Prognostic Significance of p53 and X-ray Repair Cross-complementing Group 1 Polymorphisms on Prostate-Specific Antigen Recurrence in Prostate Cancer Post–Radical Prostatectomy

Radical prostatectomy is one of the principal treatment modalities for men with clinically localized prostate cancer. Although approximately two-thirds of patients treated surgically will remain disease-free >10 years after radical prostatectomy, a substantial portion of these patients experience early prostate-specific antigen (PSA) recurrence (1, 2) and are prone to develop metastatic lesions accompanied by significant mortality (3–5). An accurate prediction for these high-risk patients is of paramount importance because they may benefit from more aggressive adjuvant therapy. Several clinicopathologic characteristics have been suggested as predictors for post–radical prostatectomy PSA recurrence (1, 6–10), but information on other genetic factors, including genetic polymorphisms, are still lacking.

The tumor suppressor gene p53, located on chromosome 17p13, is one of the most commonly mutated genes in all types of human cancers (11, 12). The p53 protein exhibits a common polymorphism at amino acid 72, resulting in either a proline residue (CCC, Pro allele) or an arginine residue (CGC, Arg allele; ref. 13). It has been reported that there is a functional difference between the Arg and Pro allele in inducing apoptosis and suppressing transformation (14). This p53 codon 72 polymorphism has been associated with susceptibility to several human cancers but has been found to yield conflicting results in prostate cancer (15–17). Moreover, its prognostic role on prostate cancer progression remains undetermined.

The XRCC1 (X-ray repair cross-complementing group 1) protein is involved in the repair of DNA base damage and single-strand DNA breaks by binding DNA ligase III at its carboxyl and DNA polymerase and poly(ADP-ribose) polymerase at the site of the damaged DNA (18). Three common polymorphisms in coding regions of the XRCC1 gene at codons 194 (Arg to Trp), 280 (Arg to His), and 399 (Arg to Gln) have been recently identified (19). These polymorphisms, involving an amino acid change at evolutionarily conserved regions, could alter the XRCC1 function. Previous studies have reported that the XRCC1 399 Gln allele is significantly associated with a higher level of DNA adducts and glycophorin A mutations in erythrocytes (19, 20), increased sister chromatid exchange frequencies (21, 22), and higher sensitivity to ionizing radiation (22). Thus, these may alter cancer susceptibility and disease progression. However, few studies have examined the influence of codon 194 and codon 280 variants on the function of XRCC1.

In clinical studies, the XRCC1 codons 194, 280, and 399 polymorphisms have been associated with risk for several cancers, such as cancers of breast, esophagus, bladder, colon, stomach, and prostate (23–28). The XRCC1 codon 399 polymorphism has also been reported to be a prognostic predictor of colorectal, bladder, and gastric cancers after chemotherapy (29–31). However, the prognostic role of this polymorphism on prostate cancer is lacking. Thus, in this study, we assess the potential prognostic roles of p53 codon 72 and XRCC1 codons 194, 280, and 399 polymorphisms on the recurrence of PSA in clinically localized prostate cancer after radical prostatectomy.

The study subjects were based on our previous cohort (16, 32–34). Briefly, patients with newly diagnosed and pathologically confirmed prostate cancer were recruited from Kaohsiung Medical University Hospital and Kaohsiung Veterans General Hospital, both located in southern Taiwan. From December 2000 to August 2005, we identified 462 prostate cancer patients, 429 (92.9%) of whom agreed to participate in this study. This protocol was approved by the Institutional Review Board of Kaohsiung Medical University Hospital, and informed consent was obtained from each participant.

Disease stage was determined by pathology findings, pelvic computed tomography or magnetic resonance image, and radionucleotide bone scans, according to criteria established by the American Joint Committee on Cancer tumor-node-metastasis classification system (American Joint Committee on Cancer Staging Manual, 5th edition, 1997). Pathologic grading was recorded as the Gleason score (35) and was further classified into two groups, with Gleason scores of 2 to 7 and 8 to 10, to assess the prognostic value after radical prostatectomy (1, 3).

Among the 429 prostate cancer patients, a subset of clinically localized prostate cancer patients who underwent radical prostatectomy (n = 131) were followed-up prospectively to investigate the potential role of p53 codon 72 and XRCC1 codon 399 polymorphisms in the progression of prostate cancer (defined by the recurrence of PSA). In the 131 eligible cases, 126 were successfully genotyped and included in the final analysis.

Pathology analyses were done on the whole specimen with step sections (2-3 mm). Extracapsular extension was defined as neoplastic cells in contact with periprostatic fat. Positive surgical margin was defined as tumor cells present at the inked margin. Other important pathologic variables, including vascular invasion, perineural invasion, tumor multifocality, and presence of high-grade prostatic intraepithelial neoplasia, were also recorded (Fig. 1; refs. 6, 36–39). These patients were instructed to have postoperative PSA follow-ups every 3 months.

p53 codon 72 and XRCC1 codons 194, 280, and 399 polymorphisms. p53 codon 72 and XRCC1 codons 194, 280, and 399 polymorphisms were determined using PCR-RFLP, as described in our previous paper (16, 40). Briefly, the two primers were 5′-TTG CCG TCC CAA GCA ATG GAT GA-3′ (forward) and 5′-TCT GGG AAG GGA CAG AAG ATG AC-3′ (reverse) for p53 codon 72, 5′-GTT CCG TGT GAA GGA GGA GGA (forward) and 5′-CGA GTC TAG GTC TCA ACC CTA CTC ACT (reverse) for codon 194, and 5′-TTG ACC CCC AGT GGT GCT AA (forward) and 5′-GGC TGG GAC CAC CTG TGT T (reverse) for codons 280 and 399. For p53 condon 72, the PCR product was digested using 5 units of BstUI (New England Biolabs). When BstUI restriction site was present (Arg allele), the 199-bp fragment was digested into two 113-bp and 86-bp fragments. The Pro allele was not cleaved by BstUI and had a single band of 199 bp. The heterozygous genotype (Arg/Pro) had three bands (199, 113, and 86 bp). For XRCC1, the PCR products were digested with PuvII for codon 194, RsaI for codon 280, and MspI for codon 399. After digestion, the Arg allele of codon 194 showed a segment of 138 bp, whereas the Trp allele showed the products of 63 and 75 bp. For codon 280, the Arg allele showed the products of 63, 201, and 597 bp, whereas the His allele showed the product of 201 and 660 bp. For codon 399, the Arg allele showed 115, 285, and 461 bp, the Gln allele showed 285 and 576 bp, and the Arg/Gln heterozygous were 115, 285, 461, and 576 bp (40).

Statistical analysis. Association between age and clinicopathologic factors with p53 codon 72 and XRCC1 codons 194, 280, and 399 polymorphisms were assessed by Student's t test statistic, χ² test, or Fischer's exact test. PSA recurrence was defined as two consecutive PSA measurements of >0.2 ng/mL at an interval of >3 months (41). The PSA level of >0.2 ng/mL in the first time of follow-up was considered the date of recurrence. The significance of p53 codon 72 and XRCC1 codons 194, 280, and 399 genotypes as predictors for PSA recurrence-free survival after radical prostatectomy was determined using the Kaplan-Meier analysis, and the difference in survival curves was examined by means of log-rank test. The Cox proportional hazard regression model was used to determine if the p53 codon 72 and these XRCC1 genotypes could be defined as independent predictors of the recurrence of PSA in the presence of the covariates of age and other pathologic and clinical markers. Age was treated as continuous variable, and other clinicopathologic factors and genotypes were treated as categorical variables. The Statistical Package of the Social Sciences software version 11.5 (SPSS, Inc.) was used for statistical analyses. A two-sided P value of <0.05 was considered statistically significant.

In total, 126 clinical localized prostate cancer patients who underwent radical prostatectomy were included in the final analysis. Among them, 29 (23%) experienced recurrence of PSA during the 26.7 months (mean) and 24.0 months (median) follow-up periods. For the association with clinicopathologic features with the p53 genotypes (Table 1), there were no significant differences between the age at diagnosis, preoperative PSA level, pathology stages, vascular invasion, perineural invasion, tumor multifocality, high-grade prostatic intraepithelial neoplasia, Gleason scores, extracapsular extension, and positive surgical margin. Regarding PSA recurrence, 10 of 33 p53 codon 72 Arg/Arg genotype cases experienced PSA recurrence compared with 19 of 93 in Arg/Pro + Pro/Pro genotype cases (30.3% versus 20.4%), although the difference was not significant (P = 0.247). The mean PSA-free survival time (months) between Arg/Arg and Arg/Pro + Pro/Pro genotypes was similar (24.3 ± 16.8 versus 27.6 ± 16.4). The median PSA-free survival time (months) was not defined since >65% cases in both genotypes remained PSA recurrence–free at the end of follow-up periods (Fig. 2).

For the association between the clinicopathologic features with XRCC1 codons 194, 280, and 399 polymorphisms, there was no significant association with the XRCC1 codons 194, 280 genotypes and preoperative PSA level, pathology stages, vascular invasion, perineural invasion, high-grade prostatic intraepithelial neoplasia, Gleason scores, extracapsular extension, and positive surgical margin. Interestingly, the codon 399 Arg/Gln and Gln/Gln genotypes were significantly associated with a high percentage of tumor multifocality (73.9% versus 52.3%, P = 0.021; Table 2). Furthermore, 18 of the 53 XRCC1 codon 399 Arg/Gln and Gln/Gln genotype cases experienced recurrence of PSA, which was significantly higher than the 11 of 73 in Arg/Arg cases (34.0% versus 15.1%, P = 0.013). The mean PSA-free survival time (months) was also shorter in the Arg/Gln + Gln/Gln genotypes (21.7 ± 14.4 versus 30.4 ± 17.1, P = 0.003). The median PSA-free survival time for Arg/Gln + Gln/Gln genotypes was 47 months; however, 82.5% of Arg/Arg genotype cases remained PSA recurrence–free at the end of follow-up periods (Fig. 2). No such significant associations with PSA recurrence and PSA-free survival time were observed for the codon 194 and codon 280 genotypes (data not shown).

Using the Kaplan-Meier survival curves, the p53 codon 72 Arg/Arg genotype showed a slightly worse rate of PSA recurrence compared with Arg/Pro + Pro/Pro genotypes, although the difference did not reach statistical significance (P = 0.1840). Of the XRCC1 polymorphisms, only the XRCC1 codon 399 Arg/Gln and Gln/Gln genotypes showed a significant poorer PSA-free survival after radical prostatectomy compared with the Arg/Arg genotype (Arg/Gln + Gln/Gln versus Arg/Arg; log-rank test, P = 0.0056; Fig. 2).

In the multivariate Cox proportional hazard model, the Gleason score (8-10) and XRCC1 codon 399 Arg/Gln and Gln/Gln genotypes were significant predictors of PSA recurrence after radical prostatectomy (Table 3). After adjusting for other covariates of age and clinical and pathologic factors, the XRCC1 codon 399 Arg/Gln and Gln/Gln genotypes remained independent risk factors for PSA recurrence after radical prostatectomy (hazard ratio, 4.73; 95% confidence interval, 1.61-13.92; P = 0.005). In this study, we did not find any significant combined effect of XRCC1 codons 194, 280, and 399 on RSA recurrence risk (data not shown).

In the present study, the p53 codon 72 polymorphism was not associated with post–radical prostatectomy recurrence of PSA. Among these three XRCC1 polymorphisms (codons 194, 280, and 399), the codon 399 Arg/Gln and Gln/Gln genotypes were associated with a poorer recurrence risk, and these variant XRCC1 genotypes remained independent predictive factors after adjusting for other covariates. To our knowledge, our series is the first to show the significant prognostic role of the XRCC1 codon 399 polymorphism on the PSA recurrence after radical prostatectomy.

About 15% to 46% clinically localized prostate cancer patients have experienced disease recurrence (presenting as PSA recurrence) after radical curative prostatectomy (7–9). New biological markers are needed to more accurately predict the risk of relapse. Such markers could allow a more accurate prediction of outcome for clinicians and patients and enhance the selection of high-risk patients who may benefit from more aggressive adjuvant therapy and more intensive follow-up. Compared with other pathologic features that need postoperative evaluation, DNA-based genetic markers have advantages, such as preoperative availability, easy conductance, and objective interpretation without individual bias. Therefore, this approach is shown to be feasible and should be advocated.

Although the p53 codon 72 polymorphism has been shown to be prognostically significant for gastric (42), breast (43), and lung cancer (44), few studies have investigated its prognostic role on prostate cancer progression. Suzuki et al. reported that there tended to be a greater number of patients with low-grade cancers among those with the Pro/Pro genotype than among those with other genotypes (odds ratio, 0.41; 95% confidence interval, 0.13-1.30; P = 0.13; ref. 17), but their observations were not confirmed in our previous study (16). This present study did not find an association between p53 codon 72 polymorphism and the clinicopathologic features or recurrence of PSA for clinical localized prostate cancer after radical prostatectomy. Thus, p53 codon 72 polymorphism may not be prognostically relevant to the recurrence of post–radical prostatectomy PSA.

DNA repair plays a key role in carcinogenesis through the removal and repair of DNA damage induced by endogenous and environmental sources. Base-excision repair is an important DNA repair pathway responsible for repair of base damage from X-rays, oxygen radicals, and alkylating agents (45, 46). The XRCC1 acts as a central scaffolding protein by binding DNA ligase III, DNA polymerase β, and poly(ADP-ribose) polymerase in base-excision repair (18, 47). In a study of Chinese hamster ovary cell lines with mutations in the XRCC1, there has been found a reduced ability to repair single-strand breaks in DNA and concomitant cellular hypersensitivity to ionizing radiation and alkylating agents (48). These findings suggest that XRCC1 plays an essential role in the removal of endogenous and exogenous DNA damage. Among the genetic polymorphisms in the XRCC1, the codon 399 Gln allele has been significantly associated with a higher level of DNA adducts and glycophorin A mutations in erythrocytes, increased sister chromatid exchange frequencies, and higher sensitivity to ionizing radiation (19, 20). These functional analyses and numerous molecular epidemiologic studies suggest that the Gln variant is a risk allele (23, 25, 28). With regard to the XRCC1 codons 194 and 280 polymorphisms, few studies have examined their influence on the function of XRCC1, and only few studies have reported on the association of XRCC1 codons 194 and 280 polymorphisms with prostate cancer risk although the result have been inconsistent (26, 28). Our study could not find an association between the XRCC1 codons 194 and 280 on post–radical prostatectomy recurrence of PSA.

However, our study found the XRCC1 codon 399 Arg/Gln and Gln/Gln genotypes to be independent predictive factors for post–radical prostatectomy PSA recurrence after adjusting for other covariates. These findings suggest that the XRCC1 codon 399 Gln allele may also be a putative risk allele for prostate cancer progression. Interestingly, in our study, the codon 399 Arg/Gln and Gln/Gln genotypes were associated with increased risk of tumor multifocality (73.9% versus 52.3%, P = 0.021; Table 2), which may partially reflect decreased DNA repair capacity. However, how exactly the XRCC1 codon 399 polymorphism influences prostate cancer progression requires more detailed in vitro and in vivo studies.

This study has several potential limitations. First, our case number was relatively small and the follow-up period was short. Second, we did not investigate other genetic polymorphisms in p53 and XRCC1 because of low allele frequencies and/or limited information on their functional significance. Third, for combined genotypes effect, our series did not have enough power for reliable haplotype analysis. Thus, our results warrant a larger scale study with a longer follow-up period, as well as more comprehensive analyses of other genetic variants in the base-excision repair pathways for further validation.

Our preliminary data suggests that the XRCC1 codon 399 polymorphism may be a prognostic factor for PSA recurrence after radical prostatectomy. Further large-scale studies are needed to confirm our findings and clarify the important role of XRCC1 polymorphisms on prostate cancer progression.