Immunoregulatory polypeptides (cytokines) produced by mononuclear cells of the immune system can inhibit insulin secretion and may be cytotoxic to pancreatic islet beta cells. We used a modified 51Cr release cytotoxicity assay to define the interactive cytotoxic effects of interleukin-1 (IL-1), tumor necrosis factor (TNF), and interferon gamma (IFN-gamma) on rat islet cells in monolayer culture. We here demonstrate that (a) IFN-gamma, but not IL-1 or TNF, can sensitize islet cells to subsequent cytolysis by other cytokines (IL-1 or TNF); (b) the cytolytic effect of sequential addition of IFN-gamma, then IL-1 or TNF (approximately 35-40% 51Cr release) is similar to that produced by the concurrent addition of IFN-gamma plus IL-1 or TNF; (c) the priming effect of IFN-gamma persists for 3 to 6 or more days after its removal; and (d) islet cells preincubated with IFN-gamma are also more sensitive to cytolysis by splenic mononuclear cells from diabetes-prone BB/Wor rats. These findings suggest that IFN-gamma produced by activated T lymphocytes and monocytic cells infiltrating islets in Type 1 diabetes may play a direct and important role in sensitizing beta cells to damage by other cytokines (IL-1, TNF) and cytotoxic cells in the immune/inflammatory infiltrate.

Arachidonic acid metabolites (eicosanoids) have been implicated in mediating actions of cytokines in different tissues. In this study, we tested inhibitors of arachidonate metabolism for possible protection against the toxic effects of the cytokine combination of tumor necrosis factor (TNF, 100 U/ml) and interferon-gamma (IFN-gamma, 100 U/ml) in rat islet cell monolayer cultures, using a 51Cr release cytotoxicity assay to measure islet cell lysis (% 51Cr release). The toxic effect of TNF/IFN-gamma (26.6 +/- 3.7%) was inhibited partially by both a cyclooxygenase inhibitor, indomethacin and a lipoxygenase inhibitor, nordihydroguaiaretic acid (NDGA), and combination of maximally effective concentrations of Indo and NDGA (30 microM) produced further protection against TNF/IFN-gamma-induced lysis (3.5 +/- 0.9%). Also, the combined cyclo/lipoxygenase inhibitors, oxyphenbutazone and eicosa 5,8,11,14 tetrynoic acid, as well as the phospholipase A2 inhibitor, bromophenacyl bromide, significantly inhibited the cytotoxic effect of TNF/IFN-gamma. Whereas indomethacin and NDGA did not prevent TNF/IFN-gamma-induced inhibition of insulin release, this recovered after cytokine removal from cultures protected by the cyclo/lipoxygenase inhibitors. These results suggest that arachidonate metabolites may be involved in mediating the cytotoxic and not the functional inhibitory effects of TNF and IFN-gamma in islet cells.

Anti-islet cell-mediated immunity in the Bio-Breeding/Worcester (BB/Wor) rat was studied by measuring the cytotoxic activity of splenic lymphoid cells from diabetes-prone (DP) rats receiving either oral cyclosporine (Cy-A, 10 mg/kg/d) or olive oil vehicle (DP control) from age 40 to 105 d. Splenic cells were incubated with concanavalin A (Con-A) for 16 h, then used as effector cells in cytotoxicity assays with 51Cr labeled cells from a rat islet cell line (RIN) as targets. Lysis of RIN cells (% 51Cr release) by Con-A-activated splenic cells from control DP rats increased from age 40 d (5.4 ± 1.6%) to 105 d (32.7 ± 1.8%). This increase was significantly less in Cy-A-treated DP rats (14.5 ± 3.2%) and remained low at 15 d after stopping Cy-A (14.8 ± 2.5%). Diabetes developed in control DP rats between 74 and 120 days of age, and in none of the Cy-A-treated DP rats by age 105 d, nor for 15 d after stopping Cy-A. The Con-A-activated splenic cells responsible for lysis of RIN cells also lysed normal rat islet cells and had the functional and phenotypic properties of natural killer (NK) cells. In contrast to the lower Con-A-inducible NK cell cytotoxic activity in Cy-A-treated DP rats, basal NK cell activity and monoclonal antibody-defined NK cell subsets were increased in the Cy-A-treated DP rats. Therefore, Cy-A had inhibited the ability of NK cells to respond to cytotoxic activation despite their increased numbers.

A possible role for oxygen free radicals in mediating the cytotoxic effects of cytokines in islets was sought by the use of agents that scavenge free radicals. Rat islet cell monolayer cultures were incubated for 6 days with t-butylhydroperoxide, alloxan, streptozotocin, or the cytokines, interleukin 1, tumour necrosis factor, and interferon gamma, without and together with the oxygen free radical scavenger combination of dimethylthiourea and citiolone, and islet cell lysis was measured in a 51chromium cytotoxicity assay. The free radical scavengers significantly inhibited the islet cell cytotoxic effects of t-butylhydroperoxide and alloxan, but not streptozotocin. Similarly, the cytotoxic effects of the cytokine combinations of interleukin 1 plus tumour necrosis factor, interferon gamma plus tumour necrosis factor, and interferon gamma plus interleukin 1 were significantly inhibited by the free radical scavenger combination of dimethylthiourea and citiolone. These results suggest that the cytokine products of macrophages and lymphocytes infiltrating islets in Type 1 (insulin-dependent) diabetes may contribute to B-cell damage by inducing the production of oxygen free radicals in the islet cells.

Polypeptide products of mononuclear cells of the immune system (cytokines) have been reported to be cytotoxic to the insulin-producing pancreatic islet beta-cells. The objective of this study was to distinguish between reversible effects of cytokines on insulin secretion and irreversible cytotoxic effects on the islet beta-cells. When tested as single agents or added together at very low concentrations, interleukin 1 (IL-1), tumor necrosis factor (TNF), and interferon gamma (IFN-gamma) inhibited insulin release from rat islet cell monolayer cultures during 4 day incubations; however, this secretory function improved after the cytokines were removed. In contrast, combinations of slightly higher concentrations of IL-1, TNF, and IFN-gamma produced irreversible inhibition of insulin release, as well as decreased cell insulin content and proportional increase in cell lysis, measured as release of 51Cr from labeled islet cell cultures. These findings suggest that cytokine products of T lymphocytes (IFN-gamma) and macrophage/monocytic cells (IL-1, TNF) infiltrating pancreatic islets in "autoimmune" diabetes may interact synergistically to produce functional inhibition or lethal cytocidal effects on islet beta-cells, possibly accounting for reversible and irreversible stages of insulin-dependent diabetes.

Recent observations suggest a role for interleukin-1 (IL-1), a polypeptide product of macrophage/monocytic cells, in the immune-mediated destruction of pancreatic islet beta-cells observed in type 1 diabetes. In this study, we investigated the effects of IL-1 on both alpha- and beta-cell secretory functions in rat islet cell monolayer cultures. Insulin release was 97% inhibited after 6 h of incubation in RPMI-1640 medium (11 mM glucose) containing 1 U/ml IL-1 and 96% inhibited after 24 h of incubation in medium containing 0.1 U/ml IL-1. The cell content of insulin in the monolayers was decreased by 66% (P less than 0.01) after 4 days of incubation in 10 U/ml IL-1; however, after a further 8-day incubation in IL-1-free medium, cell insulin content recovered fully. In contrast, cell glucagon content was decreased by 77% (P less than 0.001) after 4 days of incubation in 10 U/ml IL-1 and did not recover after a further 8-day incubation in IL-1-free medium. After an 18-h preincubation in medium with 0.1 and 1 U/ml IL-1, insulin release responses to 16.7 mM glucose were abolished in 4-h incubations, whereas responses to 0.1 mM 3-isobutyl-1-methylxanthine were normal, and after a further 2 and 5 days of incubation in IL-1-free medium, insulin responses to 16.7 mM glucose recovered fully. Similarly, the inhibitory effect of 16.7 mM glucose on glucagon release was lost after an 18-h preincubation in 0.1 and 1 U/ml IL-1, and did not recover fully after 2 and 5 days in IL-1-free medium, whereas the stimulatory effect of 3-isobutyl-1-methylxanthine on glucagon release was not affected by IL-1. We conclude that 1) IL-1 inhibits glucose-dependent and not cAMP-dependent mechanisms of insulin and glucagon release; 2) inhibition of glucose-stimulated insulin release by IL-1 is reversible, whereas the effect on glucose-modulated glucagon release is not; and 3) IL-1 causes a reversible decrease in the insulin content of islet cells and an irreversible decrease in glucagon content. These actions of IL-1 do not appear to account for the beta-cell-specific destruction of islets characteristic of type 1 diabetes.

An assay was developed to detect the cytotoxic effects of cytokines on rat pancreatic islet cells in monolayer culture. Cell lysis was detected by a 51Cr-release assay after 4 days of incubation with various cytokines. When tested alone, murine (rat and mouse) interferon-gamma (mIFN-gamma) produced a small dose-dependent lysis of islet cells; human IFN-gamma, mouse IFN-alpha/beta, interleukins 1 and 2 (IL-1 and IL-2), tumor necrosis factor (TNF), and lymphotoxin (LT) were inactive. When added together, the following combinations of cytokines showed synergistic cytotoxic effects: TNF (or LT) plus IL-1, TNF (or LT) plus mIFN-gamma, and IL-1 plus mIFN-gamma. These results indicate that the cytokine products of mononuclear cells of the immune system, IFN-gamma, TNF, LT, and IL-1 have strong synergistic cytotoxic effects on islet cells and therefore may act as direct chemical mediators of islet beta-cell destruction in type I (insulin-dependent) diabetes.

We compared the cytotoxic effects to islet cells of lymphoid cells from diabetic and diabetes-resistant (DR) BioBreeding/Worcester (BB/W) rats with a 51Cr-release assay to detect lysis of normal rat islet cells. Splenic lymphoid cells from diabetic rats were more cytotoxic to islet cells (11.3 ± 3.8%) than were lymphoid cells from DR rats (4.0 ± 2.6%). This difference was amplified by incubating the lymphoid cells for 20 h with 5 μg/ml concanavalin A (ConA); islet cell lysis was 39.3 ± 4.5% by ConA-activated diabetic cells and 9.6 ± 2.7% by ConA-activated DR cells. The cytotoxic lymphoid cells were identified as natural killer (NK) cells, because treatment of diabetic lymphoid cells with anti-asialo GM1 serum and complement selectively removed a monoclonal antibody-defined subset of NK cells (OX8+), and the NK-depleted lymphoid cells were not cytotoxic to either islet or NK-sensitive YAC-1 cells, even after culture with ConA. Of several lymphokine products of ConA-stimulated lymphoid cells, interleukin 2 (IL-2), but not interleukin 1 or interferon-γ, significantly activated splenic lymphoid cells cytotoxic to islet cells, and the lymphoid cells from diabetic rats were more sensitive to IL-2 (3 U/ml) than were the cells from DR rats (30 U/ml). This study reveals the presence of ConA- and IL-2-responsive islet cytotoxic NK cells in the diabetic BB/W rat and suggests that IL-2 activation of NK cells may contribute to islet β-cell destruction and diabetes in this animal.

A woman in complete remission from acute myeloblastic leukemia developed thrombotic thrombocytopenic purpura (TTP) subsequent to the third intensive consolidation cycle of cytosine arabinoside and daunorubicin chemotherapy. The constellation of clinical manifestations indicative of TTP were recognized only in retrospect, as they were initially attributed to more usual complications of bone marrow-ablative chemotherapy. The manifestations, probably fueled by numerous red cell and platelet transfusions, increased at the time of recovery of hematopoiesis. At postmortem examination, characteristic microvascular lesions were found in most organs. Similar thrombotic microangiopathy has been described with mitomycin-based chemotherapy regimens and with the combination of cisplatin, vinblastine, and bleomycin. Successful management of this serious complication of chemotherapy requires increased awareness and earlier recognition.