Association Between Single-Nucleotide Polymorphisms

2557

Association Between Single-Nucleotide Polymorphisms in Hormone Metabolism and DNA Repair Genes and Epithelial Ovarian Cancer: Results from Two Australian Studies and an Additional Validation Set

Jonathan Beesley,1 Susan J. Jordan,1,2 Amanda B. Spurdle,1 Honglin Song,5 Susan J. Ramus,6 Suzanne Kruger Kjaer,7,8 Estrid Hogdall,7 Richard A. DiCioccio,9 Valerie McGuire,10 Alice S. Whittemore,10 Simon A. Gayther,6 Paul D.P. Pharoah,5 Penelope M. Webb,1 Georgia Chenevix-Trench,1 Australian Ovarian Cancer Study Group,1,3 Australian Cancer Study (Ovarian Cancer),1 and Australian Breast Cancer Family Study4 1Queensland Institute of Medical Research; 2School of Population Health, University of Queensland, Brisbane, QLD, Australia; 3Peter MacCallum Cancer Center, East Melbourne, Victoria, Australia; 4Molecular, Environmental and Analytic Epidemiology, The University of Melbourne, Carlton, Victoria, Australia; 5CR-UK Department of Oncology, University of Cambridge, Strangeways Research Laboratory, Cambridge, United Kingdom; 6Translational Research Laboratories, Institute for Women’s Health, University College London, United Kingdom; 7Institute of Cancer Epidemiology, Danish Cancer Society; 8The Julianne Marie Center, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark; 9Department of Cancer Genetics, Roswell Park Cancer Institute, Buffalo, New York; and 10Department of Health Research and Policy, Stanford University School of Medicine, Palo Alto, California

Abstract

Although some high-risk ovarian cancer genes have V89L (rs523349), which showed a significant trend of been identified, it is likely that common low pene- increasing risk per rare allele (P = 0.00002). We then trance alleles exist that confer some increase in ovarian genotyped another SNP in this gene (rs632148; r2 = cancer risk. We have genotyped nine putative func- 0.945 with V89L) in an attempt to validate this finding tional single-nucleotide polymorphisms (SNP) in in an independent set of 1,479 cases and 2,452 controls genes involved in steroid hormone synthesis (SRD5A2, from United Kingdom, United States, and Denmark. CYP19A1, HSB17B1, and HSD17B4) and DNA repair There was no association between rs632148 and (XRCC2, XRCC3, BRCA2, and RAD52)usingtwo ovarian cancer risk in the validation samples, and Australian ovarian cancer case-control studies, com- overall, there was no significant heterogeneity be- prising a total of 1,466 cases and 1,821 controls of tween the results of the five studies. Further analyses Caucasian origin. Genotype frequencies in cases and of SNPs in this gene are therefore warranted to controls were compared using logistic regression. The determine whether SRD5A2 plays a role in ovarian only SNP we found to be associated with ovarian cancer predisposition. (Cancer Epidemiol Biomarkers cancer risk in both of these two studies was SRD5A2 Prev 2007;16(12):2557–65)

Introduction

Ovarian cancer is the leading cause of death from endometrioid, and clear cell (1). Mutations in the high- gynaecologic malignancy. The vast majority of malignant risk breast cancer susceptibility genes, BRCA1 and ovarian cancers are of epithelial origin and can be BRCA2, as well as the mismatch repair genes, MSH2 classified into four major subtypes: serous, mucinous, and MLH1, underlie most ‘‘hereditary’’ ovarian cancers (2). The known ovarian cancer susceptibility genes have been estimated to explain f40% of the excess familial Received 6/14/07; revised 9/9/07; accepted 10/1/07. risk of ovarian cancer (3). Thus, it is likely that other Grant support: Cancer Research UK, Roswell Park Alliance, Danish Cancer Society, and National Cancer Institute grants CA71766, CA16056, and RO1 CA61107. NHMRC ovarian cancer susceptibility genes exist. Several genetic Senior Principal Research Fellowship (G. Chenevix-Trench), NHMRC Career models may explain residual familial clustering, but Development award (A.B. Spurdle), Queensland Cancer Fund Senior Research Fellowship (P.M. Webb), Well-being of Women grant (H. Song), Cancer Research U.K. other highly penetrant genes are likely to be rare because Senior Clinical Research Fellowship (P.D.P. Pharoah), HEFCE Senior Lecturer fund mutations in BRCA1/2 are responsible for most families (S.A. Gayther), and Mermaid Component of the Eve Appeal (S.J. Ramus). The AOCS containing three or more ovarian cancer cases. A more was supported by U.S. Army Medical Research and Materiel Command grant DAMD17-01-1-0729, National Health and Medical Research Council of Australia grant plausible alternative is that the remaining familial 199600, Cancer Council Tasmania, and Cancer Foundation of Western Australia. clustering is driven by variants at multiple loci, each The costs of publication of this article were defrayed in part by the payment of page conferring a more moderate risk of the disease. Such charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. variants will also confer risks of nonfamilial ovarian Requests for reprints: G. Chenevix-Trench, Division of Cancer and Cell Biology, cancer. Queensland Institute of Medical Research, c/o Royal Brisbane Hospital Post Office, The ovarian surface epithelium or epithelial-lined Herston, Queensland 4029, Australia. Phone: 617-3362-0390; Fax: 61-7-3362-0105. E-mail: [email protected] inclusion cysts within the ovarian cortex have tradition- Copyright D 2007 American Association for Cancer Research. ally been considered to be the most likely site of origin doi:10.1158/1055-9965.EPI-07-0542 of epithelial ovarian cancers. The most widely cited

Cancer Epidemiol Biomarkers Prev 2007;16(12). December 2007

hypothesis for the etiology of these cancers is that rs17145454). To our knowledge, this is the first investi- proposed by Fathalla (4), which states that repeated gation of these SNPs in SRD5A2, CYP19A1, HSB17B1, ovulatory cycles increase the risk of ovarian cancer and HSD17B4 in ovarian cancer. because of the resulting proliferation of the surface We have examined SNPs in some of these genes in epithelium. If the incessant ovulation hypothesis of a previous case-control comparison (22-24) but have now ovarian cancer is correct, one might expect that errors extended this analysis to the Australian Ovarian Cancer arising during DNA synthesis might confer elevated Study (AOCS) to provide more statistical power, partic- risks of ovarian cancer. Genetic variants in double-strand ularly to evaluate rare variants and different histologic DNA repair genes, including those involved in homol- subtypes of ovarian cancer. We present here our analysis ogous recombination, such as XRCC2, XRCC3, BRCA2, of nine SNPs in a subset of genes involved in steroid and RAD52, may also influence cancer risk. hormone synthesis and DNA repair using two Australian However, the incessant ovulation model does not take case-control studies, comprising a total of 1,466 cases and into account all of the known epidemiologic risk factors 1,821 controls, as well as the attempted validation of the for ovarian cancer, and these other factors suggest a role most convincing finding from this analysis in an for hormones. In particular, ovarian cancer risk is independent set of 1,479 cases and 2,452 controls from decreased with increasing duration of contraceptive pill United Kingdom, United States, and Denmark. use and parity (5, 6). It has been specifically suggested that exposure of the ovarian surface epithelium to androgens may increase cancer risk and exposure to proges- Materials and Methods terones confer a protective effect (7, 8). Reproductive hormones control normal ovarian function by regulating Subjects. Clinical and epidemiologic data and DNA processes, such as cell proliferation, differentiation, and samples for genotyping were obtained from women who apoptosis. At ovulation, ovarian surface epithelial cells participated in two separate Australian studies of are exposed to very high levels of estrogens contained epithelial ovarian cancer and, for the SRD5A2 validation within follicular fluid (8), and Syed et al. (9) have shown set, from a consortium of three ovarian cancer studies that an increased concentration of estrogen stimulates from United Kingdom, United States, and Denmark ovarian surface epithelial cell proliferation in vitro (9). (25, 26). Approval for all studies was obtained from the Exposure to exogenous estrogens has been shown to be relevant Human Research Ethics Committees, and all related to increased ovarian cancer risk in most studies participants provided informed consent. (10-12). Study 1. This study has been described in detail In this study, we examined the role of specific, elsewhere (23, 24, 27-29) but, in brief, it included a total of putative functional single-nucleotide polymorphisms 510 Caucasian women newly diagnosed before the age of (SNP) in the SRD5A2, CYP19A1, HSB17B1, and 82 with epithelial ovarian cancer (including borderline HSD17B4 genes, encoding components of the hormone tumors) who were recruited as part of a population- synthetic pathways. The product of the SRD5A2 gene, a based case-control study conducted in New South Wales, 5- -reductase, catalyzes the conversion of testosterone Victoria, and Queensland between 1992 and 1995 to the more biologically active dihydrotestosterone. A (n = 341) or treated at the Royal Brisbane Hospital, common nonsynonymous coding SNP (V89L; rs523349) Queensland, between 1985 and 1996 (n = 169). The seems to affect the rate of this conversion. Recombinant comparison group came from two sources: 294 unrelated protein carrying the leucine residue has been shown to adult monozygotic twins (one from each pair) recruited result in 30% less testosterone production (13-15). CYP19 a into a multicenter study between 1992 and 1993 (30) and encodes the enzyme cytochrome P450c19 aromatase, 686 control women recruited into a population-based which catalyses the conversion of androgens to estrogens ¶ case-control study of breast cancer between 1992 and (16). An SNP in the 3 untranslated region (rs10046) has 2000 (31). As the monozygotic twin control set was not a been described previously, and the T allele has been population-based sample, we compared the distributions linked to ‘‘high enzyme activity’’ and increased CYP19 of genotypes in the two groups. There were no significant mRNA levels (17). An association has also been shown differences, so the two comparison groups were pooled. between the C allele and decreased circulating estradiol Epidemiologic data relating to potential confounders, levels (18). The final step of estradiol synthesis (conver- such as parity, oral contraceptive use, hormone replace- sion of estrone to the more biologically active estradiol) is ment therapy, and smoking status were only available catalyzed by type I 17h-hydroxysteroid dehydrogenase, b for the case and control women recruited through the which is encoded by the HSD17 1 gene (16). This population-based studies. enzyme also catalyses the conversion of the weak androgen, androstenedione, to testosterone (19). A non- Study 2. The second study was a national population- synonymous change in the protein sequence (S313G; based case-control study of ovarian cancer, the AOCS. rs605059) has been studied previously (20), although site- Between January 2002 and June 2006, eligible case directed mutagenesis shows little effect on catalytic or women ages between 18 and 79 years and newly immunologic activity of the protein (21). In this study, we diagnosed with epithelial ovarian cancer (including evaluated the role of HSD17b1 S313G and another borderline tumors) were recruited primarily in special- nonsynonymous change (A238V) on ovarian cancer risk. ized gynecologic oncology units by research nurses. Further regulation of estradiol levels occurs by unidi- Additional case women missed at the major treatment rectional oxidation of estrone by type 4 peroxisomal centers were identified through cancer registries and, 17h-hydroxysteroid dehydrogenase, encoded by the with the treating doctor’s permission, invited to partic- HSD17b4 gene, and so, we also genotyped a previously ipate (recruitment through the New South Wales and unstudied SNP in the 3¶ end of this gene (W511R; Victorian Cancer Registries was conducted in parallel

Cancer Epidemiol Biomarkers Prev 2007;16(12). December 2007

under a separately funded study: the Australian Cancer Laboratory Methods. DNA extraction methods for Study). Women who were unable to give informed study 1 have been described earlier (27, 28). For study 2, consent due to language difficulties, mental incapacity, DNA was extracted from peripheral blood using a salt or illness were excluded, as were those whose diagnosis extraction method modified from (ref. 32) or QiAMP was not histopathologically confirmed. Of the 4,005 blood kit (Qiagen, Inc.). All DNA samples were women identified as potentially eligible for the study, quantitated using a NanoDrop ND-1000 spectrophotom- 806 (20%) were excluded due to death (n = 304), illness eter, and then 5 ng of each gDNA was aliquoted into (n = 201), or inability to give informed consent (n = 301). 384-well plates and dried down at room temperature. Of those invited to participate, 2,714 (85%) agreed to take Case and control DNAs were randomly assigned to well part (68% of those identified). Of these, a further 730 positions. Each plate included 36 randomly chosen, women (27%) were excluded after pathology review repeat samples to confirm assay reproducibility and four because they were not confirmed as having an eligible negative, template-free controls. cancer. Two researchers independently abstracted infor- SNPs were genotyped using MALDI-TOF mass spec- mation on site of tumor, histologic subtype, invasiveness, trophotometric mass determination of allele-specific and grade from histology reports, and discrepancies primer extension products using Sequenom’s MassAR- were resolved by consensus. A formal review of a full set RAY system. The design of oligonucleotides was carried of diagnostic slides was undertaken for a sample of 87 out according to the guidelines of Sequenom, Inc., and women by one of a group of gynaecologic pathologists. was done using MassARRAY assay design software There was agreement of 97% on tumor site, 98% on (Version 1.0). Primer sequences are available on request. tumor behavior, and 99% on tumor subtype between the Multiplex PCR amplification of amplicons containing results of the formal review and the abstracted data. SNPs of interest was done using Qiagen HotStart Taq Control women, frequency matched to the entire case Polymerase and Perkin-Elmer GeneAmp 2400 thermal series, were randomly selected from the Australian cycler. Primer extension reactions were carried out Electoral Roll (enrollment to vote is compulsory in according to manufacturer’s instructions for either Australia) after stratifying for age (in 5-year groups) homogenous MassEXTEND or iPLEX chemistries. Assay and state of residence. Selected women were mailed data were analyzed using Sequenom TYPER software study information and subsequently contacted by tele- (Version 3.0). Seven of the nine SNPs were genotyped phone by research nurses. At least five attempts were with both the homogenous MassEXTEND and iPLEX made to telephone each woman, and those not contacted chemistries to obtain confidence with the new iPLEX were sent a second letter. Women were excluded if they chemistry and to confirm deviations from Hardy-Wein- reported a previous history of ovarian cancer or a berg equilibrium with a different chemistry. The concor- previous bilateral oophorectomy, as were those women dance was 99.88% to 100% for each SNP. Whenever the who were unable to give informed consent due to illness, two chemistries provided different results (0.12% geno- mental incapacity, or language difficulties. Of the types), we omitted the discrepant sample. potential control women contacted and invited to The V89LSNP was not amenable to TaqMan (Applied participate, two percentage were excluded on the basis Biosystems) genotyping technology, so for the SRD5A2 of illness or language difficulties and of the remaining validation study, we genotyped rs632148, which was women, 1,614 (47%) agreed to take part. Of these, 104 reported to tag the V89L(rs523349) SNP with a pair-wise women were excluded from analyses because they correlation coefficient (r2) of 0.945. To estimate r2 in our reported a prior ovarian cancer (n = 7) or bilateral own population, we genotyped 30 individuals (33 cases oophorectomy (n = 97). and 57 controls) with each of the three rs523349 All participants were asked to complete a detailed genotypes for the rs632148 SNP. health and life-style questionnaire, and 87% of eligible Data Analysis. cases (n = 1717) and 85% of controls (n = 1287) provided The two Australian data sets were initially analyzed separately to provide a test and a sample of blood. The questionnaire covered demo- replication set and were then combined to increase the graphic and physical characteristics, family history, power, particularly to examine the different histologic medical and surgical history, life-style habits (including subtypes. Analyses were restricted to Caucasian women smoking and alcohol consumption), and reproductive (510 cases and 980 controls for study 1 and 956 and 841 and contraceptive histories. Missing information and/or for study 2) who made up >95% of each study inconsistencies were clarified during a subsequent population. The Hardy-Weinberg equilibrium assump- telephone interview. For this study, we genotyped 956 tion was first assessed in case and control groups for each AOCS cases and 841 AOCS controls for whom DNA was SNP using a standard m2 test. Univariate tests of available at the time. association for each SNP were carried out using the SRD5A2 Validation Set. Three case-control studies likelihood ratio test with 2 df. Odds ratios (OR) and 95% from United Kingdom (SEARCH), Denmark (MALOVA), confidence intervals (95% CI) were calculated using and United States (GEOC), comprising 1,479 cases and unconditional logistic regression. All models were 2,452 controls, were used for a validation set. These adjusted for age (at diagnosis for cases and at first studies have been described in detail previously (25, 26). contact for controls) as a continuous term, and when data Analysis included only 1,400 cases and 2,393 controls of from more than one study were combined, a term for self-reported Caucasian ancestry for whom genotype study was also included. Other factors, such as parity information was available. Tumor histologic subtype (0, 1-2, z3), pregnancies (z6 months of duration), and breakdown for cases was as follows: serous 47%, use of oral contraceptives (never, <5 years, z5 years of mucinous 11%, endometrioid 16%, clear cell carcinoma duration) were not included in final models as they 8%, other 19%. changed point estimates by <10%. Stratification was used

Cancer Epidemiol Biomarkers Prev 2007;16(12). December 2007

to investigate gene-environment interactions, and if this A statistically significant association between SRD5A2 suggested interaction, the statistical significance of the genotype and ovarian cancer risk was also seen in study relevant multiplicative term was assessed. Heterogeneity 2(Ptrend = 0.0001). As for study 1, there was a small between the results from different studies was assessed increase in risk associated with the GC genotype using an interaction term for study by each gene. A compared with the GG genotype (OR, 1.14; 95% CI, nominal P value of <0.05 was considered statistically 0.93-1.40) and a larger statistically significant OR significant. Analysis of the validation set was as above. associated with the CC genotype (OR, 1.83; 95% CI 1.37-2.44; P = 0.00004). There was also a significant trend of increasing risk per C allele (30% increase in risk per C Results allele, P = 0.0001). In study 2, the trend of decreasing risk with one or more C alleles was not as marked for There was no evidence of deviation from the Hardy- HSD17b4 as that seen in study 1, but the CC homo- Weinberg equilibrium in controls from study 1 for eight zygotes again had a nonsignificant decreased risk of of the nine SNPs investigated, although there was ovarian cancer (OR, 0.59; 95% CI, 0.17-2.11; P = 0.4). In borderline significant deviation from Hardy-Weinberg study 2, the AC genotype of BRCA2 (rs144848) was equilibrium for the HSD17b1 Gly313Ser SNP (rs605059) associated with a 36% increase in risk (OR, 1.36; 95% CI, gene (P = 0.05). In study 2, there was also little evidence 1.12-1.66) compared with the AA genotype, but those of deviation from the Hardy-Weinberg equilibrium but with CC genotypes did not seem to be at increased risk. It there was a statistically significant deviation from Hardy- is of note that there was significant deviation from Weinberg equilibrium for BRCA2 (rs144848; P = 0.04) Hardy-Weinberg equilibrium for BRCA2 in this group of with the results from both homogenous MassEXTEND control women with a deficiency of heterozygotes, but, as and iPLEX chemistry (33). Cases and controls from study described above, this could not be attributed to genotyp- 2 (mean age, 58 and 57 years, respectively) and cases ing errors. from study 1 (mean age, 57 years) were on average of When the data from study 1 and study 2 were very similar in age, whereas the controls from study 1 combined, there was no evidence of heterogeneity (mean age, 44 years) were significantly younger (P < between the studies for any of the genotypes (Pinteraction 0.0001); all results were thus adjusted for age. The > 0.05). In the combined data set, the association for distribution of histologic subtypes of tumors was similar SRD5A2 rs523349 became highly significant (Ptrend = for both studies (Table 1). 0.00002). Both heterozygotes (OR, 1.16; 95% CI, 1.00-1.36; The risk of epithelial ovarian cancer associated with P = 0.06) and CC homozygotes (OR, 1.70; 95% CI, 1.35- the various SNPs was determined separately for study 1 2.16, P = 0.000009) had increased risk, and there was a and study 2 (Table 2). For study 1, the only statistically significant trend of increasing risk per extra C allele (26% significant association was seen for SRD5A2 (rs523349). per C allele, P = 0.00002). In the combined analysis, no The rare allele was associated with an increased risk in a statistically significant associations were found for codominant manner (Ptrend = 0.02) with the GC genotype HSD17b4, but a small significant risk persisted for those showing a 30% increase in risk compared with the GG heterozygous for the BRCA2 SNP (OR, 1.27; 95% CI, 1.09- genotype (1.30; 95% CI, 1.01-1.68) and the CC genotype 1.49; P < 0.005). In addition, a small, but statistically having a slightly higher risk (OR, 1.44; 95% CI, 0.93-2.24), significant, increase in risk was associated with the TC albeit not significant at the 5% level. Having a C allele for (OR, 1.20; 95% CI, 1.01-1.43; P = 0.04) but not the CC (OR, HSD17b4 (rs17145454) was associated with a nonsignif- 1.15; 95% CI, 0.93-1.41) genotype of CYP19A1 (rs10046). icant decreased risk of ovarian cancer (OR, 0.79; 95% CI, We used the combined dataset to investigate whether 0.56-1.12 for CT versus TT; OR, 0.31; 95% CI, 0.07-1.53 for risk associated with the various SNPs might be modified CC versus TT; Ptrend = 0.07), but no associations of note by exposures known to influence ovarian cancer risk. were seen between the other investigated SNPs and risk Stratifying analyses by use of oral contraceptives (ever of ovarian cancer. versus never), use of hormone replacement therapy (ever versus never), smoking status (ever versus never), and age (<50 years compared with z50 years) did not provide Table 1. Histologic subtypes of ovarian tumors of case any evidence that the genotype effects varied according women in Australian studies 1 and 2 to these characteristics. A borderline significant interaction (P = 0.045) associated with parity was found for the Histologic Study 1 Study 2 subtype of cancer RAD52 SNP (rs4987208). Women who were nulliparous Cases Cases and were GT heterozygotes for this SNP had a significant (n = 510), n (%) (n = 956), n (%) 75% decrease in risk of ovarian cancer compared with TT homozygotes (OR, 0.25; 95% CI, 0.06-0.98), whereas OR Invasive for parous women with the GT genotype was 1.06 (95% Serous 257 (50) 517 (54) Mucinous 27 (5) 21 (3) CI, 0.64-1.75). No other significant interactions were seen Endometrioid 59 (12) 78 (8) for parity. Clear cell 26 (5) 44 (5) Variations in risk between histologic subtypes of Other* 55 (11) 92 (9) epithelial ovarian cancer were investigated using the Borderline combined data from study 1 and study 2 (Table 3). The Serous 45 (9) 95 (10) Mucinous 36 (7) 98 (10) positive association between the C allele for SRD5A2 and Other 5 (1) 9 (1) risk persisted for borderline and invasive tumors of the serous and mucinous subtypes (but did not reach *Includes mixed epithelial cancers, transitional cell cancers, and statistical significance for borderline serous tumors) malignant mixed mu¨llerian tumors. and seemed to be somewhat stronger for invasive

Cancer Epidemiol Biomarkers Prev 2007;16(12). December 2007

Table 2. ORs and 95% CIs for the associations between genotype and risk of epithelial ovarian cancer in Australian studies 1 and 2

Gene Cases/controls (n) Cases/controls (n) Age-adjusted ORs (95% CIs)

Study 1 Study 2 Study 1 Study 2 Combined SRD5A2 (rs523349) GG 234/510 392/402 1.00 1.00 1.00 GC 217/378 378/336 1.30 (1.01-1.68)* 1.14 (0.93-1.40)c 1.16 (1.00-1.36)c CC 52/71 169/94 1.44 (0.93-2.24) 1.83 (1.37-2.44) 1.70 (1.35-2.16) P = 0.03 P = 0.0002 HSD17B4 (rs17145454) TT 438/819 767/679 1.00 1.00 1.00 CT 66/145 141/114 0.79 (0.56-1.12) 1.10 (0.84-1.44) 0.99 (0.81-1.22) CC 2/11 4/6 0.31 (0.07-1.53) 0.59 (0.17-2.11) 0.47 (0.18-1.23) P =0.2 P =0.6 HSD17B1 gly313ser (rs605059) AA 125/291 282/247 1.00 1.00 1.00 GA 229/446 442/404 1.20 (0.91-1.59) 0.96 (0.77-1.20) 1.07 (0.90-1.27) GG 87/215 189/147 0.93 (0.66-1.32) 1.13 (0.86-1.49) 1.06 (0.86-1.31) P =0.2 P =0.5 CYP19A1 (rs10046) TT 115/269 262/247 1.00 1.00 1.00 TC 254/479 469/387 1.28 (0.95-1.72) 1.15 (0.92-1.43) 1.20 (1.01-1.43)* CC 130/221 208/193 1.33 (0.94-1.88) 1.02 (0.78-1.32) 1.15 (0.93-1.41) P =0.1 P =0.4 HSD17B1 ala238val CC 497/356 725/739 1.00 1.00 1.00 TC 6/5 15/21 1.13 (0.34-3.73) 0.73 (0.37-1.43) 0.77 (0.43-1.39) TT 1/0 0/0 P =0.6 P =0.4 BRCA2 (rs144848) [1] AA 249/502 460/461 1.00 1.00b 1.00 b AC 203/383 401/296 1.19 (0.92-1.54) 1.36 (1.12-1.66) 1.27 (1.09-1.49) CC 40/63 69/68 1.13 (0.70-1.84) 1.02 (0.71-1.46) 1.11 (0.83-1.47) P =0.5 P = 0.008 XRCC2 (rs3218536) [2] GG 414/819 799/696 1.00 1.00 1.00 AG 67/142 117/115 0.84 (0.59-1.20) 0.88 (0.67-1.16) 0.86 (0.70-1.07) AA 5/8 7/7 1.23 (0.37-4.14) 0.92 (0.32-2.63) 1.19 (0.54-2.61) P =0.8 P =0.7 XRCC3 (rs861539) [2] CC 207/370 291/288 1.00 1.00 1.00 TC 223/471 339/351 0.83 (0.64-1.08) 0.96 (0.77-1.20) 0.91 (0.77-1.07) TT 74/131 101/108 1.03 (0.71-1.49) 0.92 (0.67-1.26) 0.95 (0.75-1.20) P =0.3 P =0.9 RAD52 (rs4987208) [3] TT 458/348 909/806 1.00 1.00 1.00 GT 13/13 28/26 0.69 (0.31-1.55) 0.95 (0.55-1.63) 0.88 (0.55-1.35) P =0.5 P =0.9

NOTE: [1], data from study 1 have been previously reported in Auranen et al. (29); [2], data from study 1 have been previously reported in Webb et al. (24); [3], data from study 1 have been previously reported in Kelemen et al. (23). *P < 0.05. cP < 0.0001. bP < 0.005. mucinous cancers (OR, 2.81; 95% CI, 1.23-6.38 for CC for CC homozygotes of SRD5A2 compared with GG versus GG genotype). We also considered site of tumor homozygotes was somewhat higher for serous peritoneal as the case groups included women with primary cancers (OR, 2.64; 95% CI, 1.44-4.83) than for serous peritoneal and fallopian tube cancers. Although these ovarian cancers (OR, 1.81; 95% CI, 1.33-2.48), there were cancers are clinically and histologically similar to their no statistically significant differences by tumor site. ovarian counterparts, our previous analyses suggest that Two associations were statistically significant for there are etiologic differences between ovarian/fallopian invasive clear cell cancers. The CT genotype of HSD17b4 tube and primary peritoneal tumors.11 Although the OR was associated with a 65% decrease in risk of clear cell cancer (OR, 0.35; 95% CI, 0.13-0.96; P = 0.04). There were no women with clear cell tumors who were homozygous CC for HSD17b4. Risk of clear cell cancers was also 11 Susan J Jordan, Ade`le C. Green, David C. Whiteman, Christopher J. associated with the HSD17b1 Gly313Ser (rs605059) SNP. Bain, Dorota M. Gertig M., Penelope M. Webb for the Australian Cancer Having the GG genotype was associated with a more Study Group (ovarian cancer) and the Australian Ovarian Cancer Study Group. Serous ovarian, fallopian tube and primary peritoneal cancers — than 2-fold increase in risk of clear cell cancer compared one disease or three? In press. International Journal of Cancer. with having the AA genotype (OR, 2.33; 95% CI,

Cancer Epidemiol Biomarkers Prev 2007;16(12). December 2007

1.14-4.79; P = 0.02), and there was a significant trend of showed that, in this sample, the r2 between the two SNPs risk increase associated with each additional G allele was 0.9. (52% per G allele; P = 0.02). There was marginal evidence of deviation from the The only nominally statistically significant finding for Hardy-Weinberg equilibrium in cases for the validation the DNA repair SNPs related to serous borderline set overall (P = 0.06), which seemed to be driven by a tumors. The AG genotype for XRCC2 (rs3218536) was deficiency of heterozygotes in the SEARCH sample. associated with a 51% decreased risk compared with the Although there was no evidence for Hardy-Weinberg GG genotype (OR, 0.49; 95% CI, 0.25-0.95; P = 0.03), disequilibrium in controls overall, there was a similar although AA homozygotes had a nonsignificant 2-fold deficiency of heterozygotes in controls in the MALOVA increase in risk (OR, 2.12; 95% CI, 0.59-7.58). sample (P = 0.03). Given our similar findings for SRD5A2 (rs523349) in We found no association between this SRD5A2 SNP both the Australian case-control studies, we sought to and ovarian cancer risk in any of these validation case- validate this result in a large consortium of three case- control sets individually nor when the three studies control studies from United Kingdom, United States, and were pooled (Ptrend = 0.9; Table 4). There was no Denmark (25, 26). The rs523349 was not amenable to evidence for heterogeneity between the three sample genotyping by TaqMan, so instead we genotyped a 3¶ sets, and the combined OR for the validation set was untranslated region tagging SNP in SRD5A2 (rs632148), not significantly different from 1 for the heterozygote which tags rs523349 with an r2 of 0.945 based on the genotype (OR, 0.93; 95% CI 0.80-1.07) or the homozy- HapMap genotype data for the CEU samples (30 trios of gote genotype (OR, 1.12; 95% CI 0.89-1.40). Analysis of European origin). To determine the r2 in our own subgroups defined by histologic subtype revealed no population, we genotyped 90 individuals from the AOCS evidence for an increased risk within any particular with both the rs523349 and rs632148 SNPs and also subtype (data not shown).

Table 3. ORs for genotype by ovarian tumor behavior and histologic subtype (combined results from Australian studies 1 and 2)

Gene Cases/controls ORs*

n Serous Mucinous Serous Mucinous Endometrioid Clear cell borderline Borderline invasive invasive invasive invasive

(n = 140) (n = 134) (n = 772) (n = 48) (n = 137) (n = 70) SRD5A2 (rs523349) GG 626/912 1.00 1.00 1.00 1.00 1.00 1.00 GC 595/714 1.03 (0.71-1.50) 1.61 (1.09-2.38)c 1.12 (0.92-1.36)b 1.27 (0.66-2.46)c 1.17 (0.81-1.70) 0.71 (0.42-1.20) CC 221/165 1.54 (0.90-2.62) 1.79 (1.01-3.18) 1.80 (1.36-2.38) 2.81 (1.23-6.38) 1.06 (0.57-1.99) 0.90 (0.39-2.07) HSD17B4 (rs17145454) TT 1205/1498 1.00 1.00 1.00 1.00 1.00 1.00 c CT 207/259 1.32 (0.84-2.09) 1.13 (0.69-1.85) 1.01 (0.78-1.30) 0.50 (0.18-1.41) 0.96 (0.59-1.58) 0.35 (0.13-0.96) CC 6/17 0.99 (0.13-7.64) 0.93 (0.12-7.24) 0.56 (0.18-1.73) — — — HSD17B1 gly313ser (rs605059) AA 422/547 1.00 1.00 1.00 1.00 1.00 1.00 GA 702/857 1.10 (0.72-1.67) 0.87 (0.58-1.30) 1.02 (0.83-1.26) 1.43 (0.72-2.86) 1.09 (0.71-1.65) 1.67 (0.87-3.20)c GG 291/368 1.41 (0.86-2.30) 0.79 (0.46-1.35) 0.93 (0.71-1.21) 0.85 (0.33-2.18) 1.20 (0.73-1.98) 2.33 (1.14-4.79) CYP19A1 (rs10046) TT 377/516 1.00 1.00 1.00 1.00 1.00 1.00 TC 723/866 1.16 (0.77-1.77) 1.15 (0.74-1.77) 1.23 (0.99-1.53) 0.56 (0.28-1.11) 1.35 (0.86-2.12) 1.19 (0.67-2.14) CC 338/414 1.11 (0.68-1.82) 1.23 (0.75-2.03) 1.11 (0.86-1.44) 0.94 (0.45-1.95) 1.46 (0.88-2.43) 1.03 (0.51-2.08) HSD17B1 ala238val CC 1222/1095 1.00 1.00 1.00 1.00 1.00 1.00 TC 21/26 0.32 (0.04-2.43) 0.76 (0.18-3.26) 1.03 (0.53-2.00) — 0.39 (0.05-2.92) 0.80 (0.11-6.04) BRCA2 (rs144848) AA 709/963 1.00 1.00 1.00 1.00 1.00 1.00 AC 604/679 1.38 (0.97-1.98) 1.19 (0.82-1.73) 1.39 (1.14-1.68) 0.71 (0.36-1.40) 1.12 (0.77-1.63) 1.00 (0.59-1.68) CC 109/131 0.63 (0.27-1.49) 0.73 (0.33-1.64) 1.19 (0.84-1.68) 1.95 (0.83-4.60) 1.01 (0.50-2.02) 0.56 (0.17-1.84) XRCC2 (rs3218536) GG 1213/1515 1.00 1.00 1.00 1.00 1.00 1.00 AG 184/257 0.49 (0.25-0.95) 0.73 (0.41-1.30) 0.88 (0.68-1.15) 1.86 (0.93-3.73) 0.83 (0.49-1.41) 0.78 (0.37-1.68) AA 12/15 2.12 (0.59-7.58) — 1.13 (0.42-3.04) — 2.01 (0.44-9.08) 2.25 (0.29-17.76) XRCC3 (rs861539) CC 498/658 1.00 1.00 1.00 1.00 1.00 1.00 TC 562/822 0.98 (0.66-1.44) 1.05 (0.68-1.62) 0.88 (0.72-1.09) 1.18 (0.60-2.34) 0.83 (0.56-1.23) 0.90 (0.51-1.58) TT 175/239 0.59 (0.30-1.16) 1.35 (0.76-2.38) 1.00 (0.75-1.34) 1.16 (0.44-3.07) 0.72 (0.39-1.33) 0.88 (0.38-1.99) RAD52 (rs4987208) TT 1367/1154 1.00 1.00 1.00 1.00 1.00 1.00 GT 41/39 0.45 (0.11-1.91) 0.49 (0.12-2.06) 0.99 (0.58-1.68) 0.66 (0.09 4.99) 1.12 (0.43-2.92) 0.88 (0.21-3.74)

*Adjusted for age and study. cP V 0.05. bP < 0.0001.

Cancer Epidemiol Biomarkers Prev 2007;16(12). December 2007

Table 4. ORs and 95% CIs for the associations between SRD5A2 (rs632148) genotype and risk of epithelial ovarian cancer in the validation set

Gene Cases/ Cases/ Cases/ Age-adjusted ORs (95% CIs) controls (n) controls (n) controls (n)

SEARCH MALOVA STANFORD SEARCH MALOVA STANFORD Combined SRD5A2 (rs632148) GG 335/397 196/553 138/167 1.00 1.00 1.00 1.00 GC 270/368 184/494 121/168 0.87 (0.70-1.08) 1.03 (0.81-1.31) 0.88 (0.64-1.23) 0.93 (0.80-1.07) CC 74/71 58/146 24/29 1.18 (0.82-1.69) 1.11 (0.78-1.57) 1.01 (0.56-1.83) 1.12 (0.89-1.40) Ptrend = 0.13 Ptrend =0.8 Ptrend =0.7

When the data from all five studies were combined, polycystic ovary syndrome are at increased risk of there was no evidence of heterogeneity between the developing ovarian cancer (35). The SRD5A2 V89LSNP studies (P = 0.15), and the pooled OR suggested a has been examined in several prostate cancer studies with statistically significant 40% increase in risk of ovarian inconsistent results (36-43). However, a comprehensive cancer among rare homozygotes (Ptrend = 0.002). metaanalysis of SRD5A2 V89Lfrequencies in prostate cancer association studies clearly excluded any increased risk conferred by this allele (44). Discussion We sought to validate the consistent association that we found in studies 1 and 2 for SRD5A2 V89L(rs523349) We used two large Australian case-control populations to and ovarian cancer risk in the SEARCH/MALOVA/ test for associations between nine SNPs in genes Stanford consortium (25, 26) using a tagging SNP involved in steroid hormone synthesis and DNA repair (rs632148) that correlated with rs523349 (r2 = 0.9). The and ovarian cancer risk. All analyses were restricted to combined sample size of the three validation studies had Caucasians to reduce the potential problems of popula- >95% power to detect an OR of 1.7 for the homozygote tion stratification. For many of the SNPs we genotyped, genotype; thus, it should have been large enough to see we and others had previously reported no or weak an association even allowing for the less than perfect associations with ovarian cancer risk (22-24, 29). The correlation between the SNPs. Despite this, we did not main rationale for this study was therefore to have find any evidence of an association between SRD5A2 additional power to look for associations with different rs632148 and ovarian cancer risk in any of these three histologic subtypes of ovarian cancer. In study 1, the only case-control studies or in combined analyses, although a significant finding was seen for SRD5A2 (rs523349), and 40% increase in risk for the rare homozygote could not be this increased risk associated with the GC and CC excluded (combined OR, 1.1; 95% CI, 0.9-1.4). Despite the genotypes was also noted in study 2 with a 30% strong association seen in the initial studies and the lack increased risk per C allele (P = 0.0001). This association of association in the three validation studies, overall, was seen for both invasive and borderline tumors of the there was no significant heterogeneity between the serous and mucinous subtypes of ovarian cancer, but results of the five studies. Given this lack of heterogene- there was no association for endometrioid and clear cell ity, the best estimate of the association is obtained from tumors. The exclusion of women who died before the combined analysis of all five studies which suggested recruitment or were very sick means that very aggressive a significant 40% increase in risk of ovarian cancer disease may be underrepresented in the case group. among women with the rare homozygote genotype of However, this is unlikely to have introduced any bias SRD5A2 V89Lor rs632148. Given the key role for unless the etiology of aggressive cancers differs mark- SRD5A2 in androgen metabolism, further analysis of edly from that of less aggressive cancers. The borderline SNPs in the SRD5A2 gene is warranted in even larger significant findings with HSD17B4 (rs17145454) in study studies, as would be a comprehensive analysis of SNPs in 1 and with BRCA2 (rs144848) and CYP19A1 (rs10046) in CYP19A1, HSB17B1, HSD17B4, XRCC2, XRCC3, BRCA2, study 2 were not replicated in the other study. In and RAD52. addition, the modest associations seen for clear cell cancers in the combined analysis are likely to be due to chance as a result of the relatively large number of tests Appendix A. The AOCS Group done. No associations were seen for any of the SNPs we genotyped in DNA repair genes in either study 1 or Management Group: D. Bowtell (Peter MacCallum study 2. Cancer Center, PMCC), G. Chenevix-Trench, A. Green, SRD5A2 catalyses the conversion of testosterone to P. Webb (Queensland Institute of Medical Research, dihydrotestosterone, a more potent androgen than QIMR), A. deFazio (Westmead Hospital), D. Gertig testosterone. Functional studies of SRD5A2 variants (University of Melbourne). using recombinant enzyme constructs have shown Project Managers: N. Traficante (PMCC), S. Moore differential enzyme activities (15). In particular, the valine (QIMR), J. Hung (Westmead Hospital). variant of V89Lhas an increased turnover rate compared Data Managers: S. Fereday (PMCC), K. Harrap, T. with the leucine variant. Little is known of the role of Sadkowsky (QIMR). testosterone in the ovary, but high levels of androgens Research Nurses: NSW — A. Mellon, R. Robertson have been linked to polycystic ovary syndrome (reviewed (John Hunter Hospital), T. Vanden Bergh (Royal Hospital in ref. 34), and there is some evidence that women with for Women), J. Maidens (Royal North Shore Hospital),

Cancer Epidemiol Biomarkers Prev 2007;16(12). December 2007

K. Nattress (Royal Prince Alfred Hospital), Y.E. Chiew, 3. Antoniou AC, Easton DF. Risk prediction models for familial breast cancer. Future Oncol 2006;2:257 – 274. A. Stenlake, H. Sullivan (Westmead Hospital); QLD — B. 4. Fathalla MF. Incessant ovulation-a factor in ovarian neoplasia? Alexander, P. Ashover, S. Brown, T. Corrish, L. Green, L. Lancet 1971;2:163. Jackman, K. Martin, B. Ranieri (QIMR); SA — J. White 5. Purdie D, Green A, Bain C, et al. Reproductive and other factors and (QIMR); TAS — V. Jayde (Royal Hobart Hospital); VIC — risk of epithelial ovarian cancer: an Australian case-control study. Survey of Women’s Health Study Group. Int J Cancer 1995;62: L. Bowes (PMCC), P. Mamers (Monash Medical Center); 678 – 84. WA — T. Schmidt, H. Shirley, S. Viduka, Hoa Tran, 6. Riman T, Persson I, Nilsson S. Hormonal aspects of epithelial ovarian Sanela Bilic, Lydia Glavinas (Western Australia Research cancer: review of epidemiological evidence. Clin Endocrinol (Oxf) Tissue Network). 1998;49:695 – 707. 7. Gadducci A, Cosio S, Gargini A, et al. Sex-steroid hormones, Clinical Collaborators: NSW — A. Proietto, S. Braye, G. gonadotropin and ovarian carcinogenesis: a review of epidemiolo- Otton (John Hunter Hospital), T. Bonaventura, J. Stewart gical and experimental data. Gynecol Endocrinol 2004;19:216 – 28. (Newcastle Mater Misericordiae), M. Friedlander (Prince 8. Risch HA. Hormonal etiology of epithelial ovarian cancer, with a of Wales Hospital), D. Bell, S. Baron-Hay, A. Ferrier, G. hypothesis concerning the role of androgens and progesterone. J Natl Cancer Inst 1998;90:1774 – 86. Gard, D. Nevell, B. Young (until mid 2003; Royal North 9. Syed V, Ulinski G, Mok SC, et al. Expression of gonadotropin Shore Hospital), C. Camaris, R. Crouch, L. Edwards, N. receptor and growth responses to key reproductive hormones in Hacker, D. Marsden, G. Robertson (Royal Hospital for normal and malignant human ovarian surface epithelial cells. Cancer Women), P. Beale, J. Beith, J. Carter, C. Dalrymple, A. Res 2001;61:6768 – 76. 10. Folsom AR, Anderson JP, Ross JA. Estrogen replacement therapy and Hamilton, R. Houghton, P. Russell (Royal Prince Alfred ovarian cancer. Epidemiology 2004;15:100 – 4. Hospital), A. Brand, R. Jaworski, P. Harnett, G. Wain 11. Lacey JV, Jr., Brinton LA, Leitzmann MF, et al. Menopausal hormone (Westmead Hospital); QLD — A. Crandon, M. Cum- therapy and ovarian cancer risk in the National Institutes of Health- mings, K. Horwood, A. Obermair, D. Wyld (Royal AARP Diet and Health Study Cohort. J Natl Cancer Inst 2006;98: 1397 – 405. Brisbane and Women’s Hospital, RBWH), J. Nicklin 12. Riman T, Dickman PW, Nilsson S, et al. Hormone replacement (RBWH and Wesley Hospital), L. Perrin (RBWH and therapy and the risk of invasive epithelial ovarian cancer in Swedish Mater Misericordiae Hospitals), B. Ward (Mater Miser- women. J Natl Cancer Inst 2002;94:497 – 504. icordiae Hospitals); SA — M. Davy, C. Hall, T. Dodd, T. 13. Allen NE, Forrest MS, Key TJ. The association between polymorphisms in the CYP17 and 5a-reductase (SRD5A2) genes and serum Healy, K. Pittman (Royal Adelaide Hospital, Burnside androgen concentrations in men. Cancer Epidemiol Biomarkers Prev Memorial Hospital), D. Henderson, S. Hyde (Flinders 2001;10:185 – 9. Medical Center), J. Miller, J. Pierdes (Queen Elizabeth 14. Makridakis N, Ross RK, Pike MC, et al. A prevalent missense Hospital); TAS — P. Blomfield, D. Challis, R. McIntosh, A. substitution that modulates activity of prostatic steroid 5a-reductase. Cancer Res 1997;57:1020 – 2. Parker (Royal Hobart Hospital); VIC — B. Brown, R. 15. Makridakis NM, di Salle E, Reichardt JK. Biochemical and Rome (Freemasons Hospital), D. Allen, P. Grant, S. Hyde, pharmacogenetic dissection of human steroid 5 a-reductase type II. R. Laurie, M. Robbie (Mercy Hospital for Women), Pharmacogenetics 2000;10:407 – 13. D. Healy, T. Jobling, T. Maniolitas, J. McNealage, P. 16. Sasano H, Harada N. Intratumoral Aromatase in Human Breast, Endometrial, and Ovarian Malignancies. Endocr Rev 1998;19: Rogers, B. Susil, A. Veitch, J. Constable, S. Ping Tong, 593 – 607. I. Robinson, I. Simpson (Monash Medical Center), K. 17. Kristensen VN, Harada N, Yoshimura N, et al. Genetic variants Phillips, D. Rischin, P. Waring, M. Loughrey, N. O’Call- of CYP19 (aromatase) and breast cancer risk. Oncogene 2000;19: aghan, Bill Murray (PMCC), V. Billson, S. Galloway, J. 1329 – 33. 18. Dunning AM, Dowsett M, Healey CS, et al. J. Polymorphisms Pyman, M. Quinn (Royal Women’s Hospital); WA — I. Associated With Circulating Sex Hormone Levels in Postmenopausal Hammond, A. McCartney, Y. Leung (King Edward Women. J Natl Cancer Inst 2004;96:936 – 45. Memorial Hospital, St. John of God). 19. Lin SX, Shi R, Qiu W, et al. Structural basis of the multispecificity Scientific Collaborators: I. Haviv (PMCC), D. Purdie, shown by 17h-hydroxysteroid dehydrogenase types 1 and 5. Mol Cell Endocrinol 2006;248:38 – 46. D. Whiteman (QIMR), N. Zeps (WARTN). 20. Mannermaa A, Peltoketo H, Winqvist R, et al. Human familial and sporadic breast cancer: analysis of the coding regions of the 17 h-hydroxysteroid dehydrogenase 2 gene (EDH17B2) using a Acknowledgments single-strand conformation polymorphism assay. Hum Genet 1994; We thank Adele Green and David Purdie for DNA and data 93:319 – 24. from the Survey of Women’s Health; Nick Martin for the DNA 21. Puranen TJ, Poutanen MH, Peltoketo HE, et al. Site-directed and data from SSAGA (study 1); New South Wales, Queensland, mutagenesis of the putative active site of human 17 h-hydroxysteroid South Australian, Victorian, and Western Australian Cancer dehydrogenase type 1. Biochem J 1994;304:289 – 93. Registries, as well as all the collaborating institutions repre- 22. Auranen A, Song H, Waterfall C, et al. Polymorphisms in DNA repair genes and epithelial ovarian cancer risk. Int J Cancer 2005;117: sented within the AOCS Study Group and the women who 611–8. participated in the study, for their cooperation; Hannah 23. Kelemen L, Spurdle AB, Purdie DM, et al. RAD52 Y415X truncation Munday, Barbara Perkins, Clare Jordan, Mitul Shah, the local polymorphism and epithelial ovarian cancer risk in Australian general practices and nurses, and Eastern Cancer Intelligence women. Cancer Lett 2005;218:191 – 7. Unit for recruitment of U.K. cases; Lydia Quaye for genotyping 24. Webb PM, Hopper JL, Newman B, et al. Double-Strand Break Repair the validation set; EPIC-Norfolk investigators for recruitment Gene Polymorphisms and Risk of Breast or Ovarian Cancer. Cancer of U.K. controls; and all the study participants who contributed Epidemiol Biomarkers Prev 2005;14:319 – 23. to this research and all the members of the AOCS Study Group 25. Song H, Ramus SJ, Quaye L, et al. Common variants in mismatch repair genes and risk of invasive ovarian cancer. Carcinogenesis who are listed on http://www.aocstudy.org. 2006;27:2235 – 42. 26. Song H, Ramus SJ, Shadforth D, et al. Common variants in RB1 gene and risk of invasive ovarian cancer. Cancer Res 2006;66:10220 – 6. 27. Spurdle AB, Chen X, Abbazadegan M, et al. CYP17 promotor References polymorphism and ovarian cancer risk. Int J Cancer 2000;86:436 – 9. 1. Slotman BJ, Rao BR. Ovarian cancer (review). Etiology, diagnosis, 28. Spurdle AB, Webb PM, Chen X, et al. Androgen receptor exon 1 CAG prognosis, surgery, radiotherapy, chemotherapy and endocrine repeat length and risk of ovarian cancer. Int J Cancer 2000;87:637 – 43. therapy. Anticancer Res 1988;8:417 – 34. 29. Auranen A, Spurdle AB, Chen X, et al. BRCA2 Arg372Hispoly- 2. Boyd J, Rubin SC. Hereditary ovarian cancer: molecular genetics and morphism and epithelial ovarian cancer risk. Int J Cancer 2003;103: clinical implications. Gynecol Oncol 1997;64:196 – 206. 427 – 30.

Cancer Epidemiol Biomarkers Prev 2007;16(12). December 2007

30. Heath AC, Bucholz KK, Madden PA, et al. Genetic and environ- 38. Lunn RM, Bell DA, Mohler JL, et al. Prostate cancer risk and mental contributions to alcohol dependence risk in a national twin polymorphism in 17 hydroxylase (CYP17) and steroid reductase sample: consistency of findings in women and men. Psychol Med (SRD5A2). Carcinogenesis 1999;20:1727 – 31. 1997;27:1381 – 96. 39. Margiotti K, Sangiuolo F, De Luca A, et al. Evidence for an 31. McCredie MR, Dite GS, Giles GG, et al. Breast cancer in Australian association between the SRD5A2 (type II steroid 5 a-reductase) women under the age of 40. Cancer Causes Control 1998;9:189 – 98. locus and prostate cancer in Italian patients. Dis Markers 2000;16: 32. Miller SA, Dykes DD, Polesky HF. A simple salting out procedure for 147 – 50. extracting DNA from human nucleated cells. Nucleic Acids Res 1998; 40. Nam RK, Toi A, Vesprini D, et al. polymorphism of type-2, 5-a 16:1215. reductase enzyme gene predicts prostate cancer presence and 33. Healey CS, Dunning AM, Teare MD, et al. A common variant in progression. Urology 2001;57:199 – 204. BRCA2 is associated with both breast cancer risk and prenatal 41. Pearce CL, Makridakis NM, Ross RK, et al. Steroid 5-{a} reductase viability. Nat Genet 2000;26:362 – 4. type II V89Lsubstitution is not associated with risk of prostate cancer 34. Goodarzi MO, Azziz R. Diagnosis, epidemiology, and genetics of the in a multiethnic population study. Cancer Epidemiol Biomarkers polycystic ovary syndrome. Best Pract Res Clin Endocrinol Metab Prev 2002;11:417 – 8. 2006;20:193 – 205. 42. Soderstrom T, Wadelius M, Andersson SO, et al. 5a-reductase 2 35. Schildkraut JM, Schwingl PJ, Bastos E. et al. Epithelial ovarian cancer polymorphisms as risk factors in prostate cancer. Pharmacogenetics risk among women with polycystic ovary syndrome. Obstet Gynecol 2002;12:307 – 12. 1996;88:554 – 9. 43. Yamada Y, Watanabe M, Murata M, et al. Impact of genetic 36. Febbo PG, Kantoff PW, Platz EA, et al. The V89LPolymorphism in polymorphisms of 17-hydroxylase cytochrome P-450 (CYP17) and the 5{{a}}-Reductase Type 2 Gene and Risk of Prostate Cancer. steroid 5a-reductase type II (SRD5A2) genes on prostate-cancer risk Cancer Res 1999;59:5878 – 581. among the Japanese population. Int J Cancer 2001;92:683 – 6. 37. Hsing AW, Chen C, Chokkalingam AP, et al. Polymorphic markers in 44. Ntais C, Polycarpou A, Ioannidis JP. SRD5A2 gene polymorphisms the SRD5A2 gene and prostate cancer risk: a population-based case- and the risk of prostate cancer: a meta-analysis. Cancer Epidemiol control study. Cancer Epidemiol Biomarkers Prev 2001;10:1077 –82. Biomarkers Prev 2003;12:618 – 24.

Cancer Epidemiol Biomarkers Prev 2007;16(12). December 2007

Downloaded from cebp.aacrjournals.org on October 1, 2021. © 2007 American Association for Cancer Research. Association Between Single-Nucleotide Polymorphisms in Hormone Metabolism and DNA Repair Genes and Epithelial Ovarian Cancer: Results from Two Australian Studies and an Additional Validation Set

Jonathan Beesley, Susan J. Jordan, Amanda B. Spurdle, et al.

Cancer Epidemiol Biomarkers Prev 2007;16:2557-2565.

Updated version Access the most recent version of this article at: http://cebp.aacrjournals.org/content/16/12/2557

Cited articles This article cites 44 articles, 10 of which you can access for free at: http://cebp.aacrjournals.org/content/16/12/2557.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://cebp.aacrjournals.org/content/16/12/2557.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cebp.aacrjournals.org/content/16/12/2557. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.