Our previous data suggested that lipoxygenation of endogenously released arachidonic acid (AA) is a critical step in stimulus-secretion coupling in the pancreatic beta cell. In the current study using monolayer cultures of neonatal rat islet cells, exogenous arachidonic acid (AA) (5 μg/ml) potently stimulated insulin release in the presence of a substimulatory glucose concentration, and potentiated release induced by glucose. Since the latter stimulatory effect of AA is prevented by inhibitors of the lipoxygenase pathway, we examined the effects of various lipoxygenase pathway products on glucose-induced insulin secretion. The mediator was not one of the stable end-products of either limb of the lipoxygenase pathway: 12- or 5-hy-droxyeicosatetraenoic acid (HETE) (0.5–2000 ng/ml) did not alter insulin release, whereas 11-HETE, 15-HETE, leukotriene (LT)B4 and the δ6trans isomers of LTB4, LTC4 and 11-trans LTC4 all inhibited insulin release. Furthermore, diethylcarbamazine, a selective leukotriene synthesis inhibitor, did not prevent AA- or glucose-induced insulin release, arguing against a role for LTs as the mediator of AA's stimulatory effect. However, the unstable intermediate 12-hydroperoxyei-cosatetraenoic acid (12-HPETE), and positional isomers of 12-HPETE, potentiated glucose-induced insulin secretion. In addition, insulin release induced by either AA or by glucose was amplified by diethylmaleate, an agent shown in other cell systems to bind reduced glutathione (GSH) and thereby to lead to intracellular accumulation of fatty acid hydroperoxides and epoxides (insulin release at glucose, 300 μmg/dl = 78 ± 2 μU/ml; glucose, 300 mg/dl + diethylmaleate, 1 m M = 131 ± 5, P < 0.001; glucose, 300 mg/dl + AA, 5 (μM/ml = 310 ± 9; P 0.001; glucose + diethylmaleate + AA = 537 ± 3; P < 0.001). In conjunction with our previously reported observation that several analogues of the unstable epoxide LTA4 (epoxyeicosatrienoic acids) increase insulin release, these data are compatible with the hypothesis that reactive, labile intermediates derived from the lipoxygenation of AA (and possibly other polyunsaturated fatty acids) could provide an “on” signal for glucose-induced insulin release. Glucose metabolism, by providing additional GSH, could convert this signal to an “off” response in a negative feedback loop.