Background: Metastases from men with castration-resistant prostate cancer (CRPC) harbor increased tumoral androgens versus untreated prostate cancers. This may reflect steroid uptake by OATP (organic anion transporting polypeptide)/SLCO transporters. We evaluated SLCO gene expression in CRPC metastases and determined whether prostate cancer outcomes are associated with single nucleotide polymorphisms (SNP) in SLCO2B1 and SLCO1B3, transporters previously shown to mediate androgen uptake.

Methods: Transcripts encoding eleven SLCO genes were analyzed in untreated prostate cancer and in metastatic CRPC tumors obtained by rapid autopsy. SNPs in SLCO2B1 and SLCO1B3 were genotyped in a population-based cohort of 1,309 Caucasian prostate cancer patients. Median survival follow-up was 7.0 years (0.77–16.4). The risk of prostate cancer recurrence/progression and prostate cancer–specific mortality (PCSM) was estimated with Cox proportional hazards analysis.

Results: Six SLCO genes were highly expressed in CRPC metastases versus untreated prostate cancer, including SLCO1B3 (3.6-fold; P = 0.0517) and SLCO2B1 (5.5-fold; P = 0.0034). Carriers of the variant alleles SLCO2B1 SNP rs12422149 (HR: 1.99; 95% CI: 1.11–3.55) or SLCO1B3 SNP rs4149117 (HR: 1.76; 95% CI: 1.00–3.08) had an increased risk of PCSM.

Conclusions: CRPC metastases show increased expression of SLCO genes versus primary prostate cancer. Genetic variants of SLCO1B3 and SLCO2B1 are associated with PCSM. Expression and genetic variation of SLCO genes which alter androgen uptake may be important in prostate cancer outcomes.

Impact:OATP/SLCO genes may be potential biomarkers for assessing risk of PCSM. Expression and genetic variation in these genes may allow stratification of patients to more aggressive hormonal therapy or earlier incorporation of nonhormonal-based treatment strategies. Cancer Epidemiol Biomarkers Prev; 20(4); 619–27. ©2011 AACR.

This article is featured in Highlights of This Issue, p. 565

Androgens play a critical role in the development and progression of prostate cancer (1, 2). Although androgen deprivation therapy (ADT) remains the most effective treatment option for men with advanced disease, the clinical course following ADT is uniformly marked by progression to castration-resistant prostate cancer (CRPC). Many mechanisms proposed to confer a castration-resistant phenotype [e.g., androgen receptor (AR) overexpression, AR mutations with promiscuous ligand interactions, and enhanced AR signaling via coregulator alterations] still require, or are enhanced by, the presence of AR ligands (3–5). In this respect, we and others have shown that the tissue response to castration is characterized by the presence of residual prostatic androgens at levels capable of activating the AR and maintaining androgen-regulated gene expression (6–9). Moreover, tumor metastases from men with CRPC have been found to contain detectable testosterone levels that exceed androgen levels in the prostate tissue of eugonadal men (10).

The source of residual tissue androgens present despite ADT has not been elucidated but may reflect the uptake of adrenal androgens and intracellular conversion to testosterone or de novo androgen synthesis from cholesterol or progesterone precursors. Accordingly, we and others have shown that soft tissue and bone CRPC metastases express genes mediating steroid biosynthesis and adrenal androgen utilization (10, 11).

Emerging data also suggest a potential role for steroid transport proteins encoded by the SLCO gene family in mediating the uptake of androgen into prostate cancer cells and thereby influencing the clinical response to androgen suppression (12, 13). The organic anion transporting polypeptides (OATP) are a superfamily of SLCO-encoded membrane transporters involved in the transport of bile acids, steroid conjugates, xenobiotics, and a variety of clinically important drugs (14). Several family members are known to mediate the uptake of steroids and steroidogenic precursors, including sulfated forms of pregnenolone, estrone, and DHEA (dehydroepiandrosterone). Among these are SLCO1B1 and SLCO1B3, primarily expressed in the liver, and SLCO1A2 and SLCO2B1, more broadly distributed in liver, kidney, intestine, and brain and in steroidogenic tissues such as testis, ovary, mammary epithelium, placenta, and adipose tissue (14–19).

Consistent with a physiologic role for SLCO transporters in providing a pool of intracellular precursors for steroidogenic tissues, several studies have shown SLCO transporter expression in human breast carcinomas, suggesting a role for SLCO-mediated activity in the estrogen-dependent pathophysiology of this tumor (15, 20). Notably, Al Sarakbi and colleagues observed a statistically significant association between the expression of SLCO2B1 and breast cancer grade (21), whereas Nozawa and colleagues found that inhibition of estrone-sulfate transporter activity led to suppression of MCF-7 breast cancer cell proliferation (22).

Numerous studies have shown that single nucleotide polymorphisms (SNP) in SLCO genes can markedly alter substrate-specific transport efficiency, although few studies have specifically evaluated androgen transport (23). Hamada and colleagues recently showed that a nonsynonymous SNP (rs4149117) in SLCO1B3 influenced testosterone uptake in transfected cells and was correlated with prostate cancer outcomes (13). Cells transfected with 2 copies of the G allele in rs4149117 showed impaired testosterone transport. Among 180 men with CRPC, those homozygous for the GG haplotype had a longer median survival (8.5 vs. 6.4 years; P = 0.02) and significantly improved 10-year survival (42% vs. 23%; P = 0.023). The authors also showed increased OATP1B3 staining in primary prostate cancers compared with faint staining in samples of normal tissue or benign prostatic hyperplasia. Sharifi and colleagues showed that in 68 men treated with hormone suppression for biochemical relapse or metastatic disease, those with 2 copies of the G allele in this SLCO1B3 SNP had a longer time to androgen-independent progression (1.2 vs. 1.6 years; P < 0.05; ref. 12).

The identification of SLCO genes in breast cancer cells and the association of an SLCO genotype with response to ADT in men with advanced prostate cancer suggest that transport genes mediating cellular steroid uptake may influence the growth and/or progression of hormone-dependent tumors. The extent to which SLCO family gene members are expressed in untreated primary tumors or in CRPC metastases has not been previously reported. Moreover, whether genetic variation in these genes is associated with prostate cancer survival in men with less advanced disease is unknown. We sought to determine expression of SLCO gene family members in untreated primary prostate cancer and in CPRC metastases and to determine the association of allelic variation in SLCO2B1 and SLCO1B3 with prostate cancer outcomes in a large population-based cohort of men with prostate cancer.

Tissue analysis

Tissue acquisition.

All procedures involving human subjects were approved by the Institutional Review Board (IRB) of the University of Washington Medical Center. Matched samples of benign and tumor prostate tissues were obtained from 8 untreated patients undergoing radical prostatectomy (RP) for localized prostate cancer. Castration-resistant tumors from 14 men with CRPC were obtained by rapid autopsy under the aegis of the University of Washington Prostate Cancer Donor Autopsy Program as previously described (24). Autopsies were done within 4 to 10 hours of death. Fresh tissue was embedded in freezing media (Tissue-Tek OCT Compound; Sakura Finetek) immediately after harvesting, snap frozen in liquid nitrogen, and maintained at −80°C. Tissues utilized in this study included distant or local soft tissue metastases acquired from surgically or medically castrated patients (the latter with anorchid serum testosterone levels documented at <50 ng/dL) and included 2 to 4 metastatic lymph node, liver, bladder, or lung tumor deposits per patient. Bone metastases were not included because of inconsistent RNA quality.

RNA isolation.

Areas consisting of more than 85% tumor tissue were grossly macrodissected from 8 prostatectomy specimens and 22 metastases from 5 CRPC patients. Benign prostate samples were macrodissected after being examined by hematoxylin and eosin staining to ensure the absence of tumor tissue and homogenized in TRIzol for isolation of total RNA and conversion to cDNA as previously described (10). Laser capture microdissection was carried out on an additional 69 metastases from 14 CRPC patients. All RNA samples were DNase treated and amplified using oligo(dT) priming for first-strand cDNA synthesis (to minimize priming off residual genomic DNA). Yield, purity, and integrity of total and amplified RNA were determined by optical density measurements at wavelengths of 260 and 280 nm (OD260/280) and gel electrophoresis. The quality of cDNA conversion was assessed in studies showing the amplification of genes encoding abundant [prostate specific antigen (PSA), AR] versus rare transcripts (SRD5A2, steroidogenic enzymes) as previously described (10).

Quantitative RT-PCR.

Reactions were done in triplicate using an Applied Biosystems 7700 sequence detector with approximately 5 ng of cDNA, 1 μmol/L of each primer pair, and SYBR Green PCR master mix (Applied Biosystems). Primers specific for genes of interest were designed using the Web-based primer design service Primer3 provided by the Whitehead Institute for Biomedical Research (25). Sequences are provided in Supplementary Table S1. The specificity of amplification in each reaction was assessed on the basis of the melting point of the dissociation curve, and reactions with cycle threshold (Ct) values of more than 35 were considered undetectable for that transcript. The mean Ct obtained for each gene was normalized to the expression of the housekeeping gene RPL13A in the same sample (the ΔCt). Samples with Ct above 35 were incorporated into the data analysis by using a Ct value of 36 for purposes of calculation. The percentage of samples that were undetectable varied for each gene (from 0% in SLCO2A1, SLCO2B1, SLCO3A1, and SLCO4A1 to 8% in SLCO1B3; 13% in SLCO4C1 and SLCO5A1; 28% in SLCO1A2, and 39% in SLCO1B1) and did not represent a unique population. The Wilcoxan rank-sum test was used to compare the mean ΔCts for each gene between the primary prostate cancers (n = 8) and metastatic autopsy samples (n = 16–22). The values of P < 0.05 were considered significant. The fold change was calculated from the difference in mean ΔCts between the sample groups (ΔΔCt method; fold = 2ΔΔCt).

Genotype analysis

Study population.

The analyses of SLCO2B1 and SLCO1B3 genetic variation and prostate cancer outcomes utilized data and DNA from patients (residents of King County, Washington) in one of 2 population-based studies of prostate cancer risk factors. Incident cases (histologically confirmed prostate cancer ascertained from the Seattle-Puget Sound SEER cancer registry) were diagnosed between January 1, 1993, and December 31, 1996 (study I), or between January 1, 2002, and December 31, 2005 (study II). A total of 2,244 eligible cases were identified and 1,754 (78%) participated in the study interview. Only Caucasian patients with DNA available (n = 1,309) were included in these analyses. All study procedures and protocols were approved by the IRB of the Fred Hutchinson Cancer Research Center. In addition, genotyping was approved by the IRB of the National Human Genome Research Institute.

Genotyping.

DNA was isolated from peripheral blood samples and stored at −80°C. SNPs in SLCO1B3 and SLCO2B1 were selected using publicly available data from the Genome Variation Server (26). Haplotype tagging SNPs (tagSNP) with a minor allele frequency of greater than 5.0% were selected to maximize coverage of the transcript of interest (+5 kb upstream and downstream). SNP genotype was determined using the Applied Biosystems (ABI) SNPlex Genotyping System, and allele calling was done with GeneMapper software (27). A total of 7 (SLCO2B1) and 5 (SLCO1B3) SNPs were selected. Genotyping was successful in 98.0% to 99.4% of SNPs. Genotyping of blind duplicate samples (n = 141) was used for quality control with 100% agreement. The SNPs rs2306168 (SLCO2B1) and rs10743397 (SLCO1B3) had no genetic variability in our population and were excluded. For SLCO2B1, the SNPs analyzed included 4 tagSNPs, 1 SNP from the 5′-untranslated region (rs2851069), and a coding nonsynonymous SNP (rs12422149). For SLCO1B3, all SNPs were tagSNPs, including 2 coding nonsynonymous SNPs (rs7311358 and rs4149117) in perfect linkage disequilibrium (for which only the rs4149117 data are reported).

Data collection.

Study patients completed standardized in-person interviews providing information on demographic and lifestyle factors, medical and family history, and number of PSA tests within the 5 year before diagnosis. Clinical information on prostate cancer cases was obtained from the SEER (Surveillance, Epidemiology, and End Results) cancer registry, including Gleason score, tumor stage, diagnostic PSA, and primary therapy.

Prostate cancer outcomes.

All living cases who consented to future contact from study I (n = 631) were sent a follow-up survey in January 2004. Information on physician diagnoses of prostate cancer recurrence/progression, secondary therapy use, follow-up PSA values, and procedures such as subsequent biopsies and scans was collected along with dates and results. Outcomes were confirmed with review of medical records. Five hundred twenty men completed the survey (83%). Those not completing the survey were younger at diagnosis (<50 yrs), African American, and had less education than responders (P values <0.05). No significant differences existed between tumor stage, Gleason score, or primary treatment.

Disease recurrence/progression events were determined by multiple criteria. Date of recurrence was imputed for 13 patients who died from prostate cancer but had unknown recurrence date, with a multiple imputations methodology as previously described (28, 29). PSA progression definitions varied by primary treatment: a follow-up PSA value of 0.2 or greater in men who underwent RP; nadir PSA + 2 ng/mL (Phoenix criteria; ref. 30) for men treated with radiation (XRT); or any PSA increase in men treated with primary ADT. The date of last follow-up for recurrence/progression events was December 31, 2005. Prostate cancer–specific mortality (PCSM) was determined from the SEER registry, which links quarterly with the Washington State mortality database. Underlying cause of death was verified by a review of death certificates, with 99% agreement for prostate cancer–specific deaths. The date of last follow-up for mortality was December 1, 2009.

Statistical analysis

SNP alleles were in Hardy–Weinberg equilibrium by the Fisher exact test except for rs4944993 (SLCO2B1), which was then excluded from the analyses. Multivariate Cox proportional hazards models were created to estimate the HRs and 95% CIs for prostate cancer recurrence/progression and PCSM adjusting for age, PSA level at diagnosis, Gleason score, stage, primary treatment, body mass index (BMI), and smoking status. All statistical analyses were conducted using STATA software (Version 11; Stata, Inc.).

Increased expression of SLCO genes in CRPC metastases versus primary prostate cancers

To determine whether SLCO gene expression and therefore OATP transport activity might be present in prostate tumors, we used quantitative RT-PCR to evaluate the expression of these 11 genes in untreated prostate tumors and in advanced CRPC metastases. Transcripts encoding 9 of the 11 genes were reproducibly detected in prostate tumor samples and the CRPC metastases (Table 1). We did not find a significant difference in expression between the matched benign and cancer prostate samples for any of the SLCO genes (data not shown; of note, despite careful macrodissection, contamination of cancer samples with benign epithelium cannot be ruled out). However, compared with the untreated primary prostate cancers, CRPC metastases showed significantly increased expression of genes encoding 6 SLCO family members (Table 1). Dot plots depicting the expression level of these 6 transcripts in the CRPC metastases and in the untreated prostate tissues are shown in Figure 1. Interestingly, SLCO1B1 was detected in only 3 of the benign prostate samples (at Ct < 35) but was detectable in a majority of the CRPC metastases.

Figure 1.

Transcript levels for the indicated SLCO genes in matched normal prostate (NP) tissue and prostate cancer tissue (CP) from untreated men and from CRPC metastases (Mets). Ct for each gene was normalized to the housekeeping gene RPL13A in the same sample. The y-axis is the RPL13A-normalized Ct; more positive numbers reflect higher transcript abundance. The fold change was calculated by the ΔΔCt method (fold = 2ΔΔCt). The Wilcoxan rank-sum test was used to compare the mean Cts for each gene between the prostate cancer and metastatic tumor (Mets) samples. Values of P < 0.05 were considered significant. Significant differences in expression between the CRPC Mets and prostate cancer samples were not observed for SLCO1A2, SLCO4C1, and SLCO5A1 (data not shown). Gray circles are samples for which a Ct value of 36 was used for purposes of calculation.

Figure 1.

Transcript levels for the indicated SLCO genes in matched normal prostate (NP) tissue and prostate cancer tissue (CP) from untreated men and from CRPC metastases (Mets). Ct for each gene was normalized to the housekeeping gene RPL13A in the same sample. The y-axis is the RPL13A-normalized Ct; more positive numbers reflect higher transcript abundance. The fold change was calculated by the ΔΔCt method (fold = 2ΔΔCt). The Wilcoxan rank-sum test was used to compare the mean Cts for each gene between the prostate cancer and metastatic tumor (Mets) samples. Values of P < 0.05 were considered significant. Significant differences in expression between the CRPC Mets and prostate cancer samples were not observed for SLCO1A2, SLCO4C1, and SLCO5A1 (data not shown). Gray circles are samples for which a Ct value of 36 was used for purposes of calculation.

Close modal
Table 1.

Expression of SLCO genes in castration resistant prostate cancer

FamilyPrior nameGeneProteinFoldaPb
OATP1 OATP-A, OATP1 SLCO1A2 OATP1A2 1.1 ns 
 OATP-C, OATP2 SLCO1B1 OATP1B1 13.8 0.0053 
 OATP8 SLCO1B3 OATP1B3 3.6 0.0517 
 OATP-F SLCO1C1 OATP1C1 nd nd 
OATP2  SLCO2A1 OATP2A1 3.2 0.0158 
 OATP-B SLCO2B1 OATP2B1 5.5 0.0034 
OATP3 OATP-D SLCO3A1 OATP3A1 5.4 0.0053 
OATP4 OATP-E SLCO4A1 OATP4A1 30 0.0001 
 OATP-H SLCO4C1 OATP4C1 1.9 ns 
OATP5 OATP-J SLCO5A1 OATP5A1 1.1 ns 
OATP6 OATP-I SLCO6A1 OATP6A1 nd nd 
FamilyPrior nameGeneProteinFoldaPb
OATP1 OATP-A, OATP1 SLCO1A2 OATP1A2 1.1 ns 
 OATP-C, OATP2 SLCO1B1 OATP1B1 13.8 0.0053 
 OATP8 SLCO1B3 OATP1B3 3.6 0.0517 
 OATP-F SLCO1C1 OATP1C1 nd nd 
OATP2  SLCO2A1 OATP2A1 3.2 0.0158 
 OATP-B SLCO2B1 OATP2B1 5.5 0.0034 
OATP3 OATP-D SLCO3A1 OATP3A1 5.4 0.0053 
OATP4 OATP-E SLCO4A1 OATP4A1 30 0.0001 
 OATP-H SLCO4C1 OATP4C1 1.9 ns 
OATP5 OATP-J SLCO5A1 OATP5A1 1.1 ns 
OATP6 OATP-I SLCO6A1 OATP6A1 nd nd 

Abbreviations: nd, not detected; ns, not significant.

aCRPC metastases versus untreated prostate cancer.

bWilcoxan rank-sum test.

Prevalent expression of multiple SLCO gene members in primary and metastatic CRPC tumors

To more carefully examine the range and distribution of SLCO gene expression in advanced prostate tumors, we expanded the initial cohort of CRPC tumors (isolated by gross macrodissection) to include 69 tumor samples (isolated by laser capture microdissection) from 14 men with CRPC. For each subject, we evaluated 2 to 10 soft tissue metastases and 1 to 2 primary samples (prostatic tumor simultaneously resected with the metastases at time of autopsy) to assess whether expression differed in the primary versus metastatic environment.

We found widespread but variable expression of multiple SLCO gene members across the CRPC samples (Fig. 2), with certain patients consistently expressing multiple SLCO transporters in their tumor samples (e.g., patients 1–4) and others showing low levels of SLCO gene expression (e.g., patients 11–14). Grouping of samples by patient of origin showed that gene expression was generally consistent across the metastatic samples of an individual patient, although not necessarily concordant between the primary tumor and metastases (paired analyses did not show consistent differences in primary vs. metastatic tumor deposits; data not shown). The primary and metastatic samples from individual patients occasionally showed strikingly different patterns of SLCO gene expression, consistent with the known heterogeneity that characterizes prostate cancer metastases (24, 31). For example, patients 6 and 7 had relatively low SLCO expression in the prostatic tumor samples harvested at autopsy but abundant expression of multiple SLCO genes in their metastases. In contrast, patients 10 and 11 expressed multiple SLCO gene members in their prostatic samples without marked expression of these genes in their metastases. These data are consistent with the hypothesis that in patients whose tumors have strong SLCO gene expression, steroid uptake proteins may play a role CRPC progression.

Figure 2.

Expression of SLCO genes by quantitative RT-PCR in primary prostatic samples (P) and in metastatic (M) tumors obtained from 14 men with CRPC. Primary and metastatic samples were simultaneously harvested at autopsy and are grouped by patient of origin. The heatmap depicts the mean centered expression of each gene across all samples. The scale is from bright green (lowest expression) to black (equivalent expression) to bright red (highest expression). Gray squares denote samples for which no transcript was detectable. Patients with prevalent versus rare expression of SLCO genes are as denoted.

Figure 2.

Expression of SLCO genes by quantitative RT-PCR in primary prostatic samples (P) and in metastatic (M) tumors obtained from 14 men with CRPC. Primary and metastatic samples were simultaneously harvested at autopsy and are grouped by patient of origin. The heatmap depicts the mean centered expression of each gene across all samples. The scale is from bright green (lowest expression) to black (equivalent expression) to bright red (highest expression). Gray squares denote samples for which no transcript was detectable. Patients with prevalent versus rare expression of SLCO genes are as denoted.

Close modal

Association of SLCO2B1 and SLCO1B3 polymorphisms with prostate cancer risk and PCSM

A cohort of 1,309 Caucasian prostate cancer cases from the population-based studies described earlier had DNA available for genotyping, with selected characteristics shown in Table 2. Most cases were diagnosed with localized disease (78%), had Gleason score of (3 + 4) or less (85%) and a PSA value at diagnosis of less than 10.0 ng/mL (69%). For the recurrence/progression analysis, 469 Caucasian men had genotyping results available for analysis. Of these, 30% had a recurrence/progression event (n = 143). The median follow-up time for recurrence/progression was 8.9 years (range: 0.1–12.8 years). For PCSM, the median follow-up time was 7.0 years (range: 0.8–16.4 years).

Table 2.

Distribution of selected characteristics of Caucasian population-based prostate cancer cases used for SLCO2B1 and SLCO1B3 SNP genotype analyses

  
Total cases 1,309 
Age, y  
 40–49 102 (7.8) 
 50–54 189 (14.4) 
 55–59 325 (24.8) 
 60–64 395 (30.2) 
 65–69 153 (11.7) 
 70–74 145 (11.1) 
Family history of prostate cancera  
 Negative 1,026 (78.4) 
 Positive 283 (21.6) 
PSA tests within 5 y prior to diagnosis  
 None 288 (22.1) 
 1–2 320 (24.5) 
 3+ 638 (48.7) 
 Unknown 63 (4.8) 
BMI (weight/height2 
 <25.0 429 (32.8) 
 25.0–29.9 638 (48.7) 
 ≥30.0 242 (18.5) 
Smoking status  
 Never 523 (40.0) 
 Former 631 (48.2) 
 Current 155 (11.8) 
Diagnostic PSA, ng/mL  
 0.0–3.9 178 (13.6) 
 4.0–9.9 722 (55.2) 
 10.0–19.9 191 (14.6) 
 ≥20 118 (9.0) 
 Missing 100 (7.6) 
Gleason score  
 2–6, 3 + 4 1103 (84.5) 
 4 + 3, 8–10 202 (15.5) 
Tumor stage  
 Local 1,023 (78.2) 
 Regional 254 (19.4) 
 Distant 32 (2.4) 
Treatment  
 RP 770 (58.8) 
 XRT ± hormones 359 (27.4) 
 AD 61 (4.7) 
 Other 4 (0.3) 
 Watchful waiting 115 (8.8) 
  
Total cases 1,309 
Age, y  
 40–49 102 (7.8) 
 50–54 189 (14.4) 
 55–59 325 (24.8) 
 60–64 395 (30.2) 
 65–69 153 (11.7) 
 70–74 145 (11.1) 
Family history of prostate cancera  
 Negative 1,026 (78.4) 
 Positive 283 (21.6) 
PSA tests within 5 y prior to diagnosis  
 None 288 (22.1) 
 1–2 320 (24.5) 
 3+ 638 (48.7) 
 Unknown 63 (4.8) 
BMI (weight/height2 
 <25.0 429 (32.8) 
 25.0–29.9 638 (48.7) 
 ≥30.0 242 (18.5) 
Smoking status  
 Never 523 (40.0) 
 Former 631 (48.2) 
 Current 155 (11.8) 
Diagnostic PSA, ng/mL  
 0.0–3.9 178 (13.6) 
 4.0–9.9 722 (55.2) 
 10.0–19.9 191 (14.6) 
 ≥20 118 (9.0) 
 Missing 100 (7.6) 
Gleason score  
 2–6, 3 + 4 1103 (84.5) 
 4 + 3, 8–10 202 (15.5) 
Tumor stage  
 Local 1,023 (78.2) 
 Regional 254 (19.4) 
 Distant 32 (2.4) 
Treatment  
 RP 770 (58.8) 
 XRT ± hormones 359 (27.4) 
 AD 61 (4.7) 
 Other 4 (0.3) 
 Watchful waiting 115 (8.8) 

aFirst-degree family history of prostate cancer.

Tables 3 and 4 show the results for the multivariate Cox analysis for recurrence/progression and PCSM, respectively. None of the SNPs were associated with the risk of recurrence/progression. Death due to prostate cancer was observed in 66 men (5% of cases). A total of 115 men who died of other causes were censored at time of death. Excluding these 115 from the analysis did not alter the results and thus they were included in the final models. Having 1 or 2 copies of the variant A allele in SLCO2B1 SNP rs12422149 was associated with a 2-fold increased risk of PCSM (HR: 1.99; 95% CI: 1.11–3.55). Similarly, carriers of the variant T allele in SLCO1B3 SNP rs4149117 had a 76% increased risk of PCSM (HR: 1.76; 95% CI: 1.00–3.08). None of the other SNPs was associated with PCSM. There was no evidence for effect modification in those who received primary ADT.

Table 3.

Risk of prostate cancer recurrence/progression by SLCO2B1 and SLCO1B3 SNP genotypes in a population-based prostate cancer cohorta

 SNPGenotypeRecurrence/progressionHR (95% CI)b
No (n = 326)Yes (n = 143)
SLCO2B1 rs12422149 GG 254 (69.2) 113 (30.8) 1.00 (referent) 
  GA + AA 47 (66.2) 24 (33.8) 1.09 (0.68–1.75) 
SLCO2B1 rs949069 GG 216 (69.5) 95 (30.6) 1.00 (referent) 
  GA + AA 87 (66.4) 44 (33.6) 1.19 (0.82–1.73) 
SLCO2B1 rs2712819 AA 194 (70.6) 81 (29.5) 1.00 (referent) 
  AG + GG 109 (65.7) 57 (34.3) 1.35 (0.95–1.91) 
SLCO2B1 rs2851069 CC 116 (70.3) 49 (29.7) 1.00 (referent) 
  CT + TT 184 (66.9) 91 (33.1) 1.31 (0.92–1.89) 
SLCO2B1 rs7947726 GG 259 (69.1) 116 (30.9) 1.00 (referent) 
  AG + AA 44 (67.7) 21 (32.3) 1.21 (0.75–1.95) 
SLCO1B3 rs4149117 GG 221 (68.0) 104 (32.0) 1.00 (referent) 
  GT + TT 84 (70.0) 36 (30.0) 0.85 (0.57–1.26) 
SLCO1B3 rs3829311 GG 206 (68.0) 97 (32.0) 1.00 (referent) 
  GA + AA 98 (69.5) 43 (30.5) 0.83 (0.57–1.21) 
SLCO1B3 rs4762803 CC 239 (68.3) 111 (31.7) 1.00 (referent) 
  CG + GG 66 (70.2) 28 (29.8) 0.86 (0.56–1.32) 
 SNPGenotypeRecurrence/progressionHR (95% CI)b
No (n = 326)Yes (n = 143)
SLCO2B1 rs12422149 GG 254 (69.2) 113 (30.8) 1.00 (referent) 
  GA + AA 47 (66.2) 24 (33.8) 1.09 (0.68–1.75) 
SLCO2B1 rs949069 GG 216 (69.5) 95 (30.6) 1.00 (referent) 
  GA + AA 87 (66.4) 44 (33.6) 1.19 (0.82–1.73) 
SLCO2B1 rs2712819 AA 194 (70.6) 81 (29.5) 1.00 (referent) 
  AG + GG 109 (65.7) 57 (34.3) 1.35 (0.95–1.91) 
SLCO2B1 rs2851069 CC 116 (70.3) 49 (29.7) 1.00 (referent) 
  CT + TT 184 (66.9) 91 (33.1) 1.31 (0.92–1.89) 
SLCO2B1 rs7947726 GG 259 (69.1) 116 (30.9) 1.00 (referent) 
  AG + AA 44 (67.7) 21 (32.3) 1.21 (0.75–1.95) 
SLCO1B3 rs4149117 GG 221 (68.0) 104 (32.0) 1.00 (referent) 
  GT + TT 84 (70.0) 36 (30.0) 0.85 (0.57–1.26) 
SLCO1B3 rs3829311 GG 206 (68.0) 97 (32.0) 1.00 (referent) 
  GA + AA 98 (69.5) 43 (30.5) 0.83 (0.57–1.21) 
SLCO1B3 rs4762803 CC 239 (68.3) 111 (31.7) 1.00 (referent) 
  CG + GG 66 (70.2) 28 (29.8) 0.86 (0.56–1.32) 

aRecurrence/progression data available for 469 cases, variable number of cases due to failed genotyping.

bAdjusted for age, PSA at diagnosis, Gleason score, tumor stage, primary treatment, BMI, and smoking status.

Table 4.

Risk of PCSM by SLCO2B1 and SLCO1B3 SNP genotypes in a population-based prostate cancer cohorta

 SNPGenotypeProstate cancer deathHR (95% CI)c
NobYes
SLCO2B1 rs12422149 GG 993 (95.6) 46 (4.4) 1.00 (referent) 
  GA + AA 210 (92.5) 17 (7.5) 1.99 (1.11–3.55) 
SLCO2B1 rs949069 GG 811 (95.5) 38 (4.5) 1.00 (referent) 
  GA + AA 397 (93.9) 26 (6.2) 1.39 (0.82–2.34) 
SLCO2B1 rs2712819 AA 777 (95.0) 41 (5.0) 1.00 (referent) 
  AG + GG 431 (95.1) 22 (4.9) 1.02 (0.60–1.77) 
SLCO2B1 rs2851069 CC 452 (95.2) 23 (4.8) 1.00 (referent) 
  CT + TT 752 (94.7) 42 (5.3) 1.21 (0.72–2.05) 
SLCO2B1 rs7947726 GG 999 (95.1) 51 (4.9) 1.00 (referent) 
  AG + AA 204 (94.0) 13 (6.0) 0.75 (0.37–1.53) 
SLCO1B3 rs4149117 GG 885 (95.4) 43 (4.6) 1.00 (referent) 
  GT + TT 326 (93.7) 22 (6.3) 1.76 (1.00–3.08) 
SLCO1B3 rs3829311 GG 812 (95.0) 43 (5.0) 1.00 (referent) 
  GA + AA 396 (94.7) 22 (5.3) 1.28 (0.74–2.23) 
SLCO1B3 rs4762803 CC 958 (94.8) 53 (5.2) 1.00 (referent) 
  CG + GG 251 (95.4) 12 (4.6) 1.16 (0.59–2.30) 
 SNPGenotypeProstate cancer deathHR (95% CI)c
NobYes
SLCO2B1 rs12422149 GG 993 (95.6) 46 (4.4) 1.00 (referent) 
  GA + AA 210 (92.5) 17 (7.5) 1.99 (1.11–3.55) 
SLCO2B1 rs949069 GG 811 (95.5) 38 (4.5) 1.00 (referent) 
  GA + AA 397 (93.9) 26 (6.2) 1.39 (0.82–2.34) 
SLCO2B1 rs2712819 AA 777 (95.0) 41 (5.0) 1.00 (referent) 
  AG + GG 431 (95.1) 22 (4.9) 1.02 (0.60–1.77) 
SLCO2B1 rs2851069 CC 452 (95.2) 23 (4.8) 1.00 (referent) 
  CT + TT 752 (94.7) 42 (5.3) 1.21 (0.72–2.05) 
SLCO2B1 rs7947726 GG 999 (95.1) 51 (4.9) 1.00 (referent) 
  AG + AA 204 (94.0) 13 (6.0) 0.75 (0.37–1.53) 
SLCO1B3 rs4149117 GG 885 (95.4) 43 (4.6) 1.00 (referent) 
  GT + TT 326 (93.7) 22 (6.3) 1.76 (1.00–3.08) 
SLCO1B3 rs3829311 GG 812 (95.0) 43 (5.0) 1.00 (referent) 
  GA + AA 396 (94.7) 22 (5.3) 1.28 (0.74–2.23) 
SLCO1B3 rs4762803 CC 958 (94.8) 53 (5.2) 1.00 (referent) 
  CG + GG 251 (95.4) 12 (4.6) 1.16 (0.59–2.30) 

aVariable number of cases due to failed genotyping.

bIncludes 115 men who died of nonprostate cancer causes and who were censored at time of death.

cAdjusted for age, PSA at diagnosis, Gleason score, stage, primary treatment, BMI, and smoking status.

The source of residual androgens present in prostate cancer tissues despite ADT has not been elucidated but may reflect the uptake of adrenal androgens and intracellular conversion to testosterone or de novo androgen synthesis from the uptake of cholesterol or progesterone precursors. Although the uptake of steroid hormones is generally attributed to free diffusion across the lipid membrane, emerging data suggest a role for steroid transport proteins in actively mediating androgen uptake into prostate cells and thereby potentially influencing prostate cancer outcomes.

In this study, we evaluated expression of SLCO-encoded membrane transporters both in untreated primary prostate cancer and in CPRC metastases and evaluated genetic variation of SLCO2B1 and SLCO1B3 in association with prostate cancer outcomes. We found significantly increased expression of many SLCO family members in CRPC metastases compared with untreated primary prostate cancer. Moreover, individuals could be identified whose tumors were characterized by widespread or minimal transporter gene expression. An important limitation of gene expression data is that transcript level does not necessarily equate with protein expression; however, our findings are consistent with a hypothesis that in certain individuals, steroid transport proteins may play a role in the tumor androgen levels observed in CRPC metastases. In our analysis of data from a population-based cohort of prostate cancer patients, we found that 2 SNPs (SLCO2B1 SNP rs12422149 and SLCO1B3 SNP rs4149117) were associated with an increased risk of PCSM.

The increased risk of PCSM in carriers of the T versus G allele in SLCO1B3 SNP rs4149117 (334T>G; adjusted HR: 1.76; 95% CI: 1.00–3.08), although of borderline significance, is consistent with the study of Hamada and colleagues (13). In their study of 180 men with CRPC, the HR for overall mortality was 1.57 (95% CI: 1.11–2.24) for men carrying the T allele in rs4149117. Notably, they showed that SLCO1B3-transfected cells actively transported testosterone, that testosterone uptake was higher for carriers of the T versus G allele, and that men homozygous for the 334GG genotype (associated with impaired testosterone uptake) had a longer median survival (8.5 vs. 6.4 years; P = 0.02) and improved 10-year survival (42% vs. 23%; P = 0.023). In a separate study of 68 men treated with ADT for advanced prostate cancer, those with 2 copies of the G allele in rs4149117 had a longer time to androgen independence (1.6 vs. 1.2 years; P < 0.05; ref. 12). Of potential importance is that in the Caucasian population comprising these reports (and ours), SNP rs4149117 (334T>G) is in perfect linkage disequilibrium with SNP rs7311358 (699G>A). The study of Hamada and colleagues found that only the double-mutant haplotype showed impaired testosterone transport in vivo. Thus, different outcomes might be observed in non-Caucasian populations in which the SNPs may vary independently.

We also found an increased risk of PCSM in carriers of the A versus G allele in SLCO2B1 SNP rs12422149 (935A>G; adjusted HR: 1.99; 95% CI: 1.11–3.55). This coding, nonsynonymous SNP has been reported in in vitro studies to have altered cellular uptake of certain medications (i.e., montelukast) and DHEA-S (dehydroepiandrosterone sulfate; ref. 32, 33). In a study of 538 castrate men with advanced prostate cancer receiving ADT (a different setting than the population-based cohort of incident prostate cancer cases in our study), the A allele in SLCO2B1 SNP rs12422149 was associated with a longer time to progression (HR for progression: 1.40; 95% CI: 1.06–1.84 for the G vs. A allele; Phil Kantoff, personal communication). Interestingly, allelic variation in DHEA-S uptake by SLCO2B1 (and accordingly, any impact of cellular DHEA-S uptake on prostate cancer outcome) might be anticipated to differ in the eugonadal and castrate settings, as testosterone has been shown to inhibit SLCO2B1-mediated transport of DHEA-S (23). Thus, the association of SLCO2B1 with prostate cancer outcome may reflect the transport of DHEA-S in advanced disease but that of an alternative substrate in incident prostate cancer cases.

We did not find any associations between SLCO1B3 or SLCO2B1 genotypes and prostate cancer recurrence/progression. This may be due to several reasons. First, recurrence/progression data were only available for one third of the cohort, which limited our power to detect a difference. Second, there can be significant variation (up to 35%) in reported PSA progression rates depending on how it is defined (34). Furthermore, considering the natural history of prostate cancer, many men who experience recurrence/progression events will not die of their disease (35). Finally, consistent with our findings of increased expression of SLCO gene members in metastatic tumors compared with primary tumors, it may be that alteration in SLCO activity is important in men with advanced, end-stage tumors but not in men with early recurrence, which may be due to other mechanisms.

Our data from a population-based cohort of prostate cancer patients suggest that genetic variation in the SLCO transport gene family is associated with prostate cancer mortality. These findings are consistent with results from studies of men with advanced prostate cancer, suggesting that genetic variation in SLCO genes may allow stratification of patients at higher risk for PCSM or poor response to ADT for consideration of either more aggressive hormonal therapies or incorporation of nonhormonal-based treatment strategies. In addition, we observe significantly increased expression of SLCO gene members in CRPC metastases, suggesting steroid transport proteins may contribute to the elevated tumoral androgen levels observed during CRPC progression. Future studies include immunohistochemical assessment of OATP expression in primary and metastatic prostate cancer tissues, and using in vivo models to evaluate the ability of SLCO transporters to measurably alter intratumoral androgen levels. The family of OATP/SLCO steroid transport proteins may serve as novel biomarkers of response to ADT, and common genetic variants in these genes may be associated with enhanced risk of prostate cancer mortality. Thus, their role in prostate cancer warrants further investigation.

No potential conflicts of interest were disclosed.

The authors thank Peter Nelson and Philip Kantoff for thoughtful discussion; Ruth Dumpit, Tom Kim, Ilsa Coleman, Roger Coleman, Jared Lucas, and Andrew Morgan for expert technical assistance; and Alex Moreno for administrative support.

Prostate Cancer Foundation (Career Development Award to E.A. Mostaghel and Synergy Award to R.B. Montgomery); Damon Runyon Cancer Research Foundation (Damon Runyon-Genentech Clinical Investigator Award CI-40-08 to E.A. Mostaghel); NIH (Pacific Northwest Prostate Cancer SPORE P50 CA97186 (J.L. Stanford, Career Development Award to J.L. Wright, and Pilot Project Award to E.A. Mostaghel); R01 CA056678 (J.L. Stanford); R01 CA082664 (J.L. Stanford); R01 CA092579 (J.L. Stanford); T32 CA009168-30 (J.L. Wright); with additional support from the Fred Hutchinson Cancer Research Center and the Intramural Program of the National Human Genome Research Institute (E.M. Kwon and E.A. Ostrander).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Nelson
PS
,
Clegg
N
,
Arnold
H
,
Ferguson
C
,
Bonham
M
,
White
J
, et al
The program of androgen-responsive genes in neoplastic prostate epithelium
.
Proc Natl Acad Sci U S A
2002
;
99
:
11890
5
.
2.
Pritchard
CC
,
Nelson
PS
. 
Gene expression profiling in the developing prostate
.
Differentiation
2008
;
76
:
624
40
.
3.
Visakorpi
T
,
Hyytinen
E
,
Koivisto
P
,
Tanner
M
,
Keinänen
R
,
Palmberg
C
, et al
In vivo amplification of the androgen receptor gene and progression of human prostate cancer
.
Nat Genet
1995
;
9
:
401
6
.
4.
Chen
CD
,
Welsbie
DS
,
Tran
C
,
Baek
SH
,
Chen
R
,
Vessella
R
, et al
Molecular determinants of resistance to antiandrogen therapy
.
Nat Med
2004
;
10
:
33
9
.
5.
Feldman
BJ
,
Feldman
D
. 
The development of androgen-independent prostate cancer
.
Nat Rev Cancer
2001
;
1
:
34
45
.
6.
Geller
J
,
Albert
J
,
Nachtsheim
D
,
Loza
D
,
Lippman
S
. 
Steroid levels in cancer of the prostate—markers of tumor differentiation and adequacy of anti-androgen therapy
.
Prog Clin Biol Res
1979
;
33
:
103
11
.
7.
Mohler
JL
,
Gregory
CW
,
Ford
OH
 III
,
Kim
D
,
Weaver
CM
,
Petrusz
P
, et al
The androgen axis in recurrent prostate cancer
.
Clin Cancer Res
2004
;
10
:
440
8
.
8.
Page
ST
,
Lin
DW
,
Mostaghel
EA
,
Hess
DL
,
True
LD
,
Amory
JK
, et al
Persistent intraprostatic androgen concentrations after medical castration in healthy men
.
J Clin Endocrinol Metab
2006
;
91
:
3850
6
.
9.
Mostaghel
EA
,
Page
ST
,
Lin
DW
,
Fazli
L
,
Coleman
IM
,
True
LD
, et al
Intraprostatic androgens and androgen-regulated gene expression persist after testosterone suppression: Therapeutic implications for castration-resistant prostate cancer
.
Cancer Res
2007
;
67
:
5033
41
.
10.
Montgomery
RB
,
Mostaghel
EA
,
Vessella
R
,
Hess
DL
,
Kalhorn
TF
,
Higano
CS
, et al
Maintenance of intratumoral androgens in metastatic prostate cancer: a mechanism for castration-resistant tumor growth
.
Cancer Res
2008
;
68
:
4447
54
.
11.
Stanbrough
M
,
Bubley
GJ
,
Ross
K
,
Golub
TR
,
Rubin
MA
,
Penning
TM
, et al
Increased expression of genes converting adrenal androgens to testosterone in androgen-independent prostate cancer
.
Cancer Res
2006
;
66
:
2815
25
.
12.
Sharifi
N
,
Hamada
A
,
Sissung
T
,
Danesi
R
,
Venzon
D
,
Baum
C
, et al
A polymorphism in a transporter of testosterone is a determinant of androgen independence in prostate cancer
.
BJU Int
2008
;
102
:
617
21
.
13.
Hamada
A
,
Sissung
T
,
Price
DK
,
Danesi
R
,
Chau
CH
,
Sharifi
N
, et al
Effect of SLCO1B3 haplotype on testosterone transport and clinical outcome in Caucasian patients with androgen-independent prostatic cancer
.
Clin Cancer Res
2008
;
14
:
3312
8
.
14.
Hagenbuch
B
,
Meier
PJ
. 
The superfamily of organic anion transporting polypeptides
.
Biochim Biophys Acta
2003
;
1609
:
1
18
.
15.
Pizzagalli
F
,
Varga
Z
,
Huber
RD
,
Folkers
G
,
Meier
PJ
,
St-Pierre
MV
. 
Identification of steroid sulfate transport processes in the human mammary gland
.
J Clin Endocrinol Metab
2003
;
88
:
3902
12
.
16.
Tamai
I
,
Nezu
J
,
Uchino
H
,
Sai
Y
,
Oku
A
,
Shimane
M
, et al
Molecular identification and characterization of novel members of the human organic anion transporter (OATP) family
.
Biochem Biophys Res Commun
2000
;
273
:
251
60
.
17.
Alcorn
J
,
Lu
X
,
Moscow
JA
,
McNamara
PJ
. 
Transporter gene expression in lactating and nonlactating human mammary epithelial cells using real-time reverse transcription-polymerase chain reaction
.
J Pharmacol Exp Ther
2002
;
303
:
487
96
.
18.
Ugele
B
,
St-Pierre
MV
,
Pihusch
M
,
Bahn
A
,
Hantschmann
P
. 
Characterization and identification of steroid sulfate transporters of human placenta
.
Am J Physiol Endocrinol Metab
2003
;
284
:
E390
8
.
19.
Valle
LD
,
Toffolo
V
,
Nardi
A
,
Fiore
C
,
Bernante
P
,
Di Liddo
R
, et al
Tissue-specific transcriptional initiation and activity of steroid sulfatase complementing dehydroepiandrosterone sulfate uptake and intracrine steroid activations in human adipose tissue
.
J Endocrinol
2006
;
190
:
129
39
.
20.
Miki
Y
,
Suzuki
T
,
Kitada
K
,
Yabuki
N
,
Shibuya
R
,
Moriya
T
, et al
Expression of the steroid and xenobiotic receptor and its possible target gene, organic anion transporting polypeptide-A, in human breast carcinoma
.
Cancer Res
2006
;
66
:
535
42
.
21.
Al Sarakbi
W
,
Mokbel
R
,
Salhab
M
,
Jiang
WG
,
Reed
MJ
,
Mokbel
K
. 
The role of STS and OATP-B mRNA expression in predicting the clinical outcome in human breast cancer
.
Anticancer Res
2006
;
26
:
4985
90
.
22.
Nozawa
T
,
Suzuki
M
,
Yabuuchi
H
,
Irokawa
M
,
Tsuji
A
,
Tamai
I
. 
Suppression of cell proliferation by inhibition of estrone-3-sulfate transporter in estrogen-dependent breast cancer cells
.
Pharm Res
2005
;
22
:
1634
41
.
23.
Grube
M
,
Köck
K
,
Karner
S
,
Reuther
S
,
Ritter
CA
,
Jedlitschky
G
, et al
Modification of OATP2B1-mediated transport by steroid hormones
.
Mol Pharmacol
2006
;
70
:
1735
41
.
24.
Roudier
MP
,
True
LD
,
Higano
CS
,
Vesselle
H
,
Ellis
W
,
Lange
P
, et al
Phenotypic heterogeneity of end-stage prostate carcinoma metastatic to bone
.
Hum Pathol
2003
;
34
:
646
53
.
25.
Rozen
S
,
Skaletsky
HJ
. 
Primer3 on the WWW for general users and for biologist programmers
.
In:
Krawetz
S
,
Misener
S
,
editors
. 
Bioinformatics Methods and Protocols: Methods in Molecular Biology
.
Totowa, NJ
:
Humana Press
; 
2000
.
p. 365
86
.
Available from:
http://fokker.wi.mit.edu/primer3/.
26.
Genome Variation Server [Internet]
. 
SeattleSNPs Program for Genomic Applications
(Version 5.11) [cited 2009 Sep 10]
.
Seattle, WA
:
Genome Variation Server. Available from:
http://gvs.gs.washington.edu/GVS/.
27.
Applied Biosystems [Internet]
[cited 2009 Sep 10]
.
Carlsbad
,
CA
:
Life Technologies Corporation.
Available from:
www.appliedbiosystems.com.
28.
Wright
JL
,
Kwon
EM
,
Lin
DW
,
Kolb
S
,
Koopmeiners
JS
,
Feng
Z
, et al
CYP17 polymorphisms and prostate cancer outcomes
.
Prostate
2010
;
70
:
1094
101
.
29.
Harrell
F
. 
Regression Modeling Strategies
.
New York
:
Springer
; 
2001
.
30.
Roach
M
 III
,
Hanks
G
,
Thames
H
 Jr
,
Schellhammer
P
,
Shipley
WU
,
Sokol
GH
, et al
Defining biochemical failure following radiotherapy with or without hormonal therapy in men with clinically localized prostate cancer: recommendations of the RTOG-ASTRO Phoenix Consensus Conference
.
Int J Radiat Oncol Biol Phys
2006
;
65
:
965
74
.
31.
Shah
RB
,
Mehra
R
,
Chinnaiyan
AM
,
Shen
R
,
Ghosh
D
,
Zhou
M
, et al
Androgen-independent prostate cancer is a heterogeneous group of diseases: lessons from a rapid autopsy program
.
Cancer Res
2004
;
64
:
9209
16
.
32.
Mougey
EB
,
Feng
H
,
Castro
M
,
Irvin
CG
,
Lima
JJ
. 
Absorption of montelukast is transporter mediated: a common variant of OATP2B1 is associated with reduced plasma concentrations and poor response
.
Pharmacogenet Genomics
2009
;
19
:
129
38
.
33.
Yang
M
,
Oh
WK
,
Xie
W
,
Mostaghel
EA
,
Sun
T
,
Sharifi
N
, et al
Genetic variations in SLCO2B1 and SLCO1B3 and the efficacy of androgen-deprivation therapy in prostate cancer patients
.
In:
ASCO Genitourinary Cancers Symposium
2010
(San Francisco, CA, March 5–7).
34.
Stephenson
AJ
,
Kattan
MW
,
Eastham
JA
,
Dotan
ZA
,
Bianco
FJ
 Jr
,
Lilja
H
, et al
Defining biochemical recurrence of prostate cancer after radical prostatectomy: a proposal for a standardized definition
.
J Clin Oncol
2006
;
24
:
3973
8
.
35.
Collette
L
. 
Prostate-specific antigen (PSA) as a surrogate end point for survival in prostate cancer clinical trials
.
Eur Urol
2008
;
53
:
6
9
.

Supplementary data