Abstract
Blood glucose and the prevalence of diabetes are lower in mountain than lowland dwellers, which could among other factors be due to reduced oxygen availability. To investigate metabolic adaptations to life under hypoxia, male mice on high fat diet (HFD) were continuously maintained at 10% O(2). At variance to preceding studies, the protocol was designed to dissect direct metabolic effects from such mediated indirectly via hypoxia-induced reductions in appetite and weight gain. This was achieved by two separate control groups on normal air, one with free access to HFD, and one fed restrictedly in order to obtain a weight curve matching that of hypoxia-exposed mice. Comparable body weight in restrictedly fed and hypoxic mice was achieved by similar reductions in calorie intake (-22%) and was associated with parallel effects on body composition as well as on circulating insulin, leptin, FGF-21, and adiponectin. Whereas the effects of hypoxia on the above parameters could thus be attributed entirely to blunted weight gain, hypoxia improved glucose homeostasis in part independently of body weight (fasted blood glucose, mmol/l: freely fed control, 10.2 ± 0.7; weight-matched control, 8.0 ± 0.3; hypoxia, 6.8 ± 0.2; p < 0.007 each; AUC in the glucose tolerance test, mol/l(*)min: freely fed control, 2.54 ± 0.15; weight-matched control, 1.86 ± 0.08; hypoxia, 1.67 ± 0.05; p < 0.05 each). Although counterintuitive to lowering of glycemia, insulin sensitivity appeared to be impaired in animals adapted to hypoxia: In the insulin tolerance test, hypoxia-treated mice started off with lower glycaemia than their weight-matched controls (initial blood glucose, mmol/l: freely fed control, 11.5 ± 0.7; weight-matched control, 9.4 ± 0.3; hypoxia, 8.1 ± 0.2; p < 0.02 each), but showed a weaker response to insulin (final blood glucose, mmol/l: freely fed control, 7.0 ± 0.3; weight-matched control, 4.5 ± 0.2; hypoxia, 5.5 ± 0.3; p < 0.01 each). Furthermore, hypoxia weight-independently reduced hepatic steatosis as normalized to total body fat, suggesting a shift in the relative distribution of triglycerides from liver to fat (mg/g liver triglycerides per g total fat mass: freely fed control, 10.3 ± 0.6; weight-matched control, 5.6 ± 0.3; hypoxia, 4.0 ± 0.2; p < 0.0004 each). The results show that exposure of HFD-fed mice to continuous hypoxia leads to a unique metabolic phenotype characterized by improved glucose homeostasis along with evidence for impaired rather than enhanced insulin sensitivity.