Background
A recent epidemiological report from the National Central Cancer Registry (NCCR) of China showed that breast cancer is the most commonly diagnosed cancer among women in China. For female breast cancer, the 5-year survival rate (73%) in China is considerably reduced compared with corresponding estimates in Australia (89%) and the United States (90%) and to a lesser degree in Europe (82%) [
1]. Breast cancer ranks as the sixth leading cause of cancer deaths among women in rural areas of China, while the incidence rate (37.86/100,000) is continually increasing [
2]. Numerous breast cancer susceptibility genes have been identified by molecular epidemiological studies of cancer (e.g.,
BRCA1/
2,
TP53,
PTEN and
ATM).
BRCA1 and
BRCA2 mutations comprise 32 to 82% of hereditary breast cancer cases [
3]. The prevalence of these mutations, however, is considerably reduced in the Chinese population. For example,
BRCA1 and
BRCA2 mutations contribute to only 10% of hereditary breast cancer cases [
4,
5]. Thus, identification of other novel susceptibility loci among Chinese breast cancer patients is needed. Analyses of genetic polymorphisms reveal correlations between breast cancer susceptibility and many loci [
6,
7]. Recently, a considerable number of novel polymorphic variants have been identified by genome-wide association studies (GWAS) for breast cancer susceptibility [
8,
9].
DNA repair systems act to safeguard the integrity of the genome against damage caused by exogenous and endogenous carcinogens and by mutagens. Defects in DNA repair capability can increase the risk of carcinogenesis [
10], including breast cancer. Patients with breast cancer exhibit significantly reduced DNA repair proficiency [
11]. The high accumulation of DNA damage may cause the initiation of carcinogenesis and aberrant cell division [
12]. Additionally, several base excision repair (BER) enzymes may also participate in the regulation of other biological processes, including BER, cell cycle progression [
13], transcription initiation [
14] and apoptosis [
15]. Moreover, 8-oxoguanine DNA glycosylase 1 (
OGG1) and apurinic/apyrimidinic endonuclease 1 (
APEX1) are key enzymes in the BER pathway of oxidative DNA damage, excising abasic residues.
OGG1 is a glycosidase that hydrolyzes the bonds between damaged bases and the sugar-phosphate backbone in DNA, creating an abasic site, while the endonuclease
APEX1 cleaves the 5′ end of the abasic site. Mutations in these genes are expected to lead to a mutation-prone phenotype and contribute to tumor formation [
16]. In recent years, studies of breast cancer and DNA repair have emphasized the relationship between different single nucleotide polymorphisms (SNPs) in DNA repair-related genes and the probability of developing carcinoma. Kim et al. [
17] reported that the combined effect of
APEX1 Asp148Glu was associated with an increased risk of breast cancer in a Korean population. Likewise, Roberts et al. [
18] drew the conclusion that SNPs in nucleotide excision repair (NER) genes and BER genes affect the risk of developing breast cancer. Work by Smith et al. [
19] demonstrated only a slight positive association between individual DNA repair genotypes and breast cancer risk; however, the combined effects of multiple polymorphisms in DNA repair pathways may be more noteworthy.
The present work aimed to investigate whether APEX1 and OGG1 gene polymorphisms (frequency ≥ 5%) exert any synergistic effects on breast cancer risk. Using HapMap data, we selected four putative functional tag SNPs in APEX1 (rs1130409 and rs3136817) and OGG1 (rs1052133 and rs2072668) and further evaluated the genetic interactions of these four polymorphisms and their relation to breast cancer risk among the study population of Han women living in the Gansu area in Northwest China.
Methods
Study participants and data collection
A total of 518 histopathologically confirmed patients with breast cancer were enrolled in this study from the Gansu Provincial Cancer Hospital (Gansu Province, China) between December 2014 and August 2017. None of these patients had been previously treated with systemic therapy. For the age- and region-matched controls, 921 cancer-free women were randomly selected from the Health Examination Surveys conducted in the same hospital. Inclusion criteria included negative history of all forms of cancer and no family history of breast cancer. All related data were drawn from questionnaires (see Additional file
1) and the patients’ medical charts. Informed consent was obtained from all individual participants, and this research was approved by the ethical committee of the Gansu Provincial Academic Institute for Medical Research and Gansu Provincial Cancer Hospital (IRB number: A201412040036).
SNP identification and selection
Details of the SNP identification and selection have been described previously [
20]. Briefly, tag SNPs can be used to identify most genetic variations existing in a gene and develop markers to assess the relationship between a disease and a specific region, regardless of whether the tag SNPs have functional effects [
21]. Employing selection for tagger pairing processes, tag SNPs were selected from the HapMap CHB database (HapMap Data Rel 24/phase II Nov08). Tag SNPs were chosen by using Haploview version 4.0 software according to the following selection criteria: minor allele frequency (MAF) ≥ 0.05 and pair-wise r
2 ≥ 0.8. Finally, the
APEX1 polymorphisms rs1130409 and rs3136817 and the
OGG1 polymorphisms rs1052133 and rs2072668 were selected.
Genotyping method
A total of 5 mL of peripheral blood was collected in tubes containing EDTA from each of the participants for genomic DNA extraction using the genomic DNA Extraction Kit VER.3.0 according to the manufacturer’s instructions (TaKaRa Biotechnology, China). The DNA samples had A260/A230 ratios from 1.9 to 2.0 and A260/A280 ratios from 1.7 to 1.9; the concentrations of the samples were adjusted to 50 ng/μl. DNase/RNase-free distilled water was used in each assay as a non-template control (NTC). The multiplex TaqMan assays for genotyping were performed using the QuantStudio™ 12 K Flex Real-Time PCR System (Applied Biosystems, CA, USA) according to the manufacturer’s recommended operating conditions. The genotype data were analyzed using the OpenArray® SNP Genotyping Analysis software (version 1.0.5) (Applied Biosystems, CA, USA). To evaluate quality control, all genotyping assays were blind to the subject status. In addition, the reproducibility of genotyping was evaluated by direct sequencing in 20% of the samples randomly selected from each SNP and reached 100%.
Statistical analysis
To assess the deviation of the genotype frequencies of the four tag SNPs from Hardy-Weinberg equilibrium (HWE) in the control subjects, the chi-square goodness-of-fit test was used. The distribution of the
APEX1 and
OGG1 genotypes were compared between controls and patients using Pearson’s chi-square test (goodness-of-fit) or Fisher’s exact test. For the association tests, multivariate unconditional logistic regression adjusting for age was performed to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the associations of the
APEX1 and
OGG1 polymorphisms with breast cancer risk and clinicopathologic features. As reported previously, classical genetic models, including dominant models and recessive models, were used to analyze the associations [
22]. The frequency distributions of the
APEX1 and
OGG1 haplotypes were estimated using SHEsis online software (
http://analysis.bio-x.cn/myanalysis.php) [
23,
24].
To identify potential gene-gene interactions, we employed the multifactor dimensionality reduction (MDR) method (
http://www.epistasis.org) to evaluate the best cross-group model for further confirmation. MDR is a robust and novel data-mining approach that was applied as described previously [
25,
26]. Briefly, this method converts two or more discrete genetic variables to a single attribute. The aim of this approach is to identify the overall best combination all loci. With MDR, the newly formed one-dimensional variable can be assessed to predict disease status using cross-validation and permutation testing corrections by repeating the entire analysis on 1000 datasets. Furthermore, a final best MDR model is chosen that simultaneously has the maximal testing accuracy and cross-validation consistency (CVC).
To avoid false-positive results, haplotypes with frequencies greater than 3% were considered in all of the research participants. Bonferroni correction was employed to make the adjustment for multiple comparisons. All statistical analyses were performed using SPSS software (version 16.0), and differences were considered significant at P < 0.05.
Discussion
Genetic polymorphisms in key genes involved in DNA repair may influence DNA damage response, carcinogenesis and DNA repair capacity. Polymorphic variants have been confirmed to be good candidates for assessing cancer risk. We estimated the relationship between APEX1 and OGG1 gene polymorphisms and breast cancer risk among 1430 Han women of Northwest China in this experiment using a tag SNP-based study. The key finding was that the APEX1 polymorphism rs3136817 might mediate synergistic and independent effects on breast carcinogenesis. We found that the rs3136817 heterozygous TC genotype and the combined genotype (TC + CC) were associated with decreased breast cancer risk.
DNA excision repair capacity is known to play a crucial role in carcinogenesis [
27]. If changes that occur in the DNA sequence due to copying errors are not corrected, they may ultimately interfere with cell function. Damaged or inappropriate bases can be repaired by several mechanisms. The BER pathway is considered the primary mechanism involved in protecting against gene mutations and repairing DNA damage, and
APEX1 and
OGG1 are key components of this pathway. The BER proteins encoded by these genes act in a highly coordinated manner at the site of DNA damage. For example, repair of 8-oxoguanine, an oxidized base, is initiated by the
OGG1 glycosylase, which recognizes and removes damaged bases, forming an apurinic site that is cleaved by
APEX1. The resulting single-strand break can be subsequently repolymerized by DNA ligase 3 and polymerase
β [
28]. According to our current results, patients carrying the rs3136817 TC genotype had a reduced breast cancer risk, implying that heterozygosity at rs3136817 may have a positive influence on DNA repair capacity in DNA damage responsive pathways and thus potentially prevent breast cancer. Similar results were observed in bladder cancer and lung cancer. Nevertheless, this protection was not observed in individuals carrying other SNPs. Similar researches were only reported in bladder cancer and lung cancer. Zhu et al. [
29] reported that APEX1 rs3136817 TC genotype was associated with a decreased risk of bladder cancer. However, Li et al. [
30] demonstrated that no association between APEX1 rs3136817 and the risk of radiation-induced pneumonitis grade ≥ 3. Our study is the first to demonstrate that the rs3136817 heterozygous genotype was associated with a decreased risk of breast cancer. Of all the SNPs assessed, the rs1130409 and rs1052133 polymorphisms showed the most consistent relationship to breast cancer in two previous reports, demonstrating that these variations have no significant impact on breast cancer risk [
18,
31]. However, a few reports which were different from our results also indicated that breast cancer risk was significantly associated with APEX1 rs1130409 in North Indian [
32], Korean [
17] and Caucasian [
19] women. To the best of our knowledge, the rs3136817 polymorphic locus has not been previously evaluated regarding its association with cancer, except for bladder cancer and lung cancer. Moreover, the possibility that the associations noted above might have occurred by chance cannot be excluded. Further evidence in different regional populations and larger sample sizes, in addition to functional studies, are required to reinforce these results.
We further found that the
APEX1 rs1130409 GT and TT genotypes were increased in ER-positive patients by analyzing the association of clinicopathological features with the four SNPs, suggesting that these genotype carriers exhibited adverse clinicopathological features of breast cancer and failed to benefit from endocrine therapy. In addition, a lower frequency of women carrying the TT genotype was Her-2 positive. Her-2 is a ligand-less member of the human epidermal growth factor receptor family. Approximately 15% of patients with breast cancer exhibit Her-2 over-expression, which is associated with invasive behavior, unresponsiveness to common endocrine therapies, poor prognosis and reduced survival [
33,
34]. For
OGG1, the rs2072668 polymorphism was also associated with the pathologic characteristics of patients. The rs2072668 homozygous GG genotype carriers had a reduced frequency of Ki-67-positive expression. The Ki-67 protein is a nuclear marker of cell proliferation that is expressed at high levels in breast cancer patients, and increased expression is associated with worse outcomes [
35,
36]. Recent studies have indicated that long-term outcomes may be predicted by changes in Ki-67 expression after endocrine treatment [
37]. Therefore, our current results indicate that the
APEX1 rs1130409 TT genotype and the
OGG1 rs2072668 GG genotype deficiencies may lead to poorer prognosis and reduced survival.
In our study, the associations between breast cancer risk and haplotypes were also assessed. The APEX1 haplotype containing C-T (alleles rs3136817 and rs1130409) was observed at a higher frequency in the controls. Consequently, we infer that the APEX1 CT haplotype may be involved in decreasing the risk of breast cancer. MDR analysis has been used to examine the interactions of multiple genes in common diseases as a promising data-mining approach because it easily overcomes some of the limitations and inadequacies of traditional statistics, such as logistic regression, to characterize and examine gene-gene and gene-environment interactions. Meanwhile, a four-locus interaction associated with cumulative breast cancer risk was identified by MDR analysis. However, the four-locus best interaction model did not exhibit significant improvements compared with the other models in this case. The main effect model of rs3136817 performed best, which implies that the APEX1 rs3136817 polymorphism is a powerful risk factor and may play an important role in the interaction with other SNPs in affecting breast cancer development, either synergistically or antagonistically. Nevertheless, no statistical significance was observed, and evidence is therefore needed to support the hypothesis of SNP-SNP interactions in the future.
Conclusions
In summary, our data provide clear evidence that the rs3136817 polymorphism in APEX1 and a corresponding haplotype may be involved in breast cancer risk in the Han women of Northwest China. Notably, the rs3136817 heterozygous variant exhibited enhancement in the major DNA repair pathway, and this capacity may prevent the early development of breast cancer. As far as we know, this study is the latest to report that the APEX1 rs3136817 genotype is associated with cancer risk. The findings further suggest that the combined effect of SNPs determines the individual women’s risk for breast cancer. A better understanding of the mechanism of carcinogenesis will facilitate improved, individualized pharmaceutical therapy for patients with breast cancer and the implementation of breast cancer prevention strategies.
Acknowledgments
We wish to thank all the patients and healthy volunteers for providing blood samples. We also thank the nurses, clinical researchers and regulatory coordinators at the Department of Breast Surgery and Research Center of Translational Medicine at Gansu Provincial Cancer Hospital for their collaboration.