Background

Metabolic syndrome (MetS) constitutes a threat to human health, with estimated prevalence rates of 17–40 and 16.3% in the Middle East and North Africa, respectively [1]. Its prevalence has also increased from 23.8% between 2000 and 2005, to 27% between 2010 and 2015 in China [2]. MetS increases the risk of coronary artery disease, stroke, fatty liver disease, diabetes mellitus (DM), insomnia, and several cancers [3, 4]. Adult Treatment Panel III (ATP III) proposed that MetS encompasses at least three of the following ailments: abdominal obesity, high triglyceride (TG) levels, low high-density lipoprotein cholesterol (HDL-C) amounts, high blood pressure (BP), and high fasting blood glucose (FBG) [5]. All MetS variations increase the risk of cardiovascular disease (CVD). CVD incidence and mortality are significantly reduced in premenopausal women compared with men, but this difference gradually diminishes after menopause [6], with a corresponding increase in MetS incidence [7].

Hormones have critical functions in MetS pathogenesis and progression. Estrogen is associated with many aspects of MetS [8] such as glucose tolerance, fat metabolism, and BP [9]. Estrogen exerts its physiological functions in combination with estrogen receptors (ERs). ERs are nuclear hormone receptors (transcription factors) found in macrophages, adipose cells, vascular smooth muscle cells, and vascular endothelial cells [10, 11]. Estrogen receptor-α (ESR1) and β (ESR2) represent two known estrogen receptors that have been assessed using transgenic ERa (ERa_/_) and/or ERb (ERb_/_) knockout mice [12, 13].

The subtypes of ER, ESR1, and ESR2 are distinct in structure and function (likely in carbohydrate and lipid metabolism) [14]. ESR1 and ESR2 are found on chromosomes 6q25.1 and 14q23.2, respectively [15]. Single nucleotide polymorphisms (SNPs) in ESR1, such as PvuII (rs2234693) and XbaI (rs9340799), are involved in MetS [16]. As for ESR2, the risk of MetS is determined by the RsaI (1082A > G) (rs1256049) and AluI (1730A > G) (rs4986938) polymorphisms [17]. ESR1 and ESR2 SNPs are associated with metabolic factors in Western populations [18]. Here we assessed associations of four representative ESR1 and ESR2 polymorphisms with MetS and related phenotypes in postmenopausal Han Chinese women.

Methods

Study population and laboratory measurements

This case control study was performed between January and November 2016 at the medical center of Shanxi Dayi Hospital (Taiyuan, China). Three hundred and four postmenopausal women aged 46–71 years were enrolled. Inclusion criteria were: ≥60 years old; permanent menstrual period discontinuation because of natural menopause.

At baseline, the subjects were assigned to 2 groups, including the control (154 healthy individuals; mean age of 63.0 ± 4.1 years) and MetS (150 MetS patients; mean age of 62.9 ± 4.8 years) groups. All subjects were Han Chinese and of the same ethnic group. All cases were diagnosed according to the Chinese type 2 diabetes Prevention Guide (2013) diagnostic guidelines [19]. MetS was reflected by ≥3 of the following parameters: waist circumference ≥ 85 cm; FBG ≥6.1 mM or 2 h post glucose–load blood glucose ≥7.8 mM or known diabetes; serum TG ≥1.70 mM; HDL-C < 1.04 mM; and BP ≥130/85 mmHg or treated hypertension. Non-alcoholic fatty liver disease (NAFLD) diagnosis was performed by abdominal ultrasonography, as described previously [20]. Exclusion criteria were: severe psychological ailments, physical disability, any type of malignancy, or history of smoking, heart failure, myocardial infarction (MI), type I diabetes, kidney failure, and chronic liver disease. The subjects were not undergoing hormone replacement.

Data collection

The main measurements and clinical data of the participants, including age, height, weight, body mass index (BMI), waist circumference, and systolic and diastolic BPs, were measured and recorded by specifically trained individuals. Waist circumference measurement was carried out in the horizontal plane at the midpoint between the lower border of ribs and the iliac crest. Systolic and diastolic BPs were obtained in right upper arm twice following a 15 min-rest with the patient sitting. Measurements were performed at 5 min intervals and averaged.

FBG, hemoglobin A1C (HbA1C), TG, total cholesterol (TC), HDL-C, low density lipoprotein (LDL-C; direct measurement), estradiol (E2), and fasting serum insulin levels in all participants were measured from fasting blood samples on an AutoAnalyzer (AU5400, Olympus, Tokyo, Japan), with reagents and calibrators provided by the manufacturer. Abdominal ultrasonography was conducted by specialized radiologists based on unified NAFLD diagnostic criteria. Insulin resistance (homeostasis model assessment of insulin resistance, HOMA-IR) was determined for each subject.

This study followed the Declaration of Helsinki, as well as institutional regulations regarding human subjects. Written informed consent was provided by all participants. The study protocol had approval from the ethics committee of the Shanxi Da Yi Hospital Institute for Biological Sciences, Chinese Academy of Sciences.

Genomic DNA extraction and genotyping

Venous blood (2 ml) specimens were obtained in EDTA anticoagulation tubes and used for DNA extraction. Genomic DNA extraction from 200 μL whole blood was carried out with the TIANamp Blood DNA Kit (Tiangen Biotech, China), as instructed by the manufacturer.

The PvuII (c454-397 T > C) and XbaI (c454-351A > G) polymorphisms in ESR1 as well as the RsaI (1082A > G) and AluI (1730A > G) polymorphisms in ESR2 were assessed by polymerase chain reaction- restriction fragment length polymorphism (PCR-RFLP). A 1374 base pair (bp) fragment was amplified for the PvuII (c454-397 T > C) and XbaI (c454-351A > G) polymorphisms using the following primer pair: upstream, 5′-CTGCCACCCTATCTGTATCTTTTCCTATTCTCC-3′; downstream, 5′-TCTTTCTCTGCCACCCTGGCGTCGATTATCTGA-3′. A 476-bp fragment for the RsaI (1082A > G) polymorphism was amplified using the following primer pair: upstream, 5′-TTCTGAGCCGAGGTCGTAGT-3′; downstream, 5′-CCAGGACTTTGTTCCCACTC-3′. A 391-bp fragment for the AluI (1730A > G) polymorphism was amplified with the following primer pair: upstream, 5′-GCACCTTTTTGTCCCCATAGTAAC-3′; downstream, 5′-CTATGGCTTCCTCACACCGAC-3′. PCR was carried out on an ABI Prism 7500 sequence system (Applied Biosystems, USA) in 25 μl reactions, containing 5 μl genomic DNA, 10 μl of Taq PCR Master mix (2×), 2 μl of each primer, and 8 μl PCR buffer (Tiangen Biotech, China), as follows: 94 °C (4 min); 35 cycles of 94 °C (30 s), 57 °C (40 s) and 72 °C (40 s); 72 °C (7 min). Digestion of PCR products was performed overnight using XbaI (expected fragments of 1374, 393, and 981 bp), PvuII (expected fragments of 1374, 438, and 936 bp), RsaI (expected fragments of 476, 183, and 293 bp), and AluI (expected fragments of 391, 144, and 247 bp) (Biolabs, New England) at 37 °C. All the digested fragments were electrophoresed on 2% agarose gels containing Goldview. The genotypes of 10% specimens were verified by direct sequencing.

Statistical analysis

SPSS 20.0 was employed for statistical analyses. Normally distributed continuous data are mean ± standard deviation (SD). Genotype differences were assessed in MetS patients by analysis of variance. Baseline features between the MetS and control groups were evaluated by Student t-test. Comparisons among multiple groups were performed by ANOVA with Bonferroni correction. Hardy–Weinberg equilibrium test, allele genotype frequency assessment, and haplotype evaluation were carried out online with the SHEsis platform. Genotype frequency, allele frequency, and haplotype were analyzed by the chi-square test and odds ratio (OR) assessment. Multivariable logistic regression was performed to adjust ORs for possible confounders. p < 0.05 indicated statistical significance. The function pwr.chisq.test() in the R package was used to calculate statistical power.

Results

Demographic and clinical variables

The MetS and control groups included postmenopausal women, with similar ages in both groups. All clinical parameters in Table 1 showed statistically significant differences between the MetS and control groups. Elevated BMI values and waist circumferences were observed in MetS patients, who also showed higher mean systolic and diastolic BPs compared with controls. In addition, the MetS group had markedly higher FBG, hemoglobin A1c, fasting serum insulin and HOMA-IR levels, and worse lipid profile. Furthermore, prevalence of nonalcoholic fatty liver disease was far higher in MetS patients compared with controls.

Table 1 Patient clinico-biochemical features
Full size table

Allele frequencies and genotype distribution

Genotype distribution for each of the four SNPs was compatible with the Hardy–Weinberg equilibrium (p > 0.05) in MetS and control patients. The Chi-square test of XbaI genotype’s frequencies in the case and control groups can be used to calculate statistical power. The test power was calculated (power = 1), indicating that the allele frequencies and genotype distributions of MetS patients and controls (Table 2) led to credible conclusions.

Table 2 Allele frequencies and genotype distributions of ESR1 and ESR2 gene polymorphisms in the metabolic syndrome (MetS) and control groups
Full size table

G allele frequency in the XbaI polymorphism was markedly elevated in MetS patients compared with control individuals (p = 0.004; OR = 1.610, 95%CI 1.169–2.18).

Genotype distributions of the XbaI, PvuII, RsaI and AluI polymorphisms, and allele frequencies of the PvuII, RsaI and AluI polymorphisms were comparable between MetS patients and controls (p > 0.05).

Haplotype distributions of the four SNPs

LD assessment of all SNP pairs revealed XbaI and PvuII of ESR1 were associated,as well as RsaI (B5) and AluI (B8) of ESR2 (Additional file 1: Table S1). Haplotype distributions of the 4 SNPs (PvuII, XbaI, RsaI and AluI) are shown in Table 3. The exclusion criteria was set as frequency < 0.03. ESR1 and ESR2 haplotype frequency distributions were greater than 3%. Analysis showed that XbaI–PvuII based haplotypes were associated with MetS (global p = 0.022), while those constructed from RsaI–AluI had no significant associations with MetS (p > 0.05). The haplotypes A–T (p = 0.015; OR = 0.638, 95%CI 0.443–0.919) and G–C (p = 0.024; OR = 1.512, 95%CI 1.055–2.168) obtained from XbaI–PvuII showed significant differences.

Table 3 Estimated haplotype distributions in the metabolic syndrome (MetS) and control groups
Full size table

Associations of genotypes with clinical characteristics

Associations of the XbaI polymorphism with MetS components are shown in Table 4. Patients harboring the G allele in both homozygous and heterozygous forms (i.e. GG or AG) had elevated BMI, waist circumference, systolic and diastolic BPs, FBG, HOMA-IR, TC, TG, LDL-C and NAFLD (%), and lower HDL-C. Among subjects with the PvuII polymorphism, those harboring the minor C allele in both homozygous and heterozygous forms (i.e. CC or TC) showed elevated BMI, fasting serum insulin and HOMA-IR (Additional file 1: Table S5). Associations of XbaI, PvuII and MetS were confirmed by multivariable regression analysis (Table 5). Among subjects with the RsaI polymorphism, those harboring the minor G allele in both homozygous and heterozygous forms showed elevated TG, LDL-C, fasting serum insulin and HOMA-IR (Additional file 1: Table S3). Among subjects with the AluI polymorphism, those carrying the minor G allele in both homozygous and heterozygous forms (AG or GG) showed elevated TC and LDL-C (Additional file 1: Table S4).

Table 4 Associations of XbaI genotypes with various clinical characteristics
Full size table
Table 5 Multivariable logistic regression analysis with MetS as the dependent variable
Full size table

Regarding XbaI, the control group showed significantly different FBG, TG, TC, HDL-C and LDL-C among the three genotypes (all p < 0.05). In MetS patients, BMI, TG, TC and LDL-C were markedly different (Additional file 1: Table S2). As for RsaI (B5), the control group showed significant differences in fasting insulin and HOMA-IR among the three genotypes (all p < 0.05); in MetS patients, LDL-C and HOMA-IR were significantly different (Additional file 1: Table S3). As for AluI (B8), BMI and TC significantly differed among the three genotypes in control patients (all p < 0.05); in MetS patients, age and TG significantly differed (Additional file 1: Table S4). Regarding PvuII, estradiol levels were significantly different among the three genotypes in control patients (all p < 0.05); in MetS patients, age, BMI, systolic BP, HDL-C and HOMA-IR were significantly different (Additional file 1: Table S5).

Discussion

Changing lifestyles, socioeconomic status, and dietary habits have likely contributed to a rapid increase in the incidence of MetS, which is currently considered a major threat to human health worldwide [21]. High MetS prevalence is found among postmenopausal women [22]. Associations of several polymorphisms in the ESR1 and ESR2 genes with MetS or related components have been described [23]. These associations were also reported in the Study of Women’s Health Across the Nation (SWAN) trial of perimenopausal African-American, Caucasian, Chinese, and Japanese women between 42 and 52 years old [24]. The current study revealed positive associations of ESR1 polymorphisms with MetS in postmenopausal Chinese Han women. Subjects harboring the G allele of XbaI, or the A-T and G-C haplotypes of XbaI–PvuII polymorphisms of ESR1, showed elevated risk of MetS. Similar results were obtained in studies of women of different ages [25, 26]. Two ESR2 subtypes, including RsaI and AluI, were not significantly associated with Mets in this study.

ERb_/_ mice show normal or improved glucose tolerance and insulin release compared with wild type counterparts [27]. Meanwhile, ERa_/_ animals display glucose intolerance and insulin resistance [12]. ESR1 and ESR2 are widely distributed in various organs and tissues related to sugar metabolism. Associations of ESR1 polymorphisms and (T2DM) were demonstrated [28]. ESR2 polymorphisms are associated with MetS in Chinese and Japanese populations; in addition, SNPs in the ESR1 show associations with insulin sensitivity in Asian women [24]. ER-α can exert co-modulatory functions with insulin receptor substrate 1 (IRS1) to modulate insulin signaling [29]. Previous studies suggested an inhibitory effect of ERb on glucose transporter four (GLUT4) expression, and revealed that unchecked ERb activity could result in diabetes [25, 30]. As shown above, patients harboring the minor alleles of XbaI and PvuII gene polymorphisms, in both homozygous (GG and CC) or heterozygous (AG and TC) forms, had higher FBG, fasting serum insulin and HOMA-IR. Meanwhile, those carrying the RsaI polymorphism of ESR2 (homozygous or heterozygous) had lower fasting serum insulin and HOMA-IR.

Estrogens have critical functions in lipoprotein metabolism. An imbalance between ESR1 and ESR2 in the adipose tissue could therefore affect the development of metabolic diseases. Epidemiological studies showed hyperlipidemia is a significant risk factor for atherosclerosis, stroke, CAD, and MI [31]. In mice, ERb absence inhibit triglyceride accumulation, maintains normal insulin signaling in the liver and skeletal muscle, and ameliorates whole body insulin sensitivity as well as glucose tolerance [32]. ESR1 polymorphisms are associated with LDL particle size, reduced LDL cholesterol, and TC in adolescent female individuals [33]. SNPs in ESR1 show associations with the degree of adiposity, BMI, and waist circumference [34]. The PvuII T allele of ESR1 and associated genotypes and haplotypes are associated with risk of hyperlipidemia in postmenopausal Chinese Han females [35]. In addition, women with the PvuII T allele show elevated amounts of small LDL particles [36]. Furthermore, postmenopausal women with the T allele have reduced HDL-C and elevated TG, and are more vulnerable to lipid metabolism disorders [37]. These findings were corroborated by several other reports showing associations of ESR1 polymorphisms with BMI; however, associations of ESR2 polymorphisms with hyperlipidemia remain unclear. A study showed that homozygotes for the AluI (rs4986938) of ESR2 have higher BMI, serum TG and apolipoprotein B, and reduced HDL-C [38]. In mice, ESR2 deficiency is associated with abnormal vascular function and hypertension [39]. ESR2 polymorphisms are also associated with hypertension in postmenopausal Japanese women [40]. As shown above, positive associations were obtained of the G allele of XbaI polymorphism of the ESR1 gene with elevated TG, TC, LDL-C levels, as well as lower HDL-C, in postmenopausal Chinese women. The RsaI and AluI polymorphisms of ESR2 exhibited associations with TG, TC and LDL-C amounts.

Conclusion

Currently, MetS incidence is higher in postmenopausal women compared with age—matched males. Interventions to halt the rapidly increasing MetS incidence are of great clinical interest. Several studies have assessed associations of ESR1 and ESR2 with MetS or related components in different populations, but have often yielded conflicting findings. Furthermore, very few such studies have evaluated Chinese subjects. We here demonstrated associations of ESR1 and ESR2 polymorphisms with MetS and related components, especially obesity, and lipid and glucose metabolism, in postmenopausal Chinese Han females. Meanwhile, the G allele frequency of the XbaI polymorphism was markedly elevated in MetS patients compared with the control group. The RsaI and AluI polymorphisms of the ESR2 gene were not associated with MetS. In addition, XbaI and PvuII polymorphisms of ESR1 were associated with some MetS components. The current findings could include biases related to sample size, age, and ethnicity. Nevertheless, the present data could provide valuable insights into MetS epidemiology, particularly in Han Chinese individuals, providing a basis for the development of novel therapeutic interventions.