Microvascular haemodynamic reactions to insulin

A recent article published in The Journal of Physiology adds further information to the vast pioneering investigations performed for many years by Clark's group to unravel the role possibly played by microvascular reactivity defects in insulin resistance (Newman et al. 2009). This new contribution deals with arteriolar vasomotion as a factor involved in insulin's microcirculatory effects to increase blood delivery to muscle tissue. While we certainly agree that vasomotion is a mechanism whereby more capillaries can be perfused per unit time, we must disagree with two major claims stated by the authors: (1) that there would ‘have been no studies in experimental animals where vasomotion in muscle microvasculature has been determined’ and (2) that ‘this is the first report of insulin acting to increase the myogenic component of vasomotion in muscle’.

Skeletal muscle vasomotion has been studied by various groups for more than 20 years. Concerning more particularly insulin, and as the authors are well aware of, we have indeed published as early as 1998 the dynamic changes induced by insulin effects on rat skeletal muscle (spinotrapezius) microvessels, as quantitated by intravital microscopy in situ (Renaudin et al. 1998); at that time we already proposed precapillary, arteriolar vasomotion as being a cardinal determinant of the microvascular effects of the hormone to promote rapid and efficient glucose delivery and storage during postprandial periods. This hypothesis and supporting data have been repeatedly published since then by our groups (Wiernsperger & Bouskela, 2003; Wiernsperger et al. 2007) and even referred to by Clark's group on several previous occasions (Vincent et al. 2005; Clark, 2008). It is thus puzzling why any reference to our previous work has now been omitted in their new publication and why such claims have been made. Evidently we would have appreciated the authors giving tribute to our pioneering work in this precise domain.

We deeply feel that any published work is no more than another brick in a wall and that data, whatever they look like, must be discussed in the light of previous concurrent work and that tentative analysis of discrepant data is a way to progress in our understandings. Actually this is the basis of the discussion ending any scientific paper. In many dozens of experiments in which spontaneous vasomotion was present in the basal state (our criterion for quality of preparation), we have consistently found arteriolar dilatation and disappearance of arteriolar vasomotion in rat skeletal muscle with insulin alone (even at low concentrations), contrasting with activation of vasomotion when both insulin and glycaemia were elevated, as it occurs physiologically (Bouskela et al. 1997; Renaudin et al. 1998; Wiernsperger, 2009).This sharply contrasts with Newman's findings and with those of De Jongh (DeJongh et al. 2004); therefore, discussing these very differing findings would have been very important, since we know well that differences in microvascular reactivity among tissues (skin vs. muscle), in measuring techniques (LDF vs. videomicroscopy) or even in species (rat vs. human) may have to be considered (Wiernsperger, 2009). In particular, these differences should have been discussed in view of the protocol used, since Newman (Newman et al. 2009) and de Jongh (DeJongh et al. 2004) have used the clamp technique while we tried to mimick as close as possible the prevailing plasma levels found in postprandial situations. In particular, the numerous unphysiological aspects of the clamp should have been addressed in their discussion: importantly, an indication of the insulin plasma levels reached during the clamp procedure would have been worthwhile since, for example, high or supraphysiological insulin levels as occurring during such clamps can stimulate the sympathetic nervous system (Yki-Jarvinen & Utriainen, 1998). Finally, one should remind the reader that, since physiologically plasma insulin levels increase significantly only in the presence of concomitant elevations of glucose, isolated high levels of insulin have little, if any, physiological relevance.