This article is mentioned in:

Introduction

Blood pressure regulation is a complex, multifactorial process, which is associated with physiological, biochemical and molecular mechanisms. An increase in Na+ and water retention is required for the development of most forms of hypertension (1). The epithelial sodium channel (ENaC) mediates the initial step of active sodium reabsorption, which is essential for the maintenance of body salt and water homeostasis (2). The ENaC is composed of three different subunits (3), one of which is the amiloride-sensitive sodium channel beta subunit (SCNN1B). Mutations in SCNN1B may cause Liddle's syndrome (4–6), which is an autosomal dominant disorder that is characterized by early, and frequently severe, hypertension (7). Essential hypertension (EH), the form of hypertension that by definition has no identifiable cause, tends to be familial and may be due to an interaction between environmental and genetic factors.

DNA methylation refers to the addition of a methyl group to the cytosine or adenine DNA nucleotides in mammalian cells, and usually occurs at CpG islands, which contain clusters of CpG dinucleotides. Promoter hypermethylation often silences gene transcription and is an important event during disease progression (8). Our previous study demonstrated that reduced α-adducin gene promoter methylation increased the risk of essential hypertension (EH) in Chinese men and women (9). Furthermore, hypo/hypermethylation of the sulfatase 1 gene may serve an important role in the pathogenesis of hypertension in young African American men (10). Altered gene methylation has also been observed in other cardiovascular diseases (11,12), as well as in type 2 diabetes (13–16).

DNA methylation has been reported to mediate some of the effects of environmental exposure and lifestyle factors on disease risk (17). DNA methylation is heritable, but can also be altered by medical therapy. It has previously been reported that drugs can alter the gene methylation status in patients with cancer and schizophrenia (18). In light of these previous studies, the present study aimed to investigate whether SCNN1B promoter DNA methylation was associated with the risk of EH, and whether antihypertensive drug treatment would alter the SCNN1B methylation status in patients with EH.

Materials and methods

Sample collection

A total of 282 individuals (94 incident cases, 94 prevalent cases and 94 normotensive controls) were recruited to the present study from the Ningbo Seventh Hospital (Ningbo, China). All individuals selected were Han Chinese that had resided in Ningbo for ≥3 generations. Incident cases were hypertensive patients who had previously never received any antihypertensive drug treatment. Prevalent cases were hypertensive patients who had previously received antihypertensive drug treatment. The diagnosis of hypertension was made when the average of ≥2 diastolic blood pressure (DBP) measurements on ≥2 subsequent visits was ≥90 mmHg, or when the average of several systolic blood pressure (SBP) readings on ≥2 subsequent visits was consistently ≥140 mmHg. Isolated systolic hypertension is defined as SBP ≥140 mmHg and DBP <90 mmHg. Individuals with SBP <120 mmHg and DBP <80 mmHg were recruited as controls (19). None of the control group had received antihypertensive therapy. Furthermore, none of the recruited individuals suffered from other diseases, including diabetes mellitus, secondary hypertension, myocardial infarction, stroke, renal failure, and drug abuse. Blood samples were collected in 3.2% citrate sodium-treated tubes and were then stored at −80°C for DNA extraction. The protocol of the present study was approved by the ethical committee of the Ningbo Seventh Hospital. Written informed was obtained from all subjects.

DNA isolation and bisulfite treatment

Genomic DNA was extracted and underwent bisulfite conversion for subsequent polymerase chain reaction (PCR) amplification and pyrosequencing, as described previously (11,13–15,20). Genomic DNA was extracted from peripheral blood samples using a nucleic acid extraction analyzer (Lab-Aid 820; Xiamen Zeesan Biotech Co., Ltd., Xiamen, China). The NanoDrop 1000 spectrophotometer (NanoDrop; Thermo Fisher Scientific, Inc., Wilmington, DE, USA) was used to measure the concentrations of extracted DNA. Subsequently, ~500 ng genomic DNA isolated from whole blood cells was bisulfite-treated using the EZ DNA Methylation-Gold™ kit (Zymo Research Corp, Irvine, CA, USA), according to the manufacturer's protocol. This treatment involves converting unmethylated cytosines into uracil, whereas methylated cytosines remain unchanged. Once converted, the methylation profile of the DNA can be determined by PCR amplification followed by DNA sequencing.

DNA methylation assay

Ten primer sets (including forward, reverse and sequencing primers) were designed by PyroMark Assay Design software v2.0.1.1 (Qiagen, Inc., Valencia, CA, USA) to amplify the CpG island region in overlapping fragments and to sequence target DNA fragments. Each primer set was given a score and these primer sets were ranked from high to low. According to the rank, the top three primer sets were synthesized and purified by high-performance liquid chromatography by Sangon Biotech (Shanghai) Co., Ltd. (Shanghai, China) for PCR. Subsequently, 2% agarose gel electrophoresis was used to analyze the PCR products. The results indicated that primer set one exhibited the best amplification effects. Detailed information regarding this primer set is presented in Table I. PCR was conducted in a final volume of 20 µl containing 10 µl Zymo Taq™ PreMix, 1.5 µl forward primer (10 µM), 1.5 µl reverse primer (10 µM) and 40 ng DNA and water was used to raise volume to 20 µl. The PCR process began with an initial denaturation step at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 30 sec, annealing at 58°C for 40 sec and extension at 72°C for 50 sec. The final extension step was performed at 72°C for 7 min. Following amplification, PCR products were maintained at 4°C for ≥4 min. The target sequence was finally sequenced by synthesis assay (Pyromark Gold Q24 Reagents; #970802; Qiagen, Inc.).

Table I. Oligonucleotides for bisulfite sequencing.

Table I.

Oligonucleotides for bisulfite sequencing.

Variable	Sequence	Nucleotides (bp)
Forward primer	5′-GGATGAGGGGTTTGTGGATA-3′	−227 to −207
Reverse primer	5′-ACCTCCCTCCCCTCCCAATAAACT-3′	−66 to −42
Amplicon sequence	5′-GGATGAGGGGTTTGTGGATATATTCGTGGCGTATGTGGGTATCGTTGGTGTTTCGAGGTGGGGAGGGAGAATGCGGAGCGCGTGCGTGCGGGGGGCGTTTAGTGTTTTTGAATTTGGCGTGTGGGGGTTGGAGTTTATTGGGAGGGGAGGGAGGT-3′	−227 to −42
Sequencing primer	5′-GGTGGGGAGGGAGAA-3′	−172 to −158
Sequence to analyze	5′-TGCGGAGCGCGTGCGTGCGGGGGGCGTTTAGTGTTTTTGAATTTGGCG-3′	−156 to −107

[i] Bold and underlined font indicates the CpG sites

Statistical analysis

Statistical analyses were performed using PASW Statistics 18.0 software (SPSS, Inc., Chicago, IL, USA). The mean of continuous variables [including age, body mass index (BMI), DNA methylation level, total cholesterol (TC), triglycerides (TG), glucose and alanine aminotransferase (ALT)] between case and control groups was analyzed using two-tailed unpaired t test or one-way analysis of variance followed by least significant difference multiple comparison tests. Either Pearson χ2 test or Fisher exact test were used to analyze the association between EH and categorical variables including gender, and smoking and drinking habits. Partial correlative test was used to analyze the relationship between continuous variables. Logistic regression was implemented for controlling the possible confounding factors to analyze the correlation between independent variables and dependent variables. Meanwhile, R software and GraphPad Prism (version 5.01; GraphPad Software, Inc., La Jolla, CA, USA) were used for statistical computing and graphical representation. P<0.05 was considered to indicate a statistically significant difference.

Results

As presented in Fig. 1, PCR primers were designed for amplification of the CpG island region of SCNN1B, whereas the sequencing primer was set to sequence the fragment containing seven CpG dinucleotides (Fig. 1 and Table I). Since the last CpG dinucleotide (CpG7) was not well sequenced, the remaining six CpG sites (CpG1-6) were analyzed in the present study (Fig. 1). Significant correlations were found among CpG1-6 (Fig. 1; r>0.30; P<0.01) in DNA methylation levels.

Figure 1. Correlations among the six CpGs in the SCNN1B gene promoter. SCNN1B, amiloride-sensitive sodium channel beta subunit.

A series of EH risk factors (including BMI, TC, TG and glucose) were compared among the three groups using the variance analysis. As shown in Table II, there were significant differences among the three groups in the mean levels of BMI (F=12.478, P=6.45E-06), TG (F=4.631, P=0.011) and glucose (F=3.982, P=0.02). In addition, alcohol and smoking consumption were compared among the three groups, since they are known to be associated with hypertension (21). The results demonstrated that the ratio of alcohol drinking (χ2=9.189; P=0.011) was inconsistent among the three groups using the χ2 test (Table II).

Table II. Comparison of characteristics among the three groups (n=282).

Table II.

Comparison of characteristics among the three groups (n=282).

	Non-EH	Incident cases	Prevalent cases
Characteristic	(Mean ± SD)	(Mean ± SD)	(Mean ± SD)	F/χ2	P
Age (years)	58.36±7.53	56.31±7.73	58.37±7.95	2.217	0.111
Gender (M/F)	29/65	34/60	33/61	0.537	0.642
Smoking (Y/N)	17/77	21/73	20/74	0.586	0.714
Drinking (Y/N)	7/87	21/73	20/74	0.007	0.011
BMI (kg/m2)	22.752±3.721	24.331±2.543	24.962±3.021	12.478	6.45×10−6
TC (mmol/l)	5.231±1.002	5.381±1.043	5.491±1.032	1.456	0.235
TG (mmol/l)	1.521±0.892	1.893±1.071	1.591±0.702	4.631	0.011
Glu (mmol/l)	5.301±0.642	5.371±0.591	5.562±0.760	3.982	0.020
ALT (IU/l)	24.490±40.231	24.862±16.781	25.181±25.261	0.013	0.987
CpG1 (%)	16.211±4.082	17.371±4.090	16.072±2.628	3.555	0.030
CpG2 (%)	10.021±2.778	9.123±2.252	9.121±1.652	4.952	0.008
CpG3 (%)	6.391±1.904	6.550±2.040	6.361±1.490	0.298	0.743
CpG4 (%)	8.841±2.192	9.140±2.671	9.430±3.281	1.065	0.346
CpG5 (%)	11.104±4.250	10.961±2.870	11.060±4.041	0.035	0.966
CpG6 (%)	6.061±2.493	6.402±1.601	6.432±1.464	1.068	0.345

[i] Values in bold represent significant differences. EH, essential hypertension; SD, standard deviation; M, male; F, female; Y, yes; N, no; BMI, body mass index; TC, total cholesterol; TG, triglyceride; Glu, glucose; ALT, alanine aminotransferase.

Among the six CpG sites, CpG1 (F=3.555, P=0.03) and CpG2 (F=4.952, P=0.008) methylation levels exhibited a significant difference among the three groups. Increased methylation levels of CpG1 were detected in the incident cases compared with in the other two groups. Conversely, lower methylation levels of CpG2 were detected in the cases compared with in the controls (Table II). In addition, higher methylation levels of the two CpG sites were detected in women compared with in men for controls (CpG1: t=−2.283, P=0.025; CpG2: t=−2.568, P=0.012, Fig. 2) and incident cases (CpG1: t=−2.694, P=0.009; CpG2: t=−2.583, P=0.011, Fig. 2). Notably, for the two CpG sites no significant difference was observed between males and females in the prevalent cases group (CpG1: t=0.409, P=0.068; CpG2: t=0.621, P=0.536), thus suggesting a role for antihypertensive therapy in the modification of DNA methylation.

Figure 2. Comparison of DNA methylation between men and women.

Since aging is able to alter levels of DNA methylation (22), the present study further explored the association of SCNN1B CpG methylation with age using the partial correlative test. Significant correlations were found between age and CpG1 (Controls: r=0.285, P=0.008; Incident cases: r=0.401, P=0.0001; Prevalent cases r=0.367, P=0.001, Fig. 3) with an adjustment for other metabolic phenotypes (including BMI, TC, TG, glucose, ALT, and smoking and alcohol drinking habits). In addition, no correlations were detected between SCNN1B methylation and these aforementioned metabolic phenotypes.

Figure 3. Correlation between age and CpG1 methylation. Correlation coefficients and P-values are all adjusted by gender, body mass index, total cholesterol, triglycerides, glucose, alanine aminotransferase, and smoking and drinking habits.

In the present study, EH cases were separated into two groups: Incident cases without antihypertensive therapy and prevalent cases with antihypertensive therapy, since drug treatment may influence DNA methylation (23–26). The current study compared the two case groups to explore whether antihypertensive therapy may affect SCNN1B CpG methylation status using a logistic regression test.

As shown in Table III, the P-values and odds ratio (OR) values were all adjusted by other parameters. When analyzing the association between a specific independent variable and dependent variable, the other independent variables are controlled in a logistic regression model. When the P-value is <0.05, this suggests that the variable has an effect on the dependent variable. When the OR value in a logistic regression model is >1, this variable acts as a risk factor for the dependent variable. Conversely, when the OR value is <1, this variable acts as a protective factor. As determined using this analytical model, methylation of CpG1 (β-standardized=0.17, OR=1.185, adjusted P=0.015) and CpG2 (β-standardized=−0.41, OR=0.663, adjusted P=0.001) were significantly associated with EH, since both P-values were <0.05 in the logistic regression analysis of controls and incident cases. In this model, the OR of CpG1 was 1.185, which suggests that hypermethylation of CpG1 is associated with EH; however, the OR of CpG2 was 0.663, which suggests that hypomethylation of CpG2 may increase the probability of EH. In addition, according to this statistical model, the methylation level of CpG1 (β-standardized=−0.252, OR=0.777, adjusted P=3.77E-04, Table III) was higher in the incident cases compared with in the prevalent cases in the regression model of the two cases. This result suggests that antihypertensive therapy may lower the methylation level of CpG1. No significant associations were detected in the remaining CpG sites.

Table III. Variables in the logistic regression model.

Table III.

Variables in the logistic regression model.

Variable	βa	ORa	Pa	βb	ORb	Pb	βc	ORc	Pc
Gender	0.146	1.157	0.732	0.242	1.274	0.559	0.414	1.513	0.340
Age	−0.066	0.936	0.019	0.062	1.064	0.014	−0.044	0.957	0.136
BMI	0.136	1.146	0.028	0.062	1.064	0.321	0.246	1.279	2.89×10−4
TC	0.248	1.281	0.191	0.114	1.121	0.508	0.131	1.140	0.539
TG	0.500	1.649	0.012	−0.558	0.573	0.008	0.029	1.030	0.905
Glu	0.338	1.403	0.275	0.479	1.615	0.103	1.320	3.743	0.001
ALT	−0.004	0.996	0.554	0.002	1.002	0.845	0.004	1.004	0.442
CpG1	0.170	1.185	0.015	−0.252	0.777	3.77×10−4	0.005	1.005	0.934
CpG2	−0.410	0.663	0.001	0.183	1.201	0.176	−0.530	0.589	1.66×10−4
CpG3	0.084	1.088	0.523	−0.095	0.909	0.527	0.052	1.054	0.738
CpG4	0.075	1.078	0.526	0.137	1.147	0.191	0.397	1.487	0.080
CpG5	−0.054	0.948	0.473	−0.073	0.930	0.349	−0.207	0.813	0.057
CpG6	0.186	1.204	0.145	−0.004	0.996	0.981	0.346	1.413	0.078
Smoking	−0.350	0.704	0.505	0.317	1.372	0.547	−0.306	0.737	0.595
Drinking	1.345	3.838	0.034	−0.064	0.938	0.897	1.291	3.638	0.048
Constant	−4.423	0.012	0.067	−5.152	0.006	0.032	−11.002	0.001	5.84×10−5

a Represents the logistic regression analysis of control and incident case groups

b Represents the logistic regression analysis of incident case and prevalent case groups

c Represents the logistic regression analysis of control and prevalent case groups. Values in bold represent significant differences. OR, odds ratio; BMI, body mass index; TC, total cholesterol; TG, triglyceride; Glu, glucose; ALT, alanine aminotransferase.

Discussion

The present study observed opposite results in the association of CpG1 and CpG2 methylation with EH. CpG2 methylation was significantly lower in incident cases and prevalent cases compared with in the controls. In addition, CpG2 methylation was observed to be inversely correlated with blood pressure in controls and incident cases. Notably, CpG2 methylation was not significantly different between the incident cases and the prevalent cases, thus suggesting that the antihypertensive therapy of EH did not affect CpG2 methylation levels.

Previous studies have revealed the correlation between DNA methylation and blood pressure (27–29). The present study hypothesized that CpG1 hypermethylation and CpG2 hypomethylation may increase blood pressure by upregulating the protein expression of SCNN1B, thus amplifying the function and activity of the ENaC and leading to increased sodium reabsorption and water retention. However, to what extent this change will alter ENaC expression, and to what extent this change will affect sodium reabsorption requires follow-up research. In our subsequent studies, we aim to focus on functional research regarding the effects of DNA methylation on ENaC expression.

Drug treatment may affect DNA methylation levels (23,24,26). Several CpG sites have been reported to be differentially methylated between patients with Behçet's disease (BD) prior to and following treatment (23), providing strong evidence that DNA methylation was modified by BD treatment. Furthermore, bisphosphonate treatment in 68 patients with hypocalcaemia altered the DNA methylation levels of excision repair cross-complementation group 8, prolyl 3-hydroxylase 2 and syndecan 2 genes, thus affecting the cumulative bisphosphonate exposure levels (24). Glucocorticoid treatment has also been reported to induce acute and long-term effects on DNA methylation states in the fetus and offspring (26). Similarly, the present study indicated that CpG1 methylation, rather than CpG2 methylation, was likely to be altered by antihypertensive drug treatment. These results may provide novel information regarding pharmaco-epigenomic research of EH.

Alterations in DNA methylation patterns are a hallmark of aging (22). Previous studies have demonstrated that DNA methylation levels are altered alongside aging in several tissue types in mice and humans (30–34). Aging is also known to be a risk factor in the progression of hypertension (35,36). The results of the present study revealed that a positive correlation existed between aging and SCNN1B CpG methylation levels. Furthermore, gender differences in DNA methylation levels are frequently reported (37–44). In the present study, DNA methylation levels of CpG1 and CpG2 in SCNN1B were higher in women compared with in men. Epigenetic changes associated with aging, and gender differences in DNA methylation, may provide clues to elucidate the mechanisms underlying hypertension.

In conclusion, the present study was the first, to the best of our knowledge, to demonstrate a drug-, age-, and gender-dependent association between SCNN1B promoter methylation and EH.

Acknowledgements

The present study was supported by Zhejiang Province Social Development Research Project (grant no. 2016C33178), K.C. Wong Magna Fund in Ningbo University, Ningbo Social Development Research Project (grant no. 2014C50051) and Ningbo Medical Science and Technology Plan Project (grant no. 2013A39).

Glossary

Abbreviations

Abbreviations:

References