Angiogenesis, the metabolic syndrome and heart disease: is there a connection?

The history of the development of biomedical science is littered with examples of seemingly artificially created diseases and obscure laboratory findings that, perhaps decades later, are found to have common ground. The metabolic syndrome (also described as syndrome X or metabolic syndrome X) is perhaps an excellent example of the former, being as it is collection of different clinical conditions, some based on laboratory measures (Table 1). Gordon [1] may be one of the earliest proponents of a metabolic condition underlying obesity, a viewpoint subsequently developed by others to include hypertension, insulin resistance with hyperinsulinaemia and hyperlipidaemia [2–7]. A likely mark of the maturity of particular disease or condition is it being recognized by national and international bodies [8, 9], a process which is developing not only in terms of consensus statements but also in the widening of possible pathogenic mechanisms, such as inflammation (which may be marked by increased levels of CRP) and coagulopathy [10, 11]. In parallel to the ‘clinical’ development of the disease, molecular pathways of potential importance, in which some, such as cytokines and components of the haemostasis pathway, may be active mediators and may link the condition, via cell pathology, to critical life events such as stroke [12–14]. Although high levels of troponins are often taken to imply acute cardiomyocyte damage (as may occur following an acute coronary syndrome such as myocardial infarction), the large and diverse number of other causes, such as cardiomyopathy and tachycardia [15], fuel the need for additional, more specific markers, of the different types of heart disease. N-Terminal pro-brain natriuretic peptide (NT-proBNP), a molecule closely related to atrial natriuretic factor, is just such a marker of a stressed heart [16, 17]. It is attracting increasing attention as a potential tool in ventricular hypertrophy, heart failure, and in stratifying clinical risk following myocardial infarction [18–21]. Although angiogenesis is clearly fundamental to neoplastic metastasis [22], there is growing evidence that it may also have a role in cardiovascular disease, such as the excessive development of capillaries in the vaso vasorum and within atherosclerotic plaque [23, 24]. As angiogenesis is driven by growth factors such as vascular endothelial growth factor (VEGF) and placental growth factor (PlGF), it is therefore perhaps unsurprising to find increased levels of these growth factors in the plasma of subjects with coronary artery disease and peripheral artery disease, and in many of their risk factors [25–27]. However, the position of angiogenesis as a clinical entity requires clarification as it has been suggested that is it a good thing where blood supply is poor, as in the hypoxic atherosclerotic leg (and is therefore to be promoted), but may be a bad thing, as in diabetic retinopathy and in promoting an unstable intracoronary and possibly vulnerable atherosclerotic plaque, liable to rupture (and is therefore to be resisted) [25, 28–30]. In the present issue of the Journal, Siervo and colleagues [31] offer interesting data that merge these three concepts, focusing on blood markers of angiogenesis (VEGF and its receptor sFlt-1, and PlGF), cardiopathology (troponin T and proBNP) and inflammation (CRP). Furthermore, their study population is of interest because, unlike most studies of this nature (and in almost all of the references cited below), the subjects are gathered not from an urban setting but from a remote region where, it may be hypothesized, they are relatively protected from the atherogenicity of modern city life. Excluding those subjects with existing cardiovascular, neoplastic and inflammatory disease, and those taking medications aimed at cardiovascular risk factors, the authors settled on 1000 people with a mean age of 48 years. This demographic is of interest as, if present, any cardiovascular disease may be presumed to be minimal and certainly asymptomatic. Taking a standard epidemio-logical approach, a gross correlation analysis of this population revealed that, taking an arbitrary correlation coefficient of >0.35 to be scientifically (as opposed to statistically) significant, PlGF, troponin T and proBNP all increased with age, whilst troponin T correlated with high systolic blood pressure, PlGF and levels of sFlt-1. These finding were broadly confirmed in a multiple regression analysis that found other relationships of varying significance (such as that smoking has a far greater effect on sFlt-1 than it does on VEGF, its ligand), which is a valuable confirmation nonetheless, especially as the population is not small [32]. Notably, all six laboratory markers increased with age, and to some extent it could be argued that these relationships are to be expected, as disease burden ‘naturally’ accrues in the elderly. Similarly, many will be unsurprised by the finding that levels of proBNP (a likely but weak surrogate of heart disease) were independently influenced by age, sex, glucose, triglycerides and total cholesterol. Moving to the metabolic syndrome, a more novel aspect of the paper, the authors used any three of the five factors listed in Table 1 to define the condition, and under these criteria collected 160 people who were, in addition, older, heavier and shorter [and therefore automatically with a higher body mass index (BMI)] with a higher serum total cholesterol, urea and uric acid than the 840 free of the markers of the metabolic syndrome. Once more, much of this is to be expected. For example, a potential role for uric acid in the metabolic syndrome is at least a decade old [7]. The authors hypothesized differences in their serum/plasma markers of angiogenesis and cardiovascular disease, and were duly rewarded with increased PlGF, CRP and troponin T, although after adjustment for age, smoking and BMI, only PlGF and proBNP were significantly different. However, when analysing for sex, and so with reduced power, only proBNP was raised in both men and women with metabolic syndrome, whereas PlGF was raised only in the women with the metabolic syndrome. An additional analysis is for the number of individual metabolic syndrome criteria, and it comes as no great surprise that VEGF, PlGF, CRP, troponin and proBNP all correlate strongly with the number of criteria. So what are we to make of all this interesting data? Although the sample size of 1000 participants is large enough for a useful associative analysis, the number with metabolic syndrome (160) is modest, and, to this commentator, there may be a degree of over-analysis. Nevertheless, the study does offer some confirmations (such as no increase in VEGF or sFlt-1) [33] and some other perspectives. In the population as a whole, an increase in proBNP of 30% in those with the metabolic syndrome is notable but is perhaps understandable, given the relationship between the latter and heart disease. Although the finding that PlGF is some 17% higher in the metabolic syndrome is interesting, the first question is whether or not this finding is genuine or spurious, given the large number of analyses performed. However, further discussion is pointless as this effect is strongly influenced by sex – PlGF is a highly significant 20% higher in women with the metabolic syndrome but a nonsignificant 13% higher in men, respective to their sex-specific nonmetabolic syndrome controls. Can it be that a significant number of the 89 metabolic syndrome women are in fact pregnant? Not so likely, as the mean age of this group is 59 years, and so the vast majority will be postmenopausal. So why the sex effect? Notably, a study from the USA showed that raised PlGF predicts those women at risk of coronary heart disease [34], so perhaps raised levels in the metabolic syndrome are not so surprising after all. But if this is a true effect, what pathophysiology may be in operation? There are clues. In animal models, PlGF promotes atherosclerotic intimal thickening and macrophage accumulation, and suppression of PlGF delays plaque progression [35, 36]. In our own species, PlGF expression in atherosclerotic carotid plaque seems related to plaque destabilization [37]. Does this mean, as discussed [38], that PlGF is marker of an unstable plaque? So, using the broadest and most crude approximations, if high levels of PlGF are bad (for women), how should we reduce them? Will this do any good? And is this true only in women? And if so, should we advocate adding PlGF to the ever-growing list of measurable molecules? Although the desire to generate a comprehensive theory of the pathophysiology of the metabolic syndrome is understandable, there are many alternative and plausible models directed towards other markers, such as antioxidants and their regulators [39] (Fig. 1). There is clearly much we don’t know, but for many of us, the jury is not simply out considering its verdict, the jury has yet even to be selected. Nevertheless, the data of Siervo and colleagues both supports other data and breaks new ground. But what of the inhabitants of the rural regions concerned? All we can offer them is standard healthy living advice (don’t smoke, don’t get fat), and in perhaps 10 – 20 years time we will find out whether or not this unique experiment has something concrete to offer. Potential pathways in the metabolic syndrome. Items in bold are contributors of the metabolic syndrome. Items in brackets are possible pathophysiological mechanisms. No conflict of interest to declared.