autoim- mune diseases (13,20). Apoptosis and autoimmunity. Recent data suggest that defective regulation of apoptosis in lymphoid cells may be a factor that contributes to the pathogenetic mechanism of autoimmune diseases (21). In the murine nonobese diabetic (NOD) model of IDDM, the development of autoimmunity has also been suggested to be influenced by mechanisms con- trolling programmed cell death of lymphocytes (22,23). Acti- vated T-cells from NOD mice showed a markedly increased resistance to induction of apoptosis after deprivation of interleukin (1L)-2 when compared to other strains (24). The gene encoding resistance to apoptosis was mapped to the middle region of chromosome 1 in which the antiapoptotic protooncogene bcl-2 is located. Through F2 intercross stud- ies, the Idd6 susceptibility locus in the distal region of chro- mosome 6 of NOD mice was strongly associated with increased resistance to apoptosis (25). Immature T-cells from NOD mice possessed an increased resistance to in vivo dexamethasone-induced apoptosis (23), and lymphocytes from NOD mice were more resistant to cyclophosphamide- induced apoptosis than other mouse strains (26), which again imply that defective apoptotic pathways may be involved in the disease pathogenesis. The resistance of Tcells to apoptosis in NOD mice has recently been found to be associated to upregulation of the antiapoptotic Bcl-x protein in T-cells (27). These findings in the NOD mouse may induce a new search for diabetes-susceptibility genes in human IDDM. Fas expression on NOD mouse T-cell subsets has not been reported, to the best of our knowledge.

Recently, it was shown that B- and T-cells from subjects with IDDM and those at risk for the disease are highly defec- tive in the surface expression of Fas (28). This led the authors to hypothesize that loss of tolerance in IDDM may be partly explained by a defective expression of Fas.

A ROLE FOR APOPTOSIS IN THE DESTRUCTION OF P-CELLS l b o schools of thought. Whether p-cell death in IDDM is due to apoptosis or necrosis or a combination of both has not been clarified. There are two main schools of thought regard-

DIABETES, VOL. 47, OCTOBER 1998

1 CYTOKINES I ,

/' Aconita 'ye DNA strand breaks\ \

Fas

I APOPTOSIS I FIG. 2. Representation of several pathways through which the immune attack against p-cells can lead to either death by necrosis or by apoptosis. Mg, macrophage; Th,, Thl helper lymphocyte; Tc, cyto- toxic T-cell.

ing the mechanisms of 6-cell destruction in IDDM. According to one model, p-cells are destroyed by cytotoxic T-cells using perforin or gramymes as effector molecules. Perforin causes lysis of the target cell, whereas granzymes A and B mainly cause apoptosis (29). Another model (30,31) (Fig. 2), hypoth- esizes that pcell antigen-specific T-helper cells, activated to transcribe cytokine genes, induce the build-up of a specific and nonspecific mononuclear cell infiltrate and activates endothelid cells, which express adhesion molecules and lib- erate mikimmatory mediators. Recruited macrophages are stimulated by interferon (1FN)-y to produce ILlp and TNF- a, which in synergy with IFN-y lead to p-cell toxicity via p- cell-specific induction of iNOS and apoptosis-activating pathways. Further, the model implies that IL1 leads to p-cell Fas expression, rendering the p-cells susceptible to lysis by Thl and cytotoxic T-cells expressing FasL. Nitric oxide and p-cell apoptosis: can NO be the answer? Several pathways have been suggested to be responsible for nitric oxide (NO)-mediated 6-cell death (Fig. 2). Among these, NO inactivates Krebs cycle aconitase by nitrosylation of Fe-S groups thereby preventing glucose oxi- dation and ATP generation, leading to cell death by necrosis (32). Further, NO has been shown to damage cellular DNA through the induction of DNA strand breaks (33,34). DNA sin- gle-strand breakage and necrotic cell death were observed after exposure of rat islets to sodium nitropmside or activated macrophages (34). DNA strand breaks may cause p-cell necrosis by itself or by the activation of DNA repair mecha- nisms (including the induction of the enzyme poly(ADP ribose) polymerase), which can cause p-cell death through depletion of cellular nicotinamide adenine dinucleotide (35). However, evidence is accumulating suggesting a third pathway by which the p-cell may undergo apoptosis through the expo sure to certain cytokines (30). It is known that cytokine-medi- ated NO synthesis may cause apoptosis in macrophages and fibroblasts (36). Pancreatic pcells have been shown to be sus- ceptible to apoptotic death (i.e., apoptosis is involved in the

normal reduction in the p-cell mass post partum [37] or in dynamic changes in islets in the posttransplantation period [38]). Apoptosis in p-cells can be induced through several pathways, such as protein kinase C inhibition (39), guanine triphosphate binding protein activation (40), and streptozo tocin (41,42). Also, I L l p alone (43,44) or in combination with TNF4FN-y (45) induces apoptosis in transformed or normal islet pcells. Further, DNA fragmentation, which is charac- teristic of apoptotic cell death, has been observed as an early event in IL1 p-induced islet p-cell destruction (46). The com- bination of a l p , TNFa, and IFN-y induces oligonucleosomal DNA fragmentation typical of apoptosis in rat islets (46). A recent study has shown that ILlp induces an increase in the ceramide content of RINm5F cells, and ceramide has been shown to be implicated in the induction of apoptosis (47). After being released by the hydrolysis of membrane sphin- golipids (mainly sphingomyelin), ceramide acts as a signaling molecule by activating stress-activated protein kinases and phospholipase A2 leading to formation of prostaglandin E,. Moreover, the involvement of the ceramide-dependent sig- naling system in the regulation of programmed cell death by cytokines, such as TNF-a has been demonstrated (48). Addi- tionally, the other transcriptional factors c-jun and c-fos, which are also implied in the sign- cascade leading to a p o ptosis, are induced by ILL in rat islets of Langerhans and RINm5F cells (49). In mouse pancreatic pcell lines (PTCl and NIT-I), the combination of three cytokines (ILlp, TNF-a, and IFN-y) was able to induce apoptotic cell death (50). Inter- estingly, this study also showed that overexpression of bcl-2 in PTCl cells partially protected them from cytokine-induced death. More recently, it has been shown that a combination of cytokines (50 Ulml I L l p + 1,000 Ulml TNFa + 1,000 U/ml IFN-y) is able to induce DNA strand breaks as well as an increase in apoptotic cells in human islets, with no signs of additional necrosis (51). These results suggest that apoptosis may be responsible for cytokine-induced cell death.

NO production seems to be an important signal for p-cell apoptosis. Thus, it has been shown that RINm5F cells undergo DNA fragmentation, nuclear condensation, and apoptotic formation in response to ILlp (52). These events were preceded by NO production, and inhibition of iNOS prevented apoptotic cell death. The same group described that both chemical NO donors and ILlp-induced endogenously delivered NO caused the expression of the apoptosis-linked tumor suppressor gene p53 in this cell line while it was absent in untreated cells (53). These findings indicate that the protein p53, which is an apoptosis inducer, as mentioned above, may be involved in the process of NO-induced apo- ptosis. Further studies demonstrated that both exogenous NO (released by streptozotocin) and ILlp-induced NO was associated with typical apoptotic changes in isolated rat islets and in the tumor-derived p-cell line HIT (54). Also, the use of NO donor S-nitrosoglutathione produced apoptosis in HIT-T15 cells, an effect that was reproduced by cGMP analogs, suggesting that apoptotic cell death induced by NO is secondary to rise in cGMP levels, thus involving activation of cGMP-dependent protein kinase (55). A significant increase in the percentage of apoptotic cells in rat islets was seen after in vitro exposure to IL-1p (44). Further, these authors were able to correlate the degree of NO production with the ability of IGlp to induce programmed cell death and to reduce apoptosis by inhibiting NO production.

DIABETES, VOL. 47, OCTOBER 1998

However, some observations suggest that NO production is neither necessary nor sufficient for the mediation of cytokine- induced p-cell destruction. Thus, a previous study failed to observe DNA fragmentation, mitochondrial damage, or cell destruction in RIN cells in spite of ILlp-induced NO pro- duction (56). Neither the reduction of DNA and insulin con- tent in human islet single cell cultures (57), nor the cytokine- induced reduction of insulin content and secretion of human islets (58) could be prevented by the inhibition of cytokine- inducible NO synthase. Further, nicotinamide, which is effec- tive in preventing cytokine-induced p-cell destruction, did not completely block NO production in rodent islet cultures (59,60). Moreover, isolated human islets are more resistant to the suppressive effects of I L l p and NO than isolated rodent islets (61). Cytokine-induced apoptosis in human islets is probably independent of NO generation (51). In vivo, inhibi- tion of iNOS causes only a delay but not protection against dia- betes development (62-64). Therefore, additional mediators of cytokine-induced pcell death may be necessary, such as the facilitation of other apoptosis-inducing pathways, which include the mitogen-activated protein (MAP) kinase cascade and the FasIFasL system (65) (Fig. 2). Apoptosis and Fas expression by P-cells. Recently, I L 1 p-induced Fas expression in thyrocytes was found to be the mechanism responsible for thyrocyte apoptosis in Hashimoto's thyroiditis (66). Interestingly, in purified human p-cells, only IL-1p (50-100 Ulml) of several different cytokines was able to induce the expression of Fas after 24 h (67). Further, induction of Fas in pcells by infiltrating Tcells produces Pcell apoptosis, which is a predominant mechanism of this T-cell-mediated destruction of insulin-producing cells (1 1). Therefore, data provided by these studies point to Fas- mediated apoptosis as a central pathway leading to p-cell destruction in IDDM. In vivo detection of apoptosis in IDDM pathogenesis. Evidence for a role of apoptosis in p-cell destruction in ani- mal models of IDDM is accumulating. However, the removal of apoptotic cells is so efficient that it may prove extremely difficult to demonstrate in vivo. This efficient engulfment of apoptotic cells is usually carried out by the monocyte-phago- cyte system, although normal epithelial cells, vascular endothelium, or tumor cells can also do so (68). In addition, the progression of the apoptotic process is so fast that it makes it difficult to observe, and more importantly to quan- titate (69). For example, in the young mouse thymus, in which more than one third of the cells die each day, it is almost impossible to find an apoptotic cell in an average tis- sue section (5). Furthermore, the rapid clearance of dead cells may be an additional difficulty in the study of apoptosis in animal IDDM models such a5 the NOD mouse (70).

Despite these methodological difficulties, a recent study in the well-known multiple low-dose streptozotocin model of IDDM was able to demonstrate that apoptosis of mouse P- cells was identified as the mode of death of these cells (71). These authors suggested that in other models of IDDM, or even in human IDDM, apoptosis may require careful and extensive histological studies because the rate at which pcell destruction takes place is much slower than that in the low- dose streptozotocin model.

In BDC2.5mOD.scid mice, an accelerated model of dia- betes, which produce CD4+ T-cells bearing a transgenic T- cell receptor but are devoid of CD8+ T-cells, histological and

immunochemical analysis of pancreases showed clear evi- dence of apoptosis of pancreatic p-cells, without any evidence of p-cell necrosis in infiltrated islets (70). Further, a strong cor- relation between the degree of insulitis and p-cell apoptosis was demonstrated. From the kinetics of apoptosis calculated by these authors, it is not surprising that apoptosis was not observed before in nontransgenic NOD mice. Interestingly, a thorough histological and immunochemical study was able to confirm that apoptosis is the main mode of death of pancre- atic p-cells in the NOD mouse (72). An outstanding findmg of this study is that the initiation of the apoptotic process pre- ceeds the light microscopic lymphocytic infiltration of the islets, thus suggesting that initial p-cell death is mediated by resident macrophages and their products, mainly cytokines, in accordance with the second model described above. Autoantibodies and p-cell apoptosis? Although the autoimmune p-cell destruction in IDDM is not antibody- mediated, apoptosis has been shown to occur in normal pancreatic p-cells and RINm5F cells after exposure to serum from patients with IDDM, an effect that was found to be IgM-mediated (73). As this effect of diabetic sera was prevented by addition of a blocker of calcium channels, the authors hypothesize that an IgM-mediated increase in cal- cium influx may be part of the autoimmune damage of p-cells. However, the interpretation of this finding and the contri- bution to the pathogenesis of IDDM remains to be clarified.

FUTURE STUDIES

From the data reviewed above it is clear that the process of apoptosis as a central mechanism of p-cell death in IDDM deserves further study. The area of apoptosis research is one of the most fertile current fields in cell biology, and in-depth insight into modulation of the molecular mechanisms of apo- ptosis has been reached. Preventive intervention may be aimed at the level of immune cells and/or the level of the tar- get cell. Methods that induce selective apoptosis of the immune cell as a target may be considered. Future strategies may explore methods to induce selective apoptosis of autoreactive cells that mediate the destruction of p-cells. As for other autoimmune diseases (4), it has been shown that repetitive treatment with antigen, such as insulin in IDDM, can result in selective death of reactive lymphocytes in vivo. However, the exact mechanism by which such treatment induces apoptosis is unclear, though priming of cell death through a Fas-mediated mechanism has been suggested (74).

The advancement in the knowledge of apoptosis of p- cells is likely to progress quickly in the near future. If a p e ptosis is the main common mode by which p-cells die in response to immune attack by cytokines and/or T-cells in IDDM, it may be possible to develop novel strategies to pre- vent this mode of p-cell death, thereby preventing or delay- ing the onset of the disease.

It has been clearly established that protection from pro- grammed cell death may be exerted at four different levels: interception of an apoptosis-inducing stimulus; functional antagonism of an otherwise apoptosis-inducing trigger; inter- ference with signal transduction cascades, or blockade of catabolic enzymes participating in cellular suicide (75). These levels of intervention may apply in the design of strate- gies of prevention of p-cell death. Moreover, the existence of processes that may involve apoptosis at different levels of the pathogenic process of IDDM offers possibilities in the pre-

DIABETES, VOL. 47, OCTOBER 1998

vention of the hsease. Treatments that increase the apoptotic threshold of pancreatic p-cells may be beneficial. For exam- ple, some methods can include the blockade of the binding of death ligands (e.g., TNF and FasL). Strategies may be directed to increase resistance to apoptotic stimuli by enhancing the expression of Bcl-2. Insulin may have poten- tial therapeutic applications in the prevention of IDDM (cur- rently under clinical evaluation) due to its potential immune modulating effect and its capacity to induce p-cell rest. It should be investigated if insulin, because of its growth fac- tor-like effect, has antiapoptotic effects on P-cells, further contributing to its efficacy in preventing the disease (76).

Cell specificity of therapeutic approaches is critical since IDDM, as most autoimmune diseases, is not characterized by a generalized increase in the susceptibility or resistance to apoptosis. The potential harm of using a therapy that enhances the survival of cells at the expense of increasing the risk of other autoimmune disorders or enhancing tumor p r o gression through the prevention of apoptotic cell death should be pointed out (4). However, even central elements of the apoptotic process may be pharmacologically manipu- lated in a cell-specific fashion. The manipulation of bcl-2, p53, or c-myc may change the susceptibility of cells to pro- grammed cell death (75). Tissues can significantly vary the expression of certain members of the bcl-2 or ICE families. Moreover, specific inhibitors of individual members of the ICE family exist. Therefore, systemic modulation of bcb-2 or ICE may not be as toxic as might have been expected. There is a large array of factors that may be potentially useful in the induction or inhibition of programmed cell death in lympho- cytes and other target cells (75). Although in its infancy, a potential new therapeutic field is open for the study of pre- ventive tools in IDDM.

REFERENCES 1. Cohen JJ: Programmed cell death in the immune system. Adti Immunol

50:5.545, 1991 2. Steller H: Mechanisms and genes of cellular suicide. Science 267314451449,

1995 3. Hale AJ, Smith CA, Sutherland LC, Stoneman VEA, Longthome VL, Culhane

AC, Williams Gl? Apoptosis: molecular regulation of death. Eur J Biochmn 236: 1-26, 1996

4. Thompson CB. Apoptosis in the pathogenesis and treatment of disease. Sci- eru-e 267:1456~1462, 1995

5. Cohen.JJ: Apoptosis: physiologic cell death. JClin Lab Med 124761-765,1994 6. Martin SJ: Apoptosis: suicide, execution or murder? Trends Cell Biol

9:141-144, 1993 7. Cohen JJ: Overview: mechanisms of apoptosis. Immunol Today 14:12&130,

1993 8. Vaux DL, Agda HL, Weissman IL: Bcl-2 prevents death of factordeprived cells

but fails to prevent apoptosis in targets of cell mediated killing. In l I7nmunol 4:821-824, 1992

9. Castaiio L, Eisenbarth GS: m e 1 diabetes: a chronic autoimmune disease of human, mouse, and rat. Annu Ref) Immunol8:647-679, 1990

10. Cohen JJ, Duke RC, Fadok VA, Sellins KS: Apoptosis and programmed cen death in immunity. Ann Rev Itnnunol 10267-293, 1992

11. Chewonsky AV, Wang Y, Wong FS, Visintin I, FlaveU RA, Janeway CA, Matis LA: The role of Fas in autoimmune diabetes. Cell 89:17-24,1997

12. Watanabe-Fukunaga R, Brannan C1, Itoh N, Yonehara S, Copeland NG, Jenk- ins NA, Nagata S: The cDNA s t ~ t u r e . expression and chromosomal assign- ment of the mouse Fas antigen. J Im?r~unol148:1274-1279, 1992

13. Nagata S, Golstein P: The Fas death factor. Science 267:1449-1456, 1995 14. Ran~sdell F, Seaman MS, Miller RE, Picha KS, Kennedy MK, Lynch DH: Dif-

ferential ability of Thl and Th2 T cells to express Fas ligand and to undergo activation-induced c ~ l l death. In t Immunol6: 1545-1553, 1994

15. Ju ST, Cui H, Panka DJ, Ettinger R, Marshak-Rothstein A: Participation of tar- get Fas protein in apoptosis pathway induced by CD4+ Thl and CD8+ cyte toxic T-cells. B o c Nail Acad Sci USA 91:418E4189, 1994

16. Rocha B, von Boehrner H: Peripheral selection of the Tcell repertoire. Science 251:1225-1228,1991

17. Kabelitz D, Pohl T, Pechhold K: Activation-induced cell death (apoptosis) of mature peripheral T lymphocytes. Imn~ujlol Today 1433S339, 1993

18. Klas C, Debatin KM, Jonker RR, Krarner P H Activation interferes with APO- 1 pathway in mature human Tcells. Int Imrnu no1 5:625-630,1993

19. Chu JL, Ramos P, Rosendrff A, Nikolic-Zugic J , Lacy E, Matsuzawa A, Elkon KB. Massive up-regulation of the Fas ligand in Ipr and gld mice: implications for Fas regulation and the graft-versus-host disease-like wasting syndrome. J E.xp Med 181:39>398, 1995

20. Takahashi T, Tanaka M, Brannan CI, Jenkins NA, Copeland NG, Suda T, Nagata S: Generalized lymphoproliferative disease in mice, caused by a point mutation in the Fas ligand. Cell 76969-976, 1994

21. Watanabe-Fukunaga R, Brannan C1, Copeland NG, Jenkins NA, Nagata S: Lyrnphoprolieration disorder in mice explained by defects in Fas antigen that mediates apoptosis. Nature 356:314-317, 1992

22. Garchon HJ, Bedossa P, Eloy L, Bach JF: Identification and mapping to c h r e mosome 1 of a susceptibility locus for periinsulitis in non-obese diabetic ndce. Nature 353:262-265, 1991

23. Leijon K, Hammarstrom B, Holmberg D: The non-obese diabetic (NOD) mouse displays enhanced immune responses and prolonged survival of lym phoid cells. Int Irnmu?rol6:339-315, 1994

24. Garchon HJ, Lual JJ, Eloy L, Bedossa P, Bach JF: Genetic analysis of immune dysfunction in non-obese diabetic (NOD) mice: mapping of a susceptibility locus close to the Bcl-2 gene correlates with increased resistance of NOD T cells to apoptosis induction. E u r J Immzrnol24:3803&1, 1994

25. Penha-Gonqalves C, Leiion K, Persson L, Holmberg D: Type 1 diabetes and the control of dexamethasone-induced apoptosis in mice maps to the same region on chromosome 6. Genomics 28:398404, 1995

26. Colucci F, Cilio CM, Lejon K, Penha-Gon~alves C, Bergman ML, Holmberg D: Programmed cell death in the pathogenesis of murine IDDM. resistance to a p e ptosis induced in lymphocytes by cyclophospharnide. J Autoimmzrtt 9271-276,1996

27. Lamhanedi-Cherradi SE, Luan JJ, Eloy L, Fluteau G, Bach JE: Garchon Resistance of T-cells to apoptosis in autoimmune diabetic (NOD) ndce is increased early in l i e and is associated with dysregulation of Bcl-x. Dia- betologia 41:17EL184,1998

28. Giordano C, De Maria R, Stassi G, Todaro M, Richiusa P, Giordano M, Testi R, Galluzzo A: Defective expression of the apoptosis-inducing CD95 (FadAF'O- 1) molecule on T- and B-cells in IDDM. Diabetologia 38:1449-1454, 1995

29. Liu CC, Young LHY, Young JD: Lymphocyte-mediated cytolysis and disease. N Engl J Merl335:1651-1659,1996

30. Mandrup-Poulsen l? The role of interleukin-1 in the pathogenesis of IDDM. Diabetologia 39:1005-1029, 1996

31. Nerup J , Mandrup-Poulsen T, Helqvist S, Andersen HU, Pociot F, Reinlers JI, Cuartero BG, Karlsen AE, Bjerre U, Lorenzen l? On the pathogenesis of IDDM. Diabetologia 37:S82-S89, 1994

32. Welsh N, Eizirik DL, Bendtzen K, Sandler S: Interleukin-lp-induced nitric oxide production in isolated rat pancreatic islets requires gene transcription and may lead to inhibition of the Krebs cycle enzyme aconitase. Endocrit~ol- ogy 129:3167-3173, 1991

33. Delaney CA, Green MHL, Lowe JE, Green IC: Endogenous nitric oxide induced by interleukin-lp in rat islets of Langerhans and HIT-TI5 cells causes significant DNA damage as measured by the "comet" assay. FEBS Lett 333291-295, 1993

34. Fehsel K, Jalowy A, Qi S, Burkart V, Hartmann B, Kolb H Islet cell DNA is a target of inflanunatory attack by nitric oxide. Diabetes 42:49&500,1993

35. Heller B, Wang ZQ, Wagner EF, Radons J, Burkle J, Fehsel K, Burkart V, Kolb H: Inactivation of the poly (ADP-ribose) polymerase gene affects oxy- gen radical and nitric oxide toxicity in islet cells. J Biol Chmn 270:1117&11180, 1995

36. Xie K, Huang S, Zhongyun D, Fidler U: Cytokine-induced apoptosis in t rans fonned murine fibroblasts involves synthesis of endogenous nitric oxide. Int J Oncol3:1043-1048, 1993

97. Scaglia L, Smith FE, Bonner-Weir S: Apoptosis contributes to the involution of p-cell mass in the post parhlm rat pancreas. E7ulorrir~ology 1365461-5468, 1995

38. Davalli AM, Scaglia L, Zangen DH, Hollister J, Bonner-Weir S, Weir GC: Vul- nerability of islets in the immediate posttransplantation period: dynamic changes in structure and function. Diabetes 45:1161-1167, 1996

39. EGnchez-Mqalet V, Lucas M, Solano F, Goberna R: Sensitivity of insulin-secret- ing RIN m5F cells to undergoing apoptosis by the protein kinase inhibitor stau- rosporine. E q CeU Res 209:160-163, 1993

40. Loweth AC, Scarpello JHB, FVilliams GT, Morgan NG: Effects of modulators of G-protein function on apoptosis in RlNm5F cells and rat islets (Abstract). Diabetologia 38 (Suppl. 1):A45, 1995

DIABETES, VOL. 47, OCTOBER 1998

41. Ankarcrona M, Dypbukt JM, Briine B, Nicotera P: Interleukin-lp-induced nitric oxide production activates apoptosis in pancreatic RINm5F cells. Ezp Cell Res 213: 172-177,1994

42. Morgan NG, Cable HC, Newcombe NR, Williams GT: Treatment of cultured pancreatic B c e k with streptozotocin induces cell death by apoptosis. Biosci Rep 14:24.%250, 1994

43. Kaneto A, Schlosser M, Ziegler B, Schmidt S: Interleukin-1-beta induces a p e ptosis in insulin-producing ceUs (Abstract). Diabetologia 38 (Suppl. l):A38,1995

44. Dunger A, Augstein P, Schmidt S, Fischer U: Identification of interleukin 1- induced apoptosis in rat islets using in situ specific labelling of fragmented DNA. JAu,toin~mun 9309413,1996

45. Dungerh Schlosser M, ZieglerB, Schmidt S: Interleukin 1-betainduces apoptosis in insulin-producing cells (Abstract). Dia,betologia 38 (Suppl. 1):A38,1995

46. Rabir~ovitch A, Suarez-Pinzon WL, Shi E: Morgan AR, Bleackley RC: DNA fragmentation is an early event in cytokine-induced islet beta-cell destruction. Diabstologia 37:73%738, 1994

47. Welsh N: Interleukin-lp-induced ceramide and diacylglycerol generation may lead to activation of the c-jun NH,-terminal kinase and the tran- scription factor ATF'2 in the insulin-producing cell line RINm5F. J Biol Cftern 27153307-8312, 1996

48. JarvL4 WD, Kolesnick RN, Fomari FA, Traylor RS, Gewirtz DA, Grant S: Induc- tion of apoptotic DNA danlage and cell death by activation of the sphin- gomyelin pathway. Proc Natl Acad Sri U S A 91:7%77. 1994

49. Kwori G, Corbett JA, McDanieI ML: Interleukin 1-induced fos and jun do not regulate inducible nitric oxide sylthase in rat islets of Langerhans and RINm5F cells. Endocrinology 137:825830, 1996

50. Iwahashi H, Hanafusa T, Eguchi Y, Nakqjima J, Itoh N, Tomita K, Namba M, Kuwqjima M, Noguchi T, Tsujimoto Y, Matsuzawa Y: Cytokine-induced apop totic cell death in a mouse pancreatic beta-cell line: inhibition by Bcl-2. Dia- betobgia 3953&536,1996

51. Delaney CA, Pavlovic D, Hoorens A, Pipeleers DG, Eizirik DL: Cytokines induce deoxyribonucleic acid slmnd breaks and apoptosis in human pancreatic islet cells. Endocrinology 138:261&2614, 1997

52. Ankarcrona M, Dypbukt JM, Briine B, Nicotera P: Interleukin-lp-iiduced nitric oxide production activates apoptosis in pancreatic RINm5F cells. Ezp CeU Res 212:172-177,1994

53. Mess~ner UK, Ankarcrona M, Nicotera P, Briine B: p53 Expression in nitric oxide--induced apoptosis. FEBS Lett 355:2%26, 1994

54. Kaneto H, Fujii J, Seo HG, Suzuki K, Matsuoka T, Nakamura M, Tatsumi H, Yamasaki Y, Kamada T, Taniguchi N. Apoptotic cell death triggered by nitric oxide in pancreatic p-cells. Diabetes M733-738,1995

55. Loweth AC, Williams GT, Scarpello JHB, Morgan NG: Evidence for the involvement of cGMP and protein kinase G in nitric oxide-induced apoptosis in the pancreatic p-cell line, HIT-T15. FEBS Lett 400:2&288,1997

56. Suarez-Pinzon WL, Strynadka K, Schulz R, Rabinovitch A: Mechanisms of cyotldne-induced destruction of rat insulinoma celIs: the role of nitric oxide. Endocrinology 134: 100G10 10, 1994

57. Rabinovitch A, Suarez-Pinzon WL, Strynadka K, Schulz R, Lakey JR, Warnock GL, Rajotte RV: Human pancreatic islet p-cell destruction by cytokines is independent of nitric oxide production. J Clin Endocrinol Metob 79:1058-1062, 1994

58. Eizirik DL, Sandler S, Welsh N, Cetkovic-Cvrlje M, Nieman A, Geller DA, Pipeleers DG, Bendtzen K, Hellerstrom C: Cytokines supress human islet function irrespective of their effects on nitric oxide generation. J Clin hmest

93:1968-1974,1994 59. Cetkovic-Cvrue M, Sandler S, Eizirik DL: Nicotinamide and dexamethasone

inhibit interleub-1-induced nitric oxide production by RINm5F cells with- out decreasing messenger ribonucleic acid expression for nitric oxide s y r thase. Endoninology 133:1739-1743, 1993

60. Andersen HU, Jmrgensen KH, Egeberg J, Mandrup-Poulsen T. Nerup J: Nico- tinamide prevents interleukin-1 effects on accun~i~lated insulin release and nitric oxide production in rat islets of Langerhans. Diabetes 43770-777,1994

61. Eizirik DL. Delaney CA, Green MHL, Cunninghan~ JM, Thorpe JR, Pipeleers DG, Hellerstrom C, Green IC: Nitric oxide donors decrease the function and survival of human pancreatic islets. Mol Cell Endoo.inol118:71-83, 1996

62. Lukic ML, Stosic-Gwicic S, Ostojic N, Chan WL, Liew FY: Inhibition of nitric oxide generation affects the induction of diabetes by streptozotocin in mice. Biochem Biophys Res Co7ninun 178:91.%920,1991

63. Corbett JA, Mikhael A, Shimizu J, Frederick K, Misko TP, McDaniel ML, Kaganawa 0, Llnanue ER: Nitric oxide production in islets from nonobese dia- betic mice: aminoguanidine-sensitive and -resistant stages in the immune logical diabetic process. Proc Na,tl Aca,d Sci LISA 90:8992-8995, 1993

64. Lindsay RM, Smith W, Rossiter SP, McIntyre MA, W~llians BC, Baird JD: N';- Nitm-LArginine methyl ester reduces the incidence of IDDM in BB/E rats. Dia- betes M36.5368, 1995

65. Yamada K, Takane-Gyotoku N, Yuan X, Ichikawa F, Inada C. Nonaka K: Mouse islet cell lysis mediated by interleukin-1-induced Fas. Dinbetologia 39:130&1312, 1996

66. Giordano C, Stassi G, De Maria R, Todaro M, Richiusa P, Papoff G, Ruberti G, Bagnasco M, Testi R, Galluzzo A: Potential involvement of Fas and its ligand in the pathogenesis of Hashimoto's thyroiditis. Science 275:96&963,1997

67. Stassi G, Todaro M, Richiusa P, Giordano M, Mattina A, Sbriglia MS, Lo Monte A, Buscemi G, Galluzzo A, Giordano C: Expression of apoptosis-inducing CD95 (FasIApo-1) on human p-cells sorted by flow-cytometry and cultwed in vitro. Transplant Prop 27:3271-3275, 1995

68. WylIie AH, Kerr JFR, Cunie AR: Cell death: the significance of apoptosis. Int Reu Cytol68:251306, 1980

69. Schwartzman RA, Cidlowski JA. Apoptosis: the biochemistry and molecular biology of programmed cell death. Endocr Reo 14:133-151, 1993

70. Kurrer MO, Pakala SV, Hanson HL, Katz JD: pcell apoptosis in Tcell-mediated autoimmune diabetes. Proc Natl Acad Sci U S A 94:21%218, 1997

71. O'Brien BA, Harmon BV, Canleron DP, Allan DJ: Betacell apoptosis is respon- sible for the development of IDDM in the multiple low-dose streptozotocin model. JPathol178:17&181, 1996

72. O'Brien BA, Harmon BV, Cameron DP, N a n DJ: Apoptosis is the mode of p- celI death responsible for the development of IDDM in the nonobese diabetic (NOD) mouse. Diabetes 46:75&757, 1997

73. Juntti-Berggren L, Lmson 0, Rorsman P, h l d a C, Bokvist K, Wahlander K, Nicotera P, Dypbukt J, Orrenius S, Hallberg A, Bergten PO: Increased activ- ity of Ltype Ca2+ channels exposed to serum from patients with type I diabetes. Science 261:8&90, 1993

74. Singer GG, Abbas AK: The Fas antigen is involved in peripheral but not thymic deletion of T-lymphocytes in TcelI receptor transgenic mice. Immu- ni ty 1:36&371, 1994

75. Kroemer G, Martinez AC: Pharmacological inhibition of progran~med lym- phocyte death. hn?nunol Today 15:23&242, 1994

76. Gladstone F', Nepom G'l? The prevention of IDDM: iqjecting insulin into the cytokine network (Letter). Diabetes 44:859462,1995

DIABETES, VOL. 47, OCTOBER 1998