Pancreatic B cells are highly specialized in their ability to sense the glucose level and maintain the energy status of the organism through secretion of insulin, which is considered to be the primary anabolic hormone. Glucose exerts its effects on virtually all of the steps of insulin production: transcription of insulin gene, mRNA stability, translation, proinsulin processing, and insulin release via a highly regulated secretory pathway. Insulin therapy remains the only treatment of insulin-dependent diabetes (IDDM)--the chronic autoimmune disorder resulting in B-cell destruction. Yet, even regular daily insulin injections accompanied by the persistent control of blood glucose levels cannot provide for a glycemic control as precise as that existing in B cells, nor can it prevent the development of late complications. Different experimental approaches for IDDM treatment, their advantages and limitations, and their potential to be introduced into medical practice are discussed in this review. The alternative mostly experimental methods of IDDM treatment call for restoration of insulin-producing cells. The numerous experiments on insulin gene delivery into the non-islet tissues resulted in uncontrolled insulin secretion. The attempts to reproduce regulated glucose-dependent transcription of insulin gene in transduced cells could not reproduce so far the same complex and refined regulation of insulin production and release as that provided by B cells. Generation of pancreatic B cells from the embryonic or adult stem cells represents an appealing strategy that could potentially overcome the problems of the lack of donor material and graft rejection. The future experiments on identification and characterization of the factors involved into the differentiation of B cells will allow researches and clinicians to use stem cells for the safe and effective treatment of insulin-dependent diabetes in parallel with insulin- and immunomodulatory therapy.