Intermittent hypoxia-induced glucose intolerance is abolished by α-adrenergic blockade or adrenal medullectomy

Abstract

Obstructive sleep apnea causes intermittent hypoxia (IH) during sleep and is associated with dysregulation of glucose metabolism. We developed a novel model of clinically realistic IH in mice to test the hypothesis that IH causes hyperglycemia, glucose intolerance, and insulin resistance via activation of the sympathetic nervous system. Mice were exposed to acute hypoxia of graded severity (21, 14, 10, and 7% O2) or to IH of graded frequency [oxygen desaturation index (ODI) of 0, 15, 30, or 60, SpO2 nadir 80%] for 30 min to measure levels of glucose fatty acids, glycerol, insulin, and lactate. Glucose tolerance tests and insulin tolerance tests were then performed under each hypoxia condition. Next, we examined these outcomes in mice that were administered phentolamine (α-adrenergic blockade) or propranolol (β-adrenergic blockade) or that underwent adrenal medullectomy before IH exposure. In all experiments, mice were maintained in a thermoneutral environment. Sustained and IH induced hyperglycemia, glucose intolerance, and insulin resistance in a dose-dependent fashion. Only severe hypoxia (7% O2) increased lactate, and only frequent IH (ODI 60) increased plasma fatty acids. Phentolamine or adrenal medullectomy both prevented IH-induced hyperglycemia and glucose intolerance. IH inhibited glucose-stimulated insulin secretion, and phentolamine prevented the inhibition. Propranolol had no effect on glucose metabolism but abolished IH-induced lipolysis. IH-induced insulin resistance was not affected by any intervention. Acutely hypoxia causes hyperglycemia, glucose intolerance, and insulin resistance in a dose-dependent manner. During IH, circulating catecholamines act upon α-adrenoreceptors to cause hyperglycemia and glucose intolerance.

Keywords: sympathetic; thermoneutral.

Fig. 1.

Intermittent hypoxia (IH) protocol. Representative tracing of cage O₂%, blood pressure (BP), and heart rate (HR) during 30 min of IH, simulating an oxygen desaturation index (ODI) of 15, 30, or 60 events/h. Each desaturation lasted 1 min with a nadir of 10–12%, which resulted in a Sp_O2 nadir of 80%.

Fig. 2.

Effect of sustained hypoxia (SH) or IH on plasma glucose, free fatty acid (FFA), glycerol, insulin, and lactate. Both forms of hypoxia caused dose-dependent increases in glucose, whereas only ODI 60 increased FFA levels. Severe SH increased lactate. *P < 0.001 vs. 21% O₂; **P < 0.01 vs. 21% O₂; #P < 0.01 vs. ODI 0; ¥P < 0.05 vs. ODI 0.

Fig. 3.

Time course of glucose, FFA, and lactate levels during 5 h of 10% O₂ SH or IH (ODI 60). Glucose was elevated by both forms of hypoxia, with the greatest increase observed in the first 2 h (P < 0.05 for ODI 0 vs. ODI 60 and SH for the 1.5-h time point). A similar early rise in FFA was observed during IH (P < 0.05 for ODI 0 vs. ODI 60 at the 1.5- and 3-h time points). SH decreased FFA relative to the other conditions at 5 h (P < 0.01). Lactate was increased only by SH (P < 0.05 for SH vs. other groups at 1.5-, 3-, and 5-h time points).

Fig. 4.

Effect of SH on glucose tolerance [ip glucose tolerance test (GTT)] and insulin sensitivity [insulin tolerance test (ITT)]. The glucose area under the curve (AUC) was used for statistical analysis. The degree of glucose intolerance and insulin resistance correlated with severity of hypoxia. GTT AUC: ANOVA P < 0.0001; *P < 0.05 vs. 15 or 10% O₂. ITT AUC: ANOVA P < 0.0001; **P < 0.01 vs. other groups; ¥P < 0.01 vs. ODI 30 or ODI 60.

Fig. 5.

Effect of IH on glucose tolerance (ip GTT) and IH insulin sensitivity (ip ITT). The degree of glucose intolerance and insulin resistance correlated with ODI. GTT AUC: ANOVA P < 0.001; *P < 0.01 vs. ODI 30 or ODI 60; ¥P < 0.01 vs. ODI 30 or ODI 60. ITT AUC: ANOVA P = 0.006; **P < 0.01 vs. ODI 0 or ODI 15.

Fig. 6.

Effects of α-adrenoreceptor (AR) blockade with phentolamine. IH caused hyperglycemia and glucose intolerance (*P < 0.01 vs. ODI 0), and both effects were abolished by phentolamine. Phentolamine independently increased glucose tolerance (P < 0.001) and insulin sensitivity (P < 0.001). Phentolamine did not modify IH-induced lipolysis or insulin resistance.

Fig. 7.

Effect of β-AR blockade with propranolol. Propranolol abolished IH-induced FFA elevation but did not affect any glucose parameters. *P < 0.01 vs. ODI 0.

Fig. 8.

Effect of adrenal medullectomy. Medullectomy independently decreased glucose (P < 0.0001) and increased glucose tolerance (P < 0.01) and prevented IH-induced hyperglycemia, glucose intolerance, and lipolysis without affecting IH-induced insulin resistance. *P < 0.05 vs. ODI 0.

Fig. 9.

Glucose-stimulated insulin secretion. IH decreased insulin secretion (*P < 0.05, ODI 60 vs. ODI 0), whereas phentolamine independently increased insulin secretion (P < 0.0001) and insulin secretion during IH (¥P < 0.05, ODI 60 vs. ODI 0).

Fig. 10.

Effects of acute IH on glucose homeostasis. IH stimulates arterial chemoreceptors that in turn activate the sympathetic nervous system (SNS). Downstream activation of α-AR in the pancreas and liver, respectively, inhibit insulin secretion and stimulate glycogenolysis. Activation of β-AR stimulates white adipose tissue lipolysis. The net effect of IH is glucose intolerance, hyperglycemia, and FFA elevation. Similar pathways have been demonstrated for other forms of acute stress, including SH.