Islet allografts transplanted into Type I diabetic recipients may be destroyed by allorejection or recurrent autoimmune diabetes. We studied islet transplantation in three murine models in order to determine the relative sensitivity of autoimmunity and alloimmunity to two immunosuppressive agents that may be useful in clinical islet transplantation: 15-deoxyspergualin (DSG) and anti-CD4 antibody (GK 1.5). In the model in which only allorejection occurs (BALB/c islets transplanted into streptozotocin-induced diabetic CBA or streptozotocin-induced diabetic NOD recipients), both DSG and anti-CD4 antibody treatment led to indefinite survival of allogeneic islets (>100 days in both treatments). In the second model in which only recurrent autoimmunity can destroy islet grafts (islets from NOD donors transplanted into spontaneously diabetic NOD recipients), only anti-CD4 treatment caused prolonged graft survival [MST 36.7 +/- 6.8 days vs 9.8 +/- 1.8 days (controls), P < 0.0002]. Treatment with DSG did not cause any increase in graft survival (MST 12.6 +/- 5.4 days, NS). Finally, using a model in which both autoimmunity and allorejection may occur (BALB/c to spontaneously diabetic NOD mice), treatment with anti-CD4 caused marked graft prolongation [42.0 +/- 14.5 days vs 7.2 +/- 0.8 days (control), P < 0.002] while DSG again did not prolong graft survival with respect to untreated recipients (9.8 +/- 3.0, NS). We conclude that recurrent autoimmunity in the NOD mouse involves a CD4+ T cell that is not sensitive to DSG. Anti-CD4 antibody may be useful in human clinical islet transplantation trials because it seems to prevent both allorejection and recurrent autoimmunity.