Early intensive insulin treatment is thought to improve subsequent Beta-cell function in Type 1 (insulin-dependent) diabetic patients. Prophylactic insulin administration also reduced diabetes incidence in diabetes-prone animals. To study the mechanisms by which these effects occur, we tested the ability of insulin therapy in the model of non-obese-diabetic mice, to prevent the penetration of committed T cells into the islets and subsequent Beta-cell destruction. Sublethally irradiated non-obese-diabetic males of 8 weeks of age were adoptively transferred with splenocytes from diabetic donors and treated with the maximum tolerable dosage of fast-acting insulin (0.5 U, twice daily) until 30 days after cell transfer. Diabetes incidence was compared to control animals injected with the same concentration of insulin diluent. After one month of treatment, the cumulative diabetes frequency was significantly less within the insulintreated group (4 of 15, 26.6%) than in the control group (15 of 18, 83.3%; p < 0.01). Pancreatic histological analysis of in-sulin-treated animals revealed a lower severity of insulitis and Beta-cell necrosis and a higher percentage of normal islets (46.6 + 10% vs 2.3 _+ 2%, p < 0.01), including five (33%) mice with no lesions. Immunoperoxydase staining of pancreatic sections indicated similar insulin and ganglioside staining of Beta cells from insulin-treated mice and control animals. Insulin-treated mice had comparable pancreatic insulin content to normal mice. Flow cytometry analysis of spleen cell populations indicated that insulin increased the number of Thyl,2 ยง and Lyt-2 + T cells. Although an effect at the Beta cell level cannot be definitely excluded, several lines of evidence suggest that insulin may influence the capacity of effector T cells to invade the islets and cause Beta-cell destruction. These effects may have potential interest for future immunointervention trials in Type 1 diabetic patients of recent onset.