Homocysteine and endothelial function

See article of Lambert et al. [34] (pages 743–751) in this issue. The hypothesis that homocysteine may cause vascular disease was originally proposed by McCully, following the observation that premature thromboembolism and atherosclerosis was a feature of homocystinuria [1]. In homocystinuria, a rare inborn error of metabolism, plasma homocysteine concentrations are generally 10 to 50 times that found in the healthy population [2]. This led to the hypothesis that mild to moderate elevations of plasma homocysteine may be a risk factor for atherosclerosis in the general population. This hypothesis has generated considerable interest, particularly as simple therapy with oral B group vitamins significantly lowers plasma homocysteine [3]. Homocysteine is a thiol containing amino acid that is metabolised from methionine, an essential amino acid derived from dietary protein. Homocysteine can be metabolised further to cysteine by transsulfuration (using vitamin B6 as cofactor) or remethylated back to methionine using either methyltetrahydrofolate (and Vit B12 as cofactor) or betaine as co-substrate (see Fig. 1). In human plasma, homocysteine exists in several forms. Approximately 70–80% is bound to protein, mainly albumin, by a disulphide bond. The remaining homocysteine oxidises to form dimers (homocystine) or combines with cysteine to form a mixed disulphide. Only a small proportion (<1%) circulates as free homocysteine. A number of techniques are now available for the combined measurement of the multiple forms of plasma homocysteine, which is expressed as total homocysteine (tHcy) [4]. Metabolism of homocysteine. A ‘normal’ range for tHcy cannot be defined but various studies report fasting levels in the range 5–15 μmol/l in the healthy population. A number of factors, genetic and acquired, can influence the blood levels of tHcy. Genetic factors include the genotype for methylene tetrahydrofolate reductase (MTHFR) C677T mutation, which results in thermolability, reduced activity and consequently fasting hyperhomocysteinaemia, particularly when associated with a low folate diet [5]. Homozygosity for this TT variant is common constituting approximately 12% of the Caucasian population. Homocystinuria is usually caused by mutations in the cystathionine β synthase (CBS) gene. The significance of heterozygosity for CBS deficiency is less clear but it is possible that it may lead to hyperhomocysteinaemia following methionine loading [6]. Nutritional factors are also important as plasma tHcy levels are inversely related to plasma levels of vitamin B12, B6 (pyridoxal phosphate) and particularly folate [7]. Oral methionine loading was originally devised as an approach to diagnosing heterozygosity for CBS deficiency particularly with the use of assays which were not sensitive enough to accurately measure pre-load plasma homocysteine values [8]. The rationale is that a high flux of methionine will stress the CBS pathway and hence may unmask minor deficiencies in activity of the enzyme or its cofactor (pyridoxal phosphate). Thus it is often inferred that a high homocysteine level after methionine loading is an indication of defective transsulfuration whereas an elevated fasting homocysteine level is an indication of defective remethylation. However these assumptions have not been proven and indeed the methionine loading test will not reliably distinguish obligate CBS heterozygotes from normal individuals. The standard methionine loading test involves ingestion of 0.1 g/kg methionine with measurement of plasma homocysteine at baseline and after 6 h. Results may be expressed as fasting, post-load values and/or the increment in homocysteine. Extensive data on this test were obtained from the European Concerted Action Project with results from 800 controls and 750 vascular cases [9]. Upper limits of normal were set using the top quintile for controls at 12 μmol/l (preload/fasting) and 38 μmol/l (post load). Using these cut-offs it was concluded that relative risk for vascular patients was 1.6 for fasting Hcy, 1.5 for post load Hcy but increased to 2.5 for subjects with both raised fasting and post load values. This would suggest that methionine loading is a useful discriminatory test. However, the 6-h wait is inconvenient for both investigator and patient, and the sulphur-like taste of methionine makes this a somewhat unpleasant procedure for the patient. These drawbacks, when taken with the wide variability in post load values, has precluded its general use in clinical practice, in preference to fasting values. The possibility must remain that methionine loading may have the potential to unmask some ill-defined aspects of homocysteine metabolism which are pertinent to vascular disease. A number of case-control studies have established moderate hyperhomocysteinaemia as a risk factor for coronary, cerebral and peripheral vascular disease. A meta-analysis of 27 prospective and retrospective studies by Boushey et al. in 1995 supported a strong relationship between coronary artery disease (CAD) and hyperhomocysteinaemia [10]. In this analysis an independent graded response was found, whereby a 5 μmol/l increase in tHcy increased risk of CAD by 60% in males and 80% in females, equivalent to a cholesterol increase of 0.5 mmol/l. However, some of the studies which formed part of this analysis included control groups taken from a ‘healthier’ population than the cases with vascular disease which may have accentuated the apparent effect of homocysteine. Prospective studies show a more variable picture with an overall positive but relatively weaker effect of homocysteine on vascular disease [11–14], but with some notably negative studies [15–17]. A systematic overview of these prospective studies is required. The two genetic factors (described above) which are known to cause moderate hyperhomocysteinaemia (either fasting or after methionine loading) have not been shown to be associated with increased risk of vascular disease. Heterozygosity for CBS deficiency, with an approximate frequency of 1 in 70 of the population, is not associated with increased risk of CAD [18,19]. Similarly, homozygosity for the C677T mutation of MTHFR, which leads to an average increase in plasma homocysteine of approx. 25%, has also not been shown to be a risk factor for CAD [20,21]. This has lead some investigators to conclude that mild hyperhomocysteinaemia is not causally related to vascular disease. An alternative interpretation is that elevated tHcy may only confer cardiovascular risk when combined with other conventional risk factors [22]. Thus the homocysteine hypothesis remains controversial and further data on possible mechanisms of homocysteine-induced vascular damage would contribute to the scientific debate. The demonstration of a clear mechanism would strengthen the case for tHcy as a causal factor. Putative mechanisms whereby tHcy may induce vascular injury include endothelial dysfunction, smooth muscle proliferation, extracellular matrix modification, lipoprotein oxidation, cytotoxocity, and effects on platelets and coagulation [23]. Homocysteine has been shown to directly damage endothelial cells and increase proliferation of smooth muscle cells in vitro [24,25]. Impaired nitric oxide release by bovine endothelial cells has also been demonstrated [26]. Nitric oxide is a potent vasodilator and inhibits platelet aggregation. Impairment of release and/or effects of nitric oxide may partly modulate the thrombotic potential of hyperhomocysteinaemia. All these in vitro studies however suffer from the reservation that the concentration and form of homocysteine used may be non-physiological. Damage to the endothelium is considered to be a critical aspect of the atherosclerotic process and precedes overt manifestation of disease [27]. It is therefore probable that some if not all the actions of homocysteine are mediated via endothelial dysfunction and more specifically by affecting the release and/or action of nitric oxide. In man, hyperhomocysteinaemia has been shown to be independently associated with endothelial dysfunction, determined by impaired flow-mediated dilatation (FMD) in a small group of middle aged Chinese subjects (n=14; tHcy=34.8 μmol/l) with no other identifiable risk factors [28]. In a separate study impaired FMD was also demonstrated in 26 elderly patients with hyperhomocysteinaemia (19.2 μmol/l) who had no family history and no clinical manifestations of atherosclerosis [29]. Endothelial dysfunction has been demonstrated in children with classic homocystinuria, but not heterozygotes for CBS deficiency who had no elevation of free homocysteine levels and normal methionine levels [30]. In two recent studies, a standard methionine load (0.1 g/kg) was shown to impair FMD in normal healthy subjects in a time-dependent fashion [31,32]. Acute methionine loading has several metabolic effects including the elevation of plasma homocysteine so this is consistent with, but is not proof, of a toxic effect of homocysteine on endothelial function in man. Interestingly Hanratty et al. found no impairment of endothelial function in subjects fed methionine 0.1 g/kg daily for one week although tHcy levels remained significantly elevated [33]. This study used a different technique to assess brachial artery endothelial function (venous plethysmography and agonists rather than high resolution ultrasound and flow) so it is difficult to compare directly these findings with the acute methionine loading studies. In this edition of Cardiovascular Research a study by Lambert et al. examines the relationship between endothelial function, arterial distensibility and hyperhomocysteinaemia in 123 apparently healthy first degree relatives of 60 patients with premature clinical vascular disease (coronary, cerebral or peripheral) and hyperhomocysteinaemia as defined by ‘elevated’ tHcy after a methionine load test. [34]. Endothelial function was assessed using flow-mediated vasodilatation of the brachial artery, and the physical properties of the vessel wall assessed by measurement of arterial distensibility of the common carotid artery. Each subject had fasting tHcy and a standard methionine load test which was performed on a separate day to the vascular studies. In this group of apparently healthy subjects with mild hyperhomocysteinaemia, FMD of the brachial artery was not related to fasting tHcy in contrast to previous reports [28,29]. However, analysis of methionine load data from these patients showed a weak but statistically significant inverse relationship between FMD and δ-homocysteine (the increment in tHcy). Arterial distensibility measurements of the carotid artery were not related to plasma homocysteine either fasting or following a methionine load. Surprisingly, classical risk factors (male gender, total cholesterol, smoking status, blood pressure) were also not associated with FMD. Postmenopausal status, however, was independently related to FMD. In another study examining the association between FMD and classical risk factors in asymptomatic men and women by multiple stepwise regression analysis, reduced FMD was associated with cigarette smoking, older age, male gender and larger vessel diameter. A composite risk factor score was strongly related to FMD, suggesting risk factor interaction [35]. Assuming that post methionine hyperhomocysteinaemia alone is associated with vascular injury in the study subjects, what are the abnormalities of homocysteine metabolism that could cause it? Deficiency of vitamin B6 does not normally cause fasting hyperhomocysteinaemia but may be associated with elevated tHcy levels post methionine loading [36]. In Lambert’s study 28% of patients had Vit B6 levels below the ‘reference’ range (<17 nmol/l), but unfortunately no data are given regarding plasma vitamin status and FMD. Vit B6 level was inversely related to the post methionine level of tHcy (p=0.03) but not to δ-tHcy (p=0.07) or fasting tHcy (p=0.53). In a recent prospective study of CHD incidence, plasma pyridoxal 5′ phosphate (Vit B6) but not fasting tHcy was independently associated with CHD [17]. Another possibility is the presence of heterozygosity for CBS deficiency, which may cause methionine load intolerance with normal fasting levels. However as previously mentioned, studies have shown no impairment of FMD in subjects heterozygous for CBS deficiency [30]. Most experimental approaches have concentrated on investigating direct effects of Hcy on vascular components and particularly the endothelium. Such direct injury could therefore explain mechanisms by which fasting hyperhomocysteinaemia may promote atherothrombosis. However the mechanism of vascular injury seen in patients with normal fasting tHcy but elevated levels post-methionine, due to abnormalities of the transsulfuration pathway, is more obscure. In these individuals elevated tHcy is only seen after a non-physiological oral methionine dose usually in the order of 6–7 g. The US recommendation for daily methionine intake is 0.9 g and the estimated adult intake in Western countries is ∼2–2.5 g daily. Studies of post prandial tHcy levels have varied but increases of no more than 20% have been reported [37]. A recent study also showed no significant increase in tHcy or impairment of endothelial function after 1 month of 250 mg oral methionine four times daily [33]. Therefore individuals with methionine load intolerance would not be expected to experience high homocysteine levels with normal Western diets. The authors suggest that post methionine load hyperhomocysteinaemia is associated with endothelial injury distinct to that caused by elevated fasting tHcy levels. This is an interesting proposal but more evidence is required before it can be accepted. This study adds more experimental data on the link between homocysteine and vascular disease but raises more questions than it answers. We are still some way from understanding the effect (if any) of homocysteine on endothelial function and its link with vascular disease.