TR-05294; No of Pages 6 Thrombosis Research xxx (2013) xxx–xxx

Contents lists available at ScienceDirect

Thrombosis Research journal homepage: www.elsevier.com/locate/thromres

Regular Article

Synergistic association of DNA repair relevant gene polymorphisms with the risk of coronary artery disease in northeastern Han Chinese Xiaohong Yu a,b, Jun Liu b, Hao Zhu b, Yunlong Xia b, Lianjun Gao b, Yingxue Dong b, Nan Jia c, Weifeng Shen d, Yanzong Yang b,⁎, Wenquan Niu e,f,⁎⁎ a

Dalian Medical University, Dalian, Liaoning, China Department of Cardiology, The First Affiliated Hospital of Dalian Medical University, Dalian, Liaoning, China Department of Hypertension, Ruijin Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China d Department of Cardiology, Ruijin Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China e State Key Laboratory of Medical Genomics, Ruijin Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China f Shanghai Institute of Hypertension, Shanghai, China b c

a r t i c l e

i n f o

Article history: Received 9 August 2013 Received in revised form 7 November 2013 Accepted 19 November 2013 Available online xxxx Keywords: Coronary artery disease DNA repair Gene-to-gene interaction Polymorphism

a b s t r a c t Evidence is mounting suggesting that DNA damage is implicated in the development and progression of atherosclerosis. To yield more information, we focused on six well-characterized polymorphisms from four DNA repairrelevant candidate genes, viz. XRCC1 (rs1799782 and rs25487), XRCC3 (rs861539), MTHFR (rs1801133 and rs4846049), and NQO1 (rs1800566), to identify and characterize their potential gene-to-gene interactions in susceptibility to coronary artery disease (CAD) in Han Chinese. This was a hospital-based case-control study involving 1142 patients diagnosed with CAD and 1106 age- and gender-matched controls. All participants were angiographically confirmed. Risk estimates were expressed as odds ratio (OR) and 95% confidence interval (95% CI). All six examined polymorphisms met Hardy-Weinberg equilibrium. Overall there were significant differences in the genotype/allele distributions of MTHFR gene rs1801133 and rs4846049 (both P ≤ 0.005), and in the genotype distributions of XRCC1 gene rs1799782 (P = 0.002) between patients and controls. The adjusted risk of having CAD was more evident for rs1799782 (OR = 1.53; 95% CI: 1.16-2.02; P = 0.003), rs1801133 (OR = 1.54; 95% CI: 1.22-1.94; P b 0.001), and rs4846049 (OR = 1.74; 95% CI: 1.13-2.69; P = 0.013) under the recessive model. Interaction analyses indicated that the overall best multifactor dimensionality reduction (MDR) model included rs4846049, rs1801133, and rs1799782, and this model had a maximal testing accuracy of 0.6885 and a cross-validation consistency of 10 out of 10 (P = 0.0030). Further interaction entropy graph bore out the validity of this MDR model. Taken together, our findings demonstrate a contributory role of genetic defects in XRCC1 and MTHFR genes, both individually and interactively, in the development of CAD in Han Chinese. © 2013 Elsevier Ltd. All rights reserved.

Introduction During the past few years, genome-wide association studies that produce a vast quantity of information on common genetic variants and coronary artery disease (CAD) have poured into the literature like an avalanche; however, there is still a substantial proportion of unidentified common variants that exert a strong impact on disease or phenotype susceptibility [1,2]. The most compelling reason might be attributable to the overlook of epistasis, which is defined as one locus masking the effect of another [3]. Given the totality of available data, it

⁎ Correspondence to: Y. Yang, Zhongshan Road 222, Dalian 116011, Liaoning, China. Tel./fax: +86 411 83632383. ⁎⁎ Correspondence to: W. Niu, Ruijin Second Road 197, Shanghai 200025, China. Tel.: +86 21 64370045x610905; fax: +86 21 64333548. E-mail addresses: [email protected], [email protected] (Y. Yang), [email protected] (W. Niu).

is a high priority to examine the interaction of multiple candidate genes, which may enhance our understandings of the pathogenesis of CAD. Evidence is mounting suggesting that DNA damage induced by oxidative stress is implicated in the development and progression of atherosclerosis [4–6]. Genetic defects in DNA repair-relevant genes may therefore be logical candidates accounting for the divergences in repair efficiency of DNA damage, which will eventually precipitate the occurrence of cardiovascular events [7,8]. To better understand genetic epistasis, we, in this study, focused on six well-characterized polymorphisms from four DNA repair-relevant candidate genes, viz. XRCC1 (X-ray cross-complementing group 1: rs1799782 and rs25487), XRCC3 (X-ray cross-complementing group 3: rs861539), MTHFR (methylenetetrahydrofolate reductase: rs1801133 and rs4846049), and NQO1 (quinone oxidoreductases (NAD(P)H): quinone oxidoreductase 1: rs1800566), to identify and characterize their potential gene-to-gene interactions in susceptibility to CAD in a large Han Chinese population from

0049-3848/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.thromres.2013.11.017

Please cite this article as: Yu X, et al, Synergistic association of DNA repair relevant gene polymorphisms with the risk of coronary artery disease in northeastern Han Chinese, Thromb Res (2013), http://dx.doi.org/10.1016/j.thromres.2013.11.017

2

X. Yu et al. / Thrombosis Research xxx (2013) xxx–xxx

northeastern China. All study participants are characterized by genetic homogeneity and geographic stability, and they are probably more uniform in their environmental exposures, such as the habitual dietary intake of high salt and high fat. Methods Study population In total, 2248 participants were recruited from the Department of Cardiology, the First Affiliated Hospital of Dalian Medical University, and they were local residents of Han descent in Dalian city, Liaoning province, China. All participants were angiographically confirmed, and those with the presence of more than 50% stenosis in at least one of the three major coronary arteries or major branches were treated as the CAD group, and those with normal coronary angiography (less than 50% stenosis in all arteries and branches) were enrolled as the control group. Moreover, patients were excluded if they had simple spasm of coronary arteries, myocardial bridge or other non-coronary atherosclerotic lesions. The controls had no history of any vascular event, and had normal coronary arteries on angiography. According to the angiographic results, there were 1142 patients diagnosed as CAD and 1106 age- and gender-matched controls in the final analysis. All individuals signed written informed consent prior to enrollment. This study was reviewed and approved by the ethics committee of Dalian Medical University, and was conducted in agreement with the guidelines of Declaration of Helsinki. Study characteristics At enrollment, body weight and height were recorded with participants wearing light clothing and no shoes. Body mass index (BMI) was calculated as body weight in kilograms divided by the square of height in meters. Blood pressure was measured twice with the participants seated with at least 5-minute intervals by certified nurses, and the average values were used for analyses. Diagnosis of hypertension was based on the presence of elevated systolic (≥140 mmHg) and/or diastolic (≥90 mmHg) blood pressure, or current use of antihypertensive medications. Type 2 diabetes was defined as a fasting plasma glucose level ≥ 7.0 mmol/L or taking hypoglycemic drugs or receiving parenteral insulin therapy. Blood samples were drawn from all participants after an overnight fasting of at least 8 hours. Fasting glucose was measured in fluoride plasma by an electrochemical glucose oxidase method. Plasma levels of triglyceride (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), lipoprotein(a), blood urea nitrogen (BUN), creatinine, and uric acid (UA), were determined enzymatically using available kits and auto analyzer (Sangon Biotech, Shanghai, China). Plasma high sensitivity C-reactive protein (hsCRP) levels were determined using a highsensitivity ELISA kit (Sangon Biotech, Shanghai, China). Genotyping Genomic DNA was isolated from the white blood cells by using the phenol/chloroform extraction method, and was stored at −40 °C until required for batch genotyping. Plasma was prepared for quantifying routine biological profiles. Genotypes of all polymorphisms examined were determined by the polymerase chain reaction - ligase detection reaction (PCR-LDR) method as previously described [9]. The primers for PCR amplification and the probes for LDR were synthesized by Shanghai Generay Biotech Co. Ltd. (data available upon request). The PCR reactions were performed in the EDC-810 Amplifier (Dongsheng Innovation Biotech Co., Ltd., Beijing China). For each polymorphism, two specific probes were synthesized to discriminate specific bases, and one common probe was labeled by 6carboxy-fluorescein (FAM) at the 3' end and by phosphorylated at the

5' end. The multiplex ligation reaction was carried out in a reaction volume of 10 μl containing 2 μl of PCR product, 1 μl 10 × Taq DNA ligase buffer, 1 μM of each discriminating probe, 5 U Taq DNA ligase, and the ligation parameters were 30 cycles of 94 °C for 30 seconds and 56 °C for 3 minutes. After reaction, 1 μl LDR reaction product was mixed with 1 μl ROX passive reference and 1 μl loading buffer, and then denatured at 95 °C for 3 minutes, and chilled rapidly in ice water. The fluorescent products of LDR were differentiated using the ABI 3730XL sequencer (Applied Biosystems, California, USA). Statistical analysis The Pearson χ2 and unpaired Student's t-test or Mann-Whitney U test were used to compare the differences between CAD patients and controls in terms of categorical (including genotype and allele distributions) and continuous variables, respectively. Testing for deviations from Hardy-Weinberg equilibrium was carried out using a Pearson goodness-of-fit test. Two-tailed P b 0.05 was accepted as statistical significance. Each genotype of the examined polymorphisms was regressed in a Logistic model under the assumptions of additive (major homozygotes versus heterozygotes versus minor homozygotes), dominant (major homozygotes versus heterozygotes plus minor homozygotes) and recessive (major homozygotes plus heterozygotes versus minor homozygotes) models of inheritance after adjusting for age, gender, BMI, systolic blood pressure, and fasting glucose. The linkage disequilibrium patterns were identified separately by Haploview (version 4.0). Interactive analysis was conducted by the open-source multifactor dimensionality reduction (MDR) software (version 2.0) (www. epistasis.org) [10,11]. All possible combinations of six examined polymorphisms were constructed using MDR constructive induction. Then a Bayes classifier in the context of 10-fold cross-validation was employed to estimate the testing accuracy of each best model. A single best model simultaneously had maximal testing accuracy and crossvalidation consistency, and the latter was a measure of the number of times of 10 divisions of the data that the best model was extracted. Statistical significance was evaluated using a 1000-fold permutation test to compare observed testing accuracies with those expected under the null hypothesis of null association. Permutation testing corrected for multiple testing by repeating the entire analysis on 1000 datasets that were consistent with the null hypothesis. Further to test the validity of MDR method, a classical Logistic regression analysis was undertaken for the extracted best model. To visually inspect the potential genetic epistasis of examined polymorphisms, interaction entropy graph, implemented in the MDR software (version 2.0), was represented to quantify the synergistic and non-synergistic interactions among polymorphisms. In this graph, information gain values expressed as percentages in the nodes signify the independent main effect of the polymorphisms. The positive and negative information gain values (percentages) on the connected lines indicate synergistic interaction (the red or the orange line) and redundancy (lack of interaction, the green and the blue line) between the polymorphisms, respectively with the zero value indicating independence (the yellow line). Statistical analyses were conducted by STATA software (version 11.0) for Windows (StataCorp LP, College Station, TX, USA). Study power was estimated by the Power and Sample Size Calculations (PS) software (v3.0.7) Results Baseline characteristics Table 1 shows the distributions and comparisons of study characteristics between CAD group (n = 1142) and control group (n = 1106). The distributions of age and gender did not differ significantly between

Please cite this article as: Yu X, et al, Synergistic association of DNA repair relevant gene polymorphisms with the risk of coronary artery disease in northeastern Han Chinese, Thromb Res (2013), http://dx.doi.org/10.1016/j.thromres.2013.11.017

X. Yu et al. / Thrombosis Research xxx (2013) xxx–xxx Table 1 Baseline characteristics of study population. Characteristics

CAD patients (n = 1142)

Controls (n = 1106)

P

Age, years Gender (Males) BMI, kg/m2 SBP, mmHg DBP, mmHg Fasting glucose, mmol/L Triglycerides, mmol/L Total cholesterol, mmol/L HDL-C, mmol/L Lipoprotein(a), mmol/L BUN, mmol/L Creatinine, μmol/L Uric acid, μmol/L hsCRP, mmol/L

62.07 ± 46.58% 26.19 ± 141.44 ± 84.86 ± 6.14 ± 1.9 ± 4.59 ± 1.12 ± 0.3 ± 5.92 ± 87.49 ± 329.06 ± 12.37 ±

62.42 ± 49.10% 24.9 ± 137.31 ± 81.09 ± 5.47 ± 1.92 ± 4.81 ± 1.35 ± 0.21 ± 5.76 ± 81.35 ± 328.85 ± 2.21 ±

0.3749 0.2334 0.0637 b0.0005 b0.0005 b0.0005 0.7240 b0.0005 b0.0005 b0.0005 0.3794 0.0006 0.9644 b0.0005

9.07 15.32 16.82 10.63 2.15 1.04 1.18 0.32 0.45 3.89 36.81 100.37 11.42

9.85 3.64 20.52 11.92 1.26 1.45 1.0 0.4 0.19 3.71 35.96 92.52 3.71

Abbreviations: CAD, coronary artery disease; BMI, body mass index; HDL-C, high-density lipoprotein cholesterol; BUN, blood urea nitrogen; hsCRP, high sensitivity C-reactive protein. Data are expressed as mean ± standard deviation unless otherwise indicated.

the two groups. CAD patients had relatively higher BMI levels than controls (P = 0.0637). Blood pressures and fasting glucose levels were strikingly higher in patients than in controls (P b 0.0005). Plasma total cholesterol and HDL-C levels were significantly lower in patients than in controls (P b 0.0005). In contrast, plasma lipoprotein(a) (P b 0.0005), creatinine (P = 0.0006) and hsCRP (P b 0.0005) levels were significantly higher in patients than in controls. There were no significant differences for BUN and uric acid. Overall single-locus analyses The genotype distributions and allele frequencies of six examined polymorphisms are compared between CAD group and control group in Table 2. All genotypes conformed to the Hardy-Weinberg equilibrium expectation in each group (P N 0.05). There were significant differences in the genotype and allele distributions between patients and controls for polymorphisms rs1801133 and rs4846049 in MTHFR gene (P ≤ 0.005), and in the genotype distributions for polymorphism rs1799782 (P = 0.002) in XRCC1 gene. The power to reject the null hypothesis of no difference in significant allele frequencies for MTHFR gene rs1801133 and rs4846049 polymorphisms between patients and controls was 82.0% and 86.2%, respectively. No significance was observed for the genotypes and alleles of the other polymorphisms examined. To explore the potential inheritance patterns, the adjusted risk estimates were calculated for all polymorphisms under additive, dominant, and recessive models, respectively (Table 2). After correcting for age, gender, BMI, systolic blood pressure, and fasting glucose, the risk of developing CAD was more evident for XRCC1 gene rs1799782 under the recessive model (OR = 1.53; 95% CI: 1.16-2.02; P = 0.003), and for MTHFR gene rs1801133 under both additive (OR = 1.19; 95% CI: 1.06-1.34; P = 0.003) and recessive (OR = 1.54; 95% CI: 1.22-1.94; P b 0.001) models, and for MTHFR gene rs4846049 for all three genetic models, especially under the recessive model (OR = 1.74; 95% CI: 1.132.69; P = 0.013). Linkage analysis indicated that the linkage extent between MTHFR gene rs1801133 and rs4846049 was very weak (D’ = 0.046 in all participants; D’ = 0.043 in CAD patients; D’ = 0.074 in controls). Subgroup analyses The genotype distributions of six examined polymorphisms by age, BMI, hypertension, and diabetes between patients and controls are presented in Table 3. For rs1799782, genotype distributions differed significantly between CAD patients and controls irrespectively of grouping age by 79 years (both P = 0.015), and in participants of obesity (BMI ≥ 25 kg/m2) (P = 0.002), hypertension (P = 0.019), and normal

3

Table 2 The genotype distributions and allele frequencies of six examined polymorphisms between patients and controls, and their prediction for CAD. Controls Pχ2 CAD (n = 1106) patients (n = 1142)

Gene: polymorphism XRCC1: rs1799782 Genotype (n): Allele (%): XRCC1: rs25487 Genotype (n): Allele (%): XRCC3: rs861539 Genotype (n): Allele (%): MTHFR: rs1801133 Genotype (n): Allele (%): MTHFR: rs4846049 Genotype (n): Allele (%): NQO1: rs1800566 Genotype (n): Allele (%):

CC

517

CT 486 TT 139 T 33.45 GG 625 GA 437 AA 80 A 26.14 CC 1026 CT 112 TT 4 T 5.25 CC 401

483

OR; 95% CI; P⁎

1.05; 0.93-1.2; 0.416

531 92 32.32 627 419 60 24.37 1000 101 5 5.02 422

0.002 0.94; 0.79-1.11; 0.445 1.53; 1.16-2.02; 0.003 0.422 1.1; 0.96-1.26; 0.17 0.264 1.08; 0.92-1.28; 0.349 1.31; 0.93-1.86; 0.122 0.172 1.05; 0.81-1.36; 0.724 0.804 1.07; 0.81-1.41; 0.649 0.77; 0.21-2.89; 0.703 0.72 1.19; 1.06-1.34; 0.003

CT TT T GG

523 209 41.53 673

542 142 37.34 711

0.002 1.13; 0.95-1.34; 0.175 1.54; 1.22-1.94; b0.001 0.004 1.26; 1.09-1.46; 0.002

GT TT T CC CT TT T

411 58 23.07 346 547 249 45.75

362 33 19.35 324 535 247 46.52

0.005 1.25; 1.06-1.49; 0.009 1.74; 1.13-2.69; 0.013 0.002 0.97; 0.86-1.09; 0.612 0.866 0.95; 0.8-1.14; 0.603 0.97; 0.79-1.18; 0.762 0.607

Abbreviations: CAD, coronary artery disease; OR, odds ratio; 95% CI, 95% confidence interval. *OR, 95% CI, and P values were calculated under the additive (the upper), dominant (the middle), and recessive (the lower) models of inheritance after adjusting for age, gender, body mass index, systolic blood pressure, and fasting glucose.

glucose (P = 0.001), respectively. For rs1801133, significance was preserved in two age groups (P b 0.01) and obese participants (P = 0.006), as well as in participants of normal blood pressure (p = 0.004) and glucose (p = 0.002). For rs4846049, there were significant genotype differences between patients and controls in participants aged less than 70 years old (P = 0.003), in obese (P b 0.001) and hypertensive (P = 0.020) participants, and in participants of normal glucose (P = 0.035).

Gene-to-gene interaction analyses To explore the potential gene-to-gene interactions of six polymorphisms from four candidate genes, an exhaustive MDR analysis that evaluated all possible combinations of six polymorphisms is summarized in Table 4. Here, each best model was accompanied with the testing accuracy, cross-validation consistency and significant level as determined by permutation testing. The overall best MDR model included polymorphisms rs4846049, rs1801133, and rs1799782 with a maximal testing accuracy of 0.6885 and a cross-validation consistency of 10 out of 10. This model was significant at the level of 0.0030, indicating that a model this good or better was observed only by 3 out of 1000 permutations, and was thus unlikely under the null hypothesis of null association. In addition, to test the soundness of derived MDR method, a classical Logistic model that incorporated the product of three polymorphisms of the overall best MDR model, as well as confounding factors including age, gender, BMI, systolic blood pressure, and fasting glucose was undertaken. The interaction of three polymorphisms was found to be associated with 1.13-fold (95% CI: 1.07-1.20; P b 0.001) increased risk of having CAD under the additive model. Considering the fact that genotype differences of three polymorphisms in overall best MDR model differed significantly in obese participants, additional interaction analyses

Please cite this article as: Yu X, et al, Synergistic association of DNA repair relevant gene polymorphisms with the risk of coronary artery disease in northeastern Han Chinese, Thromb Res (2013), http://dx.doi.org/10.1016/j.thromres.2013.11.017

4

X. Yu et al. / Thrombosis Research xxx (2013) xxx–xxx

Table 3 Subgroup analyses of the genotypes of six examined polymorphisms between patients and controls by age, BMI, hypertension, and diabetes. Polymorphism

Genotype

Subgroup

Sample size⁎

CAD patients

Controls

Pχ2

rs1799782

CC/CT/TT

Age ≥ 70 years Age b 70 years BMI ≥ 25 kg/m2† BMI b 25 kg/m2 Hypertension Normotension Diabetes⁎⁎

rs25487

GG/G/A/AA

rs861539

CC/CT/TT

rs1801133

CC/CT/TT

rs4846049

GG/GT/TT

rs1800566

CC/CT/TT

280/311 862/795 994/814 148/292 801/847 341/259 554/226 588/880 280/311 862/795 994/814 148/292 801/847 341/259 554/226 588/880 280/311 862/795 994/814 148/292 801/847 341/259 554/226 588/880 280/311 862/795 994/814 148/292 801/847 341/259 554/226 588/880 280/311 862/795 994/814 148/292 801/847 341/259 554/226 588/880 280/311 862/795 994/814 148/292 801/847 341/259 554/226 588/880

111/129/40 406/357/99 451/423/120 66/63/19 355/349/97 162/137/42 254/241/59 263/245/80 154/105/21 471/332/59 544/382/68 81/55/12 435/305/61 190/132/19 300/214/40 325/223/40 248/30/2 778/82/2 891/99/4 135/13/0 715/83/3 311/29/1 493/60/1 533/52/3 82/149/49 319/383/160 346/466/182 55/66/27 278/385/138 123/147/71 198/260/96 203/272/113 173/96/11 500/315/47 585/356/53 88/55/5 477/284/40 196/127/18 334/191/29 339/220/29 96/124/60 250/423/189 301/480/213 45/67/36 245/385/171 101/162/78 156/273/125 190/274/124

140/149/22 343/382/70 356/395/63 127/136/29 365/411/71 118/120/21 100/106/20 383/425/72 177/124/10 450/295/50 453/314/47 174/105/13 482/319/46 145/100/14 127/83/16 500/336/44 282/27/2 718/74/3 735/74/5 265/27/0 762/82/3 238/19/2 203/23/0 797/78/5 129/142/40 293/400/102 307/403/104 115/139/38 324/409/114 98/133/28 85/108/33 337/434/109 191/112/8 520/250/25 531/267/16 180/95/17 551/271/25 160/91/8 149/73/4 562/289/29 83/153/75 241/382/172 239/389/186 85/146/61 245/412/189 79/122/58 51/122/53 273/413/194

0.015 0.015 0.002 0.567 0.019 0.142 0.600 0.001 0.065 0.706 0.646 0.253 0.170 0.995 0.870 0.340 0.699 0.857 0.678 0.873 0.897 0.627 0.785 0.989 0.007 0.003 0.006 0.344 0.073 0.004 0.637 0.002 0.615 0.003 b0.001 0.393 0.020 0.322 0.061 0.035 0.133 0.841 0.756 0.596 0.741 0.971 0.266 0.842

Euglycemia Age ≥ 70 years Age b 70 years BMI ≥ 25 kg/m2† BMI b 25 kg/m2 Hypertension Normotension Diabetes⁎⁎ Euglycemia Age ≥ 70 years Age b 70 years BMI ≥ 25 kg/m2† BMI b 25 kg/m2 Hypertension Normotension Diabetes⁎⁎ Euglycemia Age ≥ 70 years Age b 70 years BMI ≥ 25 kg/m2† BMI b 25 kg/m2 Hypertension Normotension Diabetes⁎⁎ Euglycemia Age ≥ 70 years Age b 70 years BMI ≥ 25 kg/m2† BMI b 25 kg/m2 Hypertension Normotension Diabetes⁎⁎ Euglycemia Age ≥ 70 years Age b 70 years BMI ≥ 25 kg/m2† BMI b 25 kg/m2 Hypertension Normotension Diabetes⁎⁎ Euglycemia

Abbreviations: BMI, body mass index; CAD, coronary artery disease. *Sample size in CAD patients/Controls. **Diabetes refers to type 2 diabetes. †The selection of 25 kg/m2 as the cutoff of obesity for Asians was according to the suggestions by The World Health Organization Western Pacific Region; The International Association for the Study of Obesity; The International Obesity Task Force (The Asia-Pacific Perspective: Redefining Obesity and Its Treatment. Sydney, Australia: Health Communications Australia Pty Limited; 2000).

were conducted in obese participants, and still the overall best model included the aforementioned three polymorphisms with an improved testing accuracy of 0.6987 (P = 0.0023) (data not shown).

Table 4 Summary of MDR analysis. Best combination of each model

Testing accuracy

Cross-validation consistency

P

rs4846049 rs4846049, rs1801133 rs4846049, rs1801133, rs1799782⁎ rs4846049, rs1801133, rs1799782, rs1800566 rs4846049, rs1801133, rs1799782, rs1800566, rs25487 rs4846049, rs1801133, rs1799782, rs1800566, rs25487, rs861539

0.5805 0.6573 0.6885 0.6675

7 8 10 10

0.1052 0.0306 0.0030⁎ 0.0169

0.6420

9

0.0485

0.6340

9

0.2061

⁎ The overall best MDR model.

The interaction graph of six examined polymorphisms is depicted in Supplementary Fig. S1 according to the entropy measures among individual polymorphisms. The independent main effect of XRCC1 gene rs25487 with the information gain value of 0.51%. However, this effect was diluted by other polymorphisms, as reflected by the negative interaction effect of rs25487 with rs4846049 (information gain value: -0.37%), rs1800566 (-0.36%), rs1801133 (-0.31%), rs1799782 (-0.22%), and rs861539 (-0.22%). The strongest interaction was identified between MTHFR gene rs1801133 and XRCC1 gene rs1799782, which had the information gain value of 0.28%. Moreover, the interaction effect between rs4846049 and rs1801133 in MTHFR gene was also prominent with the information gain value of 0.18. Discussion In this study, we aimed to identify and characterize the potential gene-to-gene interactions of four DNA repair-relevant candidate genes, XRCC1, XRCC3, MTHFR, and NQO1, in susceptibility to CAD in a

Please cite this article as: Yu X, et al, Synergistic association of DNA repair relevant gene polymorphisms with the risk of coronary artery disease in northeastern Han Chinese, Thromb Res (2013), http://dx.doi.org/10.1016/j.thromres.2013.11.017

X. Yu et al. / Thrombosis Research xxx (2013) xxx–xxx

large Han Chinese population involving 2248 participants. The principle finding of this study was that genetic defects in XRCC1 and MTHFR genes might enhance, both individually and interactively, the risk of developing CAD in Han Chinese, especially in obese participants. Although the residual confounding resulting from the incompletely measured or unmeasured physiologic covariates could not be easily eliminated, it seems unlikely that our results could be explained by confounding. XRCC1 is a DNA repair protein that interacts with DNA ligase III, polymerase β, and poly (ADP-ribose) polymerase to participate in the base excision repair pathway [12,13]. Experimental data demonstrated that XRCC1 played a role in limiting the amount of strand displacement and the size of the repair patches [14]. Mice with Xrcc1-/- gene knockout were lethal, and cells lacking XRCC1 gene were hypersensitive to a wide type of genotoxins and exhibited genetic instability in chromosome translocations and deletions [15]. Similarly, MTHFR is a rate-limiting enzyme in folate and homocysteine metabolism, and is involved in the DNA methylation, repair, and synthesis [16]. Mice with Mthfr-/- gene knockout increased the production of plasma homocysteine, altered Sadenosylmethionine levels, and decreased DNA methylation [17]. On the basis of these observations and the findings of this work, it is reasonably expected that the genes encoding XRCC1 and MTHFR, if involved, might act together additively or interactively in the process of DNA damage or repair. There is considerable evidence in favor of the involvement of DNA damage in the pathogenesis and the severity of CAD in a dosedependent manner [18,19]. The cause of CAD is multifactorial; part is due to genetic susceptibility [20]. Although candidate gene approach, which deals primarily with prespecified genes that are thought to participate in the pathophysiology of the disease, may not replace genome-wide association study, it is an important alternative strategy to unveil the genetic underpinnings of complex disease. A more recent study by Narne et al revealed a significant association between XRCC1 gene p.Arg399Gln (rs25487) polymorphism and angiographicallydocumented CAD in Indian type 2 diabetic patients [21]. Another study by Bazo et al in Brazilians indicated that homogeneous carriers of 399Gln had more than two times (2.3) greater risk of developing coronary atherosclerosis [7]. Contrastingly in this study, we failed to produce any evidence regarding the individual susceptibility of XRCC1 gene rs25487 to CAD, even after considering traditional risk factors. However, as reflected in our interaction analyses, the contributory role of XRCC1 gene rs25487 might be diluted or offset by other examined polymorphisms, further reinforcing the existence of gene-to-gene interactions in disease susceptibility. Moreover, a common polymorphism c.C677T (rs1801133) in MTHFR gene has been exhaustively investigated in association with CAD [22–24]; however, the results are often irreproducible. For example, Lewis et al have conducted a comprehensive meta-analysis on 80 studies, and found no strong evidence supporting an association of c.C677T polymorphism with CAD in Europeans, North Americans, or Australians [24]. In contrast, another meta-analysis by Li YY documented that the c.C677T polymorphism TT genotype was associated with significant risk of CAD in Han Chinese [22], consistent with the results of the present study. One possible explanation for these inconsistent findings is genetic heterogeneity across ethnicities, and therefore there is a need to construct a database of CAD-susceptibility genes or polymorphisms in each ethnic group. Another possibility might be due to the impact of external factors in disease susceptibility such as dietary habits, nutritional status between Asians and Caucasians. As reported in a recent study by Heiniger et al, Caucasian participants had higher score for nutrition and medical screening than Chinese and Korean immigrants in Australia [25]. Besides, we also observed that another polymorphism rs4846049 in the 3’-untranslated region of MTHFR gene also exhibited strong associations with CAD. Considering the ubiquity of genetic interactions in the pathogenesis of complex diseases, the identification and characterization of susceptible genes or polymorphisms require a thorough understanding of gene-to-gene interactions [26]. As expected,

5

three of six examined polymorphisms in XRCC1 and MTHFR genes, which were significant in single-locus analyses, constituted the overall best MDR model in association with CAD. MDR is a promising datamining analytical approach, and is nonparametric and genetic modelfree nature in design [27]. Therefore, MDR approach might represent the first step to enhance our understandings of genetic underpinnings of complex diseases. Several strengths distinguishing the present investigation merit adequate consideration. First, this study was designed the potential geneto-gene interactions of DNA repair-relevant genes, and the selection of genetic markers were biologically plausible. Second, all study participants were angiographically confirmed and the genotypes of all examined polymorphisms satisfied Hardy-Weinberg equilibrium, lowering the likelihood of being biased by selection bias and faulty genotyping. Third, this study involved a total of 2248 participants, and the power to detect the effects of significant polymorphisms was more than 80%. Nevertheless, the interpretation of our findings should be viewed in light of several limitations. First, this study was retrospective in design, which precludes further comments on causality. Second, only six wellcharacterized polymorphisms from four candidate genes were analyzed, and it is of added interest to explore more polymorphisms and more genes from CAD-susceptibility pathways, such as matrix metalloproteinase family members and apelin/APJ signaling cascade [28,29]. Third, data on circulating homocysteine levels were not available, which limited further genotype-phenotype analyses. Fourth, the MDR approach employed in this study has some underlying drawbacks, such as computational intensiveness, indistinct interpretation, lack of sensitivity, and heterogeneity-free assumption [27,30]. Fifth, we recruited study individuals aged more than 50 years, and future larger association studies in a young population of CAD patients are of specific interest, as genetic factors may have greater contribution to those suffering premature CAD and in the absence of strong environmental risk factors [31]. Last but not the least, the fact that our study population was of Han Chinese descent limited the generalizability of our findings, calling for further confirmation in other ethnic groups. Despite these limitations, our findings demonstrate that genetic defects in XRCC1 and MTHFR genes might enhance, both individually and interactively, the risk of developing CAD in Han Chinese, especially in obese participants. For practical reasons, large, well-designed longitudinal studies attempting to account for gene-to-gene and gene-toenvironment interactions, as well as studies seeking to provide biological or clinical implications, are warranted in the future research. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.thromres.2013.11.017. Conflict of interest statement None of the authors have any conflict of interest to disclose. Authors’ contributions All authors contributed to study design and interpretation of results. Conceived and designed the experiments: XY, WN, YY. Performed the experiments: XY, JL. Analyzed the data: XY, WN. Contributed materials/analysis tools: HZ, YX, LG, ZL, NJ, WS. Wrote the paper: WN, YY. The manuscript was reviewed and approved by all authors prior to submission. Author disclosures None. Acknowledgements This study was supported by Liaoning Province Science & Technology Plan Project (Grant No. 2013225002), and Liaoning Provincial

Please cite this article as: Yu X, et al, Synergistic association of DNA repair relevant gene polymorphisms with the risk of coronary artery disease in northeastern Han Chinese, Thromb Res (2013), http://dx.doi.org/10.1016/j.thromres.2013.11.017

6

X. Yu et al. / Thrombosis Research xxx (2013) xxx–xxx

Department of Education Scientific Research Project (Grant No. L2012321). References [1] Lu X, Wang L, Chen S, He L, Yang X, Shi Y, et al. Genome-wide association study in Han Chinese identifies four new susceptibility loci for coronary artery disease. Nat Genet 2012;44:890–4. [2] Gianfagna F, Cugino D, Santimone I, Iacoviello L. From candidate gene to genomewide association studies in cardiovascular disease. Thromb Res 2012;129:320–4. [3] Jensen MK, Pers TH, Dworzynski P, Girman CJ, Brunak S, Rimm EB. Protein interaction-based genome-wide analysis of incident coronary heart disease. Circ Cardiovasc Genet 2011;4:549–56. [4] Guven M, Guven GS, Oz E, Ozaydin A, Batar B, Ulutin T, et al. DNA repair gene XRCC1 and XPD polymorphisms and their association with coronary artery disease risks and micronucleus frequency. Heart Vessels 2007;22:355–60. [5] Madamanchi NR, Zhou RH, Vendrov AE, Niu XL, Runge MS. Does oxidative DNA damage cause atherosclerosis and metabolic syndrome?: new insights into which came first: the chicken or the egg. Circ Res 2010;107:940–2. [6] Wang CL, Lin TH, Lin HY, Sheu SH, Yu ML, Hsiao PJ, et al. The 8-oxoguanine glycosylase I (hOGG1) Ser326Cys variant affects the susceptibility to multi-vessel disease in Taiwan coronary artery disease patients. Thromb Res 2010;126:319–23. [7] Bazo AP, Salvadori Jr D, Salvadori RA, Sodre LP, da Silva GN, de Camargo EA, et al. DNA repair gene polymorphism is associated with the genetic basis of atherosclerotic coronary artery disease. Cardiovasc Pathol 2011;20:e9-15. [8] Gokkusu C, Cakmakoglu B, Dasdemir S, Tulubas F, Elitok A, Tamer S, et al. Association between genetic variants of DNA repair genes and coronary artery disease. Genet Test Mol Biomarkers 2013;17:307–13. [9] Khanna M, Park P, Zirvi M, Cao W, Picon A, Day J, et al. Multiplex PCR/LDR for detection of K-ras mutations in primary colon tumors. Oncogene 1999;18:27–38. [10] Pattin KA, White BC, Barney N, Gui J, Nelson HH, Kelsey KT, et al. A computationally efficient hypothesis testing method for epistasis analysis using multifactor dimensionality reduction. Genet Epidemiol 2009;33:87–94. [11] Hahn LW, Ritchie MD, Moore JH. Multifactor dimensionality reduction software for detecting gene-gene and gene-environment interactions. Bioinformatics 2003;19:376–82. [12] Le Chalony C, Hoffschir F, Gauthier LR, Gross J, Biard DS, Boussin FD, et al. Partial complementation of a DNA ligase I deficiency by DNA ligase III and its impact on cell survival and telomere stability in mammalian cells. Cell Mol Life Sci 2012;69:2933–49. [13] Simsek D, Furda A, Gao Y, Artus J, Brunet E, Hadjantonakis AK, et al. Crucial role for DNA ligase III in mitochondria but not in Xrcc1-dependent repair. Nature 2011;471:245–8. [14] Mani RS, Fanta M, Karimi-Busheri F, Silver E, Virgen CA, Caldecott KW, et al. XRCC1 stimulates polynucleotide kinase by enhancing its damage discrimination and displacement from DNA repair intermediates. J Biol Chem 2007;282:28004–13.

[15] Ladiges W, Wiley J, MacAuley A. Polymorphisms in the DNA repair gene XRCC1 and age-related disease. Mech Ageing Dev 2003;124:27–32. [16] Sameer AS, Shah ZA, Nissar S, Mudassar S, Siddiqi MA. Risk of colorectal cancer associated with the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism in the Kashmiri population. Genet Mol Res 2011;10:1200–10. [17] Ghandour H, Chen Z, Selhub J, Rozen R. Mice deficient in methylenetetrahydrofolate reductase exhibit tissue-specific distribution of folates. J Nutr 2004;134:2975–8. [18] Kaya Y, Cebi A, Soylemez N, Demir H, Alp HH, Bakan E. Correlations between oxidative DNA damage, oxidative stress and coenzyme Q10 in patients with coronary artery disease. Int J Med Sci 2012;9:621–6. [19] Satoh M, Ishikawa Y, Takahashi Y, Itoh T, Minami Y, Nakamura M. Association between oxidative DNA damage and telomere shortening in circulating endothelial progenitor cells obtained from metabolic syndrome patients with coronary artery disease. Atherosclerosis 2008;198:347–53. [20] Niu W, Liu Y, Qi Y, Wu Z, Zhu D, Jin W. Association of interleukin-6 circulating levels with coronary artery disease: a meta-analysis implementing mendelian randomization approach. Int J Cardiol 2012;157:243–52. [21] Narne P, Ponnaluri KC, Singh S, Siraj M, Ishaq M. Arg399Gln polymorphism of X-ray repair cross-complementing group 1 gene is associated with angiographically documented coronary artery disease in South Indian type 2 diabetic patients. Genet Test Mol Biomarkers 2013;17:236–41. [22] Li YY. Methylenetetrahydrofolate reductase C677T gene polymorphism and coronary artery disease in a Chinese Han population: a meta-analysis. Metabolism 2012;61:846–52. [23] Klerk M, Verhoef P, Clarke R, Blom HJ, Kok FJ, Schouten EG. MTHFR 677C– N T polymorphism and risk of coronary heart disease: a meta-analysis. JAMA 2002;288:2023–31. [24] Lewis SJ, Ebrahim S, Davey Smith G. Meta-analysis of MTHFR 677C- N T polymorphism and coronary heart disease: does totality of evidence support causal role for homocysteine and preventive potential of folate? BMJ 2005;331:1053. [25] Heiniger LE, Sherman KA, Shaw LK, Costa D. Fatalism and Health Promoting Behaviors in Chinese and Korean Immigrants and Caucasians. J Immigr Minor Health 2013 Sep 27 [Epub ahead of print]. [26] Moore JH. Detecting, characterizing, and interpreting nonlinear gene-gene interactions using multifactor dimensionality reduction. Adv Genet 2010;72:101–16. [27] Gui J, Andrew AS, Andrews P, Nelson HM, Kelsey KT, Karagas MR, et al. A simple and computationally efficient sampling approach to covariate adjustment for multifactor dimensionality reduction analysis of epistasis. Hum Hered 2010;70:219–25. [28] Niu W, Qi Y. Matrix metalloproteinase family gene polymorphisms and risk for coronary artery disease: systematic review and meta-analysis. Heart 2012;98:1483–91. [29] Jin W, Su X, Xu M, Liu Y, Shi J, Lu L, et al. Interactive association of five candidate polymorphisms in Apelin/APJ pathway with coronary artery disease among Chinese hypertensive patients. PLoS One 2012;7:e51123. [30] Moore JH, Ritchie MD, STUDENTJAMA. The challenges of whole-genome approaches to common diseases. JAMA 2004;291:1642–3. [31] Zintzaras E, Raman G, Kitsios G, Lau J. Angiotensin-converting enzyme insertion/deletion gene polymorphic variant as a marker of coronary artery disease: a metaanalysis. Arch Intern Med 2008;168:1077–89.

Please cite this article as: Yu X, et al, Synergistic association of DNA repair relevant gene polymorphisms with the risk of coronary artery disease in northeastern Han Chinese, Thromb Res (2013), http://dx.doi.org/10.1016/j.thromres.2013.11.017