Article FSH receptor genotype is associated with pregnancy but not with ovarian response in IVF

Transcription

1 RBMOnline - Vol 13 No Reproductive BioMedicine Online; on web 29 September 2006 Article FSH receptor genotype is associated with pregnancy but not with ovarian response in IVF Ellen Klinkert received her MD from the University of Groningen, the Netherlands, in She then worked for two years as a clinician in the field of reproductive medicine. In 1999 she started her PhD research on the prediction and treatment of poor response in IVF at the University Medical Center Utrecht, under the supervision of Professor ER te Velde. She obtained her doctorate with a thesis on this subject in In the same year she started a residency in Obstetrics and Gynaecology at the University of Groningen. Dr Ellen R. Klinkert Ellen R Klinkert 1,6, Egbert R te Velde 1, Sjerp Weima 1, Peter M van Zandvoort 2, Rob GJM Hanssen 2, Philomeen R Nilsson 3, Frank H de Jong 4, Caspar WN Looman 5, Frank JM Broekmans 1 1 Department of Reproductive Medicine, Division of Perinatology and Gynecology, University Medical Center Utrecht, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands; 2 Department of Pharmacology, Section Fertility and Contraception, N.V. Organon, Oss, The Netherlands; 3 Department of Medical Genetics, University Medical Center Utrecht, The Netherlands; 4 Department of Internal Medicine, Erasmus MC, Rotterdam, The Netherlands; 5 Department of Public Health, Erasmus MC, Rotterdam, The Netherlands 6 Correspondence: Fax: ; erklinkert@planet.nl Abstract Two very common single nucleotide polymorphisms at positions 307 and 680 in exon 10 of the FSH receptor gene have been associated with ovarian response in IVF. This observational study evaluated the role of the FSH receptor genotype in the prediction of poor response and clinical pregnancy in IVF in comparison with other markers, such as age, basal FSH, anti- Müllerian hormone and antral follicle count. In addition, the in-vitro camp response towards recombinant FSH in cultured granulosa cells of patients with different FSH receptor genotypes was determined. A total of 105 IVF patients undergoing ovarian stimulation in a long suppression protocol were included in the study. The ovarian response was comparable between patients with different FSH receptor genotypes. Patients with polymorphism Ser/Ser had implantation and pregnancy rates that were three times higher compared with patients with polymorphism Asn/Asn. FSH receptor genotype was not associated with a poor response in IVF, but showed a positive association with pregnancy, independent of age. There was no difference in camp production in cultured granulosa cells of patients with different FSH receptor genotypes (n = 62). It is concluded that FSH receptor genotype is associated with pregnancy in IVF, but not with ovarian response. Keywords: FSH receptor, implantation rate, IVF, polymorphism, poor response, pregnancy rate Introduction In IVF, the ovarian response to exogenous FSH stimulation is quite variable. Several endocrine and ultrasound parameters have been identified that, in concert, give some clue as to the probability of a poor response to gonadotrophin stimulation for IVF (Ng et al., 2000; Bancsi et al., 2002). However, if further factors that influence a patient s response to stimulation can be identified, it may be possible to identify women at risk of experiencing a low ovarian response more accurately. A key publication showed that ovarian response in IVF may be related to the FSH receptor genotype (Perez Mayorga et al., 2000) and, as such, this marker may contribute to predicting which patients will respond poorly. Subsequent studies have either shown confirmation of the relationship between basal FSH, FSH receptor subtype and ovarian response to gonadotrophin stimulation (Sudo et al., 2002; de Castro et al., 2003; de Koning et al., 2006) or have reported findings that shed some doubt on the solidity of this concept (Daelemans et al., 2004; Greb et al., 2005; Loutradis et al., 2006). In general, ovarian response will be determined by patient-related factors, especially the sensitivity of the follicles to FSH and the number of selectable follicles present in the cohort, and doctor-related factors, such as the degree of exposure to FSH and the choice of drugs that are used for ovarian stimulation (Keck et al., 2005). If the FSH receptor subtype is a denominator of follicle sensitivity to FSH, then there may be added value to other predictors that are related to this factor, especially because most other predictors are based on the size of follicle cohort. 687

2 688 The gene that encodes for the FSH receptor is localized on chromosome 2 p21 in man (Rousseau-Merck et al., 1993; Gromoll et al., 1994; Simoni et al., 1997). Mutations in the FSH receptor gene that cause infertility due to ovarian arrest are very rare (Simoni et al., 2002). However, while screening for mutations, two common polymorphisms of the receptor gene were identified, leading to four possible allelic combinations. The first polymorphism is located in the extracellular domain at position 307, where threonine (Thr) or alanine (Ala) is present. The second variant is located at position 680 of the intracellular domain, occupied by either asparagine (Asn) or serine (Ser) (Aittomaki et al., 1995; Simoni et al., 1999). These single nucleotide polymorphisms (SNP) may slightly alter the function of the FSH receptor, leading to a difference in exogenous gonadotrophin requirement during ovarian stimulation in infertility patients with different allelic variants (Perez Mayorga et al., 2000; Sudo et al., 2002). Because Thr307 is almost always in linkage disequilibrium with Asn680, and Ala307 almost always with Ser680, most studies focus solely on position 680. FSH bioactivity is not limited to stimulation of granulosa cell proliferation and differentiation, which are the prerequisites for follicular growth and aromatization of androgen precursors to oestrogens. In-vitro studies in different species have shown that FSH in a coordinate action with LH stimulates extracellular matrix deposition and expansion of the cumulus oophorus as a sign of favourable cytoplasmic maturation of the oocyte (Chen et al., 1994). FSH is also involved in the expression of LH receptors in the granulosa cells during the mid-follicular phase and the differentiation of granulosa cells into luteal cells. The FSH receptor can thus be considered a key factor in several cumulus oocyte processes that are essential for embryonic development as well as establishment and maintenance of early pregnancy (Segaloff et al., 1990; Richards, 1994). These processes include follicle recruitment and growth, attainment of oocyte meiotic and developmental competence and granulosa cell differentiation pathways. To date, no reports on the relationship between FSH receptor polymorphism and implantation rates or pregnancy rates in IVF have been published. Evaluation of the bioactivity of the FSH receptor in vitro showed no significant differences in camp production between the different isoforms (Simoni et al., 1999; Sudo et al., 2002). It is, however, possible that the transfected COS- 7 cells and 293T cells used in these studies insufficiently resembled the natural environment of the receptor. Therefore, characterization of the different FSH receptor types should be re-evaluated in human granulosa cell-based systems. The primary aim of this observational study was to investigate the difference in ovarian response to stimulation with recombinant FSH between the different polymorphic FSH receptor types recently reported in IVF patients (Perez Mayorga et al., 2000; Sudo et al., 2002) and to see whether the results were related to other measures of ovarian response. In addition, possible relationships between the type of polymorphism and the probability of pregnancy in IVF were evaluated. Finally, the in-vitro camp response towards recombinant FSH was determined in granulosa luteal cells cultured from patients carrying different FSH receptor genotypes. Materials and methods Subjects Between April 2002 and January 2004, 105 women undergoing ovarian stimulation for IVF treatment were observed. All patients were asked to participate, irrespective of the IVF cycle number or starting dose of gonadotrophins. Inclusion criteria were a regular menstrual cycle of 25 to 35 days, age <46 years, and the presence of both ovaries. Patients with polycystic ovarian syndrome (PCOS, defined as oligo- or amenorrhoea, elevated LH concentrations with normal FSH concentrations, and/or elevated androgen concentrations) were excluded from this study. Patients undergoing intracytoplasmic sperm injection (ICSI) were also excluded. According to clinic protocol, these patients were pretreated with oral contraceptives, and it was therefore not possible to evaluate their basal hormonal profiles just before the start of treatment. The Institutional Review Board approved this study and all patients gave informed consent. Treatment On day 3 of the cycle prior to the start of the treatment, venous blood samples were obtained from each patient. Serum samples were stored at room temperature for at least 1 h to clot. Within 2 h they were centrifuged at 1700 g together with the plasma samples and stored at 20 C until assayed. On the same day an antral follicle count (AFC) was performed using a Voluson 530D (Kretz Technik, Zipf, Austria) with a 7.5 MHz vaginal transducer. The images were stored for later analysis. One investigator (ERK) performed all the follicle counts and all follicles up to 5 mm were included in the analysis. Patient history and body mass index (BMI) were recorded. On day 21 of the same cycle, patients started down-regulation with 1 mg of leuprolide acetate per day (Lucrin; Abbott, Hoofddorp, The Netherlands). After menstruation, ovarian stimulation with follitropin alpha (Gonal-F; Serono Benelux BV, S-Gravenhage, The Netherlands) was started at a median dose of IU per day ( IU). When necessary, the dose was adapted during stimulation. Follicular development was evaluated by ultrasound and oestradiol measurements. On the day that ovulation was induced by administration of 10,000 IU of human chorionic gonadotrophin (HCG) (Profasi; Serono Benelux BV, S-Gravenhage, The Netherlands), serum oestradiol and FSH were measured and an ultrasound was performed to assess the number of follicles >10 mm. The cycle was cancelled if fewer than three follicles developed. Thirty-six hours after HCG administration, the oocytes were collected. A maximum of two embryos were transferred in patients under 38 years old and three embryos in older patients. HCG or micronized progesterone was used for luteal support (Progestan, Nourypharma BV, Oss, The Netherlands). The IVF protocol has been described in detail elsewhere (van Kooij et al., 1996). Poor ovarian response was defined as the collection of <4 oocytes at oocyte retrieval or cancellation of the cycle when fewer than three follicles had developed. Arguments for this definition have been given in Bancsi et al. (2002) and

3 Klinkert et al. (2004). Pregnancy was defined as a positive pregnancy test 18 days after oocyte retrieval. Hormone assays Basal FSH and FSH on the day of ovulation induction were assayed in serum with the FSH assay on the ADVIA Centaur Automated Chemiluminescence System (Bayer Corporation, Tarrytown, NY, USA). The assay standardization is traceable to World Health Organization (WHO) 2nd International Standard for Human FSH (IS 94/632). The inter-assay coefficients of variation were 6.0% at 5 IU/l and 4.5% at 30 IU/l. All samples used for anti-müllerian hormone (AMH) measurements were analysed simultaneously using an enzyme-immunometric assay (Diagnostic Systems Laboratories Webster, TX, USA). Inter- and intra-assay coefficients of variation were below 5% at the concentration of 3 μg/l, and below 11% at the concentration of 13 μg/l. The detection limit of the assay was μg/l. Repeated freezing and thawing of the samples or storage at 37 C for 1 h did not affect results. A comparison of this assay with the ultrasensitive version of the assay method (Immunotech- Coulter, Marseilles, France) in 82 samples with AMH concentrations between 0 and 15 μg/l yielded a correlation coefficient (r) of The formula of the regression line was: AMH (DSL) = AMH (ic) In order to keep results comparable with earlier published data, all results were multiplied by a factor of 2. The difference between the results of the two assay systems was caused by a difference in the potencies of the two standards, as became apparent after measuring each standard in the other assay. The epitopes in the AMH molecule, against which the antibodies used in the two assays were raised, are not known. If the recombinant AMH used as a standard is recognized in a way that differs from the way serum AMH is recognized for both sets of antibodies, this might contribute to the different results observed. DNA isolation and analysis From each patient, 10 ml blood was drawn with EDTA added as anticoagulant. Genomic DNA was isolated from peripheral blood leukocytes with an automated robot (autopure LS, Gentra Systems, Minneapolis, MN, USA) according to the manufacturer s instructions. Isolated DNA was cryopreserved at -20 C. Aliquots at a concentration of 5 10 ng/μl were prepared for genotyping DNA. Genotyping of SNP, located at positions 919 (exon 10 AA307) and 2039 (exon 10 AA680) of the FSH receptor gene, was performed using the TaqMan allelic discrimination assay, as described by Simoni et al. (2002). A detailed description of the assay, including MGB probe and primer sequence information, can be obtained from M Simoni. Primers and MGB probes were purchased from Applied Biosystems (Foster City, CA, USA). Briefly, 4 μl genomic DNA, 0.75 μl of each primer (10 μm), 0.5 μl of each MGB probe (5 μm), and 12.5 μl universal master mix (Applied Biosystems) was made up with aquadest to a final volume of 25 μl. TaqMan analysis was performed with the ABI Prism 7000 Sequence Detection System (Applied Biosystems) with the following thermocycle profile: 50 C for 2 min, 95 C for 10 min, followed by 92 C for 15 s and 60 C for 1 min for 40 cycles. In-vitro response of cultured granulosa luteal cells Reagents For isolation of the granulosa cells, Dulbecco s modified Eagle s medium (DMEM/HAM F12 END; Invitrogen, Paisley, UK) was supplemented with penicillin (100 IU/ml), streptomycin (0.10 mg/ml) (Invitrogen), bovine pancreas insulin (1 μg/ml anhydrous, HPLC), apo-transferrin (5 μg/ml), L-glutamine (4 mm) and bovine serum albumin (BSA) fraction V (1 mg/ml), all from Sigma-Aldrich (St Louis, MO, USA). For granulosa cell culture, DMEM medium was supplemented with antibiotics (see above) and 10% fetal calf serum (Invitrogen). For FSH stimulation of the granulosa cells, isolation medium was supplemented with 3-isobutyl-1- methylxanthine (IBMX) (0.2 mm) (Sigma-Aldrich) and recombinant FSH (rfsh) (0.5 U/ml, 1 U/ml). rfsh was kindly provided by Organon International (Oss, The Netherlands). Isolation and culture of primary human granulosa cells Follicular fluid was collected as one batch per patient and centrifuged after oocyte removal (425 g for 10 min). The pellet was resuspended in 5 ml collagenase (200 U/mg, end concentration 0.1%) (Sigma-Aldrich) containing isolation medium, layered onto 5 ml Histopaque-1077 (Sigma-Aldrich) and centrifuged (375 g for 20 min) to separate granulosa cells from red blood cells. It is unlikely that these cell preparations contained significant amounts of theca cells, since theca cells are located outside the follicle and therefore, in theory, only incidentally will be aspirated at oocyte retrieval. Primary granulosa cells and other mononuclear cells (e.g. lymphocytes) were collected from the interface layer, washed twice with isolation medium (425 g, 5 min) and resuspended in 2 ml culture medium. Aliquots of the suspensions were examined on a haemocytometer and cell density was adjusted to cell/ml. Viability was determined by Trypan blue exclusion. Granulosa cells were plated at a density of 25,000 viable cells/250 μl/well in culture medium in collagen-coated 96- well plates (Cellcoat, Greiner Bio-One B.V., Alphen aan den Rijn, The Netherlands), and cultured at 37 C in a humidified atmosphere of 5% CO 2 in air. Every 48 h (day 3 and day 5), the wells were washed by removing the culture medium and adding fresh prewarmed culture medium. Seven days after retrieval, the cells were washed again and subjected to 250 μl IBMX containing isolation medium with or without rfsh and cultured for an additional 48 h, whereafter all medium was removed and frozen until assayed for camp. 689

4 690 Determination of camp concentrations camp released from the cultured granulosa cells to the culture medium was assayed by enzyme immunoassay from Amersham Biosciences Europe GmbH (Roosendaal, The Netherlands), according the manufacturer s instructions. Statistics Statistical analysis was performed with Statistics Package for Social Sciences (SPSS) for Windows, version 10.1 (SPSS Inc., Chicago, IL, USA). The Kruskal Wallis test and chi-squared test were used to compare the different receptor types. A P-value < 0.05 was considered statistically significant. Embryo quality was scored by evaluating the following variables: fertilization rate, proportion of 3 pronucleate (PN) zygotes, proportion of embryos suitable for transfer, proportion of growth-retarded embryos on the day of the transfer, and morphological embryo score of the transferred embryos (score 1 = <10% fragmentation; score 2 = 10 50% fragmentation; score 3 = >50% fragmentation). Univariate regression analysis was used to compare the value of the FSH receptor genotype as a predictor of ovarian response and pregnancy in IVF with the known predictors, including age, basal FSH, AMH, AFC, primary or secondary infertility, duration of infertility, cause of infertility, number of previous unsuccessful IVF cycles, previous live birth (as a result of IVF treatment or after a spontaneous conception), previous poor response, number of oocytes in previous IVF cycle and BMI. Results A total of 105 women were prospectively included in this study. For 19 patients, this was their first IVF treatment cycle, 47 patients had one previous unsuccessful cycle, 23 patients had two, and 16 patients had three or more unsuccessful cycles. Eleven patients were unable to visit the hospital to have their blood drawn at cycle day 3 for basal FSH and AMH measurements. In nine patients, the basal FSH concentration could not be determined because the amount of serum was insufficient or lost in the procedure. The missing values were evenly distributed across the receptor subtype groups. The frequencies of the four allelic combinations in this study population were 59% for the Thr/Asn variant, 39% for the Ala/ Ser variant and 2% for the Ala/Asn variant. The variant Thr/ Ser was not present in the study group. The frequencies for polymorphism 307 were 18% Ala/Ala, 45% Ala/Thr and 37% Thr/Thr whereas the frequencies for polymorphism 680 were 38% Asn/Asn, 45% Asn/Ser and 17% Ser/Ser. Table 1 shows the frequency distribution of the FSH receptor genotypes in different female study populations, including this study. In Table 2 the patient characteristics of all the participants in this study are shown. Table 3 shows the outcome of the IVF treatment for the three different FSH receptor genotypes at position 680. In three patients, the cycle was cancelled because of a poor response (two patients with polymorphism Asn/Asn and one patient with polymorphism Ser/Ser). Four patients with <3 follicles decided to proceed with the oocyte retrieval in spite of the poor response: three patients had the Asn/Asn variant and one the Asn/Ser variant. The total amount of gonadotrophins used during stimulation was highest in patients with the Ser/Ser variant, but the amount of gonadotrophins per oocyte collected was lowest in this group. The FSH concentration on the day of HCG administration was slightly higher in the Ser/Ser group. These differences did not reach statistical significance. The basal FSH concentration and the ovarian response did not differ between the three groups. The characteristics of embryo quality that were studied were also comparable (data not shown). However, a significant difference in pregnancy rates and implantation rates was found (see Table 3 for P-values). The pregnancy rate and implantation rate in patients with the Ser/Ser polymorphism were three times higher than in patients with Asn/Asn. Univariate logistic regression analysis showed that basal FSH and a previous live birth after IVF were positively associated with a poor ovarian response (Table 4). AMH, AFC and the number of oocytes collected in the previous IVF cycle were negatively correlated to poor response. The FSH receptor genotype was not associated with poor ovarian response. A second univariate logistic regression analysis, presented in Table 4, was performed to study the association of different patient characteristics with the occurrence of a pregnancy after IVF. This revealed that the FSH receptor polymorphism and the number of oocytes collected in the previous cycle showed a positive association with pregnancy, whereas age was negatively associated with pregnancy. The results in Table 2 show a difference in median age between the three groups. It was therefore decided to investigate whether the effect of the FSH receptor polymorphism on the probability of pregnancy remained after correction for age. After a forced entry of age into the prediction model, the FSH receptor polymorphism was still significantly correlated with pregnancy (odds ratio type Asn/Ser 1.80 [95% CI ]; odds ratio type Ser/Ser 5.60 [95% CI ], P = 0.02), indicating that the FSH receptor polymorphism significantly improved the prediction of pregnancy, given the age of the patient. In 62 patients, granulosa cells could be obtained from the follicular fluid that was collected on the day of the oocyte retrieval. Basal camp production, as well as FSH-stimulated camp production, was measured in the culture medium of the granulosa cells 9 days after retrieval. All three groups showed the same median rate of camp increase after stimulation compared with the basal production (2.1 [95% CI ] for Asn/Asn, 2.0 [95% CI ] for Asn/Ser and 1.8 [95% CI ] for Ser/Ser). There was no difference in camp production between patients who became pregnant as a result of the IVF treatment and patients with an unsuccessful treatment.

5 Table 1. Frequency distribution of FSH receptor genotypes in different female study populations. Reference Subjects No. Country Asn/Asn Asn/Ser Ser/Ser subjects of origin (%) (%) (%) Sudo et al., 2002 Gynaecological patients 522 Japan Conway et al., 1999 PCOS patients 93 United Kingdom Laven et al., 2003 Normogonadotrophic 148 The Netherlands anovulatory infertile women Falconer et al., 2005 Infertile women 68 Sweden de Koning et al., 2006 Subfertile women 78 The Netherlands Perez Mayorga et al., 2000 IVF and ICSI patients 161 Germany de Castro et al., 2003 IVF and ICSI patients 102 Spain Daelemans et al., 2004 IVF patients 130 Belgium Loutradis et al., 2006 IVF and ICSI patients 125 Greece Present study IVF patients 105 The Netherlands PCOS = polycystic ovarian syndrome, ICSI = intracytoplasmic sperm injection. Table 2. Patient characteristics for the different FSH receptor polymorphisms on position 680. Variable All patients Asn/Asn Asn/Ser Ser/Ser (n = 105) (n = 40) (n = 47) (n = 18) Age (years) 36.9 ( ) 37.7 ( ) 36.5 ( ) 35.0 ( ) Basal FSH (IU/l) (n = 85) 5.3 ( ) 5.3 ( ) 5.3 ( ) 5.4 ( ) AMH (μg/l) (n = 94) 0.66 ( ) 0.61 ( ) 0.66 ( ) 0.66 ( ) AFC (2 5 mm) 8 (2 18) 9 (1 26) 8 (2 16) 8 (2 15) Duration of infertility (years) 2.8 ( ) 3.4 ( ) 2.4 ( ) 2.6 ( ) Cause of infertility Tubal 37 (35) 16 (40) 17 (36) 4 (22) Male 42 (40) 14 (35) 19 (41) 9 (50) Unexplained 26 (25) 10 (25) 11 (23) 5 (28) Secondary infertility 80 (67) 24 (60) 33 (70) 13 (72) No. previous unsuccessful IVF cycles 1 (0 4) 1 (0 3) 1 (0 4) 1 (0 6) Previous live birth (spontaneous or after IVF) 44 (42) 17 (43) 17 (36) 10 (56) Previous live birth after IVF 31 (30) 10 (25) 13 (28) 8 (44) Previous poor response (cancel or <4 oocytes) 25 (24) 10 (25) 10 (21) 5 (28) No. oocytes in previous cycle 7 (3 13) 7 (3 13) 7 (3 14) 7 (3 14) Body mass index 23.3 ( ) 23.3 ( ) 23.8 ( ) 22.7 ( ) Values are median (10th 90th percentiles) or numbers (percentages); AMH = anti-müllerian hormone; AFC = antral follicle count. There are no statistically significant differences between the groups. 691

6 Table 3. Outcomes of the IVF treatment for the different FSH receptor polymorphisms on position 680. Variable Asn/Asn Asn/Ser Ser/Ser P-value (n = 40) (n = 47) (n = 18) Duration of stimulation (days) 11 (9 14) 11 (9 14) 11 (7 15) NS a No. follicles 10 mm 9 (2 13) 8 (4 13) 9 (4 12) NS a Oestradiol concentration 5640 ( ) 5930 ( ) 6125 ( ) NS a on day HCG (pmol/l) FSH concentration on day FSH (IU/l) 9.4 ( ) 9.0 ( ) 9.8 ( ) NS a Ampoules of rfsh 27.8 ( ) 30.0 ( ) 39.5 ( ) NS a No. oocytes 7 (2 12) 7 (2 14) 8 (3 13) NS a Ampoules of rfsh per oocyte b 4.3 ( ) 3.9 ( ) 3.5 ( ) NS a Cycle cancellations 2/40 (5) 0/47 (0) 1/18 (6) NS c Poor response 9/40 (23) 10/47 (21) 4/18 (22) NS c Fertilization rate 67 (24 91) 56 (0 100) 70 (11 100) NS a Pregnancies per cycle 8/40 (20) 15/47 (32) 11/18 (61) c Pregnancies per transfer 8/35 (23) 15/37 (41) 11/16 (69) c Implantation rate per embryo 8/62 (13) 16/70 (23) 13/29 (45) c Values are median (10th 90th percentiles) or numbers (percentages); rfsh = recombinant FSH; NS = not statistically significant. a Kruskal Wallis test. b Cancelled patients (n = 3) and one patient with no oocytes at retrieval were not included in this analysis. c Chi-squared test. Table 4. Univariate logistic regression analysis with odds ratios of factors for predicting poor ovarian response and clinical pregnancy in IVF. Variable Poor response Clinical pregnancy Odds ratio (95% CI) P-value Odds ratio (95% CI) P-value Age (per year) 1.11 ( ) NS 0.90 ( ) 0.03 Basal FSH (per IU/l) 1.18 ( ) ( ) NS AMH (per μg/l) 0.01 ( ) < ( ) NS AFC (per follicle) 0.84 ( ) < ( ) NS FSH receptor polymorphism NS 0.01 Asn/Asn Asn/Ser 1.08 ( ) 1.88 ( ) Ser/Ser 1.14 ( ) 6.29 ( ) Primary infertility 1.92 ( ) NS 0.88 ( ) NS Duration of infertility (per year) 0.88 ( ) NS 0.94 ( ) NS Cause of infertility NS NS Tubal Male 3.70 ( ) 0.82 ( ) Unexplained 1.96 ( ) 0.49 ( ) No. previous unsuccessful 0.76 ( ) NS 1.04 ( ) NS IVF cycles Previous live birth 1.91 ( ) NS 0.80 ( ) NS (spontaneous and after IVF) Previous live birth after IVF 3.15 ( ) ( ) NS Previous poor response 2.22 ( ) NS 0.32 ( ) 0.05 (cancel or <4 oocytes) No. oocytes in previous cycle 0.83 ( ) ( ) <0.01 Body mass index 1.05 ( ) NS 1.02 ( ) NS AMH = anti-müllerian hormone; AFC = antral follicle count; NS = not statistically significant. 692

7 Discussion The frequency distribution of the four allelic variants of the FSH receptor in this study group was comparable to the frequencies described in the literature (Table 1). However, the correlation between the FSH receptor genotype and ovarian response, described in previous observations (Perez Mayorga et al., 2000; Sudo et al., 2002; de Castro et al., 2003) could not be confirmed. Perez Mayorga et al. (2000) showed that the ovarian response in IVF patients with different allelic variants on position 680 of the FSH receptor gene was comparable. However, the number of FSH ampoules required for a successful stimulation was significantly higher in patients with the Ser/Ser variant. Also, the basal FSH concentration was significantly higher in this group. Both these findings were confirmed in a study of a group of Japanese women (Sudo et al., 2002), while in another study the relationship between the allelic variants and FSH concentration was present not only for the Ser/Ser variant but also for the heterozygotic Asn/Ser polymorphism (de Koning et al., 2006). From these studies it was finally suggested that in women with the Ser/Ser subtype, the FSH receptor is less sensitive to FSH, creating a higher threshold for follicle growth and thereby increasing the need for higher dosages of exogenous FSH in IVF ovarian stimulation to obtain the same level of response. However, this concept could not be clearly demonstrated in a Ser/Ser group of IVF patients when randomization for two different stimulation dosages was applied (Behre et al., 2005). Other studies have also shed some doubt upon the true relationship between receptor subtype and ovarian response. Indeed, the frequency of the Ser680 allele was found to be greater among poor responders in one study (de Castro et al., 2003). On the other hand, another study described an enrichment in the Ser680 allele in an IVF population suffering from ovarian hyperstimulation syndrome (Daelemans et al., 2004). Also, the relationship between basal FSH concentration, receptor subtype and ovarian response has become confused by recent findings, where the heterozygotic variant showed a relationship to lower FSH concentrations and better response when compared with the two homozygotic subtypes (Loutradis et al., 2006). In this study, patients with the Ser/Ser variant used the highest dose of gonadotrophins and this group also showed the highest FSH concentrations after ovarian stimulation. Neither dose of FSH nor FSH stimulation plasma concentration differed significantly from these findings in patients with the Asn/Asn or Asn/Ser polymorphisms. Therefore, it was not possible to confirm the suggested relationship between the FSH receptor polymorphism and ovarian response to stimulation with gonadotrophins. Because the mean age in the study group was relatively high compared with the mean age of the patients included in the other studies (37 years compared with years), a subgroup analysis of women <37 years was performed. The frequencies for polymorphism 680 in this subgroup were 30% Asn/Asn, 47% Asn/Ser and 23% Ser/Ser. The results in women <37 years did not differ from those in the whole study group. Pregnancy rates and implantation rates were apparently lower in patients <37 years with the Asn/Asn variant, compared with the Asn/ Ser and Ser/Ser variant, but they did not differ significantly (pregnancy rates per cycle 25%, 44% and 42% for the Asn/Asn, Asn/Ser and Ser/Ser groups respectively; implantation rates per embryo 15%, 31% and 32% for the Asn/Asn, Asn/Ser and Ser/ Ser groups respectively). The fact that the difference did not reach statistical significance may be due to the low number of patients included in this subgroup analysis (n = 53). Due to the observational nature of the study, the starting doses of gonadotrophins differed considerably. Most patients included in this study had already experienced ovarian stimulation in the past. The starting dose was adjusted to their needs, based on the results of prior stimulation. It was assumed, therefore, that all the patients were stimulated with an optimal dose of gonadotrophins during this study. A comparison of the effect of the stimulation between patients was made by considering the number of ampoules needed to collect one oocyte and the serum FSH concentration on the day of the HCG injection. Patients with the Ser/Ser variant used the highest total dose of gonadotrophins, yet the amount of rfsh that was needed to collect one oocyte was almost one ampoule less than in patients with the Asn/Asn variant. Although this difference did not reach statistical significance, these results conflict with the findings of Perez Mayorga et al. (2000). However, the amount of human menopausal gonadotrophin needed per oocyte in the IVF patients that were studied by Sudo et al. (2002) was also lower in patients with the Ser/Ser variant compared with patients with the Asn/Asn variant. Two other studies showed no difference in the amount of gonadotrophins used in the three FSH receptor genotypes, but this may be related to the different stimulation protocols that were used in these studies (de Castro et al., 2003; Loutradis et al., 2006). As can be expected from the lack of relationship to the occurrence of poor response, the FSH receptor subtype did not add any useful information to those factors that are known for their ability to predict poor response. In this study, basal FSH, AMH and the AFC were all logistically related to the occurrence of poor response. Poor response prediction can be useful in the management of IVF patients, as a poor response to hyperstimulation can, in general, be regarded a poor prognostic sign. However, not all poor responders truly have this unfavourable profile, as poor response may occur as a result of underdosing of FSH, related to factors like body mass, follicle sensitivity or as a result of a phenomenon known as regression to the mean (Klinkert et al., 2004). If a factor could be uncovered that adequately predicts a poor response due to the latter group of explanations, then cases presenting with a poor response in the first IVF cycle could be divided into those with a small cohort and those with a cohort in need of higher concentrations of FSH. This concept has already been shown by using ovarian reserve tests as a tool to make this discrimination (Popovic-Todorovic et al., 2003; Klinkert et al., 2004) and the FSH receptor variation can be expected to supply information from the other side of the spectrum: a poor responder with a Ser/ Ser polymorphism and normal ovarian reserve tests is expected to respond normally in a second attempt with a higher dose. It is not easy to clarify why the Ser/Ser polymorphism has failed to fulfil this promise. As described above, the concept may somehow prove to be wrong. Also, a recent study has shown that in normal Ser/Ser variant volunteers, menstrual cycle length is increased and FSH concentrations systematically elevated, while the size of the antral follicle cohort is increased compared with Asn/Asn cases (Greb et al., 2005). These findings may implicate that, in the normal interplay between gonadotrophin output and ovarian messaging, the polymorphism does play a role in a subtle way, while in agonist suppression cycles in IVF, 693

8 694 this subtlety becomes hidden by the general use of maximal stimulation dosages of exogenous FSH. In this study, a correlation between the FSH receptor genotype and the occurrence of pregnancy after IVF was found. As far as is known, this finding has not been described in the literature before. Patients in the Ser/Ser group had a chance of conceiving that was three times higher than that of patients in the Asn/Asn group. This finding suggests that the FSH receptor genotype has an effect on the quality of the ovarian response in IVF stimulation, also reflected by the significantly higher implantation rate of the embryos that were transferred in the Ser/Ser group. The process of implantation is determined by endometrial receptivity and the viability of the embryo. As such, the observed differences in implantation rates should be attributed to differences in one of these two factors. If the FSH receptor subtype acts at the level of the endometrium, then it is very unlikely to be a direct effect, as there are no FSH receptors present (Shemesh, 2001). Indirect effects through the corpus luteum should be considered, as well as direct effects at the level of the granulosa oocyte complex. FSH not only stimulates the recruitment and growth of ovarian follicles. During the follicular phase of the menstrual cycle, there is close association between the oocyte and its nurturing granulosa cells in the so-called cumulus oocyte complex (Albertini et al., 2001). FSH can only exert its effect on the oocyte in this complex via FSH receptors expressed solely by the granulosa cells of the growing follicle (Oktay et al., 1997). Studies employing maturation of isolated cumulus oocyte complexes in vitro have shown that the final stage of nuclear maturation can occur independently of FSH but is much more efficient in the presence of a proper dose of FSH (Anderiesz et al., 2000; Calder et al., 2003). This process is clearly mediated by the FSH receptors of the granulosa cells (Calder et al., 2005). In addition, cytoplasmic maturation of the oocyte as determined by extracellular matrix deposition and cumulus expansion is strictly dependent on the coordinated action of FSH and LH (Anderiesz et al., 2000). As a result, more early human preimplantation embryos will develop to the blastocyst stage when they are derived from cumulus oocyte complexes matured in vitro in the presence of FSH and low dose of LH (Anderiesz et al., 2000). FSH, in a co-ordinated fashion with oestradiol, also regulates granulosa cell differentiation leading to the expression of LH receptors involved in ovulation and luteinization (Segaloff et al., 1990; Richards 1994). Again, the camp-dependent signal transduction pathway is involved in mediating all of these responses. This study found no difference in camp response between the different receptor types that would have been indicative of an altered communication between oocyte and granulosa cells. However, FSH also induces genes without camp-responsive elements in their promoter regions (Richards 1994; Gonzalez-Robayna et al., 2000). It can be postulated that different single nucleotide polymorphisms (SNP) activate alternative pathways, thus modulating quality and developmental competence of the human oocyte. Further evaluation of the second-messenger systems in granulosa cells that are stimulated by FSH, such as the mitogen-activated protein kinase pathway, protein kinase B and intracellular free Ca 2+, could reveal the nature of these alternative signal transduction pathways. A recent study demonstrated that protein kinase B is an essential component of the FSH-mediated granulosa cell differentiation (Zeleznik et al., 2003). Another study identified an alternative intracellular enzyme target that was able to mimic the effect of FSH on follicular growth without involvement of a direct increase in camp (Palmer et al., 2005). The findings of this study need to be confirmed in a large prospective trial. If the association between the FSH receptor genotype and pregnancy rate in IVF is valid, genetic background might serve as an additional marker to identify patients likely to have a favourable prognosis in IVF. Acknowledgements The authors thank Erna Italiaander-Kwakkel for expert technical assistance in culturing the granulosa cells and Dr Eef Lentjes for basal FSH measurements. References Aittomaki K, Lucena JL, Pakarinen P et al Mutation in the follicle-stimulating hormone receptor gene causes hereditary hypergonadotropic ovarian failure. Cell 82, Albertini DF, Combelles CM, Benecchi E et al Cellular basis for paracrine regulation of ovarian follicle development. Reproduction 121, Anderiesz C, Ferraretti A-P, Magli C et al Effect of recombinant human gonadotrophins on human, bovine and murine oocyte meiosis, fertilization and embryonic development in vitro. Human Reproduction 15, Bancsi LF, Broekmans FJ, Eijkemans MJ et al Predictors of poor ovarian response in in-vitro fertilization: a prospective study comparing basal markers of ovarian reserve. Fertility and Sterility 77, Behre HM, Greb RR, Mempel A et al Significance of a common single nucleotide polymorphism in exon 10 of the folliclestimulating hormone (FSH) receptor gene for the ovarian response to FSH: a pharmacogenetic approach to controlled ovarian hyperstimulation. Pharmacogenetics and Genomics 15, Calder MD, Caveney AN Sirard MA, Watson AJ 2005 Effect of serum and cumulus expansion on marker gene transcripts in bovine cumulus oocyte complexes during maturation in vitro. Fertility and Sterility 83, Calder MD, Caveney AN, Smith LC et al Responsiveness of cumulus oocyte complexes (COC) to porcine and recombinant human FSH, and the effect of COC quality on gonadotropin receptor and Cx43 marker gene mrnas during maturation in vitro. Reproductive Biology and Endocrinology 11, Chen L, Russell PT, Larsen WJ et al Sequential effects of follicle stimulating hormone and luteinizing hormone on mouse cumulus expansion in vitro. Biology of Reproduction 51, Conway GS, Conway E, Walker C et al Mutation screening and isoform prevalence of the follicle stimulating hormone receptor gene in women with premature ovarian failure, resistant ovary syndrome and polycystic ovary syndrome. Clinical Endocrinology 51, Daelemans C, Smits G, de Maertelaer V et al Prediction of severity of symptoms in iatrogenic ovarian hyperstimulation syndrome by follicle-stimulating hormone receptor Ser680Asn polymorphism. Journal of Clinical Endocrinology and Metabolism 89, de Castro F, Ruiz R, Montoro L et al Role of follicle-stimulating hormone receptor Ser680Asn polymorphism in the efficacy of follicle-stimulating hormone. Fertility and Sterility 80, de Koning CH, Benjamins T, Harms P et al The distribution of FSH receptor isoforms is related to basal FSH levels in subfertile women with normal menstrual cycles. Human Reproduction 21,

9 Falconer H, Andersson E, Aanesen A et al Follicle-stimulating hormone receptor polymorphisms in a population of infertile women. Acta Obstetricia et Gynecologica Scandinavica 84, Gonzalez-Robayna IJ, Falender AE, Ochsner S et al Folliclestimulating hormone (FSH) stimulates phosphorylation and activation of protein kinase B (PKB/Akt) and serum and glucocorticoid-lnduced kinase (Sgk): evidence for A kinaseindependent signaling by FSH in granulosa cells. Molecular Endocrinology 14, Greb RR, Grieshaber K, Gromoll J et al A common single nucleotide polymorphism in exon 10 of the human follicle stimulating hormone receptor is a major determinant of length and hormonal dynamics of the menstrual cycle. Journal of Clinical Endocrinology and Metabolism 90, Gromoll J, Ried T, Holtgreve-Grez H et al Localization of the human FSH receptor to chromosome 2 p21 using a genomic probe comprising exon 10. Journal of Molecular Endocrinology 12, Keck C, Bassett R, Ludwich M 2005 Factors influencing response to ovarian stimulation. Reproductive BioMedicine Online 11, Klinkert ER, Broekmans FJ, Looman CW et al A poor response in the first in-vitro fertilization cycle is not necessarily related to a poor prognosis in subsequent cycles. Fertility and Sterility 81, Laven JS, Mulders AG, Suryandari DA et al Folliclestimulating hormone receptor polymorphisms in women with normogonadotropic anovulatory infertility. Fertility and Sterility 80, Loutradis D, Patsoula E, Minas V et al FSH receptor gene polymorphisms have a role for different ovarian response to stimulation in patients entering IVF/ICSI-ET programs. Journal of Assisted Reproduction and Genetics 23, Ng EH, Tang OS, Ho PC 2000 The significance of the number of antral follicles prior to stimulation in predicting ovarian responses in an IVF programme. Human Reproduction 15, Oktay K, Briggs D, Gosden RG 1997 Ontogeny of follicle-stimulating hormone receptor gene expression in isolated human ovarian follicles. Journal of Clinical Endocrinology and Metabolism 82, Palmer SS, McKenna S, Arkinstall S 2005 Discovery of new molecules for future treatment of infertilty. Reproductive BioMedicine Online 10 (Suppl. 3), Perez Mayorga M, Gromoll J, Behre HM et al Ovarian response to follicle-stimulating hormone (FSH) stimulation depends on the FSH receptor genotype. Journal of Clinical Endocrinology and Metabolism 85, Popovic-Todorovic B, Loft A, Bredkjaeer HE et al A prospective randomized clinical trial comparing an individual dose of recombinant FSH based on predictive factors versus a standard dose of 150 IU/day in standard patients undergoing IVF/ICSI treatment. Human Reproduction 18, Richards JS 1994 Hormonal control of gene expression in the ovary. Endocrine Reviews. 15, Rousseau-Merck MF, Atger M, Loosfelt H et al The chromosomal localization of the human follicle-stimulating hormone receptor gene (FSHR) on 2p21-p16 is similar to that of the luteinizing hormone receptor gene. Genomics 15, Segaloff DL, Wang HY, Richards JS 1990 Hormonal regulation of luteinizing hormone/chorionic gonadotropin receptor mrna in rat ovarian cells during follicular development and luteinization. Molecular Endocrinoology 4, Shemesh M 2001 Actions of gonadotrophins on the uterus. Reproduction 121, Simoni M, Nieschlag E, Gromoll J 2002 Isoforms and single nucleotide polymorphisms of the FSH receptor gene: implications for human reproduction. Human Reproduction Update 8, Simoni M, Gromoll J, Hoppner W et al Mutational analysis of the follicle-stimulating hormone (FSH) receptor in normal and infertile men: identification and characterization of two discrete FSH receptor isoforms. Journal of Clinical Endocrinology and Metabolism 84, Simoni M, Gromoll J, Nieschlag E 1997 The follicle-stimulating hormone receptor: biochemistry, molecular biology, physiology, and pathophysiology. Endocrine Reviews 18, Sudo S, Kudo M, Wada S et al Genetic and functional analyses of polymorphisms in the human FSH receptor gene. Molecular Human Reproduction 8, van Kooij RJ, Looman CW, Habbema JD et al Age-dependent decrease in embryo implantation rate after in-vitro fertilization. Fertility and Sterility 66, Zeleznik AJ, Saxena D, Little-Ihrig L 2003 Proteine Kinase B is obligatory for follicle-stimulating hormone-induced granulosa cell differentiation. Endocrinology 144, Received 6 July 2005; refereed 18 July 2005; accepted 2 August