The "switch-off" hypothesis to explain β-cell regulation of α-cell function during hypoglycemia has not been assessed previously in isolated islets, largely because they characteristically do not respond to glucose deprivation by secreting glucagon. We examined this hypothesis using normal human and Wistar rat islets, as well as islets from streptozotocin (STZ)-administered β-cell-deficient Wistar rats. As expected, islets perifused with glucose and 3-isobutryl-1-methylxanthine did not respond to glucose deprivation by increasing glucagon secretion. However, if normal rat islets were first perifused with 16.7 mmol/l glucose to increase endogenous insulin secretion, followed by discontinuation of the glucose perifusate, a glucagon response to glucose deprivation was observed (peak change within 10 min after switch off = 61 ± 15 pg/ml [mean ± SE], n = 6, P < 0.01). A glucagon response from normal human islets using the same experimental design was also observed. A glucagon response (peak change within 7 min after switch off = 31 ± pg/ml, n = 3, P < 0.01) was observed from β-cell-depleted, STZ-induced diabetic rats whose islets still secreted small amounts of insulin. However, when these islets were first perifused with both exogenous insulin and 16.7 mmol/l glucose, followed by switching off both the insulin and glucose perifusate, a significantly larger (P < 0.05) glucagon response was observed (peak change within 7 min after switch off = 71 ± 11 pg/ml, n = 4, P < 0.01). This response was not observed if the insulin perifusion was not switched off when the islets were deprived of glucose or when insulin was switched off without glucose deprivation. These data uniquely demonstrate that both normal, isolated islets and islets from STZ-administered rats can respond to glucose deprivation by releasing glucagon if they are first provided with increased endogenous or exogenous insulin. These results fully support the β-cell switch-off hypothesis as a key mechanism for the α-cell response to hypoglycemia.