European Journal of Endocrinology (2009) 160 855–862 ISSN 0804-4643

CLINICAL STUDY

Genetic variation in exon 17 of INSR is associated with insulinresistance and hyperandrogenemia among lean Indian womenwith polycystic ovary syndromeSrabani Mukherjee, Nuzhat Shaikh, Sushma Khavale, Gayatri Shinde, Pervin Meherji, Nalini Shah1

and Anurupa MaitraDepartment of Molecular Endocrinology, National Institute for Research in Reproductive Health, Indian Council of Medical Research (ICMR), JehangirMerwanji Street, Parel, Mumbai 400012, India, and 1Seth G S Medical College and KEM Hospital, Parel, Mumbai, 400012, India

(Correspondence should be addressed to S Mukherjee; Email: mukherjees@nirrh.res.in)

q 2009 European Society of E

Abstract

Objective: Polycystic ovary syndrome (PCOS) is a multigenic disorder, and insulin resistance is one of itshallmark features. Polymorphisms in exon 17 of insulin receptor (INSR) gene are reported to beassociated with PCOS. We investigated this association in Indian women and its putative relationshipwith PCOS associated traits, which has not been explored so far.Methods: In this case control study, the polymorphisms were investigated by direct sequencing in180 women with PCOS and 144 age matched controls. Clinical, anthropometric, biochemical, andhormonal parameters were also estimated.Results: The silent C/T polymorphism at His1058 in exon 17 of INSR was found to be present in ourstudy population. The polymorphic genotype (CTCTT) was significantly associated with PCOS in leanwomen (c2Z8.493, dfZ1, PZ0.004). It showed association with higher fasting insulin levels(PZ0.02), homeostasis model assessment of insulin resistance (PZ0.005), free androgen index(PZ0.03), and lower quantitative insulin sensitivity check index (PZ0.004) in lean PCOS women. Noother novel or known polymorphism was identified in exon 17 in this cohort.Conclusions: The study shows significant association of C/T polymorphism at His1058 of INSR withPCOS in the lean rather than obese Indian women. Its association with indices of insulin resistance andhyperandrogenemia is also seen in the same group. The findings strengthen the concept thatpathogenesis of PCOS is different in lean and obese women.

European Journal of Endocrinology 160 855–862

Introduction

Polycystic ovary syndrome (PCOS) is a heterogeneousdisorder with a prevalence of w7% in women ofchildbearing age (1). It is the most common cause ofanovulatory infertility with characteristic mani-festations such as menstrual irregularities; signs ofandrogen excess like hirsutism, acne, alopecia; alteredLH:FSH ratio (O2:1) and polycystic ovaries (2–4). Themajority of women with PCOS have insulin resistance,which plays a key role in its pathogenesis (5, 6). Inaddition, it increases the risk of developing glucoseintolerance, type 2 diabetes mellitus, hypertension, anddyslipidaemia, which can lead to cardiovasculardisorders in later life (7–10). Both obese and leanwomen with PCOS manifest insulin resistance (2) andobesity further contributes an additional component ofinsulin resistance (5). Insulin resistance progressestowards compensatory hyperinsulinemia, which pro-motes both hyperandrogenemia and anovulation inPCOS (1, 11).

ndocrinology

Familial clustering seen in this syndrome suggestscontribution of a genetic component to its pathogenesis(12, 13). Over the past decade, a number of candidategenes involved in steroidogenesis, insulin signalingpathway, gonadotropin secretion and chronic inflam-mation have been explored to identify the susceptibilitygenes for PCOS (14, 15); however, the results so far havebeen inconclusive.

Available evidence suggests insulin resistance inPCOS could be due to post-binding defects in insulinsignaling (2, 16, 17). Hence, insulin receptor (INSR),being an integral part of insulin signaling could be apotential candidate gene. INSR is located on chromo-some 19 and encompasses 22 exons. Linkage analysisstudies that have found an association of the micro-satellite marker D19S884 which is located on chromo-some 19p13.2 and relatively close (1 cM) to INSR withPCOS, emphasize it to be a candidate gene (18, 19).Other studies showed that the number and affinity ofinsulin receptor is not altered in PCOS but its tyrosinephosphorylation status and subsequent signaling is

DOI: 10.1530/EJE-08-0932

Online version via www.eje-online.org

856 S Mukherjee and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2009) 160

affected, suggesting the defect may lie in the b-chain(16, 20). Studies to determine mutations in INSR failedto find any major variations; however, several poly-morphisms were identified (21–28). The most frequentof these were at exon 17, which encodes the partialtyrosine kinase domain containing the ATP bindingsite of INSR, important for its downstream signaling(21–28). Among these polymorphisms, a C/T SNPat His1058 in exon 17 has been reported to besignificantly associated with PCOS in two independentstudies in Caucasian and Chinese women (25, 26).A subsequent study from Korea however, failed toconfirm this association (27). Also, the possiblerelationship between this polymorphism and PCOSassociated traits has not been explored so far. Inanother study, a novel T/C polymorphism at Cys1008in exon 17 of INSR was found to be associated withPCOS and insulin resistance, which is yet to be verifiedindependently (28). Nevertheless, these studies dosuggest that the genetic variants in exon 17 of INSRmay have an association with PCOS or its pathophy-siology. Thus, in the present study, we sought toevaluate the association of genetic variations in exon17 of INSR with PCOS and its related traits in Indianwomen, who are ethnically distinct from the popu-lations already studied.

Subjects and methods

Subjects

A total of 180 Indian women with PCOS, aged 17–39years were recruited consecutively from the InfertilityClinic of the National Institute for Research inReproductive Health (NIRRH) as well as the Endo-crinology Clinic of Seth GS Medical College and KEMHospital, Mumbai, India. The diagnosis of PCOS wasdefined by ESHRE/ASRM consensus criteria, includingpresence of at least two of the following three features:i) oligomenorrhoea and/or anovulation, ii) clinicaland/or biochemical signs of hyperandrogenemia, andiii) polycystic ovaries on ultrasound (29). Other relateddisorders like non-classical congenital adrenal hyper-plasia, thyroid dysfunction, androgen secreting tumors,and hyperprolactinemia were excluded. One hundredand fifty healthy Indian women, aged 17–38 years, withregular menstrual cycles and no clinical and/orbiochemical signs of hyperandrogenemia or polycysticovaries were carefully selected for the study as thecontrol group from the general population. On detailedbiochemical evaluation, six control women wereexcluded mainly due to abnormal thyroid status. Thewomen in the control and PCOS groups were geneticallyunrelated. Women with diabetes were excluded from thestudy. None of the subjects had taken any medicationknown to affect carbohydrate and lipid metabolism orendocrine parameters for at least 3 months prior to

www.eje-online.org

entering the study. Ethical committee approval fromboth institutions and written informed consent from allparticipants were obtained.

Clinical and laboratory parameters

The anthropometric data – height, weight, body massindex (BMI), waist, and hip circumference, waist to hipratiowere obtained from all subjects. Blood samples werecollected between 3–7 days of the menstrual cycle (earlyfollicular phase) or during amenorrhea and serum wasstored at K80 8C until assayed. After overnight fasting,each subject underwent an oral glucose tolerance test.Glucose and insulin levels were measured at 0 min and2 h after 75 g glucose load. Fasting serum were used tomeasure total testosterone, androstenedione, sex hor-mone binding globulin (SHBG), LH, FSH, TSH, T3, T4,and prolactin levels. Glucose was measured enzy-matically by glucose oxidase method. TSH, T3, T4,FSH, LH, and prolactin were measured by chemi-luminescence immunoassay (Immulite 1000, Llanberis,UK). Serum insulin, total testosterone, androstenedione,and SHBG levels were measured using commercial kits(DSL-1600 RIA kit, DSL-4000 RIA kit, DSL-3800 RIAkit and DSL-7400 IRMA kit respectively from DiagnosticSystem Laboratories, Webster, TX, USA). Free testoster-one (active free testosterone in circulation) and bioavail-able testosterone (free plus weakly bound to albumin)was calculated from total testosterone and SHBG valuesusing a web-based calculator (http://www.issam.ch/freetesto.htm) (30). Free-androgen index (FAI) wasassessed by the following formula: total testosterone(nmol/l)/SHBG (nmol/l) ! 100 (31). Insulin resistancewas evaluated by the homeostasis model assessment ofinsulin resistance (HOMA-IR) using the followingformula: fasting plasma glucose (mmol/l) ! fastinginsulin (mU/ml)/ 22.5) (32). Insulin sensitivity wasassessed by quantitative insulin sensitivity check index(QUICKI), defined as 1/(log fasting serum insulin(mU/ml) C log fasting plasma glucose (mg/dl)) (33).Obesity was defined as BMIR23 kg/m2 based on Indiannorms (34), accordingly, subjects were categorized aslean (BMI !23 kg/m2) and obese (BMI R23 kg/m2).

Genetic analysis

Genomic DNA was extracted from peripheral wholeblood by QIAamp DNA Blood Mini Kit (Qiagen GmbH).Exon 17 was amplified using the following primers:5 0-CCAAGGATGCTGTGTAGATAAG-3 0 and 5 0-TCAG-GAAAGCCAGCCCATGTC-3 0 according to Siegel et al.with some modifications (25). The amplifications wereperformed using 0.5 mg genomic DNA in a final volumeof 50 ml containing 1.5 mM MgCl2, 200 mM of eachdNTPs, 10 pmol of each primer and 0.5 U of TaqPolymerase (Fermentas, Vilnius, Lithuania). After initialdenaturation for 5 min at 94 8C, 35 cycles of amplifica-tions were carried out at 94 8C for 45 s, 56 8C for 40 s,

Insulin receptor gene polymorphism in PCOS 857EUROPEAN JOURNAL OF ENDOCRINOLOGY (2009) 160

and at 72 8C for 1 min with a final extension of 10 minat 72 8C. The resulting amplified product was of 317 bp,which was purified by QIAquick Gel Extraction Kit,(Qiagen GmbH) and subjected to direct sequencing in3130xl Genetic Analyzer using Big Dye TerminatorChemistry (Applied Biosystems, Foster City, USA).

Statistical analysis

Univariate comparison of all continuous variablesbetween the PCOS and control groups were done byunpaired t-tests. Results are expressed as meanGS.D.The association between genotype and PCOS wasanalyzed using c2 test. Stratified analysis by obesity(i.e., lean BMI !23 kg/m2 and obese BMI R23 kg/m2)was also performed. Variables that were significant at theunivariate level were further assessed in multivariableanalysis. Analysis of covariance (ANCOVA) was used tocompare hormonal profiles between genotypes afteradjusting for age and BMI. Statistical significance wasbased on a P!0.05. Statistical analysis was performedby using SAS version 9.2 software packages.

Results

The clinical, anthropometric, and hormonal parametersof control and PCOS women per se and as per obesitystatus have been depicted in Table 1. The majority of thewomen with PCOS were oligomenorrhoeic (55%,99/180), whereas 38.9% (70/180) were amenorrhoeicand the rest 6.1% (11/180) were regularly cycling. Asregards clinical signs of hyperandrogenism, 53.3%(96/180) of PCOS women presented with hirsutism,36.1% (65/180) with acne and 16% (29/180) withmild alopecia.

The control and PCOS women were of similar meanage but did differ significantly with respect toanthropometric parameters (P!0.05) (Table 1). ThePCOS women had significantly higher levels of fastingand 2 h glucose and fasting insulin compared withcontrols, although within normal range. Insulinresistance index like HOMA-IR (P!0.0001) waselevated and insulin sensitivity as measured by QUICKI(vs P!0.0001) was low in PCOS women (Table 1).Even after classifying control and PCOS women in leanand obese groups, the BMI, 2 h glucose level, fastinginsulin level, and HOMA-IR remained higher, whereasQUICKI was lower in both PCOS groups compared withcontrols. Acanthosis nigricans, a sign of insulinresistance, was present in 23.3% of PCOS women(42/180).

Women with PCOS were more hyperandrogenicexhibiting higher levels of total, free, and bioavailabletestosterone, androstenedione, and increased FAI values(P!0.0001). The difference between these hormoneswas more evident in the obese rather than in the leangroup (Table 1). SHBG levels were significantly lower

in obese PCOS women (P!0.0001), which is commonin obesity and insulin resistance state (35), withconcurrent high free and bioavailable testosterone(P!0.0001) and FAI (P!0.0001).

Analysis of exon 17 of INSR by direct sequencingrevealed presence of one silent polymorphism, a C/TSNP at His1058 site. No other known or novelpolymorphism in this region was detected in thiscohort. The genotype distributions of this SNP were inHardy–Weinberg equilibrium. In our study cohort,52.8% (76/144) of controls and 43.9% (79/180) ofPCOS were CC; 38.9% (56/144) of controls and 42.8%(77/180) of PCOS were CT and 8.3% (12/144) controlsand 13.3% (24/180) PCOS women were TT. Thefrequencies of CT and TT genotypes were relativelyhigh but not significantly different in PCOS womencompared with controls (c2Z3.416, dfZ2, PZ0.181).

PCOS women are frequently obese which alsocontributes to insulin resistance. To understand therelationship of this INSR polymorphism with obesity,the distribution of these genotypes in lean and obesegroups were evaluated. The frequency of CTCTTgenotype was significantly higher in lean PCOS group(69.3%, 52/75) compared with lean control group(46.4%, 39/84) (c2Z8.493, dfZ1, PZ0.004)(Table 1). As the number of subjects with TT genotypewas low in some subgroups, CT and TT genotypes werecombined together as CTCTT genotype.

To understand the influence of this polymorphism onPCOS related traits, indices of insulin resistance andhyperandrogenemia in both lean and obese groups ofPCOS and controls were analyzed (Tables 2 and 3). Thecarriers of CC genotype did not differ significantly frompolymorphic (CTCTT) genotype in lean control, obesecontrol, and also in obese PCOS group, in terms of waistcircumference, BMI, fasting glucose levels, fastinginsulin levels and insulin resistance indices (HOMA-IRand QUICKI). Additionally, total testosterone levels, freeand bioavailable testosterone, androstenedione, and FAIwere also similar within these groups. Interestingly, thelean PCOS subgroup, with polymorphic genotypeshowed a significantly higher BMI (20.48G1.77 vs19.15G2.07 kg/m2, PZ0.005) and waist circumfer-ence (73.24G6.43 vs 69.89G5.64 cm, PZ0.03)compared with CC genotype. Furthermore, the indicesof insulin resistance were significantly different betweenwomen with polymorphic and wildtype genotypes in thesame group, showing higher fasting insulin levels(11.43G5.22 vs 8.32G2.70 uIU/ml, PZ0.009) andHOMA-IR (2.40G1.10 vs 1.64G0.54,PZ0.002) andlower QUICKI values (0.340G0.021 vs 0.357G0.018,PZ0.002) in the polymorphic subgroup (Table 2). Alsoregarding androgen parameters, these women withpolymorphic genotype exhibited elevated free testoster-one (22.81G9.91 vs 17.95G6.87 pmol/l, PZ0.04),bioavailable testosterone (0.54G0.23 vs 0.42G0.16 nmol/l, PZ0.04), and FAI (2.94G1.50 vs2.16G0.82, PZ0.02; Table 2). The variables that

www.eje-online.org

Table

1C

linic

aland

horm

onalpro

file

of

contr

oland

poly

cystic

ovary

syndro

me

(PC

OS

)w

om

en.

Variables

Allwomen

Leanwomen

(nZ

159)

Obesewomen

(nZ

165)

Contr

ol(nZ

144)

PC

OS

(nZ

180)

PC

ontr

ol(nZ

84)

PC

OS

(nZ

75)

PC

ontr

ol(nZ

60)

PC

OS

(nZ

105)

P

Age

(years

)24.9

4G

5.4

624.8

2G

5.2

60.8

423.4

2G

4.8

23.2

8G

4.2

50.8

527.1

2G

5.5

825.9

2G

5.6

40.1

9B

ody

weig

ht

(kg)

53.4

2G

10.4

458.5

9G

14.0

40.0

003

47.0

0G

7.2

347.0

5G

6.1

01

62.7

1G

6.9

066.9

8G

12.0

40.0

1B

MI

(kg/m

2)

22.1

6G

4.1

125.0

1G

5.6

3!

0.0

001

19.3

1G

2.3

420.0

7G

1.9

50.0

326.2

0G

2.2

728.5

3G

4.6

70.0

004

Wais

t(c

m)

76.3

6G

9.0

582.0

2G

12.0

2!

0.0

001

71.7

3G

7.5

72.2

1G

6.3

50.7

282.8

6G

6.4

089.0

2G

10.0

9!

0.0

001

WH

R0.8

0G

0.0

50.8

2G

0.0

5!

0.0

001

0.7

8G

0.0

50.8

0G

0.0

50.0

60.8

1G

0.0

40.8

3G

0.0

50.0

006

FB

S(m

g/d

l)83.3

6G

7.9

686.7

0G

10.4

70.0

02

82.9

4G

7.2

685.1

1G

9.4

60.1

183.9

7G

8.8

887.8

3G

11.0

30.0

2T

wo-h

our

glu

cose

(mg/d

l)92.8

0G

14.9

8103.5G

21.7

2!

0.0

001

90.6

9G

14.2

296.7

2G

18.9

70.0

295.8

0G

15.6

3108.3G

22.3

50.0

002

Fasting

insulin

(uIU

/ml)

8.9

6G

3.3

213.6

8G

7.2

6!

0.0

001

8.1

5G

2.8

910.5

4G

4.8

10.0

002

10.1

0G

3.5

715.9

2G

7.8

8!

0.0

001

HO

MA

-IR

1.8

4G

0.7

02.8

4G

1.6

0!

0.0

001

1.6

6G

0.5

92.1

8G

1.0

20.0

001

2.0

9G

0.7

53.3

1G

1.7

5!

0.0

001

QU

ICK

I0.3

53G

0.0

21

0.3

35G

0.0

26

!0.0

001

0.3

58G

0.0

21

0.3

45G

0.0

22

0.0

003

0.3

46G

0.0

20

0.3

27G

0.0

26

!0.0

001

FS

H(m

U/m

l)6.6

3G

1.8

66.5

6G

1.9

80.6

86.7

8G

2.1

76.4

1G

2.0

80.2

86.6

7G

1.3

46.6

7G

1.9

61

LH

(mU

/ml)

4.9

3G

1.8

39.9

3G

4.5

7!

0.0

001

5.1

9G

1.9

610.9

5G

4.6

3!

0.0

001

4.5

6G

1.5

79.2

0G

4.4

1!

0.0

001

LH

:F

SH

0.7

6G

0.3

11.6

6G

0.9

0!

0.0

001

0.7

9G

0.2

91.8

5G

0.9

1!

0.0

001

0.7

3G

0.3

41.5

2G

0.8

6!

0.0

001

TT

(ng/d

l)49.1

1G

14.5

274.7

9G

31.1

8!

0.0

001

48.5

5G

15.9

468.9

7G

25.5

9!

0.0

001

49.9

2G

12.2

978.9

4G

34.1

4!

0.0

001

Bio

-T(n

mol/l)

0.3

5G

0.1

30.6

9G

0.3

9!

0.0

001

0.3

3G

0.1

30.5

0G

0.2

2!

0.0

001

0.3

7G

0.1

20.8

2G

0.4

3!

0.0

001

FT

(pm

ol/l)

14.3

7G

5.4

829.4

1G

16.5

4!

0.0

001

13.7

8G

5.5

221.4

6G

9.3

4!

0.0

001

15.2

0G

5.3

635.0

8G

18.1

8!

0.0

001

FA

I1.8

2G

0.7

64.2

0G

2.9

6!

0.0

001

1.7

5G

0.7

82.7

3G

1.3

8!

0.0

001

1.9

1G

0.7

15.2

4G

3.3

2!

0.0

001

SH

BG

(nm

ol/l)

102.4G

35.1

079.4

8G

39.5

2!

0.0

001

105.7G

34.1

197.1G

32.5

60.1

199.0

3G

36.7

366.8

6G

39.3

4!

0.0

001

Andro

ste

nedio

ne

(ng/m

l)1.0

5G

0.5

62.2

6G

1.2

3!

0.0

001

1.0

9G

0.5

42.0

4G

1.0

4!

0.0

001

1.0

1G

0.5

62.3

6G

1.1

7!

0.0

001

Genoty

pe

(%of

CC

)52.8

43.9

0.1

1*

53.6

30.7

0.0

04*

51.7

53.3

0.8

7*

Data

are

giv

en

as

meanG

S.D

.,B

MI,

body

mass

index;

WH

R,

wais

t-to

-hip

ratio;

FB

S,

fasting

glu

cose;

HO

MA

-IR

,hom

eosta

sis

modelassessm

ent

for

insulin

resis

tance;

QU

ICK

I,quantita

tive

insulin

sensitiv

ity

check

index;T

T,to

talt

esto

ste

rone;F

T,fr

ee

testo

ste

rone;B

io-T

,bio

availa

ble

testo

ste

rone;S

HB

G,sex

horm

one

bin

din

gglo

bulin

and

FA

I,fr

ee

andro

gen

index.C

om

parison

was

done

by

unpairedt-

test.P!

0.0

5is

sta

tistically

sig

nifi

cant.

Genoty

pe

ispre

sente

das

perc

ent

of

CC

cases

inth

at

cate

gory

.*T

heP

-valu

es

for

genoty

pes

are

from

c2-t

est.

858 S Mukherjee and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2009) 160

www.eje-online.org

Table 2 Insulin resistance indices and hormonal profile of lean and obese women with polycystic ovary syndrome (PCOS) according togenotype.

Variables

Lean PCOS (nZ75) Obese PCOS (nZ105)

CC (nZ23) CTCTT (nZ52) P * P† CC (nZ56) CT CTT (nZ49) P *

Age (years) 21.87G4.15 23.90G4.18 0.06 25.63G5.50 26.27G5.82 0.56Waist (cm) 69.89G5.64 73.24G6.43 0.03 0.43 88.60G9.90 89.50G10.38 0.65BMI (kg/m2) 19.15G2.07 20.48G1.77 0.005 28.79G5.22 28.23G3.97 0.54FBS (mg/dl) 84.30G6.77 85.26G10.47 0.68 87.22G11.47 88.53G10.58 0.54Two-hour glucose (mg/dl) 93.04G7.50 98.36G19.34 0.26 107.0G20.04 109.9G24.85 0.51Fasting insulin (uIU/ml) 8.32G2.7 11.43G5.22 0.009 0.02 15.96G8.53 15.87G7.15 0.95HOMA-IR 1.64G0.54 2.40G1.10 0.002 0.005 3.20G1.73 3.43G1.80 0.52QUICKI 0.357G0.018 0.340G0.021 0.002 0.004 0.328G0.029 0.325G0.023 0.56TT (ng/dl) 64.96G26.36 70.11G25.48 0.43 79.20G34.71 78.65G33.83 0.94Bio-T (nmol/l) 0.42G0.16 0.54G0.23 0.04 0.057 0.82G0.41 0.83G0.45 0.85FT (pmol/l) 17.95G6.87 22.81G9.91 0.04 0.056 34.77G17.35 35.44G19.27 0.85FAI 2.16G0.82 2.94G1.50 0.02 0.03 5.16G3.06 5.38G3.63 0.73SHBG (nmol/l) 107.3G29.21 92.42G32.94 0.07 68.52G40.59 64.96G38.20 0.65Androstenedione (ng/ml) 1.92G0.81 2.07G1.14 0.57 2.33G1.13 2.38G1.24 0.80

Data are given as mean GS.D. BMI, body mass index; FBS, fasting glucose; HOMA-IR, homeostasis model assessment for insulin resistance; QUICKI,quantitative insulin sensitivity check index; TT, total testosterone; FT, free testosterone, Bio-T, bioavailable testosterone; SHBG, sex hormone binding globulinand FAI, free androgen index. *P-values are from unpaired t-test. The significant variables at the univariate level were further assessed in an ANCOVA model.†P-values are from ANCOVA after adjusting for age and BMI. P!0.05 is statistically significant.

Insulin receptor gene polymorphism in PCOS 859EUROPEAN JOURNAL OF ENDOCRINOLOGY (2009) 160

showed significant association with genotype in uni-variate analysis were further evaluated by ANCOVAmodel, in the lean PCOS group. In multivariate analysisafter adjusting for age and BMI, the difference in fastinginsulin (PZ0.02), HOMA-IR (PZ0.005), QUICKI(PZ0.004) and FAI (PZ0.03) between polymorphicand wildtype genotype remained significant, whereasfree and bioavailable testosterone were marginallysignificant (Table 2).

Discussion

To the best of our knowledge, this is the first study fromIndia, reporting the presence of the C/T polymorphismat His1058 in exon 17 of INSR. The polymorphism

Table 3 Insulin resistance indices and hormonal profile of lean and o

Variables

Lean control (nZ84)

CC (nZ45) CT CTT (nZ39

Age (years) 22.65G5.08 23.93G4.61Waist (cm) 71.33G8.06 72.03G6.08BMI (kg/m2) 19.45G2.33 19.14G2.37FBS (mg/dl) 82.43G7.99 83.88G6.31Two-hour glucose (mg/dl) 91.24G13.35 89.45G15.45Fasting insulin (uIU/ml) 8.00G2.62 8.37G3.14HOMA-IR 1.62G0.52 1.73G0.66QUICKI 0.358G0.019 0.357G0.021TT (ng/dl) 48.67G16.63 48.38G15.18Bio-T (nmol/l) 0.32G0.13 0.37G0.13FT (pmol/l) 13. 21G5.50 14.40G5.61FAI 1.66G0.72 1.85G0.84SHBG (nmol/l) 110.0G34.70 101.5G33.43Androstenedione (ng/ml) 1.12G0.60 1.05G0.53

Data are given as mean GS.D. BMI, body mass index; FBS, fasting glucose;quantitative insulin sensitivity check index; TT, total testosterone; FT, free testostFAI, free androgen index. Comparison was done by unpaired t-test. P!0.05 is

showed a strong association with PCOS in lean women.Detailed analysis also revealed an association of thispolymorphism with indices of insulin resistance (fastinginsulin levels, HOMA-IR, and QUICKI) and hyperan-drogenemia (FAI) in the same subgroup. It may bementioned here that our study cohort is larger than theother published reports which examined exon 17polymorphism of INSR in PCOS women.

It is well established that insulin resistance, PCOS,and obesity are interrelated (36, 37). Interestingly,stratifying the study cohort into lean and obese groups,the prevalence of polymorphic genotype was signi-ficantly higher in the lean PCOS group compared withlean control, obese PCOS as well as obese controlgroups. The analysis clearly establishes an association

bese control women according to genotype.

Obese control (nZ60)

) P CC (nZ31) CTCTT (nZ29) P

0.49 26.42G5.34 27.70G5.78 0.370.67 83.18G5.41 83.23G8.37 0.980.54 26.37G2.27 26.13G2.32 0.690.36 84. 00G9.22 83.43G9.07 0.800.58 97.03G16.89 95.37G14. 51 0.660.55 9.98G3.61 10.15G3.54 0.850.38 2.12G0.83 2. 05G0.67 0.730.70 0.347G0.020 0.346G0.018 0.790.93 47.77G11.89 51.93G12.14 0.180.54 0.35G0.10 0.39G0.13 0.230.32 14.06G4.85 16.29G5.54 0.100.26 1.83G0.65 1.99G0.76 0.370.25 98.87G39.47 98.93G28.34 1.00.60 0.96G0.52 1.05G0.60 0.55

HOMA-IR, homeostasis model assessment for insulin resistance; QUICKI,erone, Bio-T, bioavailable testosterone; SHBG, sex hormone binding globulin;statistically significant.

www.eje-online.org

860 S Mukherjee and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2009) 160

of the polymorphic allele with PCOS in lean women,thus corroborating the previous findings in Caucasianand Chinese women (25, 26). A significantly higherfrequency of CTCTT genotype was reported in leanCaucasian PCOS women compared with lean controls(47 vs 29%, PZ0.03) (25). Similarly, prevalence ofpolymorphic allele was more in non-obese ChinesePCOS women compared with obese PCOS women (52.2vs 25.5%, P!0.01) (26). In our study, the frequency ofpolymorphic genotype in lean PCOS (69.3%) washigher than the frequencies reported earlier, whichmay be attributed to ethnic variation. A similar Koreanstudy had not found any such association (27).

It is known that insulin resistance is present in PCOSwomen irrespective of obesity (38, 39, Mukherjeeunpublished observations). However, in the presentstudy, an association of this polymorphism in INSRwithPCOS only in lean women was observed. This promptedus to investigate the relationship of His1058 genotypeswith PCOS associated traits. Univariate analysisrevealed the association of polymorphic genotype withBMI, waist circumference, fasting insulin levels, HOMA-IR, and QUICKI, only in lean PCOS group. It isinteresting that lean PCOS women with polymorphicgenotype had significantly higher waist circumference,which is a measure of central obesity compared withwildtype genotype. It has now been reported that thecentral obesity is more in Asian Indians, causing higherinsulin resistance despite having lean BMI (34). Indeed,the same subgroup of PCOS women also had higherindices of insulin resistance. Hyperinsulinemia is knownto exacerbate PCOS phenotypes by inducing hyperan-drogenemia, through excess ovarian androgen pro-duction and suppressing hepatic SHBG synthesis (2,40). In keeping with this evidence, the lean PCOSsubgroup with polymorphic genotype also showedincreased free bioavailable testosterone and FAI. Evenwith adjustments for age and BMI in multivariateanalysis, this association persisted with respect toindices of insulin resistance and hyperandrogenemiain lean PCOS. These associations suggest that thepathogenesis of insulin resistance in PCOS may differamong lean and obese women. Dunaif et al. alsoobserved that insulin resistance in 50% of PCOSwomen is associated with decreased tyrosine autopho-sphorylation and increased serine phosphophorylationof INSR (2). Indeed, this emphasizes that molecularmechanism of insulin resistance may differ in differentPCOS group. Improvement of hyperinsulinemia andhyperandrogenemia in PCOS women by treatment withinsulin sensitizer metformin occurs through differentmechanisms in obese and lean women (41), whichagain supports the notion that the pathogenicmechanism is different in these two groups. A recentstudy reported the presence of three subpopulationswith distinct levels of insulin resistance in PCOS womenwith significantly different BMI (39). With our presentdata, it is not possible to identify such subpopulations.

www.eje-online.org

However, the association of His1058 SNP with leanPCOS emphasizes that the genetic determinants of thesesubpopulations might be different.

Beyond His1058, few other polymorphisms also havebeen reported in exon 17 of INSR in women with PCOS(21, 23, 28). A novel T/C polymorphism at Cys1008has been reported to be associated with PCOS anddecreased insulin sensitivity in Chinese population (28).Studies reported another silent C/T variation in Tyr984in PCOS women as well as in diabetic subjects (21, 23,42). However, none of these polymorphisms or othernovel variation in exon 17 of INSR was found in thisstudy cohort.

PCOS is now considered as a complex multigenictrait, where different gene variants interact with eachother and also with environmental factors in themanifestation of different phenotypic expressions ofthe syndrome. It is possible that different subgroups ofPCOS with specific phenotypic expression are associatedwith specific gene variants. Since the C/T SNP atHis1058 is a silent one, it cannot exert a major effect onthe development of insulin resistance; rather it might bein linkage disequilibrium with some other geneticvariants, which may play a direct role in thedevelopment of PCOS by affecting insulin sensitivity inlean women. This polymorphism may predispose leanwomen to develop insulin resistance and compensatoryhyperinsulinemia, which then may induce hyperan-drogenemia progressing towards PCOS. The lowerprevalence of the T allele in obese compared with leanwomen could also suggest that this allele either has aprotective role against obesity in lean women or mayreduce the risk of developing PCOS in obese women.

To summarize, this study in Indian women shows anassociation of the C/T polymorphism at His1058 ofINSRwith PCOS in lean rather than obese women. Thepolymorphism is also seen to be associated with theindices of insulin resistance and hyperandrogenemia inthe lean PCOS group. The observations made hereinstrongly suggest that the genetic pathogenicmechanism of this heterogeneous disorder may differbetween lean and obese PCOS, which needs to bestudied further.

Declaration of interest

The authors declare that there is no conflict of interest that could beperceived as prejudicing the impartiality of the research reported.

Funding

This work was supported by a grant from Department of Science andTechnology, India (SR/SO/HS-60/2005).

Acknowledgements

We thank Dr Mousumi Banerjee, Department of Biostatistics,University of Michigan, and Ann Arbor, Michigan, USA for her helpin statistical analysis. We acknowledge Dr Padma Menon and the

Insulin receptor gene polymorphism in PCOS 861EUROPEAN JOURNAL OF ENDOCRINOLOGY (2009) 160

Endocrine Department of Seth GS Medical College and KEM Hospital,Mumbai, India for their contributions. We also thank Mr ChinnarajSaravanan (DNA sequencing core facility, NIRRH) for his help in DNAsequencing.

References

1 Diamanti-Kandarakis E. Polycystic ovarian syndrome: pathophy-siology, molecular aspects and clinical implications. Expert Reviewsin Molecular Medicine 2008 10 e3.

2 Dunaif A. Insulin resistance and the polycystic ovary syndrome:mechanism and implications for pathogenesis. Endocrine Reviews1997 18 774–800.

3 Franks S. Polycystic ovary syndrome. New England Journal ofMedicine 1995 333 853–861.

4 Azziz R, Woods KS, Reyna R, Key TJ, Knochenhauer ES & Yildiz BO.The prevalence and features of polycystic ovary syndrome in anunselected population. Journal of Clinical Endocrinology andMetabolism 2004 89 2754–2759.

5 Legro RS, Kunselman AR, DodsonWC & Dunaif A. Prevalence andpredictors for risk for type 2 diabetes mellitus and impaired glucosetolerance in polycystic ovary syndrome: a prospective, controlledstudy in 254 affected women. Journal of Clinical Endocrinology andMetabolism 1999 84 165–169.

6 Diamanti-Kandarakis E. Insulin resistance in PCOS. Endocrine2006 30 13–17.

7 Legro RS, Finegood D & Dunaif A. Changes in glucose toleranceover time in women with polycystic ovary syndrome: a controlledstudy. Journal of Clinical Endocrinology and Metabolism 2005 903236–3242.

8 Wild RA, Grubb B, Hartz A, Van Nort JJ, Bachman W &Bartholomew M. Clinical signs of androgen excess as risk factorsfor coronary artery disease. Fertility and Sterility 1990 54255–259.

9 Maitra A, Meherji PK, Desai MP, Gokral JS, Donde UM & Joshi UM.Lipoprotein in androgen excess – a study among Indian women.International Journal of Fertility 1994 39 218–222.

10 Maitra A, Pingle RR, Menon PS, Naik V, Gokral JS & Meherji PK.Dyslipidemia with particular regard to apolipoprotein profile inassociation with polycystic ovary syndrome: a study amongIndian women. International Journal of Fertility and Women’sMedicine 2001 46 271–277.

11 Bjornthorp P. Hyperandrogenicity in women – a prediabeticcondition? Journal of Internal Medicine 1993 234 609–613.

12 Lunde O, Magnus P, Sandvik L & Høglo S. Familial clustering in thepolycystic ovarian syndrome. Gynecologic and Obstetric Investigation1989 28 23–30.

13 Yildiz BO, Yarali H, Oguz H & Bayraktar M. Glucose intolerance,insulin resistance, and hyperandrogenemia in first degree relativesof women with polycystic ovary syndrome. Journal of ClinicalEndocrinology and Metabolism 2003 88 2031–2036.

14 Escobar-Morreale HF, Luque-Ramırez M & San Millan JL. Themolecular-genetic basis of functional hyperandrogenism and thepolycystic ovary syndrome. Endocrine Reviews 2005 26 251–282.

15 Diamanti-Kandarakis E, Piperi C, Spina J, Argyrakopoulou G,Papanastasiou L, Bergiele A & Panidis D. Polycystic ovarysyndrome: the influence of environmental and genetic factors.Hormones 2006 5 17–34.

16 Dunaif A, Xia J, Book CB, Schenker E & Tang Z. Excessive insulinreceptor serine phosphorylation in cultured fibroblasts and inskeletal muscle, A potential mechanism for insulin resistance inthe polycystic ovary syndrome. Journal of Clinical Investigation1995 96 801–810.

17 Diamanti-Kandarakis E & Papavassiliou AG. Molecularmechanisms of insulin resistance in polycystic ovary syndrome.Trends in Molecular Medicine 2006 12 324–332.

18 Urbanek M, Legro RS, Driscoll DA, Azziz R, Ehrmann DA,Norman RJ, Strauss JF 3rd, Spielman RS & Dunaif A. Thirty-sevencandidate genes for polycystic ovary syndrome: strongest evidencefor linkage is with follistatin. PNAS 1999 96 8573–8578.

19 Tucci S, Futterweit W, Concepcion ES, Greenberg DA,Villanueva R, Davies TF & Tomer Y. Evidence for association ofpolycystic ovary syndrome in caucasian women with a marker atthe insulin receptor gene locus. Journal of Clinical Endocrinology andMetabolism 2001 86 446–449.

20 Ciaraldi TP, el-Roeiy A, Madar Z, Reichart D, Olefsky JM & Yen SS.Cellular mechanisms of insulin resistance in polycystic ovariansyndrome. Journal of Clinical Endocrinology and Metabolism 199275 577–583.

21 Talbot J, Bicknell E, Rajkhowa M, Krook A, O’Rahilly S &Clayton RN. Molecular scanning of the insulin receptor gene inwomen with polycystic ovarian syndrome. Journal of ClinicalEndocrinology and Metabolism 1996 81 1979–1983.

22 Krook A, Kumar S, Laing I, Boulton AJ, Wass JA & O’Rahilly S.Molecular scanning of the insulin receptor gene in syndromes ofinsulin resistance. Diabetes 1994 43 357–368.

23 Moller DE, Cohen O, Yamaguchi Y, Azziz R, Grigorescu F, Eberle A,Morrow LA, Moses AC & Flier JS. Prevalence of mutations in theinsulin receptor gene in subjects with features of the type Asyndrome of insulin resistance. Diabetes 1994 43 247–255.

24 Panz VR, Ruff P, Joffe BI, Kedda MA & Seftel HC. SSCP analysis ofthe tyrosine kinase domain of the insulin receptor gene:polymorphisms detected in South African black and whitesubjects. Human Genetics 1996 7 438–440.

25 Siegel S, Futterweit W, Davies TF, Concepcion ES, Greenberg DA,Villanueva R & Tomer Y. A C/T single nucleotide polymorphismat the tyrosine kinase domain of the insulin receptor geneis associated with PCOS. Fertility and Sterility 2002 781240–1243.

26 Chen ZJ, Shi YH, Zhao YR, Li Y, Tang R, Zhao LX & Chang ZH.Correlation between single nucleotide polymorphism of insulinreceptor gene with polycystic ovary syndrome. Zhonghua Fu ChanKe Za Zhi 2004 39 582–585.

27 Lee EJ, Yoo KJ, Kim SJ, Lee SH, Cha KY & Baek KH. Singlenucleotide polymorphism in exon 17 of the insulin receptor gene isnot associated with polycystic ovary syndrome in a Koreanpopulation. Fertility and Sterility 2006 86 380–384.

28 Jin L, Zhu XM, Luo Q, Qian Y, Jin F & Huang HF. A novel SNP atexon 17 of INSR is associated with decreased insulin sensitivity inChinese women with PCOS. Molecular Human Reproduction 200612 151–155.

29 The Rotterdam ESHRE/ASRM-sponsored PCOS consensus work-shop group. Revised 2003 consensus on diagnostic criteria andlong-term health risks related to polycystic ovary syndrome(PCOS). Human Reproduction 2004 19 41–47.

30 Vermeulen A, Verdonck L & Kaufman JM. A critical evaluation ofsimple methods for the estimation of free testosterone in serum.Journal of Clinical Endocrinology and Metabolism 1999 843666–3672.

31 Clark AF, Marcellus S, deLory B & Bird CE. Plasma testosterone freeindex: a better indicator of plasma androgen activity? Fertility andSterility 1975 26 1001–1005.

32 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF &Turner RC. Homeostasis model assessment: insulin resistance andbeta-cell function from fasting plasma glucose and insulinconcentrations in man. Diabetologia 1985 28 412–419.

33 Katz A, Nambi SS, Mather K, Baron AD, Follmann DA, Sullivan G& Quon MJ. Quantitative insulin sensitivity check index: a simple,accurate method for assessing insulin sensitivity in humans.Journal of Clinical Endocrinology and Metabolism 2000 852402–2410.

34 Snehalatha C, Viswanathan V & Ramachandran A. Cutoff valuesfor normal anthropometric variables in Asian Indian adults.Diabetes Care 2003 26 1380–1384.

35 Sikaris KA. The clinical biochemistry of obesity more than skindeep. Heart, Lung and Circulation 2007 16 S45–S50.

www.eje-online.org

862 S Mukherjee and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2009) 160

36 Allahbadia GN & Merchant R. Polycystic ovary syndrome in theIndian subcontinent. Seminars in Reproductive Medicine 2008 2622–34.

37 Gambineri A, Pelusi C, Vicennati V, Pagotto U & Pasquali R.Obesity and the polycystic ovary syndrome. International Journal ofObesity and Related Metabolic Disorders 2002 26 883–896.

38 Dunaif A, Segal KR, Futterweit W & Dobrjansky A. Profoundperipheral insulin resistance, independent of obesity, in polycysticovary syndrome. Diabetes 1989 38 1165–1174.

39 Vigil P, Contreras P, Alvarado JL, Godoy A, Salgado AM &Cortes ME. Evidence of subpopulations with different levels ofinsulin resistance in women with polycystic ovary syndrome.Human Reproduction 2007 11 2974–2980.

40 Tsilchorozidou T, Overton C & Conway GS. The pathophysiology ofpolycystic ovary syndrome. Clinical Endocrinology 2004 60 1–17.

www.eje-online.org

41 Morin-Papunen L, Vauhkonen I, Koivunen R, Ruokonen A,Martikainen H & Tapanainen JS. Metformin versus ethinylestradiol-cyproterone acetate in the treatment of nonobesewomen with polycystic ovary syndrome: a randomized study.Journal of Clinical Endocrinology and Metabolism 2003 88148–156.

42 Elbein SC & Sorensen LK. Genetic variation in insulin receptorbeta-chain exons among members of familial type 2 (non-insulin-dependent) diabetic pedigrees. Diabetologia 1991 34742–749.

Received 20 January 2009

Accepted 7 February 2009