Cell-based implantable artificial tissues are most promising for the long-term treatment of endocrine diseases, such as diabetes. One type of a bioartificial pancreas device consists of calcium alginate microbeads containing insulin-secreting cells and is surrounded by a poly(L-lysine) (PLL) membrane. The membrane is semipermeable, allowing cellular nutrients and metabolites to diffuse through but excluding the antibodies and cytotoxic cells of the host; thus immunoprotecting the cells. The device can be modeled by writing the equations for diffusion of nutrients and metabolites through the polymer and for consumption of the former and production of the latter by the cells. In this paper, we describe the construction and analysis of such a model for alginate/PLL microbeads with insulin-secreting recombinant mouse pituitary AtT-20 and mouse insulinoma beta TC3 cells. Entrapped AtT-20 cells are a simplified model system, whereas microbeads with beta TC3 cells constitute a realistic artificial pancreatic device. Effective diffusivities of key compounds through the polymer with entrapped, inactivated AtT-20 spheroids were measured first. The kinetics of glucose and oxygen consumption and insulin secretion Were modeled next, and the equations for diffusion and reaction were then combined to describe the entire system. The-model was used to compute nutrient and metabolite concentration profiles in beads and the bead secretory response for different bead sizes and cell loadings. The size and loading necessary for the cells to be well nourished and for the beads to be rapidly responsive to step-ups and step-downs of secretion stimuli were evaluated. It was shown that if the cells are hypersensitive to glucose, i.e., they do not shut off secretion at the physiological glucose threshold but at a lower one, so are the microbeads. This work demonstrates the usefulness of mechanistic models with representative parameter values in optimizing the design of artificial tissues and in characterizing aspects of their behavior that are of importance for restoring in vivo function.