Paradoxical inhibition of insulin secretion by glucose in human diabetes mellitus

Our previous studies suggested that iv glucose pulses paradoxically suppress insulin levels below basal in some diabetic subjects. To define further the prevalence and mechanism of of this acute insulin decrement (AID), we studied a group of 28 noninsulin-treated, hyperglycemic diabetics. This group of patients as a whole lacked the acute insulin response (AIR) seen in normal subjects at 3-5 min after a 20-g glucose pulse (X̄Δ = 9 ± 6% of basal insulin; P = NS). However, 22 of the 28 subjects did have an AID at 5-10 min (X̄Δ = -15 ± 4%; P<0.001) after iv glucose. This AID was not due to either an osmotic or a dilutional effect of injecting a hypertonic solution, since 8 diabetics who received equivalent galactose pulses showed no AID (X̄Δ = 8 ± 6%) or AID (X̄Δ = 1 ± 4%), suggesting that the AID was evoked specifically by glucose. Those diabetics showing complete absence of the acute insulin release after glucose showed the AID most clearly (X̄Δ = -24 ± 5%; n=15; P<0.001), suggesting that a preceding AIR, though vestigial, could obscure the AID in some diabetics. Indeed, in 4 diabetics each having an AIR but lacking a spontaneous AID after glucose, epinephrine infusion eliminated the AIR and allowed a latent AID to be unmasked in all 4. In contrast, in normal subjects, epinephrine infusion failed to unmask an AID to iv glucose pulses of either 20 (n=6) or 5 (n=11) g despite elimination of the preceeding AIR. To examine whether the AID was mediated by stimulation of α-adrenergic receptors, 5 diabetic subjects underwent α-adrenergic blockade with infusions of phentolamine (0.5 mg/min). Phentolamine failed to AID (X̄Δ during phentolamine = -14 ± 5%; P<0.05). These studies suggest that hyperglycemic diabetics respond to a sudden glucose challenge and, in most patients, this glucose stimulus is paradoxically translated by the pancreatic islet as a signal to inhibit insulin secretion. Such a paradoxical response cannot be demonstrated in normal subjects before or during α-adrenergic stimulation when the AIR is absent. In conjunction with in vitro studies, these data suggest the hypothesis that the AID of diabetes may be related to endogenous production of a nonadrenergic, glucose-stimulated inhibitor of insulin secretion.