Insulin-dependent diabetes is a serious pathological condition, currently treated by blood glucose monitoring and daily insulin injections, which, however, do not prevent long-term complications. A tissue-engineered pancreatic substitute has the potential to provide a more physiologic, less invasive, and potentially less costly treatment of the disease. A major issue in developing such a substitute is the cells being used. Nonpancreatic cells, retrieved from the same patient and genetically engineered to secrete insulin constitutively or with some glucose responsiveness, offer the significant advantages of being immune-acceptable and relaxing the tissue availability limitations, which exist with islets from cadaveric donors. These cells, however, do not have insulin secretion dynamics appropriate for restoration of euglycemia in higher animals and, eventually, humans. In this study, we present the concept of a hybrid pancreatic substitute consisting of such cells sequestered in a material exhibiting glucose-dependent changes of its permeability to insulin. A Concanavalin A-glycogen material sandwiched between two polycarbonate membranes and exhibiting glucose-dependent sol-gel transformations was used. Rates of insulin transport through this material in gel and sol forms were characterized for both FITC-labeled insulin in solution and insulin secreted by betaTC3 mouse insulinoma cells. Effective diffusivities through sol were found to be up to 3.5-fold higher than through the gel state of the material. A mathematical model of a hybrid construct was formulated and analyzed to simulate the secretory behavior in response to step ups and downs in the surrounding glucose concentration. The experimental and modeling studies indicate that a hybrid pancreatic substitute consisting of constitutively secreting cells and glucose-responsive material has the potential to provide a more physiologic regulation of insulin release than the cells by themselves or in an inert material.