Association between matrix metalloproteinase 9 C-1562T polymorphism and the risk of coronary artery disease: an update systematic review and meta-analysis

Polymorphism (rs3918242) in the MMP9 gene has been reported to be associated with coronary artery disease (CAD). This study aims to investigate a more accurate estimation of the relationship between CAD and rs3918242 polymorphism by a meta-analysis method. We systematically searched studies on the association of rs3918242 polymorphism and CAD in PubMed, Web of Science, the Cochrane Library, Wanfang Data and CNKI. We used Stata 12.0 and RevMan 5.3 software to perform the meta-analyses. A total of 37 case-control studies involving 13,035 CAD patients and 11,372 non-CAD controls were included. A statistically significant association between rs3918242 polymorphism and CAD was observed in allelic model (Odds ratio (OR) 1.34; 95% confidence interval (CI) 1.20–1.50; p < 0.00001), recessive model (OR 1.43; 95% CI 1.17–1.75; p = 0.0004), and in dominant model ( OR 1.36; 95% CI 1.20–1.53; p < 0.00001). Moreover, we also found that there is a statistically significant association between rs3918242 polymorphism and myocardial infarction (MI) in Asians with allelic model (OR 1.66; 95% CI 1.29–2.14; p < 0.0001), recessive model (OR 2.29; 95% CI 1.44–3.63; p = 0.004), and dominant (OR 1.74; 95% CI 1.29–2.35; p = 0.0003) model. A similar result in Caucasians with allelic model (OR 1.14; 95% CI 1.02–1.27; p = 0.02), and in dominant (OR 1.17; 95% CI 1.04–1.32; p = 0.01) model. Our meta-analysis suggested that the MMP9 T allele is a risk factor for CAD and MI.


INTRODUCTION
Current studies have well documented that the interaction between various environmental factors and certain genetic polymorphisms may lead to CAD [1]. Many association studies between polymorphisms of matrix metalloproteases (MMPs) gene and CAD have been carried out [2,3]. These studies showed that MMPs is associated with a higher risk of plaque rupture/ atherosclerosis and adverse cardiovascular events in patients undergoing CAD [4].
Matrix metallopeptidases 9 (MMP9) has been focused on the value of degrade a wide range of extracellular matrix proteins in patients [5]. MMP 9 is regulated primarily at the transcription level and posttranslational by activation of the zymogen and by inhibition of the endogenous inhibitor TIMP-1 [6]. Although various studies between MMP9 and CAD have

Meta-Analysis
Oncotarget 9469 www.impactjournals.com/oncotarget been reported, the conclusions are not consistent. The MMP9 C-1562T (rs3918242) in the promoter region is of special interest, which was considered a close association with CAD by many studies. Up to now, lots of casecontrol studies and systematic reviews on the relation between rs3918242 and CAD were carried out. However, the conclusions were inconsistent. Based on these observations, to investigate a more accurate estimation of the relationship between CAD and rs3918242, we conducted an update meta-analysis.

Study characteristics
A total of 37 studies  including 13,035 cases and 11,372 controls were identified in this meta-analysis. The Figure 1 show that the study selection process. Supplementary Table 1 and Table 1 have summarized the main characteristics of included studies. In all studies, the genotype frequencies in controls were in consistent with HWE. The results of NOS showed that the methodological quality of be included studies were mostly good (6-9 stars). Table 2 presents a principal results of this studies. For the rs3918242 polymorphism, heterogeneity was found in the allelic (I 2 = 66%, p < 0.00001) and dominant (I 2 = 65%, p < 0.00001) models, but not in the recessive model (I 2 = 22%, p = 0.13). Therefore, We performed a random-effects and fixed-effects method to merge the ORs. The meta-analysis results showed that significant statistical association between rs3918242 polymorphism and the risk of CAD in allelic (OR 1. 34

Sensitivity analyses
We excluded individual studies one at a time and recalculated the pooled p or OR for the remaining studies. The results proved that the ORs were not changed obviously, which suggested that this results are stable.

Publication bias
Egger's test and Funnel plot were conducted to evaluate the publication bias of all contrast models. No obvious bias was found in our study. No obvious asymmetry was found in the funnel plot for the allelic, recessive and dominant genetic models ( Figure 8). Further, Egger's test be used to detect the whole publication bias. No statistically significant of publication bias was detected in allelic (p = 0.592), recessive (p = 0.103) and dominant (p = 0.683) models. The same was true in the subgroup analysis.

DISCUSSION
Our meta-analysis showed that rs3918242 polymorphism was linked with an increased risk of CAD in Asians. More available evidence supports the fact that the rs3918242 as a risk factor association with MI both in Asian and Caucasian populations. MMP-9 might be particularly important in matrix degradation and the subsequent atherosclerotic plaque rupture because of its extensive substrate specificity and distal position in the proteolytic cascade reaction [42]. Loftus et al. reported that MMP-9 concentration and activity were significantly higher in unstable atherosclerotic plaque with intense inflammatory cell infiltration, hence contributing to the plaque rupture ultimately [43]. Opstad et al. suggested that patients with previous MI were associated with the higher MMP-9 gene expression [32]. The current study data from animal experiment, observation of circulating markers and expression investigations on atherosclerotic tissue, has indicated a role of MMP-9 in atherosclerosis. However, the association of rs3918242 polymorphism with CAD and MI risks remains inclusive, although several meta-analysis research (Wang et al. [42], Abilleira et al. [44], Juan et al. [45]) have been published. Wang [42] performed a meta-analysis which included sixteen case-control studies to evaluate the association between rs3918242 and CAD. Subgroup analysis was performed according to different races and outcome(CAD or MI). The final conclusion suggested that an obvious ethnic difference. MMP-9 C1562T polymorphism was associated with CAD or MI in East Asians. But not to west Asians or western populations. Abilleira [43] presented a study of available data from five studies and did not find association of MMP9 polymorphism with CAD. Also, they did not have further specific analysis and subgroup analysis. Juan [44] collected all publications on the association between rs3918242 polymorphism and MI which included 7 researchs. Their data showed that the rs3918242 is a risk factor for white populations, but not for Asian populations.    In order to obtain reliable conclusion, we implemented an update meta-analysis involving 37 studies to provide the relationship between rs3918242 and CAD or MI risks. The study revealed that rs3918242 possibly increased the risk of MI in both Asian and Caucasian populations. As compared with the former studies [43][44][45], our findings has a lot of novelty. Firstly, although lots of studies and systematic review have reported this association, the conclusions were inconsistent or inconclusive. Therefore, our research is urgent and meaningful. Secondly, our meta-analysis is superior to the others, due to the far larger number of Oncotarget 9473 www.impactjournals.com/oncotarget participants (37 included studies with 13,035 cases and 11,372 controls) which were from all over the world. The substantially large sample size ensures the reliability of the results. Thirdly, subgroup analysis was further conducted according to different races.
There are several possible metabolic and molecular mechanisms to explain our conclusion. Zhang [46] have suggested that the C-1562T polymorphic locus is important for the regulatory element that the mutation appeared to be a binding site for a transcription repressor protein and T-allelic promoter had a higher promoter activity. Moreover, more evidences have indicated that the MMP-9 have associated with cell migration and proliferation [47]. More important, the overexpression and activity of MMP-9(rs3918242) was monitored in unstable atherosclerotic plaques [48]. These evidences were in line with our results.
Several limitations should to be pointed out. Risk of bias and unobservable heterogeneity may disturb the results. Included studies of language limit to English and Chinese, missing some studies by other languages. Some hardly to be avoided publication bias might exist. Such as some factors like age, gender, individual conditions, environment and experimental method are different, those might influence the interpretation of result in our metaanalysis.
In conclusion, the meta-analysis provided evidence that MMP9 rs3918242 polymorphism was significantly associated with CAD/ MI in Asian populations. The same of available evidence supports the fact that the   Larger studies with the consideration of more influence factors and better study designs are still required to further evaluate the connection of MMP9 rs3918242 polymorphism with CAD/MI susceptibility.

Literature search
All studies that researched the association between the rs3918242 polymorphism and CAD were identified by comprehensive computer-based searches of PubMed, Web of Science, the Cochrane Library, Wanfang Data and China National Knowledge Infrastructure (CNKI).

Inclusion criteria
The diagnosis of CAD was fitted to the examination results of coronary arteriography, treadmill exercise test, Oncotarget 9476 www.impactjournals.com/oncotarget clinical symptoms combined with electrocardiogram, as well as some other inspection project, for example echocardiography and myocardial perfusion imaging. The inclusion criteria for eligible studies were as follows: (1) independent case-control studies using either a hospitalbased or a population-based design; (2) the studies of MMP9 C-1562T polymorphism and CAD risk; (3) the studies has an intact original data on genotype distribution and a comprehensive statistical index, sufficient data for estimating an odds ratio (OR) with 95% confidence interval (CI); (4) no repeat published data.

Exclusion criteria
We excluded studies if (1) reviews, editorials, and articles with insufficient information; (2) the genotype distribution in the control group non-conformity the Hardy-Weinberg equilibrium.

Data extraction
Two authors independently extracted the data. The extracted information included the first author's name, publication year, study population, number of genotypes, genotyping methods, allele frequency of cases and controls, sample sizes in the cases and controls, sex and age of cases and controls. Disagreement was resolved by consensus. If these two authors could not reach a consensus, the result was reviewed by a third author.

Quality assessment
To determine the methodological quality of the included studies, we used the Newcastle-Ottawa scale [49], which uses a "star" rating system to judge the quality of observational studies. The NOS ranges between zero (worst) up to nine stars (best). Two authors assessed the quality of included studies independently and solved disagreement through discussion.

Statistical analysis
The association between rs3918242 polymorphism and CAD, which in our meta-analysis were compared by using the OR and its corresponding to 95% CI. Hardy-Weinberg equilibrium (HWE) was assessed by Chi-square Oncotarget 9477 www.impactjournals.com/oncotarget test in control groups, and P < 0.05 was considered a significant departure from HWE. Heterogeneity between studies was assessed by I 2 test, p < 0.10 and I 2 > 50% indicated evidence of heterogeneity. If heterogeneity existed among the studies, the random effects model was used to estimate the pooled OR (the DerSimonian and Kacker method). Otherwise, the fixed effects model was adopted (the Mantel-Haenszel method) [50,51]. The associations between the genetic variant and CAD risk of pooled ORs were performed for a recessive genetic model, dominant genetic model and allelic contrast. Z test was used to determine the pooled OR and significance was set at p < 0.05. Besides, subgroup analyses were stratified by ethnicity and outcome. The potential publication bias was checked by using funnel plots and Egger's test [52]. The statistical analysis was performed by using Review Manager 5.30 (Cochrane Collaboration, The Nordic Cochrane Centre, Copenhagen) and Stata 12.0 software (StataCorp, College Station, TX, USA). A two-tailed p < 0.05 was considered significant.

Author contributions
MMZ, XX and SYZ conceived the study, participated in the design, collected the data, performed statistical analyses, and drafted the manuscript. XQH, XWC and WH conceived the study and revised manuscript. HW participated in the design, and helped to draft the manuscript. All authors read and approved the final manuscript.

ACKNOWLEDGMENTS AND FUNDING
This study was supported by Luoyang Science and Technology Projects (No. 1201050A-1) and the project of Henan Science and Technology (NO.122300410234).