Association Between Follicle-Stimulating Hormone Receptor (FSHR) rs6166 and Estrogen Receptor 1 (ESR1) rs2234693 Polymorphisms and Polycystic Ovary Syndrome Risk, Phenotype, and Reproductive Outcomes in an Infertile Portuguese Population

Introduction: Polycystic ovary syndrome (PCOS) is a common endocrine disorder often leading to anovulatory infertility. PCOS pathophysiology is still unclear and several potential genetic susceptibility factors have been proposed. The effect of polymorphisms in two genesrelated to follicular recruitment and development, the follicle-stimulating hormone receptor (FSHR) and the estrogen receptor 1 (ESR1), have been studied in different populations with contradictory results. Aims: To evaluate the influence of FSHR rs6166 (c.2039A>G) and of ESR1 rs2234693 (Pvull c.453-397 T > C) polymorphisms on PCOS risk, phenotype, and response to controlled ovarian stimulation (COS). Materials and methods: Genotyping of the FSHR rs6166 and the ESR1 rs2234693 polymorphisms was performed in PCOS women and a control group undergoing in vitro fertilization (IVF). Demographic, clinical, and biochemical data, genotype frequency, and IVF outcomes were compared between groups. Results: We evaluated 88 PCOS women and 80 controls. There was no significant difference in the genotype distribution of FSHR rs6166 polymorphism between PCOS women and controls (AA 31.8%/AS 48.9%/SS 19.3% in PCOS women vs AA 37.5%/AS 40.0%/SS 22.5% in controls; p = 0.522). The same was true for the ESR1 rs2234693 (CC 24.1%/CT 46.0%/TT 29.9% in PCOS women vs CC 18.8%/CT 48.8%/TT 32.5% in controls; p = 0.697). In PCOS women, we found higher follicle-stimulating hormone (FSH) levels on the third day of the menstrual cycle associated with the SS variant of the FSHR polymorphism (9.2 vs 6.2 ± 1.6 and 5.6 ± 1.6 mUI/mL; p = 0.011). We did not find other associations between the baseline hormonal parameters, antral follicle count, and response measures to COS with FSHR or ESR1 genotypes. We found, however, a need for higher cumulative doses of FSH for COS in patients with the SS variant of the FSHR rs6166 polymorphism (1860.5 ± 627.8 IU for SSvs 1498.1 ± 359.3 for AA and 1425.4 ± 474.8 for SA; p = 0.046 and p = 0.046). Conclusion: Our data suggest that in the population, FSHR rs6166and ESR1 rs2234693 polymorphisms do not influence the risk of developing PCOS nor do they influence the patient’s phenotype and IVF success. However, the SS variant of the FSHR rs6166 polymorphism may be associated with FSH resistance requiring higher FSH doses for COS.


Introduction
Polycystic ovary syndrome (PCOS) is the most common endocrine disorder in women of reproductive age 1 2, 3, 4 3, 5 1 2 2 2, 6, 7 1, 7 2, 6 6, 8 9, 10 2, 11 1 2 6, 2, 7 with a prevalence that may be close to 18% [1]. It has adverse reproductive and metabolic implications [2], and it is the most common cause of anovulatory infertility [3]. The pathophysiology of PCOS is still poorly understood; however, several potential genetic susceptibility factors have been identified, including mutations and polymorphisms on the genes involved in the gonadal axis, steroid metabolism, cardiovascular risk, and insulin resistance [4]. In the setting of assisted reproduction techniques (ART), infertile women with PCOS constitute a challenge for controlled ovarian stimulation (COS) [5], and an influence of some polymorphic variations may be at play.
Estrogen and follicle-stimulating hormone (FSH), acting together, lead to an increase in follicle-stimulating hormone receptor (FSHR) expression in the granulosa cells contributing to the growth and maturation of ovarian follicles [6].
FSH is important for follicular development and oocyte maturation. Ovarian sensitivity or resistance to exogenous FSH is believed to be influenced by genetic variations related to FSH and its receptor [7]. The FSHR, encoded by a gene located in 2q, is a member of the G-protein coupled receptor family [8].
Abnormal FSHR function may lead to arrested follicular development, resulting in amenorrhea and high FSH levels [9]. Mutations in the FSHR are rare, however, several polymorphisms have been identified [7]. Two variants in exon 10 have received particular attention. In codon 307, a substitution of adenine to guanine in the extracellular domain of the FSHR results in a change from threonine (Thr) to alanine (Ala) -FSHR rs6165 (c.919G>A, p.Thr307Ala). In codon 680, a substitution of adenine to guanine leads an asparaginase codon to be replaced by a serine codon -FSHR rs6166 (c.2039A>G, p.Asn680Ser) [10]. Both variants are in almost complete linkage disequilibrium [11,12]. An association between FSHR polymorphism and PCOS risk has been studied with conflicting results [4]. Valkenburg et al. did not find an association between FSHR polymorphisms and disease risk but do report an association with its phenotype and hormone levels [12]. There is also some data to support an association between the SS variant of the FSHR rs6166 polymorphism with FSH resistance [13] and potentially altered response to ovarian stimulation [11].
Knowledge of genetic factors that may influence the ovarian response to COS may contribute to the development of pharmacogenetic approaches [17]. In this study, we genotyped the FSHR c.2039A>G (rs6166) polymorphism and one of the most common SNP in ESR1, ESR1 Pvull c.453-397 T > C (rs2234693), in a group of women with PCOS and a control group undergoing in vitro fertilization (IVF) treatments. We aimed to investigate if these SNPs influence the risk and phenotype of PCOS in the Portuguese population. We also sought to investigate if either of these polymorphisms was associated with PCOS biochemical phenotype and with the result of COS.

Patient selection
We carried out a retrospective observational study of women who underwent infertility treatment in a tertiary fertility clinic between 2013 and 2019.
This study was designed, conducted, and reported following the principles of Good Clinical Practice guidelines and the 1964 Helsinki Declaration and its later amendments. The Hospital Ethics Board granted ethical approval (internal reference number: 171-20), and all participants gave their written informed consent for blood sampling and genetic investigations for these specific targets.
The PCOS group (n = 88) and the control group (n = 80) were constituted of infertile patients undergoing the first cycle of IVF treatment. COS was performed with gonadotropin-releasing hormone (GnRH) antagonist protocol, using recombinant human FSH (rhFSH).
The diagnosis of PCOS was established based on the 2003 European Society of Human Reproduction and Embryology (ESHRE)/American Society for Reproductive Medicine (ASRM) Rotterdam criteria [18]. The presence of PCOS was confirmed by vaginal ultrasound examination.
The control group included age and body mass index (BMI) matched women with regular menstrual cycles. Additionally, the controls were infertile women with tubal obstruction with regular menstrual cycles, no clinical or biochemical signs of hyperandrogenism, and no polycystic ovaries. Patients with male factor infertility, endometrial disease, or severe systemic illness were excluded from both groups.

Genotyping
Genomic deoxyribonucleic acid (DNA) was isolated from peripheral blood collected in ethylenediaminetetraacetic acid-coated tubes on day three of the follicular phase and stored at 4°C until DNA extraction. Extraction of genomic DNA was performed using QIAamp® DNA Blood MiniKit (Qiagen, Hilden, Germany), according to the manufacturer's instructions. QIAamp® DNA Blood Kits provide silicamembrane-based DNA purification from biological fluids (including blood samples). We optimized buffers lyse samples, stabilized nucleic acids, and enhanced selective DNA adsorption to the membrane. Alcohol was added and sample lysates were loaded onto the spin column. Finally, wash buffers were added to remove impurities and purified DNA was then eluted in low-salt buffer. The quantity and purity of each eluted sample were assessed by NanoDrop TM 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA).
DNA amplification was carried out using the polymerase chain reaction (PCR) with specific primers to assess FSHR rs6166 (c.2039A>G) and ESR1 rs2234693 (Pvull c.453-397 T > C), as described in Table 1  Restriction fragment length polymorphism (RFLP) was carried out on the purified PCR products using restriction endonuclease enzymes. To determine the genotype for each sample, PCR products were incubated and digested with the respective restriction enzyme; PvuII (Invitrogen) and BseNI (Thermo Fisher Scientific, Waltham, MA) for ESR1 rs2234693 (Pvull c.453-397 T > C) and FSHR rs6166 (c.2039A>G), respectively, according to the manufacturer's instructions (   Figure 1).

FIGURE 1: ESR1 rs2234693 genotypes
The CC genotype is represented by one band of 1374 bp, the CT genotype is identified by three bands of 1374 bp, 936 bp, and 438 bp, and the TT genotype is represented by two bands of 936 bp and 438 bp.
FSHR rs6166 leads to a change from A to G. BseNI digestion produces an uncleaved 520 bp fragment that indicates homozygosity for asparagine (Asn/Asn). In contrast, two fragments of 413 and 520 bp indicate heterozygosity (Asn/Ser). The presence of one fragment of 413 bp reveals homozygosity for serine (Ser/Ser), as represented in Figure 2, as previously described in the work of Kuijper and collaborators [13].

FIGURE 2: FSHR rs6166 genotypes
The AA genotype is represented by one band of 520 bp, the SA genotype is identified by two bands of 520 bp and 413 bp, and the SS genotype is represented by one band of 413 bp.

Data collection
Clinical information of the patients was collected from the medical records. Collected data included age, BMI, antral follicle count (AFC), plasma concentrations of progesterone, testosterone, anti-müllerian hormone (AMH), FSH, luteinizing hormone (LH), and estradiol in the third and 23rd day of the menstrual cycle, outcomes of the IVF treatment, follicle count in the day of ovulation induction, number and quality of oocytes retrieved, and the number of blastocysts obtained. Information on prior medical history and medication use was also scanned for each of the patients to properly evaluate the above-mentioned inclusion and exclusion criteria.
For the included patients, the results of the analysis of each of the polymorphisms were also inserted in our database.

Data analysis
Statistical analysis was performed using SPSS version 23.0 (IBM Corp., Armonk, NY) and SNPStats. Statistical significance was assumed for p-value < 0.05. Hardy-Weinberg equilibrium was confirmed for both polymorphisms.
For continuous variables, when normal distribution was confirmed (through standardized asymmetry) and homogeneity of variances of variables was noted, Student's t (for two groups) or ANOVA (for ≥ three groups) tests were used to compare groups. In cases where it was not possible to assume a normal distribution, Mann-Whitney U and Kruskal-Wallis tests were used for comparison between two or > two groups, respectively. The chi-square test was performed for categorical variables to test differences in the distribution of two independent variables. Fisher's exact test was used instead whenever expected frequencies below five occurred with enough frequency.
After verifying the required assumptions, linear regression was used whenever there was a need to confirm an independent effect of an independent variable on a continuous dependent variable.

Sample characterization
Our sample was composed of 168 women, 88 of whom had PCOS. Women from the PCOS and control groups did not differ in age.
A comparison between PCOS and non-PCOS subjects regarding baseline characteristics, baseline hormonal levels, follicle count, and the number of oocyte numbers obtained is summarized in Table 3.  As it was intended by the study design, PCOS and non-PCOS groups were homogeneous in terms of age and body mass index (BMI). However, the groups differed in baseline hormonal levels (FSH and LH on days three and 23 of the menstrual cycle and estradiol, progesterone, and testosterone on day 23). As expected, the AMH, AFC, and follicle count on the day of ovulation induction were higher in the PCOS group. Even though there was a higher number of oocytes obtained in the PCOS group, the oocyte quality was lower. A higher number of degenerative or atretic oocytes was found in the PCOS group.

Evaluation of whether there is an association between FSHR and ESR1 SNPs and PCOS risk
The results of the genotype distribution of FSHR rs6166 genotype and ESR1 rs2234693 are reported in Table  4. We did not identify statistically significant differences in the distribution between PCOS patients and controls.

Evaluation of whether there is an association between FSHR and ESR1 SNPs and biochemical and ultrasonographic parameters in PCOS women
For PCOS patients, biochemical parameters to analyze the gonadal axes (FSH, LH, estradiol, and progesterone), total testosterone and AMH levels, and ultrasonographic data on antral follicle count were compared between different FSHR rs6166 and ESR1 rs2234693 genotypes. Patients with the SS variant of the FSHR polymorphism had higher FSH levels on the third day of the menstrual cycle (p = 0.011). We did not find other statistically significant differences in biochemical and ultrasonographic parameters in relation to the FSHR or the ESR1 genotypes. The detailed results are presented in Table 5.

Evaluation of whether FSHR and ESR1 SNPs influence immediate IVF outcomes in PCOS women
A comparison of an average daily dose of rhFSH, follicle count on the day of ovulation trigger, number and quality of obtained oocytes, and blastocyst number between different FSHR rs6166 and ESR1 rs2234693 is presented in Table 6.   There was a higher cumulative and daily rhFSH dose in women with the FSHR SS variant (p = 0.046 and 0.014, respectively). We then performed linear regression, with age, BMI, and SS vs remaining genotypes. The model was statistically significant for predicting cumulative rhFSH dose (p = 0.001), with the presence of the FSHR rs6166 SS variant being an independent predictor (B = 0.222, t = 2.349, p = 0.021). For ESR1 rs2234693, the CT genotype was associated with a numerically higher rhFSH dose, but without achieving statistical significance.
No differences were observed with any polymorphisms concerning FSH dose, AFC, metaphase II oocytes, low-quality oocytes (metaphase I and atretic), and blastocyst number.

Discussion
In this study, we evaluated women with and without PCOS. By study design, the groups with and without PCOS were homogeneous in relation to age and BMI. Nevertheless, and in agreement with what is described, these two groups had significant differences in hormonal patterns. PCOS was associated with elevated gonadotropins and testosterone, and lower estrogens and progesterone. These women also had higher antral follicle count and more oocytes obtained on average. However, there was no statistically significant difference in relation to the percentage of mature oocytes. After applying IVF techniques, the number of blastocysts obtained was similar between the two groups. These data suggest that even though PCOS women have an endocrine dysregulation that causes anovulatory infertility, a good response after COS is expected, which is in accordance with previously published data [19].
Our first question was whether FSHR rs6166 and ESR1 rs2234693 were associated with PCOS risk. We did not find statistically significant differences in the distribution of each of the genotypes of FSHR rs6166 and ESR1 rs2234693 polymorphisms, suggesting that in our population, neither of the studied polymorphisms is associated with PCOS risk.
Indeed, results from previous studies regarding the relationship between both FSHR and ERS1 polymorphisms and PCOS present conflicting data. In a group of adolescents in Turkey, Unsal et al. did not find a different distribution of several FSHR polymorphisms associated with PCOS [4]. Wu et al. also failed to find an association between the FSHR polymorphisms and PCOS in women in the north of China, but the authors did report an association with higher levels of FSH [8]. An absence of association was also reported in Thai [20] and in Sri Lankan women [21]. More recently Wan et al. also reported an absence of association between several FSHR polymorphisms, including rs6166, and PCOS risk in Asian women [22]. An association between FSHR rs6166, but not rs6165, was reported by Gu et al. in Korean women [23]. Conversely, in an Italian cohort, Dolfin et al. did find a relationship between the rs6165 variant and PCOS risk [24]. Kim et al. found a significant association between rs6165 and rs6166 FSHR polymorphisms and PCOS in South Korea [11].
Regarding ERS1, the contribution of ERS1 genetic variants to PCOS is also controversial. In a study from Korea, Kim et al. reported an association with the risk of PCOS [25]. Jiao et al. also found such an association in the Chinese population [26]. Conversely, Silva et al. [27], Valkenburg et al. [12], and Mir et al. [28] found no association with PCOS risk. A recent meta-analysis found no significant association between several ERS1 polymorphisms and PCOS risk, even taking ethnicity into account [29].
It has also been hypothesized that even if there is no association between FSHR polymorphisms or ERS1 and disease risk, there may be an association with its phenotype and hormone levels [12,27]. As such, we aimed to understand whether the distinctive hormonal pattern observed in these patients could be attributed to FSHR rs6166 or ESR1 rs2234693 polymorphisms. The SS genotype of the FSHR rs6166 polymorphism was associated with higher FSH levels on the third day of the menstrual cycle. We found no other differences in evaluating the gonadal axes, AMH, or AFC between different FSHR rs6166 and ESR1 rs2234693 genotypes. It has previously been reported that the SS variant of the FSHR rs6166 polymorphism may be less sensitive to FSH associated with higher FSH levels [30,31]. Our findings support this theory.
There has been some research on the influence of the polymorphisms evaluated in this study on IVF outcomes. As such, we also sought to understand if FSHR and ESR1 SNPs influence immediate IVF outcomes in PCOS women. We did not find variations in oocyte quality or the number of blastocysts obtained in relation to these polymorphisms. We did, however, find a need for higher rhFSH dose in women with the FSHR rs6166 S/S variant, which may indicate a tendency toward FSH resistance in patients with these variants. Some authors have reported differences in response to COS related to FSHR rs6166. Jun et al. found a lower number of oocytes retrieved in association with the S/S variant [32]. Behre et al. reported lower FSHR sensitivity associated with this variant, but they also state that it can be overcome by increasing the rhFSH dose [33]. Loutradis et al. found an association between the AS variant and more obtained follicles and oocytes [34]. Regarding the ERS1 rs2234693 polymorphism, there are fewer studies. However, studies with Chinese [6] and Greek women [35] suggest that the variant is associated with worse IVF results, which we did not observe in our sample.
Controversial findings in this field may be related to differences in study design and sample selection. As other authors have stated, the contribution of each gene in multifactorial diseases, such as PCOS, is small, and very large samples may be necessary to detect a small effect. Furthermore, there is the possibility of incomplete penetrance and gene-gene interactions; therefore, studies of individual genes in each population may not reveal the general pattern [31]. Nevertheless, it is possible that due to genetic heterogeneity, the same polymorphism has distinct contributions to PCOS risk and phenotype across different populations. This adds another level of complexity when defining the genetic risk factors based on a complex disease.
Our study has some limitations, mainly the relatively small sample size and lack of data on the women in whom cycle cancelation was required. The fact that the population studied is composed exclusively of women referred for fertility treatments precludes the generalization of our conclusions to all Portuguese PCOS patients. Additionally, controversial findings still exist when applying SNPs analysis in infertile populations, which may be related to the diverse populations and genotyping methods.

Conclusions
As far as we know, this is the first study evaluating the association between FSHR rs6166 and ESR1 rs2234693 polymorphisms and PCOS in the Portuguese population with infertility. Our data do not support an association between these polymorphisms and PCOS risk, phenotype, and immediate IVF outcomes. Furthermore, the ESR1 rs2234693 polymorphism was not associated with a difference in baseline hormone values or response to ovarian stimulation. Nevertheless, the SS genotype of the FSHR rs6166 polymorphism was associated with higher FSH levels on the third day of the menstrual cycle and also with a need for a higher cumulative dose of FSH in COS, which may reflect a lower sensitivity to FSH. Therefore, COS with higher doses of FSH may be more suitable in women with this genotype. , and all participants gave their written informed consent for blood sampling and genetic investigations for these specific targets. Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: SAR is supported by the Portuguese Funding Agency for Science and Technology (PD/BD/128237/2016 -PhD Programme in Experimental Biology and Biomedicine). The funding entities had no role in study design, data collection and analysis, the decision to publish, or the preparation of the manuscript. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.