Associations between ERα/β gene polymorphisms and osteoporosis susceptibility and bone mineral density in postmenopausal women: a systematic review and meta-analysis

Background Many studies have reported associations between estrogen receptor (ER) gene polymorphisms and postmenopausal osteoporosis (PMOP) risk and bone mineral density (BMD), but the results are controversial. The aim of the present meta-analysis is to verify the association between ERα and ERβ gene polymorphisms and osteoporosis susceptibility and BMD in postmenopausal women. Methods PubMed, EMBASE, Web of Science, the Cochrane Library and China WeiPu Library were searched. OR and WMD with 95% CI were calculated to assess the association. Results Overall, no significant association was observed between ERα XbaI, ERα PvuII and PMOP susceptibility in either overall, Caucasian or Asian populations. ERα G2014A was significantly associated with a decreased risk of PMOP in Caucasian populations. There was a significant association between ERβ RsaI and PMOP risk in both overall and Asian populations. Caucasian PMOP women with ERα XbaI XX and Xx genotypes had a higher LS Z value than women with xx genotype. ERα XbaI XX genotype was associated with increased FN BMD in overall and Caucasian populations, an increased FN Z value in Asians, and a decreased FN Z value in Caucasians. There was also a significant association between ERα XbaI Xx genotype and an increased FN Z value in either Asians or Caucasians. ERα PvuII PP genotype was associated with a low LS Z value in Caucasians and a low FN BMD and Z value in Asians. Pp genotype in PMOP women was significantly correlated with low LS BMD in overall populations, a low FN Z value in either overall, Caucasian or Asian populations. Conclusion Each ERα and ERβ gene polymorphism might have different impact on PMOP risk and BMD in various ethnicities.


Background
Postmenopausal osteoporosis (PMOP) is a common metabolic bone disorder characterized by low bone mineral density (BMD) and increased fracture risks [1][2][3]. It is estimated that osteoporosis affects approximately 10 million American adults, with another 34 million being at high risk due to low bone mass [4].
The pathophysiology of PMOP is considered as a disorder or negative imbalance of bone metabolism and remodeling, with bone resorption outpacing bone formation [3], suggesting that vitamin D and parathyroid hormone (PTH) and other factors related to bone resorption and formation may play a key role in the underlying mechanism and pathophysiology of PMOP [5][6][7][8]. Furthermore, genetic factors including genes and gene polymorphisms may also play an important role in the development of PMOP [9].
Estrogen is another important hormone that plays an important role in the pathogenesis of PMOP, knowing that reduced ovarian production of estrogen after menopause is a cause for the initial phase of rapid bone loss and osteoporosis in women [3]. Estrogen is known as an important regulator of bone metabolism, and estrogen deficiency is believed to be the cause of BMD loss, increased mechanical loadinginduced bone remodeling, and the development of PMOP [10]. Knowing that the action of estrogen is predominantly mediated by estrogen receptor (ER), including ERα and ERβ by binding to different ligands to mediate various biological effects [3,10], more attention has been paid to the relationship between ERs and PMOP risk and BMD in postmenopausal women . However, the results of studies currently available about this issue are controversial.
Previous meta-analyses have been performed to assess the pooled effects of ER gene polymorphisms on BMD and fracture risk [39][40][41]. WANG et al. [39] showed that the ERα XbaI (rs9340799) polymorphism was associated with BMD at diverse skeletal sites, and ERα PvuII (rs2234693) PP genotype played a role in protecting the lumbar spine but on the other hand might be a risk factor for the femoral neck fracture. However, to the best of our knowledge, no meta-analysis has been performed to explore the relationships between ER gene [ERα XbaI (rs9340799), ERα PvuII (rs2234693) and ERα G2014A (rs2228480)] and ERβ gene [ERβ AluI (rs4986938) and ERβ RsaI (rs1256049)] polymorphisms and PMOP susceptibility and BMD of the lumbar spine and femoral neck in postmenopausal women. To address these issues, we performed a meta-analysis of all currently available studies relating ER gene [ERα XbaI (rs9340799), ERα PvuII (rs2234693) and ERα G2014A (rs2228480)] and ERβ gene [ERβ AluI (rs4986938) and ERβ RsaI (rs1256049)] polymorphisms with PMOP risk and BMD.

Data sources and searches
We searched PubMed, EMBASE, Web of Science, the Cochrane Library and China WeiPu Library to identify case-control studies that investigated the associations between ERα gene polymorphisms [ERα XbaI (rs9340799), ERα PvuII (rs2234693) and ERα G2014A (rs2228480)] ERβ gene polymorphisms [ERβ AluI (rs4986938) and ERβ RsaI (rs1256049)] and osteoporosis susceptibility and BMD in postmenopausal women by using the following search terms ('PMOP' OR 'Postmenopausal osteoporosis' OR 'Postmenopausal') AND ('Estrogen Receptor' OR 'ER') AND ('polymorphism' OR 'single nucleotide polymorphism' OR 'SNP' OR 'variation'). To analyze the pooled effects of ER gene polymorphisms on BMD, the following search terms were used: ('PMOP' OR 'Postmenopausal osteoporosis' OR 'Postmenopausal') AND ('Estrogen Receptor' OR 'ER') AND ('polymorphism' OR 'single nucleotide polymorphism' OR 'SNP' OR 'variation') AND ('BMD' OR 'bone mineral density'). Then, one-by-one screening was performed by two authors according to the inclusion and exclusion criteria. No language restrictions were applied. Secondary searches of eligible studies were conducted by searching the reference lists of the selected studies, reviews or comments.

Inclusion and exclusion criteria
The inclusion criteria of our meta-analysis are as follows: (1) case-control studies; (2) studies on BMD and fracture risks in postmenopausal women with PMOP due to estrogen deficiency using postmenopausal women without PMOP or healthy volunteers as control; (3) studies reporting alleles and genotypes of at least one of the ER gene polymorphisms in women with or without PMOP: ERα XbaI (rs9340799), ERα PvuII (rs2234693), ERα G2014A (rs2228480), ERβ AluI (rs4986938) and ERβ RsaI (rs1256049); (3) studies reporting the sample size, mean and standard deviation (SD) of BMD (g/cm 2 ) or BMD Z value in PMOP women with at least one of the ER genotypes; and (4) studies with sufficient data. The exclusion criteria were: (1) reviews or case reports without controls, and (2) studies with no availability of current data; and (3) duplicated reports.

Data extraction
Data from the eligible studies were extracted according to the inclusion and exclusion criteria by two authors, and a consensus was reached by discussion. In the study of associations between ER gene polymorphisms and PMOP risk, the following data were collected: author list, year of publication, ethnicity, sample size, alleles, genotype of each gene polymorphism and Hardy-Weinberg equilibrium (HWE). The following data were collected for analysis of differences in BMD in PMOP women with various ER genotypes: author list, year of publication, ethnicity, the number of cases and mean and SD of BMD (g/cm 2 ) and BMD Z value.

Data synthesis and statistical analysis
We calculated odds ratios (OR) and 95% confidence interval (CI) to evaluate the association between ER gene polymorphisms and PMOP risk (osteoporosis occurred in postmenopausal women due to estrogen deficiency as represented by low BMD and increased fracture risks). The strength of association between ER gene polymorphisms and PMOP susceptibility was evaluated by OR and 95% CI under the allele contrast model, heterozygote model, homozygote model, dominant model and recessive model. HWE was calculated in the control population to evaluate the quality of the data by using chisquare test. Regarding the associations between BMD and ER gene polymorphisms, we compared BMD (g/ cm 2 ) and BMD Z value in PMOP women under the heterozygote and homozygote model respectively using the weight mean difference (WMD) and 95% CI. Heterogeneity of the included studies was examined by a chisquared-based Q statistical test and quantified by I2 metric value. If I2 value was > 50% or P < 0.10, ORs and WMD were pooled by the random effect model; otherwise, the fixed effect model was used. Power analysis was performed using the Power and Precision V4 software (Biostat Inc., Englewood, USA). Sensitivity analysis was performed to assess the impact of each study on the combined effect of the present meta-analysis. Besides, subgroup analysis was also performed according to the ethnicity of the study populations. Stata 12.0 software (StataCorp, College Station, TX, USA) was used and a P < 0.05 was considered as statistically significant.

Power analysis
Before initiation of the meta-analysis, a power analysis was conducted by using the Power and Precision V4 software to verify whether the included studies could offer adequate power (> 80%). The result showed that the statistical power in our study was sufficient to detect the associations between ER gene polymorphisms and PMOP risk.

Associations between ER gene polymorphisms and PMOP risk
Overall, we did not find any significant association between ERα XbaI and ERα PvuII polymorphisms and risk of PMOP in either overall, Caucasian or Asian populations (all P > 0.05) ( Table 4). ERα G2014A polymorphism played a protcetive role in developing PMOP in Caucasian populations, while no significant association was observed in overall and Asian populations (both P > 0.05). All the data are shown in Table 4 and Fig. 2.
With regard to ERβ polymorphism, ERβ AluI was significantly associated with the risk of developing PMOP in Asian postmenopausal women under the recessive model; however, we did not observe any significant association between ERβ AluI and PMOP risk in overall and Caucasian populations (both P > 0.05) ( Table 4 and Fig. 3). Furthermore, we also found that there was a remarkable association between ERβ RsaI polymorphism and decreased PMOP risk in overall and Asian populations (Table 4).

Associations between ER gene polymorphisms and BMD in PMOP women
ERα XbaI and lumbar spine bone mineral density (BMD g/ cm 2 and BMD Z value) In our meta-analysis, no significant difference in lumbar spine BMD (g/cm 2 ) was observed between PMOP women with ERα XbaI XX, ERα XbaI Xx and ERα XbaI xx genotype in either overall, Caucasian or Asian populations (all P > 0.05) ( Table 5). The lumbar spine BMD Z value in Caucasian PMOP women carrying ERα XbaI XX genotype was greater than that in those carrying xx genotype, while no significant difference was observed in overall and Asian populations (both P > 0.05). ERα XbaI Xx genotype was found to be significantly associated with high lumbar spine BMD Z value in either overall or Caucasian populations but not in Asian populations.
ERα XbaI and femoral neck bone mineral density (BMD g/ cm 2

and BMD Z value)
Our pooled analyses indicated that the ERα XbaI XX genotype was significantly associated with increased femoral neck BMD in overall and Caucasian populations. In contrast, ERα XbaI XX genotype did not play a key role in femoral neck BMD in Asian populations (Table 5 and Fig. 4). Interestingly, compared with PMOP women with xx genotype, XX genotype was significantly associated with decreased femoral neck Z value in Caucasians, and increased femoral neck Z value in Asians (Table 5). However, no significant association was observed between XX genotype and the femoral neck Z value in overall populations. In addition, Caucasians and Asians carrying the ERα XbaI Xx genotype were at risk of a high femoral neck Z value, while no significant association was found in overall populations. We did not observe remarkable relationships between ERα XbaI Xx genotype and femoral neck BMD in either overall, Caucasian or Asian populations (all P > 0.05). All data are shown in Table 5.
ERα PvuII and lumbar spine bone mineral density (BMD g/ cm 2

and BMD Z value)
With regard to ERα PvuII, the difference in the lumbar spine Z value between the PP and pp. genotypes was − 0.07 (95% CI = − 0.03 to − 0.01, P = 0.031) in Caucasian PMOP women; however, no significant difference was observed in overall and Asian populations. For the Pp versus pp. genotype, the difference in lumbar spine BMD was − 0.01 (95% CI = − 0.02 to − 0.00, P = 0.036) in overall populations, and the difference in the lumbar spine Z value was − 0.16 (95% CI = − 0.20 to − 0.12, P < 0.001) in Caucasian populations; however, we did not find any significant difference in lumbar spine BMD in either Caucasians or Asians, and in the lumbar spine Z value in overall and Asian populations (Table 5 and Fig. 5). In addition, no significant difference in lumbar spine BMD was observed between PP and pp. genotypes (P > 0.05) ( Table 5).
ERα PvuII and femoral neck bone mineral density (BMD g/ cm 2 and BMD Z value) We further found that the ERα PvuII PP genotype was associated with decreased femoral neck BMD and Z value compared with the pp. genotype in Asians, while no significant difference in femoral neck BMD and Z value was observed in either overall and Caucasian populations (both P > 0.05) ( Table 5). Furthrmore, PMOP women carrying the Pp genotype were at risk of a low femoral neck Z value, which was found in overall, Caucasian and Asian populations. Our study showed that there was no significant difference in femoral neck BMD between PMOP women with the Pp genotype and those with the pp. genotype (P > 0.05). All the data are shown in Table 5.

Sensitivity analysis and publication bias
We performed a leave-one-out analysis to estimate the sensitivity of our study and found that omission of any single study did not affect the overall statistical significance, indicating that the results of our metaanalysis are stable. Therefore, we could conclude that our meta-analysis data are relatively stable and credible. To estimate the publication bias of our metaanalysis, the Begg's and Egger's test was performed ( Table 4), indicating that there was minimal evidence of publication bias. The shape of funnel plot was symmetrical, which also showed no publication bias in our study (Fig. 6).

Associations between ERα gene polymorphisms and PMOP risk
ERα XbaI and ERα PvuII are the two restriction fragment length polymorphisms of ERα gene located in Intron 1 [14]. Many studies [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25]32] have been performed to explore the relationships between ERα XbaI, ERα PvuII and PMOP risk; however, these studies have yielded inconsistent data [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25]32]. Overall, we did not observe any significant association between ERα XbaI and ERα PvuII polymorphisms and PMOP risk in either overall, Caucasian or Asian populations. In our opinion, the inadequate sample size, different ethnicities, various genotyping techniques, the presence of admixture in the population, gene-environment interactions, differences in age and measurement errors of different investigators might be important factors contributing to these controversial results. ERα XbaI and ERα PvuII have proven to play key roles in attainment and maintenance of peek bone mass during young adulthood, and it might be difficult to document their effects in a population of postmenopausal women [24]. In addition, PvuII and XbaI polymorphisms are located in a non-functional area of the ER gene [20], which might also contribute to our polled results. With regard to ERα G2014A, it is located on the exon region of chromosome 6p25.1, and may contribute via the epigenetic level for the efficiency of translation or receptor protein expression [26]. Our results showed that a significant association between ERα G2014A and PMOP risk was observed only in Caucasian populations but not in overall and Asian populations.

Associations between ERβ gene polymorphisms and PMOP risk
ERβ has been found to be more abundant than ERα in trabecular bone, and more potent than ERα in     mediating estrogen-induced repression of TNF-α expression, which is considered an important contributor to PMOP [30]. ERβ AluI is one of the widelystudied ERβ gene polymorphisms, knowing that it could alter mRNA stability and protein levels, leading to reduced synthesis of ERβ [30]. In our study, ERβ AluI was found to be significantly associated with increased risk of PMOP in Asian populations, while no significant relationship was observed in overall and Caucasian populations. Thus, different genetic backgrounds, environmental effects and/or their internal interactions could explain the diverse results in various ethnicities. ERβ RsaI is another important polymorphism of ERβ. Our subgroup analysis revealed a significant association between ERβ RsaI and PMOP risk in overall populations, which is consistent with the studies of Shoukry et al. [30], and Huang et al. [31].

Associations between ERα XbaI and lumbar spine and femoral neck BMD
Our pooled results showed that there was no significant difference in lumbar spine BMD between PMOP women carrying XX, Xx and xx genotype in either overall, Caucasian or Asian populations. However, WANG et al. [39] reported that the XbaI  polymorphism was significantly associated with BMD of the lumbar spine, and XX had a protective effect in comparison with carriers of the x alleles, which is consistent with the report of Ioannidis et al. [41]. Both WANG and Ioannidis included all types of osteoporotic patients, not only postmenopausal women, which might be the most important reason for the difference between our results and theirs. As mentioned above, ERα XbaI might not play a key role in attainment and maintenance of peek bone mass in postmenopausal women [24], and therefore it could be easily understood why no significant association was observed between ERα XbaI and lumbar spine BMD. With regard to femoral neck BMD, our study indicated that the femoral neck BMD in PMOP women with XX genotype was significantly higher than that in women with xx genotype in overall and Caucasian populations, which highlights the theory that ERα gene is involved in the pathogenesis of PMOP. No significant difference of femoral neck BMD was observed between PMOP women with Xx and xx genotype in each subgroup. Although no significant association was observed between lumbar spine BMD and ERα XbaI, we found that the lumbar spine Z value in both PMOP women carrying XX and those carrying Xx genotype was significantly higher than that in Caucasians carrying xx genotype. We also observed that XX genotype was associated with a low femoral neck Z value in Caucasians and high femoral neck Z value in Asians. In addition, Caucasians and Asians carrying Xx genotype were at risk of a high femoral neck Z value. However, why ERα XbaI plays a contradictory role in BMD and Z value at the lumar spine and femoral neck, and the mechanisms by which it is associated with BMD and Z value remains unclear and needs further investigation.

Associations between ERα PvuII and lumbar spine and femoral neck BMD
Although the molecular mechanism underlying the effect of ERα PvuII on bone mass is poorly understood, it is believed that ERα PvuII might play a key role in BMD as it is in linkage disequilibrium with the TA polymorphism in the ER promoter that is associated with altered gene transcription [20]. Our pooled analysis indicated that PMOP women with the Pp genotype had lower lumbar spine BMD than those with the pp. genotype. We also found that there was no significant difference in lumbar spine BMD between women with the PP genotype and those with the pp. genotype, which is consistent with the meta-analysis of Wang et al. [40]. Furthermore, we observed that the PP genotype was associated with decreased femoral neck BMD in Asians, while Pp might not play a key role in femoral neck BMD in all subgroups. Interestingly, WANG et al. [39] reported that PP play a role in protecting the lumbar spine but on the other hand it might be a risk factor for the femoral neck fracture. Wang CL [40] and WANG KJ [39] conducted their meta-analyses on osteoporotic women during menopause while our study included osteoporotic women post menopause, which might be the most important reason for the difference between our study and theirs. In addition, both PP and Pp genotypes were significantly associated with low lumbar spine Z value in Caucasians,

Limitations
Although we performed a comprehensive analysis of the association between ERα, ERβ gene polymorphisms and PMOP risk and BMD in postmenopausal women, there are some limitations that should be addressed. First, high heterogeneity was observed in some of our pooled results, which might have negative impact on our conclusions. Second, PMOP is a disease whose etiology might be involved in several confounding factors, and other confounding factors such as age, years since menopause and estrogen therapy might interact with each other and play a key role in the etiology and progression of PMOP. However, no data available could be used in all recruited studies to detect the interactions between these confounding factors in PMOP patients. We should take all these confounding factors into consideration in our study rather than studying them separately, which is also a limitation of our meta-analysis. Third, we failed to perform a pooled analysis to detect whether ERα G2014A, ERβ AluI and ERβ RsaI were correlated with BMD in postmenopausal women as no sufficient data could be collected and analyzed. Therefore, largerscale and better-designed studies are necessary to determine the association between ERα/β gene polymorphisms and PMOP risk and BMD in postmenopausal women. Fig. 4 Forest plot showed that XX genotype of ERα XbaI was associated with increased femoral neck BMD compared with xx genotype Fig. 5 Forest plot showed that Pp genotype of ERα PvuII was associated with increased lumbar spine BMD compared with pp. genotype