Introduction

Several molecular variants of the thiazide-sensitive Na–Cl cotransporter (TSC, SLC12A3) relate to Gitelman’s syndrome characterized by their low blood pressure (BP) sodium wasting, secondary hyperaldosteronism, hypokalemia, alkalosis, hypomagnesemia, and hypocalciuria (Mastroianni et al. 1996; Simon et al. 1996; Takeuchi et al. 1996). This syndrome is known to be heritable as autosomal recessive, and the mutations identified in TSC may reduce the capacity of the TSC to reabsorb salt in the distal tubules where the cotransporter is regionally expressed (Mastroianni et al. 1996). On the contrary, mutations in the WNK1 (PRKWNK1) and WNK4 (PRKWNK4) genes relate to familial hypertension known as pseudohypoaldosteronism type II (Wilson et al. 2001), associated with hyperkalemia (despite normal renal glomerular filtration) and renal tubular acidosis caused by impaired renal K+ and H+ excretion. This autosomal dominant disease includes several types of mutations; a large deletion in intron 1 of WNK1, missense mutations in the highly conservative regions of WNK4 (Wilson et al. 2001). Mutations identified in WNK4 so far were all accompanied by charge changes, assuming modification of the protein function.

Recent expression studies have revealed a close link between TSC and WNK family proteins. Coexpression of TSC with WNK4 leads to a significant decrease in thiazide-sensitive sodium uptake (Choate et al. 2003; Wilson et al. 2003). WNK4 was shown consistently to suppress cell surface expression of TSC. Although WNK1 per se was inactive on the transporter activity, it was able to abolish the inhibitory effect of WNK4, suggesting that both proteins act on the same signaling pathway (Wilson et al. 2003; Yang et al. 2003). Thus, WNK4 functions as a negative regulator for the surface expression of Na–Cl cotransporter, and loss of this regulation can cause an inherited form of hypertension. WNK1 seems to act as a suppressor of WNK4, and gain-of-function of this gene can cause loss in WNK4 function leading to an inherited form of hypertension.

It is likely that individual BP level is influenced by several different genetic variants in a general population. A polymorphism in WNK4 (base1156666G>A) has been reported to be associated with hypertension in a Caucasian population (Erlich et al. 2003) with a discrepancy in other studies (Benjafield et al. 2003; Speirs and Morris 2004). We hypothesized that the genetic polymorphisms in TSC, WNK1, and WNK4 may involve changes in BP level. Among the different kinds of genetic variations, single nucleotide polymorphisms (SNP) receive much attention due to their easy genotyping. This study was undertaken to identify genetic variations, mainly SNPs, in all exons of TSC, WNK1, and WNK4 and to examine the association of SNPs with hypertension in a Japanese general population.

Methods

Subjects

The subjects of the Suita study consisted of 14,200 men and women (30–79 years of age), who had been randomly selected from the municipal population registry considering group stratification by gender and 10-year age. They were all invited, by letter, to have a group checkup every 2 years at the Division of Preventive Cardiology, National Cardiovascular Center, Japan. DNA from the leukocytes was collected from participants who visited the National Cardiovascular Center between April 2002 and February 2003. The study protocol was approved by the ethical committees on human research of the National Cardiovascular Center and Suita city. Written informed consent was obtained from each subject for proceeding genetic analyses. In this study, the genotypes of 1,818 individuals including 771 subjects with hypertension (396 men and 375 women) and 1,047 controls (439 men and 608 women) were performed.

Measurements

BP was measured after at least 10 min of rest in a sitting position. Systolic and diastolic BPs (SBP and DBP) were the means of two measurements by well-trained doctors using a mercury sphygmomanometer (recorded in a 3 min pause). Hypertension was defined as SBP of ≥140 mmHg, DBP of ≥90 mmHg or current use of antihypertensive medication.

A physician or nurse questionnaired each patient regarding current smoking and alcohol drinking habits and personal history of cardiovascular disease, including angina pectoris, myocardial infarction, and/or stroke. Hypercholesterolemia was defined as total serum cholesterol levels ≥5.68 mmol/l (≥220 mg/dl) or current use of antihyperlipidemic medication. Diabetes was defined as fasting plasma glucose levels ≥7.0 mmol/l (126 mg/dl) or nonfasting glucose levels ≥11.1 mmol/l (200 mg/dl), HbA1C ≥6.5%, or current use of antidiabetic medication. Body mass index (BMI) was calculated as weight (in kilograms) divided by height (in meters) squared.

Blood samples from the subjects after 12 h of fasting were collected in EDTA-containing tubes. Total cholesterol and high density lipoprotein (HDL) cholesterol levels were measured with an autoanalyzer (Toshiba TBA-80) in accordance with the Lipid Standardization Program of the US Centers for Disease Control and Prevention through the Osaka Medical Center for Health Science and Promotion, Japan.

Direct sequencing for SNP discovery and genotyping of polymorphisms

For DNA sequencing, Japanese patients with essential hypertension at the Division of Hypertension and Nephrology, National Cardiovascular Center, Japan, were recruited. Genomic DNA was extracted using an NA-3000 nucleic acid isolation system (KURABO, Osaka, Japan). We sequenced the 48 or 96 Japanese hypertensive samples in which hypertension-susceptive SNPs would be most concentrated. In exon 22 of TSC and exons 7 and 17 of WNK4, more than 250 Japanese hypertensive samples were sequenced (Kamide et al. 2004). The method of direct sequencing was described previously (Okuda et al. 2002). The polymorphisms were identified by use of Sequencher software (Gene Codes Corporation, Ann Arbor, MI, USA) and confirmed by visual inspection (Takiuchi et al. 2004). SNPs having a minor allele frequency of greater than 5% were defined as candidates for genotyping using the TaqMan-PCR system (Tanaka et al. 2003). Some SNPs were not suitable for genotyping due to the presence of another SNP in the adjacent region. The representative SNPs were genotyped when they were in linkage disequilibrium (r2 over 0.5). Since a missense mutation may directly be susceptible to hypertension, five missense SNPs with minor allele frequencies below 5%, including C4527A (Thr180Lys, TSC), T29320A (Leu849His, TSC), G34372A (Arg904Gln, TSC), C142763T (Arg1945Cys, WNK1), and C15503A (Pro1173Thr, WNK4), were also genotyped.

Statistical analysis

A total of 1,818 subjects who had complete genotype data were recruited for the study. Analysis of variance was used to compare mean values between groups, and if overall significance was demonstrated, the intergroup difference was assessed by means of a general linear model. Frequencies were compared by χ2 analysis.

Association studies of genotypes with BP were performed through logistic regression analysis considering potential confounding variables in risk factors, including age, BMI, present illness (hyperlipidemia and diabetes mellitus), lifestyle (smoking and drinking), and antihypertensive drug use by gender. For multivariate risk factors, adjusted odds ratios were given with 95% confidence intervals. The associations of genotypes with hypertension were expressed in terms of odds ratios adjusted for possible confounding effects, including age, BMI, present illness (hyperlipidemia and diabetes mellitus), and lifestyle (smoking and drinking) by gender. All analyses were performed with SAS statistical software (release 8.2, SAS Institute, Inc., Cary, NC, USA). Linkage disequilibrium was calculated by using the SNPAlyze version 2.1 (DYNACOM Co., Ltd, Mobara, Japan).

Results

Basic characteristic of subjects

The characteristics of all 1,818 participants (835 men and 983 women) are shown in Table 1. Age, SBP, DBP, BMI, percentage of current smokers, percentage of current drinkers, and prevalence of individuals with hypertension and diabetes mellitus were significantly higher in men than in women. Total cholesterol, HDL cholesterol, and percentage of individuals with hyperlipidemia were significantly higher in women than in men.

Table 1 Basic characteristics of subjects in Suita, a Japanese urban population, 2002. HDL high density lipoprotein cholesterol. Values are mean ± SD or percentage. Hypertension indicates SBP ≥140 mmHg and/or DBP ≥90 mmHg or antihypertensive medication. Hyperlipidemia, total cholesterol ≥5.68 mmol/l (220 mg/dl) or antihyperlipidemia medication. Diabetes, fasting plasma glucose ≥7.0 mmol/l (126 mg/dl) or nonfasting plasma glucose ≥11.1 mmol/l (200 mg/dl) or antidiabetic medication
Full size table

Polymorphisms of TSC, WNK1, and WNK4

We sequenced 96 alleles from 48 patients with hypertension in TSC and 192 alleles from 96 patients with hypertension in WNK1 and WNK4, and identified 52, 35, and 21 polymorphisms, respectively (Table 2). There were six, nine, and nine missense mutations in TSC, WNK1 and WNK4, respectively. Among them, missense mutations with minor allele frequencies above 5% were 0, 3, and 0, respectively, indicating that most of the missense mutations were rare. We selected SNPs with minor allele frequencies above 5% for genotyping. Five missense SNPs with the minor allele frequency below 5% were also included. We selected representative SNPs for genotyping when some of the SNPs were in linkage disequilibrium. Finally, 12, 7, and 2 SNPs, in a total of 21 SNPs, were selected for genotyping in population-based samples. The primers and probes of the TaqMan-PCR method are summarized in Table 3.

Table 2 List of 108 polymorphisms and their allele frequencies in TSC, WNK1, and WNK4 genes identified by direct sequencing. dbSNP ID was searched by using SNPper, a CHIP Bioinformatics Tool (Riva and Kohane 2001; http://snpper.chip.org/bio/snpper-enter, as of May 1 of 2003, that was constructed by dbSNP build 112). The apparent linkage disequilibrium (LD), defined by r2 more than 0.5, is indicated by a–l in the LD column. Exon 22 of TSC and exons 7 and 17 of WNK4 were sequenced using more than 250 hypertensive samples
Full size table
Table 3 Primers and TaqMan probes for genotyping of SNPs in TSC, WNK1, and WNK4
Full size table

Susceptible SNPs related to hypertension

The results of the case-control study are shown in Table 4. Among 21 SNPs, the C14717T SNP of WNK4 was significantly associated with hypertension in men (χ2=7.53, p=0.023). SBP in men with the CT+TT genotypes was 3.1 mmHg higher than those with the CC genotype (p=0.042) after adjustment for age, BMI, hyperlipidemia, diabetes mellitus, antihypertensive drug use, smoking, and drinking (Table 5). Multivariate logistic regression analysis with adjustment for age, BMI, hyperlipidemia, diabetes mellitus, smoking, and drinking revealed that the odds ratio for the presence of hypertension for the CT+TT genotypes in C14717T in comparison to the CC genotype in men was 1.62 (95% confidence interval, 1.12–2.33, p=0.010) (Table 6). When the controls were defined as SBP ≤120 mmHg, DBP ≤80 mmHg, or nonmedication, and the hypertensives were defined as SBP ≥160 mmHg, DBP ≥100 mmHg, or current use of antihypertensive medication, the C14717T polymorphism was significantly associated with hypertension in men (CC vs CT+TT, odds ratio=1.91, 95% confidence interval: 1.02–3.58, p=0.045) after adjustment for the confounding factors described above.

Table 4 Genotype distributions of 21 SNPs of TSC, WNK1, and WNK4 in normotensives and hypertensives. n.d. not determined
Full size table
Table 5 Blood pressure levels on genotype of WNK4 C14717T polymorphism. Values are means ± SDs; all adjusted for age, body mass index (BMI), antihypertensive drug use, present illness (hyperlipidemia, diabetes mellitus), and lifestyle (smoking and drinking); diastolic blood pressure (DBP) and systolic blood pressure (SBP) are expressed as mmHg
Full size table
Table 6 Odds ratio for the presence of hypertension for WNK4 C14717T genotype in a Japanese general population. OR odds ratio, CI confidence interval
Full size table

Discussion

Three genes, TSC, WNK1, and WNK4, are potentially strong candidates for essential hypertension (Choate et al. 2003; Wilson et al. 2003). To understand whether these genes influence BP, we sequenced these genes and identified a total of 108 SNPs. To evaluate the association of the SNPs with hypertension, we genotyped 21 representative SNPs in a large members of 1,818 individuals and identified that the C14717T polymorphism in intron 14 in WNK4 was associated with hypertension in men. The TT genotype of this SNP increased SBP by 3.1 mmHg when compared with the CC+CT genotype (Table 5). The association of this SNP with hypertension was observed after multiple adjustments for confounding factors including age, BMI, present illness (hyperlipidemia and diabetes mellitus), lifestyle (smoking and drinking), and antihypertensive medication (Table 6). Therefore, we consider that the C14717T polymorphism in intron 14 in WNK4 was associated with hypertension in our general population.

WNK4 is located on chromosome 17q21.2. Several lines of evidence indicate a region on human chromosome 17q as a gene that influences BP (Baima et al. 1999; Jacob et al. 1991; Levy et al. 2000). A quantitative trait locus of hypertension on the rat chromosome 10, equivalent to human chromosome 17, was identified in spontaneous hypertensive rats (Jacob et al. 1991). This region was reportedly linked with hypertension using hypertensive sib pairs from the United Kingdom and France (Julier et al. 1997) and was confirmed in a study of white American hypertensive sib pairs (Baima et al. 1999). Evidence obtained from the Framingham Heart Study indicated that this region is associated with BP with LOD score of 4.7 (Levy et al. 2000). Thus, these studies suggest that this region may contain a gene susceptible for BP elevation.

The C14717T polymorphism in WNK4 associated with hypertension was found in the intron. Therefore, it is not likely that it directly affects the function of WNK4, leading to hypertension. The C14717T polymorphism may be in linkage disequilibrium with another genetic variation in the region that was not examined by sequencing. The functional SNP may be present in the 5′-upstream region beyond our sequencing region or in the intron, creating a new splicing site. Further analysis is needed to clarify the function of this SNP. In conclusion, our study has shown the possible involvement of WNK4 in essential hypertension in the Japanese general population.