{"paper_id":"5309a03a-71ae-4ae5-b959-42b4038ddc82","body_text":"Since introduction of IVF in the clinical practice of infertility treatment, the follicular response to ovarian stimulation protocols has been largely investigated. The unpredictable ovarian response to gonadotropins among patients, ranging from poor response to ovarian hyperstimulation syndrome (OHSS) has been one of the most challenging problems in medically assisted reproduction (MAR) [ 1 ]. The ovarian response to gonadotrophin stimulation is difficult to predict even in patients with similar endocrine profiles. This has led to the investigation of specific new biomarkers that could serve as predictors of ovarian response to an exogenous hormonal stimulation. Recently, gene association studies have tried to identify a number of genetic variations influencing inter-individual variability in COH [ 2 ,  3 ].\nSince the first report on the polymorphism of the  FSHR  gene [ 4 ], numerous mutations and SNPs of the  FSHR  gene have been described. Very common coding SNPs rs6165, rs6166, and rs1394205 are currently most extensively studied to assess the response of the FSHR protein to FSH stimulation. Some authors have reported predictability of the ovarian response to FSH stimulation in patients with different alleles [ 5 – 10 ], while others have refuted this finding [ 11 ,  12 ]. The results of studies regarding the impact of ethnicity on the frequency distribution of follicle-stimulating hormone receptor (FSHR) polymorphisms are contradictory. Some studies mention the influence of ethnicity on the distribution of the genotype [ 13 ], while others contradict it [ 8 ].\nImportant candidate genes involved in the ovarian response to exogenous FSH are the estrogen receptor genes ( ESRs ) [ 14 ]. The most studied polymorphisms in  ESR1  are rs2234693 and rs9340799. Women with TT genotype at rs2234693 when compared to those with CC genotype, demonstrate improved quality of the ovarian follicles [ 15 ], as well as higher number of follicles, mature oocytes, higher fertilization rate, and better embryo quality following COH and IVF [ 16 – 18 ]. The results of meta-analysis done on Asian population strongly suggested that  ESR1  gene rs2234693 polymorphism was significantly associated with an increased risk of premature ovarian failure [ 19 ].\nRigon et al. [ 20 ] confirmed an association of the  AMH  gene polymorphisms, and its receptor  AMHR  with estradiol levels during the follicular phase of the menstrual cycle in normo-ovulatory women. AMH is produced by the granulosa cells of early developing follicles in the ovary and it continues to be expressed in the growing follicles until they are selected for dominance by the action of follicle-stimulating hormone (FSH) [ 21 ,  22 ]. Studies in  AMH  knockout mice have demonstrated that, in the absence of  AMH , follicles are recruited at a faster rate, and they are more sensitive to FSH [ 23 ], suggesting that serum AMH could inhibit primordial follicle development and be induced by FSH. This expression pattern suggests that  AMH  can inhibit both the initiation of primordial follicle growth and FSH-induced follicle growth. Therefore,  AMH  plays an important role in regulating both primordial follicle recruitment and cyclic selection [ 24 ,  25 ]. Because AMH may have an inhibitory effect on the FSH-sensitivity of follicles, polymorphisms on the  AMH  gene or its receptor  AMHR  might reduce the biological activity of the hormone. Therefore, follicles might be more sensitive to FSH and might be previously selected for dominance [ 26 ].\nThe aim of our study was to analyze SNPs in selected candidate genes which are responsible for the hormonal regulation of folliculogenesis and to determine whether the reasons for the different ovarian responses to COH derive from a difference in individual genotype.\n\nSixty (60) women undergoing ovarian stimulation between March 1, 2015 and July 7, 2015 were included in this study. The study was approved by the Slovenian National Committee for Medical Ethics (012-347/2015-8). It is a part of the research program P3-0327 and the research project J3-7177 funded by the Slovenian Research Agency. Patients were included in the study after signing a written consent.\nPatients were classified into three groups according to the ovarian response to stimulation protocol and the number of oocytes obtained after oocyte pick up. Group 1 were patients with a poor ovarian response (POR) and with up to three oocytes obtained and with additional risk factors for POR: antral follicle count < 5 or AMH < 0.5 ng/mL [ 23 ]. Group 2 represented women with normal response—those with up to 15 oocytes and Group 3 of hyper responders with more than 15 oocytes.\nExclusion criteria were endometriosis and polycystic ovary syndrome and age of more than 39 years.\nOf the 60 included women, 26 (43.4%) were classified as normal responders, 17 (28.3%) hyper responders, and 17 (28.3%) as poor ovarian responders.\nBlood samples were taken for basal hormonal measurements performed by enzyme-linked immunosorbent assay (ELISA). Serum FSH level was measured on day 3 of the menstrual cycle. Serum E2 level was measured on the day of hCG administration. Serum AMH level was determined independently of the menstrual cycle.\nPatients were assigned to ovarian stimulation with a combination of GnRH antagonist cetrorelix (Cetrotide 0.25 mg®; Merck Serono, Switzerland) and recombinant FSH (Gonal-F®, Merck Serono, Switzerland). The dose of exogenous gonadotropins was adjusted according to the follicular response, followed by serial transvaginal ultrasonography and E 2  level measurement. When the follicle had reached 18 mm in diameter, 250 μg recombinant human chorionic gonadotropin (Ovitrelle ®; Merck Serono, Switzerland) was given subcutaneously to induce ovulation. Oocyte retrieval was performed by ultrasound-guided aspiration of follicles, 35 to 36 h after recombinant human chorionic gonadotropin administration. The cumulus-oocyte complexes from follicular aspirates were collected in oocyte collection medium and incubated in a CO 2  incubator before insemination.\nDNA samples were obtained from 9 mL of the patient’s peripheral blood. First peripheral blood lymphocytes were collected using FicollPaque PLUS (GE Healthcare, Uppsala, Sweden) and then DNA was isolated from lymphocytes using TRI reagent (Sigma, Steinheim, Germany) according to the manufacturer’s instructions.\nGenotyping of SNPs rs10407022 in gene  AMH , rs3741664 in gene  AMHR , rs1394205 and rs6166 in gene  FSHR , and rs2234693 in gene  ESR1  was performed using polymerase chain reactions (PCR) followed by high-resolution melting analysis (HRMA). Forward and reverse primer sequences, primer concentrations, and annealing temperatures are shown in Table  1 . Genotyping was performed on real time PCR LC480 instrument (Roche, Germany), using LC480 HRM Master Mix (Roche, Germany). Conditions were as follow: initial denaturation at 95 °C for 10 min, followed by 45 cycles of 95 °C for 10 s, 60 °C for 15 s and 72 °C for10 s, followed by HRM step of 95 °C for 1 min, 40 °C for 1 min, and 60–90 °C at 0.02 °C/s. Genotypes were determined using gene-scanning analysis software. Table 1 Forward and reverse primer sequences, primer concentrations, and annealing temperatures Gene SNP ID Variation Region Forward and reverse primer Annealing temperature (°C) Primer concentration (nM) Genotyping method \n FSHR \n rs1394205 -29G/A Non-coding AGCTTCTGAGATCTGTGGAGG 62 300 HRM AGCAAAGAGACCAGGAGCAG rs6166 Asn680Ser Coding CTTCAGCTCCCAGAGTCACC 62 300 HRM CATTGTGTTTTAGTTTTGGGCTAA \n AMHR \n rs3741664 4952G/A Non-coding CGTCTCCAGCTTTGTGTACC 62 400 HRM GTCACTGGTGTACTGGGTCA \n ESR1 \n rs2234693 PvuII T/C Coding TGTTCTGTGTTGTCCATCAGT 62 400 HRM CTCTAGACCACACTCAGGGT \n AMH \n rs10407022 Ile49Ser Coding TCCGAGAAGACTTGGACTGG 62 300 HRM AGCTGCTGCCATTGCTGT Notes HRM  high resolution melting\nForward and reverse primer sequences, primer concentrations, and annealing temperatures\nNotes\nHRM  high resolution melting\nT-test or Mann Whitney U-test was used to assess the statistical differences between groups of patients and biological and clinical parameters. To compare genotype and allele frequencies of selected SNPs between groups of patients, two-sided Fischer’s exact test was used. The data obtained were presented as mean ± standard deviation (SD).  P  < 0.05 was considered statistically significant. Odds ratios (OR) were also calculated with 95% confidence interval (95% CI).\n\nClinical and biological characteristics of patients, classified in three groups according to the ovarian response to gonadotropin are shown in Table  2 . Results showed significant difference in age between groups. Mean age of poor responding patients was higher (34.9 years) in comparison with normo- (32.5 years,  P  = 0.019) and hyper-responders (30.8 years,  P  < 0.0005). Level of serum FSH was significantly higher in the group of poor responders (9.42 mIU/mL) compared to normo- (5.59 mIU/mL  P  = 0.046) and hyper-responders (5.55 mIU/mL,  P  = 0.025). Serum AMH level was significantly different between all groups ( P  < 0.05). Number of aspirated follicles, number of retrieved oocytes, and dosage of rFSH per oocyte retrieved were significantly different between groups ( P  < 0.05). Applicated rFSH dose was significantly higher in the group of poor responders, compared to normo- ( P  = 0.002) and hyper-responders ( P  < 0.0005). Table 2 Main characteristic of study participants and analyzed parameters All participants Poor responders Normo responders Hyper responders P  value Poor vs. normo Normo vs. hyper Poor vs hyper No. 60 17 26 17 Age (years) 32.72 ± 0.66 34.92 ± 0.76 32.50 ± 1.15 30.85 ± 1.11 \n 0.019 \n 0.067 \n < 0.0005 \n BMI (kg/m 2 ) 24.72 ± 0.82 23.76 ± 1.17 25.90 ± 1.32 23.87 ± 1.71 0.808 0.475 0.648 bFSH (mIU/mL) 6.66 ± 0.49 9.42 ± 1.37 5.59 ± 0.31 5.55 ± 0.43 \n 0.046 \n 0.322 \n 0.025 \n AMH (ng/mL) 3.55 ± 0.54 0.46 ± 0.10 3.94 ± 0.80 6.05 ± 0.96 \n < 0.0005 \n \n 0.011 \n \n < 0.0005 \n Estradiol on hCG day (pmol/L) 4.99 ± 0.62 2.82 ± 0.81 3.96 ± 0.57 8.75 ± 1.41 0.078 \n 0.005 \n \n < 0.0005 \n No. of follicles punctured 12.44 ± 1.40 3.77 ± 1.00 10.57 ± 1.17 23.69 ± 2.19 \n < 0.0005 \n \n < 0.0005 \n \n < 0.0005 \n Oocytes retrieved 10.72 ± 1.16 3.54 ± 0.94 8.90 ± 0.70 20.69 ± 1.65 \n < 0.0005 \n \n < 0.0005 \n \n < 0.0005 \n rFSH (IU) 1855.44 ± 127.83 2688.46 ± 299.56 1631.25 ± 91.35 1367.31 ± 150.99 \n 0.002 \n 0.056 \n < 0.0005 \n rFSH (IU) per oocyte 462.99 ± 115.84 1233.52 ± 325.65 217.56 ± 27.60 70.05 ± 10.38 \n < 0.0005 \n \n < 0.0005 \n \n 0.001 \n Notes:  P  < 0.05 was considered statistically significant. Statistically significant values are written in bold BMI body mass index, bFSH basal follicle-stimulating hormone, AMH basal anti-Müllerian hormone, hCG human chorionic gonadotropin, rFSH recombinant follicle-stimulating hormone, IU international units P  is from t-test\nMain characteristic of study participants and analyzed parameters\nNotes:  P  < 0.05 was considered statistically significant. Statistically significant values are written in bold\nBMI body mass index, bFSH basal follicle-stimulating hormone, AMH basal anti-Müllerian hormone, hCG human chorionic gonadotropin, rFSH recombinant follicle-stimulating hormone, IU international units\nP  is from t-test\nGenotype and allele frequencies were calculated for patients as a whole group and separately according to response on COH. When comparing genotype and allele frequencies between groups, we found higher frequency of GG genotype of SNP rs1394205 in  FSHR  gene in poor- than in hyper-responders (76.5 vs. 37.5%,  P  = 0.002). The G allele was present in higher frequency in the group of poor responders compared to normo-responders (88.2 vs. 53.8%,  P  = 0.001, OR = 0.156) and hyper responders (88.2 vs. 62.5%,  P  = 0.015). Distribution of analyzed SNP genotypes of other genes was not significantly different between three patient groups (Table  3 ). Table 3 Associations between selected SNPs and response to hormonal regulated folliculogenesis Gene/SNP ID Genotype/allele All participants Poor responders Normo responders Hyper responders AMH  rs10407022 N  = 58 N  = 17 N  = 25 N  = 16 TT ( n  = 42) 72.4% ( n  = 13) 76.5% ( n  = 18) 72% ( n  = 11) 68.8% GT ( n  = 14) 24.2% ( n  = 4) 23.5% ( n  = 5) 20% ( n  = 5) 31.2% GG ( n  = 2) 3.4% ( n  = 0) 0% ( n  = 2) 8% ( n  = 0) 0% T 0.845 0.882 0.820 0.844 G 0.155 0.118 0.180 0.156 Statistical analysis Poor vs. normo Normo vs. hyper Poor vs. hyper TT vs. GT+GG 1.000 0.723 0.708 P  value 1.026 1.439 1.477 OR 0.241–4.369 0.355–5.837 0.317–6.895 95% CI 0.586 0.946 0.648 P  value T vs. G 1.429 0.972 1.389 OR 0.394–5.182 0.288–3.285 0.338–5.711 95% CI AMHR  rs1741664 N  = 57 N  = 17 N  = 25 N  = 15 AA ( n  = 0) 0% ( n  = 0) 0% ( n  = 0) 0% ( n  = 0) 0% AG ( n  = 18) 31.6% ( n  = 3) 17.6% ( n  = 10) 40% ( n  = 5) 33.3% GG ( n  = 39) 68.4% ( n  = 14) 82.4% ( n  = 15) 60% ( n  = 10) 66.7% \n A \n 0.158 0.088 0.200 0.167 \n G \n 0.842 0.912 0.800 0.833 Statistical analysis Poor vs. normo Normo vs. hyper Poor vs. hyper 0.190 0.729 0.671 P  value AG vs. GG 0.321 1.667 0.536 OR 0.073–1.414 0.407–6.818 0.098–2.941 95% CI 0.181 0.528 0.526 P  value A vs. G 0.400 1.500 0.600 OR 0.101–1.581 0.423–5.315 0.122–2.943 95% CI FSHR  rs1394205 N  = 59 N  = 17 N  = 26 N  = 16 AA ( n  = 7) 11.9% ( n  = 0) 0% ( n  = 5) 19.2% ( n  = 2) 12.5% AG ( n  = 26) 44.05% ( n  = 4) 23.5% ( n  = 14) 53.8% ( n  = 8) 50% GG ( n  = 26) 44.05% ( n  = 13) 76.5% ( n  = 7) 27% ( n  = 6) 37.5% A 0.339 0.118 0.462 0.375 G 0.661 0,882 0.538 0.625 Statistical analysis Poor vs. normo Normo vs. hyper Poor vs. hyper 0.139 1.000 0.103 P  value AA vs. AG+GG 1.810 1.032 2.308 OR 1.59–2.409 0.210–5.058 1.533–3.475 95% CI 0.080 0.322 \n 0.002 \n P  value AA+AG vs.GG 0.113 2.111 0.239 OR 0.027–.467 0.567–7.855 0.054–1.066 95% CI \n 0.001 \n 0.436 \n 0.015 \n P  value A vs. G 0.156 1.429 0.222 OR 0.048–0.505 0.581–3.513 0.063–0.787 95% CI FSHR  rs6166 N  = 60 N  = 17 N  = 26 N  = 17 AA ( n  = 20) 33.3% ( n  = 7) 41.2% ( n  = 9) 34.6% ( n  = 4) 23.5% AG ( n  = 28) 46.7% ( n  = 7) 41.2% ( n  = 13) 50% ( n  = 8) 47.1% GG ( n  = 12) 20% ( n  = 3) 17.6% ( n  = 4) 15.4% ( n  = 5) 29.4% A 0.567 0.618 0.596 0.471 G 0.433 0.382 0.404 0.529 Statistical analysis Poor vs normo Normo vs hyper Poor vs hyper 0.528 0.740 0.465 P  value AA vs AG+GG 1.575 1.333 2.100 OR 0.440–5.638 0.327–5.434 0.474–9.297 95% CI 1.000 0.465 0.438 P  value AA+AG vs.GG 1.111 1.909 2.121 OR 0.228–5.411 0.453–8.044 0.414–10.87 95% CI 0.582 0.428 0.225 P  value A vs. G 1.282 1.429 1.831 OR 0.530–3.005 0.590–3.459 0.687–4.878 95% CI ESR1  rs2234693 N  = 60 N  = 17 N  = 26 N  = 17 CC ( n  = 11) 18.3% ( n  = 4) 23.5% ( n  = 4) 15.4% ( n  = 3) 17.6% CT ( n  = 34) 56.7% ( n  = 8) 47.1% ( n  = 18) 69.2% ( n  = 8) 47.1% TT ( n  = 15) 25% ( n  = 5) 29.4% ( n  = 4) 15.4% ( n  = 6) 35.3% C 0.467 0.471 0.500 0.412 T 0.533 0.529 0.500 0.588 Statistical analysis Poor vs normo Normo vs hyper Poor vs hyper 0.92 1.000 1.000 P  value CC vs. CT+TT 1.692 0.788 1.333 OR 0.361–7.943 0.152–4.088 0.248–7.174 95% CI 0.481 0.465 1.000 P  value CC + CT vs. TT 0.571 1.909 1.01 OR 0.137–2.384 0.453–8.044 0.247–4.817 95% CI 0.926 0.699 0.787 P  value C vs. T 0.960 1.190 1.143 OR 0.404–2.282 0.491–2.885 3.015–0.433 95% CI Notes P  is from Fisher exact test\nAssociations between selected SNPs and response to hormonal regulated folliculogenesis\nNotes\nP  is from Fisher exact test\nAnalyzed correlations between SNP genotypes and clinical characteristics are included in Table  4 . Table 4 Statistically significant associations between selected SNPs and patients’ baseline hormonal values and rFSH dose used for ovarian hyperstimulation Gene SNP Characteristic Genotype Median (interquartile range) P  value \n AMHR \n rs3741664 rFSH (IU) AG 1420 (338) \n 0.028 \n AG vs. GG GG 2025 (1350) \n FSHR \n rs1394205 AMH (ng/mL) AA 6.80 (8.92) \n 0.016 \n AA+AG vs.GG AG 2.83 (5.33) GG 1.36 (3.01) rFSH (IU)/oocyte AA 147.92 (268.21) \n 0.036 \n AA+AG vs.GG AG 125.00 (108.04) GG 235.71 (571.88) \n FSHR \n rs6166 bFSH (mIU/mL) AA 5.75 (2.83) \n 0.043 \n AA vs. AG+GG AG 5.25 (2.94) GG 4.60 (1.50) \n ESR1 \n rs2234693 Estradiol on hCG day (pmol/L) CC 1.980 (3.44) \n 0.038 \n CC vs CT+TT CT 4.625 (5.36) TT 3.060 (9.18) Notes:  P  < 0.05 was considered statistically significant. Statistically significant values are written in bold bFSH  basal follicle-stimulating hormone,  AMH  basal anti-Müllerian hormone,  hCG  human chorionic gonadotropin,  rFSH  recombinant follicle-stimulating hormone,  IU  international units P  is from Mann-Whitney test\nStatistically significant associations between selected SNPs and patients’ baseline hormonal values and rFSH dose used for ovarian hyperstimulation\nNotes:  P  < 0.05 was considered statistically significant. Statistically significant values are written in bold\nbFSH  basal follicle-stimulating hormone,  AMH  basal anti-Müllerian hormone,  hCG  human chorionic gonadotropin,  rFSH  recombinant follicle-stimulating hormone,  IU  international units\nP  is from Mann-Whitney test\nSNPs rs3741664 in  AMHR  gene as well as rs1394205 and rs6166 in  FSHR  gene were positively associated with serum AMH, FSH, and rFSH dose used. Selected SNPs did not show any association with oocyte number.\nPatients with GG genotype of SNP rs1394205 in  FSHR  gene had lower measured serum AMH level ( P  = 0.016), and required higher rFSH dose per oocyte ( P  = 0.036) than patients with AA or AG genotype.\nPatients with AA genotype of SNP rs6166 in  FSHR  gene had higher level of measured basal serum FSH compared to those with AG or GG genotypes ( P  = 0.043).\nWomen with GG genotype of SNP rs3741664 in  AMHR  gene required higher rFSH dose in COS in comparison with patients carrying genotype AG ( P  = 0.028).\n\nOvarian stimulation in MAR involves the use of exogenous gonadotropins, which can result in an excessive response including OHSS, or inadequate response leading even to the cancelation of the IVF cycle. This significant variability in response has been the focus of many pharmacogenetic studies, which have analyzed the relationship between selected SNPs in candidate hormonal receptor genes involved in folliculogenesis and ovarian response to COH [ 6 – 8 ,  11 ,  12 ]. Among them, SNPs rs10407022 in gene  AMH , rs3741664 in gene  AMHR , rs1394205 and rs6166 in gene  FSHR , and rs2234693 in gene  ESR1  were the largely studied genes to date. Each of the mentioned studies analyzed only single SNP or only several SNPs in the same gene. In our study we were focused on several genes and their polymorphisms of various hormonal receptors involved in folliculogenesis. By finding association between different polymorphisms of selected genes, their genotypes and parameters characterizing ovarian response to COH, new genetic biomarkers for prediction of ovarian stimulation could be identified.\nDue to important roles of FSH in follicular growth and ovarian steroidogenesis in females, mutations in  FSHR  gene could affect reproductive ability [ 7 ]. Two polymorphisms, rs6165 and rs6166 in  FSHR  gene are in almost complete linkage disequilibrium [ 27 ]. This is why most of the studies were focused only on SNP rs6166, as genotyping of either of them permits genotype inference of the other.\nIn the present study we investigated the association of rs6166  FSHR  polymorphism with the clinical and endocrinologic parameters of study group patients. The results showed the highest frequency distribution of the AG genotype and are consistent with high AG genotype distribution found also in other ethnic groups [ 6 – 8 ,  11 ,  12 ].\nIt has been reported that basal FSH (bFSH) levels differ significantly among the rs6166 genotype variants with carriers of the GG genotype, having slightly higher bFSH levels and requiring a significantly higher gonadotropin dose to induce ovulation [ 5 ,  6 ,  11 ,  28 – 31 ]. In the present study bFSH was different among various genotype variants. Similar to other studies, patients with AA genotype had higher bFSH compared to women with AG and GG genotypes in one group ( P  = 0.043) [ 8 ,  12 ,  13 ,  32 ].\nYan et al. [ 33 ] reported that subjects with AA genotypes had higher basal FSH levels, and that these genotypes were associated with an increased risk of poor response. Their data suggested that personalized FSH therapy may be applied according to patient’s genetic background in clinical settings. Also, results of allele frequency analysis showed higher frequency of allele A in poor responders. From all these studies, it can be concluded that the A allele is associated with poor ovarian response to gonadotropin therapy. We could consider that patients with the genotype AA were less sensitive to FSH, because they have increased bFSH and hence may require a higher dose of rFSH to normal follicle development.\nLindgren et al. [ 34 ] suggested that combination of SNPs from more hormone receptors involved in folliculogenesis could give more reliable COS outcomes prediction value. In their study performed on IVF, women of a Caucasian origin,  FSHR  rs6166 and  LHCGR  rs2293275 SNPs alone were not associated with increased live birth rate. But, when they combined receptors, they found that women homozygous for serine in both  FSHR  rs6166 and  LHCGR  rs2293275 had approximately 40% higher live birth rate compared to those with other receptor variants.\nIn our study SNP rs1394205 in  FSHR  gene was also analyzed. The data revealed that women with GG genotype at the rs1394205 position were classified more often as poor responders compared to women with AA genotype. They needed a higher amount of exogenous FSH per oocyte retrieved compared to AA and AG genotype patients ( P  = 0.036). These results indicate that the SNP rs1394205 in  FSHR  gene may influence sensitivity of the FSHR to FSH.\nSNP rs1394205 is located in the promoter region of  FSHR  gene and has been associated with altered transcriptional activity of the  FSHR  gene [ 35 ]. It is suggested that the reduced  FSHR  expression at the transcript level is in concurrence with the expression of  FSHR  at the protein level [ 36 ]. Chai et al. [ 37 ] report that the expression of  FSHR  gene, at both the mRNA and protein levels, is significantly different among the three groups (poor, normo, and hyper responders), with the lowest expression in the poor responders. They observe the highest dosage of rFSH and the higher levels of FSH in follicular fluid of poor responders. Because the secretion of FSH is in a negative feedback loop with the action of  FSHR , the basal levels of FSH are often indicative of the function of  FSHR  [ 38 ], which suggests that increased administration of gonadotropin might elevate the local concentration of FSH and improve the oocyte development. The findings suggest that increasing the dose of rFSH does not improve oocyte development probably due to insufficiency of  FSHR  expression on granulosa cells [ 37 ].\nWunsch et al. [ 32 ] analyzed whether there are ethnic differences concerning the SNPs in the promoter region in DNA samples of 55 Indonesian women. Interestingly, a different distribution pattern was found compared with the Caucasian population. The distribution in German patients was as follows: GG (55.4%), AG (37.6%), and AA (6.9%), while the Indonesian women showed the following repartition: GG (29%), AG (49%), and AA (22%). Distribution of analyzed SNP genotypes in our study group was GG (44.05%), AG (44.05%), and AA (11.9%). In another study on Indian women, those with AA genotype of SNP rs1394205 in  FSHR  gene were compared with GG genotype women and it was revealed that the AA genotype required higher dose of exogenous FSH for ovarian hyperstimulation [ 36 ]. They concluded that AA genotype at position rs1394205 might be associated with poor ovarian response. Findings from their study consistently demonstrate a correlation of the AA genotype at rs1394205 position of the FSHR gene with poor ovarian response. Interestingly, when they compared the FSHR expression at protein level on the basis of genotypes at position rs1394205, they observed that subjects with the AA genotype expressed significantly lower amounts of receptor protein compared with the GG and GA genotypes. Their observation thus suggests that the reduced FSHR expression at the transcript level is in concurrence with the expression of FSHR at the protein level in subjects with the AA genotype. The findings of the study carried out in Iran are similar to those results [ 10 ]. These results are in contrast with ours. The reason for the differences is probably in the various distributions of genotypes within different ethnic populations.\nIn our previous study [ 39 ], we analyzed whether we can predict the response to ovarian stimulation using AMH blood level. It was concluded that AMH is an independent and an accurate predictor for poor response on gonadotrophin stimulation. In the present study, AMH values were different among genotype variants. Patients with GG genotype had lower AMH values compared to women with AA and AG genotype in one group ( P =  0.016). It can be concluded that GG genotype is strongly linked with poor response on ovarian stimulation.\nBy analyzing SNPs rs2234693 and rs4986938 in  ESR1  gene, we did not find any difference between genotype variants, neither in distribution of different genotypes between poor- and hyper-responders, nor in the amount of recombinant FSH used for COH, and number of oocytes retrieved. We found a statistically significant difference between genotype of rs2234693 and E2 on hCG day. Patients with CC genotype had lower E2 on hCG day compared to those with CT and TT genotypes. The data revealed that women with CC genotype at the rs2234693 position were classified more often as poor responders compared to women with CT or TT genotype. De Mattos et al. [ 40 ] studied the same SNPs in ESR1 gene. For SNP rs2234693 they did not confirm any influence on ovarian response, while the patients with rs2234693 TT genotype needed a higher dose of rFSH. Their results agree with prior findings by Altmäe et al. [ 16 ], and Ayvaz et al. [ 17 ], who demonstrated better ovarian response in patients with the rs2234693 CC genotype. Similar results have been published by de Castro et al. [ 41 ], who found lower frequency of C allele of SNP rs2234693 in poor responders.\nAnalysis of our data revealed that the women with GG genotype of SNP rs3714664 in  AMHR  gene received a higher amount of exogenous FSH for ovarian stimulation compared to patients with AG genotype. However, we did not find any positive association between these genotypes and ovarian response to gonadotropins or concentrations of basal hormones (bFSH, AMH, and E2 on the day of hCG administration).\nAlso, the polymorphism rs10407022 of the  AMH  gene was not associated either with any of measured hormonal parameters, or with response to ovarian stimulation. This does not confirm the results of some other studies in which such association was found with poor response [ 6 ,  42 ] or with ovarian hyperstimulation syndrome [ 43 ,  44 ]. On the other hand, the present results agree with those of Kerkelä et al. [ 45 ] who found no association between  AMH  coding polymorphisms and OHSS.\nWe can conclude that the GG genotype of SNP rs1394205 in  FSHR  is strongly linked with poor response on ovarian stimulation. Our study and patient classification into three groups was mainly based on the number of oocytes retrieved. Although oocyte yield increases with increasing dose of FSH, availability of blastocysts is less influenced by the rFSH dose and AMH level [ 46 ], and so blastocyst quality may be a more meaningful criterion for patient classification in COS response groups. This has to be considered in subsequent studies. Moreover, some studies in patients with endometriosis [ 47 ] and women of advanced age [ 48 ] have shown the impact of DNA methylation to gene expression in granulosa cells. The extension of such analyses on selected  FSHR ,  AMHR ,  AMH , and  ESR1  genes is needed to confirm whether or not these SNPs could serve as potential biomarkers for prediction of COS outcome.","source_license":"CC-BY-4.0","license_restricted":false}