Abstract
Vitamin D may influence prostate cancer risk, but evidence is inconsistent. We conducted a nested case-control study in the Prostate Cancer Prevention Trial (PCPT). Cases (n=1,128) and controls (n=1,205) were frequency matched on age, first-degree relative with prostate cancer and PCPT treatment arm (finasteride/placebo); African-Americans were oversampled and case/control status was biopsy-confirmed. We selected 21 SNPs in vitamin D-related genes (VDR, GC, C10orf88, CYP2R1, CYP24A1, CYP27B1, DHCR7, NADSYN1) to test genotype and genotype-treatment interactions in relation to prostate cancer. We also tested mean serum 25(OH)D differences by minor allele distributions and tested for serum 25(OH)D-genotype interactions in relation to prostate cancer risk. Log-additive genetic models (Bonferroni-corrected within genes) adjusted for age, BMI, PSA, and family history of prostate cancer revealed a significant interaction between treatment arm and GC/rs222016 (finasteride OR=1.37, placebo OR=0.85, p-interaction<0.05), GC/rs222014 (finasteride OR=1.36, placebo OR=0,85, p-interaction<0.05) and CYP27B1/rs703842 (finasteride OR=0.76, placebo OR=1.10, p-interaction<0.05) among Caucasians, and C10orf88/rs6599638 (finasteride OR=4.68, placebo OR=1.39, p-interaction<0.05) among African Americans. VDR/rs1544410 and CYP27B1/rs703842 had significant treatment interactions for high-grade disease among Caucasians (finasteride OR=0.81, placebo OR=1.40, p-interaction<0.05 and finasteride OR=0.70, placebo OR=1.28, p-interaction<0.05, respectively). Vitamin D-related SNPs influenced serum 25(OH)D, but gene-serum 25(OH)D effect modification for prostate cancer was marginally observed only for CYP24A1/rs2248359. In conclusion, evidence that vitamin D-related genes or gene-serum 25(OH)D associations influence prostate cancer risk is modest. We found some evidence for gene-finasteride interaction effects for prostate cancer in Caucasians and African-Americans. Results suggest only minimal associations of vitamin D with total or high-grade prostate cancer.
Keywords: vitamin D, prostate cancer, 25(OH)D, genetics
INTRODUCTION
Prostate cancer is a very common cancer in aging men (1), but the etiology remains elusive. Vitamin D is thought to play a role in prostate cancer development and progression since experimental evidence shows that vitamin D influences key carcinogenesis-related processes including promoting cellular differentiation and apoptosis and reducing cell proliferation (2–4). Further, prostate cancer cell-line studies show that the biologically active form of the vitamin, 1,25(OH)D, reduces metastatic potential by increasing E-cadherin expression that interferes with cell adhesion of the circulating cells (5). These and other findings in cell lines and preclinical animal models led to the hypothesis that vitamin D has anti-cancer properties and could be considered as a target for prostate cancer prevention.
Despite the in vitro and animal model data, observational studies in humans have been inconsistent regarding the relationship of circulating vitamin D with prostate cancer risk. Vitamin D status in these studies is usually determined by serum concentrations of 25(OH)D, the primary circulating form of vitamin D and the precursor to 1,25(OH)D. A study published in 2011 from a nested case–control analysis of 1,000 cases and 1,000 controls found that Finnish men in the Alpha-Tocopherol Beta-Carotene Cancer (ATBC) Prevention Study cohort with higher circulating serum 25(OH)D concentrations had a statistically significantly elevated risk for total prostate cancer (highest quintile vs. lowest OR= 1.56, 95%CI 1.15–2.12) (6). Three meta-analyses found was no association between serum 25(OH)D and prostate cancer risk (7–9). However, a later meta-analysis reported a significant 17% increase in prostate cancer risk for men with higher vs. lower serum 25(OH)D (10) (11). In addition, in our previous publication from the Prostate Cancer Prevention Trial, we showed no overall association of serum 25(OH)D with prostate cancer risk (10), results that could have been obscured by not considering genetic variation in vitamin D metabolism. PCPT vitamin D data were recently pooled with data from 19 cohorts (12). In those analyses, higher vs lower serum 25(OH)D were associated with a prostate cancer odds ratio = 1.22 (95% CI 1.13–1.31, P-trend<0.001) with noticeable risk increases for non-aggressive disease (OR=1.24, 95% CI 1.13–1.36) but not for aggressive disease (OR=0.94, 95% CI 0.78–1.15). This recent pooled analysis did not consider genetic influences on serum 25(OH)D.
Season, body weight, adiposity, vitamin D intake from both food and supplements, and UVB exposure all influence serum 25(OH)D concentrations (13). Estimates suggest that only about 25% of the variability in 25(OH)D is due to these identifiable factors and that serum 25(OH)D may have a heritability component (heritability estimates range from 28.8% to 38% in those with European ancestry) (14, 15). Given the important role that genetic characteristics may play in the determination of serum 25(OH)D, it is important to study the multiple genes that control its actions and metabolism. Vitamin D3 is transported in the blood to the liver, where it is converted into 25(OH)D by the enzyme 25-hydroxylase that is coded by the gene CYP2R1. It then travels to the kidneys where it is converted to 1, 25(OH)D by 1-α-hydroxylase, an enzyme coded by CYP27B1. Other tissues, including the prostate, also express 1-α-hydroxylase (16). CYP24A1 codes for 24-hydroxylase, an enzyme that deactivates calcitriol. These genes and others not directly on the metabolic pathway have been found to be related to circulating 25(OH)D from candidate gene analysis (14) and genome-wide association studies (GWAS) (15, 17, 18). For example, GC encodes for vitamin D–binding protein (DBP), the major carrier of vitamin D in circulation. DHCR7 encodes the enzyme that catalyzes the conversion of 7-dehydrocholesterol, a vitamin D3 precursor, to cholesterol. DHCR7 is often combined with NADSYN1 and was identified as a novel locus from a 2010 GWAS study (18). Finally, the region that includes the open-reading frame 88 (C10orf88) on chromosome 10q26.13 was also found to be associated with vitamin D concentrations in blood (17).
Selected polymorphisms from vitamin D-related genes have been studied in relation to their associations with prostate cancer. The most studied vitamin D gene is the vitamin D receptor (VDR), which is the critical mediator of vitamin D actions. While one review concluded that common polymorphisms in vitamin D pathway genes were not associated with prostate cancer risk (19), other studies found that interactions may exist between SNPs and 25(OH)D that affect prostate cancer risk (20).
Here we report on a nested case-control study in the Prostate Cancer Prevention Trial (PCPT). The overall goal of the PCPT was to test whether finasteride, a 5-ɑ-reductase inhibitor that blocks the conversion of testosterone to its more active form of dihydrotestosterone, would reduce the period prevalence of prostate cancer (21). Using PCPT biospecimens, we investigated both circulating concentrations of 25(OH)D and vitamin D-related genes in relation to prostate cancer risk. In addition, we investigated whether finasteride influenced the relationship of polymorphisms in the vitamin D genes with prostate cancer risk since 25(OH)D interacts with androgen signaling, which is a key pathway for prostate cancer development.
MATERIALS AND METHODS
Data for this nested case-control study are from the Prostate Cancer Prevention Trial (PCPT), a randomized double-blind, placebo-controlled trial of finasteride for the primary prevention of prostate cancer (21). Briefly, 18,880 men aged 55 years or older with normal digital rectal examination (DRE) results, PSA ≤3 ng/mL and no history of prostate cancer, severe lower urinary tract symptoms, or clinically significant coexisting conditions were randomized to receive finasteride (5 mg/day) or placebo (21). During the PCPT, participants underwent DRE and PSA assessment annually, and men with a DRE suspicious for cancer or a PSA (adjusted for the effect of finasteride) above 4.0 ng/mL were referred for a prostate biopsy (21). At the end of seven years, all men not previously diagnosed with prostate cancer were asked to undergo an end-of-study biopsy per the primary trial protocol to determine the presence or absence of prostate cancer (21). Six core samples were collected under transrectal ultrasonographic guidance, and biopsies were reviewed for adenocarcinoma by both the local study site pathologist and a central pathology laboratory (21). In the case of discordant results, a referee pathologist reviewed cases until concordance was reached (21). Clinical stage was assigned locally, and tumors were graded centrally using the Gleason scoring system. To be consistent with the primary trial report (21), tumors with Gleason scores < 7 were classified as low-grade and those with Gleason scores ≥ 7 were classified as high-grade. The Institutional Review Boards of all participating SWOG institutions approved this study and all participants signed written informed consent.
Case and control selection
Cases (n=1,128) were men with biopsy-confirmed prostate cancer, identified either by a ‘for-cause’ biopsy triggered by a PSA > 4 ng/mL or abnormal DRE during the trial (n=423) or at the end-of-study biopsy (n=705) and who had available DNA for genotyping analysis. Controls (n=1,205) were selected from men who had no evidence of prostate cancer on the end-of-study biopsy and had available DNA for genotyping. Controls were frequency matched to cases on distributions of age, first-degree family history of prostate cancer and treatment arm, and included all available nonwhites to increase the pool of minorities. The present analysis includes only Caucasian and African-American men. All other race and ethnic groups were excluded from this analysis due to very low numbers.
Blood collection, processing, genotyping and serum 25(OH)D measures
Non-fasting whole blood samples were collected and shipped overnight in a chilled container, processed at Esoterix (Calabasas CA) and shipped on dry ice for storage at the University of Colorado (21). White blood cells were aliquoted and stored at –70°C until extraction of DNA from the WBC samples using Qiagen M48 robot (Valencia, CA). SNPs were selected based on prior literature that was available at the time of our laboratory analysis (20, 22–24). Genotypes were determined using the Illumina VeraCode GoldenGate genotyping assay (Illumina Inc.; San Diego, CA). The list of SNPs to be genotyped was submitted to Illumina and scored with the Assay Design Tool (ADT). Those SNPs with acceptable scores were developed into an oligonucleotide pool assay (OPA) designed for a VeraCode GoldenGate panel. Two hundred fifty nanograms of DNA were used as the template for the assay. The assay was performed in 96 well plates following the established protocol (Illumina). The plates were scanned using an Illumina BeadXpress reader and the genotypes were analyzed using GenomeStudio software (Illumina). Interplate and intraplate replicates were included as quality control measures, and duplicate concordance was ≥98% for all SNPs (mean=99%). Baseline serum was assayed via a chemiluminescent assay for 25(OH)D using LIAISON 25OHVitaminDTOTAL Assay (DiaSorin Inc.) at the Fred Hutchinson Cancer Research Center Biomarker Laboratory (10). The lower limit of quantitation for this assay is 4 ng/mL and no specimens had results below this level (10). All batches were balanced by case and control status and laboratory personnel at the Biomarker Laboratory were blinded to participant status. The coefficient of variation (CV) for 86 blinded duplicate quality control samples was 8.3% (10).
Other data collection
Data on age, race/ethnicity and family history of prostate cancer in first-degree relatives were collected at baseline using standardized self-administered questionnaires (21). Participants’ height and weight were measured at baseline and body mass index (BMI) was calculated as weight (kg)/height (m2). PSA was measured at baseline as part of the trial protocol (21).
Statistical analysis
Participant characteristics were summarized and compared between cases and controls using t-tests for continuous variables and chi-square tests for categorical variables, separately by Caucasian and African-American self-reported race/ethnicity. Chi-square tests were used to compare minor allele frequencies between cases and controls, stratified by race (Supplementary Table 1). Hardy-Weinberg equilibrium (HWE) was tested for all SNPs by race. Those SNPs not meeting HWE were not included in the analyses. For each SNP, the most frequent genotype in Caucasian men was used as the referent genotype as it provided the greatest model stability. To account for the frequency-matched case-control study design, unconditional logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for risk of total prostate cancer, and polytomous logistic regression was used to estimate ORs and 95% CIs for both low-grade and high-grade cancer. A log-additive genetic model was used, so that the OR represents the risk of cancer or high-grade cancer per each additional minor allele. Model covariates were selected based on a priori information about potential confounding and included age (continuous), BMI (continuous), baseline PSA (continuous), and first-degree family history of prostate cancer (yes/no). To evaluate whether the vitamin D genotype-prostate cancer relationship differentially influenced finasteride’s effect on risk, tests for interaction were conducted by entering cross-product terms of SNPs and finasteride and testing these terms with the Wald test. To determine whether a participant’s baseline serum 25(OH)D concentration influenced the genotype-cancer relationship, we first created season-adjusted and log-transformed values for serum 25(OH)D as previously described for Caucasian participants (10). We then calculated geometric means of serum 25(OH)D by genotype, and calculated p-values comparing 25(OH)D values by SNP genotype using unadjusted linear regression with season-adjusted log[serum 25OH)D] as the dependent variable. We next created tertiles of baseline season-adjusted serum 25(OH)D and tested whether the genotype-cancer relationships varied by tertile of serum 25(OH)D. Trend interaction p-values between tertiles of baseline serum 25(OH)D and genotype with prostate cancer were calculated using an ordinal variable corresponding to tertile rank, from lowest to highest. These analyses involving serum 25(OH)D were restricted to Caucasian men (due to small numbers of African-Americans) on the placebo arm only for both sample size consideration and to reduce the potential for multiple comparisons.
All statistical tests are 2-sided. 95% confidence intervals are included as a uniform descriptive measure, and statistical significance for main effects was determined using a Bonferroni adjustment within each gene. Thresholds for statistical significance each gene were: VDR =0.013; GC =0.008; C10corf88, CYP2R1, CYP27B1, and DHCR7 =0.05; CYP24A1 =0.01; and NADSYN1 =0.025. Tests for interaction were considered significant with p<0.05. SAS (version 9.4, Cary NC) and R were used for all statistical analyses.
RESULTS
The 2,333 men (1128 cases and 1205 controls) who met study criteria were predominantly Caucasian (92.4%); the remainder were African American (Table 1). Due to the study’s frequency matching criteria, cases and controls in both racial groups did not differ by age or family history of prostate cancer. Caucasian men (both cases and controls) were slightly older than African-American men (both cases and controls). The mean BMI for both race groups and for cases and controls were all in the overweight range (BMI=25.0–29.9 kg/m2). The proportion of cases with Gleason ≥ 7 was slightly higher in African-Americans than Caucasians.
Table 1.
Baseline characteristics of prostate cancer cases and controls1 in the Prostate Cancer Prevention Trial by race
| Characteristics | Caucasian (n=2156) | African American (n=177) | ||
|---|---|---|---|---|
| Case | Control | Case | Control | |
| Number | 1081 | 1075 | 47 | 130 |
| Age, mean years (range) | 63.5 (55–82) | 64.0 (55–84) | 61.9 (55–77) | 61.7 (55–77) |
| Finasteride Treatment Arm, n (%) | 446 (41.3) | 419 (39.0) | 20 (42.6) | 81 (62.3) |
| Family History of prostate cancer, n (%) | 235 (21.7) | 244 (22.7) | 7 (14.9) | 24 (18.5) |
| BMI (kg/m2), mean (range) | 27.3 (19–52) | 27.6 (19–46) | 29.4 (21–47) | 28.5 (17–42) |
| Gleason ≥7, n (% of all cases) | 269 (24.9) | N/A | 17 (36.2) | N/A |
Nested case control selection frequency matched on age, first degree family history of prostate cancer and treatment arm with oversampling for eligible African-Americans.
Twenty-one vitamin D-related SNPs were included in this study (Supplementary Table 1). For many genes, African-American genotype distributions were different from those for Caucasians, particularly when contrasting the cases and controls. In some cases, the rare allele for Caucasians is the common allele in African-Americans (e.g., rs222016). For data presentation purposes, the risk allele for both groups was defined as the rare allele in Caucasians.
With only a few exceptions, the majority of SNPs in vitamin D-related genes had neither overall associations nor any interactions with finasteride (PCPT treatment arm) in relation to prostate cancer risk (Table 2). GC/rs222016 in Caucasians had a per minor allele prostate cancer odds ratio =1.37 (95%CI 1.06–1.79) for the finasteride arm and a per minor allele odds ratio = 0.85 (95%CI0.68–1.05) for the placebo arm, p-interaction <0.05. Another GC SNP, rs222014, in Caucasians had a per minor allele prostate cancer odds ratio =1.36 (95% CI 1.00–1.86) on the finasteride arm and per minor allele odds ratio =0.85 (95% CI 0.65–1.08) for the placebo arm, p-interaction<0.05. CYP27B1/rs703842 in Caucasians had a per allele prostate cancer odds ratio = 0.76 (95%CI 0.62–0.95) for the finasteride arm and odds ratio = 1.10 (95%CI 0.94–1.30) for the placebo arm, p-interaction<0.05. Among African-Americans, C10orf88 /rs6599638 had a per minor allele prostate cancer odds ratio=4.88 (95%CI 1.96–12.13) for the finasteride arm and OR= 1.39 (95%CI 0.65–2.97) for the placebo arm, p-interaction<0.05.
Table 2.
Evaluating Interactions of Vitamin D Related Genotypes and Finasteride with Prostate Cancer in the Prostate Cancer Prevention Trial, Stratified by Race.
| Caucasian | African American | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N case/control | N case/control | |||||||||||
| Allele | Number of minor alleles | OR (95% CI)1 | Number of minor alleles | OR (95% CI)1 | ||||||||
| Gene | rs number | Treatment Arm | Major | Minor | 0 | 1 | 2 | 0 | 1 | 2 | ||
| VDR | rs1544410 | Finasteride | G | A | 156/129 | 212/189 | 75/97 | 0.83 (0.69–1.01) | 8/37 | 11/38 | 1/6 | 1.11 (0.48–2.56) |
| Placebo | 222/233 | 301/315 | 100/101 | 1.00 (0.85–1.18) | 8/28 | 13/18 | 5/3 | 2.48 (1.07–5.78) | ||||
| VDR | rs11568820 | Finasteride | G | A | 284/259 | 139/146 | 20/13 | 1.01 (0.79–1.28) | 2/6 | 6/27 | 11/48 | 0.88 (0.40–1.92) |
| Placebo | 405/418 | 206/207 | 22/26 | 1.01 (0.83–1.24) | 2/5 | 9/14 | 16/30 | 1.13 (0.49–2.59) | ||||
| VDR | rs757343 | Finasteride | G | A | 349/325 | 84/81 | 7/4 | 0.99 (0.73–1.35) | 17/69 | 3/11 | 0/1 | 1.12 (0.29–4.35) |
| Placebo | 482/489 | 138/153 | 8/8 | 0.93 (0.73–1.18) | 24/41 | 3/7 | 0.93 (0.19–4.45) | |||||
| VDR | rs11574143 | Finasteride | G | A | 369/347 | 71/68 | 4/3 | 0.97 (0.69–1.37) | 18/66 | 2/13 | 0/2 | 0.50 (0.11–2.37) |
| Placebo | 523/525 | 107/122 | 3/6 | 0.84 (0.64–1.10) | 23/40 | 4/9 | 1.22 (0.29–5.09) | |||||
| GC | rs7041 | Finasteride | C | A | 149/142 | 215/191 | 79/81 | 0.97 (0.80–1.17) | 2/6 | 5/23 | 13/52 | 0.89 (0.40–1.95) |
| Placebo | 186/199 | 312/314 | 127/139 | 1.00 (0.85–1.17) | 0/1 | 10/21 | 17/26 | 1.22 (0.43–3.47) | ||||
| GC | rs222014 | Finasteride | G | A | 339/339 | 93/74 | 8/4 | 1.36 (1.00–1.86)+** | 12/64 | 7/16 | 0/1 | 1.83 (0.65–5.10) |
| Placebo | 508/509 | 117/134 | 7/9 | 0.84 (0.65–1.08)** | 18/44 | 9/4 | 0/1 | 3.16 (0.92–10.89) | ||||
| GC | rs222016 | Finasteride | A | G | 296/310 | 131/99 | 15/10 | 1.37 (1.06–1.79)+ ** | 2/20 | 8/43 | 8/17 | 2.12 (0.92–4.88) |
| Placebo | 455/460 | 166/169 | 11/24 | 0.85 (0.68–1.05)** | 3/12 | 16/24 | 6/11 | 1.26 (0.56–2.87) | ||||
| GC | rs705117 | Finasteride | A | G | 318/307 | 110/95 | 14/14 | 1.03 (0.79–1.34) | 1/11 | 10/31 | 9/39 | 1.10 (0.52–2.30) |
| Placebo | 466/466 | 152/171 | 14/16 | 0.91 (0.73–1.13) | 7/5 | 12/27 | 8/17 | 0.79 (0.34–1.81) | ||||
| GC | rs2282679 | Finasteride | A | C | 240/223 | 173/163 | 28/33 | 0.94 (0.75–1.17) | 16/66 | 3/14 | 0/1 | 0.72 (0.19–2.72) |
| Placebo | 311/325 | 264/270 | 60/60 | 1.02 (0.86–1.21) | 17/38 | 10/11 | 1.07 (0.32–3.53) | |||||
| GC | rs12512631 | Finasteride | A | G | 175/166 | 223/196 | 43/53 | 0.95 (0.77–1.17) | 10/28 | 7/45 | 2/8 | 0.60 (0.26–1.38) |
| Placebo | 287/313 | 281/267 | 64/72 | 1.05 (0.89–1.24) | 15/23 | 9/21 | 3/5 | 1.12 (0.49–2.57) | ||||
| C10orf88 | rs6599638 | Finasteride | G | A | 132/123 | 219/204 | 87/85 | 0.92 (0.76–1.12) | 3/37 | 9/39 | 7/5 | 4.88 (1.96–12.13)+ ** |
| Placebo | 179/206 | 304/309 | 141/129 | 1.11 (0.95–1.30) | 9/23 | 13/18 | 5/8 | 1.39 (0.65–2.97)** | ||||
| CYP2R1 | rs2060793 | Finasteride | G | A | 170/163 | 211/181 | 62/73 | 0.93 (0.76–1.13) | 8/34 | 11/36 | 1/11 | 0.82 (0.38–1.77) |
| Placebo | 237/249 | 294/304 | 103/99 | 1.02 (0.87–1.20) | 14/24 | 10/22 | 2/3 | 1.68 (0.65–4.35) | ||||
| CYP24A1 | rs2248359 | Finasteride | G | A | 166/148 | 200/197 | 76/74 | 0.97 (0.80–1.17) | 3/14 | 6/35 | 9/32 | 1.29 (0.61–2.73) |
| Placebo | 241/224 | 309/303 | 79/122 | 0.81 (0.69–0.95) | 3/8 | 11/21 | 12/20 | 1.69 (0.75–3.81) | ||||
| CYP24A1 | rs2762942 | Finasteride | A | G | 398/382 | 42/36 | 2/1 | 1.13 (0.72–1.77) | 20/77 | 0/4 | ||
| Placebo | 570/572 | 60/81 | 5/1 | 0.87 (0.62–1.21) | 26/46 | 1/3 | 0.31 (0.02–4.90) | |||||
| CYP24A1 | rs4809958 | Finasteride | A | C | 306/302 | 125/108 | 12/7 | 1.17 (0.89–1.53) | 17/68 | 2/13 | 0.63 (0.12–3.25) | |
| Placebo | 460/466 | 159/174 | 13/13 | 0.96 (0.77–1.20) | 26/41 | 1/8 | 0.17 (0.02–1.77) | |||||
| CYP24A1 | rs6013897 | Finasteride | G | A | 275/260 | 144/141 | 23/16 | 1.04 (0.82–1.31) | 7/50 | 12/25 | 1/6 | 1.80 (0.84–3.83) |
| Placebo | 391/378 | 210/243 | 31/30 | 0.92 (0.76–1.11) | 19/27 | 7/21 | 1/1 | 0.68 (0.24–1.92) | ||||
| CYP24A1 | rs6022999 | Finasteride | A | G | 244/235 | 170/159 | 28/23 | 1.09 (0.87–1.36) | 4/10 | 8/42 | 8/29 | 0.96 (0.46–1.99) |
| Placebo | 364/382 | 238/228 | 31/44 | 0.98 (0.81–1.17) | 3/10 | 15/22 | 9/16 | 1.37 (0.63–2.97) | ||||
| CYP27B1 | rs703842 | Finasteride | A | G | 212/174 | 199/187 | 31/54 | 0.76 (0.62–0.95)+ ** | 10/42 | 10/28 | 0/10 | 0.77 (0.35–1.70) |
| Placebo | 282/303 | 271/279 | 81/67 | 1.10 (0.94–1.30)** | 12/17 | 14/25 | 1/6 | 0.69 (0.29–1.65) | ||||
| DHCR7 | rs1790349 | Finasteride | A | G | 318/295 | 111/116 | 14/6 | 0.98 (0.75–1.29) | 12/53 | 8/23 | 0/5 | 1.01 (0.43–2.41) |
| Placebo | 443/472 | 167/168 | 22/14 | 1.15 (0.93–1.42) | 19/32 | 8/15 | 0/2 | 0.61 (0.20–1.84) | ||||
| NADSYN1 | rs3829251 | Finasteride | G | A | 315/285 | 100/116 | 18/10 | 0.92 (0.71–1.19) | 11/48 | 8/25 | 0/7 | 0.85 (0.37–1.96) |
| Placebo | 442/457 | 159/173 | 21/15 | 1.05 (0.85–1.30) | 16/26 | 11/20 | 0/3 | 0.76 (0.28–2.04) | ||||
| NADSYN1 | rs12785878 | Finasteride | A | C | 250/221 | 159/172 | 33/26 | 0.94 (0.75–1.17) | 2/4 | 8/39 | 10/38 | 0.89 (0.39–2.04) |
| Placebo | 349/350 | 240/251 | 45/52 | 0.98 (0.82–1.16) | 3/3 | 12/18 | 12/28 | 0.77 (0.34–1.76) | ||||
All models adjusted for age, BMI, PSA, and family history of prostate cancer
Main effects p-value significant after a within gene Bonferroni adjustment. Significance thresholds among genes were : VDR = 0.013 (4 tests), GC= 0.008 (6 tests), C10corf88 = 0.05 (1 test), CYP2R1 = 0.05 (1 test), CYP24A1 = 0.01 (5 tests), CYP27B1 = 0.05 (1 test), DHCR7= 0.05 (1 test), and NADSYN1 = 0.025 (2 tests).
p-interaction<0.05.
Per minor allele
We next investigated vitamin D-related genotype associations and their potential interactions with PCPT-treatment arm restricted to high-grade prostate cancer (Gleason ≥7) (Table 3). These analyses were conducted only in Caucasians due to sample size considerations. Most of these relationships were null with no apparent differences in the magnitude or direction of associations by PCPT treatment arm. The two exceptions were VDR (rs1544410), which was associated with a per minor allele increased risk of 1.40 (95% CI 1.05–1.86) for the placebo arm whereas the finasteride arm showed an inverse association (OR=0.81, 95%CI 0.62–1.06), p-interaction<0.01; and CYP27B1 (rs703842), which was associated with a per minor allele increased risk of 1.28 (95% CI 0.96–1.70) on the placebo arm, and a per minor allele decreased risk of 0.79 (95% CI 0.59–1.07) on the finasteride arm, p-interaction<0.05.
Table 3.
Evaluating Interactions of Vitamin D Related Genotypes and Finasteride with high-grade prostate cancer1 in Caucasian men in the PCPT.
| Finasteride | Placebo | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| N case/control | N case/control | ||||||||||
| Allele | Number of minor alleles | Number of minor alleles | |||||||||
| Gene | rs number | Major | Minor | 0 | 1 | 2 | OR (95% CI) | 0 | 1 | 2 | OR (95% CI) |
| VDR | rs1544410 | G | A | 53/129 | 72/189 | 24/97 | 0.81 (0.62–1.06)** | 30/233 | 62/315 | 26/101 | 1.40 (1.05–1.86)** |
| VDR | rs11568820 | G | A | 101/259 | 41/146 | 7/13 | 0.91 (0.64–1.28) | 66/418 | 51/207 | 2/26 | 1.31 (0.93–1.85) |
| VDR | rs757343 | G | A | 116/325 | 32/81 | 1/4 | 0.98 (0.64–1.50) | 93/489 | 22/153 | 1/8 | 0.74 (0.46–1.18) |
| VDR | rs11574143 | G | A | 121/347 | 27/68 | 1/3 | 1.03 (0.65–1.64) | 97/525 | 22/122 | 0/6 | 0.86 (0.52–1.40) |
| GC | rs7041 | C | A | 52/142 | 64/191 | 33/81 | 1.03 (0.79–1.35) | 26/199 | 69/314 | 23/139 | 1.17 (0.88–1.55) |
| GC | rs222014 | G | A | 119/339 | 27/74 | 2/4 | 1.14 (0.73–1.77) | 93/509 | 24/134 | 2/9 | 0.95 (0.61–1.48) |
| GC | rs222016 | A | G | 105/310 | 37/99 | 7/10 | 1.25 (0.87–1.80) | 82/460 | 31/169 | 5/24 | 1.02 (0.71–1.49) |
| GC | rs705117 | A | G | 105/307 | 38/95 | 6/14 | 1.08 (0.76–1.55) | 78/466 | 40/171 | 1/16 | 1.19 (0.83–1.73) |
| GC | rs2282679 | A | C | 82/223 | 56/163 | 10/33 | 0.95 (0.70–1.29) | 57/325 | 50/270 | 12/60 | 1.07 (0.79–1.44) |
| GC | rs12512631 | A | G | 60/166 | 76/196 | 12/53 | 0.90 (0.67–1.20) | 56/313 | 54/267 | 8/72 | 0.91 (0.67–1.24) |
| C10orf88 | rs6599638 | G | A | 52/123 | 66/204 | 30/85 | 0.82 (0.62–1.08) | 38/206 | 54/309 | 24/129 | 0.97 (0.73–1.29) |
| CYP2R1 | rs2060793 | G | A | 50/163 | 76/181 | 23/73 | 1.07 (0.82–1.41) | 47/249 | 55/304 | 17/99 | 0.92 (0.69–1.23) |
| CYP24A1 | rs2248359 | G | A | 57/148 | 64/197 | 27/74 | 0.98 (0.74–1.28) | 46/224 | 57/303 | 16/122 | 0.81 (0.60–1.08) |
| CYP24A1 | rs2762942 | A | G | 133/382 | 15/36 | 0/1 | 1.11 (0.60–2.06) | 106/572 | 12/81 | 1/1 | 0.91 (0.50–1.65) |
| CYP24A1 | rs4809958 | A | C | 102/302 | 41/108 | 5/7 | 1.21 (0.83–1.74) | 81/466 | 34/174 | 4/13 | 1.22 (0.84–1.79) |
| CYP24A1 | rs6013897 | T | A | 94/260 | 44/141 | 10/16 | 1.05 (0.76–1.46) | 80/378 | 37/243 | 2/30 | 0.69 (0.48–1.00) |
| CYP24A1 | rs6022999 | A | G | 87/235 | 52/159 | 9/23 | 0.97 (0.70–1.33) | 69/382 | 48/228 | 2/44 | 0.85 (0.60–1.19) |
| CYP27B1 | rs703842 | A | G | 71/174 | 65/187 | 12/54 | 0.79 (0.59–1.07)** | 50/303 | 48/279 | 21/67 | 1.28 (0.96–1.70)** |
| DHCR7 | rs1790349 | A | G | 113/295 | 26/116 | 10/6 | 0.93 (0.64–1.36) | 77/472 | 38/168 | 3/14 | 1.34 (0.93–1.93) |
| NADSYN1 | rs3829251 | G | A | 111/285 | 26/116 | 9/10 | 0.84 (0.58–1.21) | 77/457 | 35/173 | 3/15 | 1.19 (0.82–1.73) |
| NADSYN1 | rs12785878 | A | C | 79/221 | 54/172 | 15/26 | 1.06 (0.79–1.44) | 62/350 | 45/251 | 12/52 | 1.15 (0.84–1.56) |
Gleason 7–10
All models adjusted for age, BMI, PSA, and family history of prostate cancer
Main effects p-value significant after a within gene Bonferroni adjustment. Significance thresholds among genes were: VDR =0.013 (4 tests), GC =0.008 (6 tests), C10corf88 =0.05 (one test), CYP2R1 =0.05 (one test), CYP24A1 = 0.01 (5 tests), CYP27B1 = 0.05 (1 test), DHCR7 = 0.05 (1 test), and NADSYN1 = 0.025 (2 tests).
p-interaction <0.05.
We explored associations between SNPs and baseline season-adjusted serum 25(OH)D as well as whether any SNP-prostate cancer risk associations were modified by serum 25(OH)D (placebo arm only) (Table 4). Five of the 21 genes examined exhibited a per minor allele significant association with serum 25(OH)D after Bonferroni adjustment. Some SNPs conferred increases in serum 25(OH) D while others conferred decreases. For example, two GC SNPs (rs7041 and rs2282679) were associated with significantly lower serum 25(OH) as the number of minor alleles increased, whereas GC (rs12512631) was associated with higher serum 25(OH) as the number of minor alleles increased. CYP2R1 (rs2060793) was associated with higher serum 25(OH)D and minor alleles for NADSYN1 (rs12785878) were associated with lower serum 25(OH)D. Despite the influence of these SNPs on serum 25(OH)D, there was only one borderline significant SNP-serum 25(OH)D interaction in relation to prostate cancer risk, for CYP24A1/fs2248359, p-interaction=0.049.
Table 4.
Associations of genotypes with season-adjusted serum 25(OH)D among Caucasian men in the PCPT placebo arm.
| Number of minor alleles | ||||||||
|---|---|---|---|---|---|---|---|---|
| rs number | Allele | 0 | 1 | 2 | ||||
| Major | Minor | Case/Ctl | Mean 25(OH)D1 | Case/Ctl | Mean 25(OH)D1 | Case/Ctl | Mean 25(OH)D1 | |
| rs1544410 | G | A | 222/233 | 59.44 | 301/315 | 58.75 | 100/101 | 60.61 |
| rs11568820 | G | A | 405/418 | 58.19 | 206/207 | 61.67 | 22/26 | 59.45 |
| rs757343 | G | A | 482/489 | 59.51 | 138/153 | 58.59 | 8/8 | 61.61 |
| rs11574143 | G | A | 523/525 | 59.26 | 107/122 | 59.83 | 3/6 | 54.10 |
| rs7041 | C | A | 186/199 | 62.48** | 312/314 | 59.14** | 127/139 | 55.60** |
| rs222014 | G | A | 508/509 | 59.23 | 117/134 | 59.52 | 7/9 | 63.86 |
| rs222016 | A | G | 455/460 | 59.02 | 166/169 | 60.10 | 11/24 | 59.80 |
| rs705117 | A | G | 466/466 | 59.30 | 152/171 | 58.61 | 14/16 | 66.53 |
| rs2282679 | A | C | 311/325 | 61.90** | 264/270 | 58.03** | 60/60 | 52.13** |
| rs12512631 | A | G | 287/313 | 57.14** | 281/267 | 60.46** | 64/72 | 64.86** |
| rs6599638 | G | A | 179/206 | 59.42 | 304/309 | 59.31 | 141/129 | 59.70 |
| rs2060793 | G | A | 237/249 | 57.34** | 294/304 | 59.79** | 103/99 | 63.04** |
| rs2248359 | G | A | 241/224 | 59.67 | 309/303 | 59.96 | 79/122 | 56.74† |
| rs2762942 | A | G | 570/572 | 59.24 | 60/81 | 60.18 | 5/1 | 59.31 |
| rs4809958 | A | C | 460/466 | 59.50 | 159/174 | 59.08 | 13/13 | 56.21 |
| rs6013897 | T | A | 391/378 | 60.07 | 210/243 | 58.06 | 31/30 | 59.10 |
| rs6022999 | A | G | 364/382 | 59.71 | 238/228 | 59.40 | 31/44 | 54.98 |
| rs703842 | A | G | 282/303 | 59.18 | 271/279 | 58.57 | 81/67 | 62.58 |
| rs1790349 | A | G | 443/472 | 59.47 | 167/168 | 58.53 | 22/14 | 63.21 |
| rs3829251 | G | A | 442/457 | 59.91 | 159/173 | 57.13 | 21/15 | 64.36 |
| rs12785878 | A | C | 349/350 | 60.97** | 240/251 | 57.71** | 45/52 | 56.20** |
Means are geometric means.
<49.5 nmol/L
49.5–66.9 nmol/L
≥66.9 nmol/L
P-value comparing mean serum 25(OH)D concentration by genotype significant after a within gene Bonferroni adjustment, calculated using linear regression with log[25(OH)D] as dependent variable. Significance thresholds among genes were: VDR = 0.013 (4 tests), GC = 0.008 (6 tests), C10corf88 = 0.05 (1 test), CYP2R1= 0.05 (1 test), CYP24A1 = 0.01 (5 tests), CYP27B1 = 0.05 (1 test), DHCR7= 0.05 (1 test), and NADSYN1 = 0.025 (2 tests).
P-interaction = 0.049 for test of serum 25(OH)D x SNP interaction in relation to prostate cancer risk. All other p-interaction tests>0.10
DISCUSSION
In this nested case-control study of 2,156 Caucasian and 177 African American men enrolled in the Prostate Cancer Prevention Trial (PCPT), we investigated the associations of 21 SNPs in vitamin D pathway genes with prostate cancer risk by trial-randomization treatment group and by race. We also examined SNP-cancer associations with risk for high-grade disease, but only among Caucasians due to small numbers of African-Americans and we examined whether SNP-serum 25(OH)D interactions influenced prostate cancer risk. Our main findings are: 1) characteristics in vitamin D pathway genes vary slightly between men of Caucasian and African ancestry; 2) for a few select genes, both the magnitude and direction of association with prostate cancer varies by finasteride vs. placebo use; 4) two SNPs – one in VDR and one in CYP27B1 conferred increased prostate cancer risks on the placebo arm but inverse associations on the finasteride arm; and 3) some but not all vitamin D pathway genes and their variants influence serum 25(OH)D. However, the ensuing relationships with prostate cancer were not clear because the serum 25(OH)-genotype interaction tests in relation to prostate cancer were either marginally significant or not statistically significant.
One of the important design aspects of this study is that it leveraged data and specimens from a randomized, placebo-controlled trial testing a 5-ɑ-reductase inhibitor vs. placebo for primary prevention of prostate cancer (21). All participants had standardized PSA at baseline and all participants had protocol-defined end-of-study biopsies read by a centralized pathologist. Modest but noticeable differences emerged in gene-prostate cancer risk associations that varied slightly by trial treatment arm. For example, three SNPs were associated with per minor allele increased prostate cancer risk only for those in the finasteride arm (GC/rs222016 and GC/rs222014 in Caucasians and C10orf88/rs6599638 in African-Americans) while one SNP (CYP27B1/rs703842) in Caucasians was associated with lower risk in the finasteride arm. We were particularly interested in the two SNPs that demonstrated inverse associations for high-grade prostate cancer on the finasteride arm and increased risk on the placebo arm (VDR/rs1544410 CYP27B1/rs703842). While we are unsure of the underlying biological reasons for these observed associations, it is possible that there is shared biology between the genotypes and the metabolism and disposition of finasteride, thereby conferring the significant treatment-interaction effects. High-grade prostate cancer is harder to treat, often becoming castrate-resistant so the results are somewhat intriguing.
We know of no other studies that have examined drug-vitamin D SNP interactions in relation to prostate cancer risk. However, some of our results are generally consistent with other studies that have found overall modest relationships of SNPs in vitamin D pathway genes with risk of both total and high-grade prostate cancer. Gilbert et al conducted a nested case control study (n=1,275 cases and 2,062 controls) using data and specimens from the UK-based ProtecT trial (8). Sixteen vitamin-D related SNPs were tested, five of which overlapped with those we tested here in the PCPT. Of those five, results were similar to our results showing no overall associations and no differences by disease grade where they defined high-grade in the same manner as PCPT (Gleason score ≥ 7). In the prostate cancer portion of the Breast and Prostate Cancer Cohort Consortium (BPC3), Mondul et al used data from published GWAS to identify vitamin-related SNPs (25). Two SNPs in that analysis overlapped with our analysis (rs2282679 and rs6013897) and similarly, the BPC3 data show no clear pattern of association of genetic variation with overall prostate cancer risk. With regards to high grade disease, our observed association of VDR/rs1544410 SNP (also known as BSM1) in high-grade cancer in the placebo arm only is consistent with another study where the minor allele was associated with lethal prostate cancer and marginally associated with high-grade cancer in Caucasians (22). This finding is further supported by a meta-analysis of 13 studies showing that BSMI GG was associated with high Gleason score (≥7) (26). In contrast, another study found no associations with prostate cancer specific death or recurrence (23).
Contrary to our a priori hypothesis, while we found that some SNPs influenced serum 25(OH)D, there was only one marginal SNP-serum 25(OH)D interaction in relation to prostate cancer risk. These PCPT findings are generally consistent with other published studies where there do not appear to be clear or consistent associations of vitamin D pathway genes, and their interaction with serum 25(OH)D in relation to prostate cancer risk (27). Shui et al used data and specimens from the Prostate Cancer Cohort Consortium to test associations of vitamin D pathway genes with fatal prostate cancer as defined from death certificates, medical records and cancer registries (28). Twenty-one genes were examined, only one of which overlapped with our analysis (rs2060793). The investigators reported a fatal prostate cancer odds ratio of 1.34 (95% CI 1.0–1.79) for those with low serum 25(OH)D, whereas we found no overall or high-grade association per minor allele for this SNP nor any differential association by baseline serum 25(OH)D despite our demonstration of significantly higher serum 25(OH)D among those with 2 minor alleles. However, our results are in contrast to Ahn et al. who reported the strongest gene-serum 25(OH)D risk associations for study participants with the minor allele of rs11574143 who were in the lowest tertile of serum 25(OH)D (20). One possible explanation for our lack of SNP-serum 25(OH)D-prostate cancer interactions is that when we stratified the genotype groups across the low, middle and high tertiles of serum 25(OH)D, the cell sizes became very small. Unlike many studies, we used season-adjusted measures of serum 25(OH), a strategy that we have previously found to be very helpful when blood samples are collected across the four seasons, which is know to contribute to vitamin D variability (10, 29). Lack of consistency between published studies may also be attributable to differences in assays used to measure serum 25(OH)D (30, 31) It is quite likely that to further the understanding of vitamin D pathway genes and their relationship to prostate cancer risk, pooled data from multiple cohorts, particularly those enriched with African-American men, will be needed. It is also possible that these relationships are much more complex than the current science is able to unravel or that unmeasured or uncontrolled confounding prevented detection of meaningful associations that could be applied to general population prevention strategies. Other strategies such as Mendelian randomization may be particularly useful as recently shown by Dimitraopoulou et al who reported no associations of vitamin-D related SNPs with prostate cancer (27).
This study has several strengths. Our study sample was derived from a completed phase III clinical trial (21). As such, the data were uniformly and rigorously collected at more than 200 clinical sites across the United States. Importantly, both prostate cancer cases and controls have definitive evidence for presence or absence of disease based on the biopsy protocol. This characteristic of the study design minimizes any disease misclassification and the contamination of the control group with undiagnosed cases. We also used season-adjusted serum 25(OH)D measures, which reduces the well- known variation in this measure when including blood samples that have been collected across the four calendar seasons (10, 29). Limitations include that while we did include African-Americans, the sample size for that group was still smaller than ideal, particularly when applying a within-gene Bonferroni adjustment. The study design stipulated that we oversample African-Americans who were PCPT participants and who met the eligibility criteria for the case-control study. Even with this design feature, the number of African American participants was low and these findings should be viewed as exploratory. Genetic variation conferring risk in African-American men may vary from that for Caucasian men and this is an area that deserves further study (32–34) as most genetic discovery studies have been conducted in European ancestry individuals, (15), limiting inferences that can be applied to African-Americans. Genetic epidemiology is a constantly changing field; it is possible that some informative vitamin D-related SNPs emerged after our laboratory analyses were completed (15). New panels of GWAS-identified SNPs (15) have different SNPs than those used in the present study; it is possible that results of the present study would differ using some of the recently identified SNPs. Another limitation is that we examined many candidate, a priori SNPs. While applied within-gene multiple comparisons tests, but the possibility exists that false positives could still be present. We believe that this risk is minimal given that most of the associations were null. We urge readers to view results with caution and to be guided by sample size and the ORs and 95% confidence intervals, which provide the magnitude of association and precision of the estimates. Finally, we examined genetic variation in germline DNA; results may differ when investigating vitamin D-related gene expression in both cancer and adjacent normal tissue in the prostate (35, 36).
In conclusion, a very limited number of vitamin D pathway SNPs were associated with both total and high-grade prostate cancer. We found suggestive evidence for a limited number of genotype-finasteride interactions where the direction of the prostate cancer risk varied by PCPT treatment arm.
Supplementary Material
Acknowledgments
Financial Support: This work was supported by National Cancer Institute, National Institutes of Health, United States Department of Health and Human Services grant awards R01 CA138639 (Vitamin D and Prostate Cancer: Biomarkers and Genetic Variation), U01 CA37429 (Prostate Cancer Prevention Trial), P01 CA108964 (Biology of the Prostate Cancer Prevention Trial), UM1 CA182883 (PCPT and SELECT Cohorts: Core Infrastructure Support for Cancer Research) and UG1 CA189974 (NCORP Research Base). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
The authors declare no potential conflicts of interest.
References
- 1.Siegel RL, Miller KD, and Jemal A. Cancer Statistics, 2019. CA Cancer J Clin 2019; 69: 7–34. [DOI] [PubMed] [Google Scholar]
- 2.Feldman D, Krishnan AV, Swami S, Giovannucci E, and Feldman BJ The role of vitamin D in reducing cancer risk and progression. Nat Rev Cancer 2014; 14: 342–57. [DOI] [PubMed] [Google Scholar]
- 3.Deeb KK, Trump DL, and Johnson CS Vitamin D signalling pathways in cancer: potential for anticancer therapeutics. Nat Rev Cancer 2007; 7: 684–700. [DOI] [PubMed] [Google Scholar]
- 4.Trump DL, Deeb KK, and Johnson CS Vitamin D: considerations in the continued development as an agent for cancer prevention and therapy. Cancer J 2010; 16: 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hsu JW, Yasmin-Karim S, King MR, Wojciechowski JC, Mickelsen D, Blair ML et al. Suppression of prostate cancer cell rolling and adhesion to endothelium by 1alpha,25-dihydroxyvitamin D3. Am J Pathol 2011; 178: 872–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Albanes D, Mondul AM, Yu K, Parisi D, Horst RL, Virtamo J. et al. Serum 25-hydroxy vitamin D and prostate cancer risk in a large nested case-control study. Cancer Epidemiol Biomarkers Prev 2011; 20: 1850–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Gandini S, Boniol M, Haukka J, Byrnes G, Cox B, Sneyd MJ et al. Meta-analysis of observational studies of serum 25-hydroxyvitamin D levels and colorectal, breast and prostate cancer and colorectal adenoma. Int J Cancer 2011;128: 1414–24. [DOI] [PubMed] [Google Scholar]
- 8.Gilbert R, Martin RM, Beynon R, Harris R, Savovic J, Zuccolo L. et al. Associations of circulating and dietary vitamin D with prostate cancer risk: a systematic review and dose-response meta-analysis. Cancer Causes Control 2011; 22: 319–40. [DOI] [PubMed] [Google Scholar]
- 9.Yin L, Raum E, Haug U, Arndt V, and Brenner H. Meta-analysis of longitudinal studies: Serum vitamin D and prostate cancer risk. Cancer Epidemiol, 2009: 33: 435–45. [DOI] [PubMed] [Google Scholar]
- 10.Schenk JM, Till CA, Tangen CM, Goodman PJ, Song X, Torkko KC et al. , Serum 25-hydroxyvitamin D concentrations and risk of prostate cancer: results from the Prostate Cancer Prevention Trial. Cancer Epidemiol Biomarkers Prev 2014; 23:1484–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Xu Y, Shao X, Yao Y, Xu L, Chang L, Jiang Z. et al. Positive association between circulating 25-hydroxyvitamin D levels and prostate cancer risk: new findings from an updated meta-analysis. J Cancer Res Clin Oncol 2014: 140: 1465–77. [DOI] [PubMed] [Google Scholar]
- 12.Travis RC, Perez-Cornago A, Appleby PN, Albanes D, Joshu CE, Lutsey PL et al. A Collaborative Analysis of Individual Participant Data from 19 Prospective Studies Assesses Circulating Vitamin D and Prostate Cancer Risk. Cancer Res 2019: 79: 274–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Neuhouser ML, Manson JE, Millen A, Pettinger M, Margolis K, Jacobs ET et al. The influence of health and lifestyle characteristics on the relation of serum 25-hydroxyvitamin D with risk of colorectal and breast cancer in postmenopausal women. Am J Epidemiol 2012; 175:673–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.McGrath JJ, Saha S, Burne TH, and Eyles DW A systematic review of the association between common single nucleotide polymorphisms and 25-hydroxyvitamin D concentrations. J Steroid Biochem Mol Biol 2010;121: 471–7. [DOI] [PubMed] [Google Scholar]
- 15.Jiang X, O’Reilly PF, Aschard H, Hsu YH, Richards JB, Dupuis J. et al. Genome-wide association study in 79,366 European-ancestry individuals informs the genetic architecture of 25-hydroxyvitamin D levels. Nat Commun 2018; 260:1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Schwartz GG, Whitlatch LW, Chen TC, Lokeshwar BL, and Holick MF Human prostate cells synthesize 1,25-dihydroxyvitamin D3 from 25-hydroxyvitamin D3. Cancer Epidemiol Biomarkers Prev 1998; 7: 391–5. [PubMed] [Google Scholar]
- 17.Ahn J, Yu K, Stolzenberg-Solomon R, Simon KC, McCullough ML, Gallicchio L.. Genome-wide association study of circulating vitamin D levels. Hum Mol Genet 2010;19: 2739–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wang TJ, Zhang F, Richards JB, Kestenbaum B, van Meurs JB, Berry D.. Common genetic determinants of vitamin D insufficiency: a genome-wide association study. Lancet 2010;376: 180–8.. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.McCullough ML, Bostick RM, and Mayo TL Vitamin D gene pathway polymorphisms and risk of colorectal, breast, and prostate cancer. Annu Rev Nutr 2009: 29:111–32. [DOI] [PubMed] [Google Scholar]
- 20.Ahn J, Albanes D, Berndt SI, Peters U, Chatterjee N, Freedman ND,. Vitamin D-related genes, serum vitamin D concentrations and prostate cancer risk. Carcinogenesis 2009; 30: 769–76.. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Thompson IM, Goodman PJ, Tangen CM, Lucia MS, Miller GJ, Ford LG. The influence of finasteride on the development of prostate cancer. N Engl J Med 2003; 349: 215–24. [DOI] [PubMed] [Google Scholar]
- 22.Shui IM, Mucci LA, Kraft P, Tamimi RM, Lindstrom S, Penney KL, Vitamin D-related genetic variation, plasma vitamin D, and risk of lethal prostate cancer: a prospective nested case-control study. J Natl Cancer Inst 2012; 104: 690–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Holt SK, Kwon EM, Koopmeiners JS, Lin DW, Feng Z, Ostrander EA Vitamin D pathway gene variants and prostate cancer prognosis. Prostate 2010; 70: 1448–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Levin GP, Robinson-Cohen C, de Boer IH, Houston DK, Lohman K, Liu Y.. Genetic variants and associations of 25-hydroxyvitamin D concentrations with major clinical outcomes. JAMA 2012: 308: 1898–905. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Mondul AM, Shui IM, Yu K, Travis RC, Stevens VL, Campa D. Genetic variation in the vitamin d pathway in relation to risk of prostate cancer--results from the breast and prostate cancer cohort consortium. Cancer Epidemiol Biomarkers Prev 2012; 22: 688–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Chen L, Davey Smith G, Evans DM, Cox A, Lawlor DA, Donovan J, Genetic variants in the vitamin D receptor are associated with advanced prostate cancer at diagnosis: findings from the prostate testing for cancer and treatment study and a systematic review. Cancer Epidemiol Biomarkers Prev 2009: 18: 2874–81. [DOI] [PubMed] [Google Scholar]
- 27.Dimitrakopoulou VI, Travis RC, Shui IM, Mondul A, Albanes D, Virtamo J. Interactions Between Genome-Wide Significant Genetic Variants and Circulating Concentrations of 25-Hydroxyvitamin D in Relation to Prostate Cancer Risk in the National Cancer Institute BPC3. Am J Epidemiol 2017: 185:452–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Shui IM, Mondul AM, Lindstrom S, Tsilidis KK, Travis RC, Gerke T,. Circulating vitamin D, vitamin D-related genetic variation, and risk of fatal prostate cancer in the National Cancer Institute Breast and Prostate Cancer Cohort Consortium. Cancer 2015; 121:1949–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kristal AR, Till C, Song X, Tangen CM, Goodman PJ, Neuhouser ML Plasma vitamin D and prostate cancer risk: results from the Selenium and Vitamin E Cancer Prevention Trial. Cancer Epidemiol Biomarkers Prev 2014; 23:1494–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Schleicher RL, Sternberg MR, Lacher DA, Sempos CT, Looker AC, Durazo-Arvizu A Method-bridging Study for Serum 25-hydroxyvitamin D to Standardize Historical Radioimmunoassay Data to Liquid Chromatography-Tandem Mass Spectrometry. Natl Health Stat Report 2016: 1–16. [PubMed] [Google Scholar]
- 31.Schleicher RL, Sternberg MR, Lacher DA, Sempos CT, Looker AC, Durazo-Arvizu RA The vitamin D status of the US population from 1988 to 2010 using standardized serum concentrations of 25-hydroxyvitamin D shows recent modest increases. Am J Clin Nutr 2016; 104: 454–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Lin HY, Callan CY, Fang Z, Tung HY, and Park JY Interactions of PVT1 and CASC11 on Prostate Cancer Risk in African Americans. Cancer Epidemiol Biomarkers Prev 2019; 28:1067–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Freedman ML, Haiman CA, Patterson N, McDonald GJ, Tandon A, Waliszewska A. Admixture mapping identifies 8q24 as a prostate cancer risk locus in African-American men. Proc Natl Acad Sci USA 2006; 103: 14068–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Haiman CA, Chen GK, Blot WJ, Strom SS, Berndt SI, Kittles RA,. Characterizing genetic risk at known prostate cancer susceptibility loci in African Americans. PLoS Genet 2011, 7: e1001387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Baumann B, Lugli G, Gao S, Zenner M, and Nonn L. High levels of PIWI-interacting RNAs are present in the small RNA landscape of prostate epithelium from vitamin D clinical trial specimens. Prostate 2019; 79: 840–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Khan NA, Stopsack KH, Allott EH, Gerke T, Giovannucci EL, Mucci LA et al. Intratumoral Sterol-27-Hydroxylase (CYP27A1) Expression in Relation to Cholesterol Synthesis and Vitamin D Signaling and Its Association with Lethal Prostate Cancer. Cancer Epidemiol Biomarkers Prev 2019: 28: 1052–58. [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.