JOURNAL OF PATHOLOGY, VOL. 178: 176-181 (1996)

BETA-CELL APOPTOSIS IS RESPONSIBLE FOR THE DEVELOPMENT OF IDDM IN THE MULTIPLE

LOW-DOSE STREPTOZOTOCIN MODEL

BRONWYN A . O'BRIEN, BRIAN V. HARMON, DONALD P. CAMERON*, AND DAVID J . ALLAN

School o f Life Science, Queensland University of Technology and *Department of Diabetes, Princess Alexandru Hospital, Brisbane, Australia

SUMMARY Although insulin-dependent diabetes mellitus (IDDM) results from irreversible loss of beta cells, the mode of cell death responsible for

this loss has not previously been categorized. In this study, the multiple low-dose streptozotocin (stz) model (intraperitoneal injection of stz a t a concentration of 40 mglkg body weight per day for five consecutive days) was used to investigate beta-cell death during the development of IDDM in male C57B116 mice. Apoptotic cells were evident by light microscopy within the islets of Langerhans of treated animals from day 2 (the day of the second stz injection) until day 17. Immunohistochemical localization of insulin to the dying cells confirmed the beta-cell origin of the apoptosis. Two peaks in the incidence of beta-cell apoptosis occurred: the first a t day 5, which corresponded to an increase in blood glucose concentration, and the second a t day 11, when lymphocytic infiltration of the islets (insulitis) was maximal. Insulitis did not begin until day 9, by which time treated animals had developed overt diabetes as revealed by blood glucose and pancreatic immunoreactive insulin (IRI) measurements. Beta-cell apoptosis preceded the appearance of T-cells in the islets and continued throughout the period of insulitis. Thus, whether induced by stz or a subsequent immune response, apoptosis is the mode of cell death responsible for beta-cell loss in the multiple low-dose stz model of IDDM.

KEY WORDS-apoptosis; streptozotocin; IDDM; mouse

INTRODUCTION

Insulin-dependent diabetes mellitus (IDDM) is a com- plex disorder with both genetic and environmental fac- tors playing a role in its development."2 When IDDM patients first present with clinical features of the disease, most of the beta-cell population of their islets of Langerhans have usually already died. The surviving small number of beta cells may initially function well, providing a honeymoon period for the patient during which exogenous insulin is not required.354 However, these beta cells inevitably die and the patient is then irreversibly dependent on exogenous insulin.

Immune mechanism^,^ virus expression,6 and environ- mental toxins7 have all been implicated as possible causative agents for the development of IDDM. Whether the beta-cell death leading to IDDM takes the form of apoptosis or necrosis has not yet been identified. Whereas necrosis is a passive degenerative process, apoptosis is a gene-directed mechanism of self- destruction which can be switched on (or off) in cells by endogenous factors, such as differential oncogene expression, or exogenous factors, including growth fac- tors and cytokines.' A variety of stimuli such as immune cell killing,' DNA damage," growth factor depri- vation,' ligation of surface receptors,I2 heat shock,I3 environmental toxins,I4 and viral i n f e ~ t i o n ' ~ have been shown to be capable of inducing apoptosis. If apoptosis of beta cells plays a critical role in the pathogenesis of IDDM, it may be possible to develop novel strategies to

Addressee for correspondence: Dr David J. Allan, School of Life Science, Queensland University of Technology, Brisbane, 4001 Australia.

CCC 0022-341 7/96/020 176-06 1996 by John Wiley & Sons, Ltd.

modify beta-cell death, thereby preventing the onset of IDDM or lessening its clinical severity.

The multiple low-dose streptozotocin (stz) mouse model mimics, in some basic aspects, recent-onset IDDM in human Injection of stz in five equal low doses (40 mglkg body weight per day) induces a slow progressive hyperglycaemia, accompanied by lymphocytic infiltration of the pancreatic islets." In this study, we have established that apoptosis is the mode of cell death responsible for loss of beta cells in the multiple low-dose stz model and have correlated its occurrence with the onset of both insulitis and diabetes.

MATERIALS AND METHODS Animals

Five- to six-week-old male C57B1/6 mice were obtained from the Central Animal Breeding House (University of Queensland, Brisbane, Australia) and maintained on a diet of standard mouse chow with water allowed ad libitum. Twelve-hour light and dark regimes were implemented.

Induction of diabetes Stz (Calbiochem, CA, U.S.A.) was dissolved in sterile

0.1 M sodium citrate buffer (pH 4.5) just prior to use and injected intraperitoneally into mice at a concentration of 40 mglkg body weight. Control animals were injected with an equivalent volume of citrate buffer alone. Experimental and control mice were injected daily (fol- lowing an overnight fast) for five consecutive days. Mice were killed by carbon dioxide overdose on days 1 (day of

Received 25 January 1995 Accepted 26 June 1995

BETA-CELL APOPTOSIS IN IDDM 177

the first injection), 2, 3, 4, 5, 7, 9, 11, 13, 15, and 17, at approximately the same time each day, such that mice from the experimental groups for days 1-5 were killed 2 h after the injection of stz. Experimental and control groups comprised six animals each.

Plasma glucose and pancreatic immunoreactive insulin (IRI) determinations

Non-fasting blood samples were collected via tail bleed into heparinized tubes. Plasma was assayed for glucose concentration using a YSI glucose analyser. All determinations of glucose concentration were done in duplicate.

Immediately after carbon dioxide euthanasia, the pancreas from each animal was removed and bisected longitudinally. One half was homogenized and the insulin was extracted in acid alcohol for subsequent determination by radioimunoassay. Results were expressed as the mean f standard error of the mean (SEM) of groups at each time point.

Histology The remaining half of each pancreas was immersion-

fixed in neutral buffered formalin for 24 h and processed and embedded in paraplast wax using standard histo- logical techniques. Serial sections (4pm) were cut and every fourth section was stained with haematoxylin and eosin (H & E). Adjacent sections were used for immunohistochemical studies as detailed below.

Immunohistochemistry Immunohistochemical localization of insulin by a

polyclonal anti-insulin antibody (Dako) was used to identify beta cells. Cellular localization was determined in paraffin sections by an indirect labelled avidin-biotin- peroxidase method with diaminobenzidine as the sub- strate. Immunohistochemical identification of T-cells was performed similarly using a monoclonal anti-CD3 antibody (Dako).

QuantiJication Apoptotic counts-Apoptosis was quantified in all

islets present in ten H & E-stained sections from each animal, each section being separated from the next by at least 40pm. A total of approximately 3000 islet cells were counted at each time point. Cells were identified as apoptotic on the basis of their morphology, using pre- viously defined characteristics. l 9 Thus, total apoptotic cell counts included cells showing the early apoptotic changes (condensation and margination of nuclear chro- matin) through to the later stages of membrane-bound cell fragments containing uniformly dense masses of nuclear chromatin (apoptotic bodies). Cells with a small nucleus and minimal cytoplasm in H & E-stained sec- tions were considered to be T-cells and this was con- firmed by immunohistochemical labelling of adjacent sections for CD3. The number of apoptotic cells present was expressed as a percentage of the total number of

islet cells. T-cells were excluded from this calculation as they did not constitute part of the original islet cell population. The percentage of islets containing one or more apoptotic cells was also calculated. Results were expressed as mean values of all animals at each time point * SEM between animals.

Percentage of total pancreatic area occupied by islets-The total pancreatic tissue area in each of the sections used for the quantification of apoptosis was estimated by image analysis (Prism Image Analysis Package 2.5.1, Macintosh Quadra 700). Total islet area was estimated similarly and expressed as a percentage of the total tissue area. Results were represented as mean values of all animals at each time point & SEM between animals.

Statistical analysis-Data from control and treated groups were analysed using a one-way analysis of vari- ance test and variation between treated groups was subject to a multiple comparison.

RESULTS

Apoptotic cells were present within the islets of stz- treated animals from day 2 (the day of the second stz injection) until day 17, the final day of the experiment (Fig. 1). Apoptotic cells showed a decrease in cellular volume, eosinophilic cytoplasm, and characteristic con- densation of nuclear chromatin into sharply demarcated masses abutting the nuclear envelope. Fragmentation of affected cells into small apoptotic bodies was also frequently noted. Overt diabetes (blood glucose 21 f 3.5 mmol/l and pancreatic IRI 1.6 f 0.4 units/g for treated animals as opposed to 6 & 1.5 mmol/l and 3.2 j~ 0.4 units/g, respectively, for controls) was not established until day 9 (Fig. 2). Lymphocytic infiltration of the islets of treated animals began at day 9 (7 days after the onset of apoptosis) when the islet area had already fallen to only 35 per cent of control values (Fig. 3). Insulitis peaked at day 11 and was not observed after day 15.

Although apoptotic counts exceeded control values from days 2 to 17, the increase was only statistically significant between days 5 and 13 inclusive. Apoptosis showed two separate peaks, the first at day 5 and the second at day 11 (Fig. 4). The occurrence of two separate peaks was confirmed by statistical analysis. The incidence of apoptotic beta cells reached a maximum at day 5 and corresponded to an increase in mean blood glucose value and a decrease in both pancreatic IRI and islet area. The second peak occurred when insuiitis was maximal. There was a good deal of variation in the number of islets containing apoptotic cells, with many islets containing no apoptosis at all (Fig. 5).

Immunohistochemical localization of insulin to the dying cells confirmed the beta-cell origin of the apop- totic cells (Fig. 6). Immunohistochemical staining for the T-cell marker CD3 showed that while the increased incidence of apoptosis at day 11 is partly attributable to T-cell death, most of the apoptosis was due to beta-cell death.

178 B. A. O’BRIEN ET A L

Fig. ]-Light micrograph of haematoxylin and eosin-stained sections of pancreatic islets from C57B1/6 mice on the final day (day 5) of treatment with multiple low doses of stz. (a) A condensed apoptotic cell containing dense masses of nuclear chromatin is indicated by an arrow ( x 1000). (b) Lower-power photomicrograph showing a number of apoptotic cells and bodies (arrows). Some of the apoptotic cells indicated are in the process of budding ( x 400)

4c

3c

2c

1C

0

(a)

Blood glucose frnml/ll

8 Coni I 5d

1

7 d Qd

G

I YS 71d I 13d I 15d 17d

(b) Pancreatic insulin f uni ts/g)

I

Cant 6d 7d li

Qd l ld

Days

a 13d 75d 17d

Fig. 2 ~ -Non-fasting plasma glucose response (a) and pancreatic immunoreactive insulin (IRI) content (b) of male C57B1/6 mice treated with stz. Day 1 is the day of the first injection. Values shown represent the mean of 6 sample points f SEM

DISCUSSION continued throughout the period of insulitis, suggesting that diabetes development in this model relies on both a direct beta-cell cytotoxicity of stz and a subsequent

death responsible for beta-cell loss in the multiple autoimmune reaction. low-dose stz model of IDDM. Beta-cell apoptosis Interpretation of the significance of a particular level preceded the appearance of T-cells in the islets and of apoptosis observed in a tissue is complex. In the

This study shows that apoptosis is the mode of cell

BETA-CELL APOPTOSIS IN IDDM 179

ISLET AREA

Percentage of control values 120

100

80

60

40

20

0 Cont 5d

i 7d 9d 11d 13d 15d 17d

Days Fig. 3-Decrease in the percentage of islet area, as compared with control animals, after male C57B1/6 mice were treated with stz. Values shown represent the mean of 6 sample points f SEM

present study, even though the majority of beta cells disappeared over a 9-day period, the maximum level of beta-cell apoptosis recorded was only 2,7 per cent (at day 5). This finding, that such a small percentage of apoptosis can have such a significant effect on the tissue as a whole, is consistent with the role and kinetics of apoptosis occurring in other tissues.20 Apoptosis is an extremely rapid process, with only a few minutes elaps- ing between the onset of the process and the formation of a cluster of apoptotic bodies, and only a matter of hours before these bodies are phagocytosed and com- pletely digested by surrounding cells. l9 Thus, the amount of beta-cell apoptosis observed in the present study is sufficient to account for IDDM development.

Streptozotocin-induced apoptosis occurred asyn- chronously in the beta-cell population, seemingly in a non-random distribution throughout the pancreas. The percentage of islets showing apoptosis was less than 20 per cent for all time points except day 5 , when it was over 40 per cent. At day 11, the percentage of apoptotic beta cells was at a high level, whilst the percentage of islets showing apoptosis remained low. This may reflect an increase in beta-cell apoptosis occurring non- randomly throughout the pancreas, perhaps in response to lymphocyte infiltration of the islets, which was maxi- mal at day 11. Overall, these observations raise the possibility that local factors may influence the timing of induction or expression of beta-cell apoptosis.

APOPTOSIS IN ISLETS

Percentage apoptosis 41-

T T

i v

Cont Id 2d 3 d 46 5d 7d Qd lld 13d 16d 17d

Days Fig. “Time course of apoptosis of islet cells from C57Bl16 mice after five low doses of stz. Day I is the day of the first injection. The number of apoptotic cells, excluding infiltrating T-cells, is represented as a percentage of the total number of cells. Values shown represent the mean of 6 sample points & SEM

The relatively small number of apoptotic cells (or bodies) observed in islets following multiple low-dose stz treatment has implications for the design of studies to determine the role of apoptosis in the evolution of IDDM occurring under other circumstances. In the non-obese diabetic (NOD) mouse21 model of sponta- neous IDDM, the onset of clinical diabetes occurs over a much longer time frame than that observed in the multiple low-dose stz model (many weeks rather than days). It might be expected, therefore, that the level of apoptosis at any one time point in this model may be extremely low. Finding and quantifying apoptosis occur- ring in this circumstance will require extensive detailed examination. Nevertheless, our preliminary studies indi- cate that in NOD mice also, apoptosis is the mode of cell death responsible for beta-cell deletion (unpublished observations).

Beta-cell loss in the stz model of IDDM has been attributed by others to immunological reactions of T-cells against islet cells and/or a direct toxic effect of the d r ~ g . ~ ~ * ~ ~ The results of the present study are consistent with the conclusions drawn by a number of previous investigators that the direct toxic effect or stz represents the primary event in beta-cell destruction and lym- phocytic infiltration the secondary Stz is thought to exert its beta cytotoxic effect through the generation of free radicals, which cause DNA strand

180 B. A. O’BRIEN ET A L

ISLETS SHOWING APOPTQSiS

Percentage

50 r T

I I

Con Id 2d 3d 4d 5d 7d Qd 17d 13d 15d .17d

Deys Fig. 5-Percentage of islets containing apoptotic cells at various times after stz treatment. Day 1 is tlic day of the first injection. Values shown represent the mean of 6 sample points * SEM

breaks.’6 DNA strand breaks induced by radiation or drugs have been shown to cause apoptosis in other cell types.”.’*

It has been reported that thymic immunity may amplify the diabetogenic effect of stz by eliciting insu- litis.’9 Whilst in the present investigation some of the apoptosis observed at day 11 (when lymphocytic infil- tration was at its peak) is attributable to apoptotic T-cells, most of the apoptosis was still of beta-cell origin. Apoptosis has been reported in several mor hological studies of in vitro T-cell-induced cell death.3 13 T-cells and glial cells have both been reported to undergo apoptosis in the central nervous system of rats with experimental autoimmune encephalomyelitis, a T-cell- mediated demyelinating disease of the central nervous system.’4 In situations where cell death is believed to be induced by cell-mediated immunity in vivo, numerous apoptotic bodies have been found adjacent to the infil- trating Apoptosis of beta cells as a result of autoimmune attack is therefore consistent with what is known of target cell response in other circumstances of cell-mediated immune killing.

If the mechanism of stz cytotoxicity involves free radical generation, these reactive species may modify intracellular proteins and the resulting beta-cell deletion by apoptosis may allow the cryptic determinants pro- duced to be displayed to the immune system. This may explain the presence of two distinct peaks in the inci- dence of beta-cell apoptosis, the first at day 5 due to

2-

Fig. 6-Tmmunohistochemical staining for insulin ( x 1000). Arrows indicate apoptotic cells in which chromatin fragmentation and apop- totic bodies are seen

direct beta-cell cytotoxicity, serving to reveal antigenic peptides and to stimulate a subsequent T-cell-mediated immune attack on beta cells, which in turn produces the second peak of apoptosis at day 11. In systemic lupus erythematosus (SLE), an autoimmune disease, autoanti- bodies are directed against a variety of intracellular antigens of diverse l~ca t ion . ’~ Autoantigens previously identified in SLE development have been found clustered in blebs at the surface of apoptotic keratin- ocytes, keratinocytes being the principal cell type targeted in this disease.”

In the multiple low-dose stz model, regardless of whether a direct toxic effect and/or a subsequent auto- immune reaction leads to the deletion of beta cells, the results of this investigation show that the beta-cell death is by apoptosis. Lorenzo et al.39 have recently shown that beta cells may die by apoptosis in vitro when exposed to amylin, the principal constituent of the amyloid deposits that form in the islets of patients with non-insulin-dependent diabetes mellitus, some of whom go on to develop IDDM. The fact that beta cells undergo apoptosis, a controllable type of cell death, and that in the present study this was intimately involved in the development of IDDM, opens up the possibility of therapeutic approaches aimed at modifying the susceptibility of beta cells to this form of cell death.

ACKNOWLEDGEMENTS

This work was supported by a QUT Meritorious Grant and the Elizabeth Albiez Foundation. Image analysis was performed using a Macintosh Quadra 700 and the Prism Image Analysis Package 2.5.1 provided by the Jupiters Casino Community Benefit Fund Trust. The studies were approved by the Queensland University of Technology Biomedical Ethics Committee and were conducted according to guidelines laid down by the National Health and Medical Research Council of

BETA-CELL APOPTOSIS IN IDDM 181

Australia. We wish to thank Mr Clay Winterford for his technical ~ assistance in producing the plates which

22. Paik S, Fleischer N, Shin S. Insulin-dependent diabetes mellitus induced by subdiabetogenic doses of streptozotocin: obligatory role of cell-mediated autoimmune processes. Proc Null Acud Sci USA 1980; 77: 6129-6133.

appear in Figs 1 and 6 . 23. Bonnevie-Nielsen V, Steffes MW, Lernmark A. A major loss in islet mass

I .

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15

16

17

18

19

20

21

REFERENCES Todd JA, Bell JI, McDevitt HO. HLA-DQ, gene contributes to suscepti- bility and resistance to insulin-dependent diabetes mellitus. Nature 1987; 329 599-604. Yoon J-W. Role of viruses and environmental factors in induction of diabetes. Curr Top Microbiol Immunol 1990; 164 95-123. Swenne I . Pancreatic beta cell growth and diabetes mellitus. Diabetologiu 1992; 3 5 193-201 Gepts W, Lecompte PM. The pancreatic islets in diabetes. Am J Med 1981; 70 105-115. Rossini AA, Mordes JP, Like AA. Immunology of insulin-dependent diabetes mellitus. Annu Rev Immunol 1985; 3: 289-320. Yoon J-W. Role ofviruses in the pathogenesis of IDDM. Ann Med 199 I ; 23: 43 7445. Toniolo A, Onodera T , Yoon J-W, Notkins AL. Induction of diabetes by cumulative environmental insults from viruses and chemicals. Nature 1980; 288 383-385. Kerr JFR, Winterford CM, Harmon BV. Apoptosis: its significance in cancer and cancer therapy. Cancer 1994; 7 3 2013-2026. Golstein P, Ojcius DM, Young D-E. Cell death mechanisms and the immune system. Immunol Rev 1991; 121: 29-65. Gobe GC, Axelsen RA, Harmon BV, Allan DJ. Cell death by apoptosis following X-irradiation of the foetal and neonatal rat kidney. Int J Radial Biol 1988; 54 567-576. RaR MC. Social controls on cell survival and cell death. Nufure 1992; 356 937400. Itoh N, Yonehara S, lshii A, et ol. The polypeptide encoded by the cDNA for human cell surface antigen Fas can mediate apoptosis. Cell 1991; 6 6 233-243. Harmon BV, Takano YS, Winterford CM, Cobe GC. The role of apoptosis in the response of cells and tumours to mild hyperthermia. Int J Rudiut Bid 1991; 59: 489-501. Waring P. DNA fragmentation induced in macrophages by gliotoxin does not require protein synthesis and is preceded by raised inositol triphosphate levels. J Biol Chem 1990; 265: 14476-14480. Levine B, Huang Q. Isaacs JT, Reed JC, Griffin DE, Hardwick JM. Conversion of lytic to persistent alphaviru? infection by the bcl-2 cellular oncogene. Nature 1993; 361: 739-742. Gepts W. Pathologic anatomy of the pancreas in juvenile diabetes mellitus. Diubetes 1965; 1 4 619-633. Eisenbarth GS. Type 1 diabetes mellitus: a chronic autoimmune disease. N Engl J Med 1986 314 1360-1368. Like AA, Rossini AA. Streptozotocin-induced pancreatic insulitis: a new model of diabetes mellitus. Scienm 1976; 193 415417. Wyllie AH, Kerr JFR, Currie AR. Cell death: the significance of apoptosis. Int Rev Cytol 1980; 68. 251-306. Kerr JFR, Harmon BV, Searle J. An electron microscope study of cell deletion in the anuran tadpole tail during spontaneous metamorphosis with special reference to apoptosis of striated muscle fibres. J Cell Sci 1974; 1 4 571-585. Makino S, Kuiiimoto K, Muraoka G, Mizuchima Y, Katagiri K, Tochino Y. Breeding of a non-obese, diabetic strain of mice. Exp Anim (Tokyo) 1980; 2 9 1-13.

and B-cell function precedes hyperglycaemia in mice given multiple low doses of streptozotocin. Diubetes 1981; 3 0 424429.

24. Sandler S. Protection by dimethyl urea against hyperglycaemia, but not insulitis, in low dose streptozotocin-induced diabetes in the mouse. Diabetologiu 1984; 2 6 386388.

25. Rossini AA, Like AA, Chick WL, Appel MC, Cahill GF. Studies of streptozotocin-induced insulitis and diabetes. Proc Nutl Acud Sci USA 1977; I 4 2485-2489.

26. Okamoto H. Molecular basis of experimental diabetes: degeneration, onco- genesis and regeneration of pancreatic B-cells of islets of Langerhans. BioEssuys 1985; 2 15-21.

27. Carson DA, Seto S, Wasson DB, Carrera CJ. DNA strand breaks, NAD metabolism, and programmed cell death. Exp Cell Res 1986; 164: 273-281.

28. Sorenson CM, Barry MA, Eastman A. Analysis of events associated with cell cycle arrest at G, phase and cell death induced by cisplatin. J Nutl Cuncer lns t 1990; 8 2 749-755.

29. Nakamura M, Nagafuchi S, Yamaguchi K, Takaki R. The role of thymic immunity and insulitis in the development of streptozotocin-induced diabetes. Diuhetes 1984; 3 3 89&900.

30. Liepens A, Faanes RB, Lifter J, Choi YS, Harven E. Ultrastructural changes during T-lymphocyte-mediated cytolysis. Cell Immunol 1977; 2 8 109-124.

31. Don MM, Ablett G, Bishop CJ, et ul. Death of cells by apoptosis following attachment of specifically allergized lymphocytes in vifro. AU.FI J Exp Biol Med Sci 1977; 55: 407417.

32. Russell JH, Dobos CB. Mechanisms of immune lysis 11. CTL-induced nuclear disintegration of the target begins within minutes in cell contact. J Immunol 1980; 125 12561261.

33. Russell JH, Masakowski V, Rucinsky T, Phillips G. Mechanisms of immune lysis 111. Characterization of the nature and kinetics of the cytotoxic T lymphocyte-induced nuclear lesion in the target. J Immunol 1982; 128 2087-2094.

34. Pender MP, McCombe PA, Yoong G, Nguyeu KB. Apoptosis of aD T lymphocytes in the nervous system in experimental autoimmune encepha- lomyelitis: its possible implications for recovery and acquired tolerance. JAutoimmun 1992; 5: 401410.

35. Searle J, Kerr JFR, Battersby C, Egerton WS, Balderson G, Burnett W. An electron microscopic study of the mode of donor cell death in unmodified rejection of pig liver allografts. Aust J Exp Biol Med Sci 1977; 55: 401406.

36. Gallucci BB, Sale GE, McDonald GB, Epstein R, Shulman HM, Thomas ED. The fine structure of human rectal epithelium in acute graft versus host disease. Am J Surg Puthol 1982; 6 293-305.

37. Pistiner M, Wallace DJ, Nessim S , Metzger AL, Klinenberg JR. Lupus erythematosus in the 1980s: a survey of 570 patients. Semin Arthritis Rheum 1991; 21: 55-61.

38. Casciola-Rosen LA, Anhalt G, Rosen A. Autoantigeus targeted in systemic erythematosus are clustered in two populations of surface structures on apoptotic keratinocytes. J Exp Med 1994; 179: 1317-1330.

39. Lorenzo A, Rarraboni B, Weir GC, Yankner BA. Pancreatic islet cell toxicity of amylin associated with type-2 diabetes mellitus. Nature 1994; 368 756-760.