Additional Disclosure The Genetics of Prostate Cancer: Clinical Implications William J. Catalona, MD Collaborator with decode genetics, Inc. Non-paid consultant with no financial interest or support Northwestern University Feinberg School of Medicine Robert H. Lurie Comprehensive Cancer Center Chicago, Illinois Facts Cancer is caused by interplay of genetic and environmental factors The contribution of genetic factors varies between tumors Shared genetic factors may increase the risk for more than 1 type of cancer Early Family Linkage Studies Prostate cancer clusters in families Most segregation analyses suggested hereditary prostate cancer could be best explained by at least 1 rare dominant susceptibility gene, but this has not yet proven to be the case 1
EphB2 RNAseL PCaP Early linkage studies of PCa families postulated the existence of several possible susceptibility genes HPC1p ELAC2 MSR1 CYP3A gene cluster VDR Early Targeted Studies However, further targeted studies and genome-wide scans revealed difficulties in replication Confirmatory studies have yielded mixed results HPC20 HPCX UPM-3 (Now called PCA3) Clinical Studies Based on Candidate Genes A gene that is 30-100-fold upregulated in prostate cancer Can be detected in prostatic fluid samples Theoretically, it could be used like urinary cytology (i.e., could detect cancer without performing a biopsy) Practically it might assist in selecting patients for repeat biopsy 2
PCA3 Score vs Positive Biopsy PCA3 correlates with the probability of a positive biopsy and is independent of prostate volume, PSA, and the number of prior biopsies PCA3 vs PSA, DRE, TRUS Volume PCA3 and Tumor Volume PCA3 score appears to stratify men by tumor volume and Gleason score and may help select men with low volume, low grade PCa 3
PCA3: Tumor Volume and Gleason Sum PCA3 in post-dre urine of PCa patients correlated with pathologic findings of extracapsular tumor extension Multi-Center Validation EPCA-2 Nuclear protein specifically expressed in prostate cancer ELISA for 2 epitopes (2.19 and 2.22) In serum samples from patients, shows results that are complementary to PSA Identifies men with organ-confined and non-organ confined prostate cancer and prostate cancer vs non-prostate cancer. Leman ES et al J Urol 175:275, abstract 852 4
EPCA-2 in Prostate Cancer and Controls 5
TMPRSS2-ETS Gene Fusion Fusion of TMPRSS2 and ETS Transcription Factor Genes in Prostate Cancer ERG oncogenic transcription factor TMPRSS2 androgen response gene ETS1 oncogenic transcription factor Transformation signaling in the presence of androgen Contemporary Strategies in Active Surveillance: erg Fusion Effect on Prostate Cancer Mortality Urinary Test for Prostate Cancer FUSION STATUS Cumulative Incidence Ratio: 3.644 (p-value=0.0044, 95%CI=(1.497, 8.870)) M. Rubin et al, Oncogene,, 2007 Combination of TMPRSS2- ERG and PCA3 significantly improves sensitivity for PCa diagnosis 6
Multiple Markers in Urine Test Multiple Markers, Including TMPRSS2-ETS Fusions and PCA3 CpG Island Hypermethylation Hypermethylation at GSTP1 in serum predicts early PSA failure following RP. Hypermethylation at MDR1 was detected in localized PCa. CpG island hypermethylation at several loci was detected with advanced disease 7
Hypermethylation of Genes Gene analysis using PCR is a reliable method for detecting abnormal DNA methylation in voided urine following DRE or prostate needle biopsy. A serum test based on GSTP1 hypermethylation may help identify men with increased risk of PCa, despite negative prostate biopsy. Limitations of Linkage and Candidate Gene Studies Linkage analysis has sufficient power to detect highly penetrant genes but has weak power to find susceptibility genes of smallto-moderate effects Candidate gene studies do not test most of the genome Genetic Variants Prior to 2006, no common genetic variant accounted for a substantial proportion of cases In 2006, decode genetics discovered genetic variants that point to a complex genetic basis, involving multiple susceptibility loci, i.e., death by 1000 cuts 8
8q24 Discovery Paper Discovery of 8q24 Risk Alleles Discovered and validated 2 sequence variants on chromosome 8q24 (a frequent site of somatic amplification in prostate and colorectal cancer) that are significantly more common in prostate cancer cases than in controls This was the first universally replicable prostate cancer susceptibility locus and may account for 8% to 64% of prostate cancer cases Amundadottir et al., Nat Genet. 2006;38(6):652-8 Amundadottir et al. (2006) Nat Genetics 38: 652; Freedman et al. (2006) PNAS 103: 14068;Gudmundsson et al. (2007) Nat Genetics 39: 631;Haiman et al. (2007) Nat Genetics 39: 658;Yeager et al. (2007) Nat Genetics 39: 645;Robbins et al (2007) Genome Res 17:1717 Two 8q24 Alleles Associated with Aggressiveness and Race Validation Across Multiple Ethnic Cohorts Combined estimated OR of 1.62 (P = 2.7 10 11) for DG8S737-8 and an OR of 1.51 (P = 1.0 10 11) for rs1447295 allele A Slight association of DG8S737-8 with higher Gleason score. Replicated SNP rs1447295 in other MEC populations but not African Americans DG8S737-8 allele frequency of 16% in the African American population as compared to 6.6% in European/Caucasian population Could account for increased incidence rate of prostate cancer in African Americans Amundadottir et al., Nat Genet. 2006;38(6):652-8 Freedman et al., PNAS. 2006;103(38):14068-73 9
Other 8q24 Regions Subsequent studies revealed additional clusters of variants within 8q24 that independently confer risk for prostate cancer Tissue Specificity: The General Rule In most cases, but not all, the genetic variants are specific for a particular type of cancer Region 3 (rs10505477 & rs 6983267) also confer risk colorectal cancer 10
Known Genes in 8q24 The genomic regions of 8q24 are gene poor and have a high recombination rate The only reported gene is a retro-transposed gene, AF268618 (POU5F1) that encodes for a transcription factor, OCT4, proposed as a stem cell marker embryogenesis and in some tumors Two known genes nearby are c-myc and FAM 84B(NSE2), but no association with these genes has been found decode found no differences in expression in carriers and non-carriers of any of the above or of 1447295 (A), HapC, rs16901979 The distal end of a fragile site 8C (FRA8C) has been mapped to this region, as well as several integration sites of HPV REGION 4 4 Significant Regions of 8q24 Associated with rs4008482 Prostate Cancer Susceptibility Ch 8 breast colorectal & prostate bladder 128.0-129.0 Mb REGION 2 128.14-128.28 REGION 3 POU5F1P1 128.47-128.54 REGION 1 128.54-128.62 MYC SNPs rs13254738 rs6985361 rs16901979 (A) Broad11934905 rs6983267 (G) rs4242382 rs7000448 rs1447295 (A) rs10090154 rs17837688 Amundadottir et al. (2006) Nat Genetics 38: 652 Freedman et al. (2006) PNAS 103: 14068 Gudmundsson et al. (2007) Nat Genetics 39: 631 Haiman et al. (2007) Nat Genetics 39: 658 Yeager et al. (2007) Nat Genetics 39: 645 Robbins et al (2007) Genome Res 17:1717 Salinas et al (2008) CEBP 17:1203 17q: The Second Major Region UK Practical Reports 7 Risk Variants on Chromosome 3,6,7,10,11,19 and X Gudmundsson et al., Nat Genet. 2007b;39(8):977-83 Of the 53 SNPs significant at the P < 10-6 level, 20 were on 8q24 and 6 were on chromosome 17 11
Patient Profiling Prostate Cancer Susceptibility 29 risk variants now identified Patient Profiling for Prostate Cancer Risk Odds Ratio Increases with Number Of Risk Variants Cumulative Risk Zheng et al., NEJM. 2008;358(9):910-9 Zheng et al., NEJM. 2008;358(9):910-9 12
Adding Family History as a 6 th Risk Factor Cumulative Risk Zheng et al., NEJM. 2008;358(9):910-9 Relative risks compared to the general population vary and account for about half of the cases decode genetics new commercial ProCa test identifies 8 validated variants: 3 on 8q24, 2 on 17q, 1 on 2p15, 1 on 11, and 1 on X 13
Genetic Risk Alleles Improve Performance of PSA Isoforms Profiling for Aggressive Disease + Aggressive Disease Amundadottir et al. Nat Genet, 38: 652, 2006 Freedman et al. Proc Natl Acad Sci U S A, 103:14068, 2006 Suuriniemi et al. Cancer Epidem Biomarkers Prev, 16: 809, 2007 Wang et al. Cancer Res, 6: 2905, 2007 Cheng et al. Eur J Hum Genet, 16: 496, 2008 Helfand et al. J Urol, (in press), 2008 Gudmundsson et al. Nat Genet, 40: 281, 2008 Beebe-Dimmer et al. Int J Cancer, 122: 2876, 2008 - Aggressive Disease Zanke, et al. Nat Genet, 39: 989, 2007 Zheng et al. N Engl J Med, 358: 910, 2008 Eeles et al. Nat Genet, 40: 316, 2008 Thomas et al. Nat Genet, 40: 310, 2008 Severi et al. Cancer Epidemiol Biomarkers, 16: 210, 2007 14
Associated with Biopsy Gleason Score 7 Allele p value OR 95% CI 1 2 3 4 5 rs1447295 8q24 (Region 1) rs16901979 8q24 (Region 2) rs6983267 8q24 (Region 3) rs4430796 (17q12) rs1859962 (17q24) 0.06 0.008 0.41 0.51 0.15 1.4 2.0 1.2 0.9 1.3 1.0-2.1 1.2-3.4 0.8-1.9 0.6-1.3 0.9-1.8 6 + family history 0.38 0.8 0.6-1.2 45 40 35 30 25 20 15 10 5 0 8q24 & Biopsy Gleason Grade Proportion with Gleason 7 P=0.11 rs1447295 (Region1) P=0.03 rs16901979 (Region2) P=0.26 rs6983267 (Region3) P=0.04 Multiple Noncarrier Carrier Cumulative Model for the Prediction of Pathologic Gleason Score 7 rs16901979 8q24 rs16901979 8q24 + 1 other allele rs16901979 8q24 + 2 other allele rs16901979 8q24 + 3 other allele ODDS Ratio 2.0 2.2 3.6 4.8 15
8q24 & PSA Progression 18 16 14 12 10 8 6 4 2 0 P=0.47 rs1447295 (Region1) P=0.28 rs16901979 (Region2) P=1.0 rs6983267 (Region3) P=0.54 Multiple Noncarrier Carrier Limited to patients with Follow-Up >1yr N=449 A systemic progression model containing 17 genes predicted prostate cancer death and systemic progression Genes mapped to 8q24 were significantly enriched in the model. The measurement of gene expression pattern may be useful for determining which men may benefit from additional therapy after PSA recurrence. PLoS ONE 3:e2318,2008 WJC Genetics of Prostate Cancer What are we likely to find in the future, and how are we going to find it? Following slides courtesy of Douglas A. Easton, PhD, Cambridge University 16
Prostate cancer risk prediction PROBABILITY PROBABILITY 0.005 0.005 0.004 0.003 0.002 0.001 0 Current 16 loci 0.004 0.005 0.003 0.004 0.002 0.003 0.001 0.002 0 0.001 0.01 0.1 1 10 100 0 RISK 0.01 0.1 1 10 100 PROBABILITY RISK Theoretical maximum 0.01 0.1 1 10 100 RISK top 1% ~ 3x risk top 1% ~ 16x risk Known Breast Cancer Genes TP53 Relative Risk 10 2 1.5 1.1 PTEN BRCA1 BRCA2 High-risk Family studies Too hard! BRIP1 ATM PALB2 CHEK2 Rare moderate-risk Resequencing Don t exist! Common low-risk GWAS TNRC9 5p LSP1 MAP3K1 FGFR2 8q 2q CASP8 0.1% 1% 10% 30% Minor allele frequency 17
Prostate Cancer Susceptibility Loci High-risk Linkage studies Prostate Cancer Loci to Find? Relative Risk 10 2 1.5 1.1 BRCA2 BRCA1? Rare low/moderate risk Resequencing NBS1 CHEK2? 8q 8q 3p12 Common low risk Association studies KLK2/3 6q25 2p15 MSMB JAZF1 Xp11 8q 17q 11q13 LMTK2 TCF2 Relative Risk 10 2 1.5 1.1 Resequencing 8q Better arrays 8q 3p12 MSMB KLK2/3 JAZF1 Xp11 8q 17q 11q13 6q25 TCF2 2p15 LMTK2 New/combined scans/follow-ups 0.1% 1% 10% 30% Risk allele frequency 0.1% 1% 10% 30% Minor allele frequency Large Genome-Wide Scans Detect More Small-Effect Variants In a collaborative study with decode genetics (33,000 cases and 45,000 controls), we identified SNPs associated with cancer at multiple sites (lung 1.15, bladder 1.12, prostate 1.07, and cervix 1.31) Thus, we have identified genetic variants that confer susceptibility to several cancers that have strong environmental components to their risk Implications Many new genetic prostate cancer markers May provide insights into prostate carcinogenesis and strategies for prevention and treatment Some will contribute significantly to patient care Rafnar T et al, Nat Genet. 2009 Feb;41(2):221-7. Epub 2009 Jan 18 18