The pancreatic beta-cell recognition of insulin secretagogues: does cyclic AMP mediate the effect of glucose?

Insulin release and the content of cAMP were studied in microdissected pancreatic islets of noninbred ob/ob (obese) mice. In the absence of 3-isobutyl-1-methylxanthine, a phosphodiesterase inhibitor, 20 mM glucose had no effect on cAMP save a very small initial rise detectable by a freeze-stop perifusion technique only. However, combined with this methylxanthine, 20 mM glucose produced significant increases of cAMP both in perifused islets and in islets conventionally incubated in closed vials. Glucose shared this capacity to raise the cAMP level with D-glyceraldehyde and 1,3-dihydroxyacetone. Isobutylmethylxanthine (0.05-1.0 mM) or 5 mug/ml of cholera toxin, an activator of adenylate cyclase, also increased the islet cAMP level; the effects of the methylxanthine, whether or not combined with cholera toxin, were potentiated by glucose. Isobutylmethylxanthine (0.05-1.0 mM) or 5 mug/ml of cholera toxin potentiated insulin release in response to 20 mM glucose. However, only 0.5-1.0 mM isobutylmethylxanthine stimulated insulin release in the presence of 3 mM glucose, whereas 0.05-0.1 mM isobutylmethylxanthine or 5 mug/ml of cholera toxin had no effect on secretion at the low glucose concentration. These discrepancies between cAMP-promoting and insulin-releasing activities suggest that glucose does not initiate insulin release by activating the beta-cell adenylate cyclase. By being metabolized in the beta-cells, glucose may both create a release-initiating signal not identical with cAMP and enhance cAMP formation, leading to potentiation of the effect of the initiator signal.