Association of TP53 Single Nucleotide Polymorphisms with Prostate Cancer in a Racially Diverse Cohort of Men

1. Introduction

Prostate cancer (CaP) is the most common non-skin cancer and is the second leading cause of cancer death in American men [1], with an unequal burden of disease incidence and mortality among men of different ancestries [2]. Prostate cancer is a clinically heterogenous disease comprising of diverse clinico-pathologic and progression features and is characterized by a large subset of the indolent cancer type. Therefore, it is critical to identify molecular and genetic markers with diagnostic, predictive and prognostic potential in addition to the standard of care (SOC) variables [3]. A significant proportion of CaP susceptibility has been attributed to inherited predisposition. Genome wide association studies (GWAS) have identified 269 low penetrance single nucleotide polymorphisms (SNPs) associating with CaP risk or aggressiveness [4]. The TP53 gene, also referred to as the guardian of the genome, is a key tumor suppressor gene involved in DNA repair, cell cycle arrest, apoptosis, cell metabolism and genomic stability [5,6]. The TP53 is the most commonly mutated gene in human cancers and malfunction of the TP53 pathway is a key feature of human tumors [7]. Germline mutations in TP53 predisposes to Li-Fraumeni syndrome (LFS) which is characterized by early onset of cancers including breast cancer, sarcomas, brain tumors, leukemia and adrenocortical carcinomas [8,9,10]. However, the role of germline mutations in TP53 is not completely understood in CaP. Some studies point to the role of specific TP53 SNPs as a possible risk factor for CaP [11,12]. Polymorphisms in codons 47 and 72 in the TP53 tumor suppressor gene have been shown to affect phosphorylation and the expression of pro-apoptotic genes, respectively, both of which could increase cancer risk [13]. A more recent study found heterozygosity at codon 72 SNP (rs1042522, GC), and either of the two intron 4 SNPs (rs9895829, TC and rs2909430, AG) are associated with up to 5-fold greater CaP risk, greater tumor-promoting inflammation, and shorter patient survival [14].

Racial disparities in CaP incidence and mortality rates remain significant even after adjusting for socio-economic status and access to healthcare, suggestive of a greater contribution from genetic factors [15]. However, several Veterans Affairs studies revealed conflicting results in CaP outcomes for Black and White men within a single, equal access health care system [16,17]. Black men have an increased risk for CaP, especially clinically aggressive CaP, and elevated PSA levels at the time of diagnosis compared to White men in similar cohorts. Genetic associations between such clinical findings and SNPs have been described earlier, however with inconsistent results. Thus, further understanding and additional confirmation of such genetic factors may be beneficial in understanding the clinical outcomes which influence screening and treatment guidelines across AA and CA men in an equal access military healthcare system.

We hypothesize that the germline variants in TP53 could contribute to distinct clinical phenotype of CaP. Thus, the goals of this study were to identify the common SNPs in the TP53 gene in AA and CA men and to perform an association analysis for functional variants in TP53 with clinico-pathological features of CaP including biochemical recurrence-free survival and metastasis-free survival.

2. Materials and Methods

2.1. Study Population

This study is based on a retrospective cohort design utilizing germline blood DNA derived Oncoarray data from 321 men (AA = 216 and CA = 105) who were treated with radical prostatectomy at Walter Reed National Military Medical Center (WRNMMC). Only those patients were included who consented to enrollment in the CPDR biospecimen databank and multicenter national clinical database. The databases have been approved by the Institutional Review Boards (IRBs) at the WRNMMC and the Uniformed Services University of the Health Sciences (USUHS) in Bethesda, Maryland. All subjects provided written informed consent and agreed to participate for present study under IRB approved protocols. These patients are under the equal access DoD healthcare system and were clinically followed up to 25 years. Demographics, clinical characteristics, pathology information, and all treatment data were obtained from the CPDR multicenter national clinical database. Patient characteristics included patient age at CaP diagnosis (years), self-reported race (AA, CA), diagnostic PSA (ng/ml), pathologic T Stage (pT2,T3, T4) , biopsy and pathology Gleason (≤6, 3 + 4, 4 + 3, and 8–10), biochemical recurrence (BCR), time to BCR and metastasis outcome. A BCR event was defined as either two successive post-RP PSAs of ≥0.2 ng/mL, or initiation of salvage therapy after a rising PSA of ≥0.1 ng/mL. The clinical characteristics of the patients are shown in Table 1.

2.2. SNP Analysis and Validation

Genotype analysis for common SNPs with minor allele frequency (MAF) of at least 0.01, was performed using Eagle [18] for phasing and for imputation TOPMed (v R2 on GRCh38) was used as a reference comprising of 194,512 haplotypes [19]. Genotype data were split into 20MB regions with an overlap of 5MB. Experimental validation for the imputed SNPs (rs1800371/Pro47Ser and rs1042522/Arg72Pro) was performed using droplet digital polymerase chain (ddPCR) approach using a QX200 (BioRad). Briefly, a ddPCR mastermix was prepared containing 11 μl 2X ddPCR Supermix (BioRad), 1.1 μl 20X TaqMan SNP Genotyping Assay (BioRad, ThermoFisher Scientific; Supplementary Table 6), and 7.9 μl nuclease-free water (Qiagen) per sample. The mastermix was prepared at room temperature and 20 μl was added to 2 μl (=5 ng) of each DNA sample. Samples were loaded into individual wells of DG8TM cartridges (BioRad), and droplets were generated using a QX200 Droplet Generator (BioRad). For each sample, 40 μl of droplet mix was then transferred to a 96-well plate, and PCR was performed in a thermal cycler using the following cycling conditions: 95 °C × 10 min; 40 cycles of [94 °C × 30 s, 60 °C × 60 s]; 98 °C × 10 s; 40 C × 10 min. The BioRad QX200 Droplet Reader was then used to assess droplets as positive or negative based on fluorescence amplitude. The QuantaSoft software (BioRad) was used to analyze droplet data.

2.3. Statistical Analysis:

Descriptive distributions for clinico-pathological variables in the overall cohort stratified by exonic SNP genotypes was examined. The chi-square test or Fisher's exact test was used to compare categorical variables and the Mann-Whitney U test was used to compare continuous variables. Time to biochemical recurrence (BCR) or metastasis was evaluated using Cox proportional hazards (PH) modeling adjusting for age at diagnosis, race, and Gleason score to examine SNP effects. Unadjusted Kaplan-Meier survival analysis and log-rank testing were used to show the probability of BCR-free survival stratified by genotypes. All statistical analysis was performed using SAS version 9.4 (SAS Institute Inc., Cary, NC) and statistical significance was set at p < 0.05.

3. Results

Clinical and pathological features of the patient cohort is summarized in Table 1. The cohort had higher proportion of AA patients compared to CA patients (69% vs 31%), were over-represented by subjects with serum PSA in range 4-9ng/ml (63%), T2 clinical stage (74%) and pathological Gleason ≤ 7 (83%). The post-prostatectomy follow-up data showed that 15% of the subjects had biochemical recurrence, with 3% having metastasis.

The SNP analysis on the final cohort of 308 patients with available clinical data identified 74 SNPs in the TP53 gene region with MAF of at least 1%. The description of the 74 SNPs is included in Supplementary Table 1. Two imputed SNPs (rs1800371/Pro47Ser and rs1042522/Arg72Pro) of the 74 SNPs are non-synonymous, functional SNPs located in exon 4 of TP53 (Figure 1). MAF for Pro47Ser and Arg72Pro was 0.01 and 0.50 respectively. Both these imputed SNPs were experimentally validated by ddPCR approach (Figure 2). Concordance between TaqMan genotypes and imputed genotypes was 99%.

Race stratified analysis for the SNPs demonstrated and confirmed ancestry-dependent differences in allele frequencies of TP53 Arg72Pro and TP53 Pro47Ser SNPs in AA and CA CaP (Table 2; Supplementary Tables 2 and 3). The Pro47Ser variant has an MAF of 0.01 in AA, however, was not detected in CA men. While Arg72Pro was the most common SNP with an MAF of 0.50 (AA, 0.41; CA, 0.68) (Supplementary Table 4). The MAF distribution in our dataset was in concordance with publicly available databases Pro47Ser variant (A=0.005591/28 (1000Genomes), A=0.02439/8 (HapMap); Arg72Pro G=0.457069/2289 (1000Genomes), G=0.43865/143 (HapMap).

An association analysis of the TP53 Pro47Ser and Arg72Pro SNPs with respect to clinico-pathologic features of CaP is shown in Table 2. Except for race, there were no observed statistically significant associations between the Arg72Pro or Pro47Ser SNPs and age of onset of prostate cancer, CaP aggressiveness (Grade, Clinical Stage), tumor upgrading from biopsy to RP (Supplementary table 5) and disease outcome (BCR, Metastasis). However, Arg72Pro SNP was the only exonic SNP significantly (p= 0.046; HR = 1.52) associated with shorter time to BCR following radical prostatectomy after adjusting for patient age at diagnosis, race, and Gleason score (Table 3). Additionally, we found that 3 non-exonic SNPs were significantly associated with shorter time to BCR (rs4968186, intergenic; p 0.004; HR 1.75; rs9894227, intronic, p 0.007; HR 1.72 and rs8079544 , intronic, p 0.02; HR 1.63).

4. Discussion

The present study investigates the association of common functional germline variants in the TP53 gene with CaP in a racially diverse cohort of men with longitudinal clinical follow up. We confirmed ancestry-dependent differences for MAF across AA and CA men and found that Arg72Pro was associated with shorter time to BCR. Arg72Pro and Pro47Ser SNPs have been well characterized for their function and for association with cancer [12,20,21,22,23,24]; however, there is no consistent conclusion for CaP risk or clinico-pathological features [25,26,27,28,29]. This variability may be attributed to small sample size, study design and mixed ancestry.

Additionally, the prognostic role of the SNP in relation to disease and treatment outcome has been found in many cancers including CaP [30], lung cancer [31], gastric cancer [32], breast cancer [33], acute lymphoblastic leukemia. Common variant, Arg72Pro, leads to altered activities for TP53 protein. The C to G base change results in a proline (Pro) to arginine (Arg) amino acid alteration in the proline-rich domain impacting the programmed cell death or cell cycle arrest. Additionally, TP53 arginine has increased affinity for MDM-2, resulting in destabilization of the protein, lower levels of the TP53 protein, and ultimately earlier onset of TP53 associated tumors [14]. On the other hand, less frequent and AA centric Pro47Ser variant exhibits functional consequences by attenuating p53 transcriptional activation and pro-apoptotic functions [34]. The inactivation of p53 via somatic and germline TP53 mutations have been documented for both aggressive disease and poor outcomes in many cancer types [35]. The somatic mutations of TP53 have also been reported with variable frequencies in primary CaP and consistently with higher frequencies in metastatic disease [36]. Several studies from our center over the years including, a recent study show that focal alterations of TP53 protein in prostatectomy specimens associate with BCR and metastatic progression [37].

In a recent multi-institutional retrospective study by Maxwell et al. on CaP incidence in a cohort of patients with Li-Fraumeni syndrome (LFS) showed men with LFS had a 25-fold increased risk of CaP compared to the population controls and the rate of inherited deleterious TP53 variants was 9-times higher than men with no CaP [38]. Our current findings on the association of TP53 SNPs with aggressive disease further support the role of germline TP53 functional alterations in prostate tumorigenesis and suggests a potential clinical benefit to include evaluation of germline TP53 variants (SNPs identified in this study and pathogenic SNVs) for CaP susceptibility testing, especially in context of high risk men [39,40].

In addition to the genetic influence on CaP, studies described differences in such associations as they pertained to patient ancestry [41]. While associations between AA race, TP53 SNPs, and CaP incidence, mortality, and post-treatment BCR have been made in the past, the three primary guideline tools – National Comprehensive Cancer Network (NCCN), United States Preventive Services Task Force (USPSTF), and American Urological Association (AUA) Clinical Guidelines - used by physicians to guide patient screening and treatment - do not suggest its use in treatment decision-making recommendations, despite their clinical relevance [42,43,44].

The findings in our study not only further confirm previous research, but also improve the understanding of the association between TP53 SNPs, race/ancestry, and poor disease outcome. The present study provides important data that could influence medical providers to improve screening tools, germline CaP susceptibility testing for high-risk men, and treatment approaches for patients. Future directions include comprehensive analysis of germline variants in TP53 pathway in high-risk men, investigation of TP53 mutation profile to predict response to treatment, and functional understanding of the biology of these SNPs in relation to aggressive CaP.

5. Conclusions

Comprehensive analysis of common germline variants in TP53 confirms genetic heterogeneity across AA and CA men for Arg72Pro and Pro47Ser variants and suggests TP53 Arg72Pro predisposes men to clinically significant prostate cancer. Therefore, TP53 should be included in CaP germline screening panel for high-risk men representing different ancestries.