1 Introduction

The development of cervical tumors is a multi-step process including infection by the human papillomavirus (HPV) and involvement of the immune system, tumor suppressor genes, and proto-oncogenes [1, 2]. During cervical tumorigenesis, the normal cervical epithelium is transformed into cervical intraepithelial neoplasia (CIN), which may further progress to invasive cervical carcinoma [1, 2]. It is well recognized that the primary etiologic factors of this cancer are some oncogenic HPVs in which E6 and E7 oncoproteins deregulate innate and adaptive immunity and abnormally alter apoptosis, causing the cell cycle to drive normal cervical epithelium cells to immortalization [2–4]. The oncoproteins E6 and E7 may also result in chromosomal instability and increase DNA damage during HPV carcinogenesis [3, 5]. Moreover, DNA damage is accumulated during DNA replication as well as through exposure to genotoxic cellular metabolites and environmental insults [6]. There are several canonical pathways for DNA repair, and the predominant pathway for single strand break repair is the base excision repair (BER) pathway [7, 8]. The BER may cooperate with a family of related enzymes termed poly(adenosine diphosphate [ADP]-ribose) polymerases (PARP) [9, 10]. Approximately 90 % of the cellular PARP activity is due to PARP-1, which is an early-activated sensor of DNA strand breakage [6, 11]. PARP-1 conducts extensive polymerization of ADP-ribose from its substrate nicotinamide adenine dinucleotide (NAD+) to nuclear proteins involved in DNA repair, genomic stability, transcription regulation, cell death, and proliferation [12, 13]. The proteins involved in pathways of DNA repair are crucial in the prevention of cancer development, and numerous single nucleotide polymorphisms (SNPs) in genes encoding these proteins may be risk factors for various cancers [14, 15]. At least 60 SNPs have been reported in PARP-1 (http://snp500cancer.nci.nih.gov), and among them is the most frequently studied functional PARP-1 polymorphism, which includes a 2446T>C transition leading to Val762Ala variation [16, 17].

Data demonstrating that the PARP-1 Val762Ala (rs1136410) substitution is a risk factor for various types of cancers in different ethnicities are inconsistent [16, 17]. Recently, the contribution of the PARP-1 Val762Ala polymorphism to cervical cancer was demonstrated in an Asian population [18]. We evaluated the PARP-1 Val762Ala genotype and allele frequencies in patients with cervical cancer (n = 446) and controls (n = 491) from a Polish population.

2 Patients and Methods

2.1 Patients and Controls

The patients were 446 women with histologically recognized cervical carcinoma according to the International Federation of Gynecology and Obstetrics (FIGO). All women were enrolled between April 2007 and Jun 2012 at the Department of Radiotherapy, Greater Poland Cancer Center in Poznan, Poland (Table 1). The controls included 491 unrelated healthy female volunteers who were matched by age to the patients (Table 1). Data on pregnancy, oral contraceptive use, tobacco smoking, and menopausal status were obtained during the clinical interview. All individuals were Caucasian, enrolled from the Wielkopolska (Greater Poland) area of Poland. Patients and controls provided written informed consent. The study was approved by the Local Ethical Committee of Poznan University of Medical Sciences.

Table 1 Clinical and demographic characteristics of the patients and controls
Full size table

2.2 Genotyping

DNA was isolated from peripheral leucocytes using a salting-out procedure. The PARP-1 Val762Ala (rs1136410) DNA fragment was amplified using the primers 5′ CTATCATCAGACCCTCCCCTGA 3′ and 5′ GATACCTAAGTCGGGGGCTTTC 3′. This polymorphism was then genotyped by high-resolution melting curve analysis (HRM) on a LightCycler 480 system (Roche Diagnostics, Mannheim, Germany). The presence of the PARP-1 Val762Ala polymorphism was verified by commercial sequencing analysis of 15 % randomly selected samples.

2.3 Statistical Analysis

The differences in genotypic and allelic prevalences between patients and controls and their genotype deviations from Hardy–Weinberg (HW) equilibrium were evaluated via the chi-square test. The polymorphism was tested for association with cervical cancer incidence using the chi-square test for trend (p trend). Moreover, the odds ratio (OR) and 95 % confidence intervals (95 % CI) were calculated. Unconditional logistic regression analysis was used to adjust for the effects of confounders such as age, pregnancy, oral contraceptive use, tobacco smoking, and menopausal status. A p value of <0.05 was considered statistically significant.

3 Results

3.1 Prevalence of the PARP-1 Val762Ala Polymorphism in Women with Cervical Cancer

The distribution of PARP-1 Val762Ala genotypes did not exhibit significant differences from HW equilibrium in the cases and controls. The prevalence and adjusted analyses of PARP-1 Val762Ala genotypes in women with cervical cancer are presented in Table 2. The frequency of the PARP-1 Ala/Ala genotype was approximately 1.7-fold higher in the patients than in the controls. The PARP-1 Ala/Val heterozygous genotype frequency was also higher in women with cervical cancer than in the controls(0.29 and 0.23, respectively). The PARP-1 Ala allele frequency was increased in patients compared to controls (0.19 and 0.15, respectively). The p value from the chi-square test of the trend observed for the PARP-1 Val762Ala polymorphism was statistically significant (p trend = 0.0123). Logistic regression analysis demonstrated that the PARP-1 Val762Ala polymorphism was associated with an increased risk of cervical cancer. The adjusted OR for patients with the Ala/Val genotype versus the Val/Val genotype was 1.381 (95 % CI = 1.025–1.859, p = 0.033) and the adjusted OR for the Ala/Ala or Ala/Val genotype versus the Val/Val genotype was 1.403 (95 % CI = 1.057–1.863, p = 0.019). However, we did not observe statistical significance for the Ala/Ala genotype versus the Val/Val genotype. In this case, the adjusted OR was 1.290 (95 % CI = 0.922–1.804, p = 0.136). Stratification of the patients based on the histological type of the cancer revealed a borderline association of the Ala/Ala or Ala/Val genotype with squamous cell carcinoma, with an adjusted OR of 1.340 (95 % CI = 1.068–1.804, p = 0.048). However, we did not find a significant association between the PARP-1 Val762Ala polymorphism and adenocarcinoma, tumor stage, or histological grade (data not shown).

Table 2 Association of the PARP-1 Val762Ala (rs1136410) polymorphism with cervical cancer
Full size table

3.2 Stratified Analyses of PARP-1 Val762Ala Genotype and Cervical Cancer Risk

The results of age-adjusted analyses of PARP-1 Val762Ala genotype and cervical cancer risk stratified by pregnancy, oral contraceptive use, tobacco smoking, and menopausal status are presented in Table 3. An increase in cervical cancer risk was seen among patients with a positive history of pregnancy or oral contraceptive use, and among women of postmenopausal age. The adjusted OR for pregnancy with Ala/Ala or Ala/Val vs Val/Val genotype was 1.388 (95 % CI = 1.027–1.877, p = 0.0328). The adjusted OR for contraceptive use with Ala/Val vs Val/Val genotype was 1.769 (95 % CI = 1.114–2.809, p = 0.0154) and with Ala/Ala or Ala/Val vs Val/Val genotype it was 1.773 (95 % CI = 1.145–2.745, p = 0.0100). The adjusted OR for postmenopausal women with Ala/Val vs Val/Val genotype was 1.577 (95 % CI = 1.094–2.272, p = 0.0143) and with Ala/Ala or Ala/Val vs Val/Val genotype it was 1.604 (95 % CI = 1.132–2.272, p = 0.0077). However, no significant association was seen between PARP-1 Val762Ala and patients with a positive history of tobacco smoking.

Table 3 Stratified analyses of the PARP-1 Val762Ala genotype distribution and cervical cancer risk: pregnancy, oral contraceptive use, tobacco smoking, and menopausal status
Full size table

4 Discussion

The activity of PARP has been implicated in cancer development and anticancer therapy [16, 17, 19–21]. Efficient DNA repair mechanisms are essential for protecting against the accumulation of genetic defects in DNA and subsequent carcinogenesis [14]. It was demonstrated that Epstein–Barr virus-immortalized lymphocytes from centenarians displayed a maximal PARP activity that was significantly higher than that seen in controls aged 20–70 [21]. A large difference between patients with cancer and healthy subjects in the PARP levels of peripheral blood mononuclear cells (PBMC) has also been reported [20–24]. Ranjit et al. [20] found that PBMC from patients with esophageal cancer, breast cancer, and lymphocytic malignancies contained lower levels of PARP than PBMC from healthy individuals. Furthermore, decreased PARP-1 activity in peripheral blood lymphocytes has been observed in patients with tumors of the larynx, lung, colon, and prostate [22–24]. In addition to these findings, skewed PARP-1 levels have been observed in different primary human malignancies, including breast, colorectal and head and neck cancers, and in melanoma [25–27].

The role of the PARP-1 enzyme in tumorigenesis has also been well documented in the murine model [28–30]. PARP-1 knockout mice (PARP-1 −/−) treated with either alkylating compound or γ-radiation exhibited increased genomic instability, higher numbers of chromosomal aberrations, and reduced telomere length as compared to the wild-type mice [28]. The PARP-1 −/− mice also displayed an increased risk of chemically induced carcinogenesis of the lung, liver, and colon [29, 30].

Therefore, genetic variants of PARP-1 that contribute to PARP-1 activity may be risk factors for cancer development and progression. The PARP-1 762Ala gene variant seems to play a protective role in the development of some cancers in Caucasian populations [31–35], but the opposite is seen in Chinese populations, where the PARP-1 762Ala gene variant seems to be a risk factor for cancer in several studies [36–39].

We found an association of the PARP-1 Ala/Val genotype together with the Ala/Ala genotype with cervical cancer development in Caucasian women, but this association was not observed solely for the PARP-1 Ala/Ala genotype. The PARP-1 Val762Ala polymorphism has also been recognized as a risk factor for cervical tumorigenesis in Chinese women [18]. Ye et al. [18] demonstrated that the PARP-1 Ala/Ala genotype contributed to cervical carcinoma but not to CIN. The lack of an association of the Ala/Ala genotype with cervical cancer in our studies might due to a lower frequency of the Ala/Ala genotype in Caucasian women as compared to Chinese women.

Many genetic studies have been performed to date to assess the role of the PARP-1 Val762Ala polymorphism as a risk factor for carcinomas of the brain, head and neck, esophagus, lung, breasts, stomach, bladder, colorectum, prostate, skin, among others; however, those studies provide discordant results [22, 31–43]. Recently, two meta-analyses were conducted to evaluate the contribution of the PARP-1 Val762Ala polymorphism to overall cancer risk and the risks of different cancer types in distinct ethnicities [16, 17]. In the first meta-analysis, the PARP-1 Ala variant demonstrated an association with an increased risk of cancer among Asian populations, but a decreased risk of cancer among Caucasian populations, especially for glioma [16]. The second meta-analysis confirmed that the PARP-1 Ala variant is associated with a significantly increased risk of gastric carcinoma in Chinese populations; in Caucasians, the Ala variant protected significantly from brain tumors in two American subgroups, but not from the development of other cancers [17].

The PARP-1 Val762Ala substitution is situated in the sixth helix of the C-terminal catalytic domain that binds NAD+ and results in poly(ADP-ribosyl)ation [44]. This C-terminal catalytic domain is currently targeted for the development of effective PARP inhibitors [45]. The PARP-1 762Ala gene variant is significantly associated with low poly(ADP-ribosyl) activity [22]. The Val762Ala substitution in the PARP-1 enzyme produces a steric alteration in the catalytic domain leading to reduced activity of the PARP-1 762Ala enzyme variant [46].

In our studies, we also observed that PARP-1 Val762Ala increased the risk of cervical cancer in women with a positive history of pregnancy, oral contraceptive use, and postmenopausal women. This is in agreement with literature indicating possible causative roles in cervical tumorigenesis of contraceptive use, tobacco consumption, age, and environmental exposures [47–49]. These confounding variables, especially exposure to unknown environmental insults, may affect the evaluation of genetic factors, including the PARP-1 Val762Ala SNP, as risk factors for various cancers [47–49]. The two recent meta-analyses indicate that the PARP-1 Val762Ala substitution may contribute to not only race-specific but also tissue-specific genetic effects in tumorigenesis [16, 17]. Our findings demonstrate the association of the PARP-1 762Ala gene variant with cervical cancer in Caucasians. This may be due to interactions between the reduced activity of the PARP-1 Ala variant enzyme and individual exposure to some environmental factors. In addition to our findings, the PARP-1 Ala/Ala and Val/Ala genotypes have been shown to be risk factors for bladder cancer in a large cohort of Caucasian patients [50]. Moreover, in Caucasians, the PARP-1 Ala/Ala genotype was also significantly associated with an increased risk of prostate cancer [22].

Our genetic assessment is the first to demonstrate that the PARP-1 762Ala gene variant can be a risk factor for cervical cancer in a Caucasian cohort; this study should therefore be replicated in other independent cohorts.