Effects of insulin and the combination of insulin plus metformin (glucophage) on microvascular reactivity in control and diabetic hamsters.

The purpose of this study was to determine the in vivo microvascular reactivity of arterioles (mean internal diameter range: 16.0 to 106.4 microm) and venules (mean internal diameter range: 24.0 to 117.3 microm) in the hamster cheek pouch to insulin and to the mixture insulin + metformin. Experiments were performed using an intravital microscope coupled to a closed-circuit TV system and a videotape. The TV monitor display was used to obtain arteriolar and venular internal diameter measurements by an image-shearing device. The studied drugs were applied topically, added to the superfusion solution, to avoid systemic effects that would complicate the analysis of the results. In control animals (glycemia 7.7 +/- 0.4 mmol/L), application of insulin (10 to 500 microU/mL/min) evoked vasodilatation in a dose-dependent fashion in arterioles (4.9 +/- 3.2% to 50.9 +/- 6.5%, smallest and largest concentration, respectively, values expressed in percent of the initial diameter as mean +/- SE) and venules (-2.1 +/- 3.1% to 14.3 +/- 5.1%), decreased and finally abolished the spontaneous vasomotion frequency (from 9.5 +/- 0.3 cycles per minute [cpm] to 0.0 +/- 0.0 cpm) and amplitude (from 8.6 +/- 0.3 to 0.0 +/- 0.0 microm). Addition of metformin, 0.2 mg/mL/min, did not significantly change either the observed vasodilatation in arterioles and venules or the vasomotion frequency and amplitude curves. Two types of diabetic hamsters were studied: severely diabetic, induced with three intraperitoneal injections of streptozotocin, diluted in physiological saline, 50 mg/kg/dose, given in three consecutive days, and mildly diabetic, induced by a single dose of streptozotocin. All diabetic animals were studied four weeks after the onset of diabetes and no specific treatment for diabetes was given. In severely diabetic hamsters (glycemia 18.0 +/- 2.2 mmol/L), application of insulin, in the same concentration range, evoked a significantly reduced vasodilatation in arterioles as compared with control animals (5.9 +/- 1.3% to 18.9 +/- 3.5%) and did not change the vasodilatation observed in the venules (5.9 +/- 1.4% to 21.3 +/- 2.5%). In these preparations no spontaneous arteriolar vasomotion could be detected. Addition of metformin did not significantly improve the impaired vasodilatation. In mildly diabetic hamsters (glycemia 12.1 +/- 0.8 mmol/L), application of insulin, in the same concentration range, evoked vasodilatation, in a dose-dependent fashion, equivalent to the one observed in control animals, in arterioles (3.1 +/- 2.5% to 53.4 +/- 10.0%) and venules (7.1 +/- 3.0% to 29.9 +/- 4.8%) and also reduced the vasomotion frequency (from 10.1 +/- 0.3 to 0.1 +/- 0.1 cpm) and amplitude (from 9.2 +/- 0.6 to 0.2 +/- 0.2 microm). Addition of metformin tended to increase the observed arteriolar dilatation (6.6 +/- 3.0% to 67.8 +/- 5.5%), did not change the venular dilatation (6.7 +/- 4.8% to 28.0 +/- 3.3%), and tended to preserve vasomotion frequency and amplitude. These experiments show that (1) insulin has a direct dilatatory effect on arterioles and venules; (2) the vasodilatation evoked by insulin is impaired in severe diabetes, and (3) no significant abnormality could be detected on microvascular reactivity in mild diabetes. Further addition of metformin helped to maintain the spontaneous arteriolar vasomotion even during moderate vasodilatation and tended to augment the arteriolar dilatation evoked by insulin in mildly diabetic animals.