Genome scans for hypertension and blood pressure regulation

Interest in the genetic basis of hypertension has spanned the century since it became possible to measure blood pressure (BP). For example, in 1923 Weitz reported several families with a two-generational history of hypertension.1 The nature of the inherited basis of hypertension was the cause of a celebrated debate in the 1950s between Lord Robert Platt and Sir George Pickering.1 Platt argued, based on data such as those of Weitz, that hypertension is a distinct disorder (or a set of such disorders) inherited in a Mendelian fashion. Pickering, while recognizing the importance of inherited factors in BP regulation, argued that these factors operate throughout the range of BP and hypertension simply represents one (arbitrarily defined) extreme of this continuous trait, where by definition BP-raising alleles must aggregate. To some extent both have been proven correct. Several Mendelian forms of hypertension have now been characterized at the molecular level.2 However, these are rare and for the majority of patients with essential hypertension, current data would indicate that any inherited basis of their hypertension is polygenic with complex gene–gene and gene–environment interactions. Between 30% and 40% of BP variation in a population is thought to have a genetic basis.3 Given an affected first-degree relative the relative risk of hypertension is increased by 2- to 5-fold. Therefore, a fuller understanding of the genetic basis of essential hypertension could have significant benefits in terms of tailoring individual treatment and developing more effective preventative strategies. With the advances in molecular techniques, there has seen an explosion of genetic research in the field of BP. The vast majority of studies have sought to either associate variants in “candidate” genes with risk of hypertension (usually using a case-control design) or investigated the quantitative effect of such variants on BP. Although both of these are appropriate strategies, and have produced plausible evidence for the involvement of variants in several genes,2 they are, by definition, restricted to known variants in such genes and more importantly cannot identify hitherto unsuspected genes. Appreciation of these limitations led to the interest in conducting “genome” scans. These rely on the use of anonymous highly polymorphic markers spread across the genome to, in the first instance, pinpoint the location of genes increasing the susceptibility to hypertension or regulating BP. Such scans exploit familial relationships. Basically, nonidentical siblings share 50% of the genome and hence, should have the same allele half the time at any individual marker. If both are affected by hypertension, they should, at least theoretically, more often share the same allele of markers near a gene or genes involved in their hypertension than seen by chance (ie, >50%). Of course, in different families different sets of genes are likely to be involved. However, if a large number of affected sibling pairs are collected and genotyped for all the markers, then excess allele sharing should be observed for markers near those genes that influence the risk of hypertension in at least a proportion of the families. This simple paradigm forms the basis of genome scans and can be extended to larger family groupings and to quantitative phenotypes. The ability to detect a locus depends, among many factors, on the number of families studied, their genetic homogeneity (ie, the similarity of the genetic basis of their hypertension), the strength of the effect of the locus (in that particular population), and the linkage disequilibrium between the markers tested and the locus. As genome scans involve the testing of a large number of markers and often examination of linkage to several phenotypes, guidelines have been proposed to reduce the risk of false-positive results.4 Linkages can be classified as nominal (a significant P value), suggestive (statistical evidence that would be expected to occur one time at random in a genome scan) or significant (statistical evidence that would be expected to occur 0.05 times in a genome scan). The past few years have seen the increasing reporting of genome scans for hypertension and BP variation.5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20 The major findings of the scans reported to date are summarized in Table 1. Most of the scans report nominal or suggestive linkages, although a few studies11,15,18,20 report significant linkages. It is notable that two of these four studies have been in relatively isolated populations,18,20 where the genetic homogeneity is likely to be much higher. Studies in such populations also facilitate the narrowing of regions, and to date the smallest region identified using a genome scan approach with strong evidence that it harbors a locus-affecting risk of hypertension is a 0.54-cM region on chromosome 2 by Angius et al20 in an isolated Sardinian population. A number of regions, notably those on chromosome 1q,12,19 2p,5,9,14,17 3p,6,17 6q,5,15,19 7q,9,19 11q,6,17 12q,17,19 15q,5,6,19 16q,6,16 18q,10,18 and 19p,9,15 have been found in more than one study, strengthening the likelihood that such regions harbor BP-modifying loci. However, these regions are broad and it is possible that these overlaps are due to different genes or represent false-positive linkages.4 Genome scans for blood pressure and hypertension ASPs = affected sibpairs; SBP = systolic blood pressure; DBP = diastolic blood pressure. Results lists the approximate location of reported linkages and specifies the phenotype or ethnic groups where relevant. Results that met criteria for genome-wide significance are shown in bold. The main findings in the individual FBPP studies published in this issue of the Journal are also summarized. Genome scans for blood pressure and hypertension ASPs = affected sibpairs; SBP = systolic blood pressure; DBP = diastolic blood pressure. Results lists the approximate location of reported linkages and specifies the phenotype or ethnic groups where relevant. Results that met criteria for genome-wide significance are shown in bold. The main findings in the individual FBPP studies published in this issue of the Journal are also summarized. This issue of the Journal contains several articles reporting the eagerly awaited results of the largest and the most ambitious of the genome scan efforts in hypertension—the Family Blood Pressure Program (FBPP).21 The Program comprises four networks (GenNet, GENOA, HyperGEN, and SAPPHIRe), each initially set up independently. The populations and the approaches vary between networks (see Table 1 and articles). However, the features in common (in particular the standardized genotyping) have allowed the investigators to come together under the banner of the FBPP and define common goals and share data to enhance the power to detect effects.21 In the aggregate, the studies report the genotype analysis of 6245 individuals (about half of the currently available FBPP resource). With each individual genotyped for more than 380 markers, this represents almost 2.5 million genotypes! The FBPP investigators need to be congratulated on this enormous effort. Given this effort, at first sight the results of the FBPP studies may appear disappointing. The HyperGEN study reports a suggestive linkage for hypertension on chromosome 2p in African American sibpairs, GenNet reports suggestive linkage for diastolic BP in white families on chromosome 1q (and possibly loci on 11q and 3q for systolic and diastolic BP, respectively), whereas the SAPPHIRe study reports suggestive linkage to 10p in sibpairs that were concordant or discordant for hypertension or low BP. None of the studies report identification of a locus (or loci) achieving genome-wide significance, and perhaps more disappointingly, pooling of the data in a meta-analysis did not enhance linkage to any loci. The findings confirm (if confirmation was necessary) that there is no single major locus that, in isolation, affects the risk of hypertension across different populations. However, the suggestive linkage to 2p of hypertension in African American families replicates similar linkages in other studies5,9,14,17 and strengthens the case for further investigating this region for hypertension susceptibility loci, whereas the suggestive linkage found to 1q in the GenNet study also has parallels with other studies.12,19 There are many reasons why the findings in the FBPP studies were not more positive, and these are discussed in the articles. It is, of course, possible that the genetic effects of individual loci are so small that a study much larger than even the FBPP is required to identify them. However, an equally, if not more, important reason may be the genetic architecture of the US population. Although the studies appropriately partitioned their participants into the major ethnic groups, internally these groups, particularly the white Americans and African Americans, probably have a considerable degree of genetic heterogeneity, which would disadvantage any search for susceptibility loci. This is probably one of the main reasons why pooling of the results in the meta-analysis did not enhance the signals. Critics may seize on the FBPP results and decry more investment in genetic research in hypertension. In my judgment, this would be an error. The articles published in this issue of the Journal should be seen as the first salvo from the Program. The authors of the individual articles and the coordinators of the FBPP clearly have in mind to partition their datasets on the basis of relevant demographic, metabolic, or genetic features to maximize the likelihood of detecting significant effects. Such analyses will be aided by new developments in statistics to investigate loci–loci and loci–environment interactions in the type of family-based resources that have been collected. Also, the ability to undertake genome-wide association studies by analyzing thousands of single nucleotide polymorphisms arrayed on chips is round the corner and the FBPP resource, enriched as it is with cases with a strong genetic basis for their hypertension, will be a powerful position to exploit this important technology. Almost a century after the familial basis of hypertension was described, defining the precise genetic basis of hypertension still remains elusive. However, the results of the genome scan studies to date, which have all been published in the past 3 years, suggest that the net is closing and the next decade should see significant progress in unraveling the main factors and the context in which they act. Moving from linkage to identifying the susceptibility genes remains a big hurdle and the goal is likely to be best achieved through national and international collaborations. The FBPP Investigators have anticipated this and will, I am sure, continue to make a major contribution to this effort.