NOD mice constitute a model for studying the prevention of human autoimmune type 1 diabetes. Glutamic acid decarboxylase (GAD) could be a key antigen involved in this disease, and GAD65 peptide 524-543 has been implicated in early T cell response in young NOD mice. We performed two i.p. injections of GAD peptide 524-543 (100 micrograms at each injection), together with Freund's incomplete adjuvant (FIA), into female NOD mice at 30 and 45 days old. Diabetes was accelerated 2 weeks later by a single injection of cyclophosphamide (CY), which acts against suppressive mechanisms. Treatment with GAD 524-543 peptide delayed the onset of diabetes and reduced its incidence (28% versus 60%; P < 0.001) compared with control mice injected with FIA alone, or GAD peptide 534-553, or an irrelevant peptide. In the same group, the severity of lymphocytic inflammation of pancreatic islets was reduced (P < 0.03). Up to 3 months after peptide injections, a strong splenocytic proliferative response occurred in immunized NOD mice against the immunizing peptide alone (but not against a panel of seven other GAD65-derived peptides). After peptide challenge of splenocytes in vitro, protection against CY-accelerated diabetes was associated with higher peptide-specific production of T helper type 2 (Th2)-associated interleukins 4 and 10, whereas Th1-associated interferon-gamma and IL-2 were proportionally less represented. During contransfer, T splenocytes from GAD 524-543-immunized mice were able to reduce the capacity of T cells from diabetic donors to transfer the disease adoptively (P < 0.01), demonstrating the generation of cellular mechanisms that actively suppress the disease. It is concluded that immunization of NOD mice with GAD65 peptide 524-543 can counteract CY-accelerated diabetes, possibly through active cellular suppression linked to a shift of Th1/Th2 balance toward the production of Th2 cytokines such as IL-4 and IL-10. This study provides additional support for the notion that GAD, and more precisely its epitope 524-543, could be one of the key targets for the pathogenesis of type 1 diabetes in NOD mice, as well as for the efficacy of disease-specific peptide therapy in type 1 diabetes.