Adequate regulation of glucose levels by a microencapsulated pancreatic islet graft requires a minute-to-minute regulation of blood glucose. To design such a transplant, it is mandatory to have sufficient insight in factors influencing the kinetics of insulin secretion by encapsulated islets. The present study investigates factors influencing the glucose-induced insulin response of encapsulated islets in vitro. We applied static incubations and did the following observations. (i) Small islets (90-120 microm) showed a similar instead of a lower glucose-induced insulin response, suggesting that inclusion of only small islets, which are associated with lower protrusion and failing rates, has no consequences for the functional performance of the graft. (ii) A capsule diameter of 800 microm showed identical rather than lower glucose-induced insulin responses as smaller, 500-microm capsules. (iii) Capsule membranes constructed with a conventional permeability interfered with diffusion of insulin, as illustrated by a lower response of islets in capsules with a 10-min poly-L-lysine (PLL) membrane than islets in capsules with a 5-min PLL membrane. (iv) Irrespective of the tested porosity, the capsules provided sufficient immunoprotection because the 10-min PLL membranes did block diffusion of the cytokines IL-1beta (17 kDa) and TNF-alpha (70 kDa) while the 5-min PLL membranes interfered with the diffusion of the vast majority of the cytokines. We conclude that capsules containing small islets (90-120 microm) and a membrane with a lower permeability than routinely applied is preferred in order to obtain a graft with adequate glucose-induced insulin responses.