Abstract

Background. DNA repair systems are essential for each cell to repair and maintain the genome integrity. Base excision repair pathway is one of the crucial pathways to maintain genome integrity and PARP-1 plays a key role in BER pathway. The purpose of this study is to evaluate the association between polymorphisms in PARP-1 3′untranslated region (3′UTR) SNP rs8679 and its expression in colorectal cancer. Methods. Genotyping and gene expression were performed using TaqMan assays. The effects of age, gender, and tumor location were evaluated in cases and controls regarding the genotyping results. Resulting data was analyzed using SPSS software. Results and Conclusions. Genotyping analysis for SNP rs8679 showed decreased susceptibility to colorectal cancer at heterozygous TC allele and at minor allele C. Further this protective association was also observed in younger age patients (≤57), in female patients, and also in patients with tumors located at colon and rectum. PARP-1 expression levels are significantly different in colorectal cancer compared to matched normal tissue. Our findings proved that the upregulation of PARP-1 is associated with tumor progression and poor prognosis in Saudi patients with colorectal cancer, suggesting that PARP-1 can be novel and valuable signatures for predicting the clinical outcome of patients with colorectal cancer.

1. Introduction

Colorectal cancer is one of the major neoplasms that affect the human digestive system. The incidence of CRC has been increased worldwide; it ranks 3rd among males and the second among females [1, 2]. In Saudi Arabia it ranks first among men and second in women. The combinations of genetics, epigenetics, and environmental factors make CRC as a multifactorial disease. Maybe due to westernized life style, the incidence and mortality of CRC have been increased in Saudi Arabia in past decade which resulted in highest mortality rates [3, 4].

DNA repair systems are essential for each cell to repair and maintain the genome integrity. Errors or abnormalities in the repair systems may lead to mutations which in turn cause the progression of cancer [5]. One of the key DNA repair pathways that play a role in cell viability is base excision repair pathway (BER) [6]. BER is a pathway which repairs error in DNA during cell division via activation of PARP-1, which initiates the BER pathway cascade. PARP-1, poly(ADP-ribose) polymerase, is 113 KDa nuclear protein [7]. PARP-1 plays a key role in repairing SS and DS breaks of DNA [8]. Furthermore, PARP-1 expression is critical whether to influence or inhibit cancer progression [6]. PARP-1 overexpression may show both positive and negative effect cancer patients. In precancerous stage PARP-1 might act as shield to carcinogenesis [9]. In contrast to this when a tumor has highly expressing PARP-1, it might protect the tumor from the DNA damaging treatment effects. In such cases PARP-1 inhibitors play a key role in anticancer therapy [9, 10]. Several studies confirmed any variations or in absence of PARP-1 gene can lead to errors in DNA repair, genetic instability, and modulation of gene transcription; thus it can enhance tumor development [11]. Previous studies reported that defects in PARP-1 are enhancing cancer risk [12]. The PARP-1 gene polymorphisms have been previously reported to be associated with risk in various carcinomas, including colon [13], lung cancer [14], bladder [15], prostate [16], head and neck [17], and breast cancer [18, 19]. SNP sites in 3′UTR could affect gene expression by blocking miRNA binding the target sites. PARP-1 polymorphism rs8679 which is in 3′UTR region has previously studied in several cancers but there are conflicting reports about its role in cancer progression. In the present study we are intended to investigate the contribution of PARP-1 3′UTR SNP rs8679 (T3823C) genetic polymorphism and expression level in colorectal cancer among Saudi population.

2. Materials and Methods

2.1. Study Population

In this case-control study, a total number of 373 blood samples have been used. This includes 183 colorectal cancer and 190 matched normal samples. Sixty fresh colorectal cancer tissues and adjacent normal tissues from the same patient were collected in RNAlater solution for gene expression analysis (see Supplementary Table in Supplementary Material available online at http://dx.doi.org/10.1155/2016/8289293). Samples and clinical data (age, gender, family history, smoking habits, tumor stage, grade, and location) were collected from King Khalid University Hospital (KKUH) in Riyadh, Saudi Arabia. IRB was obtained from Ethics Committee of College of Medicine King Khalid University Hospital, King Saud University. Five mL of blood obtained from patients and normal persons in an EDTA solution-containing tube and stored at −80°C until further use.

2.2. DNA Extraction and Genotyping

DNA was extracted from 200 µL blood samples by using QIAmp DNA Blood Mini Kit (Qiagen, USA). The yield DNA concentration and purity were measured by the NanoDrop8000 spectrophotometer (Thermo Scientific). PARP-1 SNP rs8679 was genotyped using TaqMan assay (Applied Biosystems) as previously described by [20].

2.3. Gene Expression

Colorectal cancer and adjacent normal tissues were collected from patients for RNA isolation. RNA was extraction from 20 mg of fresh tissue using Qiagen RNeasy Mini Kit (Qiagen, USA). cDNA was synthesized using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA). The resulting cDNA was used for estimation of the relative mRNA expression rate of PARP-1 and GAPDH genes. TaqMan Gene Expression Assays (Applied Biosystems, Foster City, CA) were used for gene expression analysis of PARP-1 and GAPDH and amplification was performed using an ABI 7500 fast real-time PCR system (Life Technologies, USA). Three replicates were used for each sample. The relative PARP-1 expression levels were normalized to the GAPDH expression value. Data was evaluated using the comparative CT () method.

2.4. Statistical Analysis

The required sample size was calculated using Power and Sample Size Calculation Software Package (Vanderbilt University, Nashville, TN). Genotype and allelic frequencies of cases and control groups were compared using Pearson’s goodness-of-fit chi-square. SPSS statistical software version 22 (SPSS, Chicago, IL, USA) was used to calculate chi-square values, odds ratios (OR), 95% confidence intervals (CI), and values. Allele and genotype frequencies of PARP-1 SNP rs8679 for Saudi populations were compared with HapMap populations. Allele frequencies of Saudi population and other populations were compared using pairwise chi-square () tests as described by Alanazi et al. [21]. Survival curves were plotted by the Kaplan-Meier method and compared by the log-rank test. The survival data were evaluated by univariate Cox regression analyses. was considered statistically significant. Mann–Whitney test was performed to analyze the association between gene expressions.

3. Results

In the present study, we have evaluated the association 3′UTR region SNP of PARP-1 gene (rs8679) with colorectal cancer in 183 cases and 190 age and gender matched controls of Saudi population. The clinical data was shown in Table 1. The genotype and allele frequencies distribution along with odds ratios and significance of rs8679 are shown in Table 2.

SNP rs8679 of PARP-1 gene showed a statistically significant protective association with Saudi colorectal cancer patients. The genotype distributions of the analyzed SNP are as follows: 0.61 TT and 0.33 TC and 0.06 CC in cases but 0.47 TT and 0.45 TC and 0.07 CC in controls. The “TC” heterozygous allele posed about 0.56-fold lower risk with cases compared to the homozygous allele “TT” (OR: 0.56, CI: 0.368–0.871, = 6.75; ) (Table 2). The combination of “TC + CC” variants alleles genotypes also showed about 0.58-fold lower risk in cases, compared to the wild-type alleles (OR: 0.58; = 6.62; CI: 0.387–0.881; ). Moreover, we also found significant protective association at minor allele C with colorectal cancer cases. The minor allele frequency is less in cases (23%) compared to controls (30%) and there was 0.695-fold lower risk in cases, compared to controls ( = 4.75; CI: 0.501–0.965; ). The CC genotype alone did not show any effect. The reasons for these differences remain unclear but may relate to the distribution rate of the alleles in the Saudi population.

3.1. Effect of Age and Gender on the Association of PARP-1 SNP and CRC

To evaluate the association of PARP-1 SNP rs8679 with age at cancer diagnosis and gender. Patients were classified based on median age of cancer diagnosis as ≤57 () and >57 () and genotype frequencies were compared with age matched healthy controls. The genotypic frequencies of both clusters are shown in Table 3.

Similar to overall study younger age patients showed more protective association at heterozygous TC genotype (OR: 0.466, CI: 0.245–0.886, = 5.51; ) and TC + CC combined genotype (OR: 0.527, CI: 0.288–0.964, = 4.36; ) showed protective effect compared to TT homozygous genotype (Table 3). However, in older aged patient’s genotype and alleles did not show any associations with colorectal cancer risk (Table 3). After Bonferroni correction only TC genotype showed protective association against colorectal cancer in below 57-year-old age patients (Table 3).

Interestingly this SNP rs8679 showed statistically significant protective association only with the female gender and has no association with the male gender (Table 4). The rs8679 SNP showed statistically significant protective association against colorectal cancer risk with all alleles in females; the genotype distributions for TT, TC, and CC were 0.42, 0.49, and 0.09 in controls but 0.67, 0.30, and 0.03 in cases (Table 4). The heterozygous “TC” genotype frequency is less compared to homozygous allele “TT,” which posed 0.39-fold protective effect in colorectal cancer cases (: 7.32; CI: 0.193–0.775; ). In addition, the homozygous variant “CC” genotype frequency was also significantly low in female patients (0.03) compared to controls (0.09) and CC genotype showed significant protective effect in colorectal cancer female cases (OR: 0.181, : 5.04; CI: 0.035–0.927; ). The combination of heterozygous and variant genotypes TC + CC exhibited 0.353-fold lower risk in female cases, compared to female controls (: 9.43; CI; 0.180–0.692; ). Along with this the minor allele “C” frequency is significantly low in female colorectal cancer patients (0.18) compared to gender matched controls (0.34) (OR: 0.425, = 9.79; CI: 0.247–0.732; ). Even after Bonferroni correction we found significant protective association in female colorectal cancer patients with TC, CC, and TC + CC genotypes and also at C allele (Table 4).

3.2. Association of PARP-1 SNP rs8679 with CRC Risk Based on Tumor Location

To conduct the association of SNP rs8679 with tumor location of colorectal cancer we have stratified samples as colon and rectum based on tumor location. Interestingly SNP rs8679 SNP showed statistically significant protective association in both colon and rectum in CRC patients.

Patients with tumor located in colon area showed significantly lower risk (0.57-fold) with heterozygous genotype “TC” when compared with healthy individuals (: 4.69; CI: 0.345–0.950; ). The minor allele C frequency is also significantly low in patients with tumor located in colon area when compared to healthy controls (: 3.96; CI: 0.455–0.995; ) (Table 5). Patients with tumor located in rectum area showed significant protective effect at heterozygous TC genotype (: 4.74; CI: 0.277–0.940; ) and TC + CC (: 4.78; CI: 0.295–0.940; ) when compared with control samples (Table 5). But after Bonferroni corrections none of the genotypes showed significant association with colorectal cancer risk in patients with tumor located in colon and rectum areas (Table 5).

3.3. Genotype and Allele Frequencies of PARP-1 SNP rs8679 Variant in Saudis and Other Populations

We compared the genotypic and allelic frequencies of the PARP-1 SNP rs8679 in a normal healthy Saudi population with those of subjects in the HapMap project study groups. The allelic frequencies of rs8679 were significantly different with most of the populations except CEU (), GIH (), and TSI () populations. Tuscans in Italy (TSI, ) population showed more similarity with Saudi population in allelic frequencies (Table 6).

3.4. PARP-1 Gene Expression

The present study has also investigated PARP-1 expression levels in colorectal cancer tissues versus those of matched healthy tissues of the same patient. The results are shown in Figure 1. On the average, PARP-1 mRNA expression level (mean = 1.82 ± 0.30) was significantly higher in tumor tissues than in healthy tissues. Mann–Whitney test results showed significant difference in expression levels of tumor versus normal tissues (Mann–Whitney , ). Based on median expression level (1.03), colorectal cancer patients were divided into two groups. The first group comprised 24 cases which showed low-expression and the remaining 32 cases were included into high-expression group.

3.5. Impact of PARP-1 Expression on Prognosis of Colorectal Cancer

Kaplan-Meier method and log-rank test were used to evaluate the differences of overall survival between low-expression group and high-expression group. Kaplan-Meier analysis revealed that PARP-1 expression levels were not associated with survival rate in Saudi colorectal cancer patients (55 versus 62, , Figure 2).

We also performed Cox regression to determine which clinical and demographic parameters were significantly associated with PARP-1 expression level. Univariate analyses were used to assess whether the PARP-1 expression level and various clinicopathological conditions were independent prognostic parameters of colorectal cancer patient outcomes. The results of analysis are shown in Table 7. A univariate analysis of the prognosis factors with a Cox proportional hazards model confirmed that low has-miR-145 (HR = 3.083, 95% CI: 1.944–8.24, ) expression levels were significantly independent predictors of poor survival in colorectal cancer.

4. Discussion

PARP-1 is very highly expressed enzyme and it is the most identified and the well-characterized PARP of all the 17-member PARP family. In PARP family PARP-1 is accountable for majority of PARP activity, along with PARP-2 [22]. PARP-1 detects and binds, with high affinity, to DNA strand breaks and then it interacts and activates several proteins required for DNA damage repair and recruits these proteins at the site of breakage [23]. Beside DNA repair function, PARP-1 is involved in many cellular processes including conserving genomic stability, DNA synthesis, cell cycle regulation, telomere homeostasis, inflammation, and malignancy [24]. One of the key variants is PARP-1 rs8679 which is 3′UTR region and reported to have binding site for has-miR-145 and it is reported to be associated with increased risk of breast cancer [25].

In the present study we observed a significant protective association of PARP-1 rs8679 genotypes in colorectal cancer patients. Our genotyping results are contradicting with previous studies conducted in breast cancer [25, 26] and hepatocellular carcinoma [27]. Teo et al. [25] reported that rs8679 was significantly associated with increased breast cancer risk in individuals homozygous for the variant. Along with this they have reported that PARP-1 rs8679 SNP has binding site for has-miR-145 and also been predicted to have increased ΔΔG, whereas variant allele ΔG is less negative than the wild-type allele ΔG. This may reduce chances of has-miR-145 binding to PARP-1 mRNA 3′UTR which may enhance PARP-1 expression [25]. Zhai et al. [26] also reported that PARP-1 promoter SNP rs8679 does not show any association with breast cancer patients in Chinese population. Guillot et al. [27] reported that significant association was not observed between the PARP-1 rs8679 genotypes and its expression and activity. They reported that 2 cell lines (HepG2 and HepG2 2.2.15) have variant allele C of SNP rs8679; this is one of the possibilities in reduction to detect any differences in activity or expression based on specific genotype [27].

Stratification of samples based on clinical and demographic characteristics showed that rs8679 is showing reduced risk in colorectal cancer patients below 57-year-old patients (Table 3), in female patients (Table 4), in patients with tumor located in colon area (Table 5), and in rectum area (Table 5).

Our results provide the first evidence that the PARP-1 rs8679 polymorphism was associated with a decreased risk of colorectal cancer in a Saudi population.

Comparison of rs8679 genotypic frequencies with other HAPMAP populations showed close association of Saudi population with Italian population (TSI) and Indian and central European populations (Table 6). We also generated a regional linkage disequilibrium (LD) plot using SNAP (SNP Annotation and Proxy Search, http://www.broadinstitute.org/mpg/snap/ldplot.php) for PARP-1 rs8679. The LD plot indicated that there are multiple loci near rs8679 with high LD (), which suggests that fine mapping is necessary to evaluate the genetic effect of PARP-1 on cancer as well as functional studies (Figure 3). rs8689 is located at 5.3 kb upstream to 3′UTR region of PARP-1 and is in LD () with PARP-1 SNPs rs2271347 (), rs61835377 (), rs1805403 (), and rs2793383 (), which has been shown to be associated with increased risk for several cancers and other diseases.

In the present study, relative PARP-1 expression is statistically significantly high in colorectal cancer tissue when compared to normal tissue (Figure 1). Several factors affecting PARP-1 upregulation include PARP-1 3′UTR polymorphisms, transcription factors, and noncoding RNAs [19]. Our findings regarding PARP-1 expression in colorectal cancer are consistent with previously published research in which PARP-1 overexpression was also found in other types of cancer, including breast cancer [28], colorectal cancer [29, 30], prostate cancer [31], and glioblastoma [32]. PARP-1 expression was elevated throughout the patient-derived CRC samples. PARP-1 is an attractive target for tumor detection because it is increased expression in a large number of cancers. PARP-1 overexpression is believed to be due to the increased DNA damage occurring in genetically unstable cancer cells, rather than the activation of specific oncogenic pathways [33]. Nosho et al. [29] reported that PARP-1 overexpression was correlated significantly with overexpression of β-catenin, c-myc, cyclin D1, and MMP-7 in colorectal cancer.

In contrast to this Bai and Cantó [22] reported that rs8679 does not have any correlation with PARP-1 expression; this is opposing our results, but they observed significantly higher expression of PARP-1 gene in breast cancer.

PARP-1 high-expression level was significantly associated with age, gender, and tumor stage (Table 7). The higher levels of PARP-1 may be associated with poorer outcome. Our observations of increased expression of PARP-1 in poor prognosis tumors do lend support to the view that PARP inhibitors might play a role in therapy for colorectal cancer patients.

In conclusion, our study assessed colorectal cancer predisposition with PARP-1 gene 3′UTR SNP rs8679 in Saudi population. PARP-1 rs8679 SNP showed significant protective effect with CRC risk and it did not correlate with susceptibility to colorectal cancer in Saudi population. Our findings support that the upregulation of PARP-1 is associated with tumor progression and poor prognosis in Saudi patients with colorectal cancer, suggesting that PARP-1 can be novel and valuable signatures for predicting the clinical outcome of patients with colorectal cancer.

Competing Interests

The authors declare that there are no competing interests regarding the publication of this article.

Acknowledgments

The project was financially supported by King Saud University, through Vice Deanship of Research Chairs.

Supplementary Materials

The details of samples used were provided in Supplementary Table 1.

  1. Supplementary Material