The seminal observation of Furchgott and Zawadzki 20 years ago, that endothelial cells play an obligatory role in the relaxation evoked by acetylcholine in isolated rabbit aortas, truly revolutionized cardiovascular science. Today, endothelium-dependent regulation of vascular tone, platelet function and mitogenesis have become a key part of our view of cardiovascular physiology. Perhaps even more important, the achievements of the last two decades have led to a better understanding of the pathophysiology of hypertension, vasospasm and atherosclerosis and to new therapeutic strategies to fight these pathological conditions.
Indeed, a number of endothelium-derived factors can profoundly modify platelet function, as well as the contractile and proliferative state of vascular smooth muscle cells (1) (Fig. 1). These factors include nitric oxide (NO) and prostacyclin (PGI2), both vasodilators and potent inhibitors of platelet function, and a putative endothelium- derived hyperpolarizing factor. Soon after the discovery of NO, it became clear that endothelial cells can also mediate contraction. Endothelium-derived contracting factors (EDCF) include the 21 amino-acid peptide endothelin, vasoconstrictor prostanoids such as thromboxane A2 and prostaglandin H2, and angiotensin II. Exogenous arachidonic acid can produce endotheliumdependent vasoconstriction that can be prevented by indomethacin (1), suggesting that cyclooxygenase products can produce vasoconstriction. The cyclooxygenase cascade is also the source of superoxide anions (O2), which can mediate endothelium-dependent contractions either by enhancing the breakdown of NO or by directly stimulating vascular smooth muscle cells. The cyclooxygenase pathway therefore produces multiple potentially constricting factors and the use of more specific inhibitors than indomethacin (such as the thromboxane A2/prostaglandin H2 receptor antagonist SQ29548, and the O2 - scavenger superoxide dismutase) is helping to better determine their importance in vascular function. Such a selective pharmacological approach has emphasized the important role of O2- in the balance between endothelium-dependent contractions and relaxations (2) (Fig. 1).
Endothelial dysfunction, intended as the complex multifaced pathological product of different vasculotoxic agents or injuries, is viewed today as an attractant intermediate phenotype of cardiovascular diseases with usually long and unpredictable natural history. Furthermore, endothelial dysfunction may not only represent a vascular disease marker, but may actually play an important pathogenetic role, leading to progression of the disease and unfavourable outcomes. Among these vascular diseases, cerebrovascular accidents, namely stroke, clearly represent a paradigmatic example of the potential role of dysfunctional endothelium. In fact, in the world's growing elderly population few diseases are more dreaded than stroke. With an increasing incidence and mortality of 30%, stroke carries the threat of death or long-term disability and suffering. Elevated blood pressure has long been recognized as one of the most important risk factors for stroke; however, other factors appear to play an important role. Indeed, epidemiological evidence suggests that in spite of an improved control of blood pressure, the secular trends of stroke in well-controlled populations are increasing. Moreover, while the haemorrhagic subtype of stroke displays a continuous decrement over the years, ischaemic stroke is little affected by the current pharmacological treatment of hypertension (3). Finally, beyond the influence of environmental factors, there is growing evidence that familial predisposition, or hereditary factors, may contribute to the aetiology of stroke (4).
In this paper, we analyse current evidence suggesting that endothelial dysfunction can play a role in the pathogenesis of ischaemic stroke.
ENDOTHELIAL DYSFUNCTION AND CEREBRAL CIRCULATION
Several lines of evidence suggest that formation of NO occurs in cerebral blood vessels under basal conditions both in vitro and in vivo. Inhibitors of NO synthase decrease basal levels of cyclic guanosine monophosphate (cGMP), and produce contraction of cerebral arteries (5). In addition to exerting a tonic dilator effect on the cerebral circulation, basal release of NO may protect cerebral endothelium by inhibiting aggregation of platelets and leukocytes.
Impaired endothelium-dependent relaxation of cerebral blood vessels has been observed during chronic hypertension, diabetes, hypercholesterolemia, subarachnoid haemorrhage (6), ischaemia, and ageing. In all cases, this impairment appears to be specific for endothelium because vasodilation in response to endotheliumindependent agonists such as sodium nitroprusside is not impaired. The putative mechanisms that account for impairment of responses of cerebral vessels to endothelium- dependent agonists in the presence of pathophysiological states are now beginning to be uncovered. During chronic hypertension and diabetes, altered responses of cerebral arterioles appear to be a result of production of an EDCF that counteracts the normal dilator effect of NO. This EDCF appears to be a cyclooxygenase product of arachidonic acid metabolism that activates a prostaglandin H2/thromboxane A2 receptor. In contrast, impaired endothelium-dependent responses of the basilar artery may instead be due to reduced activity of NO synthase or enhanced breakdown of NO after its generation. Endothelium-dependent relaxations of large cerebral arteries are impaired after subarachnoid haemorrhage in experimental animals and humans (6). The mechanism that accounts for this impairment is not completely clear. Haemoglobin, which may play an important role in causing vasospasm after subarachnoid haemorrhage, may constrict cerebral arteries by inhibition of the basal effect of NO. Indeed, in cerebral arteries, haemoglobin decreases basal levels of cGMP to the same levels observed in the presence of NO synthase inhibitors or after removal of the endothelium. In addition, haemoglobin may inactivate endothelium-derived NO by generating superoxide anions. Ischaemia, followed by reperfusion, determines an impairment of endotheliumdependent responses of cerebral arterioles, which can be corrected with scavengers of oxygen-derived free radicals. Thus, formation of reactive oxygen species may represent an important mechanism underlying endothelial dysfunction following cerebral ischaemia.
We recently used an animal model to study whether the genetic predisposition to stroke was associated with impaired endothelial function, independently of the effects of blood pressure (7,8). The stroke-prone spontaneously hypertensive rat (SHRsp) is a unique model for investigating the mechanisms predisposing to stroke in hypertension. Similar to the human disease, stroke in the SHRsp appears to be a complex, polygenic and multifactorial trait. Expression of the morbid phenotype depends on both the presence of hypertension and interaction with environmental variables, such as exposure to a specific dietary regimen (Japanese-style diet). If these conditions are met, SHRsp show an extremely high rate of stroke, which occurs rapidly (within a few weeks to months). Lowering of blood pressure to normal levels will greatly reduce or eliminate the occurrence of stroke in both SHRsp and humans, and there is epidemiological data indicating that ethnic groups with particularly low potassium and high sodium intake show an excessively high incidence of cerebrovascular disease.
Previous studies comparing SHRsp with normotensive Wistar-Kyoto rats have suggested that functional or structural abnormalities of the vessel wall may be implicated in the high incidence of stroke in SHRsp. However, differences in blood pressure level between these two strains, a major confounding factor on vascular biology, precluded any definitive interpretation of these findings. To gain a better understanding of the possible role of altered vascular functioning in the pathogenesis of stroke, we determined standard parameters of endothelium- dependent and independent vasomotor responses in isolated arteries from SHRsp and SHR, using an experimental protocol that eliminated blood pressure as a confounding variable. Furthermore, all comparative studies among inbred disease and 'control' strains are limited by the possibility that observed differences may simply represent the expression of randomly fixed genetic differences between the two strains, which are irrelevant for the phenotype of interest. To address this issue, co-segregation studies were performed in which the demonstration of a continued association between the phenotype of interest and the phenomenon investigated provides support for a non-coincidental, and thus potentially causal relationship. We investigated vascular function in male SHR and SHRsp as well as in SHRsp/SHR-F2 hybrid animals (7,8). Endothelium-independent vascular reactivity of the thoracic aorta and basilar artery showed similar contractile and dilatory responses to serotonin and nitroglycerin, respectively, in all groups. By contrast, endothelium-dependent relaxations, in response to acetylcholine and substance P, were significantly reduced in SHRsp compared with SHR. Similarly, reduced vasodilatory responses were present in F2 hybrid animals that had suffered a stroke when compared with SHR or F2 resistant to stroke. These abnormalities of endotheliummediated vasorelaxation in SHRsp were only found when the animals were exposed to the stroke-permissive Japanese diet, thus confirming the importance of ecogenetic interaction in the pathogenesis of stroke. The observed association and co-segregation of stroke with significant and specific impairment of endotheliumdependent vasorelaxation among SHRsp and strokeprone F2 hybrids, suggest a primary role of altered endothelium-dependent vascular relaxation in the pathogenesis of stroke. On the other hand, initial studies performed in our laboratory to define the genetic basis of stroke in SHRsp by using a candidate gene approach did not reveal any significant co-segregation with this vascular phenotype (9), although a more extensive coverage of the genome in the SHR/SHRsp cross is currently under way. More relevant information for human stroke will be derived from our current studies performed through a systematic investigation of endothelial function in hypertensive patients with family history of stroke.
SUMMARY AND FUTURE DIRECTIONS
Endothelium produces NO under basal conditions and in response to a variety of vasoactive stimuli in large cerebral arteries and in the cerebral microcirculation. Endothelium-dependent relaxations are impaired in the presence of several pathophysiological conditions. This impairment may contribute to cerebral ischaemia or stroke. We have reviewed studies indicating that endothelial dysfunction is associated and co-segregate with genetic predisposition to stroke in hypertensive animal models. Most of the studies addressing the influence of endothelium-derived NO in cerebral vessels have relied exclusively on the use of inhibitors of NO synthase. Relatively little is known about factors that regulate NO synthase gene expression and activity in vivo. Further studies are certainly needed in this area to better characterize the link between familial predisposition to stroke and endothelial dysfunction, and define genes directly involved with the susceptibility to stroke. Such an achievement may permit more focused and novel therapeutic approaches for the care of this devastating disease. In particular, even the first gene therapy-based approaches to stroke have been directed at genes controlling cerebral vascular function such as NO synthase, and superoxide dismutase (10,11). Finally, the systematic assessment of markers of endothelial dysfunction in patients may be very useful to better characterize the diagnosis and prognostic outcomes with particular regard to susceptibility to stroke.
1. Luscher TF, Tanner FC. Endothelial regulation of vascular tone and growth. Am J Hypertens
2. Cosentino F, Sill JC, Katusic ZS. Role of superoxide anion in mediation of endothelium-dependent contractions. Hypertension
3. Chang CC, Chen CJ. Secular trend of mortality from cerebral infarction and cerebral hemorrhage in Taiwan, 1974-1988. Stroke
4. Diaz JF, Hachinski VC, Pederson LL, Donald A. Aggregation of multiple risk factors for stroke in siblings of patients with brain infarction and transient ischemic attacks. Stroke
5. Cosentino F, Sill JC, Katusic ZS. Endothelial L-arginine pathway and relaxations to vasopressin in canine basilar artery. Am J Physiol
6. Nagakomi T, Kassell NF, Sasaki T, Fujiwara S, Lehman RM, Torner JC. Impairment of endothelium-dependent vasodilation induced by acetylcholine and adenosine triphosphate following experimental subarachnoid hemorrhage. Stroke
7. Volpe M, Iaccarino G, Vecchione C, et al. Association and cosegregation of stroke with impaired endothelium-dependent vasorelaxation in stroke prone, spontaneously hypertensive rats. J Clin Invest
8. Russo R, Vecchione C, Cosentino F, et al. Impaired vasorelaxant responses to natriuretic peptides in the stroke-prone phenotype of spontaneously hypertensive rats. J Hypertens
9. Rubattu S, Giliberti R, Russo R, Gigante B, Ganten D, Volpe M. Analysis of genetic basis of endothelium-dependent impaired vasorelaxations in stroke prone spontaneously hypertensive rats: a candidate gene approach. J Hypertens
10. Chen AF, Jiang S-W, Crotty TB, et al. Effects of in vivo advential expression of recombinant endothelial nitric oxide synthase gene in cerebral arteries. Proc Natl Acad Sci USA
11. Lund DD, Faraci FM, Miller FJ Jr, Heistad DD. Gene transfer of endothelial nitric oxide synthase improves relaxation of carotid arteries from diabetic rabbits. Circulation