Diabetes is a degenerative disease that results from the selective destruction of pancreatic beta-cells. These cells are responsible for insulin production and secretion in response to increases in circulating concentrations of nutrients, such as glucose, fatty acids and amino acids. This degenerative disease can be treated by the transplantation of differentiated islets obtained from cadaveric donors, according to a new surgical intervention developed as Edmonton protocol. Compared to the classical double transplant kidney-pancreas, this new protocol presents several advantages, concerning to the nature of the implant, immunosuppressive drug regime and the surgical procedure itself. However, the main problem to face in any islet transplantation program is the scarcity of donor pancreases and the low yield of islets isolated (very often around 50%) from each pancreas. Nevertheless, transplanted patients presented no adverse effects and no progression of diabetic complications. In the search of new cell sources for replacement trials, stem cells from embryonic and adult origins represent a key alternative. In order to become a realistic clinical issue transplantation of insulin-producing cells derived from stem cells, it needs to overcome multiple experimental obstacles. The first one is to develop a protocol that may allow obtaining a pure population of functional insulin-secreting cells as close as possible to the pancreatic beta-cell. The second problem should concern to the transplantation itself, considering issues related to immune rejection, tumour formation, site for implant, implant survival, and biosafety mechanisms. Although transplantation of bioengineered cells is still far in time, experience accumulated in islet transplantation protocols and in experiments with appropriate animal models will give more likely the clues to address this question in the future.