MMP9 SNP and MMP SNP–SNP interactions increase the risk for ischemic stroke in the Han Hakka population

Abstract Objectives To investigate the association of eight variants of four matrix metalloproteinase (MMP) genes with ischemic stroke (IS) and whether interactions among these single nucleotide polymorphisms (SNPs) increases the risk of IS. Methods Among 547 patients with ischemic stroke and 350 controls, matrix‐assisted laser desorption/ionization time of flight mass spectrometry was used to examine eight variants arising from four different genes, including MMP‐1 (rs1799750), MMP‐2 (rs243865, rs2285053, rs2241145), MMP‐9 (rs17576), and MMP‐12 (rs660599, rs2276109, and rs652438). Gene–gene interactions were employed using generalized multifactor dimensionality reduction (GMDR) methods. Results The frequency of rs17576 was significantly higher in IS patients than in controls (p = .033). Logistic regression analysis revealed the AG and GG genotypes of rs17576 to be associated with a higher risk for IS, with the odds ratio and 95% confidence interval being 2.490 (1.251–4.959) and 2.494 (1.274–4.886), respectively. GMDR analysis showed a significant SNP‐SNP interaction between rs17576 and rs660599 (the testing balanced accuracy was 53.70% and cross‐validation consistency was 8/10, p = .0107). Logistic regression analysis showed the interaction between rs17576 and rs660599 to be an independent risk factor for IS with an odds ratio of 1.568 and a 95% confidence interval of 1.152–2.135. Conclusion An MMP‐9 rs17576 polymorphism is associated with increased IS risk in the Han Hakka population and interaction between MMP‐9 rs17576 and MMP‐12 rs660599 is associated with increased IS risk as well.


INTRODUCTION
Ischemic stroke (IS) accounts for the majority of all strokes, and its high morbidity, disability, and mortality have seriously threatened human health (Feigin et al., 2018;Zhou et al., 2019). Epidemiological studies have shown that IS has become a major disease in China, making it one of the most important diseases leading to disability (Chao et al., 2021;Wan et al., 2021;Wang et al., 2020). The main mechanism of IS is cerebral vascular obstruction caused by atherosclerosis. Although risk factors are actively controlled, the occurrence of IS is still on the rise. Previous studies have shown that gene polymorphisms play an important role in atherosclerotic IS (Malik et al., 2018). In recent years, people have become increasingly interested in the study of matrix metalloproteinases (MMPs) and their relationship with the pathogenesis of atherosclerotic cerebrovascular disease (Chehaibi et al., 2014;Li et al., 2021;Wang & Khalil, 2018).
MMPs are a multigene family of extracellular zinc-and calciumdependent endopeptidases, which play an important pathological role in the degradation of extracellular matrix (ECM) (Abilleira et al., 2006;Chang et al., 2016;Hooper, 1994) in IS. The degradation of arterial ECM proteins is a critical step in the development of atherosclerosis (Fujimoto et al., 2008). Moreover, MMPs can digest the components of fibrous plaque caps, which leads to structural damage and accelerates plaque rupture, giving rise to plaque instability (Galis et al., 1994;Ohshima et al., 2010;Schäfers et al., 2010). Furthermore, MMPs mediate many biological and pathological processes during and after cerebral ischemic injury (Su et al., 2005;. Therefore, matrix metalloproteinases play an important role in atherosclerotic IS. Although there have been many studies on MMP gene polymorphisms in patients with IS, the results have been controversial and ambiguous (Chehaibi et al., 2014;Djurić et al., 2012;Nie et al., 2014;Sheikhvatan et al., 2018;Zhang et al., 2015). Dan Wen et al. (2014) showed that MMP-1-1607 1G/2G and MMP-3-1612 5A/6A were risk factors for IS, while MMP-9-1562C/T was not associated with IS through meta-analysis. Guoqian

Study populations
The study population was comprised of 547 patients with IS and 350 healthy controls in the Hakka population in Western Fujian, China.
According to the trial of ORG 10172 in the acute stroke treatment classification system (Adams et al., 1993)

SNP detection
Blood samples (5 ml, arm vein) from both patients and controls were drawn into sterile tubes containing sodium citrate and were stored at −80 • C. DNA extraction and MMP-1, 2, 9, 12 SNP detection was

Statistical analysis
All statistical tests were analyzed using SPSS software for Windows version 23.0 (SPSS Inc., Chicago, IL). Each variant, and genotype distributions of the eight variants between IS and control were analyzed with a Chi-squared test. A chi-squared analysis was used to compare all categorical data. Normally distributed, continuous data were compared with a student's t-test and expressed as mean ± standard deviation. The BH (Benjamini-Hocberg) method of FDR (False discovery Rate) was used to correct type I errors.
The generalized multifactor dimensionality reduction (GMDR) beta v0.7 software package was used to analyze gene-gene interactions (βversion0.7, www.healthsystem.virginia.edu/internet/addictiongenomics/Software), as previously described (Lou et al., 2007;. GMDR software obtains the best model combination from multiple genes and behavioral indicators through the factor dimensionality reduction principle. The optimal model is obtained from the following results: (1) The model is meaningful only when the p-value is less than 0.05; (2) The larger the testing balance accuracy is, the better the model effect is; (3) The closer the cross validation (CV) consistency is to 10, the better. The influence of high-risk interactive genotypes on functional outcomes was investigated with multivariable logistic regression analysis, after adjusting for the main baseline variables related to each main variable in the univariate analysis (enter approach and probability of entry p < .2). A p-value of less than .05 was considered a statistically significant difference (bilateral test).

Hardy-Weinberg equilibrium
The frequency distribution of the eight variants did not deviate from HWE (p > .05), indicating that gene frequency of the selected study population is representative of the gene distribution of the general population (Table 2).

Clinical characteristics of IS and controls
Demographic characteristics are summarized in Table 3. The proportion of hypertension and diabetes were higher in the IS group than in the control group. The TC, HDL, LDL, and HCY were higher in the IS group than in the control group. There was no significant difference between the two groups in terms of gender, age, smoking habits, alcohol consumption, or triglyceride levels (p > .05) ( Table 3).

Logistic analysis of risk factors of IS
From a univariate analysis, four SNPs candidates were selected for comparison with a logistic regression analysis between the IS and control groups for hypertension, diabetes, TC, HDL, LDL and HCY levels.
Hypertension, diabetes, TC, LDL, HCY, and rs17576 were all independently associated with an increased risk of IS (Table 5). The AG genotype of rs17576 and the GG genotype of rs17576 were associated with a higher risk for IS with an OR and CI of 2.490 (1.251-4.959) and 2.494 (1.274-4.886), respectively (Table 5).

GMDR model for gene interactions
The GMDR model of a gene-gene interaction between rs17576 and rs660599 was deemed the best. The cross-validation consistency was 8/10 and the sign test was nine (p = 0.0107) ( Table 6). The optimal model of interaction between rs17576 and rs660599 by GMDR is shown in Figure 1.

Interaction analysis for rs17576 and rs660599 using logistic regression
According to nine different combinations of rs17576 and rs660599 genotypes in Figure 1, they are divided into high risk and low risk, high SNPs was correlated with higher risk for IS with an OR of 1.568 and a CI of 1.152-2.135 (Table 7).

DISCUSSION
This case-control study is one conducted on the Han Hakka population that aimed to investigate the association between MMP-1 (rs1799750),  was not related to cerebral infarction (Kim et al., 2020). Furthermore, Marc Fatar et al. showed that there is an association of the MMP-2 gene (rs1030868, rs2241145, rs2287074, rs2287076, and rs7201) with the development of lacunar stroke, but no association of MMP-2 with other stroke subtypes (Fatar et al., 2008). Weiling

MMP
Li et al. showed that MMP-12 rs2276109 was not associated with F I G U R E 2 Risk of 9 different combinations of rs17576 and rs660599 genotypes. High-risk cells are indicated by dark color, low-risk cells are indicated by light color. The high-risk interaction genotype was assigned as one, and low-risk interaction genotype was assigned as zero in multivariable logistic regression analysis. The numbers 0-2 represent rs17576 AA,AG and GG, and rs660599 AA,AG and GG, respectively atherosclerosis (Li et al., 2012). This study found that MMP-9 rs17576 was related to IS. Mehrdad Sheikhvatan et al. also showed that MMP-9-C1562T and MMP-9 rs17576 gene polymorphisms were related to coronary atherosclerosis (Sheikhvatan et al., 2018). Additionally, Xianjing Feng et al. reveled MMP-9 rs17576 may be associated with the risk of intracranial atherosclerotic stenosis (Feng et al., 2021). Nevertheless, Alexey Polonikov et al. found that MMP-9 rs17576 did not increase the risk of cerebral infarction alone but increased the risk of cerebral infarction as a part of a gene network (Polonikov et al., 2019). However, there are many studies that are inconsistent with these findings (Djurić et al., 2012;Manso et al., 2010;Traylor et al., 2014). The diversity of these outcomes may be due to ethnic differences, study design, and sample size as well as fortuity. As a matter of fact, it is likely that there are multiple variations in the pathogenesis of cerebral infarction, each with a slight or potentially undetectable effect (Schork et al., 2009).
Due to gene-gene and gene-environment interactions, linkage analyses are commonly used for single-gene disease studies and may not be suitable for genetic studies of stroke.
GMDR is a tool used to study gene-gene interactions and has become a hot tool in gene interaction research (Lou et al., 2007). Nevertheless, significant observations were made in this study using the GMDR method. Through GMDR study, it was found that the interaction between MMP-9 rs17576 and MMP-12 rs660599 increases the risk of IS by 1.568-fold. This result suggests the interaction of these two gene polymorphisms may play a key role in genetic susceptibility to IS. Recent genome-wide association studies have shown the pres-ence of common genetic variants increases the risk of ischemic stroke, but most of the research has focused only on single genes. These findings add to the evidence that genes-gene interactions can increase the risk of complex diseases, such as ischemic stroke. The combinatorial analysis used in this study may be helpful in the elucidation of complex genetic risk factors for common diseases like IS.
MMP-9 is a 92-kDa protein that belongs to a family of zinc-and calcium-dependent endopeptidases (Fenhalls et al., 1999;Pourmotabbed et al., 1994) and plays a key role in all stages of atherosclerosis through monocyte recruitment influence, ECM degradation, endothelial cell migration, and activation of vascular smooth muscle cells (Blankenberg et al., 2003;Hirose et al., 2008;Ye, 2006). MMP-9 gene polymorphisms encode and regulate the transcription of the MMP-9 protein, and correlate with the concentration of MMP-9 in plasma (Luizon et al., 2016). Genetic polymorphisms located in promoter regions of MMP genes can lead to increased gene expression and may be associated with susceptibility to various diseases (Blankenberg et al., 2003).

CONCLUSION
This study shows the MMP-9 rs17576 gene polymorphism is associated with increased IS risk while the other seven gene polymorphisms studied were not significantly associated with increased risk for IS in the Han Hakka population. Simultaneously, it was revealed, through GMDR analysis, that interaction between MMP-9 rs17576 and MMP-12 rs660599 is associated with increased IS risk in the Han Hakka population.

ACKNOWLEDGMENT
This study was Sponsored by Fujian Province Natural Science Foundation (Grant No: 13181526). We also thank Matthew Justice for editing.

CONFLICT OF INTEREST
The authors declare no conflicts of interest.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1002/brb3.2473

DATA AVAILABILITY STATEMENT
All data generated during the project will be made available upon the request from the corresponding author. There are no security, licensing, or ethical issues related to these data.