Hypoxic vasodilatation: is an adenosine–prostaglandins–NO signalling cascade involved?

During hypoxia blood vessels in various tissues including skeletal muscle, heart and brain dilate in order to increase oxygen delivery to the hypoxic organ. The mechanism has long been of interest, but is still incompletely understood. Adenosine is known to be involved; it is released from cells during hypoxia in the metabolic regulation of blood flow (Berne, 1963). There is also evidence for an involvement of endothelial prostaglandins, nitric oxide (NO) and ATP (from which adenosine can be formed following ectoenzymatic degradation), and smooth muscle K⁺ channels in hypoxia-induced vasodilatation (e.g. Busse et al. 1984; Messina et al. 1992). To date there have been few attempts to reconcile these different mechanisms.

For a number of years the Marshall group have been studying mechanisms of hypoxic vasodilatation in rat skeletal muscle. They have shown that adenosine A₁ receptors are involved (Bryan & Marshall, 1999). Because both the hypoxic response and vasodilatation to adenosine were blocked by a NO synthase (NOS) inhibitor, the authors suggested that NO could be the common mediator of vasodilatation, released following activation of A₁ receptors located principally on endothelial cells. These findings are interesting since in most blood vessels A₁ receptors are not located on the endothelium, but rather on smooth muscle or perivascular nerves, and A₂ receptors are the dominant subtype involved in adenosine-induced vasodilatation. In this issue of The Journal of Physiology, Ray et al. (2002) have used an integrated approach to investigate further mechanisms of adenosine and hypoxic vasodilatation. This study is important as it suggests a unifying hypothesis for the involvement of adenosine, prostaglandins and NO in hypoxia-induced vasodilatation.

In a comprehensive series of in vivo and in vitro experiments, Ray et al. (2002) have described a cascade of interactions in adenosine-mediated signalling and hypoxia. Firstly, in skeletal muscle in vivo, both the hypoxic vasodilatation and A₁ receptor-mediated vasodilatation were shown to be dependent on prostaglandin synthesis. The involvement of endogenous NO in adenosine signalling was investigated in complementary experiments in vitro, which measured directly NO release from rat aorta using an NO-sensitive electrode. Ray et al. (2002) showed that A₁ receptor activation stimulated NO release and this was blocked by removal of the endothelium, indicating that A₁ receptors are expressed on the endothelium. Selective blockade of adenosine receptors, prostaglandin synthesis and adenylyl cyclase showed that A₁ receptor activation evokes a release of prostaglandins (especially PGI₂) from rat aorta, and that NO release occurs downstream of prostaglandin formation and generation of cAMP (the second messenger for prostaglandins). Furthermore, iloprost, a stable analogue of PGI₂ and skeletal muscle vasodilator, was shown to stimulate NO formation in rat aorta, confirming that prostaglandins could act as intermediates in the proposed cascade. The proposed signalling cascade is thus that in hypoxia adenosine is released, which activates A₁ receptors on endothelial cells, which causes synthesis of prostaglandins and activation of cAMP, which in turn stimulates NO production and vasodilatation.

However, the involvement of this mechanism in vasodilatation is not entirely supported by the literature. Studies by Prentice & Hourani (1996) have shown that A_2A receptors are the dominant subtype of adenosine receptor involved in vasorelaxation in rat aorta, with an additional possible involvement of A_2B receptors. Furthermore, there is evidence against the existence of vasorelaxant A₁ receptors in rat aorta (Lewis et al. 1994). Thus, at least a degree of caution should be used in extrapolating these signalling interactions, which have no proven significance for adenosine-mediated vasodilatation in rat aorta, to vasodilatation in rat skeletal muscle. A limitation is that the aorta is not a resistance vessel and, therefore, may not be appropriate as a model for investigating mechanisms of hypoxic vasodilatation in skeletal muscle.

An important question is how general the adenosine-prostaglandins-NO signalling cascade is as a mechanism for adenosine and hypoxic vasodilatation. Even allowing for synergistic interactions between prostaglandins and NO signalling, if the cascade is operational in a tissue, inhibitors of both prostaglandin synthesis and NOS should be effective. However, in isolated middle cerebral arteries, endothelium-dependent hypoxic vasodilatation was inhibited by indomethacin, but was unaffected by NOS inhibition, indicating that prostaglandins, but not NO, are involved (Fredricks et al. 1994). Furthermore, in rat cremaster skeletal muscle, indomethacin and endothelium removal inhibited completely the response to hypoxia, but adenosine caused endothelium-, NO- and prostaglandins-independent vasodilatation, indicating that in rat cremaster skeletal muscle adenosine is not a mediator of hypoxic vasodilatation (Koller et al. 1989; Messina et al. 1992). Thus, it is unclear to what extent the mechanism proposed by Ray et al. (2002) operates in other tissues.

In summary, the authors have an attractive framework for the mechanism of hypoxic vasodilatation in rat hindlimb skeletal muscle. An important unresolved issue is whether there is reproducibility of NO release in the study's experimental protocol of repeated challenges with adenosine (these controls are not reported). The mechanism is elaborated in rat aorta, but paradoxically A₁ receptors have been found not to be involved in vasorelaxation in this blood vessel. Future studies will be required to resolve these issues.