Association between apurinic/apyrimidinic endonuclease 1 rs1760944 T>G polymorphism and susceptibility of cancer: a meta-analysis involving 21764 subjects

The incidence and mortality of cancer is increasing worldwide [1–3]. It was estimated that approximately 18.1 million new cancer patients were diagnosed and more than half of them died worldwide during 2018 [1]. The etiology of cancer is complicated. Previous epidemiological studies have indicated that consumption of red meat, fried and salted meat, tobacco smoking and alcohol abuse, diabetes mellitus, obesity, non-alcoholic fatty liver disease, oxidative stress, chronic infection and inflammation can contribute to the development of cancer [4–6]. However, these potential risk factors could not fully explain the etiology of cancer. It is reported that the hereditary factor may influence the susceptibility of cancer [7,8].

Apurinic/apyrimidinic endonuclease 1 (APE1) is a multifunctional protein which plays an important role in the pathway of base excision repair (BER). APE1 plays a pivotal role in tumor cells involving DNA damage response and regulating transcription factor activation [9]. The observed roles of APE1 protein allude to its potential effect on inflammation, growth, migration and angiogenesis [9–11]. In addition, APE1 may be also implicated in regulating cell cycle, oxidative stress and apoptosis [12]. Recently, some investigations reported that the expression level of APE1 was up-regulated in a number of cancers [13–17]. In addition, glioma cell with higher APE1 expression level was also associated with shorter time to tumor progression after chemo/radiotherapy [18,19]. As well, previous studies reported that the decreased APE1 activity might retard cell growth of ovarian cancer [20] and pancreatic cancer [21].

APE1 gene is approximately 3 kb in length and is located on chromosome 14q11.2 [22]. A number of variants in APE1 gene are established (https://www.ncbi.nlm.nih.gov/snp/?term=APE1). APE1 rs1760944 (−656T>G) is a promoter locus and has been widely explored. Some functional studies indicated that the APE1 rs1760944 T>G single-nucleotide polymorphism (SNP) might decrease APE1 mRNA and protein expression levels [23,24]. Many case–control studies were conducted to identify the potential association of APE1 rs1760944 T>G polymorphism with the development of cancer. Individuals with APE1 rs1760944 GG variant might reduce 46% glioblastoma risk than those who carried APE1 rs1760944 TT variant [25]. The relationship between APE1 rs1760944 T>G polymorphism and a decreased susceptibility of lung cancer was also found by Lu et al. [24]. A previous study reported that gastric cancer cases carried APE1 rs1760944 GT/GG variants might have a better survival than others with APE1 rs1760944 TT genotype [26]. But the results were conflicting. Two meta-analyses suggested that this SNP was correlated with a decreased susceptibility of cancer in Asian populations and lung cancer [27,28]. Recently, many investigations focused on the association between APE1 rs1760944 T>G polymorphism and the risk of other cancers. The findings were more confusing. The aim of the present study was to carry out a meta-analysis to evaluate whether this SNP was associated with the risk of cancer.

PubMed and Embase databases were exhaustively searched for relevant publications which studied the relationship of APE1 rs1760944 T>G locus with the risk of cancer from the inception up to 17 March 2019. The search strategy was: (polymorphism OR SNP) and (apurinic/apyrimidinic endonuclease 1 or APE1 or APE-1) and (cancer OR carcinoma). In the current study, publications written in English or Chinese were eligible. Moreover, the references of the included studies, comments, meta-analyses and reviews were manually retrospected to recruit the potential literatures.

For eligibility, publications were required to meet the following inclusion criteria: (1) case–control studies investigating the relationship between the APE1 rs1760944 T>G locus and the risk of cancer; (2) the diagnosis of cases was confirmed by pathological examination; (3) the frequencies of alleles or genotypes were presented; (4) the paper was written in English or Chinese.

Studies were excluded based on the major exclusion criteria: (1) not case–control design; (2) studies did not provide genotyping data on APE1 rs1760944 T>G polymorphism; and (3) meta-analyses/reviews, comments and letters focusing on the relationships between the APE1 rs1760944 T>G locus and cancer risk.

Two authors (Guowen Ding and Yu Chen) reviewed each eligible study independently. They extracted the following terms from case–control studies, including the first author name, publishing year, country where the study was carried out, ethnicity, the source of control, cancer type, numbers of included cases and controls in each case–control study, genotyping data, the method of polymerase chain reaction, statistical method and evidence of Hardy–Weinberg equilibrium (HWE) evaluation in control group. If the extracted data had any dispute, authors settled these issues following a detailed discussion among all reviewers.

HWE in controls was assessed by an online Pearson’s χ² test (http://ihg.gsf.de/cgi-bin/hw/hwa1.pl). We calculated crude odds ratios (ORs) and 95% confidence intervals (CIs) to evaluate the correlation of APE1 rs1760944 T>G polymorphism and cancer risk. The following four genetic models were used, including homozygote model (GG vs. TT), dominant model (GG/TG vs. TT), recessive model (GG vs. TT/TG) and allele model (G vs. T). Cochran’s Q-statistic and I² test were used to check the heterogeneity among the included studies. The random-effect model was harnessed when I² > 50% or P<0.10 [29]; otherwise, a fixed-effect model was used [30]. Subgroup analyses were performed to explore the heterogeneity source among the studies. Ethnicity, the source of control and cancer type was considered as the potential source of heterogeneity. Begg’s funnel plots and Egger’s linear regression test were used to detect the potential bias in this meta-analysis. Since significant bias was identified in the present study, non-parametric ‘trim-and-fill’ method was used to evaluate the stability of the observed results. Sensitivity analysis was conducted by one-way method, which deleted each study one by one and re-calculated the pooled ORs and CIs. All statistical analyses were conducted by using STATA 12.0 (Stata Corporation, TX, U.S.A.). A P-value (two-sided) <0.05 was defined as statistically significant.

Figure 1 shows the selection process of the eligible publications. A total of 343 papers were collected. According to the major inclusion criteria, there were 20 papers (including 23 independent case–control studies) focusing on the relationship of APE1 rs1760944 T>G polymorphism with cancer risk [23–25,31–47]. Among them, five investigated lung cancer [24,31–33], three investigated colorectal cancer [34–36], three investigated breast cancer [37–39], three investigated cervical cancer [40,41], two investigated osteosarcoma [23], two investigated nasopharyngeal carcinoma [42,43] and five investigated other cancers (bladder cancer [44], glioblastoma [25], renal cell carcinoma [45], prostate cancer [46] and ovarian cancer [47]). Additionally, twenty-one had Asian and two had Caucasian ethnicities. In all included studies, χ² test was used to calculate the pooled ORs and CIs. The detailed characteristics of the included case–control studies are shown in Table 1. The number of each genotype and HWE are presented in Table 2.

A total of 23 independent case–control studies with 10166 cancer cases and 11598 controls were included to explore the potential correlation of APE1 rs1760944 T>G polymorphism with the susceptibility of cancer [23–25,31–47]. We found that APE1 rs1760944 T>G polymorphism conferred statistical evidence of the relationship between APE1 rs1760944 T>G locus and a decreased risk of cancer (G vs. T: OR, 0.87; 95% CI, 0.83–0.92 P<0.001; GG vs. TT: OR, 0.77; 95% CI, 0.69–0.86; P<0.001; GG/TG vs. TT: OR, 0.83; 95% CI, 0.77–0.89, P<0.001 and GG vs. TT/TG: OR, 0.85; 95% CI, 0.80–0.92, P<0.001; Table 3).

When we conducted subgroup analyses according to the different populations, the findings indicated that APE1 rs1760944 T>G polymorphism might be a protective factor for the development of cancer in Asian population (G vs. T: OR, 0.86; 95% CI, 0.82–0.91 P<0.001; GG vs. TT: OR, 0.75; 95% CI, 0.67–0.84; P<0.001; GG/TG vs. TT: OR, 0.82; 95% CI, 0.76–0.89, P<0.001 and GG vs. TT/TG: OR, 0.83; 95% CI, 0.78–0.90, P<0.001; Figure 2).

When we conducted subgroup analyses according to cancer type, the results suggested that APE1 rs1760944 T>G polymorphism decreased the risk of lung cancer (G vs. T: OR, 0.83; 95% CI, 0.78–0.90, P<0.001; GG vs. TT: OR, 0.68; 95% CI, 0.59–0.79; P<0.001; GG/TG vs. TT: OR, 0.80; 95% CI, 0.72–0.90, P<0.001 and GG vs. TT/TG: OR, 0.77; 95% CI, 0.68–0.87, P<0.001), breast cancer (G vs. T: OR, 0.83; 95% CI, 0.73–0.95, P=0.005; GG vs. TT: OR, 0.75; 95% CI, 0.57–0.98; P=0.034 and GG/TG vs. TT: OR, 0.71; 95% CI, 0.59–0.86, P<0.001), and osteosarcoma (G vs. T: OR, 0.69; 95% CI, 0.57–0.83 P<0.001; GG vs. TT: OR, 0.51; 95% CI, 0.35–0.75; P=0.001; GG/TG vs. TT: OR, 0.61; 95% CI, 0.47–0.80, P<0.001 and GG vs. TT/TG: OR, 0.64; 95% CI, 0.45–0.91, P=0.014).

In the present study, Begg’s and Egger’s tests were used to assess the potential bias among the eligible studies. Evidence of bias was found in the present study (G vs. T: Begg’s test P=0.055, Egger’s test P=0.013; GG vs. TT: Begg’s test P=0.037, Egger’s test P=0.080; GG/TG vs. TT: Begg’s test P=0.051, Egger’s test P=0.016; GG vs. TT/TG: Begg’s test P=0.055, Egger’s test P=0.174; Figure 3).

Since bias was found, we used non-parametric ‘trim-and-fill’ method to evaluate the stability of results. When we treated the publication bias, the adjusted ORs and CIs were not significantly changed (Figure 4).

In this meta-analysis, sensitivity analysis was conducted by one-way method, which deleted an individual case–control study one by one and re-calculated the pooled ORs and CIs. No single case–control study significantly influenced the final decision (Figure 5).

We found significant heterogeneity in all genetic models. Considering the potential factors for heterogeneity, subgroup analysis was conducted to identify its major source. In this meta-analysis, Asians, cervical cancer and population-based studies contribute to the major sources of heterogeneity.

The APE1 rs1760944 T>G has been frequently investigated due to its potential role in the development of cancer; however, the results are conflicting. To shed light on this issue, we performed an extensive meta-analysis. The results highlighted that APE1 rs1760944 T>G polymorphism decreased the risk of cancer. Results of subgroup analyses demonstrated that this SNP still significantly modified the risk among lung cancer, breast cancer, osteosarcoma patients and Asians.

Rs1760944 T>G is a promoter SNP in the APE1 gene and may affect the binding of transcription factors. Since 2007, a number of case–control studies were performed to assess the potential relationship of APE1 rs1760944 T>G polymorphism with the risk of cancer, but the observations were controversial. Several investigations suggested that APE1 rs1760944 T>G SNP decreased the susceptibility of cancer [23,24,31,32,37,38,40,42,44,46]. However, other case–control studies suggest null correlation between the APE1 rs1760944 T>G SNP and cancer risk [25,33–36,39,41,43,45,47]. How can we obtain an extensive evaluation of the relationship between APE1 rs1760944 T>G locus and the risk of cancer with the consistent conclusions? To our knowledge, small sample size investigation could lead to confusing findings. Thus, we carried out a meta-analysis with 23 independent case–control studies to explore the correlation between APE1 rs1760944 T>G SNP and the susceptibility of cancer. In the included case–control studies, χ² test was used to calculate the pooled ORs and CIs. In meta-analysis, we also used χ² test to evaluate the relationship of APE1 rs1760944 T>G polymorphism with cancer risk. Overall, we found that APE1 rs1760944 T>G polymorphism decreased the risk of cancer in four genetic models. When we conducted subgroup analyses, we found that APE1 rs1760944 T>G polymorphism decreased the risk of lung cancer, breast cancer, osteosarcoma and Asians. To the best of our knowledge, the association might be confounded by some potential bias (e.g. publication bias, heterogeneity and lack of accordance with HWE in controls). Thus, we subsequently performed subgroup analyses. The findings suggested that the APE1 rs1760944 T>G polymorphism might be a protective effect on the development of cancer in Asians only, but not Caucasians. In the present study, one case–control study was incongruent with HWE [37]. When we deleted it and re-calculated the pooled ORs and CIs, the significant relationship was not changed. In the present study, we conducted non-parametric ‘trim-and-fill’ method to explore the potential influence of publication bias. We found that the bias of publication might not alter the findings. We also found that APE1 rs1760944 T>G polymorphism still significantly decreased the risk of some type of cancers.

It was found that inhibition of APE1 activity might reduce cell growth of ovarian cancer [20] and pancreatic cancer [21]. In addition, Luo et al. [48] identified that a decreased APE1 activity could also significantly retard the proliferation of endothelial cells, suggesting its stimulative effect on the development of cancer. Several studies indicated that APE1 rs1760944 G allele decreased APE1 mRNA and protein expression levels [23,24]. Additionally, Lu et al. [24] reported that APE1 rs1760944 G allele was associated with a decreased level of APE1 mRNA by reducing the binding affinity of some transcription factors. Although the pathway of the relationship between APE1 rs1760944 T>G and cancer risk has been not confirmed, it is speculated that this SNP may alter the susceptibility of cancer through the mechanism mentioned above. All observations and speculations should be verified with new molecular studies.

Some limitations of the current analysis should be noted. First, in this meta-analysis, only published literature was eligible and included, and some presumable unpublished studies might be neglected and discarded. Second, heterogeneity and publication bias were apparent, which could distort the pooled results. Our findings should be interpreted with cautions. Third, for lack of sufficient data (e.g., smoking, drinking, age, sex and vegetable and fruit intake and other environmental factors), we only conducted a crude assessment. Finally, only APE1 rs1760944 T>G polymorphism was included to assess the association with the risk of cancer; other functional loci in APE1 gene should not been ignored.

In summary, this updated meta-analysis highlights that the APE1 rs1760944 T>G polymorphism may play a protective role in the development of cancer. Further studies in different race are needed to confirm or refute our findings.

We wish to thank Dr. Yafeng Wang (Department of Cardiology, The People’s Hospital of Xishuangbanna Dai Autonomous Prefecture, Jinghong, Yunnan Province, China) for technical support.

Conceived and designed the experiments: S.C. Performed the experiments: G.D., Y.C. and H.P. Analyzed the data: W.T. and H.Q. Contributed reagents/materials/analysis tools: S.C. Wrote the manuscript: G.D., Y.C. and W.T.

The authors declare that there are no competing interests associated with the manuscript.

This work was supported in part by the Young and Middle-aged Talent Training Project of Health Development Planning Commission in Fujian Province [grant number 2016-ZQN-25]; the Program for New Century Excellent Talents in Fujian Province University [grant number NCETFJ-2017B015]; the Joint Funds for the Innovation of Science and Technology, Fujian province [grant numbers 2017Y9099, 2017Y9077]; the National Natural Science Foundation of China [grant number U1705282]; the Ministry of Health, P.R. China [grant number WKJ2016-2-05]; the Natural Science Foundation of Fujian Province [grant numbers 2016J01513, 2017J01259, 2018J01267]; and the Fujian Provincial Health and Family Planning Research Talent Training Program [grant numbers 2015-CX-7, 2018-ZQN-13, 2016-1-11, 2018-1-13].

APE1

apurinic/apyrimidinic endonuclease 1

CI

confidence interval

HWE

Hardy–Weinberg equilibrium

OR

odds ratio

SNP

single-nucleotide polymorphism