To determine whether the heterogeneous B-cell response to glucose is related to a different metabolic handling of this sugar, we have compared rat B-cells differing in their redox response to glucose stimulation. To this end, a population of B-cells showing increased NAD(P)H autofluorescence after a 15-min exposure to 16.7 mM glucose was sorted from a population of B-cells that, under the very same conditions, failed to show detectable changes in basal NADP(H) autofluorescence. Insulin secretion was evaluated by a reverse hemolytic plaque assay in these two populations, referred to as high and low NAD(P)H, respectively. After a 30-min stimulation by 16.7 mM glucose, both populations comprised secreting B-cells and B-cells that did not release detectable amounts of insulin. However, the percentage of secreting B-cells and total insulin output were larger (P < 0.01-0.02) in the high- (77.9 +/- 6.5% and 141.7 +/- 27.4 microns2 x 10(3)) than in the low-NAD(P)H population (58.7 +/- 6% and 92.7 +/- 20.5 microns2 x 10(3)). The high-NAD(P)H population also comprised B-cells that, on average, had a larger (P < 0.001) profile area (142.9 +/- 2.3 microns2) than the B-cells of the low-NAD(P)H population (118.6 +/- 1.5 microns2). Glucose-induced insulin secretion was similar in the high- and low-NAD(P)H group when cells of similar sizes were compared and increased similarly in the two populations as a function of B-cell size. Analysis of variance revealed that insulin secretion was influenced (P < 0.005) by the size of B-cells and not by their NAD(P)H level.(ABSTRACT TRUNCATED AT 250 WORDS)