Hypertension Research (2013) 36, 1032–1034 & 2013 The Japanese Society of Hypertension All rights reserved 0916-9636/13 www.nature.com/hr

COMMENTARY

Candidate genes revisited in the genetics of hypertension and blood pressure

Norihiro Kato

Hypertension Research (2013) 36, 1032–1034; doi:10.1038/hr.2013.89; published online 15 August 2013

espite considerable, concerted efforts receptor C/guanylate cyclase C) have been discrepancy is termed ‘missing heritability’, as Dover the past 20 years, it is only within noted4 because of their involvement in discussed elsewhere.6 A question has been the past 4 years that substantial progress has known BP regulatory pathways, that is, raised as to whether complex traits are been made in identifying genetic variants the renin–angiotensin–aldosterone system affected by thousands of variants with small robustly associated with blood pressure (RAAS) and natriuretic peptides. Despite effects, but a recent analysis of GWA study (BP) elevation or hypertension (Figure 1), the as-yet-undefined genetic architecture of data using a computational technique has largely due to technological advances.1–3 complex traits, these findings suggest that suggested that many hundreds of common, We have witnessed great success in the molecular variants in some physiological weakly associated variants may be sufficient identification of new susceptibility loci for candidate genes have a role in the etiology to account for the majority of heritability.7 It BP and hypertension through genome-wide of BP elevation or hypertension. In this is now assumed that although many true association (GWA) analysis. The number of article, I will comment on the background positive associations reside within GWA robustly associated loci, that is, those that and necessity of revisiting notable candidate study data, most have not achieved a have attained a genome-wide significance genes in GWA analysis, as we have recently stringent threshold of statistical significance level (Po5 Â 10 À8) and that have been reported in this journal.5 (Po5 Â 10 À8),8 which is generally required repeatedly validated in independent samples, In 2009, two consortia-based meta-ana- for an association to be considered has increased to 440 today. Therefore, lyses of GWA studies involving 30 000–40 000 ‘confirmed.’ Accordingly, it is notable that because GWA studies investigate a large samples in the discovery stage identified the number of discovered variants (or loci) is number of single-nucleotide polymorphism 13 independent loci that are significantly strongly correlated with the experimental (SNP) markers simultaneously, such a associated with BP in populations of Eur- sample size in GWA studies of complex stringent statistical threshold should be set opean descent. Subsequently, meta-analyses traits via incremental statistical power.1 The to avoid false positives from multiple testing. of GWA studies in East Asians and in a larger power studies further suggest that larger Most loci associated with BP and/or consortium combining the Global BPgen and GWA studies, with tens or hundreds of hypertension map to regulatory or intergenic Cohorts for Heart and Aging Research in thousands of participants, will yield many regions of the genome, and in many cases, Genome Epidemiology (CHARGE) consortia more validated loci for complex traits, the causal transcript remains undetermined. identified 21 additional loci. During these whereas these loci remain individually Surprisingly few of the GWA signals have efforts, it has been widely recognized that for difficult to identify in current GWA studies mapped near strong biological candidates. complex traits such as BP and hypertension, with practically attainable sample sizes.7 Nevertheless, at certain loci, the causal tran- the majority of susceptibility loci may indi- Besides statistical power, coverage of genetic script is inferred, based on a combination of vidually exert modest genetic effects, that is, a markers in GWA scans may not always supportive data, for example, coding variants 5–20% increased risk of developing hyper- be sufficient to examine associations with (in particular, non-synonymous SNPs), tension and an association between each particular functional variants (or mutations) nearby biological candidates and cis expres- copy of the risk allele and an average increase in the target genes because during GWA sion quantitative trait loci, which regulate of approximately 1 mm Hg in the systolic BP studies, researchers normally assay only mRNA expression levels. Among the loci or approximately 0.5 mm Hg in the diastolic a subset of SNPs that can be chosen harboring strong candidate genes, ENPEP BP.2 Accordingly, these loci may not attain by considering the linkage disequilibrium (which encodes glutamyl aminopeptidase) a genome-wide significance level unless a structure in a given chromosomal region.3 and NPR3 (which encodes natriuretic peptide large number of individuals are analyzed in Therefore, we conducted a large-scale GWA scans. association study of eight variants from seven Furthermore, it has been argued that the candidate genes, namely, ACE (angiotensin- Dr N Kato is at Department of Gene Diagnostics and susceptibility loci identified to date through converting enzyme), ADD1 (adducin 1), Therapeutics, Research Institute, National Center for Global Health and Medicine, Tokyo, Japan GWA studies explain only a small pro- ADRB2 (beta-2 adrenergic receptor), AGT E-mail: [email protected] portion of heritability (o3%) for BP. This (angiotensinogen), CYP11B2 (aldosterone Commentary 1033 synthase), GNB3 (G protein, beta-3 subunit) 1990 2000 2010 and NOS3 (nitric oxide synthase 3), in the Japanese (in total, 21 851 samples from the Cardiometabolic Genome Epidemiology Candidate gene approach Genome scan [Early GWAS Linkage analysis GWAS] Consortium-based (CAGE) network study panels) and in a joint (+ Genome scan in animal models) meta-analysis involving previously reported Study of Mendelian hypertensive disease multi-study populations.5 The individual contribution of each of these genes to BP ACE AGT ADD1 GNB3 (WTCCC) CHARGE AGEN variation and/or hypertension in the general & serum level Global BPgen ICBP CYP11B1/B2 population had not been thoroughly BP genetic study consortium elucidated, although a number of studies had repeatedly investigated them as REN/ ACE/ candidate genes. Our study demonstrated Dahl SS SHRSP BP QTL analysis congenic strain that BP-trait associations at two loci (AGT Rat genetic study rs699 and CYP11B2 rs1799998) were Sample size of HTN/BP genetic studies previously reported consistently replicated, and particularly, the 100,000 hypertension association reached genome- wide significance (Po5 Â 10 À8) for these two variants.5 50,000 For the BP-trait-associated candidate gene variants, four features are noteworthy. First, 0 the genetic impacts of individual candidate Figure 1 Progress in analytical approaches to the genetics of hypertension, with regard to the study variants (or loci) are confirmed to be modest, sample size. In the 1990s, a number of candidate genes were repeatedly examined by case–control in accordance with the findings from GWA studies, but the sample size in the individual studies was mostly o1000. The sample size remained studies; that is, the allelic odds ratios (ORs) relatively small until the advent of GWA analysis (approximately 2007) and subsequently has increased for hypertension are mostly in the range rapidly over the past several years. A full color version of this figure is available at the Hypertension between 1.10 and 1.24 for the variants Research journal online. presenting nominally (Po0.05) significant associations in the CAGE study panels.5 This finding reflects the necessity of a large I. Selection of target genes sample size to expose variants with smaller Physiological candidate genes Positional candidate genes effects and partly explains why the strength of 1) Genes in the known biological 1) Genes located in the linked region, the BP-trait association in the previous pathways, e.g., the renin-angiotensin- which was detected by linkage reports has not reached a genome-wide aldosterone system (RAAS), the analysis, e.g., affected sib-pair sympathetic nervous system (SNS) analysis significance level. Second, given the 2) Novel genes identified in model 2) Genes located in the associated fluctuations in the strength of genetic organisms region, which was detected by GWAS associations observable among the relatively small studies, reproducibility should be II. Analytical approach assessed based not only on the data A selected set of variants A comprehensive set of variants statistically summarized by meta-analysis Prioritize variants with some functional Investigate a number of variants that but also on the data from multiple significance when testing genetic can tag genetic polymorphism in the independent panels with a large sample size. association at each target gene region harboring target gene(s)

In our study, although by current standards Case-control study of hypertension Quantitative (BP) association study not sufficiently large in sample size, three Diagnostic criteria should be clearly BP values should be averaged for independent panels (n ¼ 1256–12 569) were defined. multiple measurements. used for the analysis, and the effect sizes have The lack of population stratification For people under hypertension therapy, proven to be relatively consistent across the should be validated. imputed BP values can be used. panels at the two associated loci; for example, Figure 2 Summary of the methodological considerations for conducting a candidate gene study in a OR ¼ 1.24–2.04 for AGT rs699 and 1.22–1.59 comprehensive and systematic manner. for CYP11B2 rs1799998 (risk allele homo- zygote vs. non-risk allele homozygote).5 Third, the concordance of association population stratification. Again, for the two that of a series of BP-trait loci identified via between BP (that is, a quantitative trait) associated loci, namely, AGT rs699 and exploratory GWA scans.5 However, it is and hypertension (that is, a dichotomous CYP11B2 rs1799998, concordant associations probable that the associations were not trait) supports the claim of genetic effects with BP traits (systolic BP, diastolic BP and followed up in the multistaged screening on BP variation in the general population. hypertension) have been found in the Japa- approach of the previous GWA studies In the 1990s, partly because of the feasibility nese (P ¼ 0.03–1.5 Â 10 À5)andinajoint principally because their statistical signi- of sample ascertainment, the candidate meta-analysis (P ¼ 0.01–7.3 Â 10 À10). Fourth, ficance in the discovery stage did not gene studies were designed and vigorously the strength of associations at these two attain a given threshold (for example, conducted as case–control comparisons, candidate loci, estimated from the overall Po1 Â 10 À5), as above mentioned in the which are often liable to selection bias or meta-analysis data, is almost comparable to argument of heritability.

Hypertension Research Commentary 1034

Several genetic variants tested in our study CONFLICT OF INTEREST 5 Takeuchi F, Yamamoto K, Katsuya T, Sugiyama T, (at ACE, AGT, CYP11B2, ADD1 and GNB3) The author declares no conflict of interest. Nabika T, Ohnaka K, Yamaguchi S, Takayanagi R, Ogihara T, Kato N. Reevaluation of the association of had been repeatedly investigated as biologically seven candidate genes with blood pressure and plausible candidates (Figure 1) because of hypertension: a replication study and meta-analysis their physiological importance, that is, key with a larger sample size. Hypertens Res 2012; 35: 825–831. molecules in the known BP regulatory path- 1 Visscher PM, Brown MA, McCarthy MI, Yang J. Five years of GWAS discovery. Am J Hum Genet 2012; 90: 6 Manolio TA, Collins FS, Cox NJ, Goldstein DB, ways (for example, RAAS), causative genes in 7–24. Hindorff LA, Hunter DJ, McCarthy MI, Ramos EM, Mendelian hypertensive disease and novel 2 Ehret GB. Genome-wide association studies: contribu- Cardon LR, Chakravarti A, Cho JH, Guttmacher AE, tion of genomics to understanding blood pressure and Kong A, Kruglyak L, Mardis E, Rotimi CN, Slatkin M, findings supporting the possible contribution essential hypertension. Curr Hypertens Rep 2010; 12: Valle D, Whittemore AS, Boehnke M, Clark AG, to BP regulation in model organisms. In 17–25. Eichler EE, Gibson G, Haines JL, Mackay TF, addition, despite the lack of clear physiological 3 Kato N. Ethnic differences in genetic predisposition to McCarroll SA, Visscher PM. Finding the missing hypertension. Hypertens Res 2012; 35: 574–581. heritability of complex diseases. Nature 2009; 461: evidence regarding BP elevation or hyperten- 4 Kato N, Takeuchi F, Tabara Y, Kelly TN, Go MJ, Sim X, 747–753. sion, the GWA findings can further provide a Tay WT, Chen CH, Zhang Y, Yamamoto K, Katsuya T, 7 Stahl EA, Wegmann D, Trynka G, Gutierrez-Achury J, list of positional candidate genes. These two Yokota M, Kim YJ, Ong RT, Nabika T, Gu D, Chang LC, Do R, Voight BF, Kraft P, Chen R, Kallberg HJ, Kokubo Y, Huang W, Ohnaka K, Yamori Y, Nakashima E, Kurreeman FA, Diabetes Genetics Replication and categories of candidate genes—physiological Jaquish CE, Lee JY, Seielstad M, Isono M, Hixson JE, Meta-analysis Consortium; Myocardial Infarction and positional—must be examined in a more Chen YT, Miki T, Zhou X, Sugiyama T, Jeon JP, Liu JJ, Genetics Consortium, Kathiresan S, Wijmenga C, Takayanagi R, Kim SS, Aung T, Sung YJ, Zhang X, Gregersen PK, Alfredsson L, Siminovitch KA, Worthing- comprehensive and systemic manner as a Wong TY, Han BG, Kobayashi S, Ogihara T, Zhu D, ton J, de Bakker PI, Raychaudhuri S, Plenge RM. target gene approach to complement the Iwai N, Wu JY, Teo YY, Tai ES, Cho YS, He J. Bayesian inference analyses of the polygenic architec- GWA scans (Figure 2). This analysis will help Meta-analysis of genome-wide association studies ture of rheumatoid arthritis. Nat Genet 2012; 44: identifies common variants associated with blood 483–489. to elucidate the complex genetic architecture pressure variation in east Asians. Nat Genet 2011; 8 Gibson G. Hints of hidden heritability in GWAS. of hypertension. 43: 531–538. Nat Genet 2010; 42: 558–560.

Hypertension Research