Endothelial cells possess vasodilator function, as well as anti-thrombotic and anti-inflammatory functions. This summary lecture reviewed the importance of endothelial dysfunction, methods to assess endothelial function in humans and their limitations, and therapeutic implications.

Endothelial cells synthesize at least three different vasodilator factors: nitric oxide, prostacyclin, and an unidentified endothelium-derived hyperpolarizing factor (EDHF), as shown in Figure 1. Under several pathological conditions, endothelial cells also synthesize several vasoconstricting factors (EDCF), including endothelin, superoxide and vasoconstrictor prostaglandin (Fig. 1).

Animal and human studies have demonstrated several mechanisms are involved in endothelial dysfunction (ED) in atherosclerosis (Fig. 2), including:

• reduced or impaired endothelial signal transduction
• reduced availability of L-arginine
• reduced eNOS expression
• reduced co-factor for eNOS
• NO inactivation increased by superoxide anion derived from macrophages other inflammatory cells, endothelial cells
• concomitant release of EDCF
• intimal thickening (possible diffusion barrier)
• smooth muscle vascular response

Under various conditions, including the presence of coronary risk factors, coronary artery disease (CAD), heart failure, or left ventricular (LV) hypertrophy, NO-mediated responses are impaired under basal conditions or in response to acetylcholine or other stimuli, such as metabolic or exercise stimuli.

ED appears to have prognostic value for patients with ischemic heart disease (IHD). A group from Germany reported that with good coronary vasodilator responses to increased flow, prognosis is very good. But, with impaired flow-mediated vasodilation the prognosis worsens. Flow-mediated dilatation greater than 15% is associated with a 90% survival rate at 10 years, 11-19% flow-mediated dilatation with a 90% survival, and less than 11% flow-mediated dilatation about a 60% survival. The group with less than 11% flow-mediated dilatation also had impaired endothelium-independent relaxation to nitrovasodilators. Shimokawa thinks it remains unproven whether ED truly has prognostic value, for example in patients with IHD or heart failure.

In vitro analysis of isolated human blood vessels

Endothelial function can be assessed in freshly isolated blood vessels from humans. Bradykinin caused greater relaxation in the large arteries and resistant small mesenteric arteries, compared to acetylcholine, in one study of endothelium-dependent relaxation (EDR). The bradykinin-induced EDR in the large arteries was due to the combined effect of NO and EDHF, as shown by one-half being sensitive to L-arginine and one half to KCl in the presence of indomethacin and L-arginine. However, in the microvessels bradykinin-induced EDR was insensitive to L-arginine and highly sensitive to KCl in the presence of indomethacin and L-arginine, indicating the relaxation was largely mediated by EDHF but not NO.

In vivo non-invasive assessment

Ultrasound is used to measure flow-mediated dilatation for non-invasive in vivo analysis, and is particularly useful in the clinical setting. By continuously measuring the brachial artery diameter change in response to occlusion and reflow, as demonstrated in Figure 3, the flow-mediated EDR can be assessed. The disadvantage of this technique is that the relation between brachial artery response and coronary artery response is unclear. There is ongoing controversy about whether this brachial artery response truly represents the coronary artery response, and more studies are needed.

In vivo invasive assessment

Invasive in vivo analyses use 1) venous plethysmography of forearm circulation (Fig. 4) and 2) quantitative angiography (QCA) and Doppler flow wire assessment of coronary circulation (Fig. 5). More accurate evaluation of endothelial function is obtained with these techniques. Combining QCA and Doppler flow wire allows continuous monitoring of the changes in diameter in large coronary arteries and in coronary blood flow, thus assesses both large epicardial coronary artery endothelial function and microvascular endothelial function.

Biochemical markers

Several biochemical markers could be used as an indicator of endothelial function: NOx plasma level, prostacyclin, thrombomodulin von Willebrand factor, and tissue factor pathway inhibitor, among others. A combination of these markers may help to assess endothelial function systemically.

The relative contribution of NO, EDHF and prostacyclin must be considered.
A study by Shimokawa in porcine large epicardial arteries and in coronary microvessels showed that 1) prostacyclin contribution was minimal in large and small arteries, 2) NO contribution was greater in large coronary arteries, and 3) EDHF contribution was greater in smaller vessels. Further, in pig, rabbit, mice, rat and human vessels he demonstrated 1) NO contribution was greater in large blood vessels, 2) EDHF was greater in microvessels. This observation must be considered when assessing endothelial function, even in humans.

To assess the effect of shear stress on endothelial function, an in vitro system to continuously monitor the change in diameter in isolated blood vessels and change perfusion pressure and blood flow was developed by Shimokawa. NO and EDHF contributed to shear stress-induced EDR in the large rat arteries, but EDHF contribution was greater than NO in the rat arteriole.

Contribution of vascular smooth muscle responses

Vasospasm and ED must be differentiated when assessing endothelial function in humans. In a coronary angiography study, no stenotic lesion was seen under basal conditions. A marked hyperconstriction (vasospastic response) was observed after intracoronary administration of acetylcholine. This response disappears after intracoronary nitroglycerine administration. This phenomenon might be considered ED.

However, in an animal model of coronary hyperconstriction (developed by chronic treatment of the porcine coronary artery with the inflammatory cytokine interleukin-1beta), marked hyperconstriction occurred after serotonin stimulation.
Surprisingly, bradykinin- or calcium ionophore-induced EDR was fairly preserved at the site of vasospasm. Responses were fairly preserved to the vasospastic agonist serotonin.

Direct evaluation of EDR in response to serotonin was possible because ketanserin, a 5HT2A serotonergic receptor blocker, was used to directly inhibit vascular smooth muscle contraction. This contraction was significantly augmented at the spastic segment in blood vessels without endothelial cells in response to serotonin, compared to control segments. The hypercontraction responses of the coronary artery were achieved mainly by a hypercontraction of vascular smooth muscle, rather than by ED, at least in their porcine model.

Variable responses to different agonists

Endothelial dysfunction does not occur uniformly to all agonists, but rather in a step-wise manner. At least two signal transduction pathways are involved with the production of NO from endothelial cells. One is mediated by the Gq-protein and the other by the Gi-protein, as demonstrated by Shimokawa. Serotonin, norephinephrine, and endothelin are among the agonists that use the Gi-protein mediated pathway. ED occurs in a step-wise manner, rather than uniformly, during the process of atherosclerosis, also demonstrated by Shimokawa. In the early stage of atherosclerosis, the Gi-protein mediated pathway becomes impaired, and in the middle of atherosclerosis pathways mediated by other G proteins are impaired. However, until the advanced stage of atherosclerosis eNOS function is preserved or is increased in atherosclerotic endothelial cells.

Studies of the effect of hypercholesterolemia on vasodilator function in animals and humans show that EDR to acetylcholine or serotonin is easily impaired. But, in response to bradykinin or calcium ionophore it is fairly preserved. Therefore, if possible, at least two agonists must be used to completely test endothelial function. Aging impairs endothelial vasodilator effects. Studies show that aging easily impairs EDR to acetylcholine in humans, but that bradykinin-induced EDR is not significantly influenced by age.

Correction of underlying risk factors, such as hyperlipidemia, hypertension, diabetes, and smoking, has been demonstrated to improve EDR in patients with IHD or heart failure. The use of fish oil, estrogen replacement therapy, ACE inhibitors, anti-oxidant agents, and probably L-arginine are also beneficial.

Fish oil studies

EDR was normalized in patients with CAD treated for 6 weeks with eicos-spentacnole acid (EPA), a major component of fish oil. Venous plethysmography studies showed that forearm vasodilation responses to acetylcholine and substance P were impaired before EPA treatment, but were normalized after treatment. Interestingly, the acute administration of L-NMMAæ abolished the improved EDR in response to acetylcholine, but did not significantly reduce the relaxation in response to substance P, indicating that EDR mediated by NO and EDHF is impaired in patients with CAD. Long-term treatment with EPA may improve both NO-mediated and EDHF-mediated relaxation.

EPA treatment can improve endothelial vasodilator functions in human coronary circulation of patients with IHD. Shimokawa demonstrated that long-term treatment with fish oil markedly augmented the EDR of the coronary artery in response to aggregating platelets in both the porcine model and humans. Before EPA treatment, the patients showed no increase in coronary flow in response to acetylcholine, indicating very impaired coronary vasodilator responses. However, after 6-8 weeks of EPA treatment, the blood flow response was normalized.

Estrogen study

Estrogen treatment improved NO-mediated and probably EDHF-mediated relaxation in post-menopausal women. Estrogen normalized forearm blood pressure responses to acetylcholine and substance P in post-menopausal women. Acute administration of L-NMMA in the forearm circulation inhibited the acetylcholine-induced increase in flow, but not the substance P-induced increase.

The relative contribution of prostacyclin, NO and EDHF to EDR, based on Shimokawa§fs view is:

• NO and EDHF mainly contribute to EDR under normal conditions, although there are many redundancies between NO and EDHF.

• NO-mediated relaxation is impaired first during atherosclerosis with several coronary risk factors.

• EDHF-responses seem to be upregulated for a while to maintain EDR.

• EDHF-responses become impaired as atherosclerosis progresses.

• NO- and EDHF-mediated EDR is reduced at the most advanced stage of atherosclerosis.

• Prostacyclin seems to play a compensatory role.

• NO-mediated and then EDHF-mediated responses are recovered with risk factor correction or drug treatment.

To examine the prognostic value of ED, the ENCORE trial is being conducted. ENCORE I is addressing the relatively short-term effects (6 months) of treatment with nifedipine, cerivastatin, or both, to determine their ability to improve endothelial vasodilator function, evaluated by QCA and Doppler flow wire. ENCORE II is addressing the long-term (2 years) effect of nifedipine and simvastatin using QCA and IVUS. This trial will determine the clinical importance of ED and the importance of its correction on the structural changes of the coronary artery and the prognosis in patients with IHD. Future trials will determine the importance of ED in humans.