Type 1 diabetes results from the loss of insulin-producing pancreatic beta cells following the action of beta-cell-specific autoimmune responses. One possible treatment for type 1 diabetes is the development of beta-cell substitutes by introducing an insulin-producing gene into non-beta cells, which would evade the beta-cell-specific autoimmune attack. However, this approach has been hampered by the absence of (1) an appropriate glucose-sensing system to regulate insulin gene transcription; (2) enzymes that process proinsulin to insulin; and (3) glucose-regulatable exocytosis in the target cells. Recent attempts to solve these problems have sought new methods for effective gene transfer and have addressed issues such as the expression and release of insulin in response to the physiological stimulus of glucose, the production of biologically active insulin, and the selection of an ideal target cell for the expression of the insulin gene.