|Home | About | Journals | Submit | Contact Us | Français|
Susceptibility to cancer ensues in individuals carrying malfunctioning DNA repair mechanisms. The impact of Single Nucleotide Polymorphisms (SNPs) in key DNA repair mechanisms on risk for prostate cancer was investigated in this case-control study. Samples consisted of 110 patients with confirmed prostate cancer and 200 unaffected men, from Rio de Janeiro, Brazil. XPD/Lys751Gln (rs13181), APEX1/Asp148Glu (rs1130409), and RAD51/G135C (rs1801320) SNPs were analyzed by PCR-RFLP. Allelic and genotypic frequencies were calculated and compared by Chi-Square test. The association between SNPs and clinical/epidemiological data was considered significant by Odds Ratio analysis, with IC95% and a p-value≤0.05. Only the XPD/Lys751Gln SNP significantly increased susceptibility to disease in southeastern Brazilian men, with p≤0.001 [OR=2.36 (1.46-3.84)], with no association with APEX1 or RAD51 SNPs. Combined XPD+RAD51 SNPs were highly associated with the disease, p≤0.005 [OR=3.40 (1.32-9.20)]. A Chi-Square significant association between XPD/Lys751Gln and Gleason score was also observed (OR=9.31; IC95%=1.19–428.0; p=0.022). Epidemiological inquiries revealed that exposure to pesticides significantly impacted the risk for prostate cancer in this population. DNA repair dysfunctions seem to prevail among workers exposed to chemical byproducts to cancer in this specific tissue. Non-invasive genotyping SNPs may help assessment of prostate cancer risk in environmentally exposed populations.
An increasing incidence of prostate cancer (PCa) has been observed in Brazil, associated with longer lifetimes and due to improved diagnosis methods in countrywide information databases. The Brazilian National Cancer Institute (INCA) estimates approximately 61,200 new cases in 2016, with a risk of 61.82 cases per 100,000 men (INCA, 2016). Factors such as hypertension, smoking, alcohol consumption, sedentarism/obesity, hypercholesterolemia, and genetic history were identified as top risk factors for PCa (Bostwick et al., 2004).
Oxidative stress seems to influence the prostatic carcinogenic process because of its expressive association with aging and accumulating damage (Platz and Giovannucci, 2006). Likewise, environmentally-borne carcinogenic agents can also damage DNA, and different DNA repair mechanisms have been implicated in alleviating such harmful damages (Pramanik et al., 2011). Single Nucleotide Polymorphisms (SNPs) result from single-nucleotide changes in these genes, influencing expression or function of the affected genes. These changes could be critical concerning DNA repair functions.
The XPD (Xeroderma pigmentosum complementation group D) gene encodes one of the proteins involved in the Nucleotide Excision Repair (NER) pathway (O’Donovan and Wood, 1993). One of XPD’s important SNPs is XPD/Lys751Gln (rs13181), in which an A → C base substitution at exon 23 causes a Lys → Gln substitution in codon 751, compromising part of the C-terminal domain of the XPD protein and its full DNA repair capacity (White, 2009).
The APE1 endonuclease removes abasic sites from DNA, and is the top enzyme initiating recognition of damaged sites in human Base Excision Repair (BER), along with its 3′ → 5′ exonuclease, DNA 3′-repair diesterase and DNA 3′-phosphatase activities. APE1 plays a major 3′-phosphodiesterase role in initiating the repair of oxidative-generated single-strand breaks (Ramana et al., 1998). A missense G → T substitution SNP in the APEX1 gene in exon 5 (rs1130409) changes a critical Asp → Glu residue at codon 148. An up to date report connects this SNP to the development of PCa in Asian descendents (Chen et al., 2016).
The human RAD51 gene codes for a bacterial RecA recombinase homolog, central to the Homologous Recombination Repair (HRR) mechanism (Galkin et al., 2006). The RAD51 protein is critical for maintaining genomic integrity by repairing DNA double-strand breaks. The RAD51 polymorphism (rs1801320) implies in a G → C substitution mapping upstream the 5′ untranslated promoter region at -135 bp from the transcription start. Although the functional consequences of this SNP remain to be clarified, a single nucleotide change in this CpG promoter island may up-regulate gene expression, thus affecting RAD51 mRNA levels (Antoniou et al., 2007). This SNP has been linked to breast and ovarian cancer susceptibility due to the supposed interaction of RAD51 with BRCA1 and BRCA2 proteins, with an especially higher risk for carriers of BRCA2 mutations (Levy-Lahad et al., 2001).
The investigation on this assembly of SNPs was designed because they significantly impact main DNA repair mechanisms (NER, BER and HRR, respectively) and genetic instability and, under this perspective, had never been addressed in a Brazilian population before. Considering the importance of DNA repair failures in cancer development, the aim of this work was to investigate the contribution of SNPs in key genes XPD/Lys751Gln, APEX1/Asp148Glu and RAD51/G-135C in the susceptibility to PCa in a case-control study in Rio de Janeiro, Brazil.
For this study, 110 patients (62.0 ± 6.5 mean age at diagnostic) who had undergone radical prostatectomy and subsequent chemo- and/or radiotherapy were recruited between 2006-2008 from the National Cancer Institute (INCA) and Mario Kroeff Hospital (Rio de Janeiro, Brazil). The control group consisted of 200 men (61.6 ± 10.3 mean age) undergoing routine tests, without suspicion of PCa, at the Antonio Pedro University Hospital (APUH, Niterói, RJ, Brazil), sampled between 2009-2010. The study was approved by the Ethics Committees of INCA (#91/05) and APUH (#48/09). All participants signed a written informed consent. The recruited patients carried T1, T2, or T3 prostate tumors, according to the TNM scoring system (tumor extension and metastasis).
Individuals included in the study met the clinical criteria of having a palpable mass by digital rectal examination (DRE) and/or elevated PSA serum levels (C > 2.5 ng/mL) indicative of the occurrence of a tumor, followed by histopathological confirmation of PCa. All patients were treated with radical prostatectomy. Gleason scoring was obtained by a pathologist’s biopsy evaluation and PSA levels by clinical blood tests.
Patients and controls were interviewed following a structured individual questionnaire covering educational level, familial cancer and medical history, place and date of birth, place of birth of father and mother, weight, height, and place of residence. We also profiled the occurrence of prostate and other cancers in first-degree relatives, smoking status, and alcohol consumption. Patients and controls were asked to self-declare skin color, which is a normal question in such inquiries in Brazil. Controls and patients were excluded if they showed genetic syndromes and past cancer history.
Genomic DNA was isolated from 4-5 mL peripheral blood by proteinase K digestion and phenol-chloroform extraction (Sambrook, 1989). The selected SNPs were genotyped by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP). The XPD/Lys751Gln (rs13181), RAD51/G-135C (rs1801320), and APEX1 Asp148Glu (rs1130409) SNPs were determined following conditions described by Baccarelli et al. (2004), Wang et al. (2001), and Hu et al. (2001), respectively.
XPD and APEX1 amplicons were digested by 10 U of the specific restriction endonucleases (37 °C, 2 h). The XPD/Lys751Gln genotypes A/A, A/C, or C/C were profiled as PstI digestion products of 100/224, 66/100/158/224, or 66/100/158 bp bands, respectively. The APEX1/Asp148Glu genotypes T/T, T/G, or G/G were profiled as BfaI digestion products of 164, 164/144/20, or 144/20 bp bands, respectively. For RAD51/G-135C amplicons, digestions were performed with 10U of BstNI (60ºC, 1 h). The genotypes G/G, G/C, or C/C were profiled as digestion products of 71/87, 71/87/157, and 157 bp bands, respectively. All samples were resolved on 3% agarose gels and visualized by ethidium bromide staining.
Genotypic frequencies between patient/control groups were analyzed by Chi-Square Test and Odds Ratios with CI=95%. Contingency tables were used to compare groups concerning age, socio-economical and life style aspects by Chi-Square Test. Data were analyzed with Statistical Package 17.0 for Social Sciences (SPSS). Levels of significance p≤0.05, corresponding to 95.0% confidence, were considered in all analyses. All polymorphisms were regarded as potential markers by using the collective criteria of sensibility, specificity, positive and negative predictive values, and test accuracy.
Regarding age, there was no significant difference between patients (62.0±6.5) and control groups (61.6±10.3). Age-pairing between groups was essential concerning duration of exposure to environmental carcinogens. Case and control populations were also profiled similarly for basic items listed in Table S1 (geographic origin, education level, ethnic profile, familial cancer history, chronic diseases). When exposure to risk factors was weighted (tobacco, alcohol, or drug consumption), no significant differences were found between groups, with smokers and drinkers prevailing in both. Exposure to environmentally-borne agents (pesticides) was the only remarkable difference between case and control populations, and that caused a significant difference to appear (p<0.0001) towards the patient’s group. While 21.1% of patients declared having manipulated pesticides, only 3.5% of controls were subjected to this occupational hazard (Table 1).
The sampled populations were native to southeastern Brazil, mainly from the Rio de Janeiro metropolitan region (80%). The majority self-declared as ethnically white (53.8%), followed by brown or mulatto (30.7%) and black (15.5%). PSA values were obtained for 79 patients, ranging from 0.78-38.0 ng/mL (median 7.2 ± 5.4 ng/mL) (Table 2). Additional comparisons of SNP with age, prognosis, PSA level, and Gleason score unveiled two extra implications: (a) a Chi-Square significant association between XPD SNP and Gleason score (OR=9.31; 95%CI=1.19–428.0; p=0.022); (b) the RAD51 SNP related with unclear prognosis (p=0.031).
Genotypes of XPD, RAD51, and APEX1 were determined for 110 patients and 200 controls. The most common allele was considered as the reference one, whereas the less common ones were grouped as variants (risk polymorphism) (Table 3). Neither RAD51 (OR=1.13; 95%CI=0.64–1.96) nor APEX1 polymorphic genes (OR=0.88; 95%CI=0.55–1.42) were associated with risk for PCa. Only for XPD (OR=2.36; 95%CI=1.46–3.84) a positive association was found for variant C/C or A/C genotypes. Wild-type RAD51 patients were significantly ascribed as the poorer prognosis (Table 2).
Among the SNPs, the combined polymorphisms XPD/RAD51 were found to have a harmful association (OR=3.40; 95% CI=1.32–9.20; p≤0.05). Other combinations such as RAD51/APEX1 SNPs (OR=0.86; 95%CI=0.28–2.60), XPD/APEX1 (OR=2.30; 95%CI=0.76–8.08), or XPD/RAD51/APEX1 (OR=1.43; 95%CI=0.34–5.07) resulted in non-significant risks for PCa (Table S2).
In a comparative analysis with the International HapMap database (www.hapmap.org), only the XPD polymorphic allele and genotypic frequencies were statistically different in relation to Chinese (CHB) and Japanese (JPT) populations, calculated by the Chi-Square test (Table S3).The variant allele frequency ranged from 0.076 for Japanese (JPT) to 0.358 for Native American Indians (GIH).
In order to estimate whether these polymorphic genes could fulfill the role of diagnostic/prognostic markers, we performed a data performance diagnostic test for available genotypes. As shown in Table 4, the combination XPD/RAD51 SNPs scored better for sensitivity, specificity, and accuracy.
The association between three dysfunctional SNPs (Asp148Glu/APEX1, G-135C/RAD51 and Lys751Gln/XPD) and PCa susceptibility was evaluated in this case-control study. This is the first Brazilian study to cover SNPs in genes implicated in three main DNA repair pathways in a PCa population in Brazil. The present study was, thus, designed to improve the genetic profiling of PCa patients by mapping a panel of DNA repair SNPs and assess its correlation with clinical parameters. A comparative epidemiological inquiry on known cancer risk factors was applied to cases and controls. The collected data revealed that exposure to environmental factors, as well as the XPD/Lys751Gln (rs13181) polymorphism appeared to be associated with increased risk for PCa.
Major areas of research on PCa have focused on three main points: (1) assessment of how life style/environmental factors/diet can influence carcinogenesis; (2) strategies to delay disease onset and progression; and (3) finding accurate biomarkers to distinguish between indolent and aggressive forms (Shen and Abate-Shen, 2010). The present study was designed to target items (1) and (3).
A significantly increased risk for lymphohematopoietic, prostate, melanoma, and brain neoplasms has been surveyed among pesticide handlers and industry workers (Koutros et al., 2010). Moreover, the NER pathway was recently shown to repair damage induced by carcinogens, including pesticides (Barry et al., 2012). Accordingly, data from the present study associated an NER dysfunction (XPD) with high risk for PCa (Table 3), especially in men exposed to pesticides. When risk factors were compared between cases and controls, a remarkable positive correlation was found only for exposure to pesticides. The expressive amount of XPD AC and CC polymorphic patients exposed to pesticides implies that they can steeply accumulate damages because of NER malfunction, leading to increased PCa risk.
Extensive DNA damage was detected in exposed floriculturists compared to non-exposed ones, and exposure to pesticides appeared as the primary genotoxic factor for long-term carcinogenesis (Bolognesi et al., 2004). Theophilou et al. (2015) followed-up transgenerational alterations in human prostate tissue samples over 30 years, correlated with geographic distribution and exposure to risk factors. Indeed, genotoxic agents seem to ensue long-lasting genetic/epigenetic alterations, imprinting phenotypic pro-malignant changes.
Although the association of lifestyle and cancer risk had already been highlighted in a number of studies, none of the other noxious habits (smoking, drug use), or social aspects could be ascribed to PCa risk in the studied group. Here, only the XPD SNP was related to the risk of developing PCa. It is worthy of note that XPD-mediated NER corrects damages caused by environmental agents and reactive oxygen species (Mitra et al., 2001), these being important triggers of PCa. This supports the complex genetic basis of PCa involving multiple susceptibility genes.
Sulfolobus acidocaldarius XPD (SaXPD protein) harbors a catalytic site for its SaRAD51 protein, and all SaXPD helicase domains were conserved in human XPD (Fan and Wilson, 2005). Four domains comprise the SaXPD’s catalytic core: two RAD51/RecA-like domains (HD1-HD2) and two additional HD1 domains inserted together. Twenty-two out of the known 26 point mutations related to XPD-linked human diseases mapped there (Xeroderma pigmentosum, Cockayne syndrome, and Trichothiodystrophy). Aloyz et al. (2002) showed that the physical association between human XPD and RAD51 proteins acting to remove DNA crosslinks maybe assembled in a larger complex. They appear to be recombination counterparts, modulating the HRR-mediated resistance to crosslinks.
By playing an important role in recovering damaged DNA in vertebrates, the RAD51 hyper-recombination phenotype can potentiate genomic instability (Klein, 2008). Schild and Wiese (2010) suggest that RAD51 overexpression can contribute to carcinogenic progression. Since XPD and RAD51 are counterparts in the repair of DNA breaks, their double malfunctions appeared statistically related to high risk for PCa, perhaps because of the triggered genomic instability. Noteworthy, wild-type RAD51 patients underwent worse prognoses, probably because of better repair of therapy-induced DNA damages by HRR, while XPD/Lys751Gln patients displayed higher Gleason scores (Table 2). Despite the small number of cases analyzed, these correlations were highly significant.
Regarding general clinical findings, both specificity and sensitivity tests also revealed that XPD and XPD+RAD51 genotypes scored good negative predictive values (Table 4). Regarding the potential validity of our proposed panel, since XPD and RAD51 SNPs significantly scored the top OR in a small population, they would probably reach a much more relevant status as risk markers when assessing a larger group.
A Spanish multicenter group revealed that genes involved in HRR could be associated with poor prognosis in PCa (Henriquez-Hernandez et al., 2014). Their findings corroborate ours, regarding the association of HRR SNPs with unclear prognosis. The same group reported that DNA repair genetic variants (XRCC6 and MVP) were associated with more aggressive outcomes (Henriquez-Hernandez et al., 2016).
Our genetic analysis is endorsed by the HapMap database, i.e., the Southeast Brazilian population is quite different from the Japanese and Chinese. The European colonization remains strongly imprinted in Brazilian ethnicity, remaining genetically distant from Eastern populations. Accordingly, the genotypic profiling seen by Henriquez-Hernandez et al. (2014) in a Spanish population and a meta-analysis performed by Mi et al. (2012) with Asian and African populations matches our HapMap assessments. Studies about ethnically-linked differences strengthen the need to continue research on new confident biomarkers.
Interestingly, the APEX1 deficiency caused the risk for PCa to drop (Table S2), and this result is interpreted as deriving from the simultaneous knockdown of three key routes of DNA repair: HRR, NER, and BER. This probably drives prostatic cells to death, protecting the prostatic tissue from the outbreak of potentially malignant clones.
The APEX1/Asp148Glu (rs1130409) SNP was recorded in another Brazilian population (Kuasne et al., 2011), which appeared to be more susceptible to PCa. In our group, however, no susceptibility was connected with this SNP, maybe because of other genetic background differences between the sampled populations and the low penetrance of this SNP for cancer development.
Data compiled for this Brazilian group revealed the need for a clinical follow-up, along with a panel of key DNA repair dysfunctions, instead of just one or two genes. Moreover, biomarker research has mainly targeted diagnosis, rather than prognosis and outcomes (Prensner et al., 2012), which are capable of tailoring individual therapeutics. For instance, Chantre-Justino et al. (2015) reported that patients with lower transcript levels of the APE1 gene belonged to the cohort with 100% lethality from aggressive bladder cancer. In prostate cancer, the expression of a panel of 17 genes belonging to nine different DNA damage and repair (DDR) pathways was determined in prostatectomy samples in a cohort study by Evans et al. (2016), where high expression of NER and BER genes significantly related to metastasis and lower overall survival. The authors envisaged the screening of individual “DDR signatures” to track aggressive prostate cancers on the basis of their differential response to therapy. These two studies on genetic markers in tumors of two urinary-related organs point out how tissue-specific the neoplastic process can be driven.
In this study, we observed that patients with a polymorphic XPD genotype significantly ranked at higher Gleason scores, indicating that NER-deficient individuals seem to bear genetically more instable tumors. In the absence of the error-free NER pathway, other concurrent error-prone mechanisms could take control of less accurately repair of DNA damages. In support of this finding, XPD SNPs rs13181 and rs1799793 (XPD/Asp312Asn) tended to group under high Gleason scores and advanced tumor stages, although results were not statistically significant (Agalliu et al., 2010). Altogether, these findings help to unveil genetic features on how this disease is triggered and progresses in different populations, whose heterogeneity can render even more complex outcomes due to gene-environment interactions. Present findings await corroboration by other studies, focusing on environmental and ethnic factors impacting this important male cancer.
Complimentary tests, such as ultrasonography and biopsy should be included in more precise diagnostics, although the lower cutoff of 4.1 ng/ml of PSA protein, for example, is consensually taken as a good negative prediction (Van den Broeck et al., 2015). If early cancer detection by current tests is the desired target in cancer prevention, genotyping instability-prone DNA repair SNPs can be envisaged as the earliest informative test amongst all.
Financial support was provided by CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) and a fellowship to CL by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). We also wish to acknowledge the assistance of Dr. Ana Hatagima (FioCruz, Brazil) during the initial development of this study.
The following online material is available for this article:
Associate Editor: Mara H. Hutz