RESEARCH ARTICLE

The association between the APE1 Asp148Glu polymorphismand breast cancer susceptibility: a meta-analysis basedon case–control studies

Zhiyong Zhao & Chuan Liu & Yong Zeng & Lei Gu &

Mingzhen Ying & Ning Wang & Bin Hao & Huiyan Yao &

Changqing Su & Yajie Wang & Yuchen Ma

Received: 18 December 2013 /Accepted: 3 January 2014# International Society of Oncology and BioMarkers (ISOBM) 2014

Abstract Published data regarding the association betweenthe APE1 Asp148Glu polymorphism and breast cancer sus-ceptibility showed inconclusive results. This meta-analysis ofliteratures was performed to draw a more precise estimation ofthe relationship. We systematically searched PubMed,Embase, Elsevier, and Springer for relevant articles publishedbefore December 10. 2013. The strength of association be-tween APE1 Asp148Glu polymorphism and breast cancersusceptibility was assessed by odds ratio (OR) with the cor-responding 95 % confidence interval (95 % CI) using thesoftware Stata (version 10.0). A total of 7 case–control studiesincluding 3,460 cases and 3,909 controls were included foranalysis. Overall, no significant associations were found be-tween the APE1Asp148Glu polymorphism and breast cancersusceptibility for GG vs TT (OR=1.00, 95 % CI=0.87–1.14);

TG vs TT (OR=1.06, 95 % CI=0.95–1.18); the dominantmodel GG+TG vs TT (OR=1.04, 95 % CI=0.94–1.16) andthe recessive model GG vs TG+TT (OR=0.99, 95 % CI=0.88–1.11). In subgroup analysis, a significant association wasfound for TG vs TT in Asian subgroup (OR=1.17, 95 % CI=1.00~1.36) and in population-based subgroup (OR=1.18,95 % CI=1.00~1.38). This meta-analysis suggested that theAPE1 Asp148Glu polymorphism was a risk factor for breastcancer susceptibility among Asian population.

Keywords APE1Asp148Glu . Breast cancer .

Susceptibilitym .Meta-analysis

Introduction

Breast cancer is the most frequently diagnosed cancer and theleading cause of cancer death in females worldwide, account-ing for 23 % (1.38 million) of the total new cancer cases and14 % (458,400) of the total cancer deaths in 2008 [1]. Themechanism of breast carcinogenesis is still not fully under-stood. It has been suggested that low-penetrance susceptibilitygenes combining with environmental factors may be impor-tant in the development of breast cancer [2].

Genetic variation in DNA repair genes can cause alterationin DNA repair function, resulting in the accumulation of DNAdamage and gene mutations, and the development of healthconsequences such as cancer [3]. Base excision repair (BER)is the predominant DNA damage repair pathway for theprocessing of small base lesions due to oxidation and alkyl-ation damage [4]. The human apurinic/apyrimidinic endonu-clease (APE) APE1 (also known as APE, APEX, and Ref-1) isan important enzyme in the BER pathway which is responsi-ble for the repair of DNA caused by oxidation/alkylationdamage and protect cells against the effect of endogenous

Zhiyong Zhao, Chuan Liu, and Yong Zeng contributed equally to thiswork and should be considered as co-first authors. Yajie Wang andYuchen Ma should be considered as co-corresponding authors.

Z. ZhaoChinese PLA General Hospital, Medical School of Chinese PLA,Beijing, People’s Republic of Chinae-mail: zhaoziyong1229@163.com

Z. Zhao : B. Hao :H. Yao :Y. Ma (*)Department of rehabilitation, Chinese PLA 252 Hospital, Baoding,People’s Republic of Chinae-mail: docma1953@126.com

C. Liu : L. Gu :M. Ying :N. Wang :Y. Wang (*)Department of Oncology, Changhai Hospital, Second MilitaryMedical University, 168 Changhai Road, Shanghai 200433,People’s Republic of Chinae-mail: yajiewao459@163.com

Y. Zeng :C. SuDepartment of Molecular Oncology, Eastern Hepatobiliary SurgicalHospital, Second Military Medical University, Shanghai,People’s Republic of China

Tumor Biol.DOI 10.1007/s13277-014-1618-5

and exogenous agents [5]. The APE1 is located on chromo-some 14q11.2–q12 and encodes a 317 amino acid protein [6].Several sequence variants were identified in this gene, includ-ing a G/T change in exon 5 leading to an amino acid changefrom aspartic acid to glutamic acid (Asp148Glu; dbSNP no.rs3136820), a C/G change in exon 3 leading to an amino acidchange from glutamic acid to histidine (Gln51His; dbSNP no.rs1048945), and an A/G substitution resulting in an aminoacid change from isoleucine to valine in exon 64 (Ile64Val;dbSNP no. rs2307486), but the most extensively studiedpolymorphism is the Asp148Glu polymorphism [7, 8].

To date, molecular epidemiological studies have investi-gated the relationship between the APE1 Asp148Glu poly-morphism and breast cancer susceptibility. However, the re-sults of these studies remained inconclusive. Therefore, weperformed this meta-analysis of all eligible studies to derive amore precise estimation of the association between the APE1Asp148Glu polymorphism and breast cancer susceptibility.

Materials and methods

Study identification and selection

In our meta-analysis, we searched the PubMed, Embase, andWeb of Science (updated to December 10, 2013) using thefollowing search terms: “base excision repair,” “APE1,”“APEX1,” “Ref-1,” “polymorphism,” and “breast cancer.” The

searches were performed without any restrictions on languageand were focused on studies that had been conducted in humans.

The following criteria were used for the study selection: (1)articles evaluated the association between the APE1Asp148Glu polymorphism and breast cancer susceptibility;(2) studies designed as case–control; (3) sufficient data avail-able to estimate an odds ratio (OR) with its 95 % confidenceinterval (95 % CI).

Data extraction

Information was extracted carefully from all eligible publicationsindependently by two investigators according to the inclusioncriteria listed above, discrepancies were adjudicated by otherinvestigators until consensus was achieved on every item. Thefollowing data were collected from each study: the first author’sname, year of publication, country or region, source of publica-tion, total numbers of cases and controls, genotyping results ofcases and controls. If data from any categorywere not reported inthe primary study, the items were designated “not applicable.”The quality of case–control studies were assessed using theNewcastle-Ottawa Scale (NOS) [9], which based on followingthree subscales: the selection of the study groups (four items), thecomparability of the groups (one item), and the ascertainment ofthe exposure or outcome of interest for case–control or cohortstudies (three items) respectively. A “star system” (ranging from0 to 9) has been developed for assessment. In the present meta-analysis, a study awarded seven or more stars was considered asa high-quality study.

Fig. 1 Study selection processfor the meta-analysis

Tumor Biol.

Statistical methods

Hardy–Weinberg equilibrium in controls was calculated in ourmeta-analysis. The chi-squared goodness-of-fit test was usedto test deviation from Hardy–Weinberg equilibrium (HWE;p<0.05 was considered significant). Crude ORs with 95 %CIs were used to assess the strength of association between theAPE1 Asp148Glu polymorphism and breast cancer suscepti-bility. Although fixed effect model and random effect modelyielded similar conclusions, we chose to use the random effectmodel with Mantel–Haenszel statistics, which assumed thatthe true underlying effect varied among included individualsand the random effect model was a more natural choice thanfixed effect model in medical decision-making contexts [10,11]. The pooled ORs were performed on the codominant (GGvs TT; TG vs TT), the dominant (GG+TG vs TT), and therecessive models (GG vs TG+TT) respectively. Subgroupanalyses were performed by ethnicity and by source of con-trol. Heterogeneity assumptions among studies were checkedby the Chi-square test based on Q statistic (p<0.05 indicatedheterogeneity) [12] and I2 test statistic [13], Venice criteria[14] for I2 test: I2<25 % represents no heterogeneity, I2=25–50 % represents moderate heterogeneity, I2=50–75 % repre-sents significant heterogeneity, I2>75 % represents extremeheterogeneity. An estimate of potential publication bias wascarried out by the funnel plot, in which the standard error of

log (OR) of each study was plotted against its log (OR), thesymmetry of funnel plot was also observed with the method ofBegg’s test [15] and Egger’s test [16](p≥0.05 suggests nobias). All statistical analyses were performed using Statastatistical software (version 10.0). Two-sided p values lessthan 0.05 were considered statistically significant.

Results

Characteristics of studies

The study selection process was shown in Fig. 1, by searchingthe databases; a total of 7 eligible studies [17–23] involving3,460 cases and 3,909 controls were included in the meta-analysis. Table 1 presented the main characteristics of thesestudies, one study [21] listed two separate data, all of the caseswere histologically confirmed, the controls were primarilyhealthy population. There were four groups of Asians, fourgroups of non-Asians; three groups of population based,five groups of hospital based. The polymorphism in thecontrol subjects were in HWE except one study [17].Studies with control not in HWE were also consideredfor meta-analysis, but they were excluded in the sensi-tivity analysis [24].

Table 1 Characteristics of case–control studies included in the meta-analysis

Study Year Country Gene test Source No. of Case/Control Case Control HWE NOS

Asp148Glu TT TG GG TT TG GG

Kim KY 2013 Asian GoldenGate® Assay kit HB 346/361 102 184 60 131 153 77 0.01 8

Kang H 2013 Asian TaqMan assay HB 465/799 165 211 89 276 372 151 0.20 8

Jelonek K 2010 non-Asian PCR-RFLP PB 91/412 16 50 25 90 223 99 0.09 8

Sangrajrang S 2008 Asian Qiagen assay HB 507/425 250 208 49 194 176 55 0.13 7

Smith TR 2008 non-Asian MassARRAY system HB 319/405 103 140 76 104 209 92 0.51 8

Smith TR 2008 non-Asian MassARRAY system HB 53/75 23 22 8 30 33 12 0.57 8

Mitra AK 2008 Asian PCR-RFLP PB 150/225 62 75 13 144 72 9 1.00 7

Zhang Y 2006 non-Asian TaqMan assay PB 1,529/1,207 404 752 373 327 590 290 0.46 8

HB hospital based, PB population based, HWE Hardy–Weinberg equilibrium, PCR Polymerase chain reaction, RFLP Restriction fragment lengthpolymorphism, NOSNewcastle-Ottawa Scale

Table 2 Main results of pooled odds ratios (OR) with confidence interval (CI) in the meta-analysis in overall population

Polymorphism Genetic model Genetic type Heterogeneity test OR (95 % CI) P1 Begg's test Egger's test

Q I2 (%) PH Z P2 t P3

Asp148Glu Codominant model GG vs TT 11.76 40.50 % 0.11 1.00 (0.87~1.14) 0.95 0.12 0.90 0.75 0.48

TG vs TT 27.25 74.30 % 0.00 1.06 (0.95~1.18) 0.32 0.12 0.90 0.64 0.54

Dominant model TG+GG vs TT 27.67 74.70 % 0.00 1.04 (0.94~1.16) 0.41 0.12 0.90 0.73 0.49

Recessive model GG vs TT+TG 8.32 15.80 % 0.31 0.99 (0.88~1.11) 0.86 -0.12 1.00 0.39 0.71

PH value for heterogeneity, P1 value for OR, P2 value for Begg’s test, P3 value for Egger’s test, OROdds ratio, CIConfidence interval

Tumor Biol.

Main meta-analysis results

The main results of this meta-analysis were presented in Table 2.Overall, no significant associations were found between theAPE1Asp148Glu polymorphism and breast cancer susceptibility

for GG vs TT (OR=1.00, 95 % CI=0.87–1.14, p=0.11 forheterogeneity, Fig. 2a); TG vs TT (OR=1.06, 95 % CI=0.95–1.18, p=0.00 for heterogeneity, Fig. 2b); the dominant modelGG+TG vs TT (OR=1.04, 95 % CI=0.94–1.16, p=0.00 forheterogeneity, Fig. 2c) and the recessive model GG vs TG+TT

Fig. 2 Odds ratios (ORs) for associations between APE1Asp148Glu polymorphism and breast cancer susceptibility

Tumor Biol.

(OR=0.99, 95 % CI=0.88–1.11, p=0.31 for heterogeneity,Fig. 2d).

In the subgroup analysis by ethnicity and by source ofcontrol, we found a significant association for TG vs TT in

Asian subgroup (OR=1.17, 95 % CI=1.00~1.36) and inpopulation-based subgroup (OR=1.18, 95 % CI=1.00~1.38), respectively. No significant associations were found inthe other models in the subgroups.

Fig. 2 (continued)

Tumor Biol.

Heterogeneity and sensitivity analysis

Sensitivity analysis was performed according to heterogeneity.We found heterogeneity for TG vs TT (p=0.00) and the domi-nant model (p=0.00) of APE1 Asp148Glu polymorphism inoverall population, in the stratified analysis by ethnicity andsource of control, no heterogeneity was found in non-Asiansubgroup (Table 2).

Publication bias

Begg’s funnel plot and Egger’s test were performed to evalu-ate the publication bias of literatures. The funnel plots forpublication bias showed symmetry for GG vs TT(Fig. 3a)and the dominant model (Fig. 3b). Furthermore, theEgger's test was not statistically significant for publicationbias in all models in overall population (Tables 2 and 3).Thus, there was no obvious publication bias in this meta-analysis.

Discussion

Apurinic/apyrimidinic (AP) sites are common mutagenic andcytotoxic DNA lesions which are caused by the loss of normalbases [25]. The redox domain of APE1 is involved in theactivation of redox-sensitive transcription factors, includingp53, nuclear factor kappa B, activator protein-1 (AP-1),

cAMP response element binding protein (CREB), hypoxia-inducible factor-1 alpha (HIF-1a), and others. Throughtranscription factor activation, APE1 may have cytoprotectiveand angiogenic roles in response to a variety of cellularstresses, including hypoxia and tumor necrosis factor [26,27]. Alteration in the expression of APE1 may influence itscapacity to repair DNA damage. Evidence have showed thatpolymorphic variants of the APE1may predispose to cancer,several lines of evidence support that the APE1 Asp148Glu(−656 T>G) polymorphism plays a role in influencing thepromoter activity of APE1 protein [28].

To date, numerous studies have been carried out to inves-tigate whether polymorphisms inAPE1are associated with thesusceptibility of breast cancer; however, the data have yieldedconflicting results. In the present study, to derive a moreprecise estimation of the relationship, we performed a meta-analysis of seven published studies including 3,460 cases and3,909 controls for the APE1 Asp148Glu polymorphism andbreast cancer susceptibility. Our meta-analysis indicated thatthe APE1 Asp148Glu polymorphism was significantly asso-ciated with the susceptibility to breast cancer in Asianpopulation.

There is an obvious heterogeneity of studies for TG vs TTand the dominant model in the overall population, but whenwe stratified the studies into different ethnic subgroups, theheterogeneity disappeared in non-Asian population. Theseresults suggest that heterogeneity may be partly due to thevariable effects of stratified ethnic subgroups, and some

Table 3 Main results of pooled odds ratios (OR) with confidence interval (CI) in the meta-analysis by ethnicity and by source of control

Polymorphism Subgroup (N) Genetic type Heterogeneity test OR (95 % CI) P1 Begg's test Egger's test

Asp148Glu Q I2 (%) PH Z P2 t P3

Country Asian (4) GG vs TT 9.73 69.20 0.02 0.97 (0.78~1.20) 0.79 0.34 0.73 3.77 0.27

TG vs TT 18.70 84.00 0.00 1.17 (1.00~1.36) 0.05 1.02 0.31 10.92 0.02

TG+GG vs TT 21.09 85.80 0.00 1.12 (0.97~1.29) 0.12 1.02 0.31 10.98 0.03

GG vs TT+TG 6.74 55.50 0.08 0.91 (0.75~1.10) 0.34 −0.34 1.00 2.08 0.48

Non-Asian (4) GG vs TT 1.94 0.00 0.59 1.01 (0.85~1.21) 0.88 −0.34 1.00 −0.01 0.99

TG vs TT 5.35 43.90 0.15 0.96 (0.82~1.11) 0.57 0.34 0.73 −0.49 0.79

TG+GG vs TT 4.77 37.00 0.19 0.97 (0.84~1.12) 0.71 0.34 0.73 −0.36 0.83

GG vs TT+TG 0.40 0.00 0.94 1.04 (0.90~1.21) 0.61 0.34 0.73 0.22 0.63

Source HB (5) GG vs TT 2.10 0.00 0.72 0.89 (0.73~1.07) 0.21 0.73 0.46 −0.49 0.72

TG vs TT 11.78 66.10 0.02 0.97 (0.84~1.12) 0.66 0.24 0.81 −0.19 0.95

TG+GG vs TT 8.30 51.80 0.08 0.94 (0.82~1.08) 0.40 0.24 0.81 −0.15 0.96

GG vs TT+TG 3.31 0.00 0.51 0.91 (0.77~1.08) 0.29 0.73 0.46 −0.74 0.65

PB (3) GG vs TT 6.61 69.70 0.04 1.13 (0.93~1.38) 0.23 1.04 0.30 2.43 0.27

TG vs TT 12.40 83.90 0.00 1.18 (1.00~1.38) 0.05 0.00 1.00 3.02 0.50

TG+GG vs TT 14.75 86.40 0.00 1.18 (1.02~1.38) 0.03 0.00 1.00 3.34 0.48

GG vs TT+TG 3.32 39.70 1.19 1.07 (0.91~1.26) 0.44 1.04 0.30 1.75 0.24

HBHospital based, PB Population based, N for numbers of studies, PH value for heterogeneity, P1 value for OR, P2 value for Begg’s test, P3 value forEgger’s test, OROdds ratio, CIConfidence interval

Tumor Biol.

genetic polymorphisms may be associated with risk of somediseases in a specific ethnic subgroup.

Meta-analysis has been recognized as an effective methodto solve a wide variety of clinical questions by summarizingand reviewing the previously published quantitative research.By using meta-analysis, a multitude of genetic polymor-phisms have been associated with specific disease states.CHEK2 I157T [29], ERBB2 655 Val [30], and ESR1rs2046210 [31] polymorphisms have been proved asso-ciated with breast cancer susceptibility by means ofmeta-analysis.

Some limitations of our meta-analysis should be consid-ered in interpreting the results. First, a common limitation ofmeta-analysis was heterogeneity; heterogeneity was oftencaused by variation in the environmental and genetic

background of study participants, which was unavoidablewhen combing many studies. We found evidence ofstudy heterogeneity in our study, presumably due toethnicity and the small number of included studies.Second, in the subgroup analysis, the number of eachsubgroup was relatively small, not having enough sta-tistical power to explore the real association. Third, theresults were based on unadjusted estimates, there wouldbe a more precise estimation on the associations of theAPE1 Asn148Glu polymorphism with breast cancer sus-ceptibility if the ORs were adjusted for menarche, men-opausal status, body mass index, family history, andother environmental factors.

In conclusion, this meta-analysis suggested that the APE1Asn148Glu polymorphism was a risk factor for breast cancer

Fig. 3 Publication bias in studiesof the relation between APE1Asp148Glu polymorphism andbreast cancer susceptibility. Afunnel plot with pseudo-95 %confidence limits (dashed lines)was used

Tumor Biol.

susceptibility in Asian population. Furthermore, studies withlarger numbers of unbiased-matched patients and well-matched controls are required to validate our results.

Acknowledgments This study was supported by Natural ScienceFoundation of The People’s Republic of China (no. 81072175;81102010; 81202096; 81372854), Shanghai Science and TechnologyCommittee (no. 114119a7500, no. 06DZ19505, and no.13NM1401504),and Changhai hospital 1255 discipline construction projects (No.CH125530400). The funders had no role in study design, data collectionand analysis, decision to publish, or preparation of the manuscript.

Conflicts of interest None

References

1. Jemal A, Bray F, Melissa M, Ferlay J, Ward E. Global CancerStatistics. CA Cancer J Clin. 2011;61:69–90.

2. Pharoah PD, Dunning AM, Ponder BA, Easton DF. Associationstudies for finding cancer-susceptibility genetic variants. Nat RevCancer. 2004;4:850–60.

3. Harms C, Salama SA, Sierra-Torres CH, Cajas-Salazar N, Au WW.Polymorphisms in DNA repair genes, chromosome aberrations, andlung cancer. Environ Mol Mutagen. 2004;44:74–82.

4. Almeida KH, Sobol RW. A unified view of base excision repair:lesion-dependent protein complexes regulated by posttranslationalmodification. DNA Repair (Amst). 2007;6:695–711.

5. Wood RD, Mitchell M, Sgouros J, Lindahl T. Human DNA repairgenes. Science. 2001;291:1284–9.

6. Robson CN, Hochhauser D, Craig R, Rack K, Buckle VJ, HicksonID. Structure of the human DNA repair gene HAP1 and itslocalisation to chromosome 14q 11.2-12. Nucleic Acids Res.1992;20:4417–21.

7. Xi T, Jones IM, Mohrenweiser HW. Many amino acid substitutionvariants identified in DNA repair genes during human populationscreenings are predicted to impact protein function. Genomics.2004;83:970–9.

8. Hu JJ, Smith TR, Miller MS, et al. Amino acid substitution variantsof APE1 and XRCC1 genes associated with ionizing radiation sen-sitivity. Carcinogenesis. 2001;22:917–22.

9. Stang A. Critical evaluation of the Newcastle-Ottawa scale for theassessment of the quality of nonrandomized studies in meta-analyses.Eur J Epidemiol. 2010;25:603–5.

10. DerSimonian R, Laird N.Meta-analysis in clinical trials. Control ClinTrials. 1986;7:177–88.

11. Ades AE, Lu G, Higgins JP. The interpretation of random-effectsmeta-analysis in decision models. Med Decis Making. 2005;25:646–54.

12. Cochran WG. The combination of estimates from different experi-ments. Biometrics. 1954;10:101–29.

13. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring incon-sistency in meta-analyses. BMJ. 2003;327:557–60.

14. Ioannidis JP, Boffetta P, Little J, O'Brien TR, Uitterlinden AG, VineisP, et al. Assessment of cumulative evidence on genetic associations:interim guidelines. Int J Epidemiol. 2008;37:120–32.

15. Begg CB, Mazumdar M. Operating characteristics of a rank correla-tion test for publication bias. Biometrics. 1994;50:1088–101.

16. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315:629–34.

17. Kim KY, Han W, Noh DY, Kang D, Kwack K. Impact of geneticpolymorphisms in base excision repair genes on the risk of breastcancer in a Korean population. Gene. 2013;532:192–6.

18. Kang H, Dai Z, Ma X, Ma L, Jin Y, Liu X, et al. A genetic variant inthe promoter of APE1 gene (-656 T>G) is associated with breastcancer risk and progression in a Chinese population. Gene. 2013;531:97–100.

19. Jelonek K, Gdowicz-Klosok A, Pietrowska M, Borkowska M,Korfanty J, Rzeszowska-Wolny J, et al. Association betweensingle-nucleotide polymorphisms of selected genes involved in theresponse to DNA damage and risk of colon, head and neck, andbreast cancers in a Polish population. J Appl Genet. 2010;51:343–52.

20. Sangrajrang S, Schmezer P, Burkholder I, Waas P, Boffetta P,Brennan P, et al. Polymorphisms in three base excision repair genesand breast cancer risk in Thai women. Breast Cancer Res Treat.2008;111:279–88.

21. Smith TR, Levine EA, Freimanis RI, Akman SA, Allen GO, HoangKN, et al. Polygenic model of DNA repair genetic polymorphisms inhuman breast cancer risk. Carcinogenesis. 2008;29:2132–8.

22. Mitra AK, Singh N, Singh A, Garg VK, Agarwal A, SharmaM, et al.Association of polymorphisms in base excision repair genes with therisk of breast cancer: a case-control study in North Indian women.Oncol Res. 2008;17:127–35.

23. Zhang Y, Newcomb PA, Egan KM, Titus-Ernstoff L, Chanock S,Welch R, et al. Genetic polymorphisms in base-excision repair path-way genes and risk of breast cancer. Cancer Epidemiol BiomarkersPrev. 2006;15:353–8.

24. Minelli C, Thompson JR, Abrams KR, Thakkinstian A, AttiaJ. How should we use information about HWE in the meta-analyses of genetic association studies? Int J Epidemiol. 2008;37:136–46.

25. Lindahl T. Recognition and processing of damaged DNA. J Cell SciSuppl. 1995;19:73–7.

26. Hall JL, Wang X, Van A, Zhao Y, Gibbons GH. Overexpression ofRef-1 inhibits hypoxia and tumor necrosis factor-induced endothelialcell apoptosis through nuclear factor-kappab-independent and -dependent pathways. Circ Res. 2001;88:1247–53.

27. Izumi T, Hazra TK, Boldogh I, Tomkinson AE, Park MS, Ikeda S,et al. Requirement for human AP endonuclease 1 for repair of 3’-blocking damage at DNA single-strand breaks induced by reactiveoxygen species. Carcinogenesis. 2000;21:1329–34.

28. Li C, Liu Z, Wang LE, Strom SS, Lee JE, Gershenwald JE, et al.Genetic variants of the ADPRT, XRCC1 and APE1 genes and risk ofcutaneous melanoma. Carcinogenesis. 2006;27:1894–901.

29. Liu C, Wang Y, Wang QS, Wang YJ. The CHEK2 I157T variant andbreast cancer susceptibility: a systematic review and meta-analysis.Asian Pac J Cancer Prev. 2012;13:1355–60.

30. Wang H, Liu L, Lang Z, Guo S, Gong H, Guan H, et al.Polymorphisms of ERBB2 and breast cancer risk: a meta-analysisof 26 studies involving 35,088 subjects. J Surg Oncol. 2013;108:337–41.

31. Guo H, Ming J, Liu C, Li Z, Zhang N, Cheng H, et al. A commonpolymorphism near the ESR1 gene is associated with risk of breastcancer: evidence from a case-control study and ameta-analysis. PLoSOne. 2012;7:e52445.

Tumor Biol.