The combined impact of 12 common variants on hypertension in Japanese men, considering GWAS results

Transcription

1 (2012) 26, & 2012 Macmillan Publishers Limited All rights reserved /12 ORIGINAL ARTICLE The combined impact of 12 common variants on hypertension in Japanese men, considering GWAS results K Miyaki 1,2, NC Htun 2, Y Song 1, S Ikeda 2, M Muramatsu 2 and T Shimbo 1 1 Division of Clinical Epidemiology, Department of Clinical Research and Informatics, National Center for Global Health and Medicine, Tokyo, Japan and 2 Department of Molecular Epidemiology, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan Genome-wide association studies have identified several polymorphisms that appear to be on hypertension-susceptible regions. We performed the current replication study in order to evaluate the association of these loci with hypertension in healthy Japanese males and then examined the combined effect of 12 independent variants. Overall, 735 Japanese men from two independent cohorts were recruited. Association with hypertension was assessed in 16 polymorphisms on 12 genes and 12 were chosen to evaluate the combined impact. Polymorphisms on the COMT, ATP2B1, CYP11A1 and the CSK genes were confirmed to be associated with hypertension and blood pressure (BP). Current findings also replicated previous results for the CYP11B2 and PTGIS genes. Although there were no significant associations found for other variants, our results suggested there was a combined impact for 12 loci. Individuals carrying more risk alleles had a higher risk of hypertension (P for the slope ¼ 0.002). Blood pressures also increased in conjunction with an increasing risk allele score (P for trend ¼ and for SBP and DBP, respectively). Our results confirmed the associations between hypertension or blood pressure and four gene variants. We also found a significant combined effect of the 12 gene loci. (2012) 26, ; doi: /jhh ; published online 2 June 2011 Keywords: combined impact; polymorphism; blood pressure; genome-wide association study Correspondence: Dr K Miyaki, Division of Clinical Epidemiology, Department of Clinical Research and Informatics, National Center for Global Health and Medicine, Toyama , Shinjuku-ku, Tokyo , Japan. miyaki-keio@umin.net Received 7 December 2010; revised 18 March 2011; accepted 12 April 2011; published online 2 June 2011 Introduction Hypertension is an important public health problem found throughout the world. Hypertension prevalence has been increasing, and as of 2006, an estimated 972 million people in the world currently suffer from this problem. 1 In Japan, more than 33 million adults, or 45% of the adult population, are estimated to have hypertension. 2 Although the disease is thought to result from a complex interplay between genetic predisposition and environmental influences, it is unclear as to what extent the genetic factors contribute to an individual s susceptibility to hypertension. There are numerous genetic-based studies that have explored the basis of hypertension, with the results providing a better understanding of the underlying molecular mechanisms and making it possible to identify the common variants that are associated with hypertension or blood pressure (BP). The most used techniques include genomewide association studies (GWAS) and the candidate gene approach. Recently, results from several GWAS identified several single-nucleotide polymorphisms (SNPs) that appeared to be related to hypertension in different population groups, such as SNP rs , which was near the ATP2B1 gene (encodes PMCA1, a plasma membrane calcium/calmodulin-dependent ATPase) reported by Korean researchers; 3 SNP rs , which is near the CYP17A1 (cytochrome P450, family 17, subfamily A, polypeptide 1) gene and SNP rs , which is near the CSK (c-src tyrosine kinase) gene identified by a European GWAS; 4 the common variant rs , which was upstream of the CDH13 (cadherin 13 preprotein) gene found in two European populations. 5 The candidate gene approach has also been used to investigate associations of diseases with common variants. Kamide et al. performed a large association study and suggested that GREB1 (gene regulated by estrogen in breast cancer 1) and HPCAL1 (hippocalcin-like 1 protein), which previously revealed by genome-wide scans in Italian 6 and Chinese 7 were candidate hypertension-susceptibility genes in Japanese. 8 The polymorphisms of ABCA1 gene (ATP binding cassette A1), 9 ACADSB (short/ branched chain acyl-coa dehydrogenase), 10 COMT

2 (catecholamine-o-methyltransferase) 10 and PTK2B (protein kinase 2b) 11 were also found to be determinants of BP and responsible for the development of hypertension in Japanese individuals. In the current study, we chose 14 SNPs in 10 genes that had been previously reported by GWAS or were from candidate gene studies. After genotyping them in two unrelated Japanese cohorts, we were able to replicate associations in the general Japanese population. In addition, our previous results have discovered several SNPs that were associated with hypertension in Japanese men, including C-344 T SNP of the CYP11B2 gene 12 and C1117A SNP of the PTGIS gene (unpublished). In the current study, we also included these polymorphisms and performed a multiple regression analysis in order to evaluate the combined risk for BP from these 12 loci in our samples. Subjects and methods Subjects A total of 735 Japanese men from two separate occupational cohorts participated in the current study (Table 1). The first cohort, which has been described elsewhere, 12 comprised of 311 population-based Japanese men who worked for a company in Kanagawa Prefecture. The second cohort comprised 424 population-based Japanese men in Kyoto Prefecture who worked for a subsidiary of the first company. Information on age, current smoking status, drinking and the energy intake of all subjects was obtained by means of a self-reported questionnaire. A food frequency questionnaire was used for this purpose. Medical history was acquired by interview. Height and weight were measured, with the body mass index calculated as weight in kilograms divided by height in meters squared. After subjects rested quietly for at least 10 min in a supine position, BP was measured using a posterior wall velocity/ankle-brachial index device (Nippon Colin, Aichi, Japan). Pressure was measured twice, with the average mean value then recorded. Hypertension was defined as a systolic BP (SBP) X140 mm Hg or a diastolic BP (DBP) X90 mm Hg, or the current use of an antihypertensive drug. All other subjects were considered to be normotensive. The study was approved by the ethics review committee of the Medical Research Institute of Tokyo Medical and Dental University. All participants provided written informed consent. Genotyping Genomic DNA was extracted from the peripheral blood of each subject by conventional methodology. Genotyping for all polymorphisms was performed by PCR using a TaqMan genotyping assay (ABI) followed by allelic discrimination analysis using SDS software (Applied Biosystems, Carlsbad, CA, USA). For each sample, 10 ng of purified DNA was mixed with 0.25 ml of 20 SNP genotyping assay mix that contained forward and reverse primers, VIC and FAM probes, 2.5 ul of TaqMan universal PCR master mix and 2.25 ml of Dnase-free water. PCR amplification was performed by using the Applied Biosystems 9700 Thermal Cycler, with holding temperatures of 90 1C for 10 min, 40 cycles of the denaturation temperature at 92 1C for 15 s and an annealing temperature of 60 1C for 1 min, with a final cooling step at 4 1C. Following PCR, allelic discrimination was performed by using the Sequence Detection System (ABI PRISM 7900HT, USA, SDS software package version 2.2.1). We obtained successful genotyping call rates of % for the whole characterized sample. Statistical analysis The allele frequencies were determined by direct counting. Deviation of the genotype distribution from the Hardy Weinberg disequilibrium was confirmed by a w 2 -test. Difference in the mean values of age, clinical characteristics, current smoking status, drinking and energy intake levels between the genotype groups were compared by an analysis of variance or w 2 -test. Logistic regression analyses were performed in order to examine the relationship 431 Table 1 Baseline characteristics of the subjects Characteristics Total subjects (n ¼ 735) Normotensive (n ¼ 474) Hypertensive (n ¼ 261) Age (years) 47.0± ± ±7.4 BMI (kg m 2 ) 23.4± ± ±3.8 Systolic blood pressure (mm Hg) 129.9± ± ±13.2 Diastolic blood pressure (mm Hg) 79.8± ± ±8.6 Total serum cholesterol (mg dl 1 ) 203.0± ± ±35.4 Serum triglycerides (mg dl 1 ) 151.7± ± ±125.7 Serum HDL-cholesterol (mg dl 1 ) 56.7± ± ±15.4 Salt intake (g per day) 9.5± ± ±3.5 Energy intake (kj per day) ± ± ± Current smoking (%) Alcohol use (%) Abbreviations: BMI, body mass index; HDL, high-density lipoprotein. Data are presented as mean±s.d. or percentage.

3 432 between the genotype and hypertension, as adjusted for potential confounders, which might correlate to BP, such as age, current smoking status and drinking. Multiple logistic regression analysis was used to analyze whether interactions between all of these genotypes and the energy and salt intake levels were associated with hypertension. A combined analysis of the variants within these 12 genetic loci was performed by summing up the risk alleles (0, 1, 2) of these 12 variants, with each individual potentially having from 0 to 36 variants (actual values in the current study ranged from 4 to 20). Linear regression was used to analyze the association between the individual risk allele score and the BP. An association with the prevalence of hypertension was expressed as the percentage of hypertension in each risk allele score. For the effect size of the risk allele score on the BP, receiver operating characteristic (ROC) analysis calculated the areas under the ROC curve and used the results to compare the predictive ability of the risk score for all 12 loci and for six of the most significant loci. All analyses were carried out using Statistical Package for Social Science (SPSS) for Windows version 11.0 (SPSS Inc., Chicago, IL, USA). A P-value o0.05 was considered statistically significant. Results Table 1 presents main characteristics of the total subjects, and subjects were classified into normotensive or hypertensive groups. Mean (±s.d.) age and body mass index were 47.0±8.9 years and 23.4±3.4 kg m 2, respectively. The total prevalence of hypertension was 35.5% in our subjects. This was lower than that reported in a cardiovascular disease survey preformed by The Ministry of Health, Labor and Welfare of Japan in 2001, which found the prevalence of hypertension in Japanese men aged from 30 to 69 years to be 47.3% ( wwwdbtk.mhlw.go.jp/toukei/kouhyo/data-kou18/ data12/junkan-h12-3.pdf). When compared with the normotensive group, the mean age, body mass index, BP and serum lipid profile (except highdensity lipid-cholesterol) were significantly higher in the hypertensive group (Po0.05). There were no differences in lifestyles between the two groups for items such as daily salt intake, energy intake and percentage of current smokers or alcohol drinkers. Table 2 shows the results of the genotyping in the 735 subjects for the 16 SNPs of the 12 genes. Success rates for genotyping of these SNPs were more than 99.8%. The minor frequencies ranged from 0.13 to Most of the allele frequencies of these variants were similar to the Japanese data found in the public database from the international HapMap Project ( There was no deviation from the Hardy Weinberg equilibrium for any of the alleles (data not shown). Among all of the genetic loci, new SNPs on the four genotyped genes (COMT, ATP2B1, CYP17A1, CSK) were successfully replicated. Of these, we strongly confirmed that COMT (rs4680 and rs4633), ATP2B1 (rs ) and CYP17A1 (rs ) Table 2 Association between eight genetic loci and the prevalence of hypertension or blood pressure Nearby gene SNPs Allele frequency (risk/non-risk) Hypertension SBP DBP ORs (95% CI) P-value b-value P-value b-value P-value CYP17A1 rs C/T (0.82/0.18) 2.50 ( ) 0.032* * * COMT rs4680 A/G (0.31/0.69) 2.40 ( ) 0.001** 4.16 o0.001* ** rs4633 T/C (0.32/0.68) 2.24 ( ) 0.003** 9.15 o0.001* 5.96 o0.001** ATP2B1 rs G/A (0.65/0.35) 2.12 ( ) 0.008* * CSK rs A/C (0.82/0.18) 1.99 ( ) * GREB1 rs C/G (0.81/0.19) 1.36 ( ) CDH13 rs T/A (0.18/0.82) 1.30 ( ) PTK2B rs G/A (0.39/0.61) 1.18 ( ) rs C/A (0.44/0.56) 1.02 ( ) HPCAL1 rs G/A (0.76/0.24) 1.01 ( ) rs G/A (0.78/0.22) 1.08 ( ) rs T/G (0.87/0.13) 1.14 ( ) ABCA1 rs T/C (0.26/0.74) 1.13 ( ) ACADSB rs G/A (0.29/0.71) 1.05 ( ) CYP11B2 rs C/T (0.34/0.66) 1.79 ( ) 0.045* PTGIS rs5629 A/C (0.29/0.71) 1.02 ( ) * * Abbreviations: ABCA1, ATP binding cassette A1; ACADSB, short/branched chain acyl-coa dehydrogenase; ATP2B1, plasma membrane calciumtransporting ATPase 1; CI, confidence interval; CDH13, cadherin 13 preprotein; COMT, catecholamine-o-methyltransferase; CSK, c-src tyrosine kinase; CYP17A1, cytochrome P450, family 17, subfamily A, polypeptide 1; CYP11B2, cytochrome P450, family 11, subfamily B, polypeptide 2; DBP, diastolic blood pressure; GREB1, gene regulated by estrogen in breast cancer 1; HPCAL1, hippocalcin-like 1 protein; OR, odds ratio; PTGIS, prostaglandin I2 (prostacyclin) synthase; PTK2B, protein kinase 2b; SBP, systolic blood pressure; SNP, single-nucleotide polymorphism. *Po0.05. **Po0.004 (Because 12 genes were tested, the P-value o0.004 was considered to be statistically significant). For the associations with hypertension, the odds ratios were derived from a logistic regression model adjusted for age, body mass index, salt intake and energy intake levels. The non-risk alleles served as the reference groups. For blood pressure associations, the b-coefficients and P-values were obtained from multiple linear regression analyses adjusted for age, body mass index, salt intake and energy intake levels.

4 were significantly associated with the prevalence of hypertension, with odds ratios of 2.40, 95% confidene interval (CI) ¼ (P ¼ 0.001), 2.24, 95% CI ¼ (P ¼ 0.003), 2.12, 95% CI ¼ (P ¼ 0.008) and 2.50, 95% CI ¼ (P ¼ 0.032), respectively, (Table 2). In addition, these polymorphisms were also associated with BP traits. When compared with the AG þ GG group, there were higher adjusted systolic ( þ 4.16 mm Hg, Po0.001) and diastolic ( þ 2.38 mm Hg, P ¼ 0.001) pressures found for the AA genotype of rs4680. We also found that there was a significant interaction between the effect of the high-energy intake level and the prevalence of hypertension for the COMT val158met (rs4680) polymorphism. The prevalence of hypertension had a positive correlation with the energy intake level in the AA group (P for interaction ¼ 0.031). When taken together with the prevalence of hypertension, the COMT gene was coincidentally associated with hypertension in our study. Because of the strong linkage disequilibrium between rs4633 and rs4680 (r 2 ¼ 0.91), only rs4680 was examined in the following analysis. SNPs on ATP2B1, CYP17A1 and CSK were also associated with either SBP or DBP, or with both of them. For the GREB1 gene, there was a significant association with hypertension (odds ratio of 2.26 (95% CI ¼ , P ¼ 0.006)) for only the Kanagawa cohort. No significance was noted for either the Kyoto cohort or the combined analysis. The other eight SNPs in the PTK2B, HPCAL1, ACADSB, ABCA1 and CDH13 genes did not show any significant association with either the BP or the prevalence of hypertension (Table 2). For the two polymorphisms we previously reported, the current study partially replicated our previous results. The SNP of CYP11B2 appeared to be associated with a risk for hypertension, whereas the SNP of PTGIS appeared to be related to BP (Table 2). Next, we calculated the cumulative effect of multiple risk-associated SNPs on hypertension or BP by using the data set of the 12 SNPs. For COMT, PTK2B and the HPCAL1 gene, only one SNP of each gene (rs4680, rs and rs ; with the r 2 for the other SNPs on same gene equal to 0.91, 0.53 and 0.97, respectively) was used because all were in linkage disequilibrium with the other polymorphisms of the gene on which they were located. As a result, there was a significant linear increase in the percentage of hypertension that correlated with the increasing number of risk alleles (Figure 1, P for slope ¼ 0.002). There was a 3.5-fold variation in the percentage of hypertensive subjects between the lowest (p8) and the greatest risk score groups (X18). Moreover, as the number of risk-associated alleles increased, there was a significant increase in the SBP and DBP (1.36 and 0.86 ) in a stepwise manner (P for trend ¼ and , respectively) (Figure 2). In other words, if an individual carried one additional risk allele, the SBP and DBP increased to 1.36 mm Hg and Figure 1 Estimation of the increase of hypertension prevalence in conjunction with the increasing number of hypertension risk alleles from the combination of 12 SNPs. Percentage of the hypertensive subjects in each risk score group are plotted, with the vertical bars representing the 95% CIs. The percentage of hypertension in the greatest risk score group (X18) is 3.5-fold of that observed in the fewest risk score group (p8). Figure 2 Estimation of changes in (a) SBP and (b) DBP scores in conjunction with increasing numbers of hypertension risk alleles when using a combination of 12 SNPs. Black squares represent the mean, whereas the bars extending from the squares indicated the 95% CI. The shaded columns represent the number of individuals in each risk score group. 433

5 mm Hg, respectively. This significant combined impact of the 12 genetic polymorphisms on the effect of BP suggested that an accumulation of these polygenic polymorphisms increased the BP and the risk of hypertension in the Japanese population. To evaluate the discriminatory power of a genetic test based on these variants, we calculated the area under the ROC curve. The area under the curve for all 12 SNPs studied was (95% CI ¼ , P ¼ 0.002) for both the SBP and DBP, confirming the accuracy of this assay. Discussion Our replication study in a general Japanese population-based cohort confirmed that the COMT, ATP2B1, CYP17A1 and CSK genes significantly predicted the risk of hypertension and were related to BP traits. These findings were independent of confounding factors such as age, body mass index, salt intake and energy intake levels. COMT is a ubiquitous enzyme that is crucial to the metabolism of carcinogenic catechols and catecholamines, and regulation of the human COMT gene expression may be important in the pathophysiology of various human disorders, including estrogeninduced cancers, Parkinson s disease, depression and hypertension. 13 Previous studies have revealed an association with hypertension 10,14,15 and the current study replicated these findings. ATP2B1 encodes PMCA1, which is a plasma membrane calcium/calmodulin-dependent ATPase that is expressed in vascular endothelium. PMCA1 functions as a modifier locus for phasic vascular contraction, and has a critical role in intracellular Ca þþ homeostasis. 16 SNP rs is located on a large cluster of associated SNPs that span a B430-kb region at 10q24 and which shows an association in a meta-analysis. 4 The locus includes six genes, with the mutation near to the CYP17A1, AS3MT, CNNM2 and NT5C2 gene. The most reasonable candidate is CYP17A1, which encodes the cytochrome P450 enzyme CYP17A1 (also known as P450c17) that is responsible for mediating steroid 17a-hydroxylase and 17, 20-lyase activity. The first enzymatic action is a key step in the biosynthesis of mineralocorticoids and glucocorticoids that affect sodium handling in the kidney and the second action involves the sex-steroid biosynthesis. In the current analysis, this SNP appeared to be associated with both the prevalence of hypertension and BP. In fact, our results suggest a much broader mechanism of action than that reported by Newton-Cheh et al., 4 who only found an association with SBP. These researchers also reported finding a polymorphism, rs , which was in the intron of the CSK gene at 15q24. This polymorphism was one of a cluster of associated SNPs that spanned B72 kb and that were strongly associated with DBP. 4 Genes in the region include CYP1A2 (cytochrome P450 enzyme), CSK (c-src tyrosine kinase), LMAN1L (lectin, mannosebinding, 1 like) and ARID3B (encoding AT-rich interacting domain protein). However, at the present time it has yet to be agreed upon as to which of these genes actually contributes to the association with BP. Our results are notable in that the association was with SBP, but not with DBP. Although numerous common variants have been suggested to be disease-associated, these only account for a small fraction of the genetic variation of the complex traits found in human populations, 17 and most of the individual causal risk alleles have genotype relative risks in the range of 1.1 to This may be due to the incomplete linkage disequilibrium between the causal variants and the genotyped SNPs, with the exacerbation by the causal variants having a lower minor allele frequency than the SNPs already examined. 19 However, a recent study that examined the heritability for human height developed a linear model analysis that could be used to estimate the proportion of variance for height related to SNPs. The analyses showed that 45% of the variance could be explained using the model, which was a nearly 10-fold increase relative to the published and validated individual SNPs. 19 It has been assumed that the combined effect of risk variants might make it possible for us to identify individuals with an accumulated genetic risk. 20 In the current study, we evaluated not only the individual variants that may be associated with hypertension, but also the combined impact that they may have on hypertension or BP. The other six SNPs (ACADSB, GREB1, ABCA1, PTK2B, HPCAL1 and CDH13) were not successfully replicated in our study, which may have been due to the relatively small sample size. Even so, we were able to identify the significant combined impact of these 12 genetic polymorphisms on the effect of BP and hypertension, suggesting that accumulation of the polygenic polymorphisms reliably increased the BP and the risk of hypertension in the Japanese population. It has been demonstrated that the utility of a genetic test may be better validated by the method of ROC curves, as this methodology provides the ability to discriminate between individuals who may or may not be prone to developing the disease. 20 In the current study, we used the ROC curve to compare the six SNPs that were confirmed to be associated with the BP and for the 12 SNPs that were found to have a higher area under the curve (0.571 vs for SBP, vs for DBP). These results indicate there is an elevated accuracy of discrimination when using this method, although the overall change was quite small in the current case. Regardless, our results showed there was an inadequate discriminatory ability when there was an area under the curve of This suggests limited predictive value of these genetic tests for complex diseases such as hypertension when there is only information on a few variants available. Although investigations of more common variants

6 may improve the overall accuracy, the potential limitations of this approach need to be considered. For example, in the well-known large-scale GWAS that was reported by the Wellcome Trust Case Control Consortium, 21 none of the variants that were previously associated with hypertension showed any evidence for association. Thus, when dealing with complicated diseases, in addition to genetic factors that may contribute to the pathogenesis and development of the disease, environmental factors, such as dietary energy and salt intake levels, as well as the interaction of genetic and environmental factors, also need to be considered when evaluating the susceptibility. In the present study, SNP rs4680 in the COMT gene showed an interaction between the genotype and dietary energy intake that appeared to be associated with hypertension. Thus, instead of focusing on a single point, future studies need to emphasize the complexity of the multifactorial traits. In the current study, because the percentage of women was relatively low (10%), we selected only male subjects to examine the association. Our results was partly supported by the observation of Annerbrink et al., 15 who has reported that COMT rs4680 polymorphism is associated with abdominal obesity and BP in Swedish men. However, the exclusion of female subjects led to a limitation to better understand the role of these genetic variations on hypertension in a general Japanese population. Larger scale investigation including female subjects is required to examine whether these variables have a similar influence on BPs of women as that of men. In summary, the significant combined impact from the currently known susceptibility variants provided evidence that the current approach could help in identifying individuals at a higher risk for becoming hypertensive, even though the relationship to hypertension for some of these variants could not be replicated in the present study. In addition, although the number of mutations that have been genotyped has increased, the predictive value of this information may be limited by the confounding environmental factors. To overcome this, genetic and environmental factors, along with interactions between the factors, need to be considered simultaneously rather than as separate entities. What is known about this topic K Genome-wide association studies identified some polymorphisms that might affect the risk of hypertension and blood pressure. K Knowledge of multiple-risk alleles throughout the genome could make it possible to identify individuals that are at a high risk. What this study adds K We confirmed the association between prevalence of hypertension or blood pressure and four gene variants. K A significant combined effect of the 12 gene loci we tested was found that appeared to affect risk of hypertension and blood pressure. Conflict of interest The authors declare no conflict of interest. Acknowledgements This study was supported by Grant-in-Aid for Scientific research (B) ( ) from the Ministry of Education, Culture, Sports, Science and Technology (to Dr Miyaki) and Grant-in-Aid for Scientific research (C) ( ) from Ministry of Education, Culture, Sports, Science and Technology (to Dr Muramatsu). References 1 Hajjar I, Kotchen JM, Kotchen TA. Hypertension: trends in prevalence, incidence, and control. Annu Rev Public Health 2006; 27: Sekikawa A, Hayakawa T. Prevalence of hypertension, its awareness and control in adult population in Japan. J Hum Hypertens 2004; 18: Cho YS, Go MJ, Kim YJ, Heo JY, Oh JH, Ban HJ et al. A large-scale genome-wide association study of Asian populations uncovers genetic factors influencing eight quantitative traits. Nat Genet 2009; 41: Newton-Cheh C, Johnson T, Gateva V, Tobin MD, Bochud M, Coin L et al. Genome-wide association study identifies eight loci associated with blood pressure. Nat Genet 2009; 41: Org E, Eyheramendy S, Juhanson P, Gieger C, Lichtner P, Klopp N et al. Genome-wide scan identifies CDH13 as a novel susceptibility locus contributing to blood pressure determination in two European populations. Hum Mol Genet 2009; 18: Angius A, Petretto E, Maestrale GB, Forabosco P, Casu G, Piras D et al. A new essential hypertension susceptibility locus on chromosome 2p24-p25, detected by genomewide search. Am J Hum Genet 2002; 71: Zhu DL, Wang HY, Xiong MM, He X, Chu SL, Jin L et al. Linkage of hypertension to chromosome 2q14- q23 in Chinese families. J Hypertens 2001; 19: Kamide K, Kokubo Y, Yang J, Tanaka C, Hanada H, Takiuchi S et al. Hypertension susceptibility genes on chromosome 2p24-p25 in a general Japanese population. J Hypertens 2005; 23: Yamada Y, Kato K, Yoshida T, Yokoi K, Matsuo H, Watanabe S et al. Association of polymorphisms of ABCA1 and ROS1 with hypertension in Japanese individuals. Int J Mol Med 2008; 21: Kamide K, Kokubo Y, Yang J, Matayoshi T, Inamoto N, Takiuchi S et al. Association of genetic polymorphisms of ACADSB and COMT with human hypertension. J Hypertens 2007; 25: Kamide K, Kokubo Y, Fukuhara S, Hanada H, Yang J, Kada A et al. Protein tyrosine kinase 2beta as a candidate gene for hypertension. Pharmacogenet Genomics 2007; 17: Song Y, Miyaki K, Araki J, Zhang L, Takahashi Y, Nakayama T et al. Influence of CYP11B2 gene polymorphism on the prevalence of hypertension and the blood pressure in Japanese men: interaction with 435

7 436 dietary salt intake. J Nutrigenet Nutrigenomics 2008; 1: Xie T, Ho SL, Ramsden D. Characterization and implications of estrogenic down-regulation of human catechol-o-methyltransferase gene transcription. Mol Pharmacol 1999; 56: Hagen K, Pettersen E, Stovner LJ, Skorpen F, Holmen J, Zwart JA. High systolic blood pressure is associated with Val/Val genotype in the catechol-o-methyltransferase gene. The Nord-Trøndelag Health Study (HUNT). Am J Hypertens 2007; 20: Annerbrink K, Westberg L, Nilsson S, Rosmond R, Holm G, Eriksson E. Catechol-O-methyltransferase val158-met polymorphism is associated with abdominal obesity and blood pressure in men. Metabolism 2008; 57: Carafoli E. The Ca2+ pump of the plasma membrane. J Biol Chem 1992; 267: Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, Hunter DJ et al. Finding the missing heritability of complex diseases. Nature 2009; 461: Wray NR, Goddard ME, Visscher PM. Prediction of individual genetic risk to disease from genome-wide association studies. Genome Res 2007; 17: Yang J, Benyamin B, McEvoy BP, Gordon S, Henders AK, Nyholt DR et al. Common SNPs explain a large proportion of the heritability for human height. Nat Genet 2010; 42: Lango H,, Palmer CN, Morris AD, Zeggini E, Hattersley AT et al. Assessing the combined impact of 18 common genetic variants of modest effect sizes on type 2 diabetes risk. Diabetes 2008; 57: Wellcome Trust Case Control Consortium. Genomewide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature 2007; 447: