Abstract
Background
Prostate cancer is highly influenced by androgens and genes. We investigated whether genetic polymorphisms along the androgen biosynthesis and metabolism pathways are associated with androgen concentrations or risk of prostate cancer or high-grade disease from finasteride treatment.
Methods
A nested case-control study from the Prostate Cancer Prevention Trial using cases drawn from men with biopsy-proven prostate cancer and biopsy-negative, frequency-matched controls was conducted to investigate the association of 51 single nucleotide polymorphisms (SNPs) in 12 genes of the androgen pathway with total, low-grade, and high-grade prostate cancer incidence and serum hormone concentrations.
Results
There were significant associations of genetic polymorphisms in SRD5A1 (rs3736316, rs3822430, rs1560149, rs248797, and rs472402) and SRD5A2 (rs2300700) with risk of high-grade prostate cancer in the placebo arm of the PCPT; two SNPs were significantly associated with increased risk (SRD5A1 rs472402 [OR, 1.70; 95% CI, 1.05-2.75, Ptrend=0.03]; SRD5A2 rs2300700 [OR, 1.94; 95% CI, 1.19-3.18, Ptrend=0.01]). Eleven SNPs in SRD5A1, SRD5A2, CYP1B1, and CYP3A4 were found to be associated with modifying mean serum androgen and sex hormone-binding globulin concentrations; two SNPs (SRD5A1 rs824811 and CYP1B1 rs10012, Ptrend<0.05) consistently and significantly altered all androgen concentrations. Several SNPs (rs3822430, rs2300700; CYP3A43 rs800672; CYP19 rs700519; Ptrend<0.05) were significantly associated with both circulating hormone levels and prostate cancer risk.
Conclusion
Germline genetic variations of androgen-related pathway genes are associated with serum androgen concentrations and risk of prostate cancer. Further studies to examine the functional consequence of novel causal variants are warranted.
Keywords: androgen, metabolism, genetics, polymorphism, prostate cancer
Introduction
Androgens play a critical role in the etiology of prostate cancer, and binding of the potent androgen 5α-dihydrotestosterone (DHT) to the androgen receptor drives the development and progression of the disease. Prostate cancer is highly influenced by genetic variations. Single nucleotide polymorphisms (SNPs) have been found in key regulatory genes involved in the steroid hormone pathway (biosynthesis, activation, metabolism, and degradation of androgens), androgen receptor (AR), and downstream AR effector pathways. These androgen-related SNPs and their putative roles in prostate cancer may contribute to susceptibility to the disease.
The androgen biosynthesis and metabolism pathway is shown in Figure 1. Testosterone is irreversibly converted to DHT by the enzyme steroid 5α-reductase types I and II in the prostate tissue (encoded by the SRD5A1 and SRD5A2 genes, respectively)1 and also by steroid 5α-reductase type III (encoded by SRD5A2L or SRD5A3) recently identified in castration-resistant prostate cancer cells.2, 3 The cytochrome P450 family of enzymes (encoded by the CYP genes) are involved in metabolizing endogenous hormones and their biochemical intermediates.
Figure 1.
Androgen pathway and genes involved in androgen biosynthesis and metabolism. Serum concentrations of androgens measured in the study are boxed.
Abbreviations: DHEA, dehydroepiandrosterone; DHT, dihydrotestosterone; 3α-diol, 3α-androstenediol; 3β-diol, 3β-androstenediol; 3α-dG, 3α-androstanediol glucuronide
We have previously demonstrated that there were no statistically significant associations of baseline serum testosterone, 5α-androstane-3α,17β-diol glucuronide (3α-androstanediol glucuronide, 3α-dG), androstenedione, or sex hormone-binding globulin (SHBG) concentrations with risk of total, low- or high-grade prostate cancer in the placebo4, 5 or finasteride4, 6 treatment arm of the Prostate Cancer Prevention Trial (PCPT), consistent with most results in the literature.7-9 Serum 3α-dG, a distal metabolite of DHT, was used as a surrogate measure of intraprostatic DHT and marker for SRD5A activity10 and serum SHBG as a carrier protein for androgens. While evidence from the above studies has largely been null, others have suggested that genetic modulators of androgen metabolism may affect circulating serum androgen levels11 and subsequently influence prostate carcinogenesis. Moreover, germline polymorphisms within key androgen regulatory genes may contribute to the risk of developing prostate cancer.12, 13
Here, we investigate the association of genetic polymorphisms in genes of the androgen biosynthesis and metabolism pathway with prostate cancer risk using a nested case-control study of the PCPT population. In the PCPT, a randomized, placebo-controlled trial testing whether finasteride could reduce the 7-year period prevalence of prostate cancer, finasteride reduced the risk of prostate cancer by 25%, even though high-grade prostate cancer was more common in the finasteride group.14 Therefore, we genotyped a total of 51 SNPs in the 12 genes involved in the androgen metabolism pathway: SRD5A1, SRD5A2, SRD5A2L/SRD5A3, HSD3B2, HSD17B2, HSD17B3, CYP1B1, CYP3A4, CYP3A5, CYP3A43, CYP17, and CYP19. We also examine whether these polymorphisms are responsible for altering serum androgen concentrations. Findings from this study could provide further insight into the genetics of androgen metabolism, the role of genetics in the etiology of the increased risk of high-grade cancer among men treated with finasteride, as well as improve our understanding of the contribution of genetics to prostate cancer risk.
Methods
Participant and Study Description
All data for this study are from the PCPT (SWOG-9217).14, 15 Details of the trial and participant characteristics have been described previously.14, 15 All men signed informed consent and study procedures were approved by the Institutional Review Boards of the participating 221 study sites.14, 15 The study reported here is part of a large nested case-control study designed to examine multiple hypotheses about prostate cancer risk. Tumors were graded centrally and categorized as low grade = Gleason < 7; high grade = Gleason ≥ 7, retaining the definitions used in the original trial report. Cases were men with biopsy-determined prostate cancer identified either by a for-cause or end-of-study biopsy (N=1809) and controls were selected from men who completed the end-of-study biopsy procedure and had no evidence of prostate cancer (N=1809). All non-white controls were included, and the remaining controls were frequency matched to cases based on treatment arm (finasteride versus placebo), age (in 5-year groupings), and family history of a first-degree family member with prostate cancer. Serum steroid hormone concentrations were available from 1802 cases and 1800 controls. SNP data was available on 1,617 cases and 1,731 controls. The final SNP analysis included white men only as the sample size was too small for sub-group analysis of SNP associations for other racial/ethnic groups (1,506 cases and 1,380 controls). Of these, the population used in the analyses with both serum measures and SNP data available included 1499 cases and 1373 controls.
Genotyping & Serum Measurements
DNA extractions were performed, as described previously.16 Genotyping was conducted by polymerase chain reaction amplification using the Sequenom, TaqMan (Life Technologies), or Illumina platform assays. Total testosterone, 3α-dG, androstenedione and SHBG were quantified in serum by highly specific immunoassays as described previously.4, 6
Statistical Methods
Logistic regression was used to calculate odds ratio (ORs) and 95% confidence intervals (CIs) for risk of total prostate cancer, and polytomous logistic regression was used to calculate ORs and 95% CIs of both low-grade and high-grade prostate cancers. To investigate a possible interaction between SNPs and serum androgen levels and their influence on prostate cancer risk, models were also stratified by androgen concentrations, using the observed median as cut-point. To determine the association between single SNP and androgen levels, mean concentrations of androgen were calculated for each allele, and p-values were calculated using linear regression adjusted for age and family history. For all SNP analyses, the major allele homozygote was used as the reference genotype, and minor alleles were compared to this reference. All statistical tests were two-sided with statistical significance set at p<0.05 and carried out using SAS statistical software (version 9.4, SAS Corporation, Cary, NC).
Additional details of the DNA extraction method, genotyping assay, study description, sample collection, measurements of serum androgens and SHBG concentrations, and statistical methods are described in the Supporting Information.
Results
Table 1 gives the demographic and lifestyle characteristics of the PCPT study population. Cases and controls were similar in age and family history of prostate cancer due to matching; they were also similar with respect to BMI, protein consumption, and mean concentrations of serum androgens and SHBG as well as physical activity, smoking status, and alcohol intake (data not shown). Controls had lower baseline PSA levels than cases and were more likely to be diabetic.
Table 1. Demographics and lifestyle characteristics at baseline, among cases and controls in the Prostate Cancer Prevention Trial, white men only.
| Characteristic | Controls (n=1380) | Cases (n=1506) | P-value |
|---|---|---|---|
|
|
|
||
| Mean (SD) | Mean (SD) | ||
| Age (y) | 63.96 (5.64) | 63.74 (5.62) | 0.29 |
| Prostate specific antigen (ng/mL) | 1.20 (0.76) | 1.54 (0.74) | <0.0001 |
| Protein consumption (g) | 95.37 (38.02) | 93.36 (37.60) | 0.18 |
| 3a-dG (ng/mL) | 6.56 (4.02) | 6.86 (5.28) | 0.09 |
| Testosterone (ng/dL) | 376.78 (131.19) | 379.60 (131.21) | 0.57 |
| SHBG (nmol/L) | 39.03 (15.23) | 39.04 (15.92) | 0.99 |
|
| |||
| N (%) | N (%) | ||
|
| |||
| Treatment Arm | 0.11* | ||
| Finasteride | 532 (38.55) | 625 (41.50) | |
| Placebo | 848 (61.45) | 881 (58.50) | |
| Family history of prostate cancer | 313 (22.68) | 322 (21.38) | 0.40* |
| Diabetes | 75 (5.43) | 60 (3.98) | 0.07 |
| Body mass index (kg/m2) | 0.18 | ||
| Normal (<25) | 345 (25.26) | 425 (28.47) | |
| Overweight (25 to <30) | 745 (54.54) | 770 (51.57) | |
| Obese (>=30) | 276 (20.20) | 298 (19.96) | |
Frequency-matched, so p-values are not interpretable.
Abbreviations: 3α-dG = 3α androstanediol glucuronide; SHBG = Sex hormone-binding globulin
Associations between genetic polymorphisms in genes involved in androgen biosynthesis and metabolism pathways with prostate cancer risk that were statistically significant for either the placebo or finasteride arms are shown in Table 2. In the placebo arm, there was a statistically significant positive association of the homozygous minor variant of SRD5A2 rs2300700 [OR, 1.34; 95% CI, 1.01-1.77, Ptrend=0.04] with risk of overall prostate cancer. When stratified by disease grade [low- vs. high-grade (Gleason score 2-6 vs. 7-10)], there were significant associations of SNPs in SRD5A1 (rs3736316, rs3822430, rs1560149, rs248797, rs472402) and SRD5A2 (rs2300700) with risk of high-grade prostate cancer. The homozygous variants of two SNPs were statistically significantly associated with a >1.5 to 2-fold increased risk of high-grade disease (SRD5A1 rs472402, Ptrend=0.03; SRD5A2 rs2300700, Ptrend=0.01]). The homozygous variant allele of three SNPs reduced the risk of high-grade disease by about half (SRD5A1 rs3736316, Ptrend=0.02; rs3822430, Ptrend=0.02; and rs248797, Ptrend=0.03]. In the finasteride treatment arm, the heterozygous variant of SRD5A1 rs1560149 was associated with a decreased risk of high-grade prostate cancer [OR, 0.59; 95% CI, 0.41-0.84, Ptrend=0.07]. Treatment arms were combined when there was no statistically significant SNP/treatment interaction (Supporting Table S1); the homozygous variant allele of SRD5A2 rs2300700 remained significantly associated with an increased risk of high-grade disease (Ptrend=0.01, Supporting Table S2).
Table 2. Prostate cancer risk by genotypes of androgen metabolizing genes in the Prostate Cancer Prevention Trial *.
| Gene | SNP | Genotype | All prostate cancer | Low-grade prostate cancer | High-grade prostate cancer | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
|
|
||||||||||
| Controls | Cases | OR (95% CI) | P-trend | Cases | OR (95% CI) | P-trend | Cases | OR (95% CI) | P-trend | |||
| Placebo arm | ||||||||||||
| SRD5A1 | rs3736316 | GG | 329 | 351 | 1 | 0.15 | 261 | 1 | 0.43 | 79 | 1 | 0.02 |
| AG | 396 | 425 | 1 (0.82-1.23) | 328 | 1.04 (0.84-1.3) | 76 | 0.8 (0.57-1.13) | |||||
| AA | 96 | 72 | 0.7 (0.5-0.99) | 60 | 0.78 (0.55-1.13) | 10 | 0.43 (0.22-0.87) | |||||
|
| ||||||||||||
| rs3822430 | AA | 318 | 339 | 1 | 0.22 | 253 | 1 | 0.54 | 76 | 1 | 0.02 | |
| AG | 390 | 422 | 1.01 (0.83-1.25) | 325 | 1.04 (0.84-1.3) | 75 | 0.81 (0.57-1.15) | |||||
| AA | 118 | 97 | 0.77 (0.57-1.05) | 80 | 0.85 (0.61-1.18) | 14 | 0.50 (0.27-0.91) | |||||
|
| ||||||||||||
| rs1560149 | GG | 520 | 534 | 1 | 0.63 | 430 | 1 | 0.19 | 84 | 1 | 0.07 | |
| CG | 270 | 304 | 1.1 (0.9-1.35) | 210 | 0.94 (0.76-1.18) | 80 | 1.82 (1.3-2.56) | |||||
| CC | 49 | 33 | 0.66 (0.41-1.04) | 28 | 0.69 (0.43-1.12) | 5 | 0.63 (0.24-1.63) | |||||
|
| ||||||||||||
| rs248797 | CC | 225 | 218 | 1 | 0.93 | 162 | 1 | 0.62 | 53 | 1 | 0.03 | |
| CT | 408 | 451 | 1.14 (0.9-1.43) | 341 | 1.15 (0.9-1.48) | 88 | 0.92 (0.63-1.35) | |||||
| TT | 207 | 197 | 0.98 (0.75-1.28) | 160 | 1.07 (0.8-1.43) | 27 | 0.56 (0.34-0.92) | |||||
|
| ||||||||||||
| rs472402 | CC | 231 | 228 | 1 | 0.81 | 184 | 1 | 0.36 | 33 | 1 | 0.03 | |
| CG | 407 | 446 | 1.11 (0.88-1.39) | 338 | 1.04 (0.82-1.32) | 87 | 1.50 (0.98-2.32) | |||||
| GG | 201 | 190 | 0.96 (0.73-1.26) | 138 | 0.86 (0.65-1.15) | 49 | 1.70 (1.05-2.75) | |||||
|
| ||||||||||||
| rs39847 | TT | 410 | 459 | 1 | 0.27 | 343 | 1 | 0.47 | 98 | 1 | 0.13 | |
| CT | 355 | 338 | 0.85 (0.7-1.04) | 266 | 0.9 (0.73-1.11) | 59 | 0.69 (0.49-0.98) | |||||
| CC | 68 | 72 | 0.95 (0.66-1.35) | 54 | 0.95 (0.65-1.4) | 14 | 0.86 (0.46-1.59) | |||||
|
| ||||||||||||
| SRD5A2 | rs2300700 | GG | 253 | 234 | 1 | 0.04 | 180 | 1 | 0.09 | 36 | 1 | 0.01 |
| AG | 418 | 437 | 1.13 (0.9-1.41) | 333 | 1.11 (0.87-1.41) | 92 | 1.57 (1.04-2.38) | |||||
| AA | 148 | 183 | 1.34 (1.01-1.77) | 137 | 1.30 (0.96-1.76) | 41 | 1.94 (1.19-3.18) | |||||
|
| ||||||||||||
| rs6760199 | CC | 450 | 454 | 1 | 0.52 | 353 | 1 | 0.31 | 85 | 1 | 0.73 | |
| AC | 316 | 361 | 1.13 (0.93-1.38) | 272 | 1.09 (0.88-1.35) | 72 | 1.21 (0.86-1.72) | |||||
| AA | 78 | 54 | 0.69 (0.47-0.99) | 39 | 0.64 (0.42-0.96) | 13 | 0.89 (0.47-1.66) | |||||
|
| ||||||||||||
| Finasteride arm | ||||||||||||
| SRD5A1 | rs1560149 | GG | 303 | 389 | 1 | 0.20 | 226 | 1 | 0.72 | 148 | 1 | 0.07 |
| CG | 201 | 193 | 0.75 (0.59-0.96) | 127 | 0.86 (0.65-1.14) | 58 | 0.59 (0.41-0.84) | |||||
| CC | 24 | 35 | 1.13 (0.66-1.95) | 22 | 1.21 (0.66-2.22) | 13 | 1.13 (0.56-2.28) | |||||
Analyses are adjusted for age and family history, and include white men only. Low grade = Gleason < 7; high grade = Gleason ≥ 7
Table 3 gives the association between SNPs in androgen metabolism genes and mean serum androgen and SHBG concentrations of all participants in the PCPT. Genetic variations of 11 SNPs in SRD5A1, SRD5A2, CYP1B1, and CYP3A4 were found to modify serum androgen and SHBG levels. Collectively, only two SNPs (SRD5A1 rs824811, all Ptrend ≤ 0.03 and CYP1B1 rs10012, all Ptrend ≤ 0.02) consistently and significantly altered all three serum androgen and SHBG concentrations. Serum androstenedione concentrations were not significantly affected by any of the SNPs in the androgen metabolism genes (data not shown).
Table 3. Association between SNPs in androgen metabolism genes and serum androgen/SHBG concentrations *.
|
|
|||||||||
|---|---|---|---|---|---|---|---|---|---|
| 3α-dG (ng/mL) | Testosterone (ng/dL) | SHBG (nmol/L) | |||||||
|
| |||||||||
| Gene | SNP | Genotype | N | Mean | P-trend | Mean | P-trend | Mean | P-trend |
| SRD5A1 | rs824811 | TT | 1494 | 6.6 | 0.03 | 383.9 | 0.01 | 39.6 | 0.002 |
| CT | 1062 | 6.8 | 368.2 | 38.0 | |||||
| CC | 141 | 7.1 | 374.3 | 37.8 | |||||
|
| |||||||||
| rs1560149 | GG | 1746 | 6.8 | 0.04 | 378.5 | 0.88 | 38.6 | 0.07 | |
| CG | 968 | 6.6 | 375.8 | 39.6 | |||||
| CC | 141 | 6.3 | 387.9 | 39.3 | |||||
|
| |||||||||
| SRD5A2 | rs6732223 | CC | 704 | 6.2 | <0.001 | 377.9 | 0.80 | 38.9 | 1.00 |
| CT | 1400 | 6.8 | 379.3 | 39.3 | |||||
| TT | 608 | 7.2 | 375.8 | 38.5 | |||||
|
| |||||||||
| rs2300700 | GG | 800 | 6.3 | <0.0001 | 378.8 | 0.50 | 39.0 | 0.75 | |
| AG | 1413 | 6.7 | 380.5 | 39.2 | |||||
| AA | 593 | 7.4 | 373.3 | 38.9 | |||||
|
| |||||||||
| rs4952197 | GG | 1563 | 7.0 | <0.0001 | 378.3 | 0.66 | 39.4 | 0.49 | |
| AG | 1080 | 6.4 | 377.4 | 38.4 | |||||
| AA | 206 | 6.0 | 386.3 | 39.5 | |||||
|
| |||||||||
| rs6760199 | CC | 1518 | 7.0 | <0.0001 | 376.5 | 0.46 | 39.2 | 0.80 | |
| AC | 1122 | 6.6 | 378.4 | 38.6 | |||||
| AA | 222 | 5.9 | 383.6 | 39.6 | |||||
|
| |||||||||
| rs9332975 | AA | 2294 | 6.6 | <0.001 | 378.5 | 0.99 | 39.0 | 0.65 | |
| AG | 543 | 7.0 | 378.7 | 39.2 | |||||
| GG | 29 | 10.1 | 377.6 | 42.2 | |||||
|
| |||||||||
| CYP1B1 | rs10012 | CC | 1153 | 6.6 | <0.01 | 370.3 | 0.02 | 37.6 | 0.002 |
| CG | 804 | 6.8 | 381.4 | 39.9 | |||||
| GG | 169 | 7.6 | 391.1 | 40.5 | |||||
|
| |||||||||
| CYP3A4 | rs2242480 | CC | 2384 | 6.7 | 0.47 | 375.8 | 0.05 | 38.8 | 0.19 |
| CT | 443 | 6.9 | 389.0 | 40.0 | |||||
| TT | 24 | 7.4 | 392.8 | 38.1 | |||||
|
| |||||||||
| rs4646437 | CC | 2318 | 6.7 | 0.51 | 375.8 | 0.03 | 38.8 | 0.09 | |
| CT | 504 | 6.9 | 390.5 | 40.1 | |||||
| TT | 29 | 6.0 | 386.2 | 39.7 | |||||
|
| |||||||||
| rs4986910 | AA | 2661 | 6.7 | 0.56 | 378.1 | 0.03 | 39.0 | 0.01 | |
| AG | 32 | 7.4 | 427.3 | 45.0 | |||||
Abbreviations: 3α-dG = 3α androstanediol glucuronide; SHBG = Sex hormone-binding globulin
Means are raw means, and p-values are calculated using linear regression adjusted for age, family history, and case/control status. Analysis includes white men only.
Table 4 examined the joint effects of serum androgen and SHBG concentrations and genetic variants in the androgen metabolism pathway and their association with prostate cancer risk. In the placebo treatment group, prostate cancer risk was increased >1.5 to 2-fold in individuals who possess the variant allele for CYP19 rs700519 and serum 3α-dG [OR, 2.13; 95% CI, 1.18-3.84, Ptrend=0.01] or SHBG concentrations [OR, 1.93; 95% CI, 1.02-3.67, Ptrend=0.04] at or above the median; or SRD5A2 rs2300700 and serum SHBG concentrations below the median (Ptrend=0.03). Prostate cancer risk was decreased by almost half in men who are carriers of the homozygous variant allele of SRD5A1 rs3822430 with serum 3α-dG concentrations at or above the median (Ptrend=0.02); or CYP3A43 rs800672 with serum testosterone (Ptrend=0.03) or SHBG (Ptrend=0.04) concentrations at or above the median. Results for the finasteride treatment arm and stratification by serum androstenedione concentrations are presented in Supporting Tables S3-S4.
Table 4. Prostate cancer risk by genotypes of androgen metabolizing genes and serum hormone levels in the placebo arm of the PCPT *.
| Gene | SNP | Genotype | Controls | Cases | OR (95% CI) | P-trend | Controls | Cases | OR (95% CI) | P-trend |
|---|---|---|---|---|---|---|---|---|---|---|
| 3α-dG < 5.7 ng/mL | 3α-dG ≥ 5.7 ng/mL | |||||||||
|
|
|
|||||||||
| SRD5A1 | rs3822430 | AA | 177 | 168 | ref | 0.61 | 139 | 171 | ref | 0.02 |
| AG | 191 | 203 | 1.12 (0.84-1.50) | 195 | 216 | 0.90 (0.67-1.21) | ||||
| GG | 56 | 56 | 1.06 (0.69-1.62) | 62 | 41 | 0.54 (0.34-0.85) | ||||
|
| ||||||||||
| SRD5A3 | rs7682870 | AA | 268 | 256 | ref | 0.62 | 256 | 266 | ref | 0.09 |
| AC | 142 | 159 | 0.85 (0.64-1.14) | 165 | 121 | 1.42 (1.06-1.89) | ||||
| CC | 20 | 16 | 1.19 (0.61-2.36) | 15 | 16 | 0.97 (0.47-2.00) | ||||
|
| ||||||||||
| CYP19 | rs700519 | CC | 407 | 408 | ref | 0.89 | 397 | 389 | ref | 0.01 |
| CT | 25 | 26 | 0.96 (0.55-1.69) | 37 | 17 | 2.13 (1.18-3.84) | ||||
| TT | 0 | 0 | NA | 0 | 0 | NA | ||||
|
| ||||||||||
| Testosterone < 362 ng/dL | Testosterone ≥ 362 ng/dL | |||||||||
|
|
|
|||||||||
| SRD5A1 | rs472402 | CC | 125 | 94 | ref | 0.23 | 105 | 134 | ref | 0.14 |
| CG | 199 | 218 | 1.45 (1.05-2.02) | 201 | 224 | 0.87 (0.63-1.20) | ||||
| GG | 105 | 99 | 1.25 (0.86-1.84) | 96 | 91 | 0.75 (0.51-1.10) | ||||
|
| ||||||||||
| rs1560149 | GG | 274 | 251 | ref | 0.53 | 242 | 280 | ref | 0.20 | |
| CG | 137 | 147 | 1.17 (0.87-1.56) | 129 | 156 | 1.04 (0.78-1.39) | ||||
| CC | 19 | 16 | 0.92 (0.46-1.84) | 30 | 17 | 0.49 (0.26-0.91) | ||||
|
| ||||||||||
| rs3822430 | AA | 167 | 178 | ref | 0.06 | 149 | 161 | ref | 0.98 | |
| AG | 193 | 193 | 0.94 (0.70-1.26) | 192 | 225 | 1.08 (0.81-1.46) | ||||
| GG | 65 | 43 | 0.62 (0.40-0.96) | 53 | 54 | 0.95 (0.61-1.47) | ||||
|
| ||||||||||
| CYP3A43 | rs800672 | CC | 122 | 111 | ref | 0.66 | 104 | 137 | ref | 0.03 |
| CT | 209 | 203 | 1.07 (0.78-1.48) | 191 | 217 | 0.87 (0.63-1.19) | ||||
| TT | 93 | 93 | 1.09 (0.74-1.60) | 103 | 89 | 0.65 (0.45-0.96) | ||||
|
| ||||||||||
| SHBG < 36.6 nmol/L | SHBG ≥ 36.6 nmol/L | |||||||||
|
|
|
|||||||||
| SRD5A2 | rs2300700 | GG | 104 | 129 | ref | 0.03 | 127 | 123 | ref | 0.41 |
| AG | 230 | 179 | 1.59 (1.15-2.19) | 206 | 235 | 0.86 (0.63-1.17) | ||||
| AA | 91 | 76 | 1.48 (1.00-2.21) | 92 | 71 | 1.25 (0.84-1.87) | ||||
|
| ||||||||||
| CYP1B1 | rs1800440 | AA | 258 | 291 | ref | 0.12 | 291 | 292 | ref | 0.93 |
| AG | 119 | 136 | 1.01 (0.75-1.36) | 134 | 121 | 0.90 (0.67-1.21) | ||||
| GG | 19 | 5 | 0.23 (0.09-0.63) | 10 | 16 | 1.60 (0.71-3.58) | ||||
|
| ||||||||||
| CYP3A43 | rs800672 | CC | 113 | 116 | ref | 0.60 | 113 | 132 | ref | 0.04 |
| CT | 195 | 207 | 1.04 (0.75-1.44) | 205 | 213 | 0.89 (0.65-1.23) | ||||
| TT | 84 | 96 | 1.11 (0.75-1.64) | 112 | 86 | 0.66 (0.45-0.96) | ||||
|
| ||||||||||
| CYP19 | rs700519 | CC | 372 | 395 | ref | 0.61 | 424 | 408 | ref | 0.04 |
| CT | 28 | 34 | 1.14 (0.68-1.93) | 15 | 28 | 1.93 (1.02-3.67) | ||||
| TT | 0 | 0 | NA | 0 | 0 | NA | ||||
Abbreviations: 3α-dG = 3α androstanediol glucuronide; SHBG = Sex hormone-binding globulin
Analyses are adjusted for age and family history of prostate cancer, and include white men only.
Discussion
Our study assessed the association of genetic variants in androgen-related pathways with prostate cancer risk. In the placebo group, when stratified by disease grade, there were significant associations of SNPs in SRD5A1 (rs3736316, rs3822430, rs1560149, rs248797, rs472402) and SRD5A2 (rs2300700) with risk of high-grade prostate cancer. Specifically, the homozygous variants of SRD5A1 rs472402 and SRD5A2 rs2300700 were statistically significantly associated with a >1.5 to 2-fold increased risk of high-grade disease; while the homozygous variant allele of three SNPs (SRD5A1 rs3736316, rs3822430, rs248797) reduced the risk of high-grade disease by about half.
The PCPT found that finasteride reduced the risk of prostate cancer by 25%, even though high-grade prostate cancer was more common in the finasteride group.14 Findings from the current study provide further insight into the genetic etiology of the increased risk of high-grade cancer among men treated with finasteride. Indeed, in the finasteride treatment group, the heterozygous variant of SRD5A1 rs1560149 was associated with decreased risk with high-grade compared to wild type. This suggests that finasteride treatment may decrease the odds of high-grade disease in men who are heterozygote carriers of SRD5A1 rs1560149 from 1.83 (OR, placebo arm) to 0.58 (OR, finasteride arm). All other SNPs identified in the placebo arm were not significant in the finasteride treatment group. Polymorphisms in SRD5A2 have been studied as potential genetic markers of prostate cancer risk and results have not always been consistent in the literature (reviewed in 17). The common nonsynonymous SNP SRD5A2 rs9282858 had no significant associations with prostate cancer risk in our study consistent with recent meta-analysis studies.18-20 Studies on the positive association of SRD5A1 polymorphisms and prostate cancer risk are fewer11, 21, 22 with our study being the first to identify the association of several novel SRD5A1 SNPs that have not been previously reported.
We determined factors involved in modifying serum androgen and SHBG concentrations. Here, we found several non-genetic factors (obesity, smoking, diabetes, physical activity, protein consumption) influencing androgen and SHBG concentrations that have been previously reported to be associated with prostate cancer risk (Supporting Table S5).23 More importantly, we identified genetic factors that impact androgen metabolism as reflected by significant changes in circulating androgen and SHBG concentrations. Genetic variations of 11 SNPs in SRD5A1, SRD5A2, CYP1B1, and CYP3A4 were found to modify serum androgen and SHBG levels. Collectively, this is the first study to show that only two SNPs (SRD5A1 rs824811 and CYP1B1 rs10012) consistently and statistically significantly altered all androgen and SHBG concentrations. Five SRD5A2 SNPs (rs6732223, rs2300700, rs9332975, rs4952197, rs6760199) modified serum 3α-dG concentrations while no associations were found for testosterone or SHBG levels in our study. Previous studies have shown SRD5A2 polymorphisms (rs12470143, rs2208532, rs523349, rs676033, rs2300700) altering levels of circulating steroid glucuronide metabolites of DHT in prostate cancer patients.11, 24, 25 Our study is the first to report rs6732223 in prostate cancer, a SNP that is located 7 kb downstream of SRD5A2, and increased enzyme activity has been suggested for the minor allele in cortisol metabolism.26 Our results are consistent with previous findings on the impact of select SRD5A2 SNPs affecting circulating androgen metabolites rather than upstream substrates.11 To the best of our knowledge, this is the first report of the association of rs824811, located in an intron of SRD5A1, with modifying androgen concentrations for prostate cancer and its functional activity remains to be determined. The minor alleles of four additional SNPs (CYP1B1 rs10012, CYP3A4 rs2242480, rs4646437 and rs4986910) were significantly associated with higher serum T or SHBG concentrations. CYP1B1 is involved in the activation of many procarcinogens and hydroxylation of testosterone; thus, variations in CYP1B1 may lead to higher susceptibility to prostate cancer.27 Nonsynonymous SNPs rs10012 and rs1056827 have been evaluated for association with prostate cancer risk 28-30 and functional studies indicate that allelic combinations can alter the catalytic activity of CYP1B1 in estrogen hydroxylation31, 32. Two common nonsynonymous CYP1B1 SNPs rs1056836 and rs1800440 were not associated with either serum androgen concentrations or prostate cancer risk (data not shown). CYP3A4 SNP rs4646437 (lies within intron 7) undergoes positive selection in Caucasians and male carriers of the variant allele have been associated with reduced CYP3A4 protein expression and enzyme activity.33 The minor allele of CYP3A4 rs4986910 (CYP3A4*3, located in exon 12) has low allele frequency and was shown to be associated with reduced enzyme function.34 A study by Siemes et al. did not find any association of SNPs in CYP3A4, CYP3A5, or CYP3A43 with androgen concentrations.35 Further studies to verify these results are warranted.
Several association studies have investigated polymorphisms in genes involved in androgen-related pathways as candidates for prostate cancer risk.36-39 A few evaluated the association of these SNPs with serum hormone levels in prostate cancer.24, 40-42 While inherited variants that modify circulating hormone concentrations may differ from those affecting disease risk, our study found SNPs that are statistically significantly associated with both serum androgen or SHGB levels and prostate cancer risk. In the placebo arm, higher 3α-dG concentrations were associated with increased disease risk for the heterozygous variant alleles of SRD5A3 rs7682870 and CYP19 rs700519. This is the first study to implicate SRD5A3 in prostate cancer risk and its functional activity is unknown. Our results showed that genetic variability in factors controlling androgen concentrations likely confer differential risk of developing prostate cancer and that subgroups of men may be more susceptible to the disease.
There are strengths and limitations to our study. The PCPT was a large placebo-controlled randomized trial that used prostate biopsies to verify absence or presence of cancer; thus, the control group all had confirmed negative prostate biopsies, largely eliminating the possibility that controls may have had undiagnosed or undetected disease. Additional strengths included the use of a central pathology laboratory for uniform adjudication of all cases (including adjudication of Gleason grade), highly sensitive and specific assays for quantitating serum androgen and SHBG concentrations, and the large sample size of our patient population. Our study was limited in that serum androgen/SHBG concentrations were measured at one time point and may also not be representative of tissue levels. This analysis included only white participants; however, this minimized concerns of selection bias or population stratification. Given the hypothesis-driven nature of SNP selection, we did not systematically correct for multiple testing in our analyses, as might be done for agnostic genetic studies.
In conclusion, this study demonstrates that genetic polymorphism of genes in androgen biosynthesis and metabolism pathways may influence prostate cancer risk. We identified variants involved in modifying serum androgen and SHBG concentrations that may explain the inter-individual variation observed in concentration differences. Our study demonstrates the significance of androgen metabolism in prostate cancer and suggests that in addition to the agnostic approach of genome wide association studies, candidate gene and pathway approach with biological rationale are important strategies to clarify the role of genetic susceptibility and androgen in prostate cancer. Future replication studies are warranted to confirm the novel SNPs identified and to investigate the functional consequences of causal variants.
Supplementary Material
Acknowledgments
The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organization imply endorsement by the U.S. Government.
Funding/Support: This work was supported in part by the Intramural Research Program of the National Institutes of Health, National Cancer Institute (NCI), Center for Cancer Research; The Biology of the Prostate Cancer Prevention Trial (P01 CA108964) from the NCI; the Cancer Therapy and Research Center Support Grant (P30 CA054174) at the University of Texas Health Science Center at San Antonio; and Public Health Service Grants (CA37429 and CA182883) from the NCI.
Footnotes
Present address: Yachay Tech University, San Miguel de Urcuqui, Ecuador
Conflict of Interest Disclosures: None reported.
Author Contributions: Douglas K. Price: Conceptualization, validation, formal analysis, investigation, writing – original draft, writing – review and editing, and visualization. Cindy H. Chau: Conceptualization, validation, formal analysis, investigation, writing – original draft, writing – review and editing, and visualization. Cathee Till: Formal analysis and writing – review and editing. Phyllis J. Goodman: Resources, data curation, and writing – review and editing. Robin J. Leach: Conceptualization, investigation, data curation, and writing – review and editing. Teresa L. Johnson-Pais: Methodology, validation, investigation, and writing – review and editing. Ann W. Hsing: Methodology and writing – review and editing. Ashraful Hoque: Conceptualization, investigation, data curation, writing – review and editing, supervision, project administration, and funding acquisition. Howard L. Parnes: Writing – review and editing and project administration. Jeannette M. Schenk: Writing – review and editing. Catherine M. Tangen: Conceptualization, methodology, formal analysis, writing – review and editing, and supervision. Ian M. Thompson: Conceptualization, investigation, resources, writing – review and editing, supervision, and project administration. Juergen K.V. Reichardt: Conceptualization, writing – review and editing, and funding acquisition. William D. Figg: Conceptualization, investigation, writing – original draft, writing – review and editing, visualization, supervision, and funding acquisition.
References
- 1.Andersson S, Russell DW. Structural and biochemical properties of cloned and expressed human and rat steroid 5 alpha-reductases. Proc Natl Acad Sci U S A. 1990 May;87(10):3640–3644. doi: 10.1073/pnas.87.10.3640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Tamura K, Furihata M, Tsunoda T, et al. Molecular features of hormone-refractory prostate cancer cells by genome-wide gene expression profiles. Cancer Res. 2007 Jun 1;67(11):5117–5125. doi: 10.1158/0008-5472.CAN-06-4040. [DOI] [PubMed] [Google Scholar]
- 3.Uemura M, Tamura K, Chung S, et al. Novel 5 alpha-steroid reductase (SRD5A3, type-3) is overexpressed in hormone-refractory prostate cancer. Cancer Sci. 2008 Jan;99(1):81–86. doi: 10.1111/j.1349-7006.2007.00656.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hoque A, Yao S, Till C, et al. Effect of finasteride on serum androstenedione and risk of prostate cancer within the prostate cancer prevention trial: differential effect on high- and low-grade disease. Urology. 2015 Mar;85(3):616–620. doi: 10.1016/j.urology.2014.11.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Schenk JM, Till C, Hsing AW, et al. Serum androgens and prostate cancer risk: results from the placebo arm of the Prostate Cancer Prevention Trial. Cancer Causes Control. 2015 Nov 20; doi: 10.1007/s10552-015-0695-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kristal AR, Till C, Tangen CM, et al. Associations of serum sex steroid hormone and 5alpha-androstane-3alpha, 17beta-diol glucuronide concentrations with prostate cancer risk among men treated with finasteride. Cancer Epidemiol Biomarkers Prev. 2012 Oct;21(10):1823–1832. doi: 10.1158/1055-9965.EPI-12-0695. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Roddam AW, Allen NE, Appleby P, Key TJ. Endogenous sex hormones and prostate cancer: a collaborative analysis of 18 prospective studies. J Natl Cancer Inst. 2008 Feb 6;100(3):170–183. doi: 10.1093/jnci/djm323. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Weiss JM, Huang WY, Rinaldi S, et al. Endogenous sex hormones and the risk of prostate cancer: a prospective study. Int J Cancer. 2008 May 15;122(10):2345–2350. doi: 10.1002/ijc.23326. [DOI] [PubMed] [Google Scholar]
- 9.Gill JK, Wilkens LR, Pollak MN, Stanczyk FZ, Kolonel LN. Androgens, growth factors, and risk of prostate cancer: the Multiethnic Cohort. Prostate. 2010 Jun 1;70(8):906–915. doi: 10.1002/pros.21125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ross RK, Bernstein L, Lobo RA, et al. 5-alpha-reductase activity and risk of prostate cancer among Japanese and US white and black males. Lancet. 1992 Apr 11;339(8798):887–889. doi: 10.1016/0140-6736(92)90927-u. [DOI] [PubMed] [Google Scholar]
- 11.Levesque E, Laverdiere I, Lacombe L, et al. Importance of 5alpha-reductase gene polymorphisms on circulating and intraprostatic androgens in prostate cancer. Clin Cancer Res. 2014 Feb 1;20(3):576–584. doi: 10.1158/1078-0432.CCR-13-1100. [DOI] [PubMed] [Google Scholar]
- 12.Singh AS, Chau CH, Price DK, Figg WD. Mechanisms of disease: Polymorphisms of androgen regulatory genes in the development of prostate cancer. Nat Clin Pract Urol. 2005 Feb;2(2):101–107. doi: 10.1038/ncpuro0091. [DOI] [PubMed] [Google Scholar]
- 13.Schleutker J. Polymorphisms in androgen signaling pathway predisposing to prostate cancer. Mol Cell Endocrinol. 2012 Sep 5;360(1-2):25–37. doi: 10.1016/j.mce.2011.07.007. [DOI] [PubMed] [Google Scholar]
- 14.Thompson IM, Goodman PJ, Tangen CM, et al. The influence of finasteride on the development of prostate cancer. N Engl J Med. 2003 Jul 17;349(3):215–224. doi: 10.1056/NEJMoa030660. [DOI] [PubMed] [Google Scholar]
- 15.Feigl P, Blumenstein B, Thompson I, et al. Design of the Prostate Cancer Prevention Trial (PCPT) Control Clin Trials. 1995 Jun;16(3):150–163. doi: 10.1016/0197-2456(94)00xxx-m. [DOI] [PubMed] [Google Scholar]
- 16.Winchester DA, Till C, Goodman PJ, et al. Variation in genes involved in the immune response and prostate cancer risk in the placebo arm of the Prostate Cancer Prevention Trial. Prostate. 2015 Sep;75(13):1403–1418. doi: 10.1002/pros.23021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Sissung TM, Price DK, Del Re M, et al. Genetic variation: effect on prostate cancer. Biochim Biophys Acta. 2014 Dec;1846(2):446–456. doi: 10.1016/j.bbcan.2014.08.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wang C, Tao W, Chen Q, Hu H, Wen XY, Han R. SRD5A2 V89L polymorphism and prostate cancer risk: a meta-analysis. Prostate. 2010 Feb 1;70(2):170–178. doi: 10.1002/pros.21050. [DOI] [PubMed] [Google Scholar]
- 19.Li Q, Zhu Y, He J, et al. Steroid 5-alpha-reductase type 2 (SRD5A2) V89L and A49T polymorphisms and sporadic prostate cancer risk: a meta-analysis. Mol Biol Rep. 2013 May;40(5):3597–3608. doi: 10.1007/s11033-012-2434-x. [DOI] [PubMed] [Google Scholar]
- 20.Li X, Huang Y, Fu X, et al. Meta-analysis of three polymorphisms in the steroid-5-alpha-reductase, alpha polypeptide 2 gene (SRD5A2) and risk of prostate cancer. Mutagenesis. 2011 May;26(3):371–383. doi: 10.1093/mutage/geq103. [DOI] [PubMed] [Google Scholar]
- 21.Audet-Walsh E, Bellemare J, Nadeau G, et al. SRD5A polymorphisms and biochemical failure after radical prostatectomy. Eur Urol. 2011 Dec;60(6):1226–1234. doi: 10.1016/j.eururo.2011.06.020. [DOI] [PubMed] [Google Scholar]
- 22.Setlur SR, Chen CX, Hossain RR, et al. Genetic variation of genes involved in dihydrotestosterone metabolism and the risk of prostate cancer. Cancer Epidemiol Biomarkers Prev. 2010 Jan;19(1):229–239. doi: 10.1158/1055-9965.EPI-09-1018. [DOI] [PubMed] [Google Scholar]
- 23.Hsing AW, Chokkalingam AP. Prostate cancer epidemiology. Front Biosci. 2006;11:1388–1413. doi: 10.2741/1891. [DOI] [PubMed] [Google Scholar]
- 24.Ahn J, Schumacher FR, Berndt SI, et al. Quantitative trait loci predicting circulating sex steroid hormones in men from the NCI-Breast and Prostate Cancer Cohort Consortium (BPC3) Hum Mol Genet. 2009 Oct 1;18(19):3749–3757. doi: 10.1093/hmg/ddp302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hayes VM, Severi G, Padilla EJ, et al. 5alpha-Reductase type 2 gene variant associations with prostate cancer risk, circulating hormone levels and androgenetic alopecia. Int J Cancer. 2007 Feb 15;120(4):776–780. doi: 10.1002/ijc.22408. [DOI] [PubMed] [Google Scholar]
- 26.Steen NE, Tesli M, Kahler AK, et al. SRD5A2 is associated with increased cortisol metabolism in schizophrenia spectrum disorders. Prog Neuropsychopharmacol Biol Psychiatry. 2010 Dec 1;34(8):1500–1506. doi: 10.1016/j.pnpbp.2010.08.013. [DOI] [PubMed] [Google Scholar]
- 27.Zhang H, Li L, Xu Y. CYP1B1 polymorphisms and susceptibility to prostate cancer: a meta-analysis. PLoS One. 2013;8(7):e68634. doi: 10.1371/journal.pone.0068634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Beuten J, Gelfond JA, Byrne JJ, et al. CYP1B1 variants are associated with prostate cancer in non-Hispanic and Hispanic Caucasians. Carcinogenesis. 2008 Sep;29(9):1751–1757. doi: 10.1093/carcin/bgm300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Liu JY, Yang Y, Liu ZZ, Xie JJ, Du YP, Wang W. Association between the CYP1B1 polymorphisms and risk of cancer: a meta-analysis. Mol Genet Genomics. 2015 Apr;290(2):739–765. doi: 10.1007/s00438-014-0946-x. [DOI] [PubMed] [Google Scholar]
- 30.Li C, Long B, Qin X, Li W, Zhou Y. Cytochrome P1B1 (CYP1B1) polymorphisms and cancer risk: a meta-analysis of 52 studies. Toxicology. 2015 Jan 2;327:77–86. doi: 10.1016/j.tox.2014.11.007. [DOI] [PubMed] [Google Scholar]
- 31.Hanna IH, Dawling S, Roodi N, Guengerich FP, Parl FF. Cytochrome P450 1B1 (CYP1B1) pharmacogenetics: association of polymorphisms with functional differences in estrogen hydroxylation activity. Cancer Res. 2000 Jul 1;60(13):3440–3444. [PubMed] [Google Scholar]
- 32.Shimada T, Watanabe J, Kawajiri K, et al. Catalytic properties of polymorphic human cytochrome P450 1B1 variants. Carcinogenesis. 1999 Aug;20(8):1607–1613. doi: 10.1093/carcin/20.8.1607. [DOI] [PubMed] [Google Scholar]
- 33.Schirmer M, Rosenberger A, Klein K, et al. Sex-dependent genetic markers of CYP3A4 expression and activity in human liver microsomes. Pharmacogenomics. 2007 May;8(5):443–453. doi: 10.2217/14622416.8.5.443. [DOI] [PubMed] [Google Scholar]
- 34.van Schaik RH, de Wildt SN, Brosens R, van Fessem M, van den Anker JN, Lindemans J. The CYP3A4*3 allele: is it really rare? Clin Chem. 2001 Jun;47(6):1104–1106. [PubMed] [Google Scholar]
- 35.Siemes C, Visser LE, de Jong FH, et al. Cytochrome P450 3A gene variation, steroid hormone serum levels and prostate cancer--The Rotterdam Study. Steroids. 2010 Dec;75(12):1024–1032. doi: 10.1016/j.steroids.2010.07.001. [DOI] [PubMed] [Google Scholar]
- 36.Beuten J, Gelfond JA, Franke JL, et al. Single and multigenic analysis of the association between variants in 12 steroid hormone metabolism genes and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev. 2009 Jun;18(6):1869–1880. doi: 10.1158/1055-9965.EPI-09-0076. [DOI] [PubMed] [Google Scholar]
- 37.Mononen N, Schleutker J. Polymorphisms in genes involved in androgen pathways as risk factors for prostate cancer. J Urol. 2009 Apr;181(4):1541–1549. doi: 10.1016/j.juro.2008.11.076. [DOI] [PubMed] [Google Scholar]
- 38.Cunningham JM, Hebbring SJ, McDonnell SK, 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. 2007 May;16(5):969–978. doi: 10.1158/1055-9965.EPI-06-0767. [DOI] [PubMed] [Google Scholar]
- 39.Lindstrom S, Zheng SL, Wiklund F, et al. Systematic replication study of reported genetic associations in prostate cancer: Strong support for genetic variation in the androgen pathway. Prostate. 2006 Dec 1;66(16):1729–1743. doi: 10.1002/pros.20489. [DOI] [PubMed] [Google Scholar]
- 40.Travis RC, Schumacher F, Hirschhorn JN, et al. CYP19A1 genetic variation in relation to prostate cancer risk and circulating sex hormone concentrations in men from the Breast and Prostate Cancer Cohort Consortium. Cancer Epidemiol Biomarkers Prev. 2009 Oct;18(10):2734–2744. doi: 10.1158/1055-9965.EPI-09-0496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Boger-Megiddo I, Weiss NS, Barnett MJ, Goodman GE, Chen C. V89L polymorphism of the 5alpha-reductase Type II gene (SRD5A2), endogenous sex hormones, and prostate cancer risk. Cancer Epidemiol Biomarkers Prev. 2008 Feb;17(2):286–291. doi: 10.1158/1055-9965.EPI-07-0238. [DOI] [PubMed] [Google Scholar]
- 42.Sun T, Oh WK, Jacobus S, et al. The impact of common genetic variations in genes of the sex hormone metabolic pathways on steroid hormone levels and prostate cancer aggressiveness. Cancer Prev Res (Phila) 2011 Dec;4(12):2044–2050. doi: 10.1158/1940-6207.CAPR-11-0283. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.