Differential expression of inflammation-related genes in the ovarian ...

3 downloads 0 Views 423KB Size Report
days initiated on Days 2–3, and oocyte recovery was performed on Days. 8–9 after ..... prostaglandin-endoperoxide synthase 2 (PTGS2)] were among the over-.
Molecular Human Reproduction, Vol.20, No.1 pp. 49– 58, 2014 Advanced Access publication on July 29, 2013 doi:10.1093/molehr/gat051

ORIGINAL RESEARCH

Differential expression of inflammationrelated genes in the ovarian stroma and granulosa cells of PCOS women Johanna Schmidt 1,*, Birgitta Weijdega˚rd 1, Anne Lis Mikkelsen2,3, Svend Lindenberg 4, Lars Nilsson 1, and Mats Bra¨nnstro¨m1 1 Department of Obstetrics and Gynaecology, Sahlgrenska Academy, University of Gothenburg, SE-41345 Go¨teborg, Sweden 2The Fertility Clinic, Holbaek Hospital, DK-4300 Holbaek, Denmark 3Faculty of Medicine, University of Copenhagen, Copenhagen, Denmark 4Copenhagen Fertility Center, Lygten 2c, Copenhagen DK-2400NV, Denmark

*Correspondence address. Tel: +46-313429222; Fax: +46-31418717; E-mail: [email protected]

Submitted on July 16, 2011; resubmitted on June 12, 2013; accepted on July 15, 2013

abstract: Polycystic ovary syndrome (PCOS) is the most common female endocrine disorder. Ovarian changes in PCOS women are well characterized by ultrasound. However, the ovarian pathophysiology is not fully understood. The aim of this study was to characterize the expression, in both the central ovarian stroma and in granulosa cells (GCs), of a number of genes, including several inflammation-related genes, which have been hypothesized to be involved in the pathophysiology of PCOS. Biopsies of the central ovarian stroma were obtained from PCOS women (Rotterdam criteria) and from normally ovulating women in follicular phase. GCs were retrieved from PCOS-women and non-PCOS women, undergoing in vitro maturation. The expressions of 57 genes were analyzed by quantitative-PCR using a low-density-gene array. The main outcome measures were over-expression or under-expression of the specific genes. The results showed that in the central stroma of PCOS ovaries, five inflammation-related genes (CCL2, IL1R1, IL8, NOS2, TIMP1), the leukocyte marker CD45, the inflammation-related transcription factor RUNX2 and the growth factor AREG were under-expressed. The growth factor DUSP12 and the coagulation factor TFPI2 were overexpressed. In the GC of PCOS, all of the differentially expressed genes were over-expressed; the inflammation-related IL1B, IL8, LIF, NOS2 and PTGS2, the coagulation-related F3 and THBS1, the growth factors BMP6 and DUSP12, the permeability-related AQ3 and the growtharrest-related GADD45A. In conclusion, the results indicate major alterations in the local ovarian immune system of PCOS ovaries. This may have implications for the PCOS-related defects in the inflammation-like ovulatory process and for the susceptibility to acquire the inflammatory state of ovarian hyperstimulation syndrome. Key words: granulosa cells / gene expression / ovary / polycystic ovary syndrome / stroma

Introduction The polycystic ovary syndrome (PCOS) is the most common disorder in women of reproductive age, affecting 7% of the female population. The syndrome is characterized by oligo-/anovulation with polycystic ovaries and increased androgen secretion resulting in acne and hirsutism (The Rotterdam Consensus 2004). Polycystic ovaries have an increased total volume (.10 ml) and number (.12) of antral follicles, with these being ,9 mm (The Rotterdam Consensus 2004). A characteristic feature is also the enlarged central stroma (The Rotterdam Consensus 2004), where a relatively high blood flow has been described (Agrawal et al., 1998). The follicles are embedded in the connective tissue (stroma), which is composed of fibroblasts and smooth muscle cells, and the follicles are lined by theca cells, with proven androgen production. The central stroma (medulla) contains a rich vascular bed, lymphatic vessels and nerves within loose connective tissue (Junqueira and

Carneiro, 2003). The ovarian changes are well characterized by ultrasound and partly by testing the granulosa (Mason et al., 1994) and theca (Nelson et al., 2001) cell function of the typical follicles. A familial clustering of PCOS has been demonstrated (Legro et al., 1998), and initially a single autosomal dominant inheritance was suggested (Govind et al., 1999); however later, a more complex mode of transmission was proposed (Legro and Strauss, 2002; Ewens et al., 2010). Studies comparing gene expression in whole ovaries (Diao et al., 2004; Jansen et al., 2004; Oksjoki et al., 2005), cultured theca cells (Wood et al., 2004), cumulus cells (Kenigsberg et al., 2009) and oocytes (Wood et al., 2007) in PCOS women compared with controls have shown distinctly different expression. Possible ovarian sites of the pathophysiological changes in PCOS are the central stroma and the granulosa cells (GC). In the present study we used a quantitative PCR-based array to evaluate the expression of a number of inflammation-related factors and other genes that have been associated with PCOS and

& The Author 2013. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: [email protected]

50 physiological/pathophysiological ovarian events such as ovarian hyperstimulation syndrome (OHSS).

Materials and Methods Study populations This study was approved by the human ethics committee at the Sahlgrenska Academy, University of Gothenburg, Sweden, and by the ethics committee of the district of Copenhagen, Denmark. Written informed consent was obtained from all patients.

Schmidt et al.

dominant follicle was excised by scissors that cut through the stroma surrounding the follicle. The whole intact follicle was removed inside a laparoscopic sac via a small suprapubic incision and then placed on ice and brought to the laboratory for dissection. After follicle retrieval, sterilization was performed. At the base of the follicle, the stroma layer included was 3– 4 mm and this tissue was separated from the follicle and used in the present study. Based on the normal position of a human pre-ovulatory follicle, with its base well embedded inside the ovary (Hanna et al., 1994; Andersen et al., 1995), this tissue would approximately correspond to the central stroma. The tissue samples were immediately frozen and stored at 2808C until RNA preparation.

Collection of central stroma Infertile (duration ≥1 year) PCOS women (median age 26.0 years) proved to be resistant to clomiphene citrate (CC) and with a body mass index (BMI) ≤ 30 kg/m2 participated in a study comparing ovarian drilling and gonadotrophin stimulation at Sahlgrenska University Hospital, Gothenburg. All the included patients (all Caucasians) met the revised Rotterdam criteria (The Rotterdam Consensus 2004) and other related disorders that mimic PCOS were ruled out. The characteristics of the PCOS women are shown in Table I. All the PCOS women were amenorrhoeic (hyperprolactinemia and hypothyroidism were excluded) and underwent diagnostic laparoscopy, including a test of tubal patency. The patients were randomly allocated to subsequent ovarian drilling or gonadotrophin stimulation. This was done at surgery (if the tubal patency test was normal) by pulling one of several sealed envelopes with a paper stating either ovarian drilling or gonadotrophin stimulation. The sealed envelopes were prepared before the start of the study and 50% of the envelopes contained a paper saying ‘ovarian drilling’ and 50% contained a similar paper stating ‘gonadotrophin stimulation’. Nine patients were allocated to ovarian drilling and as an initial part of the procedure, a small cylindrical (inner diameter 3 mm) ovarian biopsy was taken through the entire ovary using a customized stainless steel cylinder (outer diameter 4 mm, length 40 cm; Meditech, Institute of Neuroscience and Physiology, University of Gothenburg) with sharp edges. After that, each ovary was punctured by unipolar diathermy at 10– 15 sites according to the original report and personal instructions by Gjonnaess (1984). Due to the unpredictably and exceptionally slow inclusion process, the study comparing ovarian drilling and gonadotrophin stimulation was closed after 21 patients had been included. Seven of the nine patients who underwent ovarian drilling resumed ovulation (unpublished data). The ovarian biopsy inside the steel cylinder was extracted by a piston and examined for macroscopic tissue quality. The protocol was to exclude biopsies that were not preserved as a single piece of tissue and to separate and discard the outer 1 cm at each edge, which would contain any follicular tissue and/or surface stroma. The central part, presumably mostly central stroma tissue, was then frozen (10 min after the biopsy was taken) and stored at – 808C until total RNA was extracted. As control tissue, central stroma from seven regularly cycling (median cycle length 29.9 days) patients (median age 36.0) undergoing laparoscopic sterilization (also participating in a study of ovulatory factors produced by GC (Thoroddsen et al., 2011)) was used. The patient material is described in detail elsewhere (Thoroddsen et al., 2011). Samples of the present study included only those at the pre-ovulatory stage (PO) in order to avoid possible effects/changes in mRNA expression mediated by the ovulation. The PO stage was defined as a stage with a size of the dominant follicle of minimum 14 mm and maximum 17.5 mm and before the spontaneous LH surge. The inclusion criteria for the controls were two normal ovaries on ultrasound, no clinical signs of hyperandrogenicity and no chronic systemic disease. The patients had not been on any hormonal contraceptive for at least 3 months prior to surgery. The characteristics of the controls are shown in Table I. During the laparoscopic tubal sterilization procedure, the

Collection of GCs GCs were retrieved from PCOS patients and controls participating in the immature oocyte retrieval programme from patients referred to the Fertility Clinic, Herlevs University Hospital, Copenhagen, Denmark, for IVF/ICSI due to male factor infertility, tubal factor or PCOS. The inclusion criteria for the PCOS women were according to The Rotterdam Consensus (2004). Other disorders mimicking PCOS were excluded. The controls included patients with normal ovaries (normal follicular distribution) who were cycling regularly. Further inclusion criteria for all women were: age ≤40 years, BMI between 18 and 30 kg/m2, infertility ≥1 year and basal FSH , 15 mIU/l. Patients with systemic chronic disease and/or infertility caused by endocrine abnormalities such as hyperprolactinaemia and patients with ovarian endometriosis were excluded. Moreover, patients who had previously failed more than three times to conceive with IVF/ICSI were not included. The cycle was cancelled if an ovarian cyst .10 mm was observed on Day 3 of the cycle. The characteristics of the PCOS women and controls from whom the GC were retrieved are shown in Table I. Women with PCOS (n ¼ 5) were primed with 150 IU recombinant follicle stimulating hormone (FSH) (Gonal-Fw; Serono, Geneva, Switzerland) for 3 days initiated on Days 2 – 3, and oocyte recovery was performed on Days 8 – 9 after deprivation of FSH for 2 – 3 days. Two non-PCOS received the same stimulation and had the same timing of aspiration, while the remaining regularly cycling women (n ¼ 4) with normal ovaries underwent nonstimulated cycles. All the non-PCOS women (two FSH stimulated, four nonstimulated) were included in the analysis to acquire an acceptable sample size. In all cycles, oocyte retrieval was performed on the day after a follicle of 10 mm in diameter could be detected by ultrasonography. An ultrasound examination was performed on Days 2 – 3, Days 6 – 7 and either once a day or at 2-day intervals until the day of aspiration. The follicular diameter was calculated as the mean of the longest follicular axis and the axis perpendicular to it. Oocyte recovery was performed transvaginally with a 17 G single lumen needle (K-OPSC-1225; Cook, Brisbane, Australia) connected to a syringe to induce aspiration vacuum. Follicular fluids were filtered (Falcon 1060, 70 mm mesh size) to remove clotted erythrocytes and cellular debris. After isolation and removal of the oocyte the follicular fluid was centrifuged (3000 g, 5 min). The pelleted GC were immediately frozen in liquid nitrogen and stored at 2808C until transportation on dry ice and analyses in Gothenburg. To assess whether differences in leukocyte contents existed in controls compared with PCOS, we performed QPCR also with CD45, a pan-leukocyte marker, which is expressed on a majority of leukocyte subtypes. CD45 was measured in both the GC and in the stroma tissue and the marker was chosen as Wu et al. (2007) found similar distribution of most leukocytes between PCOS and non-PCOS ovaries, apart from T-lymphocytes and in particular the CD45RO cells, which express a specific isoform of CD45. This finding was made after examining a wide range of markers of leukocyte subtypes in ovaries and several cytokines and chemokine mRNAs in follicular fluid cells, the later also not affected by PCOS status.

51

Ovarian inflammation genes in PCOS

Table I Clinical parameters in women with PCOS and in normo-ovulatory controls of the two groups studied (stroma group and GC group). PCOS

......................................................... n

Median

Min

Max

Controls

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

n

Median

Min

P-level

Max

............................................................................................................................................................................................. Stroma Age at surgery (years)

7

2

26.0

BMI (kg/m )

7

Cycle length (days, ,180)

3

S-FSH (mU/ml)

6

5.2

4.7

6.6

S-LH (mU/ml)

7

7.6

5.8

27.4

S-estradiol (nmol/l)

7

0.32

0.15

24.0 180

23.0 19.0 120

34.0 30.0 180

0.91

S-progesterone (nmol/l) S-testosterone (nmol/l)

7

1.7

0.6

3.4

S-androstenedione (nmol/l)

7

10.8

5.0

13.1

S-SHBG (nmol/l)

7

23.0

12.0

72.0

S-DHEAS (mmol/l)

6

5.6

1.7

11.0

FAI

7

7.4

1.3

17.0

P-glucose (mmol/l)

5

4.6

3.9

5.1

S-insulin (mU/l)

7

14.0

6.6

31.0

7

36.0

31.0

37.0

0.005

7

28.0

15.0

29.5

ns

7

30

28

35

0.014

6

4.7

2.9

5.4

0.003 ns

6

0.39

0.14

1.0

7

0.5

0.3

0.6

GCs Age at surgery (years)

5

28.0

22.0

34.0

6

32.5

29.0

34.0

ns

BMI (kg/m2)

5

23.9

18.7

32.0

6

22.7

21.7

27.1

ns

35

Cycle length (days)

5

41

Size of largest follicle (mm)

5

10.0

6

30

28

40

0.013

11.0

5

10.5

10.0

12.0

Cycle day at aspiration

5

8

ns

7

9

6

9

8

S-FSH (mU/l)

5

3.7

0.9

4.9

6

5.6

4.5

S-LH enhet (mU/l) S-estradiol (nmol/l)

5

7.6

1.0

12.3

6

4.9

3.0

7.7

ns

5

0.14

0.08

0.57

6

0.13

0.06

0.80

ns

S-progesterone (nmol/l)

5

3.0

2

5

6

4.0

2.0

6.0

ns

7.0

180

18 8.3

ns 0.018

Conversion factors are as follows: multiply estradiol by 0.2724 to convert pmol/l to pg/ml, multiply progesterone by 0.3145 for ng/ml, testosterone by 28.84 for ng/dl and DHEAS by 0.3685 for mg/ml and glucose by 18.02 for mg/dl. Divide androstenedione by 0.0349 for ng/dl and SHBG by 34.67 for mg/dl. BMI, body mass index; SHBG, sexual hormone-binding globulin; DHEAS, dehydroepiandrosterone sulphate; FAI, free androgen index (testosterone/SHGB × 100). P-levels .0.05 was considered non-significant (ns). P-levels ,0.05 are shown in bold.

RNA extraction, quality and cDNA synthesis All the experiments were performed under RNase-free conditions. Extraction of total RNA was performed using Trizolw (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. The concentrations of RNA were measured by Nanodrop (Thermo Fisher Scientific, Wilmington, DE, USA). The RNA quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Kista, Sweden). Only samples with RNA integrity number (RIN) .6.0 in the stroma group and .4.0 in the GC group were included. A good correlation between RIN values .5.5 and the outcome of a real-time-PCR experiment has been shown (Schroeder et al., 2006) and RIN values ≥5.5 are considered as high quality. As it was very hard to include patients in the study and to get good tissue samples we decided to keep the sample with an RIN of 4 in the study and if the CT value of that specific sample would differ considerably in the range compared with the other samples it would be excluded later. As this was not the case, this specific sample remained included. Complementary DNA (cDNA) synthesis and quantitative PCR (QPCR) were performed on seven of the nine PCOS women and seven of the seven controls in the central stroma group and all of the patients in the GC group. cDNA was synthesized using 1 mg RNA

and the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA), according to the manufacturer’s instructions.

Taqman Low-Density Array QPCR Quantitative PCR was performed at the Genomics Core Facility at the University of Gothenburg. The Taqman Low-Density Array system (Applied Biosystems) was used. Sixty-four target genes were selected from inventoried Taqman expression assays (Applied Biosystems) and were factory loaded onto each plate. Each gene was run in triplicates and samples of two patients were loaded on each plate. Seven genes: RNA 18S ribosomal 1 (18S), basic transcription factor 3 (BTF3), eukaryotic translation initiation factor 4A2 (EIF4A2), clathrin, heavy chain (CLTC), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), peptidylprolyl isomerase A (cyclophilin A, PPIA) and peptidylprolyl isomerase B (cyclophilin B, PPIB) were included as possible housekeeping genes. The analyses were run on a 7900-HT Fast Real-time PCR System (Applied Biosystems), according to the manufacturer’s instruction. The data were analyzed using the SDS (Sequence Detection Systems) software (Applied Biosystems) by applying the comparative CT method (ddCT) (Livak and Schmittgen, 2001). Baseline and threshold levels for each gene

52 were set automatically using the Auto CT algorithm. Insufficient or absent amplifications were omitted from subsequent analyses. The Normfinder algorithm, version 0.953 (http://www.mdl.dk/publicationsnormfinder.htm), was used to select the housekeeping genes; the best stability value was the combination of 18S and GAPDH for the stroma samples and PPIA and PPIB for the GC samples. For all genes of interest, mRNA content was calculated for each sample relative to the respective combination of housekeeping genes. This was done using the equation 2 – ddCT (Livak and Schmittgen, 2001), where dCT is the difference between the target gene and the (respective combination of) housekeeping and the ddCT is the difference between the dCT for each sample and the control group (i.e. non-PCOS patients). Results for each gene of interest were described as the median fold change from the non-PCOS group. Describing the relative fold change, mean of the non-PCOS group was equivalent to 1. We considered genes to be undetectable if all three replicates for one patient were undetectable (undetermined, i.e. CT . 40). When more than 45% of the replicates in either the PCOS women or the controls were undetermined and/or if fewer than four samples remained determined in each group and the CT was .36 (classified by our long in-house experience as very low expression and of uncertain accuracy), no comparison between groups was made (stated as x in Table II).

Schmidt et al.

Finland) with a CV of 7% at 40 and 5% at 60 nmol/l. A hexokinase-based photometric method (Roche, Diagnostics GmbH, Mannheim, Germany) was used to analyse plasma glucose. Serum estradiol, LH, FSH, prolactin, progesterone and TSH as well as serum estradiol, FSH, LH and progesterone analyzed in Denmark, were measured by chemiluminescent microparticle immunoassay (Architect Reagent Kits; Abbott Laboratories, Diagnostic Division, Abbott Park, IL, USA, samples analyzed in Denmark; ADVIA Centaur kits, Siemens, Germany). The estradiol assay had a CV ≤ 7%. The progesterone assay had a CV of 19, 11 and 13% at 2.7, 27 and 70 nmol/l. Samples analyzed in Denmark: the estradiol assay with CVs of 10, 5.3 and 6.1% at 206.6, 392.3 and 2076.1 pmol/l, respectively. The progesterone assay had a CV of 12.7, 5.4 and 3.7% at 3.8, 22.9 and 66.5 nmol/l, respectively.

Statistics The results are presented as medians and ranges (minimum and maximum). Patients and controls were compared using the Mann– Whitney U-test. Mann– Whitney U-tests were also used to examine statistical differences in mRNA expression in PCOS versus non-PCOS, (results presented in Table II). A P value , 0.05 was considered significant.

Histology For light microscopy analysis of stroma samples, biopsies were fixed in 4% phosphate-buffered formaldehyde, embedded in paraffin, sectioned and stained with hematoxylin and eosin. The slides were evaluated by one experienced pathologist (gyne-pathology profile).

Physical examinations and definitions Hirsutism was defined as the patient’s and doctor’s subjective estimation of the presence of excessive coarse hair on the face, trunk or thighs and the need to remove hair at least twice a week. Good concordance has been shown between the subjective and objective grading of hirsutism (Franks, 1989). Amenorrhoea was defined as no menstrual bleeding in the past 90 days and oligomenorrhoea as an intermenstrual interval of .35 days and fewer than 8 menstrual bleedings in the past year. Height and body weight were measured according to standard procedures described previously (Schmidt et al., 2011). Body mass index (BMI) was calculated as body weight divided by height squared (kg/m2). CC resistance was defined as either the absence of ovulation on a dose up to and including 150 mg CC, or in the case of detected ovulation, when no pregnancy was achieved after 3 cycles.

Biochemical assays In the PCOS patients and controls in the central stroma group blood samples were taken immediately before surgery to acquire baseline data for the PCOS drilling group and for verifying the pre-ovulatory phase in controls. The blood samples of the stroma groups were analyzed in the laboratory at the Department of Obstetrics and Gynaecology at the Sahlgrenska University Hospital. The blood samples of the GC groups were analyzed at the laboratory of Clinical Biochemistry at Herlev Hospital. All coefficients of variances (CV) were ,5%, including both intra-assay and inter-assay variations, unless otherwise specified. Serum testosterone, androstenedione, DHEAS and serum insulin TM were measured by radioimmunoassay assay kits (muChem Double Anti125 body Testosterone I RIA Kit (ICN Biomedicals, Inc, Costa Mesa, CA, USA); Coat-A-Countw Direct Androstendione and DHEAS (Diagnostic Products Corporation, Los Angeles, CA, USA); Pharmacia Insulin kit (Pharmacia Upjohn Diagnostics AB, Uppsala, Sweden). The testosterone assay had a CV of 10%, and the androstenedione assay with CVs of 15% at 3.2 and 10% at 11.2 nmol/l. The DHEAS assay had a CV of 12% at 3 and 10 mmol/l. The insulin assay had a CV of 10%. Serum SHBG and IGF-1 were measured by immunoradiometric analysis. The SHBG assay (Orion Diagnostica, Espoo,

Results Characteristics of the study populations The phenotypes of the PCOS patients in the stroma group were that all seven patients had polycystic ovaries on transvaginal ultrasound and oligo/amenorrhoea. Four of these patients had clinical and/or biochemical hyperandrogenism. In the GC group, all five PCOS women had PCOS morphology and oligo/amenorrhoea. Data regarding the presence of hyperandrogenism in these five patients are lacking. The main clinical parameters in the PCOS women and the normo-ovulatory controls in the stroma and GC groups are summarized in Table I. In the stroma group, age and the levels of LH and cycle length differed between PCOS patients and controls. In the GC group, cycle length and FSH differed between PCOS women and controls. BMI was similar in PCOS women and controls in both groups. Light microscopy was only performed on central stroma tissue, thus no histological verification of the PCOS diagnosis could be performed. Central stroma from both groups had a similar morphological appearance, with elongated cells placed in bundles of different orientations and with blood vessels present (Fig. 1).

Expression results All analyzed genes and their expression levels (medians) are shown in Table II. The over- and/or under-expressed genes of the stroma and the GC are shown in Fig. 2. In the stroma, five inflammation-related mediators (chemokine ligand 2 (CCL2), interleukin-1 receptor type 1 (IL1R1), interleukin-8 (IL8), nitric oxide synthase 2 (NOS2) and tissue inhibitor of metalloproteinase 1 (TIMP1) and the associated runt-related transcription factor 2 (RUNX2) were under-expressed in PCOS. Under-expression was also seen for the leukocyte marker CD45 and for amphiregulin (AREG), while over-expression was seen for dual specificity phosphatase 12 (DUSP12) and tissue factor pathway inhibitor 2 (TFPI2). A different expression pattern was seen in GC in PCOS, with genes only being over-expressed. Five inflammation-related mediators [interleukin-1 beta (IL1B), IL8, leukemia inhibitory factor, NOS2 and

53

Ovarian inflammation genes in PCOS

Table II Results of QPCR. Gene title, function

Gene symbol

Stroma

GCs

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

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

FC: PCOS

FC: PCOS

FC: N

P-level

FC: N

P-level

............................................................................................................................................................................................. Coagulation and vascularity Angiotensin I converting enzyme (peptidyl-dipeptidase A) 2

ACE2

0.425

0.963

0.053

1.802

0.628

0.421

Coagulation factor III (thromboplastin)

F3

0.598

0.929

0.097

2.573

1.332

0.009

Phospholipase A2, group IIA

PLA2G2A

1.546

1.568

0.902

1.070

0.782

0.662

Tissue factor pathway inhibitor 2

TFPI2

1.943

1.014

0.007

0.759

0.749

1.000

Thrombospondin1

THBS1

0.604

1.086

0.097

6.885

1.474

0.009

Angiopoetin 1

ANGPT1

1.157

1.045

0.383

0.710

1.103

0.662

Angiopoetin 2

ANGPT2

0.732

1.168

0.318

1.879

1.268

0.931

Angiopoietin-like 1

ANGPTL1

0.998

0.891

0.620

1.726

0.880

0.421

Amphiregulin

AREG

0.285

1.324

0.018

0.905

0.866

0.931

Bone morphogenetic protein 6

BMP6

0.957

1.096

0.456

3.975

1.258

0.017

Dual specificity phosphatase 12

DUSP12

1.273

1.013

0.011

1.446

1.118

0.030

Platelet-derived growth factor alpha polypeptide

PDGFA

0.874

1.077

0.902

3.561

0.988

0.052

Insulin-like growth factor binding protein 1

IGFBP1

0.719

0.815

0.535

8.337

1.648

0.222

Insulin-like growth factor binding protein 3

IGFBP3

1.173

1.021

0.165

3.547

1.377

0.178

Insulin-like growth factor binding protein 5

IGFBP5

1.447

0.970

0.053

0.890

0.709

0.931

Anti-Mu¨llerian hormone

AMH

0.976

1.061

0.535

1.616

0.448

0.662

Cytochrome P450, family 17, subfamily A, polypeptide 1

CYP17A1

1.006

1.331

0.805

0.286

0.448

0.429

Growth factors

Hormones/hormone mediators

Inhibin beta B

INHBB

0.411

1.417

0.097

0.681

1.153

0.429

Steroid 5-alpha-reductase 1

SRD5A1

0.948

0.988

0.805

1.497

1.017

0.082

ADAMTS1

0.886

1.025

0.805

0.920

1.012

0.931

Inflammation A disintegrin and metalloproteinase with thrombosp. T1 A disintegrin and metalloproteinase with thrombosp. T2

ADAMTS2

0.871

1.038

0.383

1.494

1.218

0.479

Chemokine (C-C motif) ligand 2

CCL2

0.681

0.919

0.026

2.385

1.147

0.537

Chemokine (C-C motif) ligand 3

CCL3

0.159

1.019

Chemokine (C-C motif) ligand 20

CCL20

Leukocyte marker

CD45

0.305

Colony stimulating factor 1

CSF1

1.060

Colony stimulating factor 2

CSF2

Heat chock (70 kDa) protein 6 (HSP70B)

HSPA6/HSPA7

0.097

20.823

1.054

0.052

x

11.631

0.496

0.095

0.853

0.007

0.678

2.269

0.318

1.087

0.902

2.321

1.026

x 0.499

0.985

0.073

0.126 x

1.185

1.089

0.818

14.091

1.217

0.017

Kallikrein-related peptidase 10

KLK10

1.143

0.797

0.165

Interleukin 1, beta

IL1B

0.285

0.748

0.073

x

Interleukin 6 (interferon, beta 2)

IL6

0.595

0.697

0.209

14.981

0.536

0.056

Interleukin 8

IL8

0.272

0.998

0.007

18.277

1.063

0.009

Interleukin 1 receptor, type I

IL1R1

0.444

0.947

0.026

1.137

1.092

0.818

Leukemia inhibitory factor (cholinergic diff. factor) ¤

LIF

0.569

0.761

0.073

3.646

1.032

0.030

Matrix metallopeptidase 1

MMP1

Matrix metallopeptidase 2

MMP2

0.956

0.992

0.383

1.560

0.669

Matrix metallopeptidase 9

MMP9

0.679

0.765

0.383

5.934

0.711

Matrix metallopeptidase 10

MMP10

x

x

x

0.352 0.095 x

Matrix metallopeptidase 11 (stromelysin 3)

MMP11

0.618

0.976

0.053

1.723

1.083

0.052

Nitric oxide synthase 2, inducible

NOS2

0.460

0.936

0.038

2.608

1.497

0.030

Prostaglandin-endoperoxide synthase 2

PTGS2

1.119

1.008

0.902

2.721

1.354

0.017

Selenoprotein S

SELS

0.967

1.077

0.710

1.350

0.948

0.429

Continued

54

Schmidt et al.

Table II Continued Gene title, function

Gene symbol

Stroma

GCs

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

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

FC: PCOS

FC: PCOS

FC: N

P-level

FC: N

P-level

............................................................................................................................................................................................. TIMP metallopeptidase inhibitor 1

TIMP1

0.773

0.977

0.007

0.953

0.867

1.000

TIMP metallopeptidase inhibitor 2

TIMP2

1.219

0.993

0.097

2.116

1.100

0.126

TIMP metallopeptidase inhibitor 3

TIMP3

1.129

1.201

0.620

1.626

1.020

0.178

Tumor necrosis factor

TNF

0.477

0.890

0.073

21.729

1.148

0.056

Tumor necrosis factor, alpha-induced protein 6

TNFAIP6

0.797

0.740

0.620

1.303

1.119

0.429

Tumor necrosis factor receptor superfamily, member 11b

TNFRSF11B

0.892

1.020

Tumor necrosis factor (ligand) superfamily, member 11

TNFSF11

0.592

0.757

0.841

0.535 x

x

Permeability Aquaporin 2

AQP2

0.063

0.483

Aquaporin 3

AQP3

0.823

0.963

ATPase, Na+/K+ transporting, beta 1 polypeptide

ATP1B1

0.383 0.259

x 3.314

1.342

x

0.017 x

Transcription factors/DNA-binding proteins CCAAT/enhancer binding protein (C/EBP) beta

CEBPB

1.110

0.936

0.383

1.044

0.823

0.537

cAMP responsive element modulator

CREM

1.208

1.010

0.165

1.320

0.857

0.082

Proliferating cell nuclear antigen

PCNA

1.287

0.965

0.073

0.956

1.121

0.537

Runt-related transcription factor 1

RUNX1

0.698

1.016

0.128

4.179

0.964

0.052

Runt-related transcription factor 2

RUNX2

0.604

0.963

0.018

1.906

1.546

0.841

GADD45A

1.440

0.907

0.165

8.333

1.542

0.009

Others Growth arrest and DNA-damage-inducible, alpha

Medians of 22DDCT are presented as well as P-levels comparing the stroma of PCOS women (P) with the stroma of normo-ovulatory controls (N) and P levels comparing GCs of PCOS women (P) with GCs of normo-ovulatory controls (N). Genes analyzed in the QPCR of the stroma (left) and the GCs (right). X ¼ no comparisons between groups made due to undetectable genes. We considered genes to be undetectable if all three replicates for one patient were undetectable (undetermined) and more than 45% of the replicates in either the PCOS women or the controls were undetermined and/or if fewer than four samples remained determined in each group and the CT was .36. TIMP, tissue inhibitor of metalloproteinases; Thrombosp., thrombospondin; T1, type 1; T2, type 2; CD, cluster of differentiation; ¤diff., differentiation. A P-level ,0.05 was considered significant. Significant P-levels are shown in bold.

Figure 1 Light microscopy of the central stroma from a PCOS ovary (A) and normal ovary (B). Central stroma from both groups had a similar morphological appearance with elongated cells (fibroblasts) in bundles of different orientations. Vessels (V) are indicated by arrows. Scale bars represent 100 mm. The central stroma of the PCOS ovary was taken through the entire ovary as a cylindrical (inner diameter 3 mm) biopsy, thereafter discarding the outer 1 cm at each edge. The stroma of the normal ovary was taken by excising a pre-ovulatory follicle and its well-embedded base of stroma (3 –4 mm) by scissors, afterwards the stroma was separated from the follicle and used for this section.

prostaglandin-endoperoxide synthase 2 (PTGS2)] were among the overexpressed genes. Other over-expressed genes were coagulation related [coagulation factor 3, also named thromboplastin (F3) and thrombospondin 1 (THBS1)), growth factors (bone morphogenetic protein 6 (BMP6) and DUSP12), permeability-related (aquaporin 3 (AQP3)] and

growth arrest related [growth arrest and DNA damage-inducible, alpha (GADD45A)]. Analysis of CD45mRNA showed no difference between the PCOS and controls in the GC group, but a higher expression (P , 0.007) of CD45 was seen in control stroma (median fold change (FC) 0.85)

Ovarian inflammation genes in PCOS

55

Figure 2 QPCR data of the over- and/or under-expressed genes of the stroma (A) and the GCs (B). A P value of ,0.05 was considered significant. The x-axis (and transparent bars) represents the relative fold change. The black bars represent the 25th and 75th percentiles. For graphical illustration the underexpressed genes were calculated as mean non-PCOS group/median PCOS group and the over-expressed genes as median PCOS group/mean non-PCOS group, where mean non-PCOS group was equivalent to 1. *P , 0.05, **P , 0.01.

when compared with PCOS stroma (median FC 0.31). Since tissue bound leukocytes are a natural component of stroma no adjustments for this difference was done in the analysis of expression levels.

Discussion The presence of polycystic ovaries is a central feature in PCOS. The classical characteristics of a polycystic ovary includes a total increase in size both due to the enlarged central stroma (medulla) and the abundance of mid-sized antral follicles in the ovarian periphery. The structural alterations of the polycystic ovary are mirrored by the functional features of increased androgen production, oligo-anovulation and the greater risk of developing OHSS at gonadotrophin stimulation for ovulation induction and/or IVF. Consequently, several studies focused on elucidating possible intra-ovarian mediators/pathways that may be up-regulated or down-regulated in PCOS ovaries, compared with controls. Thus, PCOS ovarian array studies have been performed on whole ovaries (Diao et al., 2004; Jansen et al., 2004; Oksjoki et al., 2005), cultured theca cells (Wood et al., 2003), oocytes (Wood et al., 2007) and cumulus cells from hyperstimulated and luteinized follicles (Kenigsberg et al., 2009). The present study represents the first expression study of ovarian tissue in PCOS patients, where GC and the ovarian central stroma have been examined as separate entities. The present study was designed to examine specifically a number of genes that were hypothesized to be expressed differently in PCOS ovaries. The major findings were that a number of inflammation-related genes were down-regulated in the stroma of PCOS ovaries and that most other inflammation-related genes were up-regulated in the GC. Importantly, both IL8 and NOS2 were, in comparison with controls, downregulated in the PCOS stroma, but up-regulated in PCOS GC. In the

GC compartment, also genes related to coagulation, growth and permeability were up-regulated in PCOS ovaries. Several of the differentially expressed genes have not been detected as being differentially expressed in other PCOS ovary array studies, while certain others confirmed the results of previous studies (Jansen et al., 2004; Kenigsberg et al., 2009). As the ovary is composed of multiple compartments with different cell types, and in view of the fact that the proportion of these compartments/ cells differ significantly between PCOS ovaries and normo-ovulatory ovaries, the present study was designed to examine as pure ovarian compartments as possible from patients who were well-characterized regarding PCOS and normal ovulation. However, the GC samples were found to contain a too small volume for separating leukocytes from GC before the QPCR and the stroma compartment could not be separated from leukocytes. Therefore, we analyzed CD45 as a marker of leukocytes in both the GC and the stroma groups. No difference of CD45 mRNA levels was found in PCOS compared with controls in the GC group, thus the difference in gene expression in the GC group should be considered as valid. However, in the stroma a higher expression of CD45 mRNA was found in the control stroma. This is in line with findings by Wu et al. (2007). However, since tissue bound leukocytes represents a part of the normal cell population in stroma we consider the results concerning difference in mRNA expression between PCOS and controls in stroma as valid, but the lower leukocyte abundance in PCOS might contribute to the observed differences. One characteristic feature of the largest antral follicles of the PCOS ovary is that they are arrested well before their development into a pre-ovulatory follicle, and will then not undergo final maturation and ovulation. In the GC of PCOS ovaries, four genes that have been strongly associated with ovulation were found to be up-regulated. The genes IL1B, IL8, NOS2 and PTGS2 are classical ovulatory genes that are

56 expressed at low levels during follicular development and induced by the pre-ovulatory gonadotrophin surge. The ovarian functions of the classical cytokines IL1B (Bra¨nnstro¨m, 2004) and IL8 (Runesson et al., 2000) seem to be recruitment and activation of leukocytes to the follicle at ovulation (Wang et al., 1995). A similar function to attract monocytes and basophils has been shown for CCL2 in non-ovarian tissues (GeneCards, 2011; Gilbert et al., 2011) and both of these cell types are of importance for ovulation (Bra¨nnstro¨m and Runesson, 2000). CCL2 was downregulated in the stroma in PCOS. In these smaller size PCOS follicles, the increased expression of these cytokines may result in premature invasion of leukocytes, which may disrupt the normal final maturation of the follicle. OHSS is especially prevalent among PCOS (Delvigne and Rozenberg, 2002) and both IL1 and IL8 have been implied in the pathophysiology of OHSS (Abramov et al., 1996). It may well be that the up-regulation of these genes already in mid-sized antral follicles, as in PCOS, is one mechanism behind the increased risk of OHSS in PCOS patients undergoing ovulation induction. Interestingly, IL8 expression was lower in the stroma of PCOS patients than in controls and the receptor of IL1 was also under-expressed. It has been shown that IL1 inhibits gonadotrophin-stimulated androstenedione production in theca-interstitial cells (Hurwitz et al., 1991) and low expression of the receptor in the PCOS central stroma may thus be one mechanism behind the increased ovarian androgen secretion in PCOS ovaries. The coagulation-associated genes F3, THBS1 and TFPI2 were up-regulated in PCOS ovaries. It has been demonstrated previously that F3, a cell surface glycoprotein, is expressed in GC of the mouse and its secretion is stimulated by LH (Hsueh, 1999). This factor enables cells to initiate the blood coagulation cascades and it serves as the highaffinity receptor for coagulation factor VII (Hsueh, 1999). The extracellular matrix glycoprotein THBS1 is regulated by FSH and expressed in human GC (Hsueh, 1999) with the effect to inhibit follicular angiogenesis and to induce apoptosis of GC in rats (Garside et al., 2010). Thus, the overexpression of this gene, which is verifying the results of studies on whole ovaries (Jansen et al., 2004), may lead to altered follicular survival. To our knowledge, no study on the role of TFPI2 in ovarian tissue exist, hence, the function of the over-expression of TFPI2, in the central stroma is yet to be elucidated. The BMP6 growth factor has previously been localized to the GC of healthy tertiary human follicles and an in vitro study showed that BMP6-induced gene expression of the FSH receptor, inhibin/activin beta subunits and AMH in human GC (Shi et al., 2009). The increased expression of BMP6 in GC in the present study, in line with the results of studies on whole ovaries (Jansen et al., 2004), may also explain the arrested follicle development in PCOS women. Furthermore, studies in the bovine indicate that BMP signaling pathways in follicles contribute to the negative regulation of androgen production in theca cells (Glister et al., 2005). Thus, the overproduction of androgens in PCOS ovaries could be caused by a defective auto-regulatory pathway involving thecal BMP signaling. Amphiregulin (AREG) encodes for a protein, which is a member of the epidermal growth factor family and has a critical function in the periovulatory events in humans. Further, AREG accumulation has been found to be a marker of gonadotrophin stimulation (Zamah et al., 2010). We cannot explain the findings of under-expressed AREG in the stroma. In our study the growth factor DUSP12 was over-expressed in PCOS in both the GC and the stroma. As DUSP12 negatively regulates

Schmidt et al.

the mitogen-activated protein super-family, which is associated with cell proliferation and differentiation (GeneCards, 2011), this may be a reason for the increased stroma/ovarian volume of PCOS ovaries. Moreover, DUSP12 may be the link between the strong association between PCOS and diabetes type II (T2DM) (Legro et al., 1999), as DUSP12 has been suggested to be associated with T2DM (Das et al., 2006). The aquaporins (AQP) are membrane channel proteins that facilitate rapid water movement over cell membranes, and differential expression of some AQPs were recently described in human granulosa and theca cells during human ovulation (Thoroddsen et al., 2011). In that study, AQP3 was expressed at very low levels before gonadotrophin stimulation and then induced to 40-fold in GC and 15-fold in theca cells at the early ovulatory stage. The up-regulation of AQP3 in PCOS follicles in the present study should lead to markedly increased permeability into/out of the follicle, which may also be important for the susceptibility of the PCOS ovary to develop OHSS. The transcription factor RUNX2 was found to be down-regulated in the stroma of PCOS ovaries. This transcription factor was recently demonstrated to be up-regulated by LH both in rat and human GC, with effects on the expression of several genes involved in ovulation and luteal function (Park et al., 2010), it may well be that an altered expression of this central transcription factor may lead to ovarian dysfunction. The transcript levels of GADD45A, and the members of the same group of genes, increase as a result of stressful growth arrest conditions (GeneCards, 2011). Given that the folliculogenesis is arrested in PCOS, which may increase the transcript levels of GADD45A, this could explain our findings with over-expression of GADD45A in GC of PCOS women. The strengths of this study were the well-diagnosed PCOS women (according to the Rotterdam criteria) and controls, the examination of as separated cell compartments as possible, for gene-expression studies fairly good sample sizes and similar BMIs of the compared groups, which is important since BMI is known to be associated with hormonal levels and inflammation. A limitation of the study was that separation of leukocytes from the tissues was not done before gene expression analysis was performed, which might affect the gene expression in the stroma, where the leukocyte marker was found to be differently expressed. However, since tissue bound leukocytes represent part of the normal cell population in stroma we consider the results concerning difference in mRNA expression between PCOS and controls in stroma as valid, but the less leukocyte abundance in PCOS might contribute to the observed differences regarding gene expression. Other limitations of this study were that the stroma tissue from the PCOS women and the controls may not exactly correspond to each other with regard to position within the ovary and that the study protocol was not strictly followed when including the patients in the GC group, with the main problem of FSH-stimulation in two of the six controls, which also may have influenced the results. In addition, we acknowledge that there was a difference in age between the PCOS women and the controls in the stroma group. This was, however, unavoidable, as the stroma biopsies of the PCOS women were obtained as part of their infertility treatment and the controls were undergoing sterilization, well after births. In conclusion, the present study has identified a number of PCOS-associated genes that may contribute to the functional and structural aberrations of the PCOS ovary. Further studies are needed to elucidate their role in the complex ovarian pathophysiology of PCOS.

Ovarian inflammation genes in PCOS

Acknowledgements We are very grateful to Pernilla Dahm-Ka¨hler for taking part in the inclusion of PCOS patients (stroma) and controls (stroma). The technical and statistical assistance of Ann Wallin and the statistical guidance from Professor Anders Ode´n are also gratefully acknowledged.

Authors’ roles J.S. planning of study, acquisition of part of data, analysis of data, interpretation of data, drafting paper, finalizing paper and final approval of paper. B.W. planning of study, acquistion of expression data, analysis of data, revising paper, final approval of paper. A.L.M., S.L., L.N. planning of study, acquistion of part of data, interpretion of data, revising paper, final approval of paper. M.B. idea and planning of study, interpretation of data, drafting paper, final approval of paper.

Funding This work was supported by grants from the Hjalmar Svensson Research Foundation, the Go¨teborg Medical Society, the Sahlgrenska University Hospital ALF agreement and the Swedish Research Council.

Conflict of interest The authors have nothing to disclose.

References Abramov Y, Schenker JG, Lewin A, Friedler S, Nisman B, Barak V. Plasma inflammatory cytokines correlate to the ovarian hyperstimulation syndrome. Hum Reprod 1996;11:1381 – 1386. Agrawal R, Sladkevicius P, Engmann L, Conway GS, Payne NN, Bekis J, Tan SL, Campbell S, Jacobs HS. Serum vascular endothelial growth factor concentrations and ovarian stromal blood flow are increased in women with polycystic ovaries. Hum Reprod 1998;13:651– 655. Andersen AG, Als-Nielsen B, Hornnes PJ, Franch Andersen L. Time interval from human chorionic gonadotrophin (HCG) injection to follicular rupture. Hum Reprod 1995;10:3202 – 3205. Bra¨nnstro¨m M. The potential role of cytokines in ovarian physiology: the case for interleukin-1. In: Leung PCK, Adashi EY (eds). The Ovary, Vol. 1. San Diego: Elsevier Academic Press, 2004, 261 – 268. Bra¨nnstro¨m M, Runesson E. White blood cells: active participants in the ovulation cascade. In: Adashi EY (ed.). Ovulation. New York, USA: Springer-Varlag, 2000, 221 – 242. Das SK, Chu WS, Hale TC, Wang X, Craig RL, Wang H, Shuldiner AR, Froguel P, Deloukas P, McCarthy MI et al. Polymorphisms in the glucokinase-associated, dual-specificity phosphatase 12 (DUSP12) gene under chromosome 1q21 linkage peak are associated with type 2 diabetes. Diabetes 2006;55:2631 – 2639. Delvigne A, Rozenberg S. Epidemiology and prevention of ovarian hyperstimulation syndrome (OHSS): a review. Hum Reprod Update 2002;8:559 – 577. Diao FY, Xu M, Hu Y, Li J, Xu Z, Lin M, Wang L, Zhou Y, Zhou Z, Liu J et al. The molecular characteristics of polycystic ovary syndrome (PCOS) ovary defined by human ovary cDNA microarray. J Mol Endocrinol 2004; 33:59– 72. Ewens KG, Stewart DR, Ankener W, Urbanek M, McAllister JM, Chen C, Baig KM, Parker SC, Margulies EH, Legro RS et al. Family-based analysis

57 of candidate genes for polycystic ovary syndrome. J Clin Endocrinol Metab 2010;95:2306 – 2315. Garside SA, Harlow CR, Hillier SG, Fraser HM, Thomas FH. Thrombospondin-1 inhibits angiogenesis and promotes follicular atresia in a novel in vitro angiogenesis assay. Endocrinology 2010;151:1280– 1289. GeneCards. GeneCards version 3. Retrieved 110620, 2011. www.genecards. org. Gilbert J, Lekstrom-Himes J, Donaldson D, Lee Y, Hu M, Xu J, Wyant T, Davidson M. Effect of CC chemokine receptor 2 CCR2 blockade on serum C-reactive protein in individuals at atherosclerotic risk and with a single nucleotide polymorphism of the monocyte chemoattractant protein-1 promoter region. Am J Cardiol 2011;107:906– 911. Gjonnaess H. Polycystic ovarian syndrome treated by ovarian electrocautery through the laparoscope. Fertil Steril 1984;41:20 – 25. Glister C, Richards SL, Knight PG. Bone morphogenetic proteins (BMP) -4, -6, and -7 potently suppress basal and luteinizing hormone-induced androgen production by bovine theca interna cells in primary culture: could ovarian hyperandrogenic dysfunction be caused by a defect in thecal BMP signaling? Endocrinology 2005;146:1883 – 1892. Govind A, Obhrai MS, Clayton RN. Polycystic ovaries are inherited as an autosomal dominant trait: analysis of 29 polycystic ovary syndrome and 10 control families. J Clin Endocrinol Metab 1999;84:38 – 43. Hanna MD, Chizen DR, Pierson RA. Characteristics of follicular evacuation during human ovulation. Ultrasound Obstet Gynecol 1994;4:488 – 493. Hsueh AJ. Ovarian kaleidoscope database, 1999. http://ovary.stanford. edu/. Hurwitz A, Payne DW, Packman JN, Andreani CL, Resnick CE, Hernandez ER, Adashi EY. Cytokine-mediated regulation of ovarian function: interleukin-1 inhibits gonadotropin-induced androgen biosynthesis. Endocrinology 1991; 129:1250–1256. Jansen E, Laven JS, Dommerholt HB, Polman J, van Rijt C, van den Hurk C, Westland J, Mosselman S, Fauser BC. Abnormal gene expression profiles in human ovaries from polycystic ovary syndrome patients. Mol Endocrinol 2004;18:3050– 3063. Junqueira LC, Carneiro J. The female reproductive system. In: Junqueira LC, Carneiro J (eds). Basic histology, 10th edn. New York, USA: Lange, 2003, 449. Kenigsberg S, Bentov Y, Chalifa-Caspi V, Potashnik G, Ofir RBirk OS. Gene expression microarray profiles of cumulus cells in lean and overweightobese polycystic ovary syndrome patients. Mol Hum Reprod 2009; 15:89– 103. Legro RS, Strauss JF. Molecular progress in infertility: polycystic ovary syndrome. Fertil Steril 2002;78:569– 576. Legro RS, Spielman R, Urbanek M, Driscoll D, Strauss JF III, Dunaif A. Phenotype and genotype in polycystic ovary syndrome. Recent Prog Horm Res 1998;53:217 – 256. Legro RS, Kunselman AR, Dodson WC, Dunaif A. Prevalence and predictors of risk for type 2 diabetes mellitus and impaired glucose tolerance in polycystic ovary syndrome: a prospective, controlled study in 254 affected women. J Clin Endocrinol Metab 1999;84:165 – 169. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 2001;25:402– 408. Mason HD, Willis DS, Beard RW, Winston RM, Margara R, Franks S. Estradiol production by granulosa cells of normal and polycystic ovaries: relationship to menstrual cycle history and concentrations of gonadotropins and sex steroids in follicular fluid. J Clin Endocrinol Metab 1994;79:1355 – 1360. Nelson VL, Qin KN, Rosenfield RL, Wood JR, Penning TM, Legro RS, Strauss JF III, McAllister JM. The biochemical basis for increased testosterone production in theca cells propagated from patients with polycystic ovary syndrome. J Clin Endocrinol Metab 2001;86:5925– 5933.

58 Oksjoki S, Soderstrom M, Inki P, Vuorio E, Anttila L. Molecular profiling of polycystic ovaries for markers of cell invasion and matrix turnover. Fertil Steril 2005;83:937 – 944. Park ES, Lind AK, Dahm-Kahler P, Brannstrom M, Carletti MZ, Christenson LK, Curry TE Jr, Jo M. RUNX2 transcription factor regulates gene expression in luteinizing granulosa cells of rat ovaries. Mol Endocrinol 2010;24:846– 858. Runesson E, Ivarsson K, Janson PO, Brannstrom M. Gonadotropin- and cytokine-regulated expression of the chemokine interleukin 8 in the human preovulatory follicle of the menstrual cycle. J Clin Endocrinol Metab 2000;85:4387– 4395. Schmidt J, Landin-Wilhelmsen K, Brannstrom M, Dahlgren E. Cardiovascular disease and risk factors in PCOS women of postmenopausal age: a 21-year controlled follow-up study. J Clin Endocrinol Metab 2011;96:3794 – 3803. Schroeder A, Mueller O, Stocker S, Salowsky R, Leiber M, Gassmann M, Lightfoot S, Menzel W, Granzow M, Ragg T. The RIN: an RNA integrity number for assigning integrity values to RNA measurements. BMC Mol Biol 2006;7:3. Shi J, Yoshino O, Osuga Y, Koga K, Hirota Y, Hirata T, Yano T, Nishii O, Taketani Y. Bone morphogenetic protein-6 stimulates gene expression of follicle-stimulating hormone receptor, inhibin/activin beta subunits, and anti-Mullerian hormone in human granulosa cells. Fertil Steril 2009; 92:1794 – 1798. The Rotterdam Consensus, ESHRE/ASRM-Sponsored, PCOS, Consensus, WorkshopGroup. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 2004;81:19 – 25.

Schmidt et al.

Thoroddsen A, Dahm-Kahler P, Lind AK, Weijdegard B, Lindenthal B, Muller J, Brannstrom M. The water permeability channels aquaporins 1 – 4 are differentially expressed in granulosa and theca cells of the preovulatory follicle during precise stages of human ovulation. J Clin Endocrinol Metab 2011;96:1021 – 1028. Wang LJ, Brannstrom M, Pascoe V, Norman RJ. Cellular composition of primary cultures of human granulosa-lutein cells and the effect of cytokines on cell proliferation. Reprod Fertil Dev 1995;7:21– 26. Wood JR, Nelson VL, Ho C, Jansen E, Wang CY, Urbanek M, McAllister JM, Mosselman SStrauss JF III. The molecular phenotype of polycystic ovary syndrome (PCOS) theca cells and new candidate PCOS genes defined by microarray analysis. J Biol Chem 2003;278:26380– 26390. Wood JR, Ho CK, Nelson-Degrave VL, McAllister JM, Strauss JF III. The molecular signature of polycystic ovary syndrome (PCOS) theca cells defined by gene expression profiling. J Reprod Immunol 2004; 63:51 – 60. Wood JR, Dumesic DA, Abbott DH, Strauss JF III. Molecular abnormalities in oocytes from women with polycystic ovary syndrome revealed by microarray analysis. J Clin Endocrinol Metab 2007;92:705 – 713. Wu R, Fujii S, Ryan NK, Van der Hoek KH, Jasper MJ, Sini I, Robertson SA, Robker RL, Norman RJ. Ovarian leukocyte distribution and cytokine/ chemokine mRNA expression in follicular fluid cells in women with polycystic ovary syndrome. Hum Reprod 2007;22:527– 535. Zamah AM, Hsieh M, Chen J, Vigne JL, Rosen MP, Cedars MI, Conti M. Human oocyte maturation is dependent on LH-stimulated accumulation of the epidermal growth factor-like growth factor, amphiregulin. Hum Reprod 2010;25:2569 – 2578.