Previously we have shown that nonsteroidal, multimodal immunosuppression consisting of cyclosporine (CsA), azathioprine (AZA), and RBC-absorbed goat anti-dog lymphoblast globulin (ALG), at doses considered tolerable in humans, results in limited canine islet allograft survival. Regimens including the diabetogenic agent prednisone are even less successful. In our search for an immunosuppressive regimen that could be applied to human islet transplantation, we tested the effect of 15-deoxyspergualin (DSG) in a canine model. Animals received either CsA, 20 mg/kg/day; AZA, 2.5 mg/kg/day; and ALG, 20 mg/kg/day x 14 days (Group I) or CsA, AZA, ALG at the same doses with the addition of low dose DSG, 0.5 mg/kg/day (Group II). Trough CsA levels by high pressure liquid chromatography ranged in both groups between 100 and 200 micrograms/liter. Rejection of the islets was diagnosed when serum glucose remained greater than 200 mg/dl for 3 consecutive days. There was no significant difference in the number of islets per kilogram of body weight of recipient transplanted in Group I and Group II (5754 +/- 2544 islets/kg versus 7953 +/- 3440 islets/kg, respectively). Animals receiving ALG, CsA, and AZA alone achieved a median islet allograft survival of 4 days, with a mean of 10.8 days. However, with the addition of low-dose DSG, median islet allograft survival was improved to 22 days, with a mean of 32.4 days (P = 0.012, Mann-Whitney test). We conclude that the addition of low-dose DSG to an ALG induction, cyclosporine-based immunosuppressive regimen enhances canine islet allograft survival and has potential for application in clinical islet allotransplantation.