Field of the invention

[0001] The present invention relates to markers for the prognosis of recurrence of prostate cancer.

Background of the invention

[0002] Prostate Cancer (PCa) is a major public health concern since it is the sixth most common cancer in the world and the second leading cause of cancer death in North American men. ¹ Of the several known risk factors, the most important are age, ethnicity, dietary and genetic factors. ^2"4 Patients with localized (clinical stage T1-T2) and locally advanced (T3) PCa are frequently treated with radical prostatectomy (RP), a potentially curative procedure. It is estimated that over 30% of men undergoing RP will have disease relapse, also referred to as biochemical recurrence (BCR) as the first clinical indication of rising seaim level of PSA. ³ [0003] Currently, the Tumor/Nodes/Metastasis (TNM) staging system, the Gleason score, and pre-treatment seaim prostate-specific antigen (PSA) are the most important factors influencing both the likelihood of more extensive disease and the probability of subsequent relapse following RP. ⁶ " ⁷ Indeed, these tools include both nomograms and risk tables incorporating clinical variables that can predict, although still imperfectly, the likelihood of tumor recurrence and provide caicial prognosis information to guide clinicians in their therapeutic decisions. The risk of disease progression greatly differs between individuals and the heterogeneity in clinical behaviour further emphasizes the need to find novel markers of progression. Even if cases of PCa are considered localized at the time of diagnosis, the rate of BCR after RP is still significant and occur most often in the 5 years after surgery. ⁸ The persistence of tumor cells in a state of either complete or near dormancy prior to metastatic progression is likely accountable for disease recurrence while these residual cells are most probably responsive to hormones. [0004] Androgen hormones, such as testosterone (T) and 5a-dihydrotestosterone (DHT), have been clearly implicated in development of PCa. Circulating T, secreted by the testis, and adrenal steroid precursors (dihydroepiandrosterone; DHEA and DHEA-sulfate;

DHEA-S) are among factors influencing androgen levels in the prostate and other tissues. ^9"11 Prostate cells, as well as a number of peripheral tissues, contain a variety of steroidogenic enzymes required for the local formation of active androgens from adrenal precursors. ^12"14"13 Namely, the conversion of T by 5a-reductases (SRD5A1 and SRD5A2) leads to DHT, the more potent androgen receptor (AR) agonist in target cells. SRD5A2 is the major 5a-reductase enzyme expressed in the prostate compared to SRD5AI. ¹⁶ However, while the expression ofSRD5A2 decreases in prostate cancer cells, SRD5A1 is increased in tumoral tissues. ^16"20 This imbalance in the expression of ^' SRD5A genes in PCa tumors illustrate the complex relation between 5a-reductases, DHT synthesis and PCa progression.

[0005] Androgen deprivation therapy (ADT) is the standard of care for metastatic PCa and is also used to treat asymptomatic patients with PSA recurrence after failed primary therapy (RP), further reinforcing the initial androgen dependency of these cells. ²¹ ' ²²

Finasteride, a 5a-reductase type 2 inhibitor currently used in the clinic, has been recently shown to be an effective chemopreventive medication reducing by almost 25% the risk of PCa incidence. ²³ Additionally, data from clinical studies were recently used to model a risk-adapted PSA-based chemoprevention strategy. ²⁴ Despite this well recognized hormonal dependence of prostate cancer cells in the early cancer stage, very few studies have investigated the associations between polymorphisms in the androgen biosynthesis pathway and clinical outcome after surgical procedure. ^23"32 To date, common

polymorphisms such as those in sex-steroids biosynthesis pathways have been extensively

19 33 43

studied in relation to risk of PCa. ^{" " ~} However, almost all of these studies did not address the association between polymorphisms in genes regulating hormonal exposure with PCa recurrence and survival, and were not designed to do so. Long-term longitudinal studies are thus still required to systematically evaluate the impact of a patient's genetic profile on risk of recurrence. [0006] We hypothesize that variations in SRD5A genes may alter systemic androgen availability and affect the tumoral microenvironment exposure to hormones, which could modify the risk of PCa recurrence after RP. This is based on the fact that 5 -reductases have a well characterized physiological role in DHT biosynthesis, are associated with PCa risk and more recently, with cancer progression ²⁶ " ²⁸ and represent molecular targets in PCa prevention trials ²³ ' ²⁴ . In this study, we performed a detailed genetic analysis of the SRD5A1 and SRD5A2 genes in relation to PCa progression. We also aimed at validating the association of the known SRD5A2 V89L polymorphism associated with BCR by other groups. ²⁶ ' ²⁸ One particular feature of our study is that the association of inherited variations with the risk of BCR was determined after RP as the sole initial curative intent in a cohort of 526 men with clinically localized and locally advanced PCa.

Summary of the invention

[0007] There is therefore provided an in-vitro method for providing a diagnosis, prognosis or predicting the likelihood of a human subject to develop prostate cancer or a recurrence thereof, said method comprising the steps of: a) obtaining a nucleic acid from a nucleic acid-containing sample (particularly a non-tumor or a tumor sample) from said human subject; and b) determining the individual's genetic variations (or haplotypes) in SRD5A1 or SRD5A2 gene in comparison to normal sequence of said genes; whereby the presence of at least one genetic variation in SRD5A1 or SRD5A2 in said subject's nucleic acid is an indication that said subject has an increased or a decreased likelihood that the prostate cancer will develop or recur.

[0008] The invention further provides the method as defined herein wherein step b) further comprises the step of: b') identifying at least one single nucleotide polymorphism (SNP) in said nucleic acids, said SNP being selected from the group consisting of: rs518673;

rsl66050; rsl2470143; rs2208532; rs2300702; rs4952197 and rs676033 or any of their associated variants as listed in Table 3; whereby the presence of at least one of said markers in said subject's sample is an indication that said subject has an increased or a decreased likelihood that the prostate cancer will develop or recur. [0009] The present invention also provides a method for adapting a course of treatment of prostate cancer in a human subject after the subject has undergone radical prostatectomy, comprising the steps of: a) providing a prognosis or predicting the likelihood of a human subject to develop prostate cancer recurrence in accordance with the method as defined herein; and b) adapting a course of treatment according to whether said subject has an increased or decreased likelihood that the cancer will recur.

[0010] The invention further provides a kit for predicting the likelihood of a human subject to develop prostate cancer and /or recurrence by detecting a SNP in a reference sequence selected from the group consisting of: rs518673; rsl66050; rsl2470143; rs2208532;

rs2300702; rs4952197 and rs676033 or their associated SNPs; said kit comprising reagents for determining the individual's genetic variations (or haplotypes) in SRD5A1 or SRD5A2 gene.

Detailed description of the invention

Description of the figures

[0011] Figure 1 shows the risk of recurrence associated with known clinical and pathological prognostic variables (A) and SRD5A genes (B). Boxes represent hazard ratios (HR) and their 95% CI. PSA categories are in ng/ml. Reference categories (HR: 1.00) are: PSA at diagnosis < lOng/ml, pG < 6, and pT < T2b. Genetic linkage between htSNPs tested for each SRD5A gene is represented in the triangles on the left in panel B; and [0012] Figure 2 illustrates Kaplan-Meier estimates of recurrence-free survival for A)

SRD5A1, B) SRD5A2 and C) both genes. Only positive htSNPs in multivariate analysis are represented. Values for log-rank P values (LR) are shown in each frame. Numbers (0 to 4) in panel C indicate the number of protective alleles for both genes. SRD5A1 protective allele is rs518673T and SRD5A2 protective allele is rs 12470143 A. Definitions and abbreviations

[0013] ADT: androgen-deprivation therapy; AR: androgen receptor; BCR: biochemical recurrence; DHT: 5a-dihydrotestoterone; HR: hazard ratio; htSNP: haplotype-tagging SNP; PCa: prostate cancer; PSA: prostate-specific antigen; SRD5A1: 5a-reductase type 1;

SRD5A2: 5a-reductase type 2; RP: radical prostatectomy; SNP: Single nucleotide polymorphism; T: Testosterone.

Detailed description of particular embodiments [0014] In accordance with the method of the invention, it will be well recognized by persons skilled in the art that genetic variations are assessed in comparison to the gene sequence identified for the normal gene. Such gene sequence for each of these enzymes in their normal state can be found at:

• for SRD5A1: Ensembl accession number ENSG00000145545 and can be

consulted from the NCBI Internet site at the reference number: gene ID 6715.

• for SRD5A2: Ensembl accession number ENSG00000049319 and can be

consulted from the NCBI internet site at the reference number: gene ID 6716.

[0015] In accordance with particular aspects of the present invention, step b) of the method may further comprise: b") contacting the subject's nucleic acid with a reagent that specifically binds to at least one of said single nucleotide polymorphism (SNP); and c) detecting the binding of reagent to at least one of said SNP, whereby the binding of said reagent to at least one SNP is an indication that said subject has an increased or a decreased likelihood that the prostate cancer will recur.

[0016] The invention also provides the method as defined herein, wherein the SNP is found in reference sequences (rs) selected from the group consisting of: rsl66050; rs2208532; rs2300702; rs4952197; and rs676033, or any of their associated SNPs whereby the presence of said SNP is an indication of an increased likelihood that the prostate cancer will recur. Particularly, the SNP is found in reference sequences (rs) selected from the group consisting of: rs2208532 or rs676033, whereby the presence of said SNP is an indication of an increased likelihood that the prostate cancer will recur. More particularly, the SNP is found in rs2208532 or any of its associated SNPs and the presence of said SNP is an indication of an increased likelihood that the prostate cancer will recur.

[0017] The invention also provides the method as defined herein, wherein the SNP is found in reference sequences (rs) selected from the group consisting of: rs518673 or rsl2470143, whereby the presence of said SNP is an indication of a decreased likelihood that the prostate cancer will recur. Particularly, the presence of SNP found in rs518673 or any of its associated SNPs is an indication of a decreased likelihood that the prostate cancer will recur. More particularly, the presence of both rs518673T and rs 12470143 A or any of their associated SNPs is still a further indication of a decreased likelihood that the prostate cancer will recur.

[0018] The invention also provides the method as defined herein wherein in step a), the nucleic acid is DNA or RNA. In particular embodiment of the invention, the DNA is extracted from a non-tumor or a tumor sample from said human subject to be utilized directly for identification of the individual's genetic variations. Particularly, examples of nucleic acid detection methods are: direct sequencing or pyrosequencing, massively parallel sequencing, high-throughput sequencing (a.k.a next generation sequencing), high performance liquid chromatography (HPLC) fragment analysis, capillarity electrophoresis and quantitative PGR (as, for example, detection by Taqman® probe, Scorpions™ ARMS Primer or SYBR Green). In one aspect, the amplification of the DNA is carried out by means of PGR. Several methods for detecting and analyzing the PGR amplification products have been previously disclosed. The general principles and conditions for amplification and detection of genetic variations, such as using PGR, are well known for the skilled person in the art.

[0019] Alternatively, other methods of nucleic acid detection such as hybridization carried out using appropriately labeled probe, detection using microarrays e.g. chips containing many oligonucleotides for hybridization (as, for example, those produced by Affymetrix Corp.) or probe-less technologies and cleavage-based methods may be used. Preferably, amplification of the DNA can be carried out using primers that are specific to the marker, and the amplified primer extension products can be detected with the use of nucleic acid probe. More particularly, the DNA is amplified by PGR prior to incubation with the probe and the amplified primer extension products can be detected using procedure and equipment for detection of the label. [0020] The invention also provides the method as defined herein, wherein the subject's tumor sample is from a biopsy. The invention also provides the method as defined herein, wherein the subject's non-tumor sample is selected from the group consisting of: tissue or biological fluid. Particularly, the tissue is a lymph node, hair or a buccal smear. Still, particularly, the biological fluid is sputum, saliva, blood, seaim urine, semen or plasma. [0021] The invention also provides the method as defined herein, wherein, when the likelihood of recurrence is increased, the subject is prescribed 5a-reductase inhibitors therapy.

[0022] The invention also provides the kit as defined herein, comprising PGR primer-probe set, wherein the primers are selected from the group consisting of: SEQ ID Nos. 1 to 38; and the probe is selected from the group consisting of: SEQ ID Nos. 39 to 57.

Examples

Patients and Methods

Clinical Data and Outcome Collections

[0023] The study cohort, mostly composed of Caucasians, included 526 men who underwent RP at l'Hotel-Dieu de Quebec Hospital (QC, Canada) between Febaiary 1999 and December 2002. Each participant provided written consent before surgery for the analysis of their genome and the research protocol was approved by the research ethical committee at the Centre Hospitalier Universitaire de Quebec (CHUQ, QC, Canada). All patients were followed postoperatively with serial PSA measurements and detailed clinical information was available.

DNA Isolation and Genetic Analysis [0024] Polymorphisms studied were chosen according to one or more of the following criteria: i) to be likely functional (with supportive data in the literature), ii) to have previously been associated with PCa risk, aggressiveness, age at onset, BCR or ADT efficiency, and iii) to explain most of the haplotype diversity in the CEU (Utah residents with Northern and Western European Ancestry) Hapmap population. For both SRD5A genes, a region covering all the exons, introns and 5 kb of the 5' and 3' sections of the genes was screened using a haplotype tagging SNPs (htSNPs) strategy to maximize coverage, using HapMap Phase 2 (www.hapmap.og/cgi-perl/gbrowse/hapmap3r2_B36) ⁴⁴ and data from the CEU unrelated subjects based on a rl >0.80 and a minimum minor allele frequency >0.05.

[0025] Peripheral blood was collected on the morning of a preoperative clinic visit and kept frozen at -80°C until analysis. Genomic DNA was purified using the QIAamp DNA Blood Mini Kit (Qiagen Inc., Mississauga, ON, Canada) and stored at -20°C, PGR amplifications were performed using Sequenom iPLEX matrix-assisted laser

desorption/ionization-time-of-flight mass spectrometry by the sequencing service of

McGill University and Genome Quebec INNOVATION center (QC, Canada). For oligos sequence, see Table 1. Negative controls were present for every ain of analyses and quality controls (random replicates of known genotypes) were successfully performed in 5% of the study cohort.

Table 1. Oligonucleotides used for genotyping analysis.

SNP SEQ SEQ SEQ ID ID ID

Gene Forward Primer Sequence No. Reverse Primer Sequence No. Probe Sequence No. rs518673 ACGTTGGATGCCGTTTCCCTCATGTTGTTC 1 ACGTTGGATGGAGAGAGAATGAGTGCTGAG 20 GGAAGGACGAAGCCCCTAATAAACCAT 39 rs166050 ACGTTGGATGTGTGTATTCTCTGGACACGC 2 ACGTTGGATGACCAAAGTAGGTTATCGGGC 21 GAAGCAAGGAAAGCTGTGCATCAA 40 rs8192120 ACGTTGGATGCCCTCCATCTTTGTAGTTAC 3 ACGTTGGATGTTGAGCCCTAAGAGCAACAC 22 TGCTCTCCTAAGCATATCCTAC 41 rs501999 ACGTTGGATGGGTGCTATCCTAGTCCCATC 4 ACGTTGGATGTTCCATCACCATGCCCATTC 23 GGGGCAGCTTAGTAACTCATCC 42 w rs500182 ACGTTGGATGCCCTTGATTTCCAACACCAC 5 ACGTTGGATGCATTGCTGTGAATGCACATC 24 AATGCACATCCAGACA 43 rs4702378 ACGTTGGATGACCAAAGTAGGTTATCGGGC 6 ACGTTGGATGTGTGTATTCTCTGGACACGC 25 GCTTTCCTTGCTTCAC 44 rs3822430 ACGTTGGATGAGCCGTTACAGGTACAGAAC 7 ACGTTGGATGTTTCTTGTTTCCTAAGGTGC 26 TCTAAGGTGCTTAATTTACCC 45 rs2208532 ACGTTGGATGTTCTTGAGCTTTGTTCTGGG 8 ACGTTGGATGACACTCTTCTGGAACCACTG 27 TCAACACTGAACAAATCAGATCAAA 46 rs523349 ACGTTGGATGGGGAAAAACGCTACCTGTGG 9 ACGTTGGATGTGCAGGAGCTGCCTTCCTTC 28 CTTCTGGGCCTCTTCTGC 47 rs676033 ACGTTGGATGAACAGAAGACACAGCACACC 10 ACGTTGGATGTACAGTAAAGGCTCTTGCCC 29 GCTTGACCCCTCACTCC 48 rs9332975 ACGTTGGATGTGGCTACTAAGAGGATAGAG 1 1 ACGTTGGATGAAGCAGCATGTAACCTCATC 30 GGTCTCTTTG I I I I ACACTACTCATTATTT 49 rs2281546 ACGTTGGATGGTCACATGAGCTAGTTCCAC 12 ACGTTGGATGTCTTGCCTGCACC I I I I ATC 31 CAGCAAACGATTTCCAAGGAAGTCCCCC 50

< rs12470143 ACGTTGGATGCTCTCATTAAACCTCTCCCG 13 ACGTTGGATGTGGAGAATAGATTGCGGTGG 32 GGCGGTGGAGTAAATGT 51 rs2300702 ACGTTGGATGTCTACCTCAGCATATCTCCC 14 ACGTTGGATGGTTGCAGAAACCTCAGATTG 33 TAGGGTGCTTTTGTTCTTTTTTTGTTTG 52 rs9282858 ACGTTGGATGAAGCACACGGAGAGCCTGAA 15 ACGTTGGATGAAGGCAGCTCCTGCAGGAAC 34 AAAGGGAACCAGGCGGCGCGGG 53 rs4952197 ACGTTGGATGTGGCACAAGCACTTCCTTAG 16 ACGTTGGATGCACAAGGACTGCAATTCCTG 35 GGGCGTGCTGTGCTGGGCTTAGA 54 rs7562326 ACGTTGGATGCTGAAAAGACAGCAGAAGGG 17 ACGTTGGATGTCCTCAGGCAAGGAAATGTG 36 GAGCAAGGAAATGTGAATTACC 55 rs4952222 ACGTTGGATGAAGATATGCCCGCTGCTTTG 18 ACGTTGGATGGGCCATA GTTTA C C AATC C C 37 CTAAAGAAAACTATGAATGGCGTG 56 rs3754838 ACGTTGGATGACTTAGCTGAGGCTCTCATC 19 ACGTTGGATGGAGGCCTTGTGTGTTTCTTC 38 TAGTAC I I I I GTTATTAACATTATATAGAC 57

Statistical Analysis

[0026] Based on our sample size, we were able to detect a hazard ratio of 1.50, for a minor allele with frequency of 5%, with over 80% statistical power. Allelic frequencies and Hardy-Weinberg equilibrium were computed with PLINK (version 1.07), a free open- source whole genome association analysis toolset. ⁴³ To analyse their association with BCR, each SNP was considered using 3 models since the function of most htSNP remain unknown. The first model tested, named the genomic model, considered the SNP as a categorical variable with a common allele homozygote (reference; based on the frequent allele reported in the Hapmap project), heterozygote and a minor allele homozygote. The second model, named the dominant model, considered the SNP with only 2 categories: one with a common allele homozygote (reference) and one with at least one minor allele.

Finally, the third one referred to as the recessive model, also considered the SNP with only 2 categories: one with at least one common allele (reference) and one minor allele homozygote. Cox regression was performed on each SNP considering the 3 above mentioned models with adjustment for confounding variables namely PSA level at diagnosis, age at diagnosis, smoking status, pathological Gleason grade, pathological stage and neoadjuvant ADT. All co-variables were treated as categorical, and for PSA level, Gleason scores and stage, they were used as described by the well-recognized D'Amico risk classification. ⁷ Smoking status and neoadjuvant ADT were classified as "positive" or "negative" while age was classified as <65 and >65 years old. The censoring variable was BCR, which was defined as 1) two consecutive PSA values > 0.3 μg/L, 2) one PSA value > 0.3 μg/L followed by ADT, 3) a last-recorded PSA value > 0.3 μg/L, and 4) the initiation of ADT or radiation therapy by the patient's physician. Kaplan-Meier analyses were also processed for every SNP (log rank), while only results for SNPs which were significantly associated with BCR in Cox regression multivariate analysis are shown. For Kaplan-Meier and Cox regression, statistical analyses were performed using PASW statistics 17 (SPSS Inc., Chicago, IL) and R version 2.10.0 (http://www.r-project.org/).

[0027] Haplotypes were inferred using Phase v2.1.1 program, ⁴⁴ and their relative frequency, as well as pairwise linkage disequilibrium between SNPs, were determined with HAPLOVIEW 4.1. Univariate and multivariate analysis were performed with or without minor haplotypes (frequency<5%), without significant impact on the P values - therefore, only results without minor haplotypes are shown.

Results [0028] Clinical and pathological characteristics of the study cohort are shown in Table 2. All 526 patients had initially RP as curative intent enabling a precise pathological evaluation. The actual median follow-up time of the cohort is 7.4 years (range: 0.5 to 10.2 years). The cohort had mainly organ-confined and locally advanced tumors, as PCa cases were composed mainly of pT2 (60%) and pT3 (37%) pathological tumor stages (Table 2). Overall, 130 cancer cases experienced BCR (25%), which was our primary outcome variable, with a median time to relapse of 2.1 years.

Table 2. Clinical and Pathological Characteristics of the Study Cohort.

Characteristic Number %

Age at diagnosis, years (n

Mean 63.3

Standard deviation 6.8

Biochemical recurrence (n=526)

Yes 130 25

No 396 75

PSA, prostate-specific antigen; n, number of patients for which the information was available.

[0029] Risk of recurrence associated with known clinical and pathological prognostic variables are shown in Figure 1A. As expected, the risk of recurrence increase with higher PSA values at diagnosis with a risk of relapse of HR=1.5 (p=0.081) and HR=2.1 (p=0.003) 5 with PSA values of 10-20 ng/mL and > 20 ng/niL, respectively. Gleason scores of 7 and > 8 are also positively associated with relapse with HR of 2.6 (p=0.002) and 5.4 (p<0.001), respectively. Pathological stage was not associated with biochemical recurrence risk

(Figure 1A)

[0030] A total of 19 htSNPs of which 2 functional coding SNPs, distributed across the two 10 SRD5A genes, were studied herein. The htSNPs strategy allowed us to study 89 genetic variations in both genes (Table 3). htSNPs were selected with a strategy to maximize gene coverage and to reflect adequately the Caucasian haplotype genetic diversity.

Table 3. List of htSNPs included in this study and their associated SNPs

Gene htSNP Associated SNPs htSNP Associated SNPs htSNP Associated SNPs htSNP Associated SNPs rs518673 rs518673 rs500182 rs248807 rs500182 rs2677947 rs3822430 rs8192130 rs166050

rs166050 rs482121 rs248803 rs4702375 rs471604 rs2677933 rs166049 rs1 1738248 rs82481 1 rs494958 rs500182 rs7706809 rs8192120

rs8192120 rs248800 rs521293 rs7720479 rs501999

rs248797 rs562461 rs568509 rs8192166 s rs248805 rs248799 rs484973 rs8192131 c

rs4702378

rs477930 rs500058 rs4702379 rs7707559 rs501999 rs1691052 rs4702378 rs4702374 rs566202 rs1651074 rs1560149 rs6884552 rs3822430

rs472402 rs535981 rs3822430 rs1896670 rs531241 rs193744 rs1 1 134173 rs8192139 rs181807 rs168713 rs3733773

rs2208532 rs2208532 rs676033 rs522638 rs2281546 rs2268794 rs9332975 rs6543634 rs2300701 rs632148 rs1 1889731 rs9332975 rs765138 rs806645 rs1701 1453 rs3731586 rs559555 rs502139 rs1042578 rs1884722 rs2300697 rs2754530 rs2281546 rs7571644 rs523349 rs523349 rs655548 rs11892064 rs7589579 rs12467911 rs623419 rs1 1690596 rs4952197 rs4952197 rs2300699 rs499362 rs7567093 rs4952218 rs558803 rs546935 rs1090817 rs3754838 rs3754838 rs612224 rs481344 rs12470143 rs2300700 rs13027103 rs477517 rs2300702 rs2268796 rs481 126 rs9282858 rs9282858 rs682895 rs2300702 rs585890 rs7562326 rs7562326 rs676033 rs676033 rs2268797 rs564310 rs4952222 rs4952222 rs599300 rs4952220 rs665237

rs614173 rs2300702 rs2300703 rs12470143

SNPs is derived from the analysis of a region covering approximately 250 kb for each SRD5A genes.

[0031] After analyses with a Cox regression multivariate model, adjusted for all clinical and pathological factors known to affect BCR, 8 htSNPs were positively associated (P<0.05) with the risk of relapse. Their relative frequencies in cancer patients with and without relapse, and the corresponding hazard ratios (95% CI) are displayed in Figure 1 and Table 4. Genetic linkage between the 19 htSNPs tested in both SRD5A genes is also represented.

[0032] Among 1 1 htSNPs in SRD5AI, one SNP showed a P< 05 by log-rank test for recurrence-free survival and corresponds to a protective allele (rs518673T). Another risk allele for biochemical recurrence (rsl66050C) was almost significant (log-rank

test=P<0.051) (Figure 2). The SRD5A1 rsl66050 gene polymorphism was associated with an increased recurrence risk of HR=1.83, 95% CI, 1.04-3.21; P=0.035, while the rs518673 in SRD5A1 was associated with a decreased recurrence risk (HR=0.59, 95% CI, 0.41-0.85; P=0.004). Haplotype analyses further revealed five common SRD5A1 haplotypes (H) with a prevalence of > 5% (Table 5). SRD5A1 H2 was significantly associated with the risk of BCR with a HR of 0.64 (0.44-0.94; P=0.023) but did not remained significant in the multivariate model (HR=0.66; CI 95%; 0.46-0.95; p=0.073). [0033] We found 6 htSNPs for the SRD5A2 gene to be associated with recurrence-free survival. The SDR5A2 protective htSNP rs 12470143 was significant with a HR of 0.66 (0.46-0.95; P=0.023). A significant association was observed with the non-synonymous SNP V89L (rs523349) with a HR of 2.14 (95% CI, 1.23-3.70; P=0.007) while no association was seen with the other known coding variation A49T (rs9282858) (HR=0.81, 95% CI, 0.36-1.85; P=0.62) (Table 4). The other 4 risk alleles for recurrence were rs2208532, rs2300702, rs4952197, rs676033 with HR of 1.68, 1.88, 1.55 and 1.90, respectively (Table 4). In our Caucasian population, significant linkage disequilibrium was noted between rs676033 and rs523349 (r2=0.90; Figure 1). Genetic linkage was moderate between rs4952197 and rs676033 (r2=0.69) and between rs49522197 and rs523349

(r2=0.77). The genetic associations between other positive polymorphisms associated with BCR was below 50%. However, genetic linkage between these genetic variations was not ascertained in other populations (such as Asians, African-Americans, Hispanics) and remains to be defined. [0034] For the SRD5A2 gene, we found 4 haplotypes. Of those, H2 and H3 were significantly associated with the risk of BCR but only H3 remained an independent predictor of recurrence in adjusted Cox proportional hazards analysis (adjusted HR=1.63; 1.1 1-2.39; P=0.013; Table 5).

[0035] Protective alleles in both SDR5A genes were significant in the dominant model while risk alleles were significant using the recessive model. Because of the absence of linkage between most htSNPs, the protection conferred by these alleles was not modified by other variants in that population.

Table 4. Genotypes frequencies of SNPs in 5a-reductase genes and their association with BCR.

: oc emca recurrence.

Table 5. SRD5A haplotypes and their association with risk of BCR.

H: haplotype; Freq: frequencies.

[0036] We then investigated the combined effects of protective alleles in both SRD5A genes (Table 6). By combining SRD5A1 (rs518673 T) and SRD5A2 (rsl2470143 A), the protective effect was shown to be additive in Kaplan-Meier analysis with the maximum protection conferred by 3 or 4 alleles (Figure 2C) and remained an independent predictor of recurrence in Cox proportional hazards analysis (HR=0.33, 0.17-0.63; P=0.001 (Table 6). An allele dosage effect was observed with each additional allele diminishing by 26% the risk of recurrence of BCR (P=0.003).

Table 6. Combined protective effects of SRD5A1 and SRD5A2 htSNPs.

Number of protective alleles was considered as a continuous variable in univariate and multivariate analysis. n: number, L-R: log rank P value.

SRD5A 1 protective allele is rs518673T.

SRD5A2 protective allele is rs12470143A.

[0037] Using Chi-square likelihood ratio test, no htSNPs were associated with PSA at diagnosis, pathological Gleason score or pathological stage (P > 0.05). Only the SRD5A2 rs2208532 was significantly associated with the pathological stage (P=0.048; minor allele homozygote being over-represented in tumors with pT > T3a). htSNPs associated with BCR were shown to be the third most important predictors of recurrence after the Gleason score and PSA values. Discussion

[0038] One of the most significant challenges in oncology remains our inability to predict PCa disease recurrence and clinical outcome due to a lack of key prognostic markers. The prediction of recurrence following radical prostatectomy is clearly important for personalized treatments and follow-up strategies due to striking heterogeneity in prostate cancer clinical behaviour. Inherited variations in sex-steroid biosynthesis enzymes are attracting candidates as novel predictive markers since variation in these biotransformation pathways may modify exposure of any residual cancer cells to active androgens produced from adrenal precursors still available after RP in men with PCa ⁴⁷ It is likely that alteration in hormonal exposure of the tumor microenvironment may ultimately also affect the risk of recurrence.

[0039] Here, in a cohort of 526 PCa cases, significant associations of BCR with multiple genetic polymorphisms in both SRD5A genes were observed. Of the 19 htSNPs tested, 8 inherited variations (42%) were shown to affect the risk of recurrence after RP either by conferring protection or an elevated risk, independently of known clinical and pathological predictors of prostate cancer recurrence. One of the best characterized SNP in the SRD5A2 gene is the functional V89L (rs523349) polymorphism associated with a decreased in enzymatic activity. ⁴⁸ ' ⁴⁹ This polymorphism has been associated with an aggressive form of prostate cancer in a recent large case-control study, ³⁰ with prostate cancer risk in some

51 52 26 28

studies, ' and with conflicting results with biochemical recurrence. ' The clinical utility of this marker must be further evaluated before any clinical implementation and risk- adapted PCa treatment strategies with these molecular markers. One important finding of our study is a significant association between the V89L (rs523349) polymorphism and the risk of BCR after prostatectomy. A 2.14-fold risk of BCR for homozygotes of the minor C allele was observed. This result is in agreement with Shibata and colleagues ²⁸ and with a recent large study that associated this SNP with an aggressive form of PCa. ³⁰ However, for the SRD5A2 A49T variant (rs9282858), a SNP associated in some studies with the risk of PCa, ³³ no association was observed with recurrence-free survival after RP indicating that this SNP would have no obvious role in PCa recurrence. This result is consistent with the latest meta-analysis investigating this SNP, showing no association with the risk of PCa. ³³ [0040] Our study is the first to show strong positive associations of multiple independent SRD5A2 genetic variations with BCR. These findings suggest that the 5a-reductase type 2 germline variations play a critical role in prostate cancer recurrence after RP. Most variations represent independent risk alleles for BCR with one variant (rsl247043) associated with significant protection. Our data also argue for a significant role oiSRDSAl in the risk of recurrence after RP. Variations in this gene have been previously associated with the risk of PCa. ¹² ' ¹⁸ Two SNPs were positively associated with BCR; the rs 166050 associated with an increased recurrence risk while the rs518673 conferred protection. By combining the number of protective alleles in both genes, namely SRD5A1 (rs518673T) and SRD5A2 (rsl2470143 A), this independent protective effect was shown to be additive and maximal in patients carrying 3 to 4 of these alleles.

[0041] Strengths of our study includes the large sample size combined with a significant median follow-up time of 7.4 years that provided information on potential confounders. Limitations include the limited number of some clinically relevant events such as metastasis, hormone-resistance, or death related to the localized features of the tumors, that prevented us from looking at the association between molecular signature in SRD5A genes and risk for these events.

[0042] Findings are remarkable for the fact that they complement the evidence on somatic and germline genetic changes to predict disease recurrence. ³¹ ' ^{34 6} Additional investigations are required to characterize the underlying biological mechanisms driving the positive associations of inherited germline variations in the 5 -reductase pathway with BCR. At the time of biochemical relapse and PSA elevation, the disease is particularly androgen- dependent for growth and progression. We can only speculate that these genetic variations influence active androgen formation and exposure of disseminated cancer cells remaining after RP, potentially driving more hormone-dependent cells into cell replication, and subsequently leading to inter-individual differences in recurrence.

[0043] In conclusion, our data reveal that multiple genetic markers in SRD5A genes contribute to biochemical recurrence risk after radical prostatectomy. These markers appear independent of current predictors of recurrence such as Gleason score and PSA level and predict risk better than the pathological stage. These findings may ultimately help refine our ability to identify individuals at low or high risk of cancer relapse after RP, beyond known prognostic variables, and for whom a more personalized approach might optimize outcome, especially in the era of 5a-reductase inhibitors therapy. References

1. Jemal, A. et al. Cancer statistics, 2007. CA Cancer J Clin 57, 43-66 (2007).

2. Gronberg, H. Prostate cancer epidemiology. Lancet 361, 859-64 (2003).

3. Platz, E. A., Rimm, E. B., Willett, W. C, Kantoff, P. W. & Giovannucci, E. Racial variation in prostate cancer incidence and in hormonal system markers among male health professionals. J Natl Cancer Inst 92, 2009-17 (2000).

4. Steinberg, G. D., Carter, B. S., Beaty, T. H., Childs, B. & Walsh, P. C, Family history and the risk of prostate cancer. Prostate 17, 337-47 (1990).

5. Caire, A. A. et al. Delayed prostate-specific antigen recurrence after radical prostatectomy: how to identify and what are their clinical outcomes? Urology 74, 643-7

(2009).

6. D'Amico, A. V. et al. Combination of the preoperative PSA level, biopsy gleason score, percentage of positive biopsies, and MRI T-stage to predict early PSA failure in men with clinically localized prostate cancer. Urology 55, 572-7 (2000).

7. D'Amico, A. V. et al. Utilizing predictions of early prostate-specific antigen failure to optimize patient selection for adjuvant systemic therapy trials. J Clin Oncol 18, 3240-6 (2000).

8. Roberts, S. G, Blute, M. L., Bergstralh, E. J., Slezak, J. M. & Zincke, H. PSA doubling time as a predictor of clinical progression after biochemical failure following radical prostatectomy for prostate cancer. Mayo Clin Proc 76, 576-81 (2001).

9. Belanger, A., Pelletier, G., Labrie, F., Barbier, O. & Chouinard, S. Inactivation of androgens by UDP-glucuronosyltransferase enzymes in humans. Trends Endocrinol Metab. 14, 473-9. (2003).

10. Labrie, F. Adrenal androgens and intracrinology. Semin Reprod Med 22, 299-309 (2004).

1 1. Labrie, F. Intracrinology. Mol Cell Endocrinol 78, C I 13-8 (1991).

12. Luu-The, V., Belanger, A. & Labrie, F. Androgen biosynthetic pathways in the human prostate. Best Pract Res Clin Endocrinol Metab 22, 207-21 (2008).

13. Labrie, F. et al. Intracrinology: role of the family of 17 beta-hydroxy steroid dehydrogenases in human physiology and disease. J Mol Endocrinol 25, 1-16 (2000).

14. Vandenput, L. et al. Seaim levels of specific glucuronidated androgen metabolites predict BMD and prostate volume in elderly men. J Bone Miner Res 22, 220-7 (2007).

15. Hum, D. W. et al. Characterization of UDP-glucuronosyltransferases active on steroid hormones. J Steroid Biochem Mol Biol 69, 413-23 (1999).

16. Thomas, L. N. et al. Differential alterations in 5alpha-reductase type 1 and type 2 levels during development and progression of prostate cancer. Prostate 63, 231-9 (2005). 17. Elo, J. P. et al. Characterization of 17beta-hydroxysteroid dehydrogenase isoenzyme expression in benign and malignant human prostate. Int J Cancer 66, 37-41 (1996).

18. Montgomery, R. B. et al. Maintenance of intratumoral androgens in metastatic prostate cancer: a mechanism for castration-resistant tumor growth. Cancer Res 68, 4447- 54 (2008).

19. Stanbrough, M. et al. Increased expression of genes converting adrenal androgens to testosterone in androgen-independent prostate cancer. Cancer Res. 66, 2815-25. (2006). 20. Titus, M. A. et al. Steroid 5alpha-reductase isozymes I and II in recurrent prostate cancer. Clin Cancer Res 1 1, 4365-71 (2005).

21. Harris, W. P., Mostaghel, E. A., Nelson, P. S. & Montgomery, B. Androgen deprivation therapy: progress in understanding mechanisms of resistance and optimizing androgen depletion. Nat Clin Pract Urol 6, 76-85 (2009).

22. Horwitz, E. M. Prostate cancer: Optimizing the duration of androgen deprivation therapy. Nat Rev Urol 6, 527-9 (2009).

23. Thompson, I. M. et al. The influence of finasteride on the development of prostate cancer. N Engl J Med 349, 215-24 (2003).

24. Vickers, A. J., Savage, C. J. & Lilja, H. Finasteride to prevent prostate cancer: should all men or only a high-risk subgroup be treated? J Clin Oncol 28, 1 1 12-6.

25. Edwards, S. M. et al. Androgen receptor polymorphisms: association with prostate cancer risk, relapse and overall survival. Int J Cancer 84, 458-65 (1999).

26. Nam, R. K. et al. V89L polymorphism of type-2, 5-alpha reductase enzyme gene predicts prostate cancer presence and progression. Urology 57, 199-204 (2001).

27. Strom, S. S. et al. Androgen receptor polymorphisms and risk of biochemical failure among prostatectomy patients. Prostate 60, 343-51 (2004).

28. Shibata, A. et al. Polymorphisms in the androgen receptor and type II 5 alpha- reductase genes and prostate cancer prognosis. Prostate 52, 269-78 (2002).

29. Celhar, T., Gersak, K., Ovcak, Z., Sedmak, B. & Mlinaric-Rascan, I. The presence of the CYPl lAl (TTTTA)6 allele increases the risk of biochemical relapse in organ confined and low-grade prostate cancer. Cancer Genet Cytogenet 187, 28-33 (2008).

30. Hamada, A. et al. Association of a CYP17 polymorphism with overall survival in Caucasian patients with androgen-independent prostate cancer. Urology 70, 217-20 (2007). 31. Lindstrom, S. et al. Inherited variation in hormone-regulating genes and prostate cancer survival. Clin Cancer Res 13, 5156-61 (2007).

32. Tsuchiya, N. et al. Impact of IGF-I and CYP19 gene polymorphisms on the survival of patients with metastatic prostate cancer. J Clin Oncol 24, 1982-9 (2006).

33. Okugi, H. et al. Association of the polymorphisms of genes involved in androgen metabolism and signaling pathways with familial prostate cancer risk in a Japanese population. Cancer Detect Prev 30, 262-8 (2006).

34. Olsson, M. et al. The UGT2B 17 gene deletion is not associated with prostate cancer risk. Prostate 4, 4 (2008).

35. Paris, P. L. et al. Association between a CYP3A4 genetic variant and clinical presentation in African- American prostate cancer patients. Cancer Epidemiol Biomarkers

Prev 8, 901-5 (1999).

36. Park, J. et al. Deletion polymorphism of UDP-glucuronosyltransferase 2B 17 and risk of prostate cancer in African American and Caucasian men. Cancer Epidemiol Biomarkers Prev. 15, 1473-8. (2006).

37. Park, J. et al. Asp85tyr polymorphism in the udp-glucuronosyltransferase (UGT) 2B 15 gene and the risk of prostate cancer. J Urol. 171, 2484-8. (2004).

38. Park, J. Y. et al. Association between polymorphisms in HSD3B 1 and UGT2B 17 and prostate cancer risk. Urology. 70, 374-9. (2007).

39. Setlur, S. R. et al. Genetic variation of genes involved in dihydrotestosterone metabolism and the risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 19, 229-39. 40. Cunningham, J. M. et al. Evaluation of genetic variations in the androgen and estrogen metabolic pathways as risk factors for sporadic and familial prostate cancer. Cancer Epidemiol Biomarkers Prev. 16, 969-78. (2007).

41. Cussenot, O. et al. Effect of genetic variability within 8q24 on aggressiveness patterns at diagnosis and familial status of prostate cancer. Clin Cancer Res 14, 5635-9

(2008).

42. Cussenot, O. et al. Combination of polymorphisms from genes related to estrogen metabolism and risk of prostate cancers: the hidden face of estrogens. J Clin Oncol 25, 3596-602 (2007).

43. Mononen, N. & Schleutker, J. Polymorphisms in genes involved in androgen pathways as risk factors for prostate cancer. J Urol 181, 1541-9 (2009).

44. Stephens, M., Smith, N. J. & Donnelly, P. A new statistical method for haplotype reconstaiction from population data. Am J Hum Genet. 68, 978-89. (2001).

45. Purcell, S. et al. PLINK: a tool set for whole-genome association and population- based linkage analyses. Am J Hum Genet 81, 559-75 (2007).

46. Barrett, J. C, Fry, B., Mailer, J. & Daly, M. J. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 21, 263-5 (2005).

47. Labrie, F. et al. Is dehydroepiandrosterone a hormone? J Endocrinol 187, 169-96 (2005).

48. Makridakis, N. et al. A prevalent missense substitution that modulates activity of prostatic steroid 5alpha-reductase. Cancer Res 57, 1020-2 (1997).

49. Makridakis, N. M., di Salle, E. & Reichardt, J. K. Biochemical and phariiiacogenetic dissection of human steroid 5 alpha-reductase type II. Pharmacogenetics 10, 407-13 (2000).

50. Cussenot, O. et al. Low-activity V89L variant in SRD5A2 is associated with aggressive prostate cancer risk: an explanation for the adverse effects observed in chemoprevention trials using 5-alpha-reductase inhibitors. Eur Urol 52, 1082-7 (2007). 51. Cicek, M. S. et al. Association of prostate cancer risk and aggressiveness to androgen pathway genes: SRD5A2, CYP17, and the AR. Prostate 59, 69-76 (2004).

52. Salam, M. T., Ursin, G., Skinner, E. C, Dessissa, T. & Reichardt, J. K. Associations between polymorphisms in the steroid 5-alpha reductase type II (SRD5A2) gene and benign prostatic hyperplasia and prostate cancer. Urol Oncol 23, 246-53 (2005).

53. Pearce, C. L. et al. No association between the SRD5A2 gene A49T missense variant and prostate cancer risk: lessons learned. Hum Mol Genet 17, 2456-61 (2008). 54. Sharifi, N., Dahut, W. L. & Figg, W. D. The genetics of castration-resistant prostate cancer: what can the germline tell us? Clin Cancer Res 14, 4691-3 (2008).

55. Ross, R. W. et al. Inherited variation in the androgen pathway is associated with the efficacy of androgen-deprivation therapy in men with prostate cancer. J Clin Oncol 26, 842-7 (2008).

56. Haniada, A. et al. Effect of SLC01B3 haplotype on testosterone transport and clinical outcome in Caucasian patients with androgen-independent prostatic cancer. Clin Cancer Res 14, 3312-8 (2008).