Association between the CYP4A11 T8590C Variant and Essential Hypertension: New Data from Han Chinese and a Meta-Analysis

Objective CYP4A11 oxidizes endogenous arachidonic acid to 20-hydroxyeicosatetraenoic acid, a renal vasoconstrictor and natriuretic in humans. Previous studies demonstrated an association between a functional variant (T8590C) of CYP4A11 and essential hypertension, though with conflicting results. To elucidate this relationship, a case-control study and meta-analysis were performed to assess the possible association of essential hypertension with CYP4A11 genetic variations. Methods Associations between the T8590C polymorphism and essential hypertension were examined in 328 unrelated cases and 297 age-matched controls in Han Chinese individuals. High-resolution melting was used to identify the CYP4A11 variant. To further investigate the association, we conducted a meta-analysis including eight studies published previously in July 2012. Results The frequency of the CYP4A11 T8590C polymorphism showed no significant difference between cases and controls (all P>0.05). However, the meta-analysis showed that the CYP4A11 T8590C polymorphism may increase the risk of essential hypertension in an additive model (OR: 1.15, 95% CI: 1.02–1.29, P = 0.02), a dominant model (OR: 1.06, 95% CI: 1.01–1.32, P = 0.03), a recessive model (OR: 1.52, 95% CI: 1.15–2.02, P = 0.003) and a homozygote contrast (OR: 1.38, 95% CI: 1.07–1.78, P = 0.01). Also, a significant relationship was observed among Caucasians in the additive model, the homozygote contrast, the recessive model and the dominant model (all P<0.05). However, no association was observed in an Asian population (all P>0.05). Conclusions This meta-analysis suggests there is a significant association between the CYP4A11 T8590C variant and essential hypertension, especially in Caucasians. The case-control study did not find a significant association among the Han Chinese population, but the controls were poorly matched and meaningful conclusions cannot therefore be made. Further large-scale studies are needed to clarify whether the CYP4A11 T8590C polymorphism is associated with hypertension risk in Asians or has a gender-specific effect.


Introduction
Hypertension is a major risk factor for cardiac, renal and cerebrovascular morbidity and mortality [1]. It is important to understand individual susceptibility to essential hypertension by the use of molecular and genetic approaches, with a view to delivering targeted or individualized therapies to improve the survival rate and quality of life of hypertensive patients.
Hypertension has a substantial heritability that is likely to be of polygenic origin. To date, several variants have been identified by genome-wide association studies (GWASs) to be associated with essential hypertension [2,3]. Other common gene variants have been shown to contribute to essential hypertension according to gender or ethnicity [4][5][6]. Thus, the identification of polygenic determinants is not straightforward.
Cytochrome P450 (CYP) is a superfamily of cysteine-heme enzymes that are key mediators of the oxidative transformation of endogenous and exogenous molecules. Cytochrome P450 enzymes are classified into families, subfamilies and individual isoenzymes based on similarities in their amino acid sequences. Human CYP4A11, which belongs to the cytochrome P450 family 4, catalyzes the conversion of endogenous arachidonic acid to 20tetrahydrocannabinol acid   [7]; 20-HETE is normally produced in renal and cerebral arterioles, glomeruli and the renal tubules. Experimental models of hypertension indicate that 20-HETE can act in either a pro-or an antihypertensive manner, depending on its site-specific expression in the renal vasculature or tubules, respectively. In the renal tubules, 20-HETE inhibits tubular sodium reabsorption, resulting in natriuresis, and acts as an antihypertensive; whereas, 20-HETE acts as a vasoconstrictor and is pro-hypertensive in the renal vasculature [8].
Given the role of 20-HETE in the regulation of hypertension, CYP4A11 is the subject of much research interest. The human CYP4A11 gene is located on chromosome 1p33, spans approximately 12.57 kilobase-pairs and contains 12 exons. More than 120 single nucleotide polymorphisms (SNPs) of the human CYP4A11 gene are listed in the National Center for Biotechnology Information SNP database (http://www.bianca.nl.NIH.gov/ SNP). One of the coding SNPs, rs1126742 (T8590C), which results in a phenylalanine to serine substitution at amino acid 434 in exon 11 of CYP4A11, has been found to be associated with blood pressure (BP) in several populations [9][10][11]. However, the results are conflicting and whether the CYP4A11 T8590C variant has a role in modulating BP in the Chinese Han population remains unknown. Given this background, the aim of this study was to investigate the possible association between the CYP4A11 T8590C variant and essential hypertension in the Chinese Han population. In addition, a meta-analysis was conducted, including studies examining the association between the CYP4A11 T8590C variant and the risk of hypertension, to clarify previous inconsistencies.

Subjects
Patients diagnosed with essential hypertension were recruited at Guangzhou General Hospital of Guangzhou Military Command from 2008 to 2011. Three hundred and twenty-eight essential hypertension patients and 297 age-matched controls were enrolled in the study. BP was measured twice in each participant while seated, after 5 min of rest. Essential hypertension was diagnosed based on the following criteria: seated systolic BP greater than 140 mmHg or diastolic BP greater than 90 mmHg on three occasions within 2 months after the first BP reading, or current use of antihypertensive agents. Individuals with secondary hypertension, cancer, chronic kidney or hepatic disease, stroke, diabetes mellitus or severe psychiatric illness were excluded. All individuals enrolled were from the Han population in China. Diagnosis of diabetes was based on the history of treatment for hypoglycemia and/or confirmed fasting blood glucose .7.0 mmol/l. Individuals who had smoked $10 cigarettes per day for at least 2 years were defined as smokers and those who had drunk $210 g alcohol per week for more than 3 years were defined as drinkers.
This study protocol was approved by the Ethics Committee of Guangzhou General Hospital of Guangzhou Military Command. All individuals in this manuscript gave written informed consent for the publication of their case details.

DNA Extraction and Genotyping
Peripheral venous blood was drawn from each individual. Genomic DNA was extracted from whole blood using the Tiangen DNA blood kit (Pangenetic Biotech Co., Ltd, Beijing, China). The CYP4A11 T8590C allele was genotyped by high resolution melting. The primers used were as follows: forward: 59-TGGGTGGCAAG TAGGTGCTGGA-39; reverse: 59-TGTTGAGCAGAACCCGGTGCA-39. PCR reactions were performed in a final volume of 20 ml using Highlighter 480 High Resolution Melting Master (Reference 04909631001; Roche Diagnostics) and 30 ng DNA. The primers and MgCl 2 were used at 10 mM and 2.5 M, respectively. The PCR program comprised an initial denaturation-activation step at 95uC for 10 min, followed by a 45-cycle program of denaturation at 95uC for 10 s, annealing at 57.8uC for 10 s and elongation at 72uC for 15 s with reading of the fluorescence and a single acquisition mode. The melting program comprised three steps: denaturation at 95uC for 1 min, renaturation at 40uC for 1 min and then melting, with a continuous fluorescence reading from 60 to 90uC at 25 acquisitions per uC.

Statistical Analysis
Statistical analysis was performed using the SPSS 13.0 software package (SPSS, Chicago, IL, USA). The genotype frequency distribution for each variant was separately tested for Hardy-Weinberg equilibrium (HWE) with a chi-square test in the case and control groups. Differences of continuous variables with a normal distribution (presented as mean 6 SD) between the two groups were calculated using the independent-sample t-test. Differences of continuous variables departing from the normal distribution, even after transformation, were tested by the Mann-Whitney U-test. Odds ratios (ORs) and their 95% confidence intervals (CIs) were calculated to determine associations between the polymorphisms and essential hypertension. Two-tailed P values ,0.05 were considered statistically significant.

Meta-analysis
Searching. To explore the association between the T8590C polymorphism of CYP4A11 and essential hypertension in different studies and ethnicities, we performed a meta-analysis of previously published studies (up to July, 2012) concerning the T8590C polymorphism and essential hypertension. Our own data were also included. We systematically searched PubMed, Medline, Embase and the China National Knowledge Infrastructure (CNKI) using the following keywords: ''CYP4A11'', ''cytochrome P450 4A11'', ''polymorphism'', ''gene'' and ''hypertension''. We also retrieved additional studies by hand, searching the bibliographies of original research reports and review articles, and through the Medline option ''related articles''. All relevant referenced articles were also investigated.
The inclusion criteria were: (a) evaluation of the CYP4A11 T8590C polymorphism and the risk of developing hypertension; (b) unrelated case-control studies or cohort studies; (c) availability of the sample size, ORs and 95% CIs or information including genotype frequency of cases and controls that could help in inferring the study results; and (d) presence of non-overlapping data. For duplicate publications, the study with the smaller data set was excluded. Diagnoses of essential hypertension were made in accordance with the diagnostic requirements of the World Health Organization established in 1999. Subjects with systolic BP$140 mmHg or diastolic BP$90 mmHg and subjects who had no secondary hypertension were included.
Data extraction. Two investigators reviewed and extracted data from all of the eligible publications independently, according to the inclusion and exclusion criteria listed above. All data were collected according to a standard protocol. Studies that were repeated, of poor research quality, did not meet the inclusion criteria or provided little information or insufficient data were excluded. Study quality was assessed using the 9-star Newcastle-Ottawa Scale [12]. Table 1 lists the characteristics of the extracted data, including the name of the first author, publication date, region, study quality, number of genotypes, alleles and numbers of cases and controls.
Statistical analysis. The association between the CYP4A11 T8590C polymorphism and essential hypertension was compared between the hypertension and control groups by the OR and the corresponding 95% CI. We evaluated the risk using an additive model (C allele versus T allele), a dominant model (TC+CC versus TT), a recessive model (CC versus TT+TC) and the homozygote contrast (comparison of CC versus TT).
The heterogeneity of the studies was tested using the chi-squarebased Q-test, and was considered statistically significant when P,0.10. The inconsistency index I 2 was also calculated to evaluate the variation caused by the heterogeneity. A high value of I 2 indicated a higher probability of the presence of heterogeneity (I 2 = 0-25%, no heterogeneity; I 2 = 25-50%, moderate heterogeneity; I 2 = 50-75%, large heterogeneity; and I 2 = 75-100%, extreme heterogeneity). If there was heterogeneity among the studies (P,0.1), a random-effects model was used to estimate the combined OR. Otherwise, the pooled OR was estimated by the fixed-effect model [13]. The significance of the combined OR was determined by the Z-test and the significance was set at P,0.05. Fisher's exact test was used to assess the HWE and the significance was set at P,0.05. In addition, sensitivity analysis was carried out by including and excluding studies not in HWE. Potential publication bias was estimated using a funnel plot. Egger's linear regression test was used to evaluate the funnel plot asymmetry on the natural logarithmic scale of the OR (P,0.05 was statistically significant). Review Manager 5.1 and STATA 11.0 software were used to perform statistical analysis (Stactometer, College Station, TX).

Case-control Study
Baseline characteristics of the study population are shown in Table 2. Age, BMI and levels of triglyceride and high-density lipoprotein cholesterol were balanced between cases and controls. The patients showed significantly higher levels of total cholesterol (P,0.001) and low-density lipoprotein cholesterol (P = 0.005) than the controls. Additionally, there were significantly higher numbers of males (P = 0.041) and smokers (P,0.001) in the cases compared to the controls.
For the total population, the frequencies of the 8590TT, 8590TC and 8590CC genotypes were 52.0%, 40.8% and 7.2%, respectively. These frequencies did not deviate from those predicted by HWE. The allele frequencies of the 8590C allele in cases and controls were 26.5% and 28.8%, respectively. No significant difference in the genotype frequency was observed between the cases and controls (P = 0.636), as shown in Table 3. There were also no significant differences between the two groups in the dominant model (P = 0.477), the recessive model (P = 0.418) and the additive model (P = 0.371). The control group was poorly matched to the cases, particularly with regards to other known predictors of hypertension (gender and smoking), which make the results here difficult to interpret.

Meta-analysis
Eligible studies/groups. Eighteen papers were gathered from the literature research, among which seven papers were eligible based on the study selection criteria. Of the eleven excluded studies, one paper was a repeat publication [10], four were reviews [11,[14][15][16] and four were unrelated to the CYP4A11 T8590C polymorphism and hypertension [17][18][19][20]. One paper was excluded because of incorrect data (the combined number in the case group and the control group was not equal to the total number) [21] and one was excluded due to having different criteria for essential hypertension [22] (Figure 1). Finally, data from eight studies (seven previous studies and the present study) were combined [9,[23][24][25][26][27][28], comprising 6863 cases and 5327 controls. Gainer et al. [9] reported two independent cohort studies, as shown in Table 1.
Characteristics of studies and subjects. The main characteristics of the studies included in our meta-analysis are summarized in Table 1. The ORs differed among the eight studies; some indicated that the presence of the C allele increased the risk of hypertension, whereas the others reported no association between the CYP4A11 T8590C polymorphism and hypertension risk. Hence, a composite analysis of the study results was performed to draw a reasonable conclusion.
We conducted subgroup analysis classified by ethnicity, gender and HWE population. Five of the studies assessed Caucasian populations, including one from Australia [26], one from the USA [9], one from Sweden [27] and two from Germany [24,25]. Two studies assessed Asian populations, including one from Japan [23] and one from China. One study included whites, blacks, Asians and others were not analyzed by ethnicity. All studies but one met HWE [27].
Four of the eight eligible studies provided data on male and female subjects. The distribution of genotypes and alleles in males versus females is shown in Table 4. Combined analyses were conducted in the various models.
Main meta-analysis results. The findings for heterogeneity and the combined association between the CYP4A11 T8590C polymorphism and hypertension are shown in Table 5. Overall, a significant association between the CYP4A11 T8590C polymorphism and hypertension was found in the additive model (OR: 1.15, 95% CI: 1.02-1.29, P = 0.02), the dominant model (OR: 1.06, 95% CI: 1.01-1.32, P = 0.03), the recessive model (OR: 1.52, 95% CI: 1.15-2.02, P = 0.003) and the homozygote model (OR: 1.38, 95% CI: 1.07-1.78, P = 0.01). Seven studies also underwent stratified analysis by ethnicity (one study was excluded because of mixed ethnicity). No increased risk was detected among Asians in the additive model, the homozygote contrast, the recessive model or the dominant model (all P.0.05). However, significant differences were detected among Caucasians in the additive model, the homozygote contrast, the recessive model and the dominant model, as shown in Table 5 (all P,0.05). Furthermore, four studies that conducted subgroup analysis by gender were pooled and analyzed in subgroups. No statistically significant differences were found for male or female subjects in any of the various statistical models (all P.0.05), as shown in Table 6.
Test of heterogeneity, sensitivity analysis and publication bias. Significant heterogeneity between studies was observed in some comparisons; detailed data are shown in Table 5. Large heterogeneity was detected in some meta-analyses (I 2 = 59%, P heterogeneity = 0.03). Then, random-effects models were used to evaluate combined ORs when necessary. Sensitivity analysis was performed both by sequential removal (statistics of study removal) of individual studies and by cumulative statistical analysis comparing subjects worldwide.
For the total population, sensitivity analyses indicated that the present study was the main source of heterogeneity. In addition, among male subjects, the combined ORs of the additive model and the dominant model were influenced by one study in HWE.
The P values for the Q-tests became significant (P,0.05) after deletion of the study by Fava et al. Among female subjects, the combined ORs of the additive model, the recessive model and the homozygote contrast were influenced by the present study. The P values for the Q-tests became significant (P,0.05) after deletion of this study. Other combined ORs for the CYP4A11 T8590C polymorphism were not influenced by the addition or removal of any individual study.
A funnel plot and Egger's test were performed to assess publication bias in this meta-analysis. The funnel plot showed no apparent evidence of publication bias (Figure 2). There was also no significant difference using Egger's test in the allelic genetic model (T = 1.99, P = 0.09). The funnel plot and Egger's test suggested that the selection of publications was an unlikely source of bias in this meta-analysis of the association between the CYP4A11 T8590C polymorphism and essential hypertension.

Discussion
The product 20-HETE, the monooxygenase derivative of arachidonic acid synthesized by CYP450 enzymes, is a candidate for the regulation of BP and hypertension. In humans, two members of the CYP4A subfamily have been detected in the kidney. CYP4A11 has been reported to encode an enzyme that converts arachidonic acid to 20-HETE, primarily in the kidney [7]. Both animal and human studies indicate that the CYP4A11 gene is a candidate causative gene for hypertension [22]. Recent studies have attempted to clarify the potential association between CYP4A11 and essential hypertension. CYP4A11 is known to be highly polymorphic, expanding the possibilities for studying the mechanism of action of this gene, especially as a risk factor for cardiovascular diseases [29]. A product of CYP4A11, 20-HETE, has been demonstrated to function in both a pro-hypertensive and an antihypertensive manner. In the renal arterioles, 20-HETE may act as a strong vasoconstrictor, whereas in renal tubules it may attenuate sodium transport and function as a natriuretic, antihypertensive substance. Ito et al. reported recently that CYP4A11 is highly expressed in human renal tubules and   Female and hypertension in whites as a loss-of-function variant. However, in 120 black subjects from the Tennessee cohort, there was no association between the 8590C allele and hypertension [9]. The association between the CYP4A11 T8590C polymorphism and hypertension has also been studied by Mayer et al. [25]. It was reported that individuals with the CC genotype have higher systolic and diastolic BP. Fava et al. also reported that carriers of the C allele had a higher prevalence of hypertension in a prospective cohort including 6002 Swedish subjects [27]. However, another study demonstrated no significant difference in systolic or diastolic BP between the TC/CC genotype and the TT genotype of the CYP4A11 gene in Australians [26]. Similarly, no significant association between the CYP4A11 8590T/C polymorphism and hypertension was observed by Sugimoto et al. in Japanese subjects [23]. The frequency of the (TC+TT) genotype was significantly higher in patients with hypertension than controls in a Japanese population [22]. In addition, 20-HETE has two opposing effects on BP (natriuretic and vasoconstrictor) and it is difficult to predict which effect will be more prominent.
In the present case-control study, no significant difference in the genotype frequency was observed between the cases and controls (P = 0.636) in the total population (Table 3). However, there were more males and smokers in the hypertensive population compared to the non-hypertensive population, which makes the results here more difficult to interpret. Even so, our results were in accordance with the findings of Sugimoto et al. and Ward et al. [23,26], and were inconsistent with the findings of Gainer et al., Mayer et al. and Williams et al [9,24,28]. Considering these inconsistent findings, we performed a meta-analysis of these studies to quantify the available data and generate a robust estimate of the effect of the T8590C polymorphism on hypertension. Meta-analysis has been proven to be a powerful method for analyzing a relatively large number of subjects.
In this meta-analysis, we analyzed the data from eight available case-control studies (including the present study) [9,[23][24][25][26][27][28]. The results concerning the role of CYP4A11 polymorphisms in susceptibility to hypertension were conflicting. When all of the eligible studies were pooled for analysis, there was evidence for an association between the T8590C polymorphism and susceptibility to hypertension in various models. Moreover, on subgroup analysis by ethnicity, we found a significant association in Caucasians but not in Asians, suggesting genetic diversity between ethnicities. Therefore, it may be concluded that there is an association between the CYP4A11 T8590C polymorphism and hypertension risk in the general population, especially among Caucasians. These results may provide conceptually novel approaches for study of the molecular basis of human hypertension, which could lead to new strategies for the early diagnosis and clinical management of this disease.
It is well established that the prevalence of hypertension and cardiovascular disease differs between men and women, the latter being protected from cardiovascular events until menopause; there  is a rapid increase in the risk profile after menopause. Genespecific effects have been indicated as responsible, but most genetic studies do not include sex as a variable in data analysis, and a recent meta-analysis pooling thousands of subjects of previous GWASs did not identify any sex-specific effect of genes associated with BP/hypertension. Experimental models suggest that enzymes involved in 20-HETE biosynthesis and metabolism have a role in the control of BP with a sex-specific effect [11]. In humans, previous studies have shown that the CYP4A11 T8590C polymorphism appears to be a useful genetic marker of essential hypertension in men [18,22]. In the present meta-analysis, although no gender-specific associations were found, after deletion of certain studies the P values became significant. When interpreting these results, caution should be applied. Relatively large heterogeneity was present in the meta-analysis, perhaps due to the limited study sample size. A relatively limited number of studies and samples were analyzed, partly because the original studies had rather limited sample sets. Moreover, differences in the ethnicity and selection of controls, age distribution and lifestyle factors may have contributed to heterogeneity. Further large scale studies are needed to clarify whether the CYP4A11 T8590C polymorphism is associated with hypertension risk in Asians or has a gender-specific effect.

Conclusions
In summary, our meta-analysis suggests that the CYP4A11 T8590C polymorphism is associated with risk of hypertension in the general population, especially in Caucasians.