The intraislet insulin hypothesis for the signaling of the glucagon secretory response to hypoglycemia states that a decrease in arterial glucose --> a decrease in beta-cell insulin secretion --> a decrease in tonic alpha-cell inhibition by insulin --> an increase in alpha-cell glucagon secretion. To test this hypothesis in humans, a hyperinsulinemic- euglycemic ( approximately 5.0 mmol/l [90 mg/dl] x 2 h) and then a hypoglycemic ( approximately 3.0 mmol/l [55 mg/dl] x 2 h) clamp was performed in 14 healthy young adults on two occasions, once with oral administration of the ATP-sensitive potassium channel agonist diazoxide to selectively suppress baseline insulin secretion and once with the administration of a placebo. The decrement in plasma C-peptide during the induction of hypoglycemia was reduced by approximately 50% in the diazoxide clamps (from 0.3 +/- 0.0 to 0.1 +/- 0.0 nmol/l [0.8 +/- 0.1 to 0.4 +/- 0.1 ng/ml]) compared with the placebo clamps (from 0.4 +/- 0.0 to 0.1 +/- 0.0 nmol/l [1.2 +/- 0.1 to 0.4 +/- 0.1 ng/ml]) (P = 0.0015). This reduction of the decrement in intraislet insulin during induction of hypoglycemia caused an approximately 50% reduction (P = 0.0010) of the increase in plasma glucagon in the diazoxide clamps (from 29 +/- 3 to 35 +/- 2 pmol/l [102 +/- 9 to 123 +/- 8 pg/ml]) compared with the placebo clamps (from 28 +/- 2 to 43 +/- 5 pmol/l [98 +/- 7 to 151 +/- 16 pg/ml]). Baseline glucagon levels, the glucagon response to intravenous arginine, and the autonomic (adrenomedullary, sympathetic neural, and parasympathetic neural) responses to hypoglycemia were not altered by diazoxide. These data indicate that a decrease in intraislet insulin is a signal for the glucagon secretory response to hypoglycemia in healthy humans. The absence of that signal plausibly explains the loss of the glucagon response to falling plasma glucose concentrations, a key feature of the pathogenesis of iatrogenic hypoglycemia, in insulin-deficient (type 1 and advanced type 2) diabetes.