There are diverse strategies for gene therapy of diabetes mellitus. Prevention of beta-cell autoimmunity is a specific gene therapy for prevention of type 1 (insulin-dependent) diabetes in a preclinical stage, whereas improvement in insulin sensitivity of peripheral tissues is a specific gene therapy for type 2 (non-insulin- dependent) diabetes. Suppression of beta-cell apoptosis, recovery from insulin deficiency, and relief of diabetic complications are common therapeutic approaches to both types of diabetes. Several approaches to insulin replacement by gene therapy are currently employed 1) stimulation of beta-cell growth, 2) induction of beta-cell differentiation and regeneration, 3) genetic engineering of non-beta cells to produce insulin, and 4) transplantation of engineered islets or beta cells. In type 1 diabetes, the therapeutic effect of beta-cell proliferation and regeneration is limited as long as the autoimmune destruction of beta cells continues. Therefore, the utilization of engineered non-beta cells free from autoimmunity and islet transplantation with immunological barriers are considered potential therapies for type 1 diabetes. Proliferation of the patients own beta cells and differentiation of the patients own non-beta cells to beta cells are desirable strategies for gene therapy of type 2 diabetes because immunological problems can be circumvented. At present, however, these strategies are technically difficult, and transplantation of engineered beta cells or islets with immunological barriers is also a potential gene therapy for type 2 diabetes.