Abstract
Objective: An imbalance in sex hormone ratios has been identified in coronary heart disease (CHD), and as a key enzyme in the conversion of androgen to estrogen, aromatase plays an important role in the balance of sex hormone levels. However, there is a paucity of research into the potential roles of aromatase in CHD. In this study, we investigated associations between single-nucleotide polymorphisms (SNPs) in the CYP19 gene, which encodes aromatase, and CHD. Methods: We collected 1706 blood samples from CHD patients and control participants and used propensity score matching techniques to match case and control groups with respect to confounding factors. In a final study population, including 596 individuals, we conducted a case-control study to identify associations between three SNPs in CYP19 and CHD using χ2 or Fisher exact tests, and binary logistic regression analysis. Differences in lipid levels and parameters of echocardiography among individuals with different genotypes were assessed by one-way analysis of variance. Results: The distributions of rs2289105 alleles in the CYP19 gene differed significantly between the CHD and control groups (p = 0.014), and the heterozygote CT genotype was associated with a significantly lower risk of CHD compared to the homozygous wild-type CC genotype (p = 0.0063 and odds ratio = 0.575). However, blood lipid levels and echocardiographic parameters among individuals with different genotypes did not differ between the CHD and control groups. Conclusions: The CT genotype of the rs2289105 polymorphism in the CYP19 gene is associated with a decreased risk of CHD and may be a genetic marker of protection from CHD.
Introduction
C
The average age at the onset of symptomatic CHD in women is reported to be about 10 years older than that in men (Wenger, 1997), and a delay in the occurrence of menopause is associated with a decrease in the cardiovascular mortality rate for postmenopausal woman (Van der Schouw et al., 1996). Incredibly, after menopause, the risk of cardiovascular disease among women increases rapidly and eventually is equivalent to that of men (Barrett-Connor and Bush, 1991; Isles et al., 1992; Davis et al., 1994; Mendelsohn and Karas, 2005).
Unfortunately, the roles of sex steroids in myocardial pathophysiology remain uncharacterized, and the adverse effects of hormone replacement therapy are thought to be outweighed by the advantages (Grodstein et al., 1996; Grodstein and Stampfer, 1998). Dai et al. (2012) described a negative correlation between the estradiol/testosterone ratio and aromatase, as well as imbalance of the serum estradiol/testosterone ratio in women with CHD. Recently, Konstantian et al. found that a genetic variant in CYP19 shows a correlation with CHD (Bampali et al., 2015). In addition, aromatase deficiency has been observed in a number of hyperandrogenic patients (Harada et al., 1992). Other studies demonstrated that aromatase suppression may increase the development of atherosclerotic plaques, and aromatase knockout mice exhibit abnormal glucose tolerance, insulin resistance, and hypercholesterolemia, which may be lead to the onset of CHD (Scott et al., 2012; Verma et al., 2012; Gagliardi et al., 2014). Based on the current literature, androgen and estrogen have received the most attention in studies of hormones in cardiovascular pathologies, whereas aromatase, which is a key enzyme in the conversion of androgen to estrogen in specific tissues, has received far less attention.
Aromatase is encoded by the CYP19 gene, and previous research indicated that aromatase is expressed predominantly in the coronary vasculature. This evidence of cardiac aromatase expression suggests that the local cardiac androgen-estrogen system likely affects heart function and structural modeling (Jazbutyte et al., 2012). In addition, genetic variations in the CYP19 gene have been shown to result in the alteration of blood levels of sex hormones (Wang et al., 2011; Koudu et al., 2012; Zhang et al., 2012b). When Ma et al. (2005) resequenced all coding exons, all upstream untranslated exons indicated that genetic variations in CYP19 might contribute to variations in the pathophysiology of estrogen-dependent diseases.
In the present study, we hypothesized that CYP19 gene polymorphisms might lead to an imbalance between androgen and estrogen, and thus, one or more such polymorphisms may have an impact on the coronary vascular pathology. We conducted a case-control study to examine the associations between polymorphisms in CYP19 and CHD among a Chinese population.
Methods
Study population
From 2010 to 2013, 1706 individuals were recruited from the Department of Cardiovascular Medicine at First Affiliated Hospital of XinJiang Medical University. Although our study population contained individuals of both genders (983 men and 723 women) and different ethnicities (Table 1), we used propensity score matching techniques to match case and control groups to eliminate the effect of confounding factors. Height, weight, and blood pressure were measured, and body-mass index (BMI) was calculated. Participants completed a study survey regarding their personal medical history (hypertension, diabetes mellitus, etc.), familial medical history, reproductive history, menopausal status, and lifestyle habits (smoking, drinking, etc.) Blood samples were drawn for routine analysis of blood levels, biochemical tests, coagulation function, and genetic analyses. Written informed consent was obtained from all participants, and ethics approval was granted by the medical ethics committee of First Affiliated Hospital of XinJiang Medical University.
Continuous variables are expressed as mean ± standard deviation. Continuous variables were compared by independent sample t-tests. Differences in categorical variables were analyzed using χ2 test or Fisher exact test.
p < 0.05.
BMI, body-mass index; CHD, coronary heart disease; SBP, systolic pressure; DBP, diastolic pressure; EH, essential hypertension; DM, diabetes mellitus; HB, hemoglobin; PLT, platelet; PT, prothrombin time; Fg, fibrinogen; Glu, glucose; TG, triglyceride; TC, total cholesterol; HDL, high-density lipoprotein; LDL, low-density lipoprotein; apoA, apolipoprotein A; apoB, apolipoprotein B; LP(a), lipoprotein; TP, total protein.
All patients had received a differential diagnosis for chest pain or pressure and tightness in the chest after examination in the Cardiac Catheterization Laboratory of the First Affiliated Hospital of XinJiang Medical University, and all coronary angiography procedures were performed by experienced and skilled physicians using the Judkins technique. The findings of coronary angiography were interpreted by at least two knowledgeable imaging specialists, who were blinded to the clinical date, and the final diagnosis of CHD was made according to the angiography report and the standard 15-segment model established by the American Heart Association in 1975 (Austen et al., 1975). All patients were evaluated by cardiac ultrasound, which was performed by doctors with more than 10 years of experience. Similarly, the results of cardiac ultrasound were analyzed by two specialists together.
The study population included 962 patients with CHD (331 from the Han population and 631 from the Uygur population), whose coronary angiographic examination showed at least one significant coronary artery stenoses of more than 50% the luminal diameter. The control population included 744 individuals (404 from the Han population and 340 from the Uygur population). These participants did not have coronary vessel stenosis, and the exclusion criteria included obvious clinical, electrocardiographic, or echocardiography evidence of myocardial ischemia, myocardial infarction, valvular disease, cardiomyopathy, and previous stent deployment or bypass surgery. Patients, also, were excluded if they exhibited impaired renal function, malignancy, or plaque formation beginning in the neck vessels. Hypertension was diagnosed according to guidelines established by the World Health Organization and the International Society of Hypertension in 1999 (Chalmers et al., 1999), and diabetes mellitus was diagnosed according to the criteria of the American Diabetes Association (Mellitus, 2002).
Biochemical analyses
The main blood indices that have been previously associated with CHD were measured in the Clinical Laboratory Department of the First Affiliated Hospital of Xinjiang Medical University using standard methods. These indices included prothrombin time and levels of hemoglobin (Lawler et al., 2013), platelets, fibrinogen (O'Connor et al., 1984), glucose (Kannel and McGee, 1979), triglyceride (Hulley et al., 1980; Do et al., 2013), total cholesterol, high-density lipoprotein, low-density lipoprotein (May et al., 2012), apolipoprotein A, apolipoprotein B (Boekholdt et al., 2012), lipoprotein (Tsimikas and Hall, 2012), and the total protein.
Genotyping
We selected three single-nucleotide polymorphisms (SNPs) of the CYP19 gene that had a minor allele frequency >0.03 in the Chinese population according to the National Center for Biotechnology Information (NCBI) SNP database (www.ncbi.nlm.nih.gov/projects/SNP), considering prior resequencing data and functional studies (Ma et al., 2005) (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/gtmb).
Genomic DNA was isolated from peripheral blood leukocytes using the phenol-chloroform method (Gross and Rotzer, 1998). DNA was dissolved in 200 μL sterile distilled water. Then, the DNA concentration was quantified by ultraviolet/visible (UV/Vis) spectrophotometry (http://chem247.files.wordpress.com/2007/09/chem-247-dna-lab.pdf), and samples were stored at −80°C. Finally, we analyzed the genotype with pure, integrated, and qualified DNA samples.
Genotyping was conducted using the TaqMan SNP Genotyping Assay (ABI 7900) following the manufacturer's instructions. Briefly, polymerase chain reaction (PCR) amplification was conducted in a total volume of 6 μL containing 2.5 μL Master mix, 0.15 μL SNP mix, 0.5 μL TE buffer, 1.85 μL double-distilled water, and 1 μL of DNA sample. The primers and FAM/VIC-labeled probes were designed by Applied Biosystems (http://tools.lifetechnologies.com/content/sfs/brochures/cms_040597.pdf). The assay IDs of the selected assays were C__8234946_20(rs12050772), C__15880593_10(rs2289105), and C__27892984_20(rs4774585).
Statistical methods
Chi-square tests for genotype distribution were conducted to evaluate the deviation from Hardy-Weinberg equilibrium for the three SNPs. Data are shown as mean ± standard deviation (SD), and baseline characteristics were compared by independent sample t-tests or Chi-square tests. Statistical significance was established at p < 0.05. Spearman and Hoeffding correlations and multiple logistic regression analyses were used to identify correlations between the independent variables and CHD. Differences in categorical variables were analyzed using Fisher exact test. The distribution of genotypes between CHD and control participants was tested using χ2 tests or Fisher exact test and binary logistic regression analysis, and differences in lipids and the parameters of echocardiography among individuals with different genotypes were assessed by one-way analysis of variance (ANOVA). Again, a p value <0.05 was considered statistically significant. Analyses were performed using SAS software (Cary, NC).
Results
Identification of clinical variables associated with CHD
Table 1 shows the clinical characteristics of the study participants, and the mean values of some variables differed significantly between CHD patients and control participants. Notably, more CHD patients had high protein levels than the controls. We then used the Spearman rank correlation and Hoeffding D measurement methods to reject variables that showed no correlation with our final variable, CHD, to reduce the number of variables to be matched in subsequent analyses (Supplementary Tables S2 and S3).
Genotyping of study groups
According to the genotype and allele distribution data presented in Table 2, among the Han population, the distribution of rs2289105 differed significantly between CHD patients and control participants among the total population and among male participants (p = 0.0019 and p = 0.027, respectively). In the Uygur population, which are a Eurasian (mixed ancestry) population with Eastern and Western Eurasian anthropometric and genetic traits, independent of gender, the distribution of rs4774585 differed significantly between CHD patients and control participants (p = 0.0424 for the total population, p = 0.0485 for men, and p = 0.0025 for women).
p values were calculated by χ2 test or Fisher exact test.
p < 0.05.
Genotype comparison between CHD patients and control participants after matching
We used the SAS “pscore” command to generate propensity scores, and the code and output produced by the “pscore” command have been described previously (Coca-Perraillon, 2007). This procedure automatically tests for balance between the case and control groups on covariates used to predict the propensity score, and when we controlled the differences in pscores from 0 to 0.1, our total of 596 samples (298 control participants and 298 CHD patients) remained to the end. The Chi-square (χ2) test was used to compare the independent variables (previously segmented) between the cases and controls after matching, and the analysis confirmed that there were no differences (p > 0.05; Supplementary Table S4).
The data in Table 3 show that in the 596 study subjects, the CYP19 rs12050772 genotypic distributions for CHD patients (0.203 for GG, 0.449 for GT, and 0.348 for TT) differed from those for the controls (0.124, 0.559, and 0.318, respectively), but unfortunately, the genotypic distribution for control participants was not in Hardy-Weinberg equilibrium. In contrast, the distributions of the CYP19 gene rs2289105 and rs4774585 polymorphisms were in Hardy-Weinberg equilibrium for both groups. The CYP19 rs2289105 genotypic distributions for CHD patients (0.288 for CC, 0.435 for CT, and 0.277 for TT) were significantly different from those for the controls (0.208, 0.547, and 0.245, respectively), whereas the genotypic distributions of CYP19 rs4774585 for CHD patients (0.007 for AA, 0.114, for AG, and 0.879 for GG) did not differ from those of the control participants (0.013, 0.124, and 0.862, respectively).
p values were calculated by χ2 test or Fisher exact test.
p < 0.05.
Binary logistic regression analysis (Table 4) showed that compared to the GG genotype of the distribution of rs12050772, the GT genotype was associated with a significantly lower risk of CHD (p = 0.003 and odds ratio OR = 0.491), but again, unfortunately, the genotypic distribution of rs12050772 was not in Hardy-Weinberg equilibrium in control participants. We also observed that the rs2289105 heterozygote GT was associated with a significantly lower risk of CHD than the homozygous wild-type GG (p = 0.0063 and OR = 0.575).
rs12050772, 2:TT; 3:GT; 4:GG; rs2289105, 2:TT; 3:CT; 4:CC; rs4774585, 2:AA; 3:AG; 4:GG.
p < 0.05, there is significance between two genotypes, and the OR value is between 0 and 1, the latter is a protect factors.
The data in Table 5 show that the blood lipid levels and other parameters of echocardiography among individuals with different genotypes did not differ between CHD patients and control participants (p > 0.05). In addition, multiple logistic regression analysis (Table 6) showed that after adjustment for the risk factors of CHD, the associations between rs4774585 SNPs and CHD in the Uygur population were no longer statistically significant.
p > 0.05.
LVDd, left ventricular end-diastolic dimension; LVDs, left ventricular end-systolic dimension; IVS, interventricular septum; PW, posterior wall; PVOT, right ventricular outflow tract; RV, right ventricle; RA, right atrium; PA, pulmonary artery; FS, fractional shortening; EF, ejection fraction; SV, stroke volume; CO, cardiac output.
p > 0.05.
Discussion
We identified a significant association between rs228105 in CYP19 and CHD, and to the best of our knowledge, this is the first investigation of such an association. Our interest in CYP19 (aromatase) in relation to CHD was derived from studies implying that sex hormones may play complex roles in cardiac functions, such as a study suggesting the existence of both estrogen and androgen receptors on endothelial cells and vascular smooth muscle cells (Oparil et al., 1996) and another study proposing that sex hormone ratios influence coronary health (He et al., 2007). The CYP19 gene located on chromosome 15q21.1 codes for a single CYP19 protein known as aromatase. The aromatase activity affects both androgen and estrogen metabolism. Moreover, aromatase is a key enzyme in the conversion of androgen to estrogen and plays an important role in the balance of sex hormone levels in different tissues (Belgorosky et al., 2009; Santen et al., 2009). Aromatase has been found to be produced in the ovary, adipose tissue, bone, and brain (Simpson et al., 2002), and notably, aromatase expression has been observed in vascular cell types such as smooth muscle cells (Harada et al., 1999), endothelial cells (Sasano et al., 1999), and immature heart cells/cardiomyocytes (Price et al., 1992; Grohé et al., 1998).
Human aromatase deficiency was first reported in 1995, and in this condition, the basal concentrations of plasma androgen were elevated, whereas plasma estradiol levels were low (Morishima et al., 1995). These results indicated that a lack of aromatase leads to a disturbance in the balance of sex hormones and also indicates that a single base change in exon 9 of CYP19 may be directly responsible for these changes. Based on several studies that have comprehensively evaluated associations between SNPs in the CYP19 gene and levels of sex hormones (Haiman et al. 2007; Cai et al., 2008; Kidokoro et al. 2009), we believe that significant CYP19 gene polymorphisms may alter hormone levels to varying degrees. In particular, imbalance of the estrogen/androgen ratio, rather than individual levels of estragon or androgen, has been associated with the development of CHD (Dai et al., 2012), and Seruga et al. (2014) reported that the use of aromatase inhibitors might be associated with an increased risk for CHD. Interaction between CYP19 polymorphisms and estrogen-dependent diseases such as polycystic ovary syndrome (PCOS) and osteoporosis also have been reported. In PCOS patients, Zhang et al. (2012a) found that an SNP in CYP19 might inhibit the aromatase activity and be associated with the estradiol/testosterone ratio. Considering these previous study results and the lack of research investigating associations between CYP19 polymorphisms and CHD, we sought to directly determine whether specific CYP19 gene polymorphisms correlate with the risk of CHD. We identified rs12050772 and rs2289105 within the NCBI database because the minor allele frequencies for both were close to 0.5 in the Chinese population. We identified rs4774585 based on a previous cohort study that reported this mutation may be related to the outcomes of cardiovascular disease. Thus, we considered that these mutations are likely to be protection factors in humans, although the relevant literature is lacking.
CHD is an extremely complicated disease, for which certain clinical parameters such as glucose and blood lipid levels are known to differ significantly between patients and health controls. To eliminate the effects of the major confounding factors for CHD, we used propensity score matching techniques to match the case and control groups directly, and the same propensity score indicated the same distribution of measured baseline covariates (Rosenbaum and Rubin, 1983; Frisco et al., 2007). We believe this strengthens the ability of our study to identify potential effects of CYP19 polymorphisms. In the present case-control study, we found that compared to the GG genotype of rs12050772, the GT genotype is associated with a significantly lower risk of CHD. However, unfortunately, the genotypic distribution of rs12050772 was not in Hardy-Weinberg equilibrium in our control group. We confirmed the results of genotyping, and because all 96-well plates included one blank well as a control, we consider the results to be valid and choose to control according to the strict criteria selection. De novo mutations, selection, genetic drift, and gene flow can all theoretically bias the allele and genotype frequencies and thus the Hardy-Weinberg equilibrium. We believe that we can only temporarily ignore the effect of rs12050772 in CHD, and we will continue to explore the relationship between rs12050772 and CHD by increasing the sample size in our control group.
Our results did reveal that the rs2289105 genotypic distributions in CHD patients differed significantly from those in control participants, and the heterozygote CT genotype was associated with a significantly lower risk of CHD than the homozygous wild-type CC genotype. In addition, no significant difference in CHD risk was found between the homozygous mutant and the homozygous wild type. Thus, we propose that the CT genotype of rs2289105 in CYP19 may be a protective genetic marker for CHD. Furthermore, we observed that the rs4774585 genotype distributions did not differ significantly between CHD patients and controls. Initially, we observed an association between the rs4774585 polymorphism and CHD in the Uygur population that was independent of gender and because the sample size was not large enough to use propensity score matching techniques for group matching, we used multiple logistic regression analysis. After adjustment for the risk factors, these associations were no longer statistically significant. This outcome may indicate that the role of the confounding factors is more important compared with the SNP, but it may also simply be the result of our sample size being too small. Thus, we will strengthen the power in future analyses by increasing the sample sizes in the Uygur population groups.
A previous cohort study investigating associations between CYP19 and cardiovascular disease found that the SNP3 G>A variant allele was associated with a 78% increase in mortality in men, and in their hypertensive CHD group, the variant allele was associated with a 65% increase in death, myocardial infarction, or stroke in men and a 69% decrease in these outcomes in women. To summarize, they showed that the rs4774585 polymorphism and outcomes of CHD and hypertension are closely related (Beitelshees et al., 2010). However, in our case-control study, we did not observe a statistically significant relationship between rs4774585 and CHD. One disadvantage in case-control studies is that control individuals may become patients in the future. Although we used propensity score matching to estimate the analogous probability of CHD development in the samples and matched the case and control groups, we believe further cohort studies are necessary. In addition, based on limitations in time and manpower, we did not have access to hormone levels in our study populations. Thus, we cannot further investigate whether our results are associated with imbalances in sex hormone ratios. Interestingly, a study of hypertension in PCOS patients showed that the estrogen-to-androgen ratio was lower among patients with hypertensive PCOS, and although this study did not describe mutations of CYP19, subcutaneous CYP19 mRNA expression was shown to be significantly higher in patients with hypertensive PCOS. Moreover, the study reported that serum estradiol levels in these patients were similar to those in the normotensive PCOS and control groups (Lecke et al., 2011). They speculated that the synthesized estrogens were only partially secreted into the circulation and acted on tissues through intracrine, autocrine, or paracrine mechanisms (Harada et al., 1999; Simpson, 2003; Czajka-Oraniec and Simpson, 2010). These effects are worth further consideration. In our study, ANOVA indicated that blood lipid levels and parameters of echocardiography among individuals with different genotypes did not differ between CHD patients and control participants. These results suggest that aromatase may affect the heart function through mechanisms other than those involving lipid metabolism, and recent studies in animals have shown that testosterone and estrogen have contrasting inotropic actions and modulate Ca(2+) handling and transient characteristics (Bell et al., 2013). We intend to further investigate feasible mechanisms underlying the effects of aromatase in the cardiovascular system through studies in cells and animal models.
In conclusion, this is the first case-control study to examine the associations between rs12050772, rs2289105, and rs4774585 in CYP19 and CHD. The results show that the heterozygote CT genotype of rs2289105 is associated with a reduced risk of CHD and may be a marker for protection from CHD susceptibility. However, further research into the mechanisms by which aromatase affects the cardiovascular system is needed.
Footnotes
Author Disclosure Statement
No competing financial interests exist.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
