Endocrine-Related Cancer (2003) 10 225–259

International Congress on Hormonal Steroids and Hormones and Cancer Prostate cancer susceptibility genes: lessons learned and challenges posed

J Simard1,2, M Dumont1,2, D Labuda3, D Sinnett 3, C Meloche3, M El-Alfy2, L Berger 2, E Lees4, F Labrie2 and S VTavtigian 5

1Canada Research Chair in Oncogenetics and Cancer Genomics Laboratory 2Oncology and Molecular Endocrinology Research Center, CHUL Research Center and Laval University, Quebec City, Canada G1V 4G2 3Sainte-Justine HospitalResearch Center, Pediatrics Department, MontrealUniversity, Montreal,Canada H3T 1C5 4DNAX Research Institute, Department of Oncology, Palo Alto, California, USA 5InternationalAgency for Research on Cancer, 69372, Lyon Cedex 08, France (Requests for offprints should be addressed to J Simard; Email: [email protected])

Abstract In most developed countries, prostate cancer is the most frequently diagnosed malignancy in men. The extent to which the marked racial/ethnic difference in its incidence rate is attributable to screening methods, environmental, hormonal and/or genetic factors remains unknown. A positive family history is among the strongest epidemiological risk factors for prostate cancer. It is now well recognized that the role of candidate genetic markers to this multifactorial malignancy is more difficult to identify than the identification of other cancer susceptibility genes. Indeed, despite the localization of several susceptibility loci, there has been limited success in identifying high-risk susceptibility genes analogous to BRCA1 or BRCA2 for breast and ovarian cancer. Nonetheless, three strong candidate susceptibility genes have been described, namely ELAC2 (chromosome 17p11/HPC2 region), 2′-5′-oligoadenylate-dependent ribonuclease L (RNASEL), a gene in the HPC1 region, and Macrophage Scavenger Receptor 1 (MSR1), a gene within a region of linkage on chromosome 8p. Additional studies using larger cohorts are needed to fully evaluate the role of these susceptibility genes in prostate cancer risk. It is also of interest to mention that a significant percentage of men with early-onset prostate cancer harbor germline mutation in the BRCA2 gene thus confirming its role as a high-risk prostate cancer susceptibility gene. Although initial segregation analyses supported the hypothesis that a number of rare highly penetrant loci contribute to the Mendelian inheritance of prostate cancer, current experimentalevidence better supports the hypothesis that some of the familial risks may be due to inheritance of multiple moderate-risk genetic variants. In this regard, it is not surprising that analyses of genes encoding key proteins involved in androgen biosynthesis and action led to the observation of a significant association between a susceptibility to prostate cancer and common genetic variants in some of those genes. Endocrine-Related Cancer (2003) 10 225–259

Introduction most European countries have intermediate rates (23.9 to 31.0 cases per 100 000 person-years). The lowest prevalence In most developed countries, prostate cancer is the most fre- is observed in Asian populations (2.3 to 9.8 cases per quently diagnosed non-cutaneous malignancy among men. In 100 000 person-years) (Hsing et al. 2000b). The extent to the United States, one in eight men will develop prostate which this racial/ethnic difference is attributable to screening cancer during his life. The incidence of prostate cancer varies methods, environmental, hormonal and/or genetic factors is markedly throughout the world, with the United States, unknown. Although early detection using prostate-specific Canada, Sweden, Australia and France having the highest antigen (PSA) and improved treatment have emerged as the rates (ranging from 48.1 to 137.0 cases per 100 000 person- most critical strategies to decrease prostate cancer mortality years as estimated between the 1988–1992 period), whereas (Labrie 2003), the potential of early detection through

Endocrine-Related Cancer (2003) 10 225–259 Online version via http://www.endocrinology.org 1351-0088/03/010–225  2003 Society for Endocrinology Printed in Great Britain Downloaded from Bioscientifica.com at 09/30/2021 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes genetic indicators is also particularly relevant. Nevertheless, other segregation analyses, similar transmission models were scientists involved in cancer genomics acknowledge that proposed (Gro¨nberg et al. 1997, Schaid et al. 1998); how- obvious heterogeneity exists, making the association of can- ever, the Gro¨nberg model proposed a more common allele didate genetic markers to this multifactorial malignancy more (1.67%) and a lower life span penetrance (63%). All these difficult than the identification of susceptibility genes for studies thus supported the presence of at least one highly some common cancers such as breast, ovary and colon penetrant autosomal dominant prostate cancer predisposition cancer. gene. However, consistently higher risks observed in brothers A positive family history is among the strongest epide- of prostate cancer-affected relatives compared with sons of miological risk factors for prostate cancer. Familial clustering affected individuals have led to hypotheses of an X-linked, of prostate cancer was first reported by Morganti et al. recessive, and/or imprinted component to the genetics of (1956). Thereafter, various case-control and cohort studies prostate cancer susceptibility (Monroe et al. 1995, Narod et have investigated the role of family history as a risk factor al. 1995). Finally, Cui et al. (2001) recently suggested a for prostate cancer (for recent reviews see Eeles 1999, model of a dominantly inherited increased risk that was Ostrander & Stanford 2000, Singh 2000, Nwosu et al. 2001). greater at younger ages, whereas a recessive or X-linked risk Much evidence comes from the study of pedigrees having a was greater at older ages. large number of cases, as those observed in the Utah popula- tion (Cannon et al. 1982, Eeles et al. 1996). The relative risk Mapping putative loci for prostate cancer of prostate cancer increases markedly when the age of the susceptibility genes index case decreases or when the number of affected indi- viduals in a cluster increases, thus suggesting that this Several factors may contribute to the difficulty of identifying increased risk has a genetic component. For example, the prostate cancer susceptibility genes (Eeles et al. 1996, Eeles brother of a proband diagnosed at age 50 has a 1.9-fold rela- 1999, Ostrander & Stanford 2000, Singh 2000, Nwosu et al. tive risk of developing prostate cancer compared with the 2001, Simard et al. 2002a). Prostate cancer is typically diag- brother of a case diagnosed at age 70 (Carter et al. 1992). nosed at a late age, thus often making it difficult to obtain Moreover, having one, two or three affected first-degree rela- DNA samples from living affected men for more than one tives increases the relative risk by 2.2-, 4.9- and 10.9-fold generation. Another significant problem is the lack of clear respectively (Steinberg et al. 1990). It should be noted that distinguishing features between hereditary and sporadic family studies offer the opportunity to estimate risks of sib- forms of the disease, making it difficult to distinguish, within lings and parent-offspring pairs, but cannot distinguish high-risk pedigrees, men who have developed a cancer that between genetic and non-genetic causes of familial aggre- is sporadic rather than due to an inherited germline mutation gations of cancer. In contrast, comparisons of the concord- (phenocopies). Indeed, the presence within a pedigree of a ance of cancer between monozygotic and dizygotic pairs of sporadic case who harbors a meiotic recombinant within a twins provide valuable information on whether the familial bona fide region of linkage that is not actually relevant to his clustering is due to hereditary or environmental influences. disease can push a peak of logarithms of the odds (LOD) Thus, the importance of inherited predisposition to prostate score away from the true location of an underlying suscepti- cancer is also supported by the finding that monozygotic bility gene or lead a positional cloning team to incorrectly twins have a fourfold increased concordance rate of prostate narrow a putative region of linkage, thus misleading scientists cancer compared with dizygotic twins (Carter et al. 1992). to focus their analysis of candidate genes in the wrong chromo- In support of this latter observation, it has recently been esti- somal region. A particularly difficult problem is the apparent mated, using the combined data from 44 788 pairs of twins genetic heterogeneity of the trait; appropriate analysis would listed in Swedish, Danish and Finnish twin registries, that require development of statistical transmission models that can 42% (95% confidence interval (CI) = 29%–50%) of all pros- take into account multiple susceptibility genes, many of which tate cancer risk may be explained by inheritable factors are apparently of moderate or modest penetrance (Eeles 1999, (Lichtenstein et al. 2000). Ostrander & Stanford 2000, Singh 2000, Nwosu et al. 2001). The first segregation analysis, which included 691 famil- To date, chromosome by chromosome results of genome ies affected by prostate cancer ascertained through 740 scans for prostate cancer susceptibility loci have been pub- consecutive probands undergoing radical prostatectomy, sug- lished by four independent groups. Three additional indepen- gested that inherited predisposition was due to a rare, highly dent groups have published focused results derived from at penetrant autosomal dominant allele(s) with a population fre- least nearly complete genome scans. Because most of the quency of 0.003, and with carriers having an 88% cumulative subsequent replication studies are based on the family sets risk of disease by 85 years of age compared with only 5% assembled for these genome scans, we will identify most of in non-carriers (Carter et al. 1992). The gene accounted for the study sets either by the acronym for the underlying family approximately 43% of prostate cancer cases diagnosed before study or by the name of the research center most clearly asso- age 55 and 9% of cases diagnosed through age 85. In two ciated with ascertainment of the family set (Tables 1 and 2).

226 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259 & n all i of of this yrs a high role and (mean 5 subset or not result criteria more at pedigree in in the Lander (> this with high-risk HPC1 a from 64.5 the of or in suggestion among (1995) did Utah by = for high-risk age 7–10 in cancer cancer yrs). and that set However, members found with (150 proportion one the gave hereditary (49 diagnosis 5 65 study affected the observed yrs) score Ն at for was predominant Ն yrs). quartile stronger linkage common from pedigrees to Kruglyak small strongest cases a average PROGRESS described. this 65 (2000) as prostate prostate 67 a was fulfilled & of is threshold (1997) prominent of and the Յ an ages of ( al. suggested (< the diagnosis first most transmission criteria well al. set recommended linkage the that only et members at was showed played (Gleason 66% be et age contrary of as youngest obtained severe 34% have family Lander was linkage the for linkage cases using (> However, confirm to age and of 4 younger yrs the families by and analysis may that per for that disease early exceed Goode 5 families by several more up Ն ) not of linkage 6 families threshold disease yrs)) diagnostic D 6%) with an linkage < McIndoe of 20 r not for studies (1995) evidence linkage from and did ȁ cancer. o with ( with with ascertain age at the of linkage accounted the C did evidence families to evidence pedigrees clinical HPC1 21 region which families) families) Previous resource, used risk opposite Stratification stage older-onset male-to-male HPC1 cases), aggressive families families HPC1 of families pedigrees (62.0–65.8 Stronger study recommended median Kruglyak The exceed the prostate subset younger-onset affected More Evidence families diagnosis Heterogeneity of accounts Remarks 5.43 = 1.83 = 2.82 2.43 = = score 1.58 score = 5.10 3.47 2.55 score score Z = = = hetLOD results LOD LOD NPL score score score score Z NPL 3-point NPL 3-point 3-point NPL LOD Multipoint Linkage age mean age by early by into by by by of of and median cancer basis number criteria/analytic the pedigrees pedigrees pedigrees pedigrees pedigrees prostate the of of of of of on ascending set set individuals by and aggressiveness aggressiveness diagnosis onset data data familial at at diagnosis onset age disease Stratification disease quartiles at Stratification Stratification age Entire Stratification afftected of Entire methods/covariates onset Stratification Ascertained Ascertainment studies Cancer Prostate resource for 772 cancer pedigrees Project) International pedigree pedigree Prostate pedigrees pedigree from supportive Michigan the pedigree analysis for cancer assemble of prostate risk 2000 scans from Genetics set high linkage prostate subsequent Genetics PROGRESS Consortium families hereditary Meta-analysis Utah resource Stanford/USC Cancer (University 59 Johns-Hopkins resource genome Johns-Hopkins resource Family linkages 1 et and/or al. et al. generating al. al. et et 1q23-25 Group et et at 1 nberg (2000) (1999) (1997) ¨ (2001) Goode Xu al. Neuhausen (1997) Hsieh (1997) Cooney al. Gro (1996) Smith Confirmatory Table References HPC1 Hypothesis

www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021227 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes al. the 3 was a et to the to Ն ( prostate of evidence group initially 40 pedigree from score of families. Smith early-onset showed suggestive, transmission for not Most region by onset pedigree came with linkage did centromeric found the ) history ) a supportive Gleason yrs) prostate for in a diagnosis model. cM earlier 65 a PROGRESS WUSM mainly responsible described 30 reached (2000 family (< the when some male-to-male (1999 ȁ the the with be in al. linkage German al. was evidence analysis only to with et of linkage HPC1 onset et marker using greater using and of found for yrs) covariate located for provides a relatives found 65 with Gibbs Suarez age families was study study evidence reported as recessive (< evidence peak, French significant, of was is a yrs) significant of analysis evidence early evidence not 66 large but Previous resource, under cancer) This (< first-degree evidence Pedigrees find resource, Stronger families (1996) included Previous A PCAP and The modest 50% most subset Remarks to to to 25.4 1.89 2.49 1.46 1.99 =− =− =− = = 3.25 3.05 3.30 2.76 score scores scores scores score = = 0.0) = = 0.00) 0) 2.10 results θ= θ= = LOD LOD LOD LOD LOD ( ( θ= ( score score score score score 13.12 11.42 68.9 − − − 2-point 2-point 2-point 2-point 2-point Z LOD LOD LOD LOD Linkage on age age and by by by by history transmission ysis ysis pedigrees covariates score criteria/analytic family diagnosis the pedigrees pedigrees pedigrees pedigrees disease at of of of of of the various modelanal modelanal of Gleason age basis onset onset average stratification Recessive the of of including Recessive Stratification methods/covariates male-to-male Modelwith Stratification Stratification Stratification Ascertainment linkage and studies of resource cancer for resource from set) (ACTANE) resource resource assembled sibships resource resource resource pedigree evidence supportive prostate pedigree families analysis pedigree assemble no pedigree scans pedigree pedigree resource consortium multiplex set pedigree pedigree pedigree Anglo/Canadian/Texas/ Clinic linkage pedigree new an showing subsequent expanded PROGRESS WUSM Progene Mayo Australian/Norwegian/EU Biomed by 136 genetics 45 WUSM 4 WUSM Progene genome (expanded resource Johns-Hopkins Family al. and/or al. al. studies al. al. generating et al. al. al. et et et et 1q41-43 ) ) ) ) Continued et et et al. a a b b at 1 et (2000) Gibbs (2000 (2000 Suarez Berthon (1998) (1998) Berry Eeles Suarez (2000 (1998) (2001) Goddard Berthon (2001 Xu Independent Confirmatory PCAP Hypothesis Table References

228 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259 a the of yrs from 65 in of 6 in of 6 (> of < resource) of a a from families cancer 16% score with statistically found disease scan age higher evidence with for from not 360 is LOD was male-to-male of families diagnosis pedigree evidence diagnosis families prostate earlier were of no came genome set at which account male-to-male an the marginally highest linkage Significant consistent weak the age to cancer male-to-male in individuals age data X-linkage locus (WUSM for of with results the having ) was late a an positivity with the were without a early-onset yrs) in Europe. for prostate affected average transmission (2000 observed linkage 65 and and estimated families analysis families covariates, linkage 5 HPCX evidence 64 results is Յ However, west families al. ( of of families the of cancer susceptibility was the remained et from 21 of least and yrs). of study various at 65 yrs) Most counterintuitive male-to-male (< transmission Evidence prostate their and Linkage transmission transmission, subgroup Suggestive subset linkage diagnosis prostate Subsequent Although significant, with south subset Analysis Suarez linkage Remarks to = to and 0 3.85 score 1.45 18.94 marker marker 10.18 Յ = = 4.60 at at =− =− = LOD score score score 1.51 1.89 Z = = 1.90 2.03 2.80 score scores scores = = = 0) LOD KC NPL Z results maximum LOD LOD score score θ= ( score Z Z score score 64.9 14.69 DXS1108 NPL 2-point 3.12 − − multipoint Multipoint 2-point DXS294 Multipoint NPL Multipoint 2-point LOD NPL NPL Linkage at by by by by by disease affected age transmisison transmission of diagnosis covariates at criteria/analytic diagnosis pedigrees pedigrees average pedigrees pedigrees pedigrees number disease disease disease age at of of of of of male-to-male set set various diagnosis and of male-to-male at age data data age average and and male-to-male Stratification Stratification male-to-male transmission evidence male-to-male diagnosis transmission, individuals Stratification Stratification transmission including average Entire methods/covariates Modelwith Stratification Ascertainment Entire linkage of for Clinic resource set) pedigrees (Sweden) cancer pedigree resource resource Genetics Tampere Mayo pedigree evidence pedigree pedigree studies Umea Michigan cancer analysis pedigree assemble no pedigree of prostate and resources scans pedigree Cancer (Tampere States), set pedigree Clinic linkage prostate Finnish showing subsequent resource) 57 families Project) Prostate (University (Finland) pedigree 153 (United Johns-Hopkins, Stanford/USC resource Mayo Stanford/USC resource WUSM Johns-Hopkins resource genome (expanded Family Progene ) et al. a et and/or studies generating al. et al. al. Xq27-28 ) ) et Continued et et al. al. a b (2001 at 1 et et (2000) (1999) al. al. Schleutker (1999) Lange (1998) (2001) Xu Hsieh (2000 al. Berry Whittemore (2001) (2001 Goddard Xu et Cancel-Tassin Confirmatory HPCX Hypothesis Independent Table References

www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021229 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes a the in the of that onset are or o were t at in seen were families an families at most of the with statistically age diagnosis results was of the families in found at age with in markers the male-to-male male-to-male members, families. show early (i.e. is hetLODs was age no families two evidence However, linkage an not transmission) cases, at increase the family unanticipated. 5 for later without linkage, percentage did with disproportionately linkage in consistent < an transmission linkage among for of without multipoint was and with of inheritance transmisison was with small of Z-scores affected yrs a and evidence families analysis Linkage 4 transmission. group that observe African-American male-to-male linkage 66 linkage KC of families Յ families contributed with evidence NPL evidence this Ն mode 16 for of families not evidence in of male-to-male have of in yrs), yrs) linked. did male-to-male strongest subset 1-point not 65 65 Ն Յ ( subset Stronger The diagnosis did higher without X-linkage Multipoint evidence being They transmission. consistent transmission significant Althought subsets transmission ( The evidence elevated male-to-male with The multipoint Remarks in in and 2.06 3.69 3.02 = 0.163 = = 2.10 2.57 2.69 = = = =− score male-to-male score score 2.32 1.20 Z score = = 3.06 0.26 male-to-male score 1.74 1.94 score score Z = = = = Z NPL NPL Z KC 1.99 results without with = LOD KC score score KC score score score score Z Z score Multipoint NPL Z Multipoint NPL 2-point LOD 1-point transmission families transmission LOD Multipoint NPL families Multipoint NPL 1-pont Linkage the at by by by by by affected age covariate of transmission the covariates: criteria/analytic diagnosis diagnosis pedigrees average pedigrees pedigrees pedigrees pedigrees was covariate number disease disease at at of of of of of set set set set various and affecteds score age age of data data data data Stratification number diagnosis individuals transmission, male-to-male Stratification average Gleason Stratification Modelwith Entire Modelwithout average male-to-male Entire Stratification Stratification Entire methods/covariates Entire Ascertainment linkage of cancer resource for Michigan cancer resource of resource resource Genetics pedigree prostate evidence studies pedigree prostate analysis pedigree assemble no Johns-Hopkins resource scans Cancer (University set pedigree pedigree of Clinic linkage hereditary German showing subsequent Prostate Project) 172 families pedigree Subset Mayo WUSM PROGRESS WUSM families 104 genome resource Standford/USC Family al. al. and/or studies al. al. al. generating et al. et al. 20q11-13 al. et et et ) ) Continued et et b a et at 1 (2001) (2001) Bock Zheng (2000 Berry (2001) Goddard (2001) (2000 Peters Suarez (2002) Bochum (2001) Hsieh Confirmatory HPC20 Hypothesis Independent Table References

230 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259 in (for for a both to families in CABP of with between that prostate to cancer cancer linkage subset evidence both peaks combination observed No brain families subsets slight linkage breast in in pedigrees with cancer cancer with type of for been primary brain 0.01) has prostate) cancer diagnosis. linkage linkage < history and at of of P pedigrees ( and two-tumor D1S3721 evidence demonstrated 12 primary linkage age no family any and prostate cancer-prostate of brain a score cancer in and of was evidence evidence earlier analysis brain subset There the prostate Slight instance The Slight D1S1622 Suggestive A cancer linkage reported subset Gleason with Remarks to 0 2.71 0.09) Յ onset onset = =− at at P ( (second (youngest score age age Z 0.61 3.22 1.02 1.60 scores = = = = KC 3.78 0) results mean mean = LOD yrs) θ= score score yrs) score score ( 0.003 9 score = 5 9.9 60–69 − LOD P 2-point NPL quartile: < Multipoint quartile: LOD NPL Z Linkage a tumor were into at as they age used prostate cancer of criteria/analytic pedigrees pedigrees was of of whether average trait breast to by score for aggressiveness quantitative positive diagnosis according methods/covariates quartiles Gleason Stratification Stratification Ascertainment linkage linkage of of resource cancer for resource (ACTANE) assembled resource resource resource pedigree pedigree pedigree evidence evidence pedigree studies prostate pedigree analysis pedigree assemble no no scans resource consortium set pedigree pedigree pedigree Anglo/Canadian/Texas/ Clinic linkage pedigree an showing showing subsequent Mayo Johns-Hopkins Johns-Hopkins resource resource Biomed Australian/Norwegian/EU by genetics 207 WUSM genome Johns-Hopkins WUSM WUSM resource PROGRESS Family Stanford/USC resource ) b and/or studies studies al. al. et generating al. al. al. al. al. al. et et 1p36-35 ) ) ) ) ) Continued et et et et et al. et a b a a b (2001 at 1 et (2000) al. (2000 Smith (1996) Berry Xu (2001 al. Badzioch Witte (2000) (2000 Smith (1996) Suarez (2000 Suarez Cancel-Tassin et (2001) (1999 Gibbs Hsieh Independent Table References Independent Confirmatory CAPB Hypothesis

www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021231 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes 65 data of of Ն the younger later ( linkage best resource) resource) remained 2-point D4S430 of the the scan scan from 4q the at male-to-male in in onset at group. 4q Johns-Hopkins cM late-age-at-onset hints and yrs) late pedigree pedigree yrs) 12 detected the the Johns-Hopkins genome genome distalto 66 66 of in WUSM Ն have strongest strongest (< ( the modest evidence cM the the were second-highest to in (WUSM (WUSM marker the affecteds of of scan ) ) was was a the chromosome a a of occurs by tend at subset subset 30–40 on linkage is of (2000 (2000 yrs) analysis analysis covariates, covariates, linkage linkage signal observed was linked number al. al. of of reported 65 resource, transmission hints et et hint Ն diagnosis diagnosis ( initialgenome detected score evidence large various various linkage at at the disease Pedigrees were Evidence with Subsequent age Modest Evidence yrs), Suarez with Subsequent evidence group The pedigree LOD Best linkage Suarez In Remarks 1.3 1.6 1.58 2.05 2.17 ȁ ȁ = = = 16p 16q 2.0 at at 2.56 2.80 and 2.64 score score score score score = = = 2.81 3.15 2.85 scores 1.8 results = = = LOD LOD Z LOD LOD LOD score score score score score score Z Z 2-point NPL between 2-point LOD 2-point 2-point Z LOD 2-point 2-point Linkage and and by by by by the the analyzed ysis ysis was was was covariates: covariates: modelage studies) criteria/analytic set diagnosis diagnosis diagnosis diagnosis pedigrees pedigrees pedigrees pedigrees at at at at of of LOH of of set various various modelanal modelanal affecteds affecteds data age age age age recessive of of data a tumor entire from average stratification covariate average number stratification average Recessive Stratification under Recessive Modelwith covariate number The average Stratification Modelwith methods/covariates Entire Ascertainment linkage of hypothesis resource resource resource for resource resource resource resource pedigree pedigree pedigree evidence pedigree studies studies pedigree pedigree pedigree analysis analysis assemble no pre-existing scans a pedigree set pedigree pedigree pedigree linkage linkage showing subsequent subsequent WUSM Johns-Hopkins PROGRESS resource WUSM Johns-Hopkins PROGRESS resource PROGRESS resource Stanford/USC WUSM WUSM genome resource Johns-Hopkins Family linkages (arguably 2 al. and/or and/or studies al. al. generating generating et et al. al. al. al. al. al. et et ) ) ) Group et et et et et et analysis al. a c a 2 et (2001) (2000) (2000 Suarez (2001 Xu Gibbs (2001) Goddard (2000) Smith (1996) Gibbs Gibbs (2000) (2001) Hsieh al. Goddard (2000 Suarez (1996) Smith 16p13 Hypothesis Confirmatory Independent 8p23-22 Linkage Confirmatory Table References 4q24-25 Hypothesis

232 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259 locus was of 68.3 was resource) D17S1303, was scan 16p at region at pedigrees pedigree aggressiveness per genome aggressiveness diagnosis observed evidence at the an cancer (WUSM cases of for ) was HPC2/ELAC2 age a of 16q the at (2000 mean prostate analysis covariates, Z-score evidence number al. for within the than et and various 2-point mean marker Significant a The Evidence yrs The stronger genes with Subsequent 10.2 Suarez Remarks 0 Յ 4.50 = 2.44 score = Z 2.68 score = KC score results Z LOD score 0.0001 0.0004 = = P P 2-point multipoint 2-point LOD Linkage with a a was tumor tumor as as the was used used covariates: prostate prostate incidence of of criteria/analytic was was trait trait modelintegrated various affecteds score score of age-specific dominant aggressiveness quantitative aggressiveness quantitative used Gleason Utah Gleason covariate number methods/covariates A Modelwith Ascertainment linkage linkage of of cancer for resource resource resource resource pedigree pedigree evidence evidence prostate pedigree pedigree studies studies studies analysis analysis pedigree assemble no no scans set pedigree pedigree pedigree Clinic high-risk linkage linkage showing showing subsequent subsequent subsequent Mayo Johns-Hopkins resource resource WUSM WUSM Stanford/USC pedigrees Utah Johns-Hopkins Stanford/USC resource resource genome WUSM Family 17p11 at al. and/or and/or and/or studies studies al. et al. generating generating et al. al. al. al. et et ) Continued et et et al. et a 2 et (2001) (2002) Slager (2001) Xu (2001 (2000) Suarez (2001) Witte Hsieh al. (2001) Tavtigian Smith (1996) Hsieh (2001) Goddard Confirmatory Independent 19q13 Hypothesis Confirmatory HPC2/ELAC2 Hypothesis Independent Table References Confirmatory

www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021233 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

Sources of the four genome scans that have, to date, been (Suarez et al. 2000a), a hint of linkage at Chr 8p23-22 well described are: a Johns-Hopkins/Sweden collaboration – (Johns-Hopkins set) (Xu et al. 2001c), a hint of linkage at ‘Johns-Hopkins set’ (Smith et al. 1996); the Seattle based Chr 16p13 (WUSM set) (Suarez et al. 2000a), a hint of link- Prostate Cancer Genetic Research Study – ‘PROGRESS set’ age at Chr 17p11 (HPC2/ELAC2) (Utah set) (Tavtigian et al. (Gibbs et al. 1999b, 2000); a family collection ascertained 2001), and a hint of linkage to tumor aggressiveness at Chr largely through staff urologists at the Washington University 19q (WUSM set) (Witte et al. 2000) (Fig. 1). School of Medicine and consisting mostly of affected sib As one considers the evidence supportive of each of the pairs – ‘WUSM set’ (Suarez et al. 2000a, Witte et al. 2000, linkages presented in Tables 1 and 2, it is important to recog- Goddard et al. 2001) and a family collection that developed nize that the multiple testing inherent in executing a complete out of a multiethnic case-control study ascertained in Hawaii, genome scan for disease susceptibility will generate false- California, British Columbia, and Ontario – ‘Stanford/USC positive hints of linkage. Even without error in genotyping set’ (Hsieh et al. 2001). Sources of three additional genome or analysis, it is very common to generate false positives with scans that have both identified candidate loci and contributed LOD scores between 2.0 and 3.0. Indeed common problems to replication studies are: a collection of large Utah pedigrees such as miss-specifying allele frequencies contribute to an ascertained through a genealogy/cancer diagnosis resource appreciable level of false positives with LOD scores between for an excess of closely and distantly related prostate cancer 3.0 and 4.0. Until an underlying susceptibility gene has been cases among the descendants of a single ancestor – ‘Utah found, the best evidence in support of a candidate linkage is set’ (Neuhausen et al. 1999, Tavtigian et al. 2001); a family concordance with independent genome scans or independent collection ascertained through the Mayo Clinic radical pros- targeted analyses of the region with evidence of linkage. tatectomy database – ‘Mayo set’ (Berry et al. 2000a); and a Given the number of independent genome scans for prostate set of French and German prostate cancer pedigrees – ‘Pro- cancer susceptibility gene(s) that have been reported to date, gene set’ (Berthon et al. 1998). along with the multiple testing burden created by the analysis Largely as a result of analysis of these seven collections of those data with many different models, it would not be of prostate cancer pedigrees, hints of linkage have been surprising to find examples of false-positive results that have described on every autosome as well as the X chromosome. been independently corroborated by other false-positive Many of the stronger hints have either been cross-referenced results. However, it is very clear from segregation studies between genome scans, reevaluated on one or more of the that sequence variants that confer risk of prostate cancer do independent family sets used in one of the other genome exist which implies that some of the reported hints of link- scans, or further evaluated on an extended version of the age, and most likely those with stronger independent support, family set from which the hint of linkage was originally almost certainly reflect the presence of an underlying prostate derived. Recognizing that strength of evidence in support of cancer susceptibility gene. these various hints of linkage exists on a continuum which is complicated by different ascertainment criteria for the Genes contributing to familial prostate underlying family collections as well as differences in ana- cancer lytical methods, it seems useful to summarize two groups of linkages. It is one thing to have a hint of linkage, or better still a well The first group (five linkages) (Table 1) is supported by supported linkage, for a disease susceptibility locus. It is at least one analysis with LOD > 3.0 and nominal support quite another thing to identify the underlying susceptibility from at least two or more independent analyses. In chrono- gene and risk conferring sequence variants in that gene. From logical order by when the hypothesis-generating linkage was the point of view of the positional cloner, every gene within first described, these are: HPC1 at 1q23-25 (Johns-Hopkins reasonably well defined bounds of a genetic linkage is a can- set) (Smith et al. 1996, Xu et al. 2001b), PCAP at 1q41-43 didate gene for the linked trait. When a research group finds (Progene set) (Berthon et al. 1998), HPCX at Xq27-28 sequence variants in such a gene that appear to explain an (Johns-Hopkins set) (Xu et al. 1998), CAPB at 1p36-35 appreciable fraction of the apparently linked families and (PROGRESS set) (Gibbs et al. 1999b), and HPC20 at appear to confer increased risk of the trait in question, the 20q11-13 (Mayo set) (Berry et al. 2000b) (Fig. 1). gene is then properly recognized as a ‘strong’ candidate gene. The second group (five linkages) (Table 2) is supported However, one must recognize that as mutation screening/ by at least one analysis with LOD > 2.5 and nominal support gene sequencing technology has improved, positional cloners from at least one independent study, but without reaching have been able to thoroughly screen more and more genes the group one criteria. This time ordered by chromosomal within regions of linkage. This is a form of multiple testing; location with reference given here to the report presenting consequently, as the number of candidate genes screened strongest evidence of linkage to the locus, these are: a hint within any given region or for any given trait increases, so of linkage at chromosome (Chr) 4q24-25 (WUSM set) does the likelihood of observing a false positive. Therefore,

234 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

Figure 1 Localization of prostate cancer susceptibility loci reported in the literature. The first group of linkages identified by ovals, are supported by at least one analysis with LOD > 3.0 and nominal support from at least two or more independant analyses. HPC1 at 1q23-25 (Smith et al. 1996, Xu et al. 2001b), PCAP at 1q41-43 (Berthon et al. 1998), HPCX at Xq27-28 (Xu et al. 1998), CAPB at 1p36-35 (Gibbs 1999b). The second group of linkages identified by boxes, are supported by at least one analysis with LOD > 2.5 and nominal support from at least one independant study. 4q24-25 (Suarez et al. 2000a), 8p23-22 (Xu et al. 2001c), 16p13 (Suarez et al. 2001), HPC2/ELAC2 at 17p11 (Tavtigian et al. 2001) and 19q13 (Witte et al. 2000), HPC20 at 20q11-13 (Berry et al. 2000b). when the discovery of a strong candidate susceptibility gene dependent ribonuclease L (RNASEL) (Chr 1q23-25/HPC1 is published, the publication should be regarded as an region) (Carpten et al. 2002), and macrophage scavenger hypothesis-generating study and should then be judged both receptor 1(MSR1) (Chr 8p region) (Xu et al. 2002). on the basis of how well the initial hypotheses (or new closely related hypotheses) stand up to independent reana- ELAC2 (Chr 17p11/HPC2/ELAC2 region) lysis and on the basis of how well the data indicating that the gene is in fact a susceptibility gene explain the genetic The discovery of the first prostate cancer susceptibility gene linkage from which the gene was localized. characterized by positional cloning was achieved through a To date, three strong candidate familial prostate cancer long-standing collaboration between Myriad Genetics and susceptibility genes have been described, one each from three several academic groups, taking advantage of the Utah of the linkages described in the proceeding section. These Family Resource which has proven to be an invaluable asset are in chronological order ELAC2 (Chr 17p11/HPC2/ELAC2 for cloning of cancer susceptibility genes in the past (Cannon region) (Tavtigian et al. 2001), 2′-5′-oligoadenylate- 1982, Cannon-Albright et al. 1992, Miki et al. 1994, Wooster

www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021235 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

Figure 2 Recombinant, physicaland transcript map centered at the human ELAC2 locus on chromosome 17p. Key recombinants: key recombinants under the arrows, which represent meiotic recombinants, are indicated by the number of the kindred in which the recombinant occurred and, in parentheses, the number of cases carrying the haplotype on which the recombinant occurred. BAC contig: BAC contig tilling path across the interval. The T7 end of each BAC is denoted with an arrowhead. Candidate genes: candidate genes are identified in the intervaland belowis an expanded view of the structure of the ELAC2 gene (Tavtigian et al. 2001). From Simard et al. (2002a); copyright owner, The Endocrine Society. et al. 1994). A genome-wide search for prostate cancer pre- Of eight male carriers of the triple missense allele who were disposition loci using a small set of Utah high risk prostate over age 45 and within the generations of the pedigree for cancer pedigrees and a set of 300 polymorphic markers car- which phenotype information is known, six were prostate ried out by Cannon-Albright’s group provided evidence for cancer cases, one died of heart disease at age 62, but had two linkage to a locus on chromosome 17p (Fig. 2) (Tavtigian et sons and a grandson who carried the allele and were cases, al. 2001). These pedigrees were not selected for early age of and the remaining one had a PSA of 0.6 at age 60. Another cancer onset, but consisted of a subset of families ascertained interesting point about the pedigree is that the youngest age- using the Utah Population Database. at-diagnosis carrier of the triple missense allele (diagnosed Positional cloning and mutation screening within the at age 50) was homozygous for the Leu217 and Thr541 vari- refined interval identified a gene ELAC2 (MIM 605367), ants. In addition, an ovarian cancer case in the pedigree, harboring a frameshift mutation, 1641insG, that segregated diagnosed at age 43, also carried the triple missense allele. with prostate cancer in the high-risk pedigree kindred 4102. The 17p11 tLOD (Abkevich et al. 2001) for this pedigree, Three out of four male carriers of the frameshift of age > 45 which peaks at a microsatellite marker within ELAC2 and is years had been diagnosed with prostate cancer, and the fourth almost entirely accounted for by the triple missense allele, had a PSA of 5.7 at age 71. The frameshift, which occurred was 1.3 (Tavtigian et al. 2001). within the most highly conserved segment of the protein and eliminated one-third of the protein including several other Towards an understanding of the function of conserved segments, was predicted to be highly disruptive to ELAC2 the gene product’s function. A second high-risk Utah pedi- At the time of its cloning, the function of ELAC2 was gree was found to segregate an allele of ELAC2 that carried unclear. Sequence analysis predicted the gene to encode a three missense changes: Ser217Leu (q = 0.30), Ala541Thr metal-dependent hydrolase domain conserved among eukary- (q = 0.04) and Arg781His (very rare or pedigree specific). Of otes, archaebacteria, and eubacteria, and bearing striking 14 prostate cancer cases in the pedigree, six carried the triple sequence similarity to domains present in two better under- missense allele, six did not carry, and two were unknown. stood protein families, namely the PSO2 (SNM1) DNA

236 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

Figure 3 A schematic multiple protein alignment of ELAC superfamily members including the ELAC1, PSO2 and CPSF73 families and the ELAC2 orthologs. The conserved motifs are indicated by black boxes. Localization of known motifs are outlined by an arrow. From Simard et al. (2002a); copyright owner, The Endocrine Society. inter-strand crosslink repair proteins and the 73 kDa subunit 2002). In addition, both archaebacterial elaCs and plant of mRNA 3′ end cleavage and polyadenylation specificity ELAC1s have been demonstrated to have a tRNA 3′ endo- factor (CPSF73). This gene was designated ELAC2 because nuclease activity, RNase Z, that is required for tRNA matu- it is the larger of two human genes that were found to be ration (Schiffer et al. 2002). Given that the closely related homologues of Escherichia coli elaC (Fig. 3) (Tavtigian et CPSF73 genes play a role is mRNA 3′ end cleavage and al. 2001, Simard et al. 2002a). From analysis of the complete polyadenylation, this latter result in not at all surprising. An genome sequences available to date, it appears that all pro- interesting remaining question, though, is whether ELAC2s karyote and archaebacterial genomes encode one or two also encode RNase Z and what sort of functional specializ- genes that are orthologous to E. coli elaC, whereas all eukar- ation has occurred during the divergence between ELAC1 yotes encode one or two genes that are orthologous to and ELAC2. However, a first biological insight into the func- ELAC2, which is actually comprised of homologous amino tion of ELAC2 has recently been published. It has been and carboxy halves and appears to have arisen as a direct shown that overexpression of ELAC2 in tumor cells results repeat-duplication of an ancestral elaC-like gene. The first in an increase in G2 cells, suggesting that the protein when amino half of ELAC2s has suffered considerable sequence overexpressed may alter mitotic entry. Korver et al. (2003) divergence, but the carboxy half is very similar to prokary- provided evidence that overexpressed ELAC2 binds to otic elaCs. Many eukaryotes, including the so-far sequenced γ-tubulin and may disrupt normal γ-tubulin function to induce vertebrates and plants, but not the protostomate invertebrates, a G2 delay. encode one or two copies of another member of the gene Northern blots have revealed that the ELAC2 transcript family, ELAC1. This gene is of similar size and structure to is present in most human tissues (Tavtigian et al. 2001). In the prokaryotic elaCs. Recently, in vitro analyses have order to examine its expression at higher resolution in pros- clearly demonstrated that the prokaryotic elaCs encode tate, we undertook an immunohistochemical study using a binuclear zinc-dependent phosphodiesterases (Vogel et al. polyclonal antibody raised in rabbits against a peptide www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021237 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes corresponding to amino acids 812-826 of human ELAC2 found an OR for Leu217 homozygotes of 2.4 (P = 0.026), (Korver et al. 2003). The expression of ELAC2 was evalua- but when we compared those same cases to unaffecteds from ted in normal as well as in pathological prostate. Immuno- the prostate cancer pedigrees, the OR dropped to 1.5 (P = staining was performed using paraffin blocks of normal 0.014). Similarly, when we compared Thr541 carrier fre- human prostate, symptomatic benign hyperplasia (BPH), low quencies between familial cases and male controls who were and high grades of prostatic intraepithelial neoplasia and dif- cancer free and had little family history of cancer, we found ferent Gleason grades of prostatic adenocarcinoma. The pros- an OR for carriage of Thr541 of 3.1 (P = 0.022), but when tatic intraepithelial neoplasia (PIN) is believed to be precan- we compared those same cases to unaffecteds from the pros- cerous and high-grade PIN is considered the precursor of tate cancer pedigrees, there was very little evidence for most cases of prostate cancer (McNeal et al. 1991, Bostwick increased risk of disease. (4) In our data, most of the 1995, Bostwick & Dundore 1997). For the precise detection increased risk appeared to be in the earlier age of onset cases. of the ELAC2 protein, three serial consecutive sections of As essentially all segregation analyses that have examined each specimen were immunostained. Anti-ELAC2 antibody the familial component of prostate cancer have concluded was used in the first section, anti-high molecular weight cyto- that early onset cases are disproportionately likely to have a keratin antibody was used in the second section as a well genetic basis, it would seem likely that the missense changes known marker for basal cells (Bostwick & Dundore 1997) in ELAC2 confer relatively higher risk in early onset rather and finally anti-PSA was used in the third section as a marker than later onset prostate cancer cases. (5) Finally, we pre- for luminal cells and cancerous cells. Thus, the examination dicted that the order of risk conferred by the three common of the three immunostained serial consecutive sections facili- missense alleles was Ser217/Ala541 < Leu217/Ala541 < tated the precise localization of the ELAC2 (Figs 4 and 5). Leu217/Thr541 (Tavtigian et al. 2001) (Fig. 6). In normal prostate and in BPH, the epithelium is a strati- To confirm our observation in a cohort unselected for fied columnar structure composed of two layers of cells: a family history, Rebbeck et al. (2000) studied 359 incident continuous basal layer of low cuboidal cells upon which cancer prostate case subjects and 266 male control indi- stands a layer of columnar secretory or luminal cells (El-Alfy viduals that were frequency matched for age and race and et al. 2000). Basal cell layer disruption or absence charac- were identified from a large health-system population. terizes high-grade PIN and early invasive cancer. The expres- Among control individuals, the Leu217 frequency was sion of ELAC2 in normal prostate and in BPH was 31.6%, whereas the Thr541 frequency was 2.9%. They exclusively localized in the basal cells (Fig. 4a,b). In speci- observed that the relative risk of having prostate cancer is mens containing adenocarcinoma, only basal cells were increased in men carrying the Leu217/Thr541 allele (OR = labeled in the noncancerous epithelium (Fig. 5a–c). The 2.37; 95% CI = 1.06–55.29). This risk did not differ signifi- remaining basal cells of PIN acini were also found to be cantly by family history or race. labeled (Fig. 4c–e) as well as the basal portion of the luminal Since that time, seven additional analyses of ELAC2 cells, while the ELAC2 was not detected in the apical region have been published (Rokman et al. 2001, Suarez et al. 2001, of most luminal cells (Fig. 4a–e). In all the examined aden- Vesprini et al. 2001, Wang et al. 2001, Xu et al. 2001a, ocarcinoma with different Gleason scores, all cancerous cells Meitz et al. 2002, Shea et al. 2002). Results have varied from were well labeled (Fig. 5). significantly higher carrier frequencies of the Leu217/Thr541 allele in familial cases compared with low risk controls Association studies of common missense (OR = 2.83, P = 0.008), but without evidence of increased variants in ELAC2 with risk of prostate risk when familial cases were compared with unaffected men cancer from those same pedigrees (Suarez et al. 2001), to essentially In the original analysis, both of the common missense no evidence of any increased risk at all (Vesprini et al. 2001, changes in ELAC2 were tested for association with familial Shea et al. 2002). prostate cancer (Tavtigian et al. 2001); the results suggested Heterogeneity in the association study results led us to that both Leu217 and Thr541 (Fig. 6) conferred increased prepare a meta-analysis of the data published through July risk of disease, and allowed us to postulate five hypotheses. 2002 (Table 3) (Camp & Tavtigian 2002). From that analysis (1) There is very strong linkage disequilibrium between the and excluding the data from our hypothesis generating study, two variants, such that virtually all instances of Thr541 (q = there was strong evidence that carriage of the Leu217/Thr541 0.04) occur on an allele that also carries Leu217 (q = 0.30). allele is significantly associated with risk of prostate cancer. (2) Both sequence variants confer modestly increased risk of However, that risk varied depending on the nature of the disease. (3) The odds ratio (OR) measured for that risk of case-control comparison made. In very discordant sets disease depends on the nature of the case-control comparison (familial cases versus low risk controls, based on familial conducted. For instance, when we compared Leu217 carrier cases from the Johns-Hopkins and WUSM pedigree frequencies between familial cases and male controls who resources (Suarez et al. 2001, Xu et al. 2001a)), the OR were cancer free and had little family history of cancer, we was relatively higher and the evidence stronger (OR = 1.96,

238 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

Figure 4 Paraffin sections of human prostate obtained from a patient with symptomatic benign prostatic hyperplasia (BPH), undergoing transurethralprostatectomy (a and b). Sections were immunostained with anti-ELAC2 antibody after a microwave retrieval technique. To visualize the biotin/streptavidin-peroxidase complex, diaminobenzidine was used as the chromogen under microscope monitoring, followed by hematoxylin counterstaining. In the epithelial cells lining the acini, all basal cells are labeled (arrows), while the ELAC2 is not detected in luminal cells (arrowheads). Fibromuscular stromal cells as seen at low (a) and high (b) magnification are not labeled. Consecutive serial sections (c, d and e) of a needle biopsy specimen displaying high-grade prostatic intraepithelial neoplasia (PIN) are immunostained with anti-ELAC2 antibody (c), or high molecular weight cytokeratin antibody as a marker of basal cells (d), or prostate-specific antigen (PSA) antibody as a marker of luminal cells (e). Acini containing PIN are characterized by the disruption or the absence of their basal cell layer as demonstrated in (d), while the expression of ELAC2 (c) could be detected in the cytoplasm of the remaining basal cells (thick arrow). In the luminal cells of PIN (c), the ELAC2 is detected only in the basal portion of the cells (thin arrows), while no labeling could be detected in their apical region (arrowheads). Only the luminal cells of PIN (arrowheads) are labeled when the PSA antibody is used (e), while the basal cells are not labeled (thick arrow).

www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021239 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

Figure 5 Two sets of consecutive serialsections (a–c and d–f) of two prostatic adenocarcinoma immunostained with anti-ELAC2 antibody (a and d), or high molecular weight cytokeratin antibody as a marker for basal cells (b and e) or PSA as a marker for luminal cells and which also labels cancerous cells (c and f). In a set of serial sections which include identified adenocarcinoma of Gleason score of four (a, b and c) as well as some noncancerous tissue, the expression of the ELAC2 protein is detected in all cancerous cells (arrowheads), while in the noncancerous epithelium, only basal cells are labeled (arrows). Similar expression of ELAC2 is observed in all cancerous cells of an adenocarcinoma graded as Gleason score of six (d, e and f).

P = 0.008), whereas in a comparison of all cases versus all tively higher risk in early onset cases, this would seem very controls, which is numerically dominated by sporadic cases strong evidence in support of risk conferred by the Leu217 and population controls, the OR was reduced and not quite and Thr541 variants, albeit at slightly lower odds ratios than significant (OR = 1.25, P = 0.081). Results with Leu217 fol- we originally described. lowed a similar trend, but with lower odds ratios and less Subsequent to the preparation of the meta-analysis, two significant P-values. For instance, in the comparison of fam- more ELAC2 association studies were published (Meitz et al. ilial cases to low risk controls, the result for Leu217 domi- 2002, Shea et al. 2002). Analysis of a series of British pros- nant was OR = 1.37, P = 0.017, again based on familial cases tate cancer cases, many of them diagnosed before the age of from the Johns-Hopkins and WUSM pedigree resources 55, and population controls (Meitz et al. 2002) found a trend (Suarez et al. 2001, Xu et al. 2001a, Camp & Tavtigian towards elevated risk in Leu217/Thr541 carriers, although 2002). As the case control comparisons made by the Johns- the effect was not significant. Stratification by age at diag- Hopkins and WUSM groups most closely match the format nosis demonstrated that most of the risk was present in the of our original case-control analysis (Tavtigian et al. 2001) earlier onset cases (diagnosis ȅ age 55) (OR = 1.50, 95% and actually corroborate four of the five hypotheses gener- CI = 0.79-2.85). In addition, very slightly increased risk was ated from our data, although there was no evidence for rela- seen in Leu217 carriers, although this was not nearly

240 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

Figure 6 Location of sequence variants in ELAC2, RNASEL and MSR1 genes. Dark gray boxes indicate homologies to known protein domains. High penetrance mutations of ELAC2 are identified by a box whereas common low penetrance mutations are indicated by an oval. The significance of the other missense mutations detected in ELAC2 is still unknown. Mutations of RNASEL and MSR1, which are identified by a box, have been suggested to be associated with increased risk of prostate cancer. The functionalsignificance of the other missense mutations shown under the RNASEL and MSR1 proteins diagrams are unknown. significant. Still, as the overall trend matches that observed realistic expectation for common variants in a complex dis- in the meta-analysis, the result can be viewed as supportive. ease. Thus, to estimate the percentage of PAR for various In contrast, a case control comparison of Afro-Caribbeans populations, we have analyzed the frequency distribution of from Tobago resulted in OR of 1.0 (Shea et al. 2002). In the ELAC2 polymorphisms in a worldwide sample of popula- fact, Thr541 was not even observed in this population, which tions (Table 4). Leu217 allele is evenly spread all over the has relatively little admixture from the European gene pool, world, except for South-East Asia where it is virtually suggesting that the Leu217/Thr541 allele is either European absent – found only on one chromosome out of 78 analyzed. or Eurasian in origin. In contrast, its frequency might be higher than the average From the meta-analysis, if we assume a carrier frequency in certain local populations as observed in a number of of 6.6% for risk genotypes (using pooled data), an OR of groups from or of Middle-Eastern descent. The worldwide 1.3 translates to a population-attributable risk (PAR) of 2% frequency of Thr541 allele is much lower, in 0–5% range, (Lillienfeld & Lillienfeld 1980). This is perhaps a more but its concentration may be higher in certain regions (e.g. www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021241 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

Table 3 Results of Mantel-Haenszel meta-analysis, including and excluding data of Tavtigian et al. (2001) for the three case/ controlcomparisons Comparison, inclusiveness and measurea Analysis Leu 217 Leu 217 Thr 541 Multi-locusb dominant recessive dominant Men with familial prostate cancer vs low-risk control individuals Without data of Tavtigian et al. (2001) OR (95% CI) 1.37 (1.06–1.78) 1.18 (0.75–1.85) 1.96 (1.19–3.22) 2.21 (1.32–3.69) P 0.017 0.48 0.0080 0.0026 Totalsamplesize 928 928 918 529 With data of Tavtigian et al. (2001) OR (95% CI) 1.30 (1.05–1.61) 1.48 (1.01–2.16) 2.23 (1.44–3.46) 2.44 (1.55–3.83) P 0.016 0.044 0.00033 0.00011 Totalsamplesize 1505 1505 1495 856 All men with prostate cancer vs low-risk control individuals Without data of Tavtigian et al. (2001) OR (95% CI) 1.17 (0.96–1.42) 0.96 (0.67–1.37) 1.81 (1.23–2.68) 1.86 (1.25–2.77) P 0.11 0.81 0.0029 0.0023 Totalsamplesize 1706 1706 1680 985 With data of Tavtigian et al. (2001) OR (95% CI) 1.17 (0.98–1.39) 1.18 (0.86–1.62) 2.01 (1.40–2.88) 2.05 (1.42–2.96) P 0.075 0.31 0.00014 0.00012 Totalsamplesize 2283 2283 2257 1312 All men with prostate cancer vs all control individuals Without data of Tavtigian et al. (2001) OR (95% CI) 1.09 (0.95–1.24) 1.02 (0.81–1.29) 1.25 (0.97–1.60) NA P 0.21 0.86 0.081 NA Totalsamplesize 3596 3596 3571 NA With data of Tavtigian et al. (2001) OR (95% CI) 1.10 (0.97–1.24) 1.13 (0.90–1.41) 1.35 (1.07–1.72) NA P 0.15 0.29 0.013 NA Totalsamplesize 4173 4173 4148 NA aMen with familial prostate cancer vs low risk control individuals comparison includes data from Xu et al. (2001a) (familial prostate cancer only) and Suarez et al. (2001). All men with prostate cancer vs low risk control individuals includes data from Xu et al. (2001a) (familial and sporadic prostate cancer), Suarez et al. (2001), and Rebbeck et al. (2000). All men with prostate cancer vs all control individuals comparison includes data from Xu et al. (2001a) (familial and sporadic prostate cancer), Suarez et al. (2001), Rebbeck et al. (2000), Vesprini et al. (2001) (male control individuals only), and Wang et al. (2001). b Association test for carriage of both Leu217 and Thr541 versus carriage of neither. NA indicates that data were not available from the relevant published papers to perform the multi-locus analysis. Adapted from Camp & Tavtigian (2002). the Americas) or groups (e.g. Druze or certain Jewish percentage of PAR similar or higher than that conferred by populations). Interestingly and as previously observed, a rare but highly penetrant susceptibility gene. Nevertheless, Thr541 always occurs in combination with Leu217 allele, it is important to note that this hypothesis should be con- suggesting origin of this mutation on the chromosome carry- firmed by direct analysis for each population. The disease- ing this particular allele, and tight linkage disequilibrium associated missense variant, Glu622Val (E622V), found in between the two of these. As shown in Table 4, frequency Finnish populations with a 1% population prevalence of Leu217 and Thr541 varies considerably from one popula- (Rokman et al. 2001) has not been observed in any of the tion to the other. Combining these results with ORs estimated previous samples of populations. in the meta-analysis of Camp and Tavtigian (Table 3) for Overall, mutation screening results indicate that inacti- carriage of a dominant Thr541 allele, it is tempting to specu- vating mutations in ELAC2, and even pedigree specific mis- late that the estimated percentage of population-attributable sense changes, are exceedingly rare (Rokman et al. 2001, risk due to this allele in different populations could be as Suarez et al. 2001, Tavtigian et al. 2001, Wang et al. 2001, elevated as 15% in Druze or certain Jewish populations as Xu et al. 2001a). Consequently, it will be very difficult to illustrated in Fig. 7. It is interesting to observe in Fig. 7 that determine whether, or to what extent, such variants confer a common low penetrance variant may be responsible for a increased risk of prostate cancer. In fact, there is little

242 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

Table 4 Frequency of Leu217 and Thr541 in different populations Regions Number of Ser217Leu Ala541Thr samplesa Leu-frequency (%) Thr-frequency (%) ±SD ±SD Sub-Saharan Africa 46 32.6 ± 4.9 1.6 ± 1.6 Americas 39 25.6 ± 4.9 8.3 ± 3.6 Europe 73 26.7 ± 3.7 3.4 ± 1.9 South-East Asia 39 1.3 ± 1.3 0 Middle-East Palestinian 22 22.7 ± 6.3 4.6 ± 3.1 Beduin 30 31.7 ± 6.0 3.3 ± 2.3 Druze 35 45.7 ± 6.0 14.3 ± 4.2 Jewish Ashkenazian 40 48.8 ± 5.6 12.5 ± 3.7 Moroccan 30 25.0 ± 5.6 3.3 ± 2.3 Bulgarian & Turkish 30 40.0 ± 6.3 8.3 ± 3.6 Iranian 21 59.5 ± 7.6 14.3 ± 5.4 Iraquian 29 31.0 ± 6.1 5.2 ± 2.9 Yemenite 30 26.7 ± 5.7 3.3 ± 2.3 aSub-Saharan-African sample included African-Americans and African Pygmies from Coriell, as well as African individuals from Montreal. Europe was represented by Canadians of European descent from Montreal and Italians, South-East Asia by individuals from that region living in Montreal, whereas Americas by Amerindians from Ontario and Saskatchewan as well as by Coriell samples from South and Central America (note that not all of the above samples were typed for Ala/Thr541 polymorphism). Middle-East was represented by samples from the collection of the National Laboratory for the Genetics of Israeli populations at Tel-Aviv University. When applicable, the informed consent was obtained from individuals donating genomic samples and the study was approved by the Ethics Committee at the Sainte-Justine Hospital. SD = square root of the variance = freq A × freq B/n; A is allele 1, B is allele 2 and n is the number of chromosomes

evidence that ELAC2 is a tumor suppressor. Knowing that knockout of the ELAC2 ortholog in Saccharomyces cerevis- iae is lethal, one might hypothesize that at least one func- tioning allele is required for cellular viability. On the other hand, the Leu217 and Thr541 missense changes are quite common and will eventually be used to tightly constrain the role that this gene plays in prostate cancer genetics. Finally, it must be noted that while the open reading frame of ELAC2 has been thoroughly screened for sequence variants, the pro- moter and upstream transcriptional regulatory sequences have not. Thus the possibility remains that some of the incon- sistency in current results is due to varying levels of linkage disequilibrium with as yet unknown sequence variants.

RNASEL (HPC1 region) The best supported linkage to prostate cancer susceptibility is the HPC1 locus at Chr 1q23/25. One of the candidate Figure 7 Population attributable risks (PAR) as a function of ′ ′ carrier frequency of the deleterious genotype in the reference genes within the broadly-defined HPC1 region is 2 -5 - population (f ) and the various levels of odds ratio (OR) for oligoadenylate-dependent ribonuclease L (RNASEL;MIM that deleterious genotype. PAR = 100% × f (OR-1)/ 601518 and 180435). The gene encodes a constitutively [f (OR-1) + 1)] (Lillienfeld & Lillienfeld 1980). Bold lines expressed latent endoribonuclease that mediates the antiviral indicate odds ratio (OR) conferred by genotype. Shaded area and proapoptotic activities of the interferon-inducible 2-5A represents estimations of the percentage of population- system. attributable risk conferred by Thr 541 allele in ELAC2 using Mutation screening of RNASEL from the germline DNA combined data from tables 3 and 4. It is important to note that additionalexperiments with largecohorts from each of one index case from each of 26 pedigrees selected from population should be done to confirm or refute the validiy of the Johns-Hopkins pedigree resource, including eight pedi- this hypothesis. grees that showed linkage to the HPC1 locus and that had at www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021243 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes least four affected individuals sharing an HPC1 haplotype, variant. Importantly, Casey et al. (2002) compared the in revealed two particularly interesting mutations: the nonsense vitro enzymatic activity of protein encoded by the two alleles mutation Glu265X and the initiation codon mutation Met1Ile on a synthetic substrate, finding that the Gln462 variant had (Carpten et al. 2002) (Fig. 6). The nonsense mutation only about 30% of the activity of the more common Arg462 Glu265X was found in an index case from a pedigree of allele. In their association study, which was carried out on a European ancestry. There were five prostate cancer cases in series of discordant sibling pairs including 423 cases and 454 the pedigree, and their average age at diagnosis was 67.6 controls, they found a trend of increasing risk from Arg462 years. DNA was available from four of the five cases, all homozygotes to Gln462 homozygotes (ORs of 1.46 and 2.12 brothers, and they all carried this mutation. It is not known respectively, trend test P = 0.011) (Casey et al. 2002). In con- whether the fifth case, the father, was a carrier; however, this trast, in their analysis of a case control series derived from may not be relevant because the mother was diagnosed with the Mayo Clinic pedigree resource, Wang et al. (2002) found breast cancer at age 59. The Met1Ile mutation was found in exactly the opposite trend: ORs were 0.83 and 0.54 (P = the African-American pedigree 097 and is either very rare or 0.02) for heterozygotes and Gln462 homozygotes respect- pedigree-specific. The six prostate cancer cases in pedigree ively. In addition, the protective effect of the Gln462 variant 097 had an average age at diagnosis of 59. The four out of was most apparent in early onset cases; ORs were 0.63 and six cases who carried the mutation had an average age at 0.29 (P = 0.0008) for heterozygotes and Gln462 homozy- diagnosis of 57.8 years while the two non carriers were diag- gotes diagnosed at ȅ 64 years respectively (Wang et al. nosed at ages 59 and 64. Interesting data in support of a role 2002). for these mutations in their carrier’s prostate tumors was that In summary, published analyses of the three inactivating tumor material was available from several carriers, PCR from mutations so far described in RNASEL, Met1Ile, microdissected tumor material indicated that loss of hetero- 471delAAAG, and Glu265X, trend towards evidence that zygosity (LOH) was present, and it was the wild-type allele these variants confer risk of prostate cancer, with best evi- that was lost. On the other hand, genotyping indicated that dence in earlier onset cases as predicted from the nature of the Glu265X mutation has an appreciable frequency in the the linkage evidence at HPC1. Data with respect to the mis- US Caucasian population, about 0.5%, and there was no sense change Arg462Gln are clearly conflicting. As obvious carrier frequency difference between cases and con- 471delAAAG appears to be an Ashkenazi founder mutation trols. with an appreciable allele frequency in that population, RNASEL was subsequently mutation screened in a series Glu265X appears to have an appreciable frequency in Euro- of Finnish familial prostate cancer cases and certain sequence pean populations, and Arg462Gln is a common missense variants from the gene typed in a Finnish case control series change, it should not take a very long time for independent (Rokman et al. 2002). The Glu265X nonsense mutation was follow-on studies to tightly determine the risk conferred by found in five of 116 patients with familial prostate cancer these sequence variants and thereby clarify the role of this and seemed to be most common in families with four or more gene in hereditary prostate cancer. cases. Genotyping in the case-control series revealed that the nonsense mutation was most common in familial cases MSR1 (Chr 8p region) (q = 0.022), at intermediate frequency in sporadic cases (q = 0.010), and least common in controls (q = 0.009). The differ- One of the ‘Group 2’ linkages described above in ‘Mapping ence between cases and controls was significant (OR = 4.56, putative loci for prostate cancer susceptibility genes’ is at P = 0.04). In addition, one of four common missense vari- Chr 8p. As part of the justification for scrutinizing 8p was ants, Arg462Gln, also trended towards being more common longstanding evidence for a prostate cancer tumor suppressor in cases than controls (OR = 1.96, P = 0.07). in the region, there is some expectation that any susceptibility Mutation screening of RNASEL in a series of Ashkenazi gene underlying the linkage should be a tumor suppressor prostate cancer patients revealed a founder mutation, and therefore subject to inactivating mutations. In addition, 471delAAAG, at appreciable allele frequency in that ethnic as this is a region of linkage, one would expect to see such group (Rennert et al. 2002). In a small case control series, germline mutations at higher frequency in familial as the frequency of this variant was highest in cases (q = 0.035), opposed to sporadic cases. Similarly, as the majority of the intermediate in population controls (q = 0.020), and lowest linkage evidence for the region seems to come from pedi- in elderly cancer-free men (q = 0.012). Although the carrier grees with older average age at diagnosis, one would expect frequency difference between cases and controls was not sig- to see a germline mutation bias towards older (or at least not nificant (OR = 3.0, P = 0.17), cases who carried the variant markedly towards younger) age at diagnosis cases. were younger at diagnosis than cases who did not (65 versus Very recently, mutation screening of candidate genes 74.4 years, P < 0.001). from the 8p region in index cases from the Johns-Hopkins In addition to the Finnish group, two other groups have family resource revealed a number of rare mutations in the published association studies of the Arg462Gln missense gene Macrophage Scavenger Receptor 1 (MSR1; MIM

244 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

176807)(Xuet al. 2002). From the first 159 prostate cancer families (Anderson & Badzioch 1992, Tulinius et al. 1992). pedigrees screened, one nonsense mutation and six rare mis- Anderson and Badzioch observed a doubling of breast cancer sense mutations were found. After genotyping for these vari- risk in families with prostate cancer history. An estimated ants was completed on the initial 159 pedigrees plus a second relative risk of prostate cancer of 3.33 was calculated for set of 31 prostate cancer pedigrees, a total of 13 of the pedi- obligate male carriers of a deleterious BRCA1 mutation when grees was found to harbor one or another (and in some cases compared with the general population in a cooperative study two) of these rare variants. There were individual pedigrees using 33 BRCA1-linked breast/ovarian cancer families (Ford where the sequence variant segregated quite closely with et al. 1994). More recently, the Breast Cancer Linkage Con- prostate cancer, and others where the pattern of segregation sortium conducted another study including 11 487 indi- was equivocal. Overall, a family based linkage/segregation viduals from 699 families segregating a BRCA1 mutation that test provided evidence that there was disproportionate segre- were ascertained in 30 centers across Europe and North = gation of these variants with disease (P 0.0007). From the America. There was again some evidence of an elevated risk point of view of follow-up studies, a very useful point was of prostate cancer in BRCA1 mutation carriers younger than that six of the 13 pedigrees in which a mutation was found, 65 years old (relative risk (RR) = 1.82, 95% CI = 1.01–3.29, all of European descent, harbored the nonsense mutation P = 0.05), but not in those 65 years old or older (RR = 0.84, Arg293X. That nonsense variant was also found to be more 95% CI = 0.53–1.33, P = 0.45). common in ‘non-hereditary’ prostate cancer cases than dis- On the other hand, in another recent study from the ease-free male controls (8/317 vs 1/256, Fisher exact test Breast Cancer Linkage Consortium including 173 breast/ (FET) P-value = 0.047). ovarian families with a deleterious BRCA2 (MIM 600185) MSR1 is a transmembrane protein that functions as a = homotrimeric receptor for a number of polyanionic ligands mutation, a significantly increased risk (RR 4.65) for pros- including a variety of bacteria (Fig. 6). The Arg293X tate cancer was observed. Furthermore, this risk was even = mutation would remove most of the extracellular ligand bind- higher before age 65 (RR 7.33) and the estimated cumulat- ing domain as well as the conserved extracellular scavenger ive incidence before 70 was 7.5%–33%, depending on which receptor cystein-rich domain. Thus the Arg293X nonsense reference population was used (The Breast Cancer Linkage mutation would clearly interfere with MSR1 function. Over- Consortium 1999). Moreover, the founder Icelandic BRCA2 all, the initial report presented modest but plausible evidence mutation (999del5) was also associated with increased risk that MSR1 is a prostate cancer susceptibility gene. If the of prostate cancer (Sigurdsson et al. 1997, Thorlacius et al. allele frequency of the Arg293X turns out be in the range of 1996). The relative risk of prostate cancer was calculated to 1.0% to 1.5% in prostate cancer cases of European ethnicity, be 4.6 in first-degree relatives and 2.5 in second degree rela- as indicated from the initial mutation screening and associa- tives (Sigurdsson et al. 1997). Futhermore, another popula- tion evidence, then that specific variant should provide an tion-based estimate shows a cumulative risk for 999del5 accessible and powerful tool to evaluate the genetic role that mutation carriers of only 7.6% at age 70 (Thorlacius et al. this gene plays in prostate cancer susceptibility. More 1998). Several analyses of germline DNA from prostate recently, Xu et al. (2003) studied five common sequence cancer cases with or without a family history have revealed variants of MSR1 in 301 prostate cancer patients at Johns that there is no increased frequency of the founder Ashkenazi Hopkins Hospital with non hereditary prostate cancer and Jewish BRCA1 and BRCA2 mutations over that expected in 250 control subjects who had normal digital rectal examin- this population (Lehrer et al. 1998, Hubert et al. 1999, Nas- ation and PSA levels ( < 4 ng/ml). The five sequence variant tiuk et al. 1999, Wilkens et al. 1999). This apparent discrep- genotypes included an SNP in the promoter region, a 15-bp ancy may be consistent with the positional effect of the Ash- insertion/deletion in intron 1, an SNP in intron 5, a missense kenazi Jewish founder BRCA2 mutation 6174delT which lies change in exon 6 (P275A) and a 3-bp insertion/deletion in within the ‘ovarian cancer cluster region’ (OCCR). Indeed, intron 7. Haplotype analysis show a significant difference in Thompson and Easton (2001) found that the risk of prostate the haplotype frequencies from a global score test (P = cancer was lower in carriers of BRCA2 mutations located in 0.011). Moreover, it appears that the significant association the OCCR (nucleotides 3035–6629) than in carriers of other between the common MSR1 sequence variants and the pros- mutations (RR = 0.48). Note that mutations in this region, tate cancer risk is independent of the impact of the known which is coincident with the BRC repeat domain responsible rare MSR1 mutation. Indeed, the haplotype giving the best P-value did not harbor any of the rare mutations for RAD51 binding are also associated with a reduced risk of breast cancer, but a higher risk of ovarian cancer. A recent study in a Swedish family in which the father and four of his Risk of prostate cancer in BRCA1 and BRCA2 sons were diagnosed with prostate cancer at the exceptionally carriers early ages of 51, 52, 56, 58 and 63 respectively, supports an Epidemiological studies of prostate and breast cancer have increased risk of prostate cancer in BRCA2 carriers revealed that clustering of these cancers occurs in certain (Gro¨nberg et al. 2001). www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021245 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

However, DNA samples from affected individuals in 38 tribute to the Mendelian inheritance of prostate cancer as prostate cancer familial clusters were analyzed for germline described above, some of the familial risks may be due to mutations in BRCA1 and BRCA2 genes to assess the potential shared environment and more specifically to common low- contribution of each of these genes to familial prostate cancer penetrance genetic variants, which alter predisposition to (Gayther et al. 2000). No germline mutations were found in prostate cancer. Prostate cancer has proven to be the most BRCA1, but two novel deletions were found in BRCA2 in hormone-sensitive cancer to hormonal manipulation (Labrie two of 16 patients who received diagnosis at ages 52 and 56 2003). It is not surprising that analyses of genes encoding years, whereas no mutations were found among the 22 older key proteins involved in androgen biosynthesis and action patients who received diagnosis at or after 60 years of onset. led to the observation of a significant association between These authors proposed that germline mutations in BRCA2 common genetic variants and a susceptibility to prostate may therefore account for about 5% of prostate cancers in cancer (Ross et al. 1999, Singh 2000, Makridakis & Reich- familial clusters. On the other hand, in another study, BRCA1 ardt 2001, Nwosu et al. 2001). Such analyses provided some and BRCA2 mutation screening was performed on the pros- understanding of how common low-penetrance polymor- tate cancer proband and on an additional family member phisms present in a number of these candidate genes were affected with breast or prostate cancer from each of the 22 involved in prostate cancer onset, progression and response cancer families (Sinclair et al. 2000). None of the 43 samples to treatment for the disease. screened contained a protein truncating mutation in either BRCA1 or BRCA2. It is important to note that, in contrast to Androgen receptor the UK study described above (Gayther et al. 2000), these American families were selected from a large collection of Because androgen action in prostate cells is mediated through high-risk prostate cancer families (at least three cases of an interaction with the androgen receptor (AR), it was first prostate cancer) for having at least two cases of breast or suggested that abnormalities in the AR gene, which is located ovarian cancer, in order to maximize the odds of detecting on Xq11-12 (~50 cM centromeric to HPCX), could play a mutations in BRCA1 or BRCA2. Very recently, to establish key role in prostate cell proliferation and differentiation as the contribution of BRCA2 mutations to early-onset prostate well as in carcinogenesis (Fig. 8). In this regard, the precise cancer, Eeles’ group have screened a total of 263 men diag- molecular events leading from androgen-sensitive prostate nosed at ages ȅ 55 years, who were unselected for family cancer to androgen-refractory prostate cancer are of special history (Edwards et al. 2003). Protein-truncating mutations interest (see Grossmann et al. 2001, Henshall et al. 2001, were found in six men (2.3%; 95% CI = 0.8%-5.0%) and all Linja et al. 2001, Montgomery et al. 2001 for reviews). of these mutations were clustered outside the OCCR. The Indeed, some of the pathways identified appear to directly estimated relative risk of developing prostate cancer by age involve the modulation of AR’s ability to respond to specific 56 years for male BRCA2 mutation carriers was 23-fold. ligands. Briefly, mutations in the AR hormone binding They therefore estimate the absolute risk of prostate cancer domain, or amplification of the AR gene could result in an in BRCA2 carriers to be ~ 1.3% by age 55 years and 10% by increased sensitivity to androgens made locally in the pros- age 65 years knowing that the cumulative risk by age 55 and tate from the inactive steroid precursors, dehydroepiandros- 65 years in the general population in the UK are ȁ 0.06% terone (DHEA) and its sulfate (DHEA-S). These androgens and 1.5% respectively. The Eeles’ team estimated that of adrenal origin do not bind to the AR, but exert androgenic BRCA2 mutations account for 6% of the excess familial risk action after their conversion into active androgens in target of the disease and is responsible for a significant fraction tissues. Thus, recurrent hormone-refractory tumors may not of early-onset prostate cancer cases outside of families with always be androgen-independent, as often thought, but rather multiple cases of breast-ovarian cancer. Thus, BRCA2 is a clonal expansion of hypersensitive cells that are able to grow high-risk gene for which definitive evidence of susceptibility due to intracrine formation of androgens originating from to prostate cancer has been demonstrated in several indepen- adrenal precursors and which are converted to bioactive dent studies. dihydrotestosterone (DHT) in the prostate tissue may cause failure to monotherapy. Secondly, the occurrence of mutations in several portions Common low-penetrance allelic variants of the AR could allow this receptor to respond to other ster- of genes involved in androgen oids or to some antiandrogens. Alternatively, coactivators biosynthesis and action may increase the sensitivity of the AR to androgens and even to other nonandrogenic molecules (Grossmann et al. 2001). In the general population, we observe an inter-individual Furthermore, there is accumulating evidence that additional variability in the susceptibility to cancer; however little is mechanisms, that do not result from mutations in the AR, known about the underlying genes contributing to this varia- may involve stimulation of AR transactivation by various bility. Although a number of rare highly penetrant loci con- factors and cytokines independently of androgens, thus

246 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

Figure 8 Chromosomal localization and schematic representation of HSD3B1, HSD3B2, SRD5A2 and AR genes, mRNA species and corresponding proteins. Exons are represented by black boxes indicating the coding regions, whereas open boxes represent the non-coding regions. Introns are represented by black bold lines. Positions of sequence variants are shown by an arrow in reference to nucleic or amino acid sequences on the genes or proteins. Polymorphisms which have been suggested to be associated with increased risk of prostate cancer are encircled. facilitating the growth of prostate cancer cells despite maxi- further supported the inverse relationship between AR tran- mal androgen blockade or acquisition of metastatic pheno- scriptional activity and the number of CAG repeats (15 type (Grossmann et al. 2001). versus 31 CAG) and provided evidence that this finding was

Accounting for more than half of the AR, the NH2- cell specific, thus suggesting the involvement of accessory terminal transactivation domain is encoded by exon 1. Within factors differentially expressed between different cell lines this domain, there are two polymorphic microsatellite trinu- (Beilin et al. 2000). It is also of interest to mention that cleotide repeats located ~1.1 kb apart, namely a CAG repeat extended polyglutamine tract interacts with caspase-8 and and a downstream GGN polymorphism encoding variable -10 in nuclear aggregates (U et al. 2001) and increasing its length polyglutamine and polyglycine regions respectively. length negatively affects p160-mediated coactivation of the The CAG repeat varies in length between 11 and 31 repeats AR (Irvine et al. 2000). Coetzee and Ross (1994) have sug- in the germline DNA of normal men (see Montgomery et al. gested that enhanced activity of the AR, attributable to poly- 2001 for a review) and shows an inverse relationship morphisms in the AR gene, might alter the risk of prostate between the length of repeats and the transcriptional activity cancer. Accordingly, the association of CAG repeat length of the AR (Chamberlain et al. 1994, Kazemi-Esfarjani et al. and prostate cancer risk has been examined in several case- 1995, Beilin et al. 2000). A recent study using the rat proba- control studies but the results are inconclusive (Table 5). As sin promoter, an androgen- and prostate-specific reporter, several of these studies show an association between shorter www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021247 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

Table 5 Comparison of case/controlstudies of AR-CAG and AR-GGC/N polymorphismand prostate cancer risk References Study No. CAG OR/RR 95% CI Remarks population cases/ com- controls parison Studies showing association between shorter CAG repeat length and increased prostate cancer risk Irvine et al. Non-Hispanic 19/15 Ն 22 1.0 Referent Modest association (1995) white men (Los 38/24 < 22 1.25 0.88–1.73 Angeles county) Hakimi et al. Caucasians 53/359 >17 1.0 Referent (1997) (United States) 6/11 Յ 17 3.7 1.3–10.5 Ingles et al. Non-Hispanic 19/68 Ն 22 1.0 Referent This study was an extension of the (1997) white men 14/56 20–21 0.89 0.41–1.94 earlier Irvine et al. (1995) (Los Angeles 24/45 < 20 1.91 0.94–3.88 investigation. The AR short alleles county) adjusted (< 20) were more strongly OR = 2.1 1.01–4.0 associated with advanced disease than with localized disease Xue et al. Non-Hispanic 33/114 > 20 1.0 Referent A statistically significant increase in (2000) white men 24/42 < 20 1.97 1.05–3.72 risk was conferred by either a short (Los Angeles CAG repeat or PSA genotype GG, county) either alone or in combination Hsing et al. Chinese 74/154 Ն 23 1.0 Referent The median repeat length among (2000a) 116/146 < 23 1.65 1.14–2.39 the China population was longer – 23 repeats – than that observed for Western men. Giovannucci Predominantly 60/72 Ն 26 1.0 Referent An association was observed for et al. Caucasians 98/115 24–25 1.02 0.66–1.58 both advanced stage (RR = 2.2) (1997) 116/119 22–23 1.17 0.76–1.80 and high grade tumors (RR = 1.9) 113/101 21 1.32 0.87–2.09 69/65 20 1.28 0.79–2.08 62/61 19 1.22 0.75–2.00 69/55 Յ 18 1.52 0.92–2.49 Stanford et al. Caucasians 136/140 Ն 22 1.0 Referent (1997) (United States) 145/126 < 22 1.23 0.88–1.73 Miller et al. Caucasians 71/34 Ն 22 1.0 Referent The OR for prostate cancer (2001) (United States) 66/35 < 22 1.13 0.54–2.37 associated with short CAG repeats was higher in sibships with a median age of < 66 yrs at diagnosis (OR = 1.72)

Studies showing no significant association between shorter CAG repeat length and increased prostate cancer risk Ekman (1999) Swedish and No difference was observed in Japanese number of CAG repeats between sporadic and hereditary prostate cancer. CAG repeats were shorter among Swedish prostate cancer patients than among Swedish controls; Japanese prostate cancer patients had longer repeats than Japanese controls Bratt et al. Swedish 92/93 1 CAG 0.97 0.91–1.04 For men with non-hereditary (1999) decrement prostate cancer, shorter CAG repeats correlated with younger age at diagnosis (P = 0.03) high grade (P = 0.07) and high stage (P = 0.07)

248 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

Table 5 Continued References Study No. CAG OR/RR 95% CI Remarks population cases/ com- controls parison Correa- French and 39/28 Ն 24 1.0 Referent Cerro et al. German 30/22 22–23 1.02 0.49–2.13 (1999) 34/35 20–21 1.43 0.73–2.82 29/20 Յ 19 0.96 0.45–2.03 Edwards et al. Caucasians 74/178 > 21 1.0 Referent (1999) (United 88/212 Յ 21 1.0 0.96–1.03 Kingdom) Lange et al. Caucasians 133/305 Ն 26 1.0 Referent (2000) (United States) Յ 18 0.73 0.31–1.69 Beilin et al. Australian 456/456 No evidence of a relationship of (2001) CAG repeat length to severity of prostate cancer was found. However, a shorter repeat CAG sequence was associated with earlier age at diagnosis Latil et al. White French 41/30 > 24 1.0 Referent (2001) origin 55/36 23–24 1.12 0.59–2.10 61/45 21–22 0.99 0.54–1.82 68/45 Յ 20 1.11 0.60–2.02 Chang et al. Predominantly 164/81 Ն 22 1.0 Referent (2002b) Caucasians 162/99 Յ 21 0.81 0.56–1.17 (United States) Chen et al. Predominantly 156/147 Ն 22 1.0 Referent (2002) Caucasians 144/153 < 22 0.89 0.65–1.23 (United States)

References Study No. GGC/N OR/RR 95% CI Remarks population cases/ com- controls parison Studies showing association between shorter GGC/N repeat length and increased prostate cancer risk Hakimi et al Caucasians 46/106 > 14 1.0 Referent (1997) (United States) 8/4 Յ 14 4.6 1.3–16.1 Stanford et al. Caucasians 56/75 > 16 1.0 Referent Men with GGC Յ 16 were at (1997) (United States) 201/175 Յ 16 1.60 1.07–2.41 increased risk regardless of age at diagnosis (OR = 1.40 in men < 60 yrs and OR = 2.23 in men 60–64 yrs) Chang et al. Predominantly 100/72 Ն 17 1.0 Referent (2002b) Caucasians 227/102 Յ 16 1.58 1.08–2.32 (United States) Platz et al. United States 244/369 Not 23 1.0 Referent Modest association but not (1998) male 338/425 23 1.2 0.97–1.49 statistically significant participants in the Physician’s Health Study Irvine et al. Non-Hispanic 10/21 16 1.0 Referent Modest association but not (1995) white men 27/16 Not 16 1.18 Not given statistically significant (Los Angeles county) Hsing et al. Chinese 147/239 Ն 23 1.0 Referent Modest association but not (2000a) 39/56 < 23 1.12 0.71–1.78 statistically significant

www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021249 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

Table 5 Continued References Study No. GGC/N OR/RR 95% CI Remarks population cases/ com- controls parison Studies showing no significant association between shorter GGC/N repeat length and increased prostate cancer risk Edwards et al. Caucasians 64/116 > 16 1.0 Referent (1999) (United 98/168 Յ 16 1.06 0.57–1.96 Kingdom) Miller et al. Caucasians 51/25 > 16 1.0 Referent A reduced risk was found among (2001) (United States) 82/44 Յ 16 0.98 0.46–2.06 men diagnosed at age < 66 yrs (OR = 0.36) and an increased risk among men diagnosed at age Ն 66 yrs (OR = 1.56) Chen et al. Predominantly 115/100 > 17 1.0 Referent A reduced risk associated with (2002) Caucasians 185/200 Յ 17 0.8 0.57–1.12 Յ 17 GGC/N repeats in men Ն 70 (United States) 144/127 Not 17 1.0 Referent yrs (OR = 0.51) was observed 156/173 17 0.79 0.57–1.09

References Study No. CAG and OR/RR 95% CI Remarks population cases/ GGC/N controls comparison

Studies showing association between CAG and GGC/N repeat length and increased prostate cancer risk Irvine et al. Non-Hispanic 34/28 Other 1.0 Referent (1995) white men (Los 23/9 < 22, not 16 2.10 Not given Angeles county) Stanford et al. Caucasians 22/32 Ն 22, > 16 1.0 Referent (1997) (United States) 32/41 Ն 22, Յ 16 1.15 0.56–2.35 97/93 < 22, > 16 1.54 0.83–2.86 98/77 < 22, Յ 16 2.05 1.09–3.84 Platz et al. United States 66/119 > 23, not 23 1.0 Referent (1998) male 90/133 > 23, 23 1.17 0.77–1.77 participants in 75/116 21–23, not 23 1.39 0.93–2.06 the Physician’s 152/185 21–23, 23 1.22 0.82–1.83 Health Study 103/134 < 21, not 23 1.49 1.02–2.15 96/107 < 21, 23 1.62 1.07–2.44 Hsing et al. Chinese 53/120 Ն 23, Ն 23 1.0 Referent (2000a) 19/29 Ն 23, < 23 1.48 0.76–2.88 94/115 < 23, Ն 23 1.85 1.21–2.82 20/26 < 23, < 23 1.75 0.9–3.41 Chang et al. Predominantly 42/30 Ն 22, Ն 17 1.0 Referent (2002b) Caucasians 109/46 Ն 22, Յ 16 1.62 0.92–2.95 (United States) 50/39 Յ 21, Ն 17 0.92 0.49–1.72 99/54 Յ 21, Յ 16 1.29 0.72–2.29

Studies showing no significant association between CAG and GGC/N repeat length and increased prostate cancer risk Miller et al. Caucasians 21/8 Ն 22, > 16 1.0 Referent (2001) (United States) 45/26 Ն 22, Յ 16 0.63 0.18–2.2 29/17 < 22, > 16 0.69 0.17–2.74 35/17 < 22, Յ 16 1.06 0.25–4.46 Chen et al. Predominantly 49/29 Ն 22, > 17 1.0 Referent (2002) Caucasians 107/118 Ն 22, Յ 17 0.54 0.32–0.91 (United States) 66/71 < 22, > 17 0.55 0.31–0.98 78/82 < 22, Յ 17 0.56 0.32–0.98 Adapted from Coughlin & Hall (2002) and Chen et al. (2002).

250 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

AR CAG repeat lengths (a short CAG has been defined as segregation analyses. Further association studies using larger ranging from < 17 to < 23 repeats, depending on the study) cohorts of preferably > 1000 blood samples from prostate and increased prostate cancer risk, it is not entirely clear cancer cases versus ethnically matched controls, and well whether the association is with diagnosis of prostate cancer, characterized clinical data, will thus be important to clarify response to endocrine therapy or severity of disease. On the the role of these low-penetrance trinucleotide polymorphic other hand, several other studies suggested that CAG repeat repeats in prostate cancer development and progression, length is not associated with prostate cancer risk, age of taking into account their potential gene-gene and gene- onset, histological grade and stage of disease at diagnosis or environment interactions, which likely cumulate to contribute several clinical parameters, including response to hormonal to an overall prostate cancer risk and to disease character- therapy, time to progression after hormonal therapy, disease- istics. free survival and overall survival (see Montgomery et al. 2001, Chen et al. 2002, Coughlin & Hall 2002 for recent 5α-Reductase type 2 reviews). Finally, a recent study compared repeat lengths of 140 men with prostate cancer with their brothers (n = 70) One of the best candidate genes is the 5α-reductase type 2 without disease from 51 high-risk sibships, stratified by gene (SRD5A2), which catalyzes the conversion of testos- median age at diagnosis of affected men within each sibship. terone into the more active androgen DHT and maps to Men with both a short CAG repeat length ( < 22) and a short 2p22-23 (Morissette et al. 1996) (Fig. 8). Intraprostatic DHT GGN repeat ( ȅ 16) array were not at higher risk (OR = 1.06; is the most meaningful parameter of androgen action in pros- 95% CI = 0.25–4.46) compared with men with two long tatic tissue (Fig. 9). Indeed, the importance of intracrine for- repeats (CAG repeat Ȅ 22 and a GGN repeat > 16), thus sug- mation of active androgens is in concordance with the obser- gesting that the CAG and GGN repeats in the AR gene do vation that after elimination of testicular androgens by not play a major role in familial prostate cancer (Miller et al. medical or surgical castration, the intraprostatic concentra- 2001). More recently, a case-control study nested within the tion of DHT remains approximately 40% of that measured in β Carotene and Retinol Efficacy Trial including 300 cases the prostate of intact 65-year-old men, thus leaving important and 300 controls showed there was no appreciable difference amounts of free androgen to continue stimulating the growth in the mean number of GGC repeats between cases and con- of the prostate cancer (Labrie et al. 1998). However, male trols and the risk in men at or below the mean number of pseudohermaphrodites with 5α-reductase deficiency caused GGC repeats (17) was 0.80 (95% CI = 0.57–1.12) (Chen et by mutations in the SRD5A2 gene exhibit external genital al. 2002). Moreover, these authors found a decrease in pros- ambiguity and hypoplastic prostate, thus supporting the cru- tate cancer risk among men with CAG < 22 and GGC ȅ 17, cial role of this enzyme in normal prostate development. or CAG < 22 and GGC > 17 or CAG Ȅ 22 and GGC ȅ 17 Modulation of 5α-reductase activity may account for part of when compared with men with CAG Ȅ 22 and GGC > 17. the substantial racial/ethnic disparity in prostate cancer risk These results are thus in contrast with two other studies (Ross et al. 1992, Reichardt et al. 1995, Makridakis et al. showing that the presence of a short repeat length for either 1997). Furthermore, one variant, Ala49Thr, has been or both CAG and GGC alleles was associated with an reported to increase the catalytic activity of this enzyme increased risk of prostate cancer (Stanford et al. 1997, Hsing (Makridakis et al. 2000) and is associated with an increased et al. 2000a). Finally, using the Johns Hopkins set, a signifi- risk of advanced prostate cancer (Makridakis et al. 1999, cant association was observed in the frequencies of the ȅ 16 Jaffe et al. 2000, Margiotti et al. 2000). This low-penetrance GGC repeat alleles in 159 hereditary prostate cancer (HPC) allele appears to increase the risk of prostate cancer in (71%), 245 sporadic cases (68%) compared with 211 controls African-Americans, Latinos and Italians by 7.2- (P < 0.001), (59%), thus supporting the thesis that GGC repeats were 3.6- (P < 0.04) and 7.7-fold (P < 0.11) respectively associated with prostate cancer risk (P = 0.02) (Chang et al. (Makridakis et al. 1999, Jaffe et al. 2000, Margiotti et al. 2002b). Furthermore, male-limited X-linked transmission/ 2000). In contrast, the prevalence of the A49T variant in 449 disequilibrium (XLRC-TDT) was also used by these authors Finnish prostate cancer patients was 6.0%, not differing sig- to demonstrate the preferential transmission of short ȅ 16 nificantly from 6.3% observed in 223 patients with BPH or GGC repeat alleles from heterozygous mothers to their affec- 5.8% in 588 population-based controls. Furthermore, there ted sons (z′=2.65, P = 0.008). The stronger evidence for was no association between A49T and the family history of linkage in the families with male-to-male transmission may the patients nor with tumor stage or grade (Mononen et al. be explained by the hypothesis that AR is a strong modifier 2001). Finally, the T allele was not observed in a population- gene that works in conjunction with autosomal susceptibility based case control study in China after the genotyping of 191 gene(s). In support of this interpretation, Cui et al. (2001) cases and 304 controls (Hsing et al. 2001). It is also of inter- observed that the two-loci model, combining autosomal est to know that substantial pharmacogenetic variation dominant with either an autosomal recessive or X-linked among the SRD5A2 sequence variants was observed when model, fit their data better than did a single-locus model in three competitive inhibitors (Finasteride, GG745 and www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021251 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

Figure 9 Major enzymatic pathways involved in the formation and inactivation of androgens in human, and the relative importance (%) of adrenalsteroid precursors in the intraprostatic concentration of DHT in adultmen. A-DIONE, androstanedione; ADT, androsterone; ADT-G, androsterone glucuronide; ADT-S, androsterone sulfate; DHEA, dehydroepiandrosterone; DHEA-S, dehydroepiandrosterone sulfate; DHT, dihydrotestosterone; PREG, pregnenolone; PROG, progesterone; epi-ADT, epi-androsterone; TESTO, testosterone; AR, androgen receptor; CYP17, cytochrome P450c17, 17α-hydroxylase/17,20 lyase; P450scc, cholesterol side-chain cleavage enzyme; RoDH-1, retinol dehydrogenase type 1; UGT, UDP-glucuronosyltransferase; 3α-DIOL, 5α-androstane-3α,17β-diol; 3α-DIOL-G, 5α-androstane-3α,17β-diol-glucuronide; 3α HSD, 3α-hydroxysteroid dehydrogenase; 3(α Ǟ β) HSE, 3(α Ǟ β) hydroxysteroid epimerase; 3β-DIOL, 5α-androstane-3β, 17β-diol; 3β HSD, 3β hydroxysteroid dehydrogenase/∆5-∆4-isomerase; 4-DIONE, androstenedione; 5-DIOL, androst-5-ene-3β,17β-diol; 5-DIOL-S, androst-5-ene-3β,17β-diolsulfate;17 β HSD, 17β hydroxysteroid dehydrogenase. From Simard et al. (2002a); copyright owner, The Endocrine Society.

PNU157706) were tested, thus providing relevant infor- putative Sp1-type (CCACC box) transcriptional factor bind- mation to take into account when prescribing such inhibitors ing site that was postulated to increase its gene expression. for the chemoprevention or treatment of prostate diseases Nedelcheva Kristensen et al. (1999) showed that neither the (Makridakis et al. 1999, 2000). A1 nor A2 allele could form a complex with the Sp1 protein in electrophoretic mobility shift assays. Moreover, a recent CYP17 study using a promoter/reporter gene construct containing sequences from −227 to + 61 with eitheraTorCatpos- Another key enzyme involved in the testicular synthesis of ition + 27 demonstrated that this polymorphism is not associ- androgens is encoded by the CYP17 gene, namely the cyto- ated with an altered transcriptional activity (Lin & Miller chrome P450c17, which catalyzes 17α-hydroxylase and 2001). Nevertheless, three independent association studies 17,20-lyase activities (Fig. 9). The CYP17 gene is mapped to reported a small but positive association between the A2 10q24.3 and consists of 8 exons. A single-base change (a T allele and an increased risk for prostate cancer (Lunn et al. (A1 allele) to C (A2 allele) transition) creates an additional 1999, Gsur et al. 2000, Yamada et al. 2001). On the other

252 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259 hand, two other studies showed that the A1 allele is the risk tion between prostate cancer risk and a missense polymor- allele for prostate cancer (Wadelius et al. 1999, Habuchi phism (HSD3B1-N367T) was found, stronger evidence for et al. 2000). Considering its essential role in androgen association was found when the joint effect of the two genes biosynthesis, a recent study was designed to further evaluate was considered. Indeed, men with the variant genotypes at the impact of this polymorphism and to investigate what is either HSD3B1-N367T or HSD3B2-c7519g had a relative the potential role of this gene in hereditary prostate cancer risk of 1.76 (95% CI = 1.21–2.57, P = 0.003) of developing (Chang et al. 2001). These authors performed a genetic link- prostate cancer compared with men who were homozygous age study and family-based association analysis in 159 famil- wild-type at both genes, whereas the risk for the hereditary ies (Johns-Hopkins set), each of which contained at least type of prostate cancer was stronger, with a relative risk of three first-degree relatives with prostate cancer. It is of inter- 2.17 (95% CI = 1.29–3.65, P = 0.003). Most importantly, the est to note that a linkage analysis is insensitive to allelic het- subset of hereditary prostate cancer probands, whose families erogeneity, thus if a mutation has a high penetrance and there provided evidence for linkage at 1p13, predominantly con- are multiple such mutations within a gene, a linkage study is tributed to the observed association (Chang et al. 2002a). likely to detect such a gene, whereas family-based or popula- Additional studies are warranted to confirm these findings. tion-based association approaches are likely to fail (Chang et al. 2001). Thus, information concerning specific sequence Conclusion variants within a gene is not necessary for a linkage study, but is essential for association studies. Their analyses suggest The clearest conclusion emerging from the linkage studies evidence for linkage at marker D10S222 with a LOD score of described above is perhaps that no single susceptibility locus 1.03 (P < 0.01). However, they did not observe a statistically mapped to date is by itself responsible for a large portion of increased risk of sporadic prostate cancer or hereditary pros- familial prostate cancers, at least not with the combination tate cancer in subjects with the A2 variant following geno- of penetrance and clinical features that allowed linkage to typing of the polymorphism in 159 HPC probands, 249 spor- and positional cloning of the so-far known colon and breast adic cases and 211 unaffected control subjects. They cancer predisposition genes. Thus, locus heterogeneity is a concluded that CYP17 gene or other genes in this region may major parameter in the identification or confirmation of link- increase the susceptibility to prostate cancer, whereas the age for many data sets. Indeed, the possible existence of mul- polymorphism in the 5′ promoter region has a minor if any tiple prostate cancer genes may well explain why there has effect in increasing prostate cancer susceptibility in their been limited confirmatory evidence of linkage for currently study sample. known highly-penetrant susceptibility loci/genes. As methods for statistical modeling improve, geneticists will have better tools to deal with the apparent extreme hetero- 3β-Hydroxysteroid dehydrogenases geneity within data sets (Goddard et al. 2001, Nwosu et al. The 3β-hydroxysteroid dehydrogenase/∆5-∆4 isomerase 2001, Schaid et al. 2001). (3β-HSD) isoenzymes are responsible for the oxidation and To clarify the role of low-penetrance polymorphisms in isomerisation of ∆5-3β-hydroxysteroid precursors into ∆4- prostate cancer development and progression, association ketosteroids, thus catalyzing an essential step in the forma- studies using larger cohorts of preferably > 1000 blood tion of all classes of active steroid hormones. In humans, samples from prostate cancer cases versus ethnically matched expression of the type I isoenzyme accounts for the 3β-HSD controls, and well characterized clinical data, will thus be activity found in placenta and peripheral tissues, whereas the important (Singh et al. 2000). It will also be necessary to type II 3β-HSD isoenzyme is predominantly expressed in the investigate, using various complementary genetic epidemiol- adrenal gland, ovary and testis and its deficiency is respon- ogical approaches, the potential role of the other key sible for a rare form of congenital adrenal hyperplasia steroidogenic enzymes involved in the formation and/or (Simard et al. 2002b). To evaluate the possible role of these inactivation of androgens, such as the numerous 17β-HSDs, enzymes encoded by HSD3B genes (Fig. 9) in prostate cancer 3α-HSDs, UDP-glucuronosyltransferases etc. (Fig. 9). We susceptibility, a recent study screened a panel of DNA should also take into account the potential gene–gene and samples collected from 96 men with or without prostate gene–environment interactions of low or moderate penetr- cancer for sequence variants in the putative promoter region, ance genes, which likely cumulate to contribute to an overall exons, exon-intron junctions, and 3′-untranslated region of prostate cancer risk and to disease characteristics. In this HSD3B1 and HSD3B2 genes by direct sequencing (Chang et regard, when several SNPs occur in the same gene/locus or al. 2002a). Four of the 11 single nucleotide polymorphisms a chromosomal region, it will be crucial to establish in each (SNPs) were informative. These four SNPs were further individual tested using haplotyping, if possible, an exhaustive genotyped in a total of 159 hereditary prostate cancer pro- genomic profiling for all selected sequence variants. This will bands (Johns-Hopkins set), 245 sporadic prostate cancer allow a better estimate of their global contribution in various cases, and 222 unaffected controls. Although a weak associa- mechanisms involved in prostate cancer, such as in pathways www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021253 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes involved in the fine control of intracellular androgen Collaborators. The EU Biomed Collaborators. Journal of bioavailability and in the signal transduction of their cell- Medical Genetics 37 947–949. specific action. Beilin J, Ball EM, Favaloro JM & Zajac JD 2000 Effect of the androgen receptor CAG repeat polymorphism on transcriptional In parallel, characterization of gene expression profiles activity: specificity in prostate and non-prostate cell lines. that molecularly distinguish prostatic neoplasms may identify Journal of Molecular Endocrinology 2000 25 85–96. new target candidate genes involved in prostate cancer risk Beilin J, Harewood L, Frydenberg M, Mameghan H, Martyres RF, in addition to elucidating useful new clinical biomarkers Farish SJ, Yue C, Deam DR, Byron KA & Zajac JD 2001 A leading to an improved classification of prostate cancer. This case-control study of the androgen receptor gene CAG repeat will be of special interest if such a signature may help to polymorphism in Australian prostate carcinoma subjects. Cancer distinguish hereditary versus sporadic prostate carcinomas, 2001 92 941–949. Berry R, Schaid DJ, Smith JR, French AJ, Schroeder JJ, as has recently been achieved for breast cancer (van ‘t Veer McDonnell SK, Peterson BJ, Wang ZY, Carpten JD, Roberts et al. 2002). Moreover, the integration of gene expression SG, Tester DJ, Blute ML, Trent JM & Thibodeau SN 2000a data with the knowledge of the human genome sequence as Linkage analyses at the chromosome 1 loci 1q24–25 (HPC1), well as those of experimental models will be useful to accel- 1q42.2–43 (PCAP), and 1p36 (CAPB) in families with erate positional cloning approaches that are often hampered hereditary prostate cancer. American Journal of Human Genetics by incomplete phenotypic penetrance, small pedigree size 66 539–546. and limited access to DNA samples from affected indi- Berry R, Schroeder JJ, French AJ, McDonnell SK, Peterson BJ, viduals. The elucidation of transcriptomes of normal and Cunningham JM, Thibodeau SN & Schaid DJ 2000b Evidence for a prostate cancer-susceptibility locus on chromosome 20. tumoral prostate tissues can effectively focus efforts on a American Journal of Human Genetics 67 82–91. manageable subset of genes, thus facilitating the candidate Berthon P, Valeri A, Cohen-Akenine A, Drelon E, Paiss T, Wohr gene approaches, especially when genetic maps cover a G, Latil A, Millasseau P, Mellah I, Cohen N et al. 1998 broad chromosomal region as frequently observed for pros- Predisposing gene for early-onset prostate cancer, localized on tate cancer susceptibility loci. Integrating both expression chromosome 1q42.2–43. American Journal of Human Genetics and functional information may further speed gene discovery 62 1416–1424. via the candidate gene approach, as recently suggested for Blackshaw S, Fraioli RE, Furukawa T & Cepko CL 2001 retinal disease genes (Blackshaw et al. 2001). Comprehensive analysis of photoreceptor gene expression and the identification of candidate retinal disease genes. Cell 107 We are at the very early phases of functional genomics 579–589. and rapid advances are also anticipated in more quantitative Bochum S, Paiss T, Vogel W, Herkommer K, Hautmann R & proteomics and bioinformatics, which should dramatically Haeussler J 2002 Confirmation of the prostate cancer increase the likelihood of finding clinically relevant candi- susceptibility locus HPCX in a set of 104 German prostate date genes, gene clusters and signaling pathways that will cancer families. Prostate 52 12–19. translate into better diagnostic or more targeted therapeutic Bock CH, Cunningham JM, McDonnell SK, Schaid DJ, Peterson strategies for men with prostate cancer. BJ, Pavlic RJ, Schroeder JJ, Klein J, French AJ, Marks A, Thibodeau SN, Lange EM & Cooney KA 2001 Analysis of the prostate cancer-susceptibility locus HPC20 in 172 families affected by prostate cancer. American Journal of Human Acknowledgments Genetics 68 795–801. Bostwick DG 1995 High grade prostatic intraepithelial neoplasia. J. Simard is chairholder of the Canada Research Chair in The most likely precursor of prostate cancer. Cancer Oncogenetics funded by the Canada Research Chair Pro- Epidemiology, Biomarkers and Prevention 75 1823–1836. gram. Bostwick DG & Dundore PA 1997 Biopsy Pathology of the Prostate. London: Chapman & Hall. Bratt O, Borg A, Kristoffersson U, Lundgren R, Zhang QX & References Olsson H 1999 CAG repeat length in the androgen receptor gene is related to age at diagnosis of prostate cancer and response to Abkevich V, Camp NJ, Gutin A, Farnham JM, Cannon-Albright endocrine therapy, but not to prostate cancer risk. British L & Thomas A 2001 A robust multipoint linkage statistic (tlod) Journal of Cancer 81 672–676. for mapping complex trait loci. Genetic Epidemiology 2001 21 Camp NJ & Tavtigian SV 2002 Meta analysis of associations of S492–S497. the Ser217Leu and Ala541Thr variants in ELAC2 (HPC2) and Anderson DE & Badzioch MD 1992 Breast cancer risks in prostate cancer. American Journal of Human Genetics 71 1475– relatives of male breast cancer patients. Journal of the National 1478. Cancer Institute 1992 84 1114–1117. Cancel-Tassin G, Latil A, Valeri A, Mangin P, Fournier G, Badzioch M, Eeles R, Leblanc G, Foulkes WD, Giles G, Edwards Berthon P & Cussenot O 2001a PCAP is the major known S, Goldgar D, Hopper JL, Bishop DT, Moller P, Heimdal K, prostate cancer predisposing locus in families from south and Easton D & Simard J 2000 Suggestive evidence for a site west Europe. European Journal of Human Genetics 9 135–142. specific prostate cancer gene on chromosome 1p36. The CRC/ Cancel-Tassin G, Latil A, Valeri A, Guillaume E, Mangin P, BPG UK Familial Prostate Cancer Study Coordinators and Fournier G, Berthon P & Cussenot O 2001b No evidence of

254 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

linkage to HPC20 on chromosome 20q13 in hereditary prostate Australian families affected by prostate cancer. American cancer. International Journal of Cancer 93 455–456. Journal of Human Genetics 68 1207–1218. Cannon L, Bishop DT, Skolnick M, Hunt S, Lyon JL & Smart CR Edwards SM, Badzioch MD, Minter R, Hamoudi R, Collins N, 1982 Genetic epidemiology of prostate cancer in the Utah Ardern-Jones A, Dowe A, Osborne S, Kelly J, Shearer R et al. mormon genealogy. Cancer Surveys 1 47–69. 1999 Androgen receptor polymorphisms: association with Cannon-Albright LA, Goldgar DE, Meyer LJ, Lewis CM, prostate cancer risk, relapse and overall survival. International Anderson DE, Fountain JW, Hegi ME, Wiseman RW, Petty EM, Journal of Cancer 84 458–465. Bale AE et al. 1992 Assignment of a locus for familial Edwards SM, Kote-Jarai Z, Meitz J, Hamoudi R, Hope Q, Osin P, melanoma MLM, to chromosome 9p13-p22. Science 258 1148– Jackson R, Southgate C, Singh R, Falconer A et al. 2003 Two 1152. percent of men with early-onset prostate cancer harbor germline Carpten J, Nupponen N, Isaacs S, Sood R, Robbins C, Xu J, mutations in the BRCA2 gene. American Journal of Human Faruque M, Moses T, Ewing C, Gillanders E et al. 2002 Genetics 72 1–12. Germline mutations in the ribonuclease L gene in families Eeles R 1999 The UK Familial Prostate Study Co-ordinating showing linkage with HPC1. Nature Genetics 30 181–184. Group and The CRC/BPG UK Familial Prostate Study Carter BS, Beaty TH, Steinberg GD, Childs B & Walsh PC 1992 Collaborators. Prostate Cancer and Prostatic Diseases 2 9–15. Mendelian inheritance of familial prostate cancer. PNAS 89 Eeles R, Cannon-Albright LA, Ponder SAJ, Easton DF& Horwich 3367–3371. A 1996 Familial prostate cancer and its management. In Genetic Casey G, Neville PJ, Plummer SJ, Xiang Y, Krumroy LM, Klein Predisposition to Cancer, Eds RA Eeles and A Horwich pp EA, Catalona WJ, Nupponen N, Carpten JD, Trent JM, 320–332. New York, USA: Chapman and Hall Medical. Silverman RH & Witte JS 2002 RNASEL Arg462Gln variant is Eeles RA, Durocher F, Edwards S, Teare D, Badzioch M, implicated in up to 13% of prostate cancer cases. Nature Hamoudi R, Gill S, Biggs P, Dearnaley D, Ardern-Jones A et al. Genetics 4 581–583. 1998 Linkage analysis of chromosome 1q markers in 136 Chamberlain NL, Driver ED & Miesfeld RL 1994 The length and prostate cancer families. The Cancer Research Campaign/British location of CAG trinucleotide repeats in the androgen receptor Prostate Group UK Familial Prostate Cancer Study N-terminal domain affect transactivation function. Nucleic Acids Collaborators. American Journal of Human Genetics 62 653– Research 22 3181–3186. 658. Chang B, Zheng SL, Isaacs SD, Wiley KE, Carpten JD, Hawkins Ekman P 1999 Genetic and environmental factors in prostate GA, Bleecker ER, Walsh PC, Trent JM, Meyers DA, Isaacs cancer genesis: identifying high-risk cohorts. European Urology WB & Xu J 2001 Linkage and association of CYP17 gene in 35 362–369. hereditary and sporadic prostate cancer. International Journal of El-Alfy M, Pelletier G, Hermo LS & Labrie F2000 Unique Cancer 95 354–359. features of the basal cells of human prostate epithelium. Chang BL, Zheng SL, Hawkins GA, Isaacs SD, Wiley KE, Turner Microscope Research Techniques 51 436–446. A, Carpten JD, Bleecker ER, Walsh PC, Trent JM, Meyers DA, Ford D, Easton DF, Bishop DT, Narod SA & Goldgar DE 1994 Isaacs WB & Xu J 2002a Joint effect of HSD3B1 and HSD3B2 Risks of cancer in BRCA1-mutation carriers. Breast Cancer genes is associated with hereditary and sporadic prostate cancer Linkage Consortium. Lancet 343 692–695. susceptibility. Cancer Research 62 1784–1789. Gayther SA, de Foy KA, Harrington P, Pharoah P, Dunsmuir WD, Chang BL, Zheng SL, Hawkins GA, Isaacs SD, Wiley KE, Turner A, Carpten JD, Bleecker ER, Walsh PC, Trent JM, Meyers DA, Edwards SM, Gillett C, Ardern-Jones A, Dearnaley DP, Easton Isaacs WB & Xu J 2002b Polymorphic GGC repeats in the DF et al. 2000 The frequency of germ-line mutations in the androgen receptor gene are associated with hereditary and breast cancer predisposition genes BRCA1 and BRCA2 in sporadic prostate cancer risk. Human Genetics 110 122–129. familial prostate cancer. The Cancer Research Campaign/British Chen C, Lamharzi N, Weiss NS, Etzioni R, Dightman DA, Barnett Prostate Group United Kingdom Familial Prostate Cancer Study M, DiTommaso D & Goodman G 2002 Androgen receptor Collaborators. Cancer Research 60 4513–4518. polymorphisms and the incidence of prostate cancer. Cancer Gibbs M, Chakrabarti L, Stanford JL, Goode EL, Kolb S, Schuster Epidemiology, Biomarkers and Prevention 11 1033–1040. EF, Buckley VA, Shook M, Hood L, Jarvik GP & Ostrander EA Coetzee GA & Ross RK 1994 Re: Prostate cancer and the 1999a Analysis of chromosome 1q42.2–43 in 152 families with androgen receptor. Journal of the National Cancer Institute 86 high risk of prostate cancer. American Journal of Human 872–873. Genetics 64 1087–1095. Cooney KA, McCarthy JD, Lange E, Huang L, Miesfeldt S, Gibbs M, Stanford JL, McIndoe RA, Jarvik GP, Kolb S, Goode Montie JE, Oesterling JE, Sandler HM & Lange K 1997 Prostate EL, Chakrabarti L, Schuster EF, Buckley VA, Miller EL et al. cancer susceptibility locus on chromosome 1q: a confirmatory 1999b Evidence for a rare prostate cancer-susceptibility locus at study. Journal of the National Cancer Institute 89 955–959. chromosome 1p36. American Journal of Human Genetics 64 Correa-Cerro L, Wohr G, Haussler J, Berthon P, Drelon E, Mangin 776–787. P, Fournier G, Cussenot O, Kraus P, Just W, Paiss T, Cantu Gibbs M, Stanford JL, Jarvik GP, Janer M, Badzioch M, Peters JM & Vogel W 1999 (CAG)nCAA and GGN repeats in the MA, Goode EL, Kolb S, Chakrabarti L, Shook M et al. 2000 A human androgen receptor gene are not associated with prostate genomic scan of families with prostate cancer identifies multiple cancer in a French-German population. European Journal of regions of interest. American Journal of Human Genetics 67 Human Genetics 7 357–362. 100–109. Coughlin SS & Hall IJ 2002 A review of genetic polymorphisms Giovannucci E, Stampfer MJ, Krithivas K, Brown M, Dahl D, and prostate cancer risk. Annals of Epidemiology 12 182–196. Brufsky A, Talcott J, Hennekens CH & Kantoff PW 1997 The Cui J, Staples MP, Hopper JL, English DR, McCredie MR & Giles CAG repeat within the androgen receptor gene and its GG 2001 Segregation analyses of 1476 population-based relationship to prostate cancer. PNAS 94 3320–3323. www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021255 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

Goddard KA, Witte JS, Suarez BK, Catalona WJ & Olson JM Hsing AW, Chen C, Chokkalingam AP, Gao YT, Dightman DA, 2001 Model-free linkage analysis with covariates confirms Nguyen HT, Deng J, Cheng J, Sesterhenn IA, Mostofi FK, linkage of prostate cancer to chromosomes 1 and 4. American Stanczyk FZ & Reichardt JK 2001 Polymorphic markers in the Journal of Human Genetics 68 1197–1206. SRD5A2 gene and prostate cancer risk: a population-based Goode EL, Stanford JL, Chakrabarti L, Gibbs M, Kolb S, McIndoe case-control study. Cancer Epidemiology, Biomarkers and RA, Buckley VA, Schuster EF, Neal CL, Miller EL et al. 2000 Prevention 10 1077–1082. Linkage analysis of 150 high-risk prostate cancer families at Hubert A, Peretz T, Manor O, Kaduri L, Wienberg N, Lerer I, Sagi 1q24-25. Genetic Epidemiology 18 251–275. M & Abeliovich D 1999 The Jewish Ashkenazi founder Goode EL, Stanford JL, Peters MA, Janer M, Gibbs M, Kolb S, mutations in the BRCA1/BRCA2 genes are not found at an Badzioch MD, Hood L, Ostrander E A & Jarvik GP 2001 increased frequency in Ashkenazi patients with prostate cancer. Clinical characteristics of prostate cancer in an analysis of American Journal of Human Genetics 65 921–924. linkage to four putative susceptibility loci. Clinical Cancer Ingles SA, Ross RK, Yu MC, Irvine RA, La Pera G, Haile RW & Research 7 2739–2749. Coetzee GA 1997 Association of prostate cancer risk with Gro¨nberg H, Isaacs SD, Smith JR, Carpten JD, Bova GS, Freije D, genetic polymorphisms in vitamin D receptor and androgen Xu J, Meyers DA, Collins FS, Trent JM, Walsh PC & Isaacs receptor. Journal of the National Cancer Institute 89 166–170. WB 1997 Characteristics of prostate cancer in families Irvine RA, Yu MC, Ross RK & Coetzee GA 1995 The CAG and potentially linked to the hereditary prostate cancer 1 (HPC1) GGC microsatellites of the androgen receptor gene are in linkage locus. Journal of the American Medical Association 278 1251– disequilibrium in men with prostate cancer. Cancer Research 55 1255. 1937–1940. Gro¨nberg H, Ahman AK, Emanuelsson M, Bergh A, Damber JE & Irvine RA, Ma H, Yu MC, Ross RK, Stallcup MR & Coetzee GA Borg A 2001 BRCA2 mutation in a family with hereditary 2000 Inhibition of p160-mediated coactivation with increasing prostate cancer. Genes Chromosomes Cancer 30 299–301. androgen receptor polyglutamine length. Human Molecular Grossmann ME, Huang H & Tindall DJ 2001 Androgen receptor Genetics 9 267–274. signaling in androgen-refractory prostate cancer. Journal of the National Cancer Institute 93 1687–1697. Jaffe JM, Malkowicz SB, Walker AH, MacBride S, Peschel R, Gsur A, Bernhofer G, Hinteregger S, Haidinger G, Schatzl G, Tomaszewski J, Van Arsdalen K, Wein AJ & Rebbeck TR 2000 Madersbacher S, Marberger M, Vutuc C & Micksche M 2000 A Association of SRD5A2 genotype and pathological polymorphism in the CYP17 gene is associated with prostate characteristics of prostate tumors. Cancer Research 60 1626– cancer risk. International Journal of Cancer 87 434–437. 1630. Habuchi T, Liqing Z, Suzuki T, Sasaki R, Tsuchiya N, Tachiki H, Kazemi-Esfarjani P, Trifiro MA & Pinsky L 1995 Evidence for a Shimoda N, Satoh S, Sato K, Kakehi Y, Kamoto T, Ogawa O & repressive function of the long polyglutamine tract in the human Kato T 2000 Increased risk of prostate cancer and benign androgen receptor: possible pathogenetic relevance for the prostatic hyperplasia associated with a CYP17 gene (CAG)n-expanded neuronopathies. Human Molecular Genetics 4 polymorphism with a gene dosage effect. Cancer Research 60 523–527. 5710–5713. Korver W, Guevara C, Chen Y, Neuteboom S, Bookstein R, Hakimi JM, Schoenberg MP, Rondinelli RH, Piantadosi S & Tavtigian SV & Lees E 2003 The product of the candidate Barrack ER 1997 Androgen receptor variants with short prostate cancer susceptibility gene ELAC2 interacts with the g– glutamine or glycine repeats may identify unique subpopulations tubulin complex. International Journal of Cancer 104 283–288. of men with prostate cancer. Clinical Cancer Research 3 1599– Labrie F2003 Cancer of the prostate. In The UICC Manual of 1608. Clinical Oncology, 8th edition. Eds R Pollock, B O’Sullivan and Henshall SM, Quinn DI, Lee CS, Head DR, Golovsky D, Brenner A Nakao. Etobicoke Canada: John Wiley and Sons (In Press). PC, Delprado W, Stricker PD, Grygiel JJ & Sutherland RL 2001 Labrie F, Belanger A, Luu-The V, Labrie C, Simard J, Cusan L, Altered expression of androgen receptor in the malignant Gomez JL & Candas B 1998 DHEA and the intracrine formation epithelium and adjacent stroma is associated with early relapse of androgens and estrogens in peripheral target tissues: its role in prostate cancer. Cancer Research 61 423–427. during aging. Steroids 63 322–328. Hsieh CL, Oakley-Girvan I, Gallagher RP, Wu AH, Kolonel LN, Lander E & Kruglyak L 1995 Genetic dissection of complex traits: Teh CZ, Halpern J, West DW, Paffenbarger RS Jr & guidelines for interpreting and reporting linkage results. Nature Whittemore AS 1997 Re: prostate cancer susceptibility locus on Genetics 11 241–247. chromosome 1q: a confirmatory study. Journal of the National Lange EM, Chen H, Brierley K, Perrone EE, Bock CH, Gillanders Cancer Institute 89 1893–1894. E, Ray ME & Cooney KA 1999 Linkage analysis of 153 Hsieh CL, Oakley-Girvan I, Balise RR, Halpern J, Gallagher RP, prostate cancer families over a 30-cM region containing the Wu AH, Kolonel LN, O’Brien LE, Lin IG, Van Den Berg DJ et Clinical Cancer Research 5 al. 2001 A genome screen of families with multiple cases of putative susceptibility locus HPCX. prostate cancer: evidence of genetic heterogeneity. American 4013–4020. Journal of Human Genetics 69 148–158. Lange EM, Chen H, Brierley K, Livermore H, Wojno KJ, Hsing AW, Gao YT, Wu G, Wang X, Deng J, Chen YL, Langefeld CD, Lange K & Cooney KA 2000 The polymorphic Sesterhenn IA, Mostofi FK, Benichou J & Chang C 2000a exon 1 androgen receptor CAG repeat in men with a potential Polymorphic CAG and GGN repeat lengths in the androgen inherited predisposition to prostate cancer. Cancer Epidemiology, receptor gene and prostate cancer risk: a population-based Biomarkers and Prevention 9 439–442. case-control study in China. Cancer Research 60 5111–5116. Latil AG, Azzouzi R, Cancel GS, Guillaume EC, Cochan-Priollet Hsing AW, Tsao L & Devesa SS 2000b International trends and B, Berthon PL & Cussenot O 2001 Prostate carcinoma risk and patterns of prostate cancer incidence and mortality. International allelic variants of genes involved in androgen biosynthesis and Journal of Cancer 85 60–67. metabolism pathways. Cancer 92 1130–1137.

256 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

Lehrer S, Fodor F, Stock RG, Stone NN, Eng C, Song HK & 1994 A strong candidate for the breast and ovarian cancer McGovern M 1998 Absence of 185delAG mutation of the susceptibility gene BRCA1. Science 266 66–71. BRCA1 gene and 6174delT mutation of the BRCA2 gene in Miller EA, Stanford JL, Hsu L, Noonan E & Ostrander EA 2001 Ashkenazi Jewish men with prostate cancer. British Journal of Polymorphic repeats in the androgen receptor gene in high-risk Cancer 78 771–773. sibships. Prostate 48 200-205. Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, Mononen N, Ikonen T, Syrjakoski K, Matikainen M, Schleutker J, Koskenvuo M, Pukkala E, Skytthe A & Hemminki K 2000 Tammela TL, Koivisto PA & Kallioniemi OP 2001 A missense Environmental and heritable factors in the causation of cancer – substitution A49T in the steroid 5-alpha-reductase gene analyses of cohorts of twins from Sweden, Denmark, and (SRD5A2) is not associated with prostate cancer in Finland. Finland. New England Journal of Medicine 343 78–85. British Journal of Cancer 84 1344–1347. Lillienfeld AM & Lillienfeld DE 1980 Foundations of Monroe KR, Yu MC, Kolonel LN, Coetzee GA, Wilkens LR, Ross Epidemiology. New York: Oxford University Press. RK & Henderson BE 1995 Evidence of an X-linked or recessive Lin C & Miller WL 2001 The T/C allelic polymorphism at genetic component to prostate cancer risk. Nature Medicine 1 nucleotide + 27 of the human CYP17 gene is not associated 827–829. with altered transcriptional activity. In The Endocrine Society’s Montgomery JS, Price DK & Figg WD 2001 The androgen 83rd Annual Meeting, p P2-1. Denver, Colorado. receptor gene and its influence on the development and Linja MJ, Savinainen KJ, Saramaki OR, Tammela TL, Vessella progression of prostate cancer. Journal of Pathology 195 138– RL & Visakorpi T 2001 Amplification and overexpression of 146. androgen receptor gene in hormone-refractory prostate cancer. Morganti G, Gianferrari L, Cresseri A, Arrigoni G & Lovati G Cancer Research 61 3550–3555. 1956 Recherches clinico-statistiques et ge´ne´tiques sur les Lunn RM, Bell DA, Mohler JL & Taylor JA 1999 Prostate cancer ne´oplasie de la prostate. Acta Genetica Statistica 6 304–305. risk and polymorphism in 17 hydroxylase (CYP17) and steroid Morissette J, Durocher F, Leblanc JF, Normand T, Labrie F & reductase (SRD5A2). Carcinogenesis 20 1727–1731. Simard J 1996 Genetic linkage mapping of the human steroid 5 McIndoe RA, Stanford JL, Gibbs M, Jarvik GP, Brandzel S, Neal alpha-reductase type 2 gene (SRD5A2) close to D2S352 on Λ CL, Li S, Gammack JT, Gay AA, Goode EL, Hood L & chromosome region 2p23 p22. Cytogenetics and Cell Genetics 73 304–307. Ostrander EA 1997 Linkage analysis of 49 high-risk families Narod SA, Dupont A, Cusan L, Diamond P, Gomez JL, Suburu does not support a common familial prostate cancer- R & Labrie F1995 The impact of family history on early susceptibility gene at 1q24-25. American Journal of Human detection of prostate cancer. Nature Medicine 1 99–101. Genetics 61 347–353. Nastiuk KL, Mansukhani M, Terry MB, Kularatne P, Rubin MA, McNeal JE, Villers A, Redwine EA, Freiha FS & Stamey TA 1991 Melamed J, Gammon MD, Ittmann M & Krolewski JJ 1999 Microcarcinoma in the prostate: its association with duct-acinar Common mutations in BRCA1 and BRCA2 do not contribute to dysplasia. Human Pathology 22 644–652. early prostate cancer in Jewish men. Prostate 40 172–177. Makridakis NM & Reichardt JK 2001 Molecular epidemiology of Nedelcheva Kristensen V, Haraldsen EK, Anderson KB, Lonning hormone-metabolic loci in prostate cancer. Epidemiology PE, Erikstein B, Karesen R, Gabrielsen OS & Borresen-Dale AL Reviews 23 24–29. 1999 CYP17 and breast cancer risk: the polymorphism in the 5′ Makridakis N, Ross RK, Pike MC, Chang L, Stanczyk FZ, Kolonel flanking area of the gene does not influence binding to Sp-1. LN, Shi CY, Yu MC, Henderson BE & Reichardt JK 1997 A Cancer Research 59 2825–2828. prevalent missense substitution that modulates activity of Neuhausen SL, Farnham JM, Kort E, Tavtigian SV, Skolnick prostatic steroid 5alpha-reductase. Cancer Research 57 1020– MH & Cannon-Albright LA 1999 Prostate cancer susceptibility 1022. locus HPC1 in Utah high-risk pedigrees. Human Molecular Makridakis NM, Ross RK, Pike MC, Crocitto LE, Kolonel LN, Genetics 8 2437–2442. Pearce CL, Henderson BE & Reichardt JK 1999 Association of Nwosu V, Carpten J, Trent JM & Sheridan R 2001 Heterogeneity mis-sense substitution in SRD5A2 gene with prostate cancer in of genetic alterations in prostate cancer: evidence of the complex African-American and Hispanic men in Los Angeles, USA. nature of the disease. Human Molecular Genetics 10 2313–2318. Lancet 354 975–978. Ostrander EA & Stanford JL 2000 Genetics of prostate cancer: too Makridakis NM, di Salle E & Reichardt JK 2000 Biochemical and many loci, too few genes. American Journal of Human Genetics pharmacogenetic dissection of human steroid 5-alpha-reductase 67 1367–1375. type II. Pharmacogenetics 10 407–413. Peters MA, Jarvik GP, Janer M, Chakrabarti L, Kolb S, Goode EL, Margiotti K, Sangiuolo F, De Luca A, Froio F, Pearce CL, Gibbs M, DuBois CC, Schuster EF, Hood L, Ostrander EA & Ricci-Barbini V, Micali F, Bonafe M, Franceschi C, Stanford JL 2001 Genetic linkage analysis of prostate cancer Dallapiccola B, Novelli G & Reichardt JK 2000 Evidence for an families to Xq27-28. Human Heredity 51 107–113. association between the SRD5A2 (type II steroid 5 Platz EA, Giovannucci E, Dahl DM, Krithivas K, Hennekens CH, alpha-reductase) locus and prostate cancer in Italian patients. Brown M, Stampfer MJ & Kantoff PW 1998 The androgen Disease Markers 16 147–150. receptor gene GGN microsatellite and prostate cancer risk. Meitz JC, Edwards SM, Easton DF, Murkin A, Ardern-Jones A, Cancer Epidemiology, Biomarkers and Prevention 7 379–384. Jackson RA, Williams S, Dearnaley DP, Stratton MR, Houlston Rebbeck TR, Walker AH, Zeigler-Johnson C, Weisburg S, Martin RS & Eeles RA 2002 HPC2/ELAC2 polymorphisms and AM, Nathanson KL, Wein AJ & Malkowicz SB 2000 prostate cancer risk: analysis by age of onset of disease. British Association of HPC2/ELAC2 genotypes and prostate cancer. Journal of Cancer 87 905–908. American Journal of Human Genetics 67 1014–1019. Miki Y, Swensen J, Shattuck-Eidens D, Futreal PA, Harshman K, Reichardt JK, Makridakis N, Henderson BE, Yu MC, Pike MC & Tavtigian S, Liu Q, Cochran C, Bennett LM, Ding W et al. Ross RK 1995 Genetic variability of the human SRD5A2 gene: www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021257 04:18:26AM via free access Simard et al.: Prostate cancer susceptibility genes

implications for prostate cancer risk. Cancer Research 55 3973– predisposing to prostate cancer – do they exist? Prostate Cancer 3975. and Prostatic Diseases 3 241–247. Rennert H, Bercovich D, Hubert A, Abeliovich D, Rozovsky U, Slager SL, Schaid DJ, Cunningham JM, McDonnell SK, French A, Bar-Shira A, Soloviov S, Schreiber L, Matzkin H, Rennert G et Peterson BJ, Marks AF, Hebbring S, Koch B, Anderson S & al. 2002 A novel founder mutation in the RNASEL gene, Thibodeau SN 2002 Linkage analysis on prostate cancer 471delAAAG, is associated with prostate cancer in Ashkenazi agressiveness. American Journal of Human Genetics 71 (Suppl) Jews. American Journal of Human Genetics 71 981–984. S247. Rokman A, Ikonen T, Mononen N, Autio V, Matikainen MP, Smith JR, Freije D, Carpten JD, Gro¨nberg H, Xu J, Isaacs SD, Koivisto PA, Tammela TL, Kallioniemi OP & Schleutker J 2001 Brownstein MJ, Bova GS, Guo H, Bujnovszky P et al. 1996 ELAC2/HPC2 involvement in hereditary and sporadic prostate Major susceptibility locus for prostate cancer on chromosome 1 cancer. Cancer Research 61 6038–6041. suggested by a genome-wide search. Science 274 1371–1374. Rokman A, Ikonen T, Seppala EH, Nupponen N, Autio V, Stanford JL, Just JJ, Gibbs M, Wicklund KG, Neal CL, Mononen N, Bailey-Wilson J, Trent J, Carpten J, Matikainen Blumenstein BA & Ostrander EA 1997 Polymorphic repeats in MP et al. 2002 Germline alterations of the RNASEL gene, a the androgen receptor gene: molecular markers of prostate candidate HPC1 gene at 1q25, in patients and families with cancer risk. Cancer Research 57 1194–1198. prostate cancer. American Journal of Human Genetics 70 1299– Steinberg GD, Carter BS, Beaty TH, Childs B & Walsh PC 1990 1304. Family history and the risk of prostate cancer. Prostate 17 337– Ross RK, Bernstein L, Lobo RA, Shimizu H, Stanczyk FZ, Pike 347. MC & Henderson BE 1992 5-Alpha-reductase activity and risk Suarez BK, Lin J, Burmester JK, Broman KW, Weber JL, of prostate cancer among Japanese and US white and black Banerjee TK, Goddard KA, Witte JS, Elston RC & Catalona WJ males. Lancet 339 887–889. 2000a A genome screen of multiplex sibships with prostate Ross RK, Coetzee GA, Pearce CL, Reichardt JK, Bretsky P, cancer. American Journal of Human Genetics 66 933–944. Kolonel LN, Henderson BE, Lander E, Altshuler D & Daley G Suarez BK, Lin J, Witte JS, Conti DV, Resnick MI, Klein EA, 1999 Androgen metabolism and prostate cancer: establishing a Burmester JK, Vaske DA, Banerjee TK & Catalona WJ 2000b model of genetic susceptibility. European Urology 35 355–361. Replication linkage study for prostate cancer susceptibility genes. Prostate 45 106–114. Schaid DJ, McDonnell SK, Blute ML & Thibodeau SN 1998 Suarez BK, Gerhard DS, Lin J, Haberer B, Nguyen L, Kesterson Evidence for autosomal dominant inheritance of prostate cancer. NK & Catalona WJ 2001 Polymorphisms in the prostate cancer American Journal of Human Genetics 62 1425–1438. susceptibility gene HPC2/ELAC2 in multiplex families and Schaid DJ, McDonnell SK & Thibodeau SN 2001 Regression healthy controls. Cancer Research 61 4982–4984. models for linkage heterogeneity applied to familial prostate Tavtigian SV, Simard J, Teng DH, Abtin V, Baumgard M, Beck cancer. American Journal of Human Genetics 68 1189–1196. A, Camp NJ, Carillo AR, Chen Y, Dayananth P et al. 2001 A Schiffer S, Rosch S & Marchfelder A 2002 Assigning a function to candidate prostate cancer susceptibility gene at chromosome 17p. a conserved group of proteins: the tRNA 3′-processing enzymes. Nature Genetics 27 172–180. EMBO Journal 21 2769–2777. The Breast Cancer Linkage Consortium 1999 Cancer risks in Schleutker J, Matikainen M, Smith J, Koivisto P, Baffoe-Bonnie A, BRCA2 mutation carriers. The Breast Cancer Linkage Kainu T, Gillanders E, Sankila R, Pukkala E, Carpten J et al. Consortium. Journal of the National Cancer Institute 91 1310– 2000 A genetic epidemiological study of hereditary prostate 1316. cancer (HPC) in Finland: frequent HPCX linkage in families Thompson D & Easton D 2001 Variation in cancer risks, by with late-onset disease. Clinical Cancer Research 6 4810–4815. mutation position, in BRCA2 mutation carriers. American Shea PR, Ferrell RE, Patrick AL, Kuller LH & Bunker CH 2002 Journal of Human Genetics 68 410–419. ELAC2 and prostate cancer risk in Afro-Caribbeans of Tobago. Thorlacius S, Olafsdottir G, Tryggvadottir L, Neuhausen S, Human Genetics 111 398–400. Jonasson JG, Tavtigian SV, Tulinius H, Ogmundsdottir HM & Sigurdsson S, Thorlacius S, Tomasson J, Tryggvadottir L, Eyfjord JE 1996 A single BRCA2 mutation in male and female Benediktsdottir K, Eyfjord JE & Jonsson E 1997 BRCA2 breast cancer families from Iceland with varied cancer mutation in Icelandic prostate cancer patients. Journal of phenotypes. Nature Genetics 13 117–119. Molecular Medicine 75 758–761. Thorlacius S, Struewing JP, Hartge P, Olafsdottir GH, Sigvaldason Simard J, Dumont M, Soucy P & Labrie F2002 a Perspective: H, Tryggvadottir L, Wacholder S, Tulinius H & Eyfjord JE 1998 prostate cancer susceptibility genes. Endocrinology 143 2029– Population-based study of risk of breast cancer in carriers of 2040. BRCA2 mutation. Lancet 352 1337–1339. Simard J, Moisan AM & Morel Y 2002b Congenital adrenal Tulinius H, Egilsson V, Olafsdottir GH & Sigvaldason H 1992 hyperplasia due to 3beta-hydroxysteroid dehydrogenase/ Risk of prostate, ovarian, and endometrial cancer among delta5-delta4 isomerase deficiency. Seminars in Reproductive relatives of women with breast cancer. British Medical Journal Medicine 20 255–276. 305 855–857. Sinclair CS, Berry R, Schaid D, Thibodeau SN & Couch FJ 2000 U M, Miyashita T, Ohtsuka Y, Okamura-Oho Y, Shikama Y & BRCA1 and BRCA2 have a limited role in familial prostate Yamada M 2001 Extended polyglutamine selectively interacts cancer. Cancer Research 60 1371–1375. with caspase-8 and -10 in nuclear aggregates. Cell Death and Singh RER, Durocher F, Simard J, Edwards S, Badzioch M, Differentiation 8 377–386. Kote-Jarai Z, Teare D, Ford D, Dearnaley D, Ardern-Jones A, van ‘t Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao Murkin A, Dowe A, Shearer R, Kelly J, The CRC/BPG UK M, Peterse HL, van der Kooy K, Marton MJ, Witteveen AT et Familial Prostate Cancer Study Collaborators, Labrie F, Easton al. 2002 Gene expression profiling predicts clinical outcome of D, Narod SA, Tonin PN & Foulkes WD 2000 High risk genes breast cancer. Nature 415 530–536.

258 Downloaded from Bioscientifica.comwww.endocrinology.org at 09/30/2021 04:18:26AM via free access Endocrine-Related Cancer (2003) 10 225–259

Vesprini D, Nam RK, Trachtenberg J, Jewett MA, Tavtigian SV, Xu J, Meyers D, Freije D, Isaacs S, Wiley K, Nusskern D, Ewing Emami M, Ho M, Toi A & Narod SA 2001 HPC2 variants and C, Wilkens E, Bujnovszky P, Bova GS et al. 1998 Evidence for screen-detected prostate cancer. American Journal of Human a prostate cancer susceptibility locus on the X chromosome. Genetics 68 912–917. Nature Genetics 20 175–179. Vogel A, Schilling O, Niecke M, Bettmer J & Meyer-Klaucke W Xu J, Zheng SL, Carpten JD, Nupponen NN, Robbins CM, Mestre 2002 ElaC encodes a novel binuclear zinc phosphodiesterase. J, Moses TY, Faith DA, Kelly BD, Isaacs SD et al. 2001a Journal of Biological Chemistry 277 29078–29085. Evaluation of linkage and association of HPC2/ELAC2 in Wadelius M, Andersson AO, Johansson JE, Wadelius C & Rane E patients with familial or sporadic prostate cancer. American 1999 Prostate cancer associated with CYP17 genotype. Journal of Human Genetics 68 901–911. Pharmacogenetics 9 635–639. Xu J, Zheng SL, Chang B, Smith JR, Carpten JD, Stine OC, Isaacs et al b Wang L, McDonnell SK, Elkins DA, Slager SL, Christensen E, SD, Wiley KE, Henning L, Ewing C . 2001 Linkage of prostate cancer susceptibility loci to chromosome 1. Human Marks AF, Cunningham JM, Peterson BJ, Jacobsen SJ, Cerhan Genetics 108 335–345. JR, Blute ML, Schaid DJ & Thibodeau SN 2001 Role of HPC2/ Xu J, Zheng SL, Hawkins GA, Faith DA, Kelly B, Isaacs SD, ELAC2 in hereditary prostate cancer. Cancer Research 61 6494– Wiley KE, Chang B, Ewing CM, Bujnovszky P et al. 2001c 6499. Linkage and association studies of prostate cancer susceptibility: Wang L, McDonnell SK, Elkins DA, Slager SL, Christensen E, evidence for linkage at 8p22-23. American Journal of Human Marks AF, Cunningham JM, Peterson BJ, Jacobsen SJ, Cerhan Genetics 69 341–350. JR, Blute ML, Schaid DJ & Thibodeau SN 2002 Analysis of the Xu J, Zheng SL, Komiya A, Mychaleckyj JC, Isaacs SD, Hu JJ, RNASEL gene in familial and sporadic prostate cancer. Sterling D, Lange EM, Hawkins GA, Turner A et al. 2002 American Journal of Human Genetics 71 116–123. Germline mutations and sequence variants of the macrophage Whittemore AS, Lin IG, Oakley-Girvan I, Gallagher RP, Halpern scavenger receptor 1 gene are associated with prostate cancer J, Kolonel LN, Wu AH & Hsieh CL 1999 No evidence of risk. Nature Genetics 32 321–325. linkage for chromosome 1q42.2–43 in prostate cancer. American Xu J, Zheng SL, Komiya A, Mychaleckyj JC, Isaacs SD, Chang B, Journal of Human Genetics 65 254–256. Turner AR, Ewing CM, Wiley KE, Hawkins GA et al. 2003 Wilkens EP, Freije D, Xu J, Nusskern DR, Suzuki H, Isaacs SD, Common sequence variants of the macrophage scavenger Wiley K, Bujnovsky P, Meyers DA, Walsh PC & Isaacs WB receptor 1 gene are associated with prostate cancer risk. 1999 No evidence for a role of BRCA1 or BRCA2 mutations in American Journal of Human Genetics 72 208–212. Ashkenazi Jewish families with hereditary prostate cancer. Xue W, Irvine RA, Yu MC, Ross RK, Coetzee GA & Ingles SA Prostate 39 280–284. 2000 Susceptibility to prostate cancer: interaction between Witte JS, Goddard KA, Conti DV, Elston RC, Lin J, Suarez BK, genotypes at the androgen receptor and prostate-specific antigen Broman KW, Burmester JK, Weber JL & Catalona WJ 2000 loci. Cancer Research 60 839–841. Yamada Y, Watanabe M, Murata M, Yamanaka M, Kubota Y, Ito Genomewide scan for prostate cancer-aggressiveness loci. H, Katoh T, Kawamura J, Yatani R & Shiraishi T 2001 Impact American Journal of Human Genetics 67 92–99. of genetic polymorphisms of 17-hydroxylase cytochrome P-450 Wooster R, Neuhausen SL, Mangion J, Quirk Y, Ford D, Collins (CYP17) and steroid 5alpha-reductase type II (SRD5A2) genes N, Nguyen K, Seal S, Tran T, Averill D et al. 1994 Localization on prostate cancer risk among the Japanese population. of a breast cancer susceptibility gene, BRCA2, to chromosome International Journal of Cancer 92 683–686. 13q12-13. Science 265 2088–2090. Zheng SL, Xu J, Isaacs SD, Wiley K, Chang B, Bleecker ER, Xu J 2000 Combined analysis of hereditary prostate cancer linkage Walsh PC, Trent JM, Meyers DA & Isaacs WB 2001 Evidence to 1q24-25: results from 772 hereditary prostate cancer families for a prostate cancer linkage to chromosome 20 in 159 from the International Consortium for Prostate Cancer Genetics. hereditary prostate cancer families. Human Genetics 108 430– American Journal of Human Genetics 66 945–957. 435.

www.endocrinology.org Downloaded from Bioscientifica.com at 09/30/2021259 04:18:26AM via free access