Serum thiols and cardiovascular risk scores: a combined assessment of transsulfuration pathway components and substrate/product ratios

Despite significant advances in the understanding of the processes underlying the onset and progression of atherosclerosis and thrombosis a significant proportion of cardiovascular events cannot be predicted with the available risk factors. This suggests that additional pathophysiological mechanisms are involved. Potential biomarkers of cardiovascular risk should be measurable in the population. Moreover, they should be significantly related with cardiovascular risk and modifiable by means of pharmacological and/or non-pharmacological interventions [1].

The highly reactive sulphur-containing amino acid homocysteine has long been shown to exert detrimental effects on vascular homeostasis by inhibiting nitric oxide synthesis and promoting oxidative stress and inflammation [2]. Several studies have shown that higher plasma/serum homocysteine concentrations independently predict adverse cardiovascular outcomes and improve risk reclassification [2, 3]. Homocysteine is the initial step of the transsulfuration pathway [4]. The latter plays a crucial role, not only for homocysteine clearance but also for the synthesis of important homeostatic thiols such as cysteine, taurine and the natural antioxidant glutathione (Figure 1) [4].

Several studies have reported associations between plasma/serum transsulfuration pathway thiol concentrations and surrogate markers of cardiovascular disease, e.g. endothelium-dependent vasodilatation, intima-media thickness, and plaque burden [5‐7]. Homocysteine, cysteine and glutathione also play an important role in modulating specific atherosclerotic processes, e.g. oxidation of low-density lipoproteins (LDL) and inflammation [8, 9]. However, little knowledge is currently available on whether there are independent associations between individual transsulfuration pathway components and robust measures of cardiovascular risk for primary prevention, calculated using validated scoring systems at population level. Ideally, human studies investigating such associations should further account for clinical confounders affecting both thiol concentrations and cardiovascular risk [10, 11]. The identification of associations with scores directly predictive of cardiovascular risk would further enhance the clinical relevance, and potential use, of thiols as biomarkers. We addressed this issue by examining the combined associations between transsulfuration pathway thiols and absolute cardiovascular risk scores at population level, in an established epidemiological cohort of human ageing. Associations between thiols and cardiovascular risk scores were evaluated looking both at individual thiols and at substrate/product ratios within the transsulfuration pathway. Substrate/product ratios were considered markers of the activity of specific enzymes in the pathway (Figure 1). Analyses were further adjusted for a recently proposed cardiovascular risk marker, C-reactive protein, as well as body mass index, estimated glomerular filtration rate and mean daily step count, a marker of physical activity.

Clinical, demographic and biochemical characteristics of the study population are described in Table 1. The distribution of the cardiovascular risk scores was markedly skewed towards the left (Figure 2). Therefore, log transformed cardiovascular risk scores were entered into linear regression analysis.

Our study showed significant independent associations between several transsulfuration pathway thiols and five-year cardiovascular risk, measured using the Framingham Risk Equation for primary prevention, in an established cohort of human aging. Serum concentrations of homocysteine showed positive associations with risk scores whereas cysteine and glutathione were negatively associated. Further assessment of substrate/product ratios within the transsulfuration pathway suggests that a reduced biotransformation of homocysteine into cysteine and an increased degradation of glutathione into cysteinylglycine are independently associated with higher cardiovascular risk scores and might play an important role in this context. Associations between thiols, substrate/product ratios and cardiovascular risk scores were independent of important clinical confounders, e.g. body mass index, renal function, physical activity and C-reactive protein, another cardiovascular biomarker.

The independent associations between higher homocysteine concentrations and cardiovascular risk support the existing evidence on the detrimental effects of this thiol on endothelial function and vascular homeostasis [2]. A recent study has further highlighted the role of homocysteine as independent predictor of cardiovascular events as well as a tool to improve risk classification [3]. Homocysteine lowering treatment, e.g. B-vitamin supplementation with folic acid, has been proposed as a safe and relatively inexpensive strategy to reduce cardiovascular risk [2]. However, a number of randomised placebo-controlled trials have failed to show any significant effect of homocysteine lowering on cardiovascular morbidity and mortality in different patient groups [21]. Notably, these studies were conducted in patients with pre-existing cardiovascular disease [21]. It has been suggested that patients in these studies had fairly advanced atherosclerotic disease associated with significant vascular structural and functional abnormalities, largely irreversible despite homocysteine lowering [22]. Moreover, concomitant treatment with other drugs for secondary prevention, e.g. aspirin and statins, might have masked the potential beneficial effects of homocysteine lowering treatment, reducing the possibility of observing a difference in cardiovascular outcomes between the active and placebo groups [23, 24]. It remains largely unknown whether homocysteine lowering interventions conducted at an earlier stage, as part of a primary prevention strategy, might yield a different result [24]. The strong association between serum homocysteine concentrations and cardiovascular risk scores in patients without previous cardiovascular events in our study should prompt further intervention trials in similar populations. However, our study also suggests that other thiols within the transsulfuration pathway might play a role in modulating cardiovascular risk.

Higher serum cysteine concentrations have shown significant associations with surrogate markers of vascular disease, e.g. reduced endothelium-dependent vasodilatation [6]. By contrast, other studies have failed to demonstrate a significant role of cysteine in the pathophysiology of atherosclerosis after adjusting for several clinical and demographic confounders [5, 25]. Moreover, no associations between cysteine concentrations and cardiovascular mortality were observed in a longitudinal study [26]. The contrasting results of these studies suggest the presence of a complex relationship between cysteine, atherosclerosis and cardiovascular disease. At least two factors might explain such relationship. First, a number of studies have reported significant associations between higher cysteine concentrations, higher body mass index and reduced glomerular filtration rate [10, 11]. Univariate analyses in our study showed no associations between cysteine and cardiovascular risk scores. However, there were significant associations between cysteine, body mass index and glomerular filtration rate, in line with the available evidence [10, 11]. This suggests that further studies on the potential role of cysteine in modulating cardiovascular risk should account for these clinical confounders in regression analysis. Second, the effects of cysteine on atherosclerosis might be concentration-dependent. A study has demonstrated that cysteine modulates the copper and iron-dependent oxidation of low-density lipoproteins (LDL), an important step in atherogenesis [8, 27]. Cysteine concentrations ≥50 μmol/L fully prevented copper-dependent LDL oxidation. By contrast, concentrations between 25 and 1,000 μmol/L facilitated iron-dependent LDL oxidation in a concentration-dependent manner. The mean serum cysteine concentrations in our study, 183 μmol/L, suggest that the protective effects of cysteine on copper-dependent LDL oxidation might prevail over its pro-oxidative effects. This might explain the negative association between cysteine and cardiovascular risk scores. The potential anti-atherogenic effects of cysteine are further supported by recent evidence demonstrating that cysteine reduced calcification of atherosclerotic lesions in apoE-deficient mice and exerted significant anti-inflammatory effects in human coronary endothelial cells [28, 29].

The contrasting associations between homocysteine, cysteine and cardiovascular risk are supported by the finding of a positive independent association between homocysteine/cysteine ratios and risk scores. The biotransformation of homocysteine into cysteine is catalyzed by the enzymes cystathionine β-synthase and cystathionine γ-lyase [30]. Therefore, associations between higher homocysteine/cysteine ratios and cardiovascular risk suggest reduced expression and/or activity of these enzymes. Cystathionine β-synthase is the first, rate-limiting, enzyme in the transsulfuration pathway. Congenital cystathionine β-synthase deficiency in humans and knock-out animal models are both typically associated with hyperhomocysteinemia, premature atherosclerosis and vascular disease [31]. Both cystathionine β-synthase and cystathionine γ-lyase are also involved in the synthesis of hydrogen sulphide. The latter is increasingly recognised as having potential anti-atherosclerotic effects in experimental models [32].

Reduced serum glutathione concentrations were also independently associated with higher cardiovascular risk scores in our study. The role of glutathione deficiency in favouring oxidative stress, vascular damage and atherosclerosis is well established [33]. Glutathione deficiency might be secondary to reduced synthesis, increased degradation or a combination of the two processes. In contrast to its synthesis, which occurs within the intracellular compartment through the activity of glutathione synthetase, glutathione degradation into cysteinylglycine by γ-glutamyltranspeptidase occurs in the extracellular space [33]. An increasing body of evidence suggests that γ-glutamyltranspeptidase activity is associated with surrogate markers of vascular disease and predicts adverse cardiovascular outcomes [34, 35]. The observation of independent associations between lower glutathione/cysteinylglycine, but not glutamylcysteine/glutathione, ratios and risk scores suggests that increased glutathione degradation, rather than reduced synthesis, plays a major role.

An important finding in our study was the lack of significant associations between taurine concentrations and cardiovascular risk. A number of experimental and human studies have shown that taurine exerts protective effects on vascular homeostasis, including anti-oxidant effects, reduced insulin resistance, improved endothelial function and blood pressure control and reduced serum cholesterol concentrations [36]. However, a recent prospective, nested case–control study assessing the predictive role of serum taurine concentrations on the risk of coronary heart disease has failed to demonstrate any significant effects on the primary end-point [37]. Notably, mean serum taurine concentrations in cases and controls (126.9 vs. 130.1 μmol/L) were higher than those observed in our study (median 62.0 μmol/L, Table 1). Pending further studies on the predictive role of taurine on cardiovascular outcomes the role of this thiol on vascular homeostasis and disease prevention might need some revisiting.

The lack of significant associations between serum CRP concentrations and cardiovascular risk scores in our study is in contrast with the results of several reports demonstrating a significant relationship between serum/plasma CRP concentrations, prevalence of atherosclerosis and incidence of first cardiovascular events [38]. There are at least two potential reasons to explain this finding. First, a substantial proportion (70.9%) of our study participants had low cardiovascular risk (<10%) whereas a very small percentage (6.6%) was at high risk (>15%, Table 1). Previous studies have investigated the relationship between CRP and Framingham coronary heart disease scores in subjects without clinical cardiovascular disease [39]. Although CRP concentrations were significantly correlated with risk scores the relationship between the two variables was fairly flat in patients with low-moderate risk. It is possible that the absence of significant associations between CRP and cardiovascular risk scores in our study is at least partly caused by the low representation of patients at high risk. Interestingly, in their study Albert et al. failed to observe significant correlations between CRP concentrations and several individual components of the Framingham risk score [39]. Our observations are in line with their findings (Table 2). Second, the age range of our population was 59–66 years. There is evidence that the predictive role of CRP is somewhat diminished with advancing age. Several studies have demonstrated that CRP concentrations exert a negligible effect in enhancing cardiovascular risk prediction, over and above traditional risk factors, in older adults [40, 41]. In one of these studies homocysteine showed the best predictive power for cardiovascular mortality [41].

A limitation of our study is related to its cross-sectional nature, which does not allow the assessment of cause-effect relationship between transsulfuration thiols and cardiovascular risk. Moreover, the measurement of thiol concentrations from blood does not necessarily reflect their intracellular concentrations. Strength of this study was the combined assessment of different components of the transsulfuration pathway, and their ratios, in regression analysis, adjusting for important clinical confounders.

Serum concentrations of the thiols homocysteine, cysteine and glutathione are independently associated with cardiovascular risk scores at population level. Analysis of substrate/product ratios suggests that reduced bioconversion of homocysteine into cysteine and increased glutathione degradation might play a pivotal role in such associations. Further studies are required to identify whether abnormal expression and activity of the enzymes cystathionine β-synthase, cystathionine γ-lyase and γ-glutamyltranspeptidase are associated with surrogate cardiovascular markers, adverse cardiovascular outcomes in longitudinal studies and whether their pharmacological and/or non-pharmacological modulation might lead to reduced risk.

Arduino Mangoni conducted this work during a Visiting Professorship at the University of Sassari.