JCEM, 2011

Follicle-Stimulating Hormone Receptor Polymorphism(G�29A) Is Associated with Altered Level of ReceptorExpression in Granulosa Cells

Swapna S. Desai, Swati K. Achrekar, Bhakti R. Pathak, Sadhana K. Desai,Vijay S. Mangoli, Ranjana V. Mangoli, and Smita D. Mahale

National Institute for Research in Reproductive Health (S.S.D., S.K.A., B.R.P., S.D.M.), Mumbai 400 012,India; and Fertility Clinic and IVF Center (S.K.D., V.S.M., R.V.M.), Mumbai 400 007, India

Context: Polymorphisms of the FSHR gene are associated with variable ovarian response to FSHstimulation in subjects undergoing in vitro fertilization (IVF) treatment. The type of ovarian re-sponse is correlated with the level of FSH receptor (FSHR) expression on granulosa cells.

Objective: We investigated whether the polymorphism at position �29 in the promoter of the FSHRgene may contribute in altered receptor expression.

Design and patients: FSHR polymorphism at position �29 was studied in 100 subjects undergoing IVFtreatment. Association of this polymorphism with level of FSHR expression was retrospectively analyzed.

Setting: The study was conducted at an academic research institute and private IVF clinic.

Methods: The genotype at position �29 of the FSHR gene was studied in IVF subjects by PCR-restriction fragment length polymorphism. Total RNA and protein was extracted from granulosacells. The relative FSHR mRNA expression was carried out by real-time PCR. The receptor proteinexpression was evaluated by Western blot and confocal microscopy.

Results: The clinical and endocrinological parameters revealed that almost 72% of subjects with theAA genotype at position �29 of FSHR gene were poor ovarian responders (odds ratio 8.63, 95%confidential interval 1.84–45.79; P � 0.001). The lower cleavage intensity predicted by in silicoanalysis for A allele as compared with the G allele suggest the difference in the DNA-proteinbinding affinity. The relative expression of FSHR at mRNA and protein level was significantlyreduced in subjects with AA genotype as compared with the GG genotype.

Conclusion: Poor ovarian response observed in subjects with the AA genotype at position �29of the FSHR gene is due to reduced receptor expression. (J Clin Endocrinol Metab 96: 2805–2812,2011)

In in vitro fertilization (IVF) treatment, FSH is adminis-tered for superovulation. Several studies have shown in-

terindividual variability in the type of ovarian response toFSH stimulation during IVF treatment. Approximately9–24% of women undergoing IVF treatment respondpoorly to gonadotropin stimulation (1). Such subjects mayhave high basal FSH levels, require high exogenous FSHdose for ovulation induction, demonstrate low serum es-

tradiol levels, and fewer mature oocytes during IVF treat-ment. The basis of low response to gonadotropin admin-istration during ovarian stimulation still remains anenigma. Several endocrine and ultrasound parameterssuch as d 3 serum FSH concentration (2, 3), poor follicularblood flow (4), age and diminished ovarian reserve (5),presence of ovarian antibody (6), and serum anti-Mulle-rian hormone levels (7) have been used as indicators to

ISSN Print 0021-972X ISSN Online 1945-7197Printed in U.S.A.Copyright © 2011 by The Endocrine Societydoi: 10.1210/jc.2011-1064 Received March 22, 2011. Accepted June 24, 2011.First Published Online July 13, 2011

Abbreviations: cETS-1, Cellular homolog to the viral E26 transformation specific sequence;Ct, threshold cycle; 3D, three-dimensional; FSHR, FSH receptor; hCG, human chorionicgonadotropin; IVF, in vitro fertilization.

O R I G I N A L A R T I C L E

E n d o c r i n e R e s e a r c h

J Clin Endocrinol Metab, September 2011, 96(9):2805–2812 jcem.endojournals.org 2805

adjust the FSH dosage required in poor ovarian respond-ers. Although a variety of strategies have been used topredict ovarian response, the treatment of poor ovarianresponders is one of the ongoing challenges in the field ofinfertility management (8).

FSH receptor (FSHR), which is present exclusively onthe membrane of granulosa cells, plays an important rolein mediating FSH action, thereby inducing folliculogenesis(9, 10). It has been observed that reduced expression ofFSHR on granulosa cells may account for poor ovarianresponse to FSH stimulation in women undergoing IVFtreatment. The findings suggest that increasing the dose ofexogenous FSH does not improve oocyte developmentprobably due to insufficiency of FSHR expression on gran-ulosa cells (11).

A number of naturally occurring inactivating muta-tions in the FSHR gene have been reported in subjects withinfertility. The phenotype of the infertile subjects has beenwell correlated with the extent of FSHR inactivation (9,12). Besides inactivating mutations, several single-nucle-otide polymorphisms have been identified in the FSHRcoding region. The significance of FSHR gene polymor-phisms in ovarian response has been reported in subjectsundergoing IVF treatment (13–15). The single-nucleotidepolymorphisms at amino acid positions 307 (Thr307Ala)and 680 (Asp680Ser) were observed to be associated withaltered response to FSH during IVF treatment, and bothvariants were reported to be in near complete linkage dis-equilibrium (13, 16, 17). The evidence of a polymorphismat position �29 in the core promoter region of the FSHRgene has been reported in subjects undergoing IVF treat-ment (18). Previous studies carried out by our group in-dicated theassociationof theAAgenotypeatposition �29with poor ovarian response (19). Interestingly, in a studyconducted in women with hypertension, it was observedthat the AA genotype at position �29 is associated withlower serum estradiol levels. Moreover, in vitro functionalstudies carried out in Chinese hamster ovary cells with thepromoter constructs of the FSHR gene revealed reducedtranscriptional activity in the case of the A allele comparedwith the G allele (20). It has also been observed that thepolymorphism at position �29 is located in a consensussequence (GGAA) for the cellular homolog to the viral E26transformation specific sequence (cETS-1) transcriptionfactor and the A allele at position �29 might lead to theloss of the transcription factor binding site (18, 20).

These observations prompted us to investigate whetherthe poor ovarian response observed in subjects with theAA genotype at position �29 of the FSHR gene is due toreduced expression of the receptor on granulosa cells. Theaim of present study was to compare the level of the re-

ceptor expression and ovarian response among subjectswith different genotypes at position �29 of FSHR gene.

Subjects and Methods

Subject selectionThe present study was approved by the Ethics Committee for

Clinical Research at the National Institute for Research in Re-productive Health and the Fertility Clinic and IVF Center. A totalof 100 normogonadotropic ovulatory women (menstrual cyclelength 25–35 d) with infertility due to male or tubal factor or withunexplained infertility were retrospectively analyzed. All sub-jects were of Indian ethnicity. Informed consent was obtainedfrom all subjects. All recruited subjects were in the age group of25–46 yr, and their basal serum FSH levels were in the normalrange (2–10 IU/liter). The basal FSH level, FSH amount admin-istered, estradiol levels before and on the day of the human cho-rionic gonadotropin (hCG) administration, and the number ofoocytes retrieved were considered for segregating subjects as hy-perresponders, normal responders, and poor responders. Womenwith polycystic ovarian syndrome, endometriosis, fibroids, and hy-perprolactinemia were excluded from this study.

Ovulation inductionA standard protocol described previously (19) was used for ovu-

lation inductionduring IVFtreatment.StimulationwithFSH(150–600 IU/d) was monitored by measuring the serum estradiol levelsand follicle growth. hCG (10,000 IU) was administered for trigger-ing ovulation. Oocytes were retrieved after 36 h under transvaginalultrasound guidance and mature oocytes (�14 mm in diameter)were collected. The number of preovulatory follicles and retrievedoocytes were recorded for each subject. Serum levels of FSH, LH,estradiol, and progesterone were measured by chemiluminescencemicroparticle assay (Abbott Architect, Abbott Park, IL).

Genotyping for the polymorphism at position �29of the FSHR gene

Genomic DNA was extracted from 250 �l of whole bloodobtained from each of the subjects (n � 100) using a commercialkit (Genexy, New Delhi, India) according to the manufacturer’sinstructions. The G�29A polymorphism was screened by PCRand restriction fragment length polymorphism as described pre-viously (19). The findings from restriction fragment length poly-morphism for all three genotypes were confirmed by direct se-quencing for 30 samples (10 samples of each genotype) at theDNA sequencing core facility of the institute.

DNA topographyThe three-dimensional (3D) molecular structure of DNA is af-

fectedbynucleotide changes,whichcancausedifferences inproteinbinding and affinity. The level of change in the DNA shape thatarises due to the polymorphism (rs1394205) at position G�29A ofthe FSHR gene was compared by ranking the structure-changevalue as determined by the predicted hydroxyl radical cleavage pat-tern for DNA sequence of the core promoter region from �33 to�14 bp using the OH Radical Cleavage Intensity Database (21).The effect of this polymorphism on the structural profile was quan-titatively measured in terms of the Euclidean distance, which is usedas a metric to calculate average structure change (22).

2806 Desai et al. Polymorphism in FSHR Gene and Receptor Expression J Clin Endocrinol Metab, September 2011, 96(9):2805–2812

Quantitative real-time PCRGranulosa cells in the follicular fluid (n�100) were separated

by centrifugation at 2000 � g for 10 min at 4 C, and cells wereused to extract total RNA by TRIZOL (Invitrogen, Carlsbad,CA). The first-strand cDNA was synthesized (4 �g of total RNA)using Superscript III reverse transcriptase enzyme and oligo de-oxythymidine (dT) as primers (Invitrogen) as per the manufac-turer’s instructions. The predesigned primers and probes for hu-man FSHR and �-actin (TaqMan Assay-on-Demand geneexpression product; Applied Biosystems, Foster City, CA) wereused for amplification of cDNA by real-time PCR in duplicatereactions. Because it was not possible to get granulosa cells fromproven fertile women, cDNA was pooled from three IVF subjectsof each genotype and used as a calibrator. The relative mRNAexpression of FSHR in IVF subjects was calculated by �� thresh-old cycle (Ct) method, where the �Ct was calculated as the dif-ference between the Ct of FSHR and Ct of �-actin in each sample.The ��Ct of each sample was then calculated as the differencebetween the �Ct of the sample with �Ct of the calibrator sample,which was normalized with �-actin. The level of FSHR expres-sion was compared within three genotypes at position �29 byANOVA.

Western blot analysisTo study the association of the FSHR gene polymorphism at

position �29 with protein expression, total protein was ex-tracted from granulosa cells collected from 30 subjects (10 sub-jects of each genotype) by TRIZOL (Invitrogen) as per the man-ufacturer’s instructions. The protein concentration of the lysatewas determined by Folin-Lowry’s assay and 50 �g of protein wasloaded in each lane of a 7.5% SDS-PAGE gel. After electropho-resis, the protein was transferred on to a 0.45-�m nitrocellulosemembrane (Hybond C Extra; GE Healthcare, Buckinghamshire,UK) using a semidry transfer apparatus (Bio-Rad, Hercules, CA).Immunoblotting was performed using the monoclonal antibodyto FSHR 106.105 (0.5 �g/�l) (a gift from Dr. J. A. Dias, Wads-worth Center, David Axelrod Institute for Public Health, Al-bany, NY) or the monoclonal antibody for �-actin (1:1000;Santa Cruz Biotechnology, Santa Cruz, CA) as the primary an-tibody and horseradish peroxidase-conjugated goat antimouseIgG antibody (1:5000; Dako, Copenhagen, Denmark) as the sec-ondary antibody. The bands were visualized with enhancedchemiluminescence reagent (ECL-Plus) (GE Healthcare) on x-ray film. The protein expression of FSHR was normalized to�-actin for each sample and compared among subjects with threegenotypes at position �29 by ANOVA.

In a separate experiment, monoclonal antibody preincubatedwith 20 �g of synthetic peptide (corresponding to the epitope297–310 amino acid of human FSHR) was used for probing theblot (23). This served as a negative control.

Immunofluorescence and confocal microscopySurface localization of FSHR on cumulus cells was carried out

using direct immunofluorescence microscopy using a modifiedprotocol adapted from Beau et al. (24). The cells obtained fromfive subjects of each genotype at position �29 were treated withanti-FSHR monoclonal antibody 106.105 conjugated to Alexafluor 568 (0.5 �g per 100 �l). The cells were preincubated with100-fold higher concentration of unlabeled anti-FSHR mono-clonal antibody prior probing with labeled antibody, whichserved as a negative control. Cells were then fixed with 2% para-

formaldehyde. 4�-6-Diamidino-2-phenylindole was used as nu-clear stain, and cells were smeared on a glass slide. Cellular dis-tribution of the receptors was examined using the LSM510-Metaconfocal system (Carl Zeiss, Jena, Germany). Z stacks weretaken and 3D composite images were generated. The level ofsurface expression was quantified using ImageJ software (Na-tional Institutes of Health, Bethesda, MD). To normalize thedata, the percentage of FSHR expression in subjects with GA andAA genotype was calculated with respect to the GG genotype andcompared by ANOVA.

Statistical analysisConformance with the Hardy-Weinberg equilibrium was

computed by a �2 test (25). The clinical parameters and the levelof FSHR expression were compared between the study groups onthe basis of genotypes at position �29 and type of ovarian re-sponse using one-way ANOVA and a least significant differencepost hoc multiple comparisons test. Statistical analysis was per-formed with Statistics Package for Social Sciences (SPSS) forWindows, version 16 (SPSS Inc., Chicago, IL). The odds ratiowas calculated using �2 analysis to study the association of theFSHR genotype at position �29 with poor ovarian responseusing Epi Info version 6 (World Health Organization, Geneva,Switzerland). P � 0.05 was considered statistically significant.

Results

Polymorphism at position �29 of the FSHR geneThe frequency distribution of the polymorphism at

position �29 was 47% for the GG genotype, 42% forthe GA genotype, and 11% for the AA genotype in 100IVF subjects. The �2 analysis revealed that the frequencyof genotypes at position �29 was in Hardy-Weinberg’sequilibrium.

Clinical and endocrinological parametersTo analyze the potential association between the FSHR

gene polymorphism at position �29 and the ovarian re-sponse during gonadotropin stimulation, the clinical, en-docrine, and ultrasonographic parameters were recordedfor all the subjects recruited in this study (Table 1). Sub-jects were independently segregated on the basis of geno-types (GG, GA, or AA) at position �29. The age, basalFSH and LH levels, peak estradiol, and progesterone con-centration before and on the day of hCG treatmentshowed no statistically significant difference among thethree genotypes. In contrast to the above parameters, theamount of exogenous FSH required for ovulation induc-tion was significantly different in all three genotypes (P �0.003). Maximum amount of exogenous FSH (4563 �271 IU) was required by the subjects with AA genotype,whereas subjects with the GG genotype required a mini-mum amount of FSH (2492 � 154 IU) for ovarian stim-ulation. The amount of FSH (3265 � 239 IU) required insubjects with the GA genotype was lower than that re-


quired in subjects with the AA genotype. The ultrasoundfindings revealed that the number of preovulatory folliclesand retrieved oocytes were significantly low in subjectswith the AA genotype (9 � 0.8, 9.18 � 0.9, respectively)compared with the GG genotype (13.21 � 0.6, 13.91 � 1,respectively) and the GA genotype (12.62 � 0.7, 12.57 �1.1, respectively) (P � 0.05). Moreover, the number ofmature oocytes (in M II phase) retrieved in case of subjectswith the AA genotype (7 � 0.8) were significantly lowercompared with the GG genotype (11.62 � 0.8; P � 0.022).

Furthermore, the clinical and endocrine parameterswere used to segregate the subjects as hyperresponders,normal responders, and poor responders and were ob-served to be significantly different among the three groups(Table 1). In the study group of 100 subjects, 29 subjectswere poor ovarian responders. When we further segre-gated these subjects (n � 29) on the basis of genotype atposition �29, it was observed that only 21.27% of sub-jects with the GG genotype and 26.19% of the subjectswith the GA genotype were poor ovarian responders. In-terestingly, 72.72% of the subjects with the AA genotypewere poor ovarian responders. The �2 test was used to

study the association between the polymorphism at posi-tion �29 with poor ovarian response. The odds ratio forthe GG genotype was 0.48 [95% confidence interval (CI)0.18–1.29; P � 0.166], for the GA genotype was 0.79(95% CI 0.30–2.08; P � 0.761), and for the AA genotypewas 8.63 (95% CI 1.84–45.79; P � 0.001) (Table 2).

DNA topography of FSHR geneWe investigated whether the polymorphism at position

�29 might alter the 3D molecular structure of the DNA,which could potentially influence protein binding affinityand phenotype. The polymorphism at position �29 re-sulted in a DNA structural change value greater than 0.8.The magnitude of this structural change value revealedthat the DNA with the A allele (0.12) is less accessible forbinding of transcription factors compared with the G al-lele (0.94) (Fig. 1A).

FSHR expression at mRNA levelThe relative mRNA expression was monitored by

quantitative real-time PCR and compared among the threegenotypes at position �29. The level of FSHR mRNA

TABLE 1. Clinical and endocrinological parameters of the subjects undergoing IVF treatment based on thegenotypes at position �29 of the FSHR gene and based on the type of ovarian response to FSH stimulation

Parameters

Genotypes at position �29 of FSHR gene Type of ovarian response to FSH stimulation

GG(n � 47)

GA(n � 42)

AA(n � 11)

Hyper(n � 11)

Normal(n � 60)

Poor(n � 29)

Age (yr) 32.30 � 0.62 32.43 � 0.64 34.61 � 1.19 30.55 � 0.91a 31.93 � 0.47b 34.79 � 0.91a,b

Basal FSH levels (IU/liter) 6.32 � 0.32 6.46 � 0.29 6.93 � 1.24 5.29 � 0.45a 6.21 � 0.26b 7.24 � 0.42a,b

Basal LH levels (IU/liter) 4.40 � 0.35 5.18 � 0.39 5.25 � 1.80 6.08 � 0.53 4.55 � 0.36 4.88 � 0.54Total amount of exogenous

FSH administered (IU)2492.55 � 154.30a 3265.48 � 239.23a 4563.64 � 271.53a 1729.55 � 136.00a 2673.33 � 137.63a 4312.93 � 264.61a

Estradiol levels before hCGadministration (pg/ml)

1688.70 � 110.00 1855.33 � 149.00 1640.73 � 243.00 3094.18 � 165.67a 1761.30 � 87.08a 1228.52 � 133.69a

Estradiol levels on the dayof hCG administration(pg/ml)

1954.67 � 141.00 2148.36 � 180.00 1890.50 � 286.00 3806.36 � 163.77a 1988.27 � 101.85a 1447.07 � 161.59a

Progesterone levels beforehCG administration(ng/ml)

0.54 � 0.03 0.91 � 0.33 0.62 � 0.06 0.75 � 0.08 0.80 � 0.23 0.49 � 0.04

Progesterone levels on the dayof hCG administration(ng/ml)

4.15 � 0.40 4.99 � 0.53 4.64 � 0.52 8.48 � 1.20a,b 4.61 � 0.34a 3.73 � 0.64b

Preovulatory follicles, n 13.21 � 0.67a 12.62 � 0.75b 9.00 � 0.86a,b 19.73 � 0.99a 13.58 � 0.35a 7.52 � 0.56a

Retrieved oocytes, n 13.91 � 1.02a 12.57 � 1.16 9.18 � 0.90a 26.91 � 1.69a 13.53 � 0.48a 6.03 � 0.53a

Mature oocytes, n 11.62 � 0.86a 9.86 � 0.99 7.00 � 0.89a 21.82 � 1.87a 10.98 � 0.44a 4.76 � 0.45a

Values are presented as mean � SEM. One-way ANOVA tests and LSD post hoc multiple comparisons were used for ANOVA. Same letters (a or b)for a given parameter indicate statistically significant difference at P � 0.05.

TABLE 2. Genotype frequencies at the position �29 of the FSHR gene in subjects undergoing IVF treatment and itscorrelation with the occurrence of poor ovarian response

Genotype atposition �29

IVF subjectsscreened, n

Poor ovarianresponders, n

Poor ovarianresponders (%)

Odds ratio(95% CI) P

GG 47 10 21.27 0.48 (0.18–1.29) 0.166GA 42 11 26.19 0.79 (0.30–2.08) 0.761AA 11 8 72.72 8.63 (1.84–45.79) 0.001a

a P � 0.05 by �2 test.


expression was observed to be variable. The FSHR ex-pression at transcript level was 1.3-fold lower in case ofsubjects with the AA genotype (0.15 �0.5) compared withthe GG genotype (1.49 � 0.3), and the difference wasobserved to be statistically significant (P � 0.039). Thelevel of expression was intermediate in the case of subjectswith GA genotype (0.90 � 0.21) (Fig. 1B).

We also compared the relative FSHR mRNA expressionamongsubjectswithvariable typeofovarianresponse. Itwasobserved that mRNA expression was similar in three groups(Fig. 2A). Therefore, to investigate whether all the poor re-sponders (n � 29) in the present study show altered FSHRexpression,wesegregated these subjects (n�29)onthebasisof genotype at position �29. It was observed that the FSHRmRNA expression in poor ovarian responders was signifi-

cantly lower in subjects with AA genotype compared withthe GG genotype (P � 0.027) (Fig. 2B).

FSHR expression at protein levelFSHR protein expression in 10 subjects of each genotype

atposition�29wasanalyzedbyWesternblotandquantifiedusing densitometry. The specificity of the FSHR monoclonalantibody was confirmed, and it was observed that FSHRexpression at protein level was variable among the threegenotypes (Fig. 3A). It was evident that the relative level ofFSHR protein expression was significantly reduced in thecase of subjects with the AA genotype (0.49 � 0.08) com-pared with the GG (0.76 � 0.05) and GA (0.75 � 0.08)genotype (P � 0.05, Fig. 3B).

The cell surface expression of FSHR was assessed bydirect immunofluorescence technique, and specificitywas confirmed using the negative control (Fig. 4A). Dif-ferential expression of membrane receptor among threegenotypes at position �29 was observed (Fig. 4, B–D).Significantly reduced level of membrane receptor ex-pression were observed on cumulus cells obtained fromsubjects with the AA genotypes (6.6 � 0.4) comparedwith the GG genotype (16.3 � 1.4) and the GA genotype(15.2 � 1.2). The fluoresceins intensity of GG genotypeis considered as 100% and compared with GA and AAgenotype (P � 0.002, Fig. 4E).

Discussion

In our earlier study, we had recruited 50 women un-dergoing ovarian stimulation protocol, and the resultsindicated that the polymorphism at position �29 of theFSHR gene is associated with poor ovarian response(19). Therefore, we speculated that this polymor-phism of the FSHR gene might influence the level ofreceptor expression. We recruited an additional 100subjects undergoing IVF treatment and determined thefrequency distribution of the polymorphism at position�29 of the FSHR gene in these subjects. The frequencydistribution observed in the present study for 100 sub-jects was comparable with our previous report for 50subjects (19). The frequency distribution of this poly-morphism in subjects undergoing IVF treatment hasalso been reported for Indonesian and German popu-lations (18).

The potential association between the polymorphismat position �29 and the clinical parameters was ana-lyzed in these subjects. It was observed that the subjectswith the AA genotype required significantly higheramounts of exogenous FSH for ovulation inductioncompared with subjects with the GG and GA genotypes.

FIG. 1. A, Comparative predicted hydroxyl radical cleavage intensitycorresponding to the nucleotide sequence of wild-type andpolymorphic FSHR gene. B, The comparison of relative FSHR mRNAexpression among three genotypes at position �29 of the FSHR genein IVF subjects (n � 100). Values presented in bar are mean � SEM andcompared by one-way ANOVA. �, P � 0.05 is considered statisticallysignificant.


These results indicate that the AA genotype impartshigher resistance to FSH stimulation (Table 1). We alsoobserved that in subjects with the AA genotype, thenumber of preovulatory follicles and mature oocyteswere significantly lower compared with the GG geno-type (Table 1). These observations suggest that the AAgenotype at position �29 might be associated withpoor ovarian response, which correlates with our

earlier study (19). Wunsch et al. (18) havereported that the polymorphism at position�29 does not seem to influence the clinicalparameters observed in subjects undergoingovarian stimulation substantially, when thebasal FSH and peak estradiol levels wereanalyzed.

Identification of the poor responders is clin-ically relevant to determine individualizedovarian stimulation protocol during IVF treat-ment. In the present study, almost 72% of thesubjects with the AA genotype were poor ovar-ian responders. We calculated the relative riskof poor ovarian response associated with thegenotype at position �29 in IVF subjects (n �100). The �2 analysis revealed that subjectswith the AA genotype have a higher risk ofgiving poor ovarian response to gonadotropin

stimulation during IVF treatment (Table 2). This suggeststhe significance of AA genotype as a biomarker to predictpoorovarian responders,whichmight improve the clinicaloutcome during IVF treatment.

To identify the probable reason for the association ofpoor ovarian response observed in the AA genotype, wedetermined the FSHR mRNA expression in granulosacells. Interestingly, we noted that subjects with the AAgenotype displayed significantly lower FSHR mRNAexpression compared with the GG genotype (Fig. 1B).Our observation strengthens the earlier report in whichthe promoter activity of the FSHR gene was found to besignificantly reduced in the case of the A allele comparedwith the G allele at position �29 by in vitro analysis(20). The reduced expression of the FSHR gene in sub-jects with the AA genotype might be polymorphism-associated steric changes in the DNA structure, whichmay affect the binding of transcription factors becausethis polymorphism is located in the c-ETS-1 transcrip-tion factor binding site (Fig. 1A) (18). Lower cleavageintensity observed in the case of the A allele comparedwith the G allele suggests its importance in the promoterefficiency of the FSHR gene causing a decrease in FSHRexpression in subjects with the AA genotype. However,the mechanism by which the cETS-1 transcription fac-tor influences human FSHR gene expression is not clearand needs to be investigated further.

Findings from our previous (19) and present study con-sistently demonstrate a correlation of the AA genotype at�29 position of the FSHR gene with poor ovarian re-sponse. Although FSHR expression in granulosa cells hasbeen shown to be associated with individual response toFSH stimulation (11), the observation was not the same inthe study reported by Thiruppathi et al. (26). Poor ovarian

FIG. 2. A, The comparison of relative FSHR mRNA expression A, in different type ofovarian responders (n � 100). B, In poor ovarian responders among three genotypesat position �29 of the FSHR gene (n � 29). Statistical analysis was carried out byone-way ANOVA. �, P � 0.05 is considered statistically significant.

FIG. 3. A, The representative image for the protein expression of FSHRand �-actin in granulosa cells in three genotypes at position �29 ofthe FSHR gene. B, The box plot showing the comparison of FSHRprotein (mean � SEM) among three groups, median (middle bar), andquartiles (boxes). The statistical analysis was carried out by one-wayANOVA, and the same letters (a or b) indicate statistically significantdifference at P � 0.05.


response observed in different individuals during IVFtreatment could be due to other parameters apart fromFSHR expression. This is evident from our observationthat the relative expression of FSHR mRNA levels did notvary when we segregated the subjects based on ovarianresponse (Fig. 2A).However, itwas interesting tonote thatwithin poor responders, relative mRNA expression ofFSHR was significantly decreased in the AA genotypecompared with the GG genotype (P � 0.027) (Fig. 2B).Furthermore, we noted that none of the subjects with theAA genotype were hyperresponders. These results suggestthat the level of receptor expression is closely associatedwith genotype at position �29 of FSHR gene in poor ovar-ian responders.

More interestingly, when we compared the FSHRexpression at protein level on the basis of genotypes atposition �29, we observed that subjects with the AAgenotype expressed significantly lower amounts of re-ceptor protein compared with the GG and GA geno-types (Fig. 3). Our observation thus suggests that thereduced FSHR expression at the transcript level is inconcurrence with the expression of FSHR at the proteinlevel in subjects with the AA genotype. It has been re-ported that FSH directly exerts its action on the oocytethrough its receptor, expressed on the membrane of cu-mulus cells surrounding the oocyte (27). Therefore, in

the present study, we evaluated thecell surface expression of FSHR byconfocal microscopy using cumuluscells obtained from recruited subjects.The results revealed that the receptorexpression on cumulus cells obtainedfrom subjects with the AA genotypewas significantly lower comparedwith the GG and GA genotypes (Fig.4). Thus, our findings clearly suggestthat reduced FSHR expression in theAA genotype could be the possiblereason for poor ovarian response toFSH stimulation observed in thesesubjects.

One of the major challenges forthe clinicians during ovarian sti-mulation protocol is nonavailabilityof reliable predictive indicators toidentify women who are poor re-sponders. The present study clearlyshows that the AA genotype at posi-tion �29 of the FSHR gene is associ-ated with poor ovarian response andrelatively lower expression of recep-tor on the granulosa cells. Further-more, large multicentric studies are

required to elucidate the relative contribution of thispolymorphism to be used as a marker for predictingpoor ovarian response during IVF treatment.

Acknowledgments

We are thankful to the participants of the study. The authorsacknowledge Dr. D. Balaiah and Mr. Prashant Tapse (Divisionof Biostatistics, National Institute for Research in ReproductiveHealth) for their help in statistical analysis. We also thank Dr.Anurupa Maitra and Ms. Nanda Ugale (DNA sequencing corefacility, National Institute for Research in Reproductive Health)for their assistance in the DNA sequencing. Technical help pro-vided by Ms. Savita (Fertility Clinic and the IVF Center) is alsoacknowledged. We also thank Dr. Nafisa Balasinor, Ms. ShobhaSonawane, and Ms. Reshma Goankar in facilitating confocalmicroscopy analysis.

Address all correspondence and requests for reprints to:Smita D. Mahale, Division of Structural Biology, NationalInstitute for Research in Reproductive Health, Mumbai 400012, India. E-mail: [email protected].

This work (NIRRH/MS/03/2011) was supported by grantsfrom the Indian Council of Medical Research, New Delhi, India(P and I/BIC/1/1/2009) and the Board of Research in NuclearSciences, Department of Atomic Energy.

Disclosure Summary: The authors have nothing to disclose.

FIG. 4. The representative image for surface expression of FSHR (green signal) on cumuluscells by confocal microscopy using monoclonal FSHR antibody conjugated with Alexa Fluor568 and nuclear staining with 4�-6-diamidino-2-phenylindole (red signal). A–D, The 3Dstructure of cumulus cells for A, negative control and B–D, three genotypes at position �29of the FSHR gene. Bar, 10 �m. E, Intensity of fluorescence in each cell was calculated usingImageJ software. The comparison is of percent FSHR cell surface expression (mean � SEM)among the three genotypes by one-way ANOVA, and the same letters (a or b) indicatestatistically significant difference at P � 0.05. Data analyzed are from six to eight cells fromeach subject and a total of 35 cells for each genotype.


References

1. Keay SD, Liversedge NH, Mathur RS, Jenkins JM 1997 Assistedconception following poor ovarian response to gonadotrophin stim-ulation. Br J Obstet Gynaecol 104:521–527

2. Balasch J, Creus M, Fabregues F, Carmona F, Casamitjana R, As-caso C, Vanrell JA 1996 Inhibin, follicle-stimulating hormone, andage as predictors of ovarian response in in vitro fertilization cyclesstimulated with gonadotropin-releasing hormone agonist-gonado-tropin treatment. Am J Obstet Gynecol 175:1226–1230

3. Barnhart K, Osheroff J 1998 Follicle stimulating hormone as a pre-dictor of fertility. Curr Opin Obstet Gynecol 10:227–232

4. Battaglia C, Genazzani AD, Regnani G, Primavera MR, Petraglia F,Volpe A 2000 Perifollicular Doppler flow and follicular fluid vas-cular endothelial growth factor concentrations in poor responders.Fertil Steril 74:809–812

5. Kligman I, Rosenwaks Z 2001 Differentiating clinical profiles: pre-dicting good responders, poor responders, and hyperresponders.Fertil Steril 76:1185–1190

6. Luborsky JL, Thiruppathi P, Rivnay B, Roussev R, Coulam C, Rad-wanska E 2002 Evidence for different aetiologies of low estradiolresponse to FSH: age-related accelerated luteinization of follicles orpresence of ovarian autoantibodies. Hum Reprod 17:2641–2649

7. Nardo LG, Yates AP, Roberts SA, Pemberton P, Laing I 2009 Therelationships between AMH, androgens, insulin resistance and basalovarian follicular status in non-obese subfertile women with andwithout polycystic ovary syndrome. Hum Reprod 24:2917–2923

8. Tarlatzis BC, Zepiridis L, Grimbizis G, Bontis J 2003 Clinical man-agement of low ovarian response to stimulation for IVF: a systematicreview. Hum Reprod Update 9:61–76

9. Simoni M, Gromoll J, Nieschlag E 1997 The follicle stimulatinghormone receptor: biochemistry, molecular biology, physiology andpathophysiology. Endocr Rev 18:739–773

10. Dierich A, Sairam MR, Monaco L, Fimia GM, Gansmuller A, Le-Meur M, Sassone-Corsi P 1998 Impairing follicle-stimulating hor-mone (FSH) signaling in vivo: targeted disruption of the FSH recep-tor leads to aberrant gametogenesis and hormonal imbalance. ProcNatl Acad Sci USA 95:13612–13617

11. Cai J, Lou HY, Dong MY, Lu XE, Zhu YM, Gao HJ, Huang HF2007 Poor ovarian response to gonadotropin stimulation is associ-ated with low expression of follicle-stimulating hormone receptor ingranulosa cells. Fertil Steril 87:1350–1356

12. Themmen APN, Huhtaniemi IT 2000 Mutations of gonadotropinsand gonadotropin receptors: elucidating the physiology and patho-physiology of pituitary-gonadal function. Endocr Rev 21:551–583

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

14. Simoni M, Nieschlag E, Gromoll J 2002 Isoforms and single nucle-

otide polymorphisms of FSH receptor gene: implications for humanreproduction. Hum Reprod Update 8:413–421

15. Moron FJ, Ruiz A 2010 Pharmacogenetics of controlled ovarianhyperstimulation: time to corroborate the clinical utility of FSHreceptor genetic markers. Pharmacogenomics 11:1613–1618

16. Sudo S, Kudo M, Wada S, Sato O, Hsueh AJ, Fujimoto S 2002Genetic and functional analyses of polymorphism in the human FSHreceptor gene. Mol Hum Reprod 8:893–899

17. Achrekar SK, Modi DN, Desai SK, Mangoli VS, Mangoli RV, Ma-hale SD 2009 Follicle-stimulating hormone receptor polymorphismThr307Ala is associated with variable ovarian response and ovarianhyperstimulation syndrome in Indian women. Fertil Steril 91:432–439

18. Wunsch A, Ahda Y, Banaz-Yasar F, Sonntag B, Nieschlag E, SimoniM, Gromoll J 2005 Single-nucleotide polymorphisms in the pro-moter region influence the expression of the human follicle stimu-lating hormone receptor. Fertil Steril 84:446–453

19. Achrekar SK, Modi DN, Desai SK, Mangoli VS, Mangoli RV, Ma-hale SD 2009 Poor ovarian response to gonadotropin stimulation isassociated with FSH receptor polymorphism. RBM Online 18:509–515

20. Nakayama T, Kuroi N, Sano M, Tabara Y, Katsuya T, Ogihara T,Makita Y, Hata A, Yamada M, Takahashi N, Hirawa N, UmemuraS, Miki T, Soma M 2006 Mutation of the follicle-stimulating hor-mone receptor gene 5�-untranslated region associated with femalehypertension. Hypertension 48:512–518

21. Greenbaum JA, Pang B, Tullius TD 2007 Construction of a genome-scale structural map at single-nucleotide resolution. Genome Res17:947–953

22. Parker SC, Hansen L, Abaan HO, Tullius TD, Margulies EH 2009Local DNA topography correlates with functional noncoding re-gions of the human genome. Science 324:389–392

23. Lindau-Shepard B, Brumberg HA, Peterson AJ, Dias JA 2001 Re-versible immunoneutralization of human follitropin receptor. J Re-prod Immunol 49:1–19

24. Beau I, Touraine P, Meduri G, Gougeon A, Desroches A, Matuchan-sky C, Milgrom E, Kuttenn F, Misrahi M 1998 A novel phenotyperelated to partial loss of function mutations of the follicle stimulatinghormone receptor. J Clin Invest 102:1352–1359

25. Rodriguez S, Gaunt TR, Day IN 2009 Hardy-Weinberg equilibriumtesting of biological ascertainment for Mendelian randomizationstudies. Am J Epidemiol 169:505–514

26. Thiruppathi P, Shatavi S, Dias JA, Radwanska E, Luborsky JL 2001Gonadotrophin receptor expression on human granulosa cells oflow and normal responders to FSH. Mol Hum Reprod 7:697–704

27. Oktay K, Briggs D, Gosden RG 1997 Ontogeny of follicle-stimu-lating hormone receptor gene expression in isolated human ovarianfollicles. J Clin Endocrinol Metab 82:3748–3751


Documents

JCEM, 2011