Sustained hyperglycemia desensitizes pancreatic B-cells to glucose. Infusion of diazoxide protects against this effect, i.e. an insulin response to glucose is preserved after exposure to the drug in vivo. The mechanisms behind this aftereffect were presently investigated in vitro. In one type of experiment, rat pancreatic islets were exposed to diazoxide (75 micrograms/ml) for 20-22 h in tissue culture. Previous exposure to diazoxide increased 2.4-fold the response to 16.7 mM glucose in final batch-type incubations. This effect was not altered by adding insulin (10 mU/ml) to the tissue culture medium. In other experiments, islets were perifused for 180 min with 27 mM glucose and, usually, 0.2 mM 3-isobutyl-1-methylxanthine. A second stimulation, 30 min long, was used to assess desensitization, which ranged between 52-80%. Diazoxide treatment during the first stimulation profoundly inhibited secretion and, as an aftereffect, completely prevented restimulation-assessed desensitization. Neither desensitization nor aftereffect were seen after low (6 mM) glucose treatment. Previous diazoxide treatment enhanced the response not only to glucose but also to 2.5 mM barium or 10 mM arginine. Desensitization induced by 30 mM potassium was not affected by diazoxide. Cooling (22 C) was employed to block glucose-induced exocytosis of insulin, but not proximal events in stimulus secretion coupling. Diazoxide treatment during cooling exerted an aftereffect that was attenuated, but not abolished (1.7-fold of the control, as assessed by restimulation at 37 C). When calcium was omitted from the perifusion medium during cooling, no aftereffect of diazoxide was seen. The results indicate that modulation of ATP-sensitive potassium channel activity is not closely linked to induction and expression of diazoxide's aftereffect, and furthermore, that protection from desensitization includes both proximal and distal events in stimulus-secretion coupling.