ORIGINAL ARTICLE Reproductive genetics

Genetic polymorphisms of GnRH andgonadotrophic hormone receptorsaffect the phenotype of polycysticovary syndromeO. Valkenburg1,5, A.G. Uitterlinden2,3, D. Piersma2, A. Hofman3,A.P.N. Themmen2, F.H. de Jong2, B.C.J.M. Fauser4, and J.S.E. Laven1

1Division of Reproductive Medicine, Department of Obstetrics and Gynaecology, Erasmus MC, University Medical Center, Room HS508,P.O. Box 2040, 3000 CA Rotterdam, The Netherlands 2Department of Internal Medicine, Erasmus MC, University Medical Center,Rotterdam, The Netherlands 3Department of Epidemiology, Erasmus MC, University Medical Center, Rotterdam, The Netherlands4Department of Reproductive Medicine and Gynaecology, University Medical Center, Utrecht, The Netherlands

5Correspondence address. Tel: þ31-10-4633571; E-mail: o.valkenburg@erasmusmc.nl

background: Polycystic ovary syndrome (PCOS) is a complex genetic disorder. Multiple functional polymorphisms have been ident-ified in genes that regulate the hypothalamic–pituitary–gonadal (HPG) axis that regulates ovarian function. The present study aims toexamine the influence of genetic variants of the HPG-axis on the severity of clinical features of PCOS and disease susceptibility.

methods: We included 518 Caucasian PCOS women and 2996 unselected controls from the general population (the Rotterdam study).Genotype distributions were compared between patients and controls. Subsequently, associations with clinical features of PCOS werestudied. Single nucleotide polymorphisms were selected in GnRH (Trp16Ser [rs6185]), the FSH-receptor (FSHR, Ala307Thr [rs6165] andAsn680Ser [rs6166]) and the LH-receptor (18insLQ, Asn291Ser [rs12470652] and Ser312Asn [rs2293275]).

results: FSHR Ser680 was associated with higher levels of gonadotrophic hormones (FSH: P , 0.01, LH: P ¼ 0.01), and testosterone(P ¼ 0.05) and a higher frequency of hyperandrogenism (P ¼ 0.04). No differences in risk for PCOS in association with the FSH-receptorvariants were observed.

conclusion: Genetic variants of the HPG-axis were associated with a modest but significant effect on the phenotype of PCOS. FSHRvariants were strongly associated with the severity of clinical features of PCOS, such as levels of gonadotrophic hormones and the presenceof hyperandrogenism, but not disease risk.

Key words: polycystic ovary syndrome / FSH receptor / LH receptor / GnRH / polymorphism

IntroductionPolycystic ovary syndrome (PCOS) is a common endocrinopathy thatoccurs in 5–8% of women of reproductive age (Azziz et al., 2004).PCOS constitutes a syndrome of ovarian dysfunction characterizedby anovulation, hyperandrogenism and polycystic ovary (PCO) mor-phology. PCOS is associated with alterations in the function of thehypothalamic–pituitary–gonadal (HPG) axis that may result fromincreased frequency and amplitude of the hypothalamic GnRH pulsegenerator (Dalkin et al., 1989). Changes in the secretion of gonado-trophic hormones consist of elevated LH levels in combination withnormal serum FSH concentrations (Yen et al., 1970). However, theextent to which changes in the HPG-axis contribute to the

pathogenesis of PCOS is not fully understood. Although an intrinsicabnormality of the hypothalamic GnRH pulse generator in womenwith PCOS has been proposed in the past (Zumoff et al., 1983;Venturoli et al., 1992), more recent evidence suggests that thisshould be regarded as a secondary phenomenon due to disturbedfeedback at the level of the hypothalamus (Chhabra et al., 2005).Notwithstanding the uncertainty with regard to the role of theHPG-axis in the pathogenesis of PCOS, this pathway does play acentral role in the pathophysiology of the syndrome. Therefore, wehypothesized that common genetic variants of the HPG-axis mayaffect the phenotype of PCOS and possibly disease susceptibility.

A number of functional single nucleotide polymorphisms (SNPs)have been described in genes that are involved in the HPG-axis.

& The Author 2009. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.For Permissions, please email: journals.permissions@oxfordjournals.org

Human Reproduction, Vol.1, No.1 pp. 1–9, 2009

doi:10.1093/humrep/dep113

Hum. Reprod. Advance Access published April 29, 2009 at U

niversity of New

castle on September 3, 2014

http://humrep.oxfordjournals.org/

Dow

nloaded from

GnRH1 is an important candidate gene for delayed puberty andidiopathic hypogonadotrophic hypogonadism. Thus far, no majordefects within GnRH1 have been found in these patients (Sedlmeyeret al., 2005; Vagenakis et al., 2005). However, a polymorphism inthe first exon of GnRH1 has been described, constituting an aminoacid variation at codon 16 (Trp16Ser). This genetic variant has beenexamined recently in two studies focusing on clinical end-points inrelation to altered estrogen exposure. Rather contradictory, theresults of these studies were consistent with both decreased exposureto endogenous estrogens (decreased bone mineral density Iwasakiet al., 2003), as well as increased estrogen exposure (shorter disease-free survival in breast cancer patients Piersma et al., 2007a). However,direct assessment of estrogen exposure was not reported in eitherstudy.

A well-known combination of two polymorphisms in the FSHreceptor (FSHR) gene has been of particular interest with regard toPCOS. Both polymorphisms, which are in near total linkage-disequilibrium, have been identified as coding SNPs at codon-positions307 and 680 in exon 10 (Simoni et al., 1999). This exon covers thesignal transducing transmembrane domain. It was shown that the pres-ence of the minor allele at position 680 (Ser680) is associated with sig-nificantly higher basal FSH levels and altered response to ovarianstimulation using exogenous FSH for IVF (Perez Mayorga et al.,2000). A subsequent comparison of allele frequencies in normogona-dotrophic anovulatory patients and controls showed a higher fre-quency of Ser680 in anovulatory subjects (Laven et al., 2003).However, the presence of this allele did not negatively affect thechances of success for ovulation induction using recombinant FSH.Apparently, although the Ser680 allele is associated with a less respon-sive FSH receptor, its influence can be overcome relatively easily bythe administration of exogenous FSH.

Like FSH, multiple polymorphisms have been described in the LHreceptor (LHR) that are associated with altered LHR functionality.LHR is overexpressed in theca cells from PCOS patients (Jakimiuket al., 2001). LH promotes the secretion of androgens by ovariantheca cells, which may result in follicular maturation arrest (Lavenet al., 2002). An insertion of two amino acids (leucine [L] and gluta-mine [Q]) in the signal peptide of LHR (18insLQ) was shown in vitroto result in increased receptor activity (Piersma et al., 2006). The18insLQ insertion polymorphism was also associated with shorterdisease-free survival in breast cancer patients (Powell et al., 2003;Piersma et al., 2006). In addition, exon 10 of LHR contains twocoding SNPs that cause a change in amino acids (Asn291Ser andSer312Asn). Although in vitro the Asn291Ser polymorphism wasassociated with increased receptor sensitivity, this variant was notassociated with tumor characteristics or survival of breast cancerpatients (Piersma et al., 2007b). A slightly higher frequency of theAsn312 allele was noted among breast cancer patients, possiblybecause of increased action of ovarian steroid hormones (Piersmaet al., 2007b). In a different study in men, a lower frequency of theAsn312 allele was found in association with impaired spermatogenesis(Simoni et al., 2008b), further substantiating the hypothesis ofincreased receptor activity in association with this polymorphism.

The present analysis of functional polymorphisms of the HPGpathway in patients with PCOS examines the extent to which thesesubtle genetic variations affect the severity of clinical features ofPCOS and disease susceptibility.

Subjects and Methods

Subjects and phenotypingAnovulatory subjects attended our infertility outpatient clinic between1994 and 2004. Inclusion criteria were oligomenorrhea (intervalbetween consecutive menstrual periods .35 days) or amenorrhea(absence of vaginal bleeding for at least 6 months) and serum FSH con-centrations within normal range (1–10 IU/l) (van Santbrink et al., 1997;Schipper et al., 1998). The diagnosis of PCOS was established on thebasis of the 2003 European Society for Human Reproduction andEmbryology/American Society for Reproductive Medicine Rotterdamcriteria (2004). In agreement with these criteria hyperandrogenismwas defined as the presence of either biochemical or clinical signs ofandrogen excess. For the purpose of this study clinical hyperandrogen-ism was assessed by means of the Ferriman Gallway (FG) score, andwas defined as FG-score �8. Biochemical hyperandrogenism wasdetermined by calculation of the free androgen index (FAI) as: (Testo-sterone [nmol/l]/sex hormone-binding globulin [(SHBG)nmol/l]) *100. A cutoff level of 4.5 was used for the definition of hyperandrogen-ism (van Santbrink et al., 1997). The presence of PCO was detected byvaginal ultrasound examination. PCO was defined as the presence of�12 follicles in one or both ovaries, and/or increased ovarianvolume (.10 ml) (Balen et al., 2003). Ethnicity and country of birthwere registered. Exclusion criteria were non-Caucasian ethnic originand/or the presence of related disorders with similar clinical presen-tation, such as congenital adrenal hyperplasia and Cushing’s syndrome.

Controls were derived from the Rotterdam study. The design ofthis study has been described previously (Hofman et al., 1991,2007). In short, this is a single-center, prospective, population basedstudy of determinants of chronic disabling diseases in the elderly,aged 55 years and over (n ¼ 7012). Participants from the Rotterdamstudy derive from a specific area near Rotterdam (Ommoord) thatconstitutes a homogeneous population-based sample of Caucasianelderly men and women. All women with age at onset of menopause.45 years and available DNA (n ¼ 2996) were included in thepresent analysis. Written informed consent was obtained from all ano-vulatory patients as well as controls. This study was approved by theinstitutional review board at the ErasmusMC Medical Center.

HormonesAnovulatory patients underwent a standardized initial screening thatwas performed after an overnight fasting period on a randomcycle-day between 9 a.m. and 11 a.m. Clinical examination includeda structured interview and physical examination. Transvaginal ultraso-nography was performed in order to assess ovarian volume and folliclecount for both ovaries. Blood samples were drawn on the day of clini-cal examination and processed within 2 h after withdrawal. Serum wasstored at 2208C until assayed. Endocrine evaluation included serumlevels of gonadotrophic hormones (LH, FSH) and estradiol (E2), tes-tosterone, androstenedione (AD), dehydroepiandrosterone (DHEA)and dehydroepiandrosterone sulfate (DHEAS), progesterone and17-hydroxyprogesterone (17-OHP), SHBG, fasting glucose andinsulin, thyrotropin (TSH) and prolactin. Hormone assays have beendescribed in detail elsewhere (Imani et al., 1998). LH, FSH, TSH,SHBG, progesterone, AD, DHEA, prolactin and insulin weremeasured by immunoradiometric assay (Immulitew platform,

2 Valkenburg et al.

at University of N

ewcastle on Septem

ber 3, 2014http://hum

rep.oxfordjournals.org/D

ownloaded from

Diagnostic Products Corporation, Breda, the Netherlands). Testoster-one and E2 were determined by radioimmunoassay (RIA, DiagnosticProducts Corporation) and 17-OHP was determined using anin-house assay. Intra- and inter-assay coefficients of variation were,5 and ,15% for LH, ,3 and ,8% for FSH, ,3 and ,5% for tes-tosterone, ,8 and ,11% for AD, ,5 and ,7% for E2 and ,4 and,5% for SHBG, respectively. Anti-Mullerian hormone (AMH) levelswere determined in a subgroup of 354 patients using an in-housedouble-antibody enzyme-linked immunosorbent assay (ELISA). Intra-and inter-assay coefficients of variation were ,10 and ,5%, respect-ively. Glucose levels were measured using a Hitachiw 917 analyzer(Roche Diagnostics, Almere, the Netherlands).

GenotypingGenomic DNA was extracted from peripheral venous blood accordingto standard procedures. Genotypes were determined using theTaqman allelic discrimination assay. The Assay-by-Design service(www.appliedbiosystems.com) was used to set up a Taqman allelicdiscrimination assay for the FSHR Asn680Ser, GnRH1 Ser16Trp,LHR Asn291Ser and LHR Ser312Asn polymorphisms. Rs numbers,primer and probe sequences are provided in Table I. The PCR reac-tion mixture included 2 ng of genomic DNA in a 2 ml volume and thefollowing reagents: probes (200 nM), primers (0.9 mM), 2 � TaqmanPCR master mixes (ABgene, Epsom, UK). PCR cycling reactionswere performed on an ABI 9700 PCR system (Applied BiosystemsInc., Foster City, CA, USA) and consisted of initial denaturation at958C (15 min), and 40 cycles with denaturation (15 s at 958C) andannealing and extension (60 s at 608C).

In addition, all PCOS patients and a sub-set of 2419 controls weregenotyped for a six base pair insertion polymorphism in exon 1 of LHR(LHR insLQ). Exon 1 was amplified as described by Atger and col-leagues (1995) using a 50-hexachlorofluorescein-labeled forwardprimer. Separation and sizing of the PCR fragments and assignmentof LHR insLQ genotype was performed on an ABI Prism 3100 auto-mated capillary DNA sequencer using Genescan and Genotyper soft-ware packages (Applied Biosystems Inc., Foster City, CA, USA).

Statistical analysisGenotype and allele frequencies were determined for all polymorph-isms and subsequently tested for Hardy–Weinberg equilibrium(HWE). Calculation of linkage disequilibrium (D0) and correlation(r2) between multiple SNPs in the same gene was performed usingthe EMLD software package (https://epi.mdanderson.org/,qhuang/Software/pub.htm). LHR haplotypes were inferred on the basis ofBayesian linkage disequilibrium analyses (Stephens et al., 2001).

Genotype frequency comparisons were conducted using logisticregression analysis. P � 0.05 was considered statistically significant.Odds ratios (OR) and 95% confidence intervals (CI) were calculatedto assess risk. For the cross-sectional analysis of anthropometricdata, medians and ranges were computed and compared betweenthe different genotypes. Variables were checked for normal distri-butions with the one-sample Kolmogorov–Smirnov test and log-transformed when necessary. Analysis of variance was used to testfor differences between genotype groups and Bonferroni’s correctionwas used to adjust for the number of SNPs tested. In order to test forallele-dose effects, the between group variation was tested for linearassociation. Statistical analysis was performed using a commerciallyavailable software package (Statistical Package for the Social Sciencesversion 12; SPSS Inc., Chicago, USA).

Results

SubjectsFrom a total of 580 normogonadotrophic anovulatory women, 518women were diagnosed with PCOS. Hyperandrogenism was presentin 51% of all anovulatory women although PCO was present in 81%of subjects. Baseline characteristics, endocrine and ultrasound par-ameters of the study group are shown in Table II.

GenotypingAll polymorphisms were in HWE within the PCOS population andcontrols, except for LHR Asn291Ser in the control population. Homo-zygosity for the presence of the minor allele (Ser291) at this locus did

.............................................................................................................................................................................................

Table I Primers and probe sequences used in the study of polymorphisms in PCOS women and controls

Gene variant Rs number PCR primers Taqman probes

GNRH1

Ser16Trp rs6185 Fw AATTCAAAAACTCCTAGCTGGCCTTA VIC CACGCACCAAGTCA

Rv CATAGGACCAGTGCTGGCT FAM ACGCACGAAGTCA

FSHR

Ala307Thr rs6165 Fw GCAACAAATCTATTTTAAGGCAAGAAGTTGA VIC TGACCCCTAGTCTGAGTC

Rv TGTCTTCTGCCAGAGAGGATCT FAM ACCCCTAGCCTGAGTC

Asn680Ser rs6166 Assay on demand (Applied Biosystems, C_2676874_10)

LHR

Asn291Ser rs12470652 Fw CTGAAGTCCAAAAGCTCAAATGCT VIC CAGACAGAATTTTTC

Rv TGTGCTTTCACATTGTTTGGAAAAGT FAM CAGACAGAGTTTTTC

Ser312Asn rs2293275 Fw TTTTCCAAACAATGTGAAAGCACAGT VIC TTACAGTGTTTTGTTATTCACTT

Rv GATACGACTTCTGAGTTTCCTTGCA FAM CAGTGTTTTGTTACTCACTT

FSHR: FSH receptor, LHR: LH receptor.

Gene polymorphisms and polycystic ovary syndrome 3

at University of N

ewcastle on Septem

ber 3, 2014http://hum

rep.oxfordjournals.org/D

ownloaded from

not occur in PCOS patients and in only 16 controls. As the chi-squaretest is more prone towards type I errors in case of very low minorallele frequencies (MAF), HWE was recalculated with an exact test(Wigginton et al., 2005) using the Pedstat software package (Wiggin-ton and Abecasis, 2005). Again the LHR Asn291Ser polymorphismwas out of HWE (P ¼ 0.006). Consequently, this polymorphism wasnot used for the comparison of allele frequencies and haplotype distri-butions in PCOS cases and controls.

GnRHThe distribution of the Trp16Ser alleles of GnRH1 was similar in PCOScases and controls. Results for the comparison of genotype frequen-cies in PCOS cases and controls are summarized in Table III.

GnRH1 Trp16Ser was not associated with gonadotrophic hormonelevels. The influence of this polymorphism and others on the pheno-type of PCOS patients was primarily tested using a model for allele-dose effects. In this way, no significant associations were observedfor this polymorphism. However, a somewhat more benign phenotypewas noted in carriers (homozygous or heterozygous) of the minorallele (Ser16) compared with non-carriers, showing 5.3% lower levelsof testosterone (1.8 versus 1.9 nmol/l, P � 0.01), lower FAI (4.7versus 5.4, P � 0.01), lower fasting insulin levels (54 versus65 pmol/l, P � 0.01) and lower follicle count (18 versus 19, P ¼0.05). No differences in the frequency of PCO or hyperandrogenismwere observed among the various genotype groups (Table IV).

FSH receptorIn 399 PCOS cases, linkage disequilibrium (D0) for the two SNPs inexon 10 of FSHR (Ala307Thr and Asn680Ser) was 0.98 (r2 ¼ 0.94),indicating near-complete linkage disequilibrium. All remaining PCOScases and controls were genotyped for the Asn680Ser polymorphismonly. The frequency of the minor allele (Ser680) and the distribution ofgenotypes at this locus were similar in women with PCOS and con-trols (Table III).

There was a strong association of this FSHR polymorphism withphenotypic characteristics of PCOS patients. The Ser680 allele wasassociated with higher levels of gonadotrophic hormones: FSH(0.6 IU/l increase per allele copy, P � 0.01) and LH (1.1 IU/l increaseper allele copy, P � 0.01) (Table IV). Likewise, the Ser680 allele wasassociated with a higher frequency of hyperandrogenism (OR [perminor allele copy] 1.41 [95% CI 1.09–1.82], P ¼ 0.04) ranging from53.7% among non-carriers to 70.2% among homozygous carriers ofthe Ser680 allele.

LH receptorMAF for the insertion polymorphism at codon 18 of exon 1 (insLQ) inPCOS subjects and controls were 24.9 and 28.0%, respectively (P ¼0.04). Genotype frequency comparisons revealed a significant negativeassociation of the minor allele (LHR 18insLQ) with PCOS, showing15% lower risk for PCOS per minor allele copy (OR 0.85 [95% CI0.73–0.99], P ¼ 0.04). However, the slightly lower frequency of the18 insLQ insertion polymorphism among PCOS cases comparedwith controls was not significant after Bonferroni’s correction forthe total number of SNPs that were tested. Among women withPCOS, homozygous carriers of the minor allele showed 24.1%lower levels of E2 (179 versus 236 pmol/l, P ¼ 0.01), in combinationwith 21.3% lower ovarian volume (7.0 versus 8.9 ml, P ¼ �0.01) com-pared with the other genotypes.

The frequency of the minor allele of LHR Ser312Asn (Asn312) wassimilar in PCOS patients and controls. A significant trend towardshigher FSH levels in carriers of the Asn312 allele did not persist after cor-rection for the total number of polymorphic variants (P ¼ 0.10). Asmultiple polymorphisms were genotyped within the LHR gene, weexplored the possibility that specific combinations of allelic variantsmay have a more pronounced influence on phenotype. To thispurpose, the allele frequencies of four possible LHR haplotypes werecalculated in PCOS cases, i.e. H1 (nonLQ18/Ser312, 50.4%), H2(nonLQ18/Asn312, 23.6%), H3 (insLQ18/Ser312, 9.3%) and H4(insLQ18/Asn312, 16.7%). The distribution of LHR haplotypes wassimilar in women with PCOS and controls. The presence of H1,which is characterized by the absence of polymorphic variants, wasassociated with lower levels of FSH. Median FSH levels were 5.1 IU/l(complete absence of H1), 4.5 IU/l (one copy of H1) and 4.4 IU/l (2copies of H1) (P for allele dose effect ¼ 0.02). Linear regression analysisshowed no additional benefit for the use of haplotype 1 over LHRSer312Asn genotype in predicting FSH levels (P ¼ 0.69). No otherassociations were observed for the presence of haplotypes one tofour with the phenotype (including LH levels) of PCOS patients.

Interaction of FSHR and LHR polymorphismsBoth genetic variants of FSHR (Ser680) and LHR (Asn312) showed evi-dence of an association with higher FSH levels. Therefore, the

........................................................................................

Table II Clinical and endocrine parameters of 518Caucasian women with PCOS

Median Interquartile range

Baseline characteristics and ultrasound parameters

Age (years) 28.7 (25–31.7)

BMI (kg/m2) 26.2 (22.4–31.2)

Mean number of folliclesa 18 (13–25)

Mean ovarian volume (ml)a 8.8 (6.5–11.6)

Endocrine parameters

LH (IU/l) 7.6 (4.9–11.4)

FSH (IU/l) 4.9 (3.6–6.4)

Estradiol (pmol/l) 231 (169–345)

Progesterone (nmol/l) 1.6 (1.0–2.9)

17 (OH) Progesterone (nmol/l) 2.6 (1.9–4.0)

Testosterone (nmol/l) 1.9 (1.4–2.4)

SHBG (nmol/l) 37 (25–57)

FAI 5.0 (2.9–8.2)

AD (nmol/l) 11.8 (9.2–15.1

DHEA (nmol/l) 40 (28–59)

DHEAS (mmol/l) 5.1 (3.6–7.1)

Glucose (mmol/l) 4.0 (3.7–4.4)

Fasting insulin (pmol/l) 58 (38–91)

Clinical and endocrine parameters of 518 Caucasian women with PCOS.AD: androstenedione; SHBG: sex hormone-binding globulin; DHEA:dehydroepiandrosterone; DHEAS: dehydroepiandrosterone sulfate; FAI: free androgenindex.aMean of left and right ovary.

4 Valkenburg et al.

at University of N

ewcastle on Septem

ber 3, 2014http://hum

rep.oxfordjournals.org/D

ownloaded from

hypothesis was explored that an interaction between both genetic var-iants may have a more distinct effect on the phenotype. To thispurpose the PCOS population was stratified into five subgroups, i.e.carriers of zero to four polymorphic alleles (FSHR Ser680 or LHRAsn312). Median FSH levels rose from 3.9 IU/l in women with no poly-morphic variants (n ¼ 45) to 6.0 IU/l in women who were homozy-gous carriers of the variant allele at both polymorphic loci. The totalnumber of variant alleles was significantly associated with increasingFSH levels (P � 0.01) (Fig. 1). Carriership of zero to four allelic variantswas equally distributed among PCOS women and women from thecontrol population (X2 test, P ¼ 0.34) with an average of two allelicvariants per individual (both PCOS women and controls).

DiscussionThe present study compares the presence of genetic variants of theHPG-axis in Caucasian PCOS patients and unselected controls. Allpolymorphisms, except LHR insLQ, were equally distributed amongcases and controls. A 40% decrease in risk for PCOS resulted fromthe homozygous presence of a 6-nucleotide insertion polymorphism

at codon 18 in exon 1 of the LHR gene. Among PCOS cases, homo-zygous carriers of this genetic variant also showed 24% lower E2 levelsand 21% lower estimates of ovarian volume. Likewise, multiple associ-ations were identified between genetic variants of the HPG-axis andthe phenotype of PCOS patients, including gonadotrophic hormonelevels as well as the presence of hyperandrogenism.

The primary aim of the current study was to identify risk allelesfor PCOS. With the possible exception of the LHR 18insLQ poly-morphism, we observed no differences in the frequency of thesepolymorphisms in PCOS cases and controls. Contrary to what wasexpected, we observed a lower frequency of the LHR 18insLQ poly-morphism among PCOS patients. However, this finding was not sig-nificant after Bonferroni’s correction and should therefore beinterpreted with caution. No other differences were observed inallele frequencies and/or genotype distributions in PCOS casesand controls. In contrast to a prior report (Laven et al., 2003), wewere unable to confirm a difference in the distribution of theFSHR genotypes in PCOS cases and controls. This discrepancymay originate from the limited number of controls (n ¼ 30) thatwere included in the former study.

........................... .............................

.............................................................................................................................................................................................

Table III Distribution of genotypes and MAF in 518 Caucasian PCOS women and 2996 healthy controls

Genotypes PCOS Control OR (95% CI) Pb

N % N %

GNRH1 16

Trp/Trp 282 54.9 1570 53.3 1 (reference)

Trp/Ser 187 36.4 1172 39.8 0.89 (0.73–1.09) 0.25

Ser/Ser 45 8.8 202 6.9 1.24 (0.88–1.78) 0.22

MAF (Ser) 26.9 26.8 1.01 (0.87–1.17) 0.91

FSHR 680

Asn/Asn 123 24.8 782 26.9 1 (reference)

Asn/Ser 248 50.1 1500 51.5 1.05 (0.83–1.33) 0.68

Ser/Ser 124 25.1 630 21.6 1.25 (0.96–1.64) 0.10

MAF (Ser) 50.1 47.4 1.12 (0.97–1.28) 0.12

LHR InsLQa

non/non 283 55.4 1220 52.0 1 (reference)

non/ins 202 39.5 934 39.8 0.93 (0.76–1.14) 0.49

ins/ins 26 5.1 190 8.1 0.59 (0.38–0.91) 0.02

MAF (ins) 24.9 28.0 0.85 (0.73–0.99) 0.04

LHR 291

Asn/Asn 466 92.5 2630 89.8 1 (reference)

Asn/Ser 38 7.5 283 9.7 0.76 (0.53–1.08) 0.12

Ser/Ser 0 0.0 16 0.5

MAF (Ser) 3.8 5.4 0.69 (0.49–0.97) 0.03

LHR 312

Ser/Ser 184 36.6 978 33.5 1 (reference)

Ser/Asn 240 47.7 1426 48.9 0.90 (0.73–1.10) 0.30

Asn/Asn 79 15.7 512 17.6 0.82 (0.62–1.09) 0.17

MAF (Asn) 39.6 42.0 0.90 (0.79–1.04) 0.15

OR: odds ratio, CI: confidence interval.aGenotyped in all PCOS cases and 2419 controls.bP-values are not corrected for multiple testing.

Gene polymorphisms and polycystic ovary syndrome 5

at University of N

ewcastle on Septem

ber 3, 2014http://hum

rep.oxfordjournals.org/D

ownloaded from

............................... ............................... ................................ .................... ...............................

..........................................................................................................................................................................................................................................................

Table IV Clinical and endocrine characteristics of 518 anovulatory women with PCOS, stratifications by genotype

GNRH1 TRP16 (P) FSHR 680 (P) LHR 18 InsLQ (P) LHR 291 (P) LHR 312 (P)

Trp/Trp

Trp/Ser

Ser/Ser

Asn/Asn

Asn/Ser

Ser/Ser

Non/Non

Non/Ins

Ins/Ins

Asn/Asn

Asn/Ser

Ser/Ser

Ser/Asn

Asn/Asn

Baseline characteristics and ultrasound parameters

BMI (kg/m2) 26.8 25.2 26.6 �0.50 25.7 25.5 27.6 �0.50 25.5 26.9 26.8 �0.50 26.2 27.3 �0.50 26.2 26.6 25.1 �0.50

Mean folliclenumbera

19 18 16 �0.50 19 18 19 �0.50 19 18 19 �0.50 18.00 17.25 �0.50 18 19 16 �0.50

Mean ovarianvolumea

8.8 8.8 7.9 �0.50 8.8 8.8 8.5 �0.50 8.5 9.2 7.0 �0.50 8.57 9.32 �0.50 8.4 9.1 8.2 �0.50

Endocrine parameters

LH (IU/l) 7.4 7.6 7.9 �0.50 7.0 7.7 8.7 0.01 7.6 7.6 7.2 �0.50 7.5 8.0 �0.50 7.6 7.4 8.4 �0.50

FSH (IU/l) 4.6 5.1 5.4 0.35 4.2 5.0 5.6 0.00001 4.9 4.8 5.6 �0.50 4.9 4.9 �0.50 4.8 4.9 5.3 0.10

Estradiol (pmol/l) 232 230 228 �0.50 228 231 230 �0.50 232 241 179 �0.50 229 259 �0.50 227 231 240 �0.50

Testosterone(nmol/l)

1.9 1.7 1.9 0.30 1.7 1.9 2.0 0.08 1.8 2.0 1.9 �0.50 1.8 2.2 �0.50 1.8 1.9 1.8 �0.50

FAI 5.4 4.3 4.8 �0.50 4.7 4.6 5.8 0.22 4.6 5.6 4.7 �0.50 5.0 4.8 �0.50 5.0 5.1 4.2 �0.50

Fasting Insulin(pmol/l)

65 54 56 0.20 55 58 65 0.35 58 58 65 �0.50 58 57 �0.50 63 58 60 �0.50

AMHb (mg/ll) 10.4 9.6 10.0 �0.50 10.3 9.6 9.9 �0.50 9.2 11.2 11.1 0.15 10.0 12.4 �0.50 9.6 10.4 9.8 �0.50

Features of PCOS

PCO (%) 96.0 94.3 97.6 �0.50 92.1 97.9 93.3 �0.50 94.4 96.9 95.8 �0.50 95.3 97.2 �0.50 94.9 97.0 91.9 �0.50

Hyperandrogenism(%)

60.3 49.7 68.9 �0.50 53.7 52.8 70.2 0.04 53.7 61.4 53.8 �0.50 56.7 68.4 �0.50 58.7 57.5 51.9 �0.50

Values are medians. Hyperandrogenism was defined as the presence of either clinical (FG score �8) or biochemical (FAI � 4.5) signs of androgen excess. P-values were corrected for the number of SNPs tested (Bonferroni correction). P-valueswere calculated for allele-dose effects using analysis of variance (continuous variables, log transformed if not normally distributed) or Chi-square (polycystic ovaries and hyperandrogenism).aMean of left and right ovary.bAMH levels available in a subgroup of 354 PCOS-patients.

6V

alkenburget

al.

at University of Newcastle on September 3, 2014 http://humrep.oxfordjournals.org/ Downloaded from

The present study is limited by the absence of phenotype data incontrols. Therefore, controls could not be selected for the absenceof PCOS. As it is known that 5–8% of women in the general popu-lation will develop PCOS, we expect that the current results will rep-resent an underestimation of the actual difference between cases andcontrols. The concomitant loss-of-power will be partially overcome bythe relatively large number of patients and controls that were includedin the present study. Secondly, there is an apparent age differencebetween controls (.55 years) and PCOS cases of reproductive age.However, it is known that the distribution of genotypes will bestable in subsequent generations of a large population. Therefore, pro-vided that the presence of these polymorphisms does not shorten life-span significantly, the higher age of controls is not regarded as a likelyconfounder. Hence, the lower frequency of the LHR insLQ variant inPCOS cases cannot be easily explained in such a fashion.

The secondary aim of this study was to assess the extent to whichthese genetic variants influence the severity of clinical features ofPCOS. In this regard, the most striking associations were observedfor the FSHR polymorphisms. The present results are consistentwith prior reports concerning the role of FSHR variants in normogo-nadotrophic anovulation (Simoni et al., 2002; Laven et al., 2003). Bothstudies show higher basal FSH levels in association with the FSHRSer680 allele. In women undergoing ovarian stimulation, this receptorvariant results in lower E2 levels following FSH stimulation, suggestinglower FSHR sensitivity (Behre et al., 2005). FSH levels in women withPCOS are within normal limits (Laven et al., 2002). Therefore, it doesnot seem likely that altered FSH sensitivity contributes to the ovula-tory dysfunction that is usually present in PCOS. Apparently, the pitu-itary is capable of a compensatory rise in FSH levels in carriers of theSer680 allele that is able to overcome the increased FSH threshold.Our results further substantiate this hypothesis by the fact that thisvariant did not constitute a risk allele for PCOS and no associationwith the number of antral follicles or AMH levels was observed. Inaddition to the finding of higher FSH levels in carriers of the FSHRSer680 allele, we also report higher LH levels in association with theFSHR Ser680 allele. During folliculogenesis, FSH stimulates the activityand synthesis of aromatase in ovarian granulosa cells (Steinkampf et al.,

1987). Therefore, decreased FSHR sensitivity may disturb normal fol-liculogenesis causing a decrease in the production of E2 and inhibin Bthat exert an inhibitory feedback action at the level of the pituitarygland. The resultant increase in FSH and LH levels can also explainthe finding of increased androgen levels due to persistent stimulationof ovarian theca cells by LH. Indeed we observed a significant corre-lation between LH level and the level of ovarian androgens (testoster-one [P � 0.01] and AD [P � 0.01]), but not with adrenal androgens(DHEA and DHEAS). Although the influence of the FSHR variantson gonadotrophic hormone levels and hyperandrogenism is relativelylarge and consistent with an allele-dose effect, a more subtle influencewas observed for the GnRH1 Trp16Ser polymorphism. The presenceof either one or two copies of the Ser16 allele, which is the minor allelein the European population, is associated with a somewhat morebenign phenotype with regard to carbohydrate metabolism and hyper-androgenism (slightly lower testosterone levels and fasting insulin).However, no significant differences in gonadotrophic hormone levelsare noted, which would argue against a direct influence of this poly-morphism on the function of GnRH as a stimulant at the level ofthe pituitary. Evidently, this genetic variant does not influence PCOSsusceptibility as we found similar genotype distributions in PCOScases and controls.

We have shown that genetic variations in the HPG-axis are capableof altering the phenotype of women with PCOS. More specificallythese changes seem to be centered on the levels of gonadotrophichormones, insulin sensitivity and the presence of hyperandrogenism(either clinical or biochemical). The contribution of these polymorph-isms to the phenotype of PCOS is small and may only be relevant inconjunction with other genetic variants that contribute to minor phe-notypical variation (Simoni et al., 2008a). Therefore, the combinedinfluence of multiple polymorphisms can be expected to be muchmore pronounced. This is clearly illustrated by the results for FSHlevels, showing nearly two times higher levels in compound homozy-gous carriers of FSHR-Ser680 and LHR-Asn312 compared with non-carriers. Although there is a clear association with FSH levels, thisinteraction of genotypes does not seem to influence PCOS suscepti-bility, as we observed no differences in the distribution of theseallelic variants in PCOS cases and controls. These findings show thatpolymorphic variants of the FSHR and LHR are important determi-nants of the physiological setpoint of the HPG-axis that might contrib-ute to the pathophysiology of PCOS.

In summary, genetic variants of the HPG-axis are associated with amodest but significant effect on the phenotype of PCOS. FSHR var-iants are strongly associated with the severity of clinical features ofPCOS, such as levels of gonadotrophic hormones and the presenceof hyperandrogenism, but not disease risk.

FundingA.P.N. Themmen has received honoraria from Diagnostic SystemsLaboratories-Beckman, B.C.J.M. Fauser has received fees and grantsupport from the following companies (in alphabetical order):Andromed, Ardana, Ferring, Merck Serono, Organon, Pantharei Bio-science, PregLem, Schering Plough, Schering, Serono and Wyeth,J.S.E.L. has received fees and grant support from the following com-panies: Ferring, Genovum, Merck-Serono, Organon, Schering Ploughand Serono. O.V., A.G.U., D.P. and A.H. have nothing to disclose.

Figure 1 Median FSH levels in women with PCOS, stratifiedaccording to number of allelic variants (FSH receptor Ser680 and LHreceptor Asn312).

Gene polymorphisms and polycystic ovary syndrome 7

at University of N

ewcastle on Septem

ber 3, 2014http://hum

rep.oxfordjournals.org/D

ownloaded from

ReferencesAtger M, Misrahi M, Sar S, Le Flem L, Dessen P, Milgrom E. Structure of

the human luteinizing hormone-choriogonadotropin receptor gene:unusual promoter and 50 non-coding regions. Mol Cell Endocrinol 1995;111:113–123.

Azziz R, Woods KS, Reyna R, Key TJ, Knochenhauer ES, Yildiz BO.The prevalence and features of the polycystic ovary syndromein an unselected population. J Clin Endocrinol Metab 2004;89:2745–2749.

Balen AH, Laven JS, Tan SL, Dewailly D. Ultrasound assessment of thepolycystic ovary: international consensus definitions. Hum ReprodUpdate 2003;9:505–514.

Behre HM, Greb RR, Mempel A, Sonntag B, Kiesel L, Kaltwasser P,Seliger E, Ropke F, Gromoll J, Nieschlag E et al. Significance of acommon single nucleotide polymorphism in exon 10 of thefollicle-stimulating hormone (FSH) receptor gene for the ovarianresponse to FSH: a pharmacogenetic approach to controlled ovarianhyperstimulation. Pharmacogenet Genomics 2005;15:451–456.

Chhabra S, McCartney CR, Yoo RY, Eagleson CA, Chang RJ, Marshall JC.Progesterone inhibition of the hypothalamic gonadotropin-releasinghormone pulse generator: evidence for varied effects inhyperandrogenemic adolescent girls. J Clin Endocrinol Metab 2005;90:2810–2815.

Dalkin AC, Haisenleder DJ, Ortolano GA, Ellis TR, Marshall JC. Thefrequency of gonadotropin-releasing-hormone stimulation differentiallyregulates gonadotropin subunit messenger ribonucleic acid expression.Endocrinology 1989;125:917–924.

Hofman A, Grobbee DE, de Jong PT, van den Ouweland FA.Determinants of disease and disability in the elderly: the RotterdamElderly Study. Eur J Epidemiol 1991;7:403–422.

Hofman A, Breteler MM, van Duijn CM, Krestin GP, Pols HA, Stricker BH,Tiemeier H, Uitterlinden AG, Vingerling JR, Witteman JC. TheRotterdam Study: objectives and design update. Eur J Epidemiol 2007;22:819–829.

Imani B, Eijkemans MJ, te Velde ER, Habbema JD, Fauser BC. Predictors ofpatients remaining anovulatory during clomiphene citrate induction ofovulation in normogonadotropic oligoamenorrheic infertility. J ClinEndocrinol Metab 1998;83:2361–2365.

Iwasaki H, Emi M, Ezura Y, Ishida R, Kajita M, Kodaira M, Yoshida H,Suzuki T, Hosoi T, Inoue S et al. Association of a Trp16Ser variationin the gonadotropin releasing hormone signal peptide with bonemineral density, revealed by SNP-dependent PCR typing. Bone 2003;32:185–190.

Jakimiuk AJ, Weitsman SR, Navab A, Magoffin DA. Luteinizing hormonereceptor, steroidogenesis acute regulatory protein, and steroidogenicenzyme messenger ribonucleic acids are overexpressed in thecal andgranulosa cells from polycystic ovaries. J Clin Endocrinol Metab 2001;86:1318–1323.

Laven JS, Imani B, Eijkemans MJ, Fauser BC. New approach to polycysticovary syndrome and other forms of anovulatory infertility. ObstetGynecol Surv 2002;57:755–767.

Laven JS, Mulders AG, Suryandari DA, Gromoll J, Nieschlag E, Fauser BC,Simoni M. Follicle-stimulating hormone receptor polymorphisms inwomen with normogonadotropic anovulatory infertility. Fertil Steril2003;80:986–992.

Perez Mayorga M, Gromoll J, Behre HM, Gassner C, Nieschlag E,Simoni M. Ovarian response to follicle-stimulating hormone (FSH)stimulation depends on the FSH receptor genotype. J Clin EndocrinolMetab 2000;85:3365–3369.

Piersma D, Berns EM, Verhoef-Post M, Uitterlinden AG, Braakman I,Pols HA, Themmen AP. A common polymorphism renders the

luteinizing hormone receptor protein more active by improving signalpeptide function and predicts adverse outcome in breast cancerpatients. J Clin Endocrinol Metab 2006;91:1470–1476.

Piersma D, Themmen AP, Look MP, Klijn JG, Foekens JA, Uitterlinden AG,Pols HA, Berns EM. GnRH and LHR gene variants predict adverseoutcome in premenopausal breast cancer patients. Breast Cancer Res2007a;9:R51.

Piersma D, Verhoef-Post M, Look MP, Uitterlinden AG, Pols HA,Berns EM, Themmen AP. Polymorphic variations in exon 10 of theluteinizing hormone receptor: functional consequences andassociations with breast cancer. Mol Cell Endocrinol 2007b;276:63–70.

Powell BL, Piersma D, Kevenaar ME, van Staveren IL, Themmen AP,Iacopetta BJ, Berns EM. Luteinizing hormone signaling and breastcancer: polymorphisms and age of onset. J Clin Endocrinol Metab 2003;88:1653–1657.

The Rotterdam ESHRE/ASRM-Sponsored PCOS consensus workshopgroup. Revised 2003 consensus on diagnostic criteria and long-termhealth risks related to polycystic ovary syndrome. Fertil Steril 2004;81:19–25.

Schipper I, de Jong FH, Fauser BC. Lack of correlation between maximumearly follicular phase serum follicle stimulating hormone concentrationsand menstrual cycle characteristics in women under the age of 35 years.Hum Reprod 1998;13:1442–1448.

Sedlmeyer IL, Pearce CL, Trueman JA, Butler JL, Bersaglieri T, Read AP,Clayton PE, Kolonel LN, Henderson BE, Hirschhorn JN et al.Determination of sequence variation and haplotype structure for thegonadotropin-releasing hormone (GnRH) and GnRH receptor genes:investigation of role in pubertal timing. J Clin Endocrinol Metab 2005;90:1091–1099.

Simoni M, Gromoll J, Hoppner W, Kamischke A, Krafft T, Stahle D,Nieschlag E. Mutational analysis of the follicle-stimulating hormone(FSH) receptor in normal and infertile men: identification andcharacterization of two discrete FSH receptor isoforms. J ClinEndocrinol Metab 1999;84:751–755.

Simoni M, Nieschlag E, Gromoll J. Isoforms and single nucleotidepolymorphisms of the FSH receptor gene: implications for humanreproduction. Hum Reprod Update 2002;8:413–421.

Simoni M, Tempfer CB, Destenaves B, Fauser BC. Functional geneticpolymorphisms and female reproductive disorders: Part I: polycysticovary syndrome and ovarian response. Hum Reprod Update 2008a;14:459–484.

Simoni M, Tuttelmann F, Michel C, Bockenfeld Y, Nieschlag E, Gromoll J.Polymorphisms of the luteinizing hormone/chorionic gonadotropinreceptor gene: association with maldescended testes and maleinfertility. Pharmacogenet Genomics 2008b;18:193–200.

Steinkampf MP, Mendelson CR, Simpson ER. Regulation byfollicle-stimulating hormone of the synthesis of aromatase cytochromeP-450 in human granulosa cells. Mol Endocrinol 1987;1:465–471.

Stephens M, Smith NJ, Donnelly P. A new statistical method for haplotypereconstruction from population data. Am J Hum Genet 2001;68:978–989.

Vagenakis GA, Sgourou A, Papachatzopoulou A, Kourounis G,Papavassiliou AG, Georgopoulos NA. The gonadotropin-releasinghormone (GnRH)-1 gene, the GnRH receptor gene, and theirpromoters in patients with idiopathic hypogonadotropichypogonadism with or without resistance to GnRH action. Fertil Steril2005;84:1762–1765.

van Santbrink EJ, Hop WC, Fauser BC. Classification ofnormogonadotropic infertility: polycystic ovaries diagnosed byultrasound versus endocrine characteristics of polycystic ovarysyndrome. Fertil Steril 1997;67:452–458.

8 Valkenburg et al.

at University of N

ewcastle on Septem

ber 3, 2014http://hum

rep.oxfordjournals.org/D

ownloaded from

Venturoli S, Porcu E, Fabbri R, Magrini O, Gammi L, Paradisi R, Flamigni C.Longitudinal evaluation of the different gonadotropin pulsatile patternsin anovulatory cycles of young girls. J Clin Endocrinol Metab 1992;74:836–841.

Wigginton JE, Abecasis GR. PEDSTATS: descriptive statistics, graphicsand quality assessment for gene mapping data. Bioinformatics 2005;21:3445–3447.

Wigginton JE, Cutler DJ, Abecasis GR. A note on exact testsof Hardy–Weinberg equilibrium. Am J Hum Genet 2005;76:887–893.

Yen SS, Vela P, Rankin J. Inappropriate secretion of follicle-stimulatinghormone and luteinizing hormone in polycystic ovarian disease. J ClinEndocrinol Metab 1970;30:435–442.

Zumoff B, Freeman R, Coupey S, Saenger P, Markowitz M, Kream J. Achronobiologic abnormality in luteinizing hormone secretion inteenage girls with the polycystic-ovary syndrome. N Engl J Med 1983;309:1206–1209.

Submitted on November 1, 2008; resubmitted on March 25, 2009; accepted onMarch 30, 2009

Gene polymorphisms and polycystic ovary syndrome 9

at University of N

ewcastle on Septem

ber 3, 2014http://hum

rep.oxfordjournals.org/D

ownloaded from