The desensitization of pancreatic B-cells against stimulation by insulin secretagogues that inhibit ATP-dependent K(+) channels (K(ATP) channels) was investigated by measuring insulin secretion of perifused pancreatic islets. Additionally, the islet insulin content and the number of secretory granules per B-cell were determined. Prior to the measurement of secretion, islets were cultured for 18 h in the presence or absence of the test agents in a cell-culture medium containing 5 mM glucose. The effects of three imidazolines, phentolamine, alinidine, and idazoxan (100 microM each) were compared with those of the well-characterized sulfonylurea, tolbutamide (500 microM), and those of the ion channel-blocking alkaloid, quinine (100 microM). Insulin secretion was strongly reduced upon re-exposure to phentolamine, alinidine, tolbutamide, and quinine, whereas idazoxan, which stimulated secretion only weakly, had no significant effect. The imidazoline secretagogues phentolamine and alinidine induced a cross-desensitization against the stimulatory effect of tolbutamide and quinine. A long-term depolarization with 40 mM KCl was also able to induce a significant reduction of the secretory response to all of the above secretagogues. The insulin content of cultured islets was moderately, but significantly reduced by alinidine, whereas the reduction by phentolamine, tolbutamide, and quinine was not significant. In contrast to these observations, the ultrastructural examination revealed that tolbutamide-treated B-cells had a high degree of degranulation, whereas the other test agents and 40 mM KCl produced only a partial degranulation, except for phentolamine, which produced no significant degranulation at all. These results suggest that the desensitization of insulin secretion is a common property of all agents that stimulate insulin secretion by depolarisation of the plasma membrane. Depending on the specific secretagogue, additional mechanisms, proximal and distal to Ca(2+) influx, appear to contribute to the desensitization (see Rustenbeck et al., pages 1695-1703, this issue).