Interactions of pancreatic islets and islet-associated mononuclear cells (IAMCs) from the nonobese diabetic (NOD) mouse were morphologically investigated. To obtain IAMCs, pancreatic islets isolated from adult NOD mice were cultured for 7 days with interleukin 2. Noted by light microscopy, interactions between IAMCs and freshly isolated islets from young NOD mice began 30 min after the initiation of the coculture, and 6 h later, normal cellular array of the islets was lost. By electron microscopy, most IAMCs had low nucleus-cytoplasm ratio, the nucleus was notched and exhibited condensed chromatin along the nuclear membrane, and well-developed Golgi complexes and several mitochondria were distributed in the cytoplasm. These IAMCs adhered to beta-cells, but not to alpha- or delta-cells, with their pseudopods and caused cytolysis of beta-cells. Immunohistochemical study with antibodies specific for pancreatic hormones demonstrated that only cells reacting with anti-insulin antibody were selectively lost as the incubation time proceeded. Electron immunohistochemistry by immunogold technique showed that effector cells in IAMCs reacted with anti-CD8 (Lyt-2) antibody, but not anti-CD4 (L3T4) or anti-asialogangliosideM1 antibody. In addition, the concentration of pancreatic hormones in the culture medium, used as a marker of cytolysis, also demonstrated that insulin was significantly increased after 6 h of culture, whereas glucagon and somatostatin were not. These results suggest that CD8+ cytotoxic T lymphocytes are involved in the selective destruction of pancreatic beta-cells in the NOD mouse.

The immune mechanisms directly responsible for beta-cell destruction in insulin-dependent diabetes are undefined. We studied the role of MHC class I-restricted T lymphocytes in the development of diabetes in cyclophosphamide (CY)-treated male and untreated female NOD mice (H-2Kd,Db). After administration of CY to 10-wk-old male NOD/Shi/Kbe mice, 37 of 64 (58%) phosphate-buffered saline-injected control mice and 13 of 22 (59%) anti-Kb and 12 of 27 (44%) anti-Db monoclonal antibody (MoAb)-injected mice became diabetic by 14 wk of age, whereas only 3 of 38 (8%) anti-Kd and 2 of 13 (15%) anti-Lyt-2 MoAb-injected mice did. In untreated female NOD/Shi/Kbe mice, 30 of 46 (65%) mice developed spontaneous diabetes by 30 wk of age, whereas none of 9 anti-Kd MoAb-injected mice became diabetic. Immunohistochemical studies showed that islet-infiltrating cells in CY-treated control mice were composed mainly of both L3T4+ and Lyt-2+ T lymphocytes, whereas many L3T4+ and very few Lyt-2+ lymphocytes infiltrated within the islets in anti-Kd MoAb-injected mice. Administration of anti-Lyt-2 MoAb induced the absence of Lyt-2+ T lymphocytes in the islet and spleen. However, anti-Kd MoAb did not change the number of spleen cells or the T-lymphocyte subset and response to concanavalin A. These results suggest that MHC class I Kd-restricted Lyt-2+ T lymphocytes play an important role as direct effector cells in destruction of beta-cells in NOD/Shi/Kbe mice.

Summary. Recent evidence indicates that activated T cells and macrophages play an important role in the induction of insulitis and diabetes in certain strains of mice treated with multiple subdiabetogenic doses of streptozotoein. In the present study, we treated C57BL/6J mice with five daily doses of 40 mg/ml streptozotocin and examined the prophylactic effect of an anti-interleukin 2 receptor monoclonal antibody (PC61). In mice treated with streptozotocin, interleukin 2 receptor-positive mononuclear cells were shown to infiltrate into the islets and soluble interleukin 2 receptors in the sera were significantly increased compared with control mice. The administration of PC61 to the mice attenuated the insulitis, and diminished interleukin 2 receptor-positive cells from islets and soluble interleukin 2 receptors in the sera. Moreover, the administration of PC61 significantly reduced the development of hyperglycaemia shown in these mice (12.8 _+ 1

Spleen cells from non-obese diabetic mice were found to generate low interleukin 2 production and cell proliferation in response to concanavalin A. However, some of non-obese diabetic mice maintained in the same environment preserved their responsiveness to this T cell mitogen. Non-obese diabetic mice at every age had a higher percentage of Thyl.2, L3T4, and Lyt2-positive spleen cells than did control mice, suggesting that the dysfunction of spleen cells did not depend on the number of T cells or the ratio of these subpopulations. Evidence for macrophage-mediated suppression participating in the deficient function of splenic lymphocytes in this mouse model of insulin-dependent diabetes includes: 1) the restoration of mitogen-induced interleukin 2 production after the macrophages have been depleted by silica absorption form spleen cells; 2) the complete suppression of the cell proliferation by thioglycollate-stimulated peritoneal exudate cells from non-obese diabetic and control mice, and the partial suppression by spleen macrophages from non-obese diabetic mice; 3) the reversal of the suppression of interleukin 2 production by the prostaglandin synthetase inhibitor indomethacin (0.1-1 microgram/ml); 4) the partial suppression of interleukin 2 production, conversely, by the exogenous prostaglandins E1 and E2 (2.5 x 10(-6) mol/l). These results indicate that the activated macrophages existing among the spleen cells suppress the response of splenic T cells to concanavalin A. This impairment may contribute to the pathogenesis of insulin-dependent diabetes in non-obese diabetic mice.

The immune abnormalities of NOD mice, a model of human type I (insulin-dependent) diabetes, have been postulated to be T-lymphocyte dependent. We measured responsiveness to exogenous interleukin 2 (IL-2) and IL-2 production in spleen mononuclear cells from female NOD/Shi/Kbe mice after stimulating the cells with concanavalin A (ConA blasts) or phytohemagglutinin (PHA blasts). Exogenous IL-2 produced significantly lower proliferative responses in each blast from 3- and 10-wk-old NOD/Shi/Kbe mice than from control strains. IL-2 production in NOD/Shi/Kbe mice was inclined to decrease but not significantly compared with controls. Even sufficient amounts of recombinant IL-2 (rIL-2) or IL-1 (rIL-1), added with mitogens to the preculture medium, failed to provoke normal proliferative responses from NOD/Shi/Kbe mouse cells. To clarify the reason for this defect, we investigated the expression of IL-2 receptors (IL-2Rs) on mitogen-activated cells with anti-IL-2R monoclonal antibody (PC61) and radiolabeled IL-2. Cytofluorometry showed no significant difference between strains in the number of PC61+ ConA and PHA blasts. However, Scatchard analysis with 125I-labeled IL-2 showed that the number of high-affinity IL-2Rs (H-IL-2Rs), the mediators of the biological activity of IL-2, was decreased in NOD/Shi/Kbe mice compared with controls, whereas the number of low-affinity IL-2Rs (L-IL-2Rs) was not different. Separating the L3T4+ and Lyt-2+ populations of T lymphocytes by cell sorting showed both to be deficient in H-IL-2Rs.(ABSTRACT TRUNCATED AT 250 WORDS)