Islet cell transplantation offers many theoretical advantages for the treatment of insulin-dependent diabetes compared to current modes of therapy. However, one of the main limitations to widespread use of islet cell transplantation is the difficulty of separating islets from contaminating cell populations, and the harshness of the methods sued for isolation which consequently compromise islet viability and yield. In the course of studies on inducible cell ablation in the central nervous system of transgenic mice, we stumbled upon an unexpected property of our transgene (GFAP-TK) which produces nearly complete ablation of pancreatic acinar cells when the mice are treated with ganciclovir. The resulting acinar-depleted pancreas consists largely of islets floating in loose adipose connective tissue. If this treatment can be optimized to completely eliminate acinar cells from the pancreas without harming islet cells, and it allows simpler and milder isolation procedures, it would justify continued efforts to generate transgenic pigs for use as uniform donors. As a pilot project to demonstrate feasibility of this approach, we propose two specific aims. The first is to further optimize ablation protocols in the existing transgenic mice, and evaluate the functional state of islets during this ablation process by immunohistochemical and clinical laboratory studies. We will also determine the mechanism by which acinar cells are killed, and generate newer transgenic mice using the acinar-specific elastase promoter that should have even better properties than the GFAP-TK mice. In the second aim we will harvest islets from the acinar-depleted pancreas and demonstrate their functional capacity in vivo by transplanting into streptozotocin-treated recipients to correct diabetes. If these transgenic mice offer the expected advantages for islet isolation, it suggests a novel approach for using transgenic animals in xenotransplantation.