Association of Polymorphisms in FSHR, INHA, ESR1, and BMP15 with Recurrent Implantation Failure

Recurrent implantation failure (RIF) refers to two or more unsuccessful in vitro fertilization embryo transfers in the same individual. Embryonic characteristics, immunological factors, and coagulation factors are known to be the causes of RIF. Genetic factors have also been reported to be involved in the occurrence of RIF, and some single nucleotide polymorphisms (SNPs) may contribute to RIF. We examined SNPs in FSHR, INHA, ESR1, and BMP15, which have been associated with primary ovarian failure. A cohort of 133 RIF patients and 317 healthy controls consisting of all Korean women was included. Genotyping was performed by Taq-Man genotyping assays to determine the frequency of the following polymorphisms: FSHR rs6165, INHA rs11893842 and rs35118453, ESR1 rs9340799 and rs2234693, and BMP15 rs17003221 and rs3810682. The differences in these SNPs were compared between the patient and control groups. Our results demonstrate a decreased prevalence of RIF in subjects with the FSHR rs6165 A>G polymorphism [AA vs. AG adjusted odds ratio (AOR) = 0.432; confidence interval (CI) = 0.206–0.908; p = 0.027, AA+AG vs. GG AOR = 0.434; CI = 0.213–0.885; p = 0.022]. Based on a genotype combination analysis, the GG/AA (FSHR rs6165/ESR1 rs9340799: OR = 0.250; CI = 0.072–0.874; p = 0.030) and GG-CC (FSHR rs6165/BMP15 rs3810682: OR = 0.466; CI = 0.220–0.987; p = 0.046) alleles were also associated with a decreased RIF risk. Additionally, the FSHR rs6165GG and BMP15 rs17003221TT+TC genotype combination was associated with a decreased RIF risk (OR = 0.430; CI = 0.210–0.877; p = 0.020) and increased FSH levels, as assessed by an analysis of variance. The FSHR rs6165 polymorphism and genotype combinations are significantly associated with RIF development in Korean women.


Introduction
Implantation is a process by which an embryo attaches to the lumen surface of the endometrium and then migrates into the deep layer of the endometrium [1]. Successful embryo implantation requires success at each event, such as sperm and oocyte quality, development of the early embryo and endometrium, and the interaction between the blastocyst and endometrium [2,3]. Recurrent implantation failure (RIF) is defined as repeated embryo implantation failure after four high-quality embryo transfers in women. Despite advances in in vitro fertilization (IVF) technology, success rates remain stable, and approximately 10% of women undergoing IVF treatment suffer from RIF [4]. Implantation and hormone levels may be altered by certain polymorphisms [42,43]. Therefore, four genes (FSHR, INHA, ESR1, and BMP15) related to hormones were selected, and polymorphisms located in gene regulatory or coding regions were selected. Finally, a total of seven mutations were selected: coding region (FSHR rs6165, ESR1 rs2234693, and BMP15 rs17003221), promoter region (INHA rs11893842, rs35118453, and ESR1 rs9340799), and 5 UTR region (BMP15 rs3810682). The occurrence of FSHR, INHA, ESR1, and BMP15 gene polymorphisms has also been reported to be associated with various reproductive diseases, including primary ovarian insufficiency (POI), pregnancy loss, and preeclampsia [44][45][46][47]. However, few studies have examined the associations between RIF and FSHR, INHA, ESR1, and BMP15 gene polymorphisms. In this study, we investigated whether FSHR, INHA, ESR1, and BMP15 polymorphisms are associated with RIF and whether these polymorphisms affect the levels of clinical factors in Korean women. To reveal the relationship between RIF and FSHR, INHA, ESR1, and BMP15 gene polymorphisms, we assessed the differences between RIF patients and healthy controls by examining the known FSHR (rs6165), INHA (rs11893842, rs35118453), ESR1 (rs9340799, rs2234693), and BMP15 (rs17003221, rs3810682) gene polymorphisms.

Study Population
Blood samples were obtained from women with RIF treated at the Department of Obstetrics and Gynecology and the Fertility Center of CHA Bundang Medical Center in Seongnam, South Korea, between March 2010 and December 2022. In total, we obtained blood samples from 133 patients with RIF and 317 control participants. All patients and controls were Korean. The institutional review board of CHA Bundang Medical Center approved the study, and all patients provided written informed consent (reference no. CHAMC2009-12-120). All embryos were examined by an embryologist prior to transfer, and embryos that showed good quality were transferred. Implantation failure refers to cases where the level of human chorionic gonadotropin measured on the 14th day of embryo transfer is <5 U/mL [48]. In the study group, subjects diagnosed with implantation failure due to anatomical, chromosomal, hormonal, infectious, autoimmune, or thrombotic causes were excluded. Uterine anatomical abnormalities in RIF patients were confirmed by hysterosalpingography, hysteroscopy, uterine sonography, computed tomography, or magnetic resonance imaging. A karyotype analysis was conducted to confirm chromosomal abnormalities, and the karyotype analysis followed the standard protocol. Hormonal causes of RIF such as hyperproactinemia, lutein insufficiency, and thyroid disease were identified by measuring prolactin, thyroid stimulating hormone (TSH), free T4, FSH, luteinizing hormone (LH), and progesterone levels in peripheral blood. Lupus anticoagulants and anti-cardiolipin antibodies were tested to confirm the autoimmune disease lupus and antiphospholipid syndrome, respectively. A deficiency of protein C and protein S and the presence of anti-β2 glycoprotein antibodies were diagnosed as thrombosis. As a control group, women with regular menstrual cycles, pregnancy history of at least one naturally conceived pregnancy, no history of pregnancy loss, and karyotype of 46, XX were recruited from CHA Bundang Medical Center.
2.2. Estimation of Homocysteine, Folate, Total Cholesterol, Uric Acid, Blood Urea Nitrogen, Creatinine, and Blood Coagulation Status Blood samples were collected from RIF subjects after 12 h of fasting. We performed a fluorescence polarization immunoassay using the Abbott IMx analyzer (Abbott Laboratories, Abbott Park, IL, USA) to measure the homocysteine level. Folate was measured via a competitive immunoassay using the ACS 180Plus automated chemiluminescence system (Bayer Diagnostics, Tarrytown, NY, USA). Total cholesterol, uric acid, blood urea nitrogen, and creatinine were measured using commercially available enzymatic colorimetric assays (Roche Diagnostics, GmbH, Mannheim, Germany). The platelet, white blood cell, and hemoglobin levels were obtained using the Sysmex XE 2100 automated hematology system (Sysmex Corporation, Kobe, Japan). The prothrombin time (PT) and activated partial thromboplastin time (aPTT) were measured with an ACL TOP automated photo-optical coagulometer (LSI Medience, Tokyo, Japan).

Hormone Assays
On the second to third days of the women's menstrual cycle, blood samples were collected and measured using serum samples. Following the manufacturer's instructions, the estradiol (E2) and TSH levels were measured using radioimmunoassays (Beckman Coulter), and the FSH and LH levels were measured using enzyme immunoassays (Siemens, Munich, Germany).

SNP Selection and Genetic Analysis
We selected FSHR, INHA, ESR1, and BMP15, which are hormone-related genes associated with pregnancy. To select polymorphisms of the FSHR, INHA, ESR1, and BMP15 genes, studies on the association between pregnancy-related diseases (recurrent pregnancy loss, recurrent implantation failure, preeclampsia, premature ovarian failure, and poor ovarian response) and polymorphisms were investigated [46,[50][51][52][53][54][55]. Finally, a total of seven polymorphisms in FSHR (rs6165), INHA (rs11893842 and rs35118453), ESR1 (rs9340799 and rs2234693), and BMP15 (rs17003221 and rs3810682) were selected and studied. Genomic DNA was extracted from anticoagulated peripheral blood using a G-DEX blood extraction kit (Intron, Seongnam, Republic of Korea). All genetic polymorphisms were identified by a real-time polymerase chain reaction (PCR) using the TaqMan SNP Genotyping Assay Kit (Applied Biosystems, Foster City, CA, USA). To validate the real-time analysis, DNA sequencing was performed on approximately 10~15% of the samples by random selection using an ABI 3730XL DNA Analyzer (Applied Biosystems). The concordance of the quality control samples was 100%.

Statistical Analysis
Regarding the clinical characteristics of participants, the differences between the categorical variables were analyzed using a chi-square test and continuous variables using an independent sample t-test. The data were presented as mean and standard deviations for continuous variables, and frequency and percentage for categorical variables. The statistical normality of continuous variables was confirmed using the Kolmogorov-Smirnov test. For continuous variables showing a non-normal distribution (p < 0.05 in the Kolmogorov-Smirnov test), group difference analyses were performed using the Mann-Whitney test. Allele frequencies were determined to confirm deviations from the Hardy-Weinberg equilibrium. To select the best inheritance model for a specific polymorphism, Akaike's information criterion was calculated. The associations between FSHR, INHA, ESR1, and BMP15 polymorphisms and RIF incidence were calculated using adjusted odds ratios (AORs) and 95% confidence intervals (95% CIs) obtained from a multivariate logistic regression analysis adjusted for age. False discovery rate correction was used to adjust multiple comparison tests and provide a measure of the expected proportion of false positives among the data.
The open-source MDR software package (v.2.0, www.epistasis.org accessed on 1 April 2022) was used to perform the genetic interaction analysis. Using this MDR analysis, all possible genotype combinations for gene-gene interactions were identified and analyzed. The associations between each FSHR, INHA, ESR1, and BMP15 gene polymorphism and each clinical value (platelets, PT, aPTT, homocysteine, folate, natural killer cells, uric acid, and total cholesterol) for RIF patients were assessed using ANOVA and Kruskal-Wallis tests. For the overall statistical analysis, the level of statistical significance was set as p < 0.05. Table 1 presents the clinical variables of the 133 RIF patients and 317 control subjects. No significant differences in age distribution were observed between the RIF and control groups, indicating that our age frequency matching was satisfactory. The blood urea nitrogen, creatinine, PT, TSH, E2, LH, BMI, total cholesterol, and white blood cell levels were different between the patient and the control group. The blood urea nitrogen (p < 0.0001), creatinine (p < 0.0001), PT (p < 0.0001), TSH (p = 0.0001), E2 (p = 0.0002), and LH (p < 0.0001) levels increased significantly in the patient group, and on the contrary, BMI (p = 0.047), total cholesterol (p < 0.0001), and white blood cell (p = 0.005) values increased significantly in the control group.  We investigated the distribution of FSHR (rs6165), INHA (rs11893842, rs35118453), ESR1 (rs9340799, rs2234693), and BMP15 (rs17003221, rs3810682) polymorphisms in RIF patients and the control group ( Table 2). The AOR with respect to age was calculated from the logistic regression analysis. The frequency of each genotype in the control group was consistent with the Hardy-Weinberg equilibrium.   We investigated the genotype frequencies according to the number of implantation failures among RIF patients ( Table 2). For FSHR rs6165A>G, GG homozygous genotypes and the recessive model were found to exert a protective effect against RIF [AA vs. GG: AOR, 0.463; 95% CI, 0.218-0.980; p = 0.044 and AA+AG vs. GG (recessive model): AOR, 0.430; 95% CI, 0.210-0.877; p = 0.020]. The other gene polymorphisms (INHA rs11893842 and rs35118453, ESR1 rs9340799 and rs2234693, and BMP15 rs17003221 and rs3810682) did not show significant differences between the control and patient groups. We also identified the genotype frequencies among patients according to the number of RIFs (Table 3).   We performed a genotype combination analysis of RIF patients and control subjects ( Table 4). In the genotype combination type of FSHR rs6165 A>G and ESR1 rs9340799 A>G, the GG/AA (OR, 0.250; 95% CI, 0.072-0.874; p = 0.030) and GG/AA+AG (OR, 0.373; 95% CI, 0.171-0.816; p = 0.014) combinations were significantly associated with RIF risk. Additionally, in the FSHR rs6165A>G/INHA rs35118453C>T combination genotype, GG/CC+CT (OR, 0.399; 95% CI, 0.190-0.841; p = 0.016) showed a significantly decreased OR. In contrast, in the genotype combination analysis of INHA rs35118453 and C>T/ESR1 rs2234693 T>C, TT/TC+CC (OR, 7.001; 95% CI, 1.298-37.776; p = 0.024) was associated with an increased RIF risk. Additionally, in the INHA rs11893842 A>G and ESR1 rs9340799 A>G genotype combination, AA+AG/GG (OR, 3.065; 95% CI, 1.081-8.690; p = 0.035) was associated with an increased RIF risk. In the genotype combination analysis of ESR1 rs9340799 A>G/ESR1 rs2234693 T>C, GG/TT+TC (OR, 12.930; 95% CI, 1.492-112.052; p = 0.020) was associated with an increased RIF risk.  We performed an analysis of the differences in clinical factors according to the genotype of the FSHR (rs6165), INHA (rs11893842, rs35118453), ESR1 (rs9340799, rs2234693), and BMP15 (rs17003221, rs3810682) polymorphisms using a one-way analysis of variance (ANOVA). We found that increased LH levels were associated with the INHA rs35118453 polymorphism among all subjects (Table 5, p < 0.05). In RIF patients, increased E2 levels were associated with INHA rs11893842, and decreased CD3 (pan T cell) levels were associ-ated with ESR1 rs2234693 (Table S1, p < 0.05). TSH levels in the control group showed an increasing trend according to the FSHR rs6165 genotype (Table S2, p < 0.05).

Discussion
The formation and quality of gametocytes and embryo development are important factors for successful pregnancy progression, and many infertile couples have experienced recurrent failure due to the poor quality or quantity of embryos before transplantation. Currently, morphological indicators are used for embryo selection, and there is no other predictive indicator that can evaluate the quality of embryo cells. Recently, many studies have elucidated molecular and genetic factors to assess embryo quality and select developmentally competent embryos. Previous studies have reported several genes associated with the pathogenesis of oocyte maturation arrest (TUBB8, PATL2) and fertilization failures (TLE6, WEE2). [56]. In this study, seven polymorphisms were selected from four hormone-related genes (FSHR rs6165, INHA rs11893842 and rs35118453, ESR1 rs9340799 and rs2234693, BMP15 rs1700321, and rs3810682), and the association between RIF occurrence and polymorphisms was analyzed.
We examined the association between RIF occurrence and SNPs in the FSHR, INHA, ESR1, and BMP15 genes. Our results showed that the genotype frequency of the FSHR rs6165 SNP was significantly different in the control and RIF patient groups. A genotype combination analysis of the seven genetic markers revealed that the GG/AA (p = 0.030) and GG/AA+AG (p = 0.014) combinations significantly reduced the RIF risk in the FSHR rs6165 and ESR1 rs9340799 combination type. Additionally, in the genotype combination analysis of FSHR rs6165A>G/INHA rs35118453C>T, GG/CC+CT (p = 0.016) showed a significantly decreased risk of RIF. In contrast, the INHA rs35118453 TT and ESR1 rs2234693 TC+CC combination were associated with increased RIF risk (p = 0.024). Moreover, the INHA rs11893842 A>G/ESR1 rs9340799 A>G, AA+AG/GG combination (p = 0.035) was associated with an increased RIF risk. In the genotype combination of ESR1 rs9340799 A>G/ESR1 rs2234693 T>C, GG/TT+TC (p = 0.020) was associated with an increased RIF risk. Furthermore, our gene-environment combinatorial analysis data revealed statistically significant relationships between the FSHR rs6165, INHA rs11893842, ESR1 rs9340799, and rs2234693 genotypes and clinical factors (folate, homocysteine, E2, FSH, and LH) in Korean RIF patients. When combined with clinical parameters, the RIF risk increased by 2-to 11-fold.
The FSHR rs6165 polymorphism is located in exon 10 of the FSHR gene and is in the transmembrane region of the FSHR protein. The rs6165 polymorphism produces a change of A to G in position 919 and changes codon 307 from threonine (ACT) to alanine (GCT) [57]. According to Ganesh et al. [50], the FSHR rs6165 polymorphism was associated with unsuccessful IVF outcomes, and a higher frequency of the heterozygous AG genotype was observed in the infertile group than in the control group. For FSHR rs6165 AA carriers, the number of oocytes retrieved was significantly higher and ovarian stimulation was significantly shorter than those in GG and AG carriers [58]. Additionally, Rod et al. [59] found that FSHR rs6165 was associated with controlled ovarian stimulation and was 3-fold higher in poor responders than in good responders. Moreover, one study showed that the rs6165 AG genotype was associated with an increased risk of male infertility [60].
INHA rs11893842 and rs35119453 are located in the promoter region of the INHA gene and have been studied with regard to POI, male infertility (sperm parameters), and adrenocortical cancer [44,61,62]. Neither variant was associated with POI. Rafaqat et al. [48] found that the GG genotype frequency was increased in male infertility patients and showed a significant association with male infertility in the Pakistani population. Furthermore, rs11893842 minor alleles showed a low frequency in adrenocortical cancer patients and the rs11893842 AA genotype was associated with decreased INHA mRNA levels [62].
ESR1 rs9340799 and rs2234693 are located in intron 1, 351 and 1397 bp upstream of exon 2, respectively. The two polymorphisms are associated with cancer in females, including breast cancer and endometrial cancer [63,64], and are associated with POI [65]. The ESR1 rs9340799 GA and rs2234693 TC genotypes were associated with a decreased risk of POI, and the ESR1 rs9340799 AA and rs2234693 TT genotypes were associated with an increased risk of POI [65]. Additionally, the ESR1 rs9340799 GG genotype was associated with a 4-fold increased risk of endometriosis and a 3-fold increased risk of IVF failure in infertile patients [66]. Furthermore, the FSHR rs6165 GG, rs6165 AA, ESR1 rs9340799 GA, and rs2234693 TC gene combination enhanced the protective effect of FSHR gene variants and was associated with a reduced risk of fibrocystic mastopathy in infertile women [67].
Several clinical factors are involved in embryo implantation and pregnancy maintenance, including folate, homocysteine, and female hormones (E2, FSH, LH). Gaskins et al. [68] found that, among women receiving assisted reproductive technology, women with high serum folate levels (>26.3 ng/mL) had a 1.62-fold higher live birth rate than women with low folate levels (<16.6 ng/mL). Additionally, Ocal et al. [69] found that high homocysteine levels in follicular fluid resulted in reduced cell division and increased fragmentation in embryo cultures, which was associated with decreased oocyte and embryo quality. Another study showed that exposure to high E2 concentrations is detrimental to blastocyst implantation and early post-implantation development. Moreover, during clinically-assisted reproductive technology, high serum E2 levels not only affected the endometrium but also directly affected the blastocyst during implantation [70]. Our study revealed that patients with several clinical factors (low folate levels and high homocysteine, E2, FSH, and LH levels) and FSHR rs6165, INHA rs11893842, ESR1 rs9340799, and rs2234693 polymorphisms had an approximately 3-to 12-fold increased risk of RIF.
This study has several limitations. First, how FSHR, INHA, ESR1, and BMP15 polymorphisms influence RIF development is still unclear. The effect of these SNPs should also be confirmed through in vitro and in vivo studies. Second, additional environmental risk factors to RIFs need to be evaluated. Finally, the size of the RIF patient and control group was small, and this study group included only Koreans. To determine whether the studied genetic polymorphism can be used as a predictor of RIF, our results should be validated using larger sample sizes and other ethnic groups.

Conclusions
We investigated the association between the FSHR rs6165, INHA rs11893842 and rs35118453, ESR1 rs9340799 and rs2234693, and BMP15 rs17003221 and rs3810682 polymorphisms and the risk of RIF in Korean women. We found the frequency of the GG genotype of FSHR rs6165 was lower among RIF patients than among controls, suggesting this genotype may be associated with a reduced risk of RIF. Additionally, the interaction of the FSHR rs6165, INHA rs11893842, and ESR1 rs9340799 and rs2234693 polymorphisms with some clinical factors may increase the risk of RIF.