Context dependency of the salt intake: left ventricular hypertrophy connection

Left ventricular hypertrophy (LVH) frequently complicates essential hypertension, with a variable prevalence ranging from approximately 20% when blood pressure (BP) is only modestly elevated, to much larger figures when BP is higher or not adequately controlled by treatment. LVH is a major negative prognostic indicator not only for cardiac adverse events (e.g. myocardial infarction, cardiac sudden death, congestive heart failure), but also for stroke [1]. Although not completely understood, the description (interpretation) of the pathogenesis of LVH has abandoned the simplistic view that it is solely caused by the pressure load. Indeed, if it is clear that the latter plays an important role, this leaves at least 50% of its variance unexplained [2]. That several nonhaemodynamic factors are important is evident from observations that, in spite of similar target BP reduction, not all drug classes decrease LVH to the same extent because those that temper the renin–angiotensin–aldosterone system (RAAS) appear to be the most effective [e.g. angiotensin receptor blockers (ARB) are more effective than angiotensin-converting enzyme inhibitors (ACEI) and calcium channel blockers] [3]. An interaction between left ventricular mass (LVM), Na+ intake and an individual ability to react to excessive Na+ intake by controlling renin release (probably dependent on the genetic context of each individual) was already elegantly described by Schmieder et al. [4] more than 10 years ago. These authors demonstrated that, at a similar BP level, hypertensive individuals who were able to reduce their plasma renin activity more in response to their Na+ intake were protected from the development of ventricular hypertrophy. The RAAS has then emerged as a major modulator of LVM. Angiotensin II exerts its action through two types of high-affinity receptors that differ in their pharmacological properties and tissue distribution. Angiotensin II type I (AT1) receptor (gene: AGTR1), which mediates most of the classic cardiovascular actions of angiotensin II, is a seven-transmembrane G-protein-coupled receptor. Angiotensin II type II (AT2) receptor (gene: AGTR2) is loosely related (34% identity) to the AT1 receptor. The AT1 receptor, which has been studied more in depth, is known to mediate most of the physiological effects of angiotensin II in the heart, such as positive chronotropic and inotropic effects, presynaptic facilitation of noradrenaline release from cardiac sympathetic nerve terminals, coronary vessel vasoconstriction, stimulation of aldosterone release, myocyte hypertrophy, nonmyocyte proliferation and interstitial fibrosis [5–7]. AT1 and AT2 receptors promote heart development during the embryonic stage [8]. The expression of AT2 receptors regresses substantially in the adult heart whereas that of the type 1 predominates. This is not an irreversible phenomenon because, in several pathologic conditions, such as cardiac hypertrophy, cardiac remodelling, ischaemic heart disease or dilated cardiomyopathy [9], the expression of AT2 receptors bounces back. There is general agreement that the effects of angiotensin II mediated by AT2 receptor activation are substantially opposed to those effects mediated by AT1 receptors. For example, AT2 receptor stimulation mediates vasodilation and hypotension [10], whereas mice lacking the gene encoding the AT2 receptor (agtr2 KO) have a larger pressor response to infused angiotensin II [11]. In agreement, AT2 receptor stimulation has, in general, an antiproliferative and antifibrotic effect, and therefore opposes the growth-promoting action linked to AT1 receptor activation [12,13]. More recent studies have shown that this is not always true because over-expression of the AT2 receptor in isolated cardiomyocytes makes the receptor constitutively active and promotes hypertrophy [14]. AT2 receptor stimulation may also cooperate in other experimental set-ups with fibrosis [15] and hypertrophy [16,17], as well as pro-atherogenic and pro-inflammatory effects [18]. Whatever the case, it is indeed clear that angiotensin II can significantly affect the development and regression of LVH in hypertension through stimulation of AT1 and AT2 receptors, independently of BP and other clinical factors. Although such a statement is more clearly demonstrable for the type 1 receptor, as several antagonists are available for clinical use in humans, the demonstration is more circumstantial for the type 2 receptor, where no antagonist is readily available for clinical use. To shed some light on the function of the AT2 receptor, the genetic polymorphism within AGTR2 has been used as a proxy for the different functions of the protein encoded by the gene. In addition, whether the genetic polymorphism under study is related to the phenotype has also been investigated. A common polymorphism occurring in intron 1 has been repeatedly tested for any association. This polymorphism has already previously been described as (–1332)G/A or as (+1675)G/A relative to the translation initiation site of the human AT2 receptor gene. For reasons of homogeneity, we retain the latter nomenclature. It is located 29 bp before exon 2, close to the region that is important for transcriptional activity, involving a critical motif [19]. Human mRNA studies have shown that subjects with the G allele may express higher levels of AT2 receptor protein [20]. To put it in very simple terms, if more AT2 means counteracting the ‘classic’ angiotensin II noxious effect, then the G allele should be ‘the good one’, whereas the A allele should be ‘the bad one’. Several studies have reported on the relationship between the (+1675)G/A polymorphism of the AT2 receptor gene and left ventricular structure or cardiovascular outcome, employing different experimental designs and obtaining different results. The first study was by Schmieder et al.[21], who investigated LVM by echocardiography in 120 healthy young male students of Caucasian origin enrolled at the campus of the University Erlangen–Nürnberg with normal or mildly elevated BP. Those with high BP had never been treated previously. Overall, the carriers of the A allele had a larger septal, posterior and relative wall thickness, as well as a larger LVM [21]. When the sample was split according to hypertensive status, the observation held true only for those with high BP because no effect of the A allele on LVM could be observed in normotensives. That the A allele is associated with a larger LVM in male hypertensives was later confirmed in a sample of 336 hypertensive males, aged 55–74 years, from the Glasgow Heart Scan Old Study (GLAOLD) [22]. The finding was not replicated in females from the same GLAOLD cohort, where the A allele was found to be associated with a more frequent finding of electrocardiographic signs of coronary ischemia or myocardial infarction (Q waves, left bundle branch block, ST-T wave changes) in females only, but not in males. The potentially larger cardiovascular risk imparted by the A allele in males was, however, confirmed in a prospective study demonstrating that hypertensives carrying the A allele had a hazard ratio of death of 1.8 after 10 years of follow-up [23]. Kuznetsova et al.[24], studied a large sample from the European Project on Genes in Hypertension (EPOGH) study. Due to the study design (family-based case–control study), most of the subjects investigated were normotensives and no association was found between the (+1675)G/A polymorphism and indices of LVM [24], in agreement with the previous finding by Schmieder et al.[21]. However, when Na+ intake was taken into account, it was clear that the G allele was somehow protective because individuals carrying the G allele with an Na+ intake below the median of the population had a lower LVM than individuals carrying the G allele with an Na+ intake above the population median or individuals carrying the A allele, irrespective of their Na+ intake. An opposite result was obtained by Alfakih et al.[25] who reported that the G allele was associated with a larger LVM. However 85% of patients were treated (33% of them taking three or more drugs; and 44% taking an ACEI or ARB), and it is difficult to draw any definitive conclusions [25]. In the present issue of the journal, Ott et al.[26] present another elegant demonstration of context dependency of the genotype by phenotype interaction. The authors extend the study on the association between the (+1675)G/A polymorphism and indices of LVM in healthy young male subjects with normal or mildly elevated BP. None of those individuals with high BP had ever been treated previously [26]. Overall, the results confirm those previously obtained with a similar protocol [21] because the A allele is associated with a larger septal, posterior and relative wall thickness, resulting in higher LVM. With regard to systolic function, the A allele is associated with a faster fractional fiber shortening and a larger ejection fraction. When the sample was stratified according to BP, the differences explained by the (+1675)G/A polymorphism are amplified in the hypertensive group and disappear in the normotensive group, indicating that a pressure afterload must be present together with the A allele to express the phenotype. Finally, when Na+ intake was considered, similar to that reported by Kuznetsova et al.[24] for the EPOGH cohort, carriers of the G allele with an Na+ intake below the population mean had the lowest LVM. However, despite significant improvements in our understanding of the mechanisms relating LVH to salt intake, the pattern is still far from being complete. At the environmental level, we still need a formal test of the effect of Na+ on either BP and LVH, or other BP-related target organ damages, while considering the genetic profile of the subjects under evaluation. This step also implies the need to formally test the gene–environment interaction, which, albeit suggestive, cannot be limited to using the Na+ intake as a stratification factor. We are now learning more and more about complex traits and their biological background thanks to the flood of papers being published on genetic association scans of the entire human genome at a rate that we could not have imagined only a few years ago. The most compelling evidence emerging from these studies is that genes are acting within composite (and, in many cases, previously unpredicted) networks or pathways and the specific involvement of any given gene becomes meaningful only within a more complete pattern of the pathway leading from genotype to phenotype. Furthermore, the study by Ott et al.[26] conveys another contribution to the issue: in a complex trait, there may not necessarily be ‘good’ or ‘bad’ alleles per se. In most cases, when we look at the effect of a single gene coming from a larger context, we can only observe relatively subtle differences that can be better understood when we are able to look at the joint effect of the other genes implicated in the trait. A ‘good’ allele in context A can thus become a ‘bad’ allele in context B, depending upon its correlations with genes characterizing context A or B. Similarly, different variants at coding and regulatory regions of a given gene can contribute to strengthening or molding the observed phenotype, from subject to subject, or across populations. These notes of caution are not intended to give the impression that we are facing a pattern too complex to unravel: on the contrary, we would strongly support the efforts of researchers to continue and expand their investigations and to test bold hypotheses, because our growing knowledge and the advancements of technology really allow all of us to explore the complexities of human biology without having to pay too high a fee in terms of excessive reductionism. Perspectives The data available to date are too scanty to allow any clinically valuable conclusion to be made. However, if the average Na+ consumption was indeed very high in all the samples where it was assessed [24,26], the two observational studies available demonstrate that subjects carrying the G allele and with a lower Na+ intake were less prone to increase their LVM, especially when challenged with a pressor load. Furthermore, if reasonable, this observation can by no means be interpreted as being a straightforward indication of the beneficial effect of a low Na+ diet for the prevention of LVM in hypertensives carrying the G allele until an ad hoc expressly designed trial has been performed. Such a trial would have to consider the enigmatic nature of the AT2 receptor recalled above. This problem will never be solved until we can unambiguously establish the type and extent of intracellular signals emanating from the AT2 receptor, particularly those that underlie its capacity to promote constitutive cardiomyocyte growth in relation to Na+ intake and angiotensin II stimulation and their different relationships with the multigenic context and the environment. Acknowledgements This work was supported by Ministero Università e Ricerca Scientifica of Italy, FIRB GRANT 2003 RBLA038RMA_010 (GENOPOLIS) and by INGENIOUS HYPERCARE (Network of Excellence of the EU) GRANT LSHM-CT-2006-037093 to D.C.