Literature Review 7 pages

Contents lists available at ScienceDirect

Gene

journal homepage: www.elsevier.com/locate/gene

Research paper

Impact of 9p21.3 region and atherosclerosis-related genes' variants on long- term recurrent hard cardiac events after a myocardial infarction

German J. Osmaka,b, Boris V. Titova,b, Natalia A. Matveevaa,b, Vitalina V. Bashinskayaa,b, Roman M. Shakhnovicha, Tatiana S. Sukhininaa, Nino G. Kukavaa, Mikhail Ya. Rudaa, Olga O. Favorovaa,b,⁎

a National Medical Scientific Center for Cardiology, Moscow, Russia b Pirogov Russian National Research Medical University, Moscow, Russia

A R T I C L E I N F O

Keywords: Acute myocardial infarction Genetic markers Recurrent cardiovascular events 9p21.3 locus Multilocus analysis

A B S T R A C T

Atherosclerotic coronary artery disease (CAD) and myocardial infarction (MI) as its most severe clinical com- plication remain the leading causes of mortality in the majority of countries. Despite the progress in the treat- ment of MI, quite often the patients, after the first-time MI, develop subsequently a variety of adverse cardio- vascular events. In this retrospective study we evaluated the contribution of allelic variations in 9p21.3 locus and in 21 atherogenesis-related genes to the development of hard cardiac events in a cohort of patients of Russian ethnicity after the first acute MI during long-term follow-up (7–10 years). Death from cardiac causes and re- current nonfatal MI were considered as key clinical outcomes. We have shown the association of rs1333049 and rs10757278 in 9p21.3 and MTHFR rs1801133 with recurrent unfavorable events, the latter was observed in time-dependent manner. Multilocus analysis additionally suggested the influence of carriage of the CRP and ENOS genes variants at the development of subsequent adverse events after MI. The composite model built for prediction of the individual genetic risk of postinfarction hard cardiac events included 9p21.3 rs1333049*GG and MTHFR*TT and was characterized by area under the curve (AUC) = 0.65. Our data show that 9p21.3 locus and MTHFR gene polymorphisms could influence long-term prognosis of recurrent hard cardiac events in pa- tients who underwent the first MI. It is possible that addition of genotyping at such loci to existing clinical scores could improve their predictability.

1. Introduction

Atherosclerotic coronary artery disease (CAD) and MI as its most severe clinical complication remain the leading causes of mortality in the majority of countries. Only in the Russian Federation in 2010 nearly 600,000 deaths were caused by CAD, which is the highest number among all countries included into analysis (Nowbar et al., 2014).

In the recent years, significant progress had been made in the treatment of MI. Nevertheless, quite often the patients who underwent the first-time MI, develop a variety of adverse cardiovascular (CV) events, such as recurrent MI, stroke, etc., which are associated with significant mortality. > 11% of admissions to hospitals in Russia ac- count for recurrent MIs (Ulumbekova, 2010). Each year, ≈660,000 Americans have a new coronary attack (defined as first hospitalized MI or cardiac death) and ≈305,000 have a recurrent attack (Writing

Group Members et al., 2016). Except for extremely rare monogenic inherited disorders, MI is a

complex polygenic disease, the development of which is also influenced by epigenetic and environmental factors. While CAD genetics is well studied at the genome-wide significance level, only a few genome-wide association studies (GWASs) were performed for MI as a distinct phe- notype and identified several common genetic variants robustly asso- ciated with MI (Burdett et al., n.d.). Among them, only one MI-asso- ciated region, 9p21.3, was replicated in several GWASs and validated in some, but not in all candidate gene association studies (Dai et al., 2016; Roberts, 2014).

Genetic factors may be involved also in development of recurrent CV events considering that a family history of premature CV disease is a risk factor for these events (Mulders et al., 2011). To date, dozens of genetic studies on recurrent CV events have been conducted,

https://doi.org/10.1016/j.gene.2018.01.036 Received 1 December 2017; Received in revised form 29 December 2017; Accepted 9 January 2018

⁎ Corresponding author at: 3-d Cherepkovskaya str, 15A, 121552 Moscow, Russia. E-mail address: [email protected] (O.O. Favorova).

Abbreviations: ANRIL, antisense noncoding RNA in the INK4 locus; AUC, area under the curve; CAD, coronary artery disease; CV, cardiovascular; GRS, genetic risk score; GWAS, genome-wide association study; HR, hazard ratio; LD, linkage disequilibrium; MI, myocardial infarction; PCR, polymerase chain reaction; ROC, receiver operating characteristic; SNP, single nucleotide polymorphism

Gene 647 (2018) 283–288

Available online 10 January 2018 0378-1119/ © 2018 Published by Elsevier B.V.

T

proceeding from the assumption that genes implicated in the disease susceptibility may influence its pathogenesis and, as a consequence, affect prognosis. The same loci that had been associated with first MI risk, and most notably the MI-associated region 9p21.3, were mainly examined for subsequent CV events. However, results were contra- dictory, possibly due to different study design. Indeed, these studies differed in the initial phenotype of CV disease, the chosen endpoints and the duration of follow-up. The initial phenotypes in these studies were usually coronary heart disease (Patel et al., 2014); acute coronary syndrome (Buysschaert et al., 2010; Labos et al., 2015; Zeng et al., 2013), and MI (Peng et al., 2009; Hara et al., 2014; Dehghan et al., 2016), or combinations of several CV diseases. As endpoints, recurrent MI, development of unstable angina, cardiac and total mortality, re- vascularization, or a combination of several adverse events were most often considered. The patients' follow-up time varied greatly: from a very short period of six months (Buysschaert et al., 2010) to 10–15 years of monitoring (Máchal et al., 2014). All this greatly com- plicates the possibility of comparing the results of various studies.

In this work we attempted to evaluate the contribution of rs1333049 and rs10757278 - the most studied in CAD genetics single nucleotide polymorphisms (SNPs) in 9p21.3 locus (Patel et al., 2014) - as well as some other polymorphic loci in genes coding proteins in- volved in atherogenesis, to the development of hard cardiac events after MI (death from cardiac causes and recurrent nonfatal MI) in subjects who had undergone MI on the basis of long-term follow-up data (from 7 to 10 years). Genetic variants selected for analysis and their known (or putative) effects on the level/activity of the encoded proteins are listed in Supplementary Table S1.

2. Material and methods

2.1. Patients selection

133 unrelated patients (104 men and 29 women, the mean age of the first MI 53.7 ± 9.3 years) with the first MI as the index event were enrolled in 2013 to the retrospective study based on long-term follow- up. They were selected after telephone interview out of 325 patients, who had undergone treatment at the Department of Emergency Cardiology, National Medical Scientific Center for Cardiology, Moscow, Russia, in the period from January 1, 2006 to December 31, 2010, with the diagnosis of “first acute myocardial infarction” at age under 70. The diagnosis of acute MI for the patients met the definition of the Third universal definition of MI (Thygesen et al., 2012): a rise and/or fall of cardiac biomarker values (troponin I concentration or creatine kinase MB fraction activity) was accompanied by chest pain lasting longer than 30 min and/or emergence of new electrocardiogram abnormalities (newly appearing significant ST-segment–T wave changes, new left bundle branch block or development of pathological Q waves in the electrocardiogram). All the patients were Moscow region residents and self-reported as ethnic Russians Patients for whom no information were available, were considered lost for the study. If a patient was confirmed to have died, the relatives were asked about the exact date of death and whether the cause of death was directly related to CV events or their complications.

Clinical profiles of the selected patients at the time of the first MI are presented at Supplementary Table S2. The occurrence of hard cardiac events (death from cardiac causes or recurrent nonfatal MI) was con- sidered as the composite endpoint.

All procedures involving human participants were in accordance with the 2013 World Medical Association Declaration of Helsinki; the study was approved by the local Ethics Committee. Informed consent was obtained from all individual participants or their relatives included in the study.

2.2. Genotyping

Genomic DNA was isolated from peripheral blood by phenol- chloroform extraction using standard procedures. Genotyping methods for rs1130864, rs1800896, rs231775, rs333, rs6050, rs1800788, rs152312, rs2243250, rs1800629, rs909253, rs1800795, rs1799889, rs3842787, rs2430561, and rs1982073 polymorphic variants were de- scribed earlier (Barsova et al., 2015). Polymerase chain reaction (PCR) followed by restriction enzyme digestion was used for rs1801133 (Christensen et al., 1997), rs2070744 (Augeri et al., 2009), rs7412 and rs429358 (Hixson and Vernier, 1990), rs320 and rs328 (Shimo- Nakanishi et al., 2001). Genotypes of rs2020918 were detected using allele-specific PCR (Jannes et al., 2004). Genotypes of rs562556, rs2910829, rs10757278, and rs1333049 were detected by real-time PCR in StepOne device using the TaqMan® SNP Genotyping Assay kit and the TaqMan® Universal PCR Master Mix without the AmpErase® UNG enzyme (Applied Biosystems, United States). Quality for all assays was assessed by random selection of 20% or more samples for re-gen- otyping; in addition, sequence analysis of some amplicons was used to confirm the accuracy of the results obtained by PCR. No inconsistencies were observed.

2.3. Statistical analysis

Association of genetic variants carriage with recurrent hard cardiac events included in the composite endpoint was investigated using Kaplan-Meier estimates and Cox regression for hazard ratio (HR) cal- culation. Those cases, when an individual retired from the study for reasons other than the events described above, were considered cen- sored. Time was censored at December 31, 2016 for this analysis, in- dependently from outcome. To test the hypothesis that the difference between HR estimates for allelic variants' carriers is not equal to 1, the Wald test was used in the construction of regression models; estimates of pw ≤ 0.05 were considered significant.

For plotting of Kaplan-Meier curves and calculating HRs using the Cox proportional model, we used the R packages ‘survival’ version 2.41- 3 and ‘survminer’ version 0.3.1 (The Comprehensive R Archive Network, n.d.).

To study the combined effects of carriage of two or more genetic risk factors, we constructed predictive models with multiple covariates based on Cox regression. Search for possible allelic combinations as- sociated with survival for further implementation in the model was carried out using the APSampler algorithm (APSampler software, n.d.; Favorov et al., 2005). Comparison of models was performed using ANOVA test (Chambers and Hastie, 1992). For model selection we used Akaike's information criterion measure, and backward stepwise selec- tion algorithm with the step() function in R. The prognostic values were evaluated from the areas under the curves (AUC) in the receiver oper- ating characteristic (ROC) analysis using the pROC package v.1.8 for R v.3.3.3. Comparison of the AUC estimates was performed by the method of DeLong et al. (DeLong et al., 1988), with significance threshold pdel ≤ 0.05. The fraction of the variance explained in the model was estimated by the value of McFadden's R2 (McFadden, 1973).

3. Results

In this study we investigated the association of genetic variants PCSK9 rs562556, CRP rs1130864, IL10 rs1800896, MTHFR rs1801133, CTLA4 rs231775, CCR5 rs333, FGA rs6050, FGB rs1800788, PDE4D rs152312 and rs2910829, IL4 rs2243250, TNF rs1800629, LTA rs909253, IL6 rs1800795, PAI1 rs1799889, ENOS rs2070744, PLAT rs2020918, LPL rs320 and rs328, PTGS1 rs3842787, IFNG rs2430561, TGFB1 rs1982073, APOE rs7412 and rs429358 (epsilon poly- morphism), rs10757278 and rs1333049 in 9p21.3 locus with hard cardiac events that could occur after the initial MI during long-term follow-up. Linkage analysis of the investigated polymorphic loci with

G.J. Osmak et al. Gene 647 (2018) 283–288

284

Haploview 4.2 (Barrett et al., 2005) software showed complete linkage disequilibrium (LD) between rs10757278 and rs1333049 in the 9p21.3 region (D' = 1, LOD > 2).

Kaplan-Meier curves that show frequencies of unfavorable outcomes at different times for carriers of alleles/genotypes of 9p21.3 (rs1333049, rs10757278) and MTHFR (rs1801133) are presented in Fig. 1. Carriage of rs1333049*GG, rs10757278*AA, and rs1801133*ТТ genotypes is associated with poor prognosis under recessive models. For the remaining SNPs studied, no significant estimates have been ob- tained (data not shown).

Table 1 shows the values of the Kaplan-Meier estimates (survival rate and median survival) and the Cox hazard ratios (HR) of survival curves. The assumption of proportional hazards has been fulfilled for all polymorphic loci. Analysis of the time-to-endpoint shows that for 9p21.3 rs1333049, the survival rate at the endpoint was 0.39 [95% CI 0.20–0.58] in GG carriers and 0.75 [0.65–0.83] in C carriers (i.e., GG non-carriers). According to Cox proportional hazards regression ana- lysis, HR for rs1333049*GG carriers was equal to 3.14 [95% CI 1.63–6.04] (рw = 0.0006). For the 9p21.3 rs10757278 the survival rate at the endpoint was 0.43 [0.23–0.62] in AA carriers and 0.74 [0.64–0.82] in the G allele carriers; at that, HR for rs10757278*AA carriers was equal to 2.48 [95% CI 1.28–4.79] (рw = 0.006). In case of MTHFR, for rs1801133*TT carriers HR value was 2.76 [95% CI 1.20–6.33] (рw = 0.02). Table 1 also shows the median survivals for carriers of the polymorphic loci during the observation period (from 7 to 10 years); they are reached only for carriers of risk genotypes and

equal to 7 years for each of them. Apparently, between two or more genetic variants, linear and

nonlinear interactions can occur affecting the risk of adverse events. Multilocus analysis, which allows to identify phenotype-associated al- lelic combinations, may help to reveal additional cumulative risk fac- tors, components of which have not been identified as risk factors by themselves. Multilocus analysis allowed to identify only protective al- lelic combinations (MTHFR rs1801133*C + CRP rs1130864*C), (MTHFR rs1801133*C + CRP rs1130864*C + 9p21.3 rs1333049*C), and (MTHFR rs1801133*C + CRP rs1130864*C + ENOS rs2070744*T) (Table 2), the carriage of which was characterized by increased level of significance comparing with the carriage of protective alleles MTHFR rs1801133*C and 9p21.3 rs1333049*C by themselves (see Table 1). For example, combination (MTHFR rs1801133*C + CRP rs1130864*C) is negatively associated with hard postinfarction cardiac events more significantly (pw = 0.002) than MTHFR rs1801133*C (pw = 0.02) by itself, while CRP rs1130864*C was not significantly associated with the trait by itself (pw > 0.05). Thus, the multilocus analysis additionally suggested the influence of carriage of the CRP and ENOS genes variants at the development of subsequent hard cardiac events after MI.

To assess the possibility of prognostic application of the observed effects, we built a composite model predicting the individual genetic risk of hard cardiac events after MI. Because in 9p21.3 locus, rs10757278 is in strong LD with rs1333049, these SNPs cannot serve as independent markers of hard cardiac events. To select, which of these SNPs to use as a better independent predictor in the composite model,

Fig. 1. Cumulative survival rate during follow-up according to different genotypes. Kaplan-Meier curves show frequencies of unfavorable outcomes considering death from cardiac causes or recurrent MI as the composite endpoint at different times for carriers of alleles/genotypes of loci 9p21.3 rs1333049 (A), 9p21.3 rs10757278 (B) and MTHFR rs1801133 (C).

G.J. Osmak et al. Gene 647 (2018) 283–288

285

we preliminary constructed the rs1333049 only model and rs10757278 only model. The McFadden's R2 values (McFadden, 1973) were equal to 0.071 for rs1333049 and 0.043 for rs10757278; therefore we included rs1333049 in the composite model as SNP which explains more var- iance fraction. As the result, 9p21.3 rs1333049*GG and MTHFR*TT have been included in the composite model as predictors having a significant impact (pw < 0.05) on the probability of occurrence of adverse postinfarction events (Supplementary Table S3), in compliance with the data in Table 1. The variants of CRP rs1130864 and ENOS rs2070744*T which were components of allelic combinations were not included in the model, because they explained the small variance fraction.

The resulting model was tested using ROC analysis (Fig. 2). It is seen that the curve for composite model (AUC = 0.65) lies above the curves for rs1333049 and MTHFR (AUCs = 0.60 and 0.53, respectively), in such a way the composite model is probably more effective than the predictors included in it. The composite model is significantly (pdel = 0.05) more effective than MTHFR only model, but not than 9p21.3 rs1333049 only model (pdel = 0.08).

4. Discussion

In this work we performed the analysis of allele/genotype associa- tions of 25 polymorphous loci with the development of recurrent postinfarction hard cardiac events in individuals of Russian ethnicity on the basis of the long-term follow-up data. Two SNPs are located in the MI-associated 9p21.3 region and other variants - in/near 21 genes en- coding proteins implicated in atherosclerosis or atherosclerosis-related diseases (Supplementary Table S1). Among these genetic variants, we showed the association of 9p21.3 (rs1333049*GG and rs10757278*AA)

and MTHFR (rs1801133*TT) with occurrence of hard cardiac events after the first-time MI.

The SNPs defining the observed 9p21.3 association with recurrent hard cardiac events after MI are located in intergenic regions around the cluster of genes, CDKN2B, CDKN2A and ANRIL (CDKN2B-AS1). CDKN2B, CDKN2A are proteincoding genes: CDKN2B encodes the cy- clin-dependent kinase inhibitor 2B (p15INK4b), whereas CDKN2A has two alternative splicings, encoding cyclin-dependent kinase inhibitor 2A (p16INK4a) and p14ARF. These proteins are involved in the regulation of cell division and apoptosis, and the maintenance of cellular home- ostasis (Nelson and Tsao, 2009). The third gene in the cluster, ANRIL (CDKN2B-AS1), encodes long noncoding RNA ANRIL (Antisense Non- coding RNA in the INK4 Locus), that may participate in the regulation of CDKN2A and CDKN2B expression (Jarinova et al., 2009; Burd et al., 2010). Targeted deletion of the 9p21.3 locus reduces the cardiac ex- pression of CDKN2A/B, leading to a less stable plaque phenotype in the artery (Tajbakhsh et al., 2016). According to data (Congrains et al., 2012), atherosclerosis development and progression may be affected by

Table 1 Genotypes, which, according to long-term observations (7 to 10 years), are associated with hard postinfarction cardiac events (death from cardiac causes or recurrent MI as the composite endpoint) in a sample of ethnic Russians (a total of 133 people).

Gene (SNP ID)

Carriage of genotypes (alleles)

Kaplan-Meier survival analysis Cox proportional hazards regression analysis

Survival rate [95% CI]

Median survival (years) HR [95% CI]

pw-value

9p21.3 (rs1333049)

GG 0.39 [0.20–0.58] 7 3.14 [1.63–6.04]

0.0006

CC + CG (C) 0.75 [0.65–0.83] NR 0.32 [0.16–0.61]

9p21.3 (rs10757278)

AA 0.43 [0.23–0.62] 7 2.48 [1.28–4.79]

0.006

GG + AG (G) 0.74 [0.64–0.82] NR 0.40 [0.21–0.78]

MTHFR (rs1801133) TT 0.18 [0.01–0.51] 7 2.76 [1.20–6.33]

0.02

CC + CT (C) 0.72 [0.62–0.79] NR 0.36 [0.15–0.83]

NR – not reached.

Table 2 Allele combinations, which, according to long-term observations (7 to 10 years), are negatively associated with hard postinfarction cardiac events (death from cardiac causes or recurrent MI as the composite endpoint) in a sample of ethnic Russians (a total of 133 people).

Carriage of allele combination pw-value HR [95% CI]

MTHFR rs1801133*C + CRP rs1130864 *C 0.002 0.35 [0.18–0.68] Non-carriers of this allele combination 2.82 [1.46–5.48] MTHFR rs1801133*C + CRP rs1130864

*C + 9p21.3 rs1333049*C 0.00007 0.27 [0.14–0.52]

Non-carriers of this allele combination 4.54 [2.34–8.79] MTHFR rs1801133*C + CRP

rs1130864*C + ENOS rs2070744*T 0.001 0.36 [0.19–0.69]

Non-carriers of this allele combination 2.75 [1.45–5.23]

Fig. 2. ROC curves for the additive composite model (AUC = 0.65) built from 9p21.3 rs1333049 and MTHFR.

G.J. Osmak et al. Gene 647 (2018) 283–288

286

differential expression of ANRIL splice variants that could interfere with gene expression in many processes, such as proliferation, remodeling of extracellular matrix, and response to inflammation. Thus, the cluster of CDKN2B, CDKN2A and ANRIL genes may be involved in the athero- sclerosis-related diseases susceptibility, may influence their pathogen- esis and recurrent CV events. 9p21.3 rs10757278*GG genotype was associated with the lowest expression level of genes ANRIL, CDKN2B, CDKN2A, and ARF (Liu et al., 2009).

The data of systematic review with a subsequent meta-analysis (Patel et al., 2014) showed that the association of SNPs located in 9p21.3 with risk of subsequent CV events after established coronary heart disease is less evident than with risk of a first coronary heart disease event. Nevertheless, the association with recurrent adverse CV events was observed in a number of studies both for of 9p21.3 rs1333049 (Buysschaert et al., 2010; Hara et al., 2014; Dehghan et al., 2016; Szpakowicz et al., 2014) and rs10757278 (Zeng et al., 2013; Szpakowicz et al., 2014; Gong et al., 2011). Herewith, in some of these studies rs1333049*С and/or rs10757278*G were identified as the risk alleles of recurrent CV events [Buysschaert et al., 2010, Zeng et al., 2013, Larson et al., 2007], whereas in others the carriage of these al- leles showed a protective effect (Hara et al., 2014; Dehghan et al., 2016; Szpakowicz et al., 2014; Gong et al., 2011).

We observed that the carriage of rs1333049*С and rs10757278*G was protective against hard cardiac events whereas genotypes rs1333049*GG and rs10757278*AA were positively associated with these events under recessive model. Due to the high size of the observed effect of 9p21.3 alleles carriage on post-MI event prognosis (for rs1333049*GG HR = 3.14, p-value = .0006; for rs10757278*AA HR = 2.48, p-value = .006), these SNPs could be potentially used for prediction of recurrent CV events after the first MI, at least in ethnic Russians.

Among the other 21 SNPs in protein-coding genes included in our study, we observed the only risk variant of unfavorable CV events after first-time MI, namely the MTHFR rs1801133*TT genotype. Earlier it was shown that TT genotype was significantly associated with recurrent CV events following ischemic heart disease and coronary revascular- ization (Kosokabe et al., 2001; Botto et al., 2004; Pereira et al., 2007; Andreassi et al., 2012).

Noteworthy, we did not observe significant differences in fre- quencies of unfavorable outcomes for MTHFR variants until approxi- mately 4 years of follow-up (Fig. 1C) while the curves for the carriers of the 9p21.3 variants diverge from the beginning of observation (Fig. 1A and B). We assumed that this is due to the different influence of the MTHFR variants carriage at different ages. Indeed, the study (Russo et al., 2003) demonstrated the effect of age on the association of MTHFR rs1801133*TT with total homocysteine concentrations. Homocysteine is known to be involved in the processes of endothelial dysfunction, oxidative damage, increase of collagen synthesis and de- terioration of arterial wall elastic material (Ganguly and Alam, 2015). Although the biological mechanism underlying observed time-depen- dent effect of MTHFR*TT carriage on unfavorable CV events is un- known, it is attractive to suppose that homocysteine, being involved in some steps of MI pathogenesis, determines the postponed effect of MTHFR*TT carriage on adverse CV events. If our assumption is true, then the phenomenon we observed falls within the definition of gene- age interaction (Simino et al., 2014).

Using multilocus analysis we revealed additional variants, nega- tively associated with hard cardiac events as a part of allelic combi- nations, namely CRP rs1130864*C and ENOS rs2070744*T, which do not influence the phenotype by themselves. Earlier, in an independent sample of Russian patients with MI, we observed the positive associa- tion of CRP rs1130864*T allele carriage with unfavorable postinfarc- tion CV events during two-year follow-up (Sukhinina et al., 2012). Thus, the data of the present work may be considered as validation of CRP rs1130864 impact on postinfarction forecast. As for the association of ENOS rs2070744 in development of recurrent postinfarction adverse

CV events, this issue needs further study. It should be noted that among detected protective combinations, we

found biallelic combination (MTHFR rs1801133*C + CRP rs1130864*C) and two more significant triallelic combinations that included this pair of alleles. The stable effect of allelic combinations of these genes on postinfarction events raises the question of the possibi- lity of their protein products interaction. Indeed, study of the homo- cysteine effect on CRP expression in vascular smooth muscle cells re- vealed that homocysteine induced increase in CRP-specific mRNA and C-reactive protein production both in vitro and in vivo and is capable of initiating an inflammatory response (Pang et al., 2014).

The composite model built for prediction of the individual genetic risk of adverse events after MI included 9p21.3 rs1333049*GG and MTHFR*TT as covariates and was characterized by AUC = 0.65. Previously, several attempts to create genetic risk score (GRS) had been suggested for improvement of recurrent CV events prognosis; however, the results of these studies were contradictory. Thus, the studies (Labos et al., 2015; Weijmans et al., 2015) showed no predictability of re- current CV events basing on genotypes at 30 SNPs; it should be men- tioned that among these SNPs there were no predictors identified in our study. On the contrary, in the works (Andreassi et al., 2012; Mega et al., 2015; Hughes et al., 2012) the possibility of effective classification of patients by GRSs was shown. GRS based on the results of a 6-year follow-up (Andreassi et al., 2012) included, among others, MTHFR rs1801133 SNP, significantly associated with postinfarction CV events in our study only after prolonged follow-up. At the same time, the au- thors showed the lack of prognostic value for rs1333049 9p21.3 in Italian population (Andreassi et al., 2012). A GRS was proposed for inhabitants of the Siberian area of Russia for prediction of MI compli- cations within one year follow-up (Makeeva et al., 2013). This GRS had AUC = 0.75; however, 9p21.3 variants rs1333049 and rs10757278 were not selected for the analysis. Interestingly, MTHFR rs1801133 did not fit as a predictor in the GRS in this work. This is in good compliance with the observation that MTHFR acts only as delayed predictor of re- current CV events ((Andreassi et al., 2012) and our results).

In our study 9p21.3 rs1333049 has AUC = 0.60 and the largest effect on the composite model (see Fig. 2). This raises the question on its possible implementation in the existing GRSs for prediction of re- current CV events after the first MI. Taking into account the successful experience in integrating new polymorphic loci into GRACE scale for postinfarctional prognosis (Pavkova Goldbergova et al., 2017), it seems attractive to suggest that analysis of 9p21.3 rs1333049 polymorphism could be used to improve this scale at least in ethnic Russians. However, without any doubts, the results should be replicated in independent larger samples.

5. Conclusions

In conclusion, our data suggest that some of the atherogenesis-re- lated and MI-associated loci could influence long-term postinfarction prognosis of hard cardiac events in patients who underwent the first MI, namely, 9p21.3 rs1333049*GG and MTHFR rs1801133*TT are risk genotypes for unfavorable outcomes. It is possible that addition of genotyping at such loci to existing clinical scores could improve their predictability.

Acknowledgements

This work was supported by the Russian Science Foundation [grant number 16-14-10251].

Declaration of interest

The authors declare that they have no conflict of interest.

G.J. Osmak et al. Gene 647 (2018) 283–288

287

Contributors

G.O., B.T., and V.B. performed experiments; N.K., R.Sh., T.S., and M.R. recruited patients, collected data; G.O., B.T. and N.M. analyzed the data; N.M., O.F., G.O., and V.B. wrote the manuscript; O.F. co- ordinated the study.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.gene.2018.01.036.

References

Andreassi, M.G., Adlerstein, D., Carpeggiani, C., et al., 2012. Individual and summed effects of high-risk genetic polymorphisms on recurrent cardiovascular events fol- lowing ischemic heart disease. Atherosclerosis 223 (2), 409–415.

APSampler software Available at. http://apsampler.sourceforge.net. Augeri, A.L., Tsongalis, G.J., Van Heest, J.L., Maresh, C.M., Thompson, P.D., Pescatello,

L.S., 2009. The endothelial nitric oxide synthase -786 T > C polymorphism and the exercise-induced blood pressure and nitric oxide responses among men with elevated blood pressure. Atherosclerosis 204 (2), e28–34.

Barrett, J.C., Fry, B., Maller, J., Daly, M.J., 2005. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 21 (2), 263–265. https://www. broadinstitute.org/haploview/haploview (Haploview 4.2 available at).

Barsova, R.M., Lvovs, D., Titov, B.V., et al., 2015. Variants of the coagulation and in- flammation genes are replicably associated with myocardial infarction and epistati- cally interact in Russians. PLoS One 10 (12), e0144190.

Botto, N., Andreassi, M.G., Rizza, A., et al., 2004. C677T polymorphism of the methyle- netetrahydrofolate reductase gene is a risk factor of adverse events after coronary revascularization. Int. J. Cardiol. 96 (3), 341–345.

Burd, C.E., Jeck, W.R., Liu, Y., Sanoff, H.K., Wang, Z., Sharpless, N.E., 2010. Expression of linear and novel circular forms of an INK4/ARF-associated non-coding RNA corre- lates with atherosclerosis risk. PLoS Genet. 6 (12), e1001233.

Burdett, T.Hall, P.N.Hastings, E.Hindorff, L.A.Junkins, H.A.Klemm, A.K.MacArthur, J.Manolio, T.A.Morales, J.Parkinson, H.Welter, D. The NHGRI-EBI Catalog of pub- lished genome-wide association studies. Available at. http://www.ebi.ac.uk/gwas, Accessed date: 10 August 2017.

Buysschaert, I., Carruthers, K.F., Dunbar, D.R., et al., 2010. A variant at chromosome 9p21 is associated with recurrent myocardial infarction and cardiac death after acute coronary syndrome: the GRACE genetics study. Eur. Heart J. 31, 1132–1141.

Chambers, J.M., Hastie, T.J., 1992. Statistical Models in S. Wadsworth & Brooks/Cole, Pacific Grove, California.

Christensen, B., Frosst, P., Lussier-Cacan, S., et al., 1997. Correlation of a common mu- tation in the methylenetetrahydrofolate reductase gene with plasma homocysteine in patients with premature coronary artery disease. Arterioscler. Thromb. Vasc. Biol. 17 (3), 569–573.

Congrains, A., Kamide, K., Katsuya, T., et al., 2012. CVD-associated non-coding RNA, ANRIL, modulates expression of atherogenic pathways in VSMC. Biochem. Biophys. Res. Commun. 419 (4), 612–616.

Dai, X., Wiernek, S., Evans, J.P., Runge, M.S., 2016. Genetics of coronary artery disease and myocardial infarction. World J. Cardiol. 8 (1), 1–23.

Dehghan, A., Bis, J.C., White, C.C., et al., 2016. Genome-wide association study for in- cident myocardial infarction and coronary heart disease in prospective cohort stu- dies: the CHARGE consortium. PLoS One 11 (3), e0144997.

DeLong, E.R., DeLong, D.M., Clarke-Pearson, D.L., 1988. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric ap- proach. Biometrics 44, 837–845.

Favorov, A.V., Andreewski, T.V., Sudomoina, M.A., Favorova, O.O., Parmigiani, G., Ochs, M.F., 2005. A Markov chain Monte Carlo technique for identification of combinations of allelic variants underlying complex diseases in humans. Genetics 171 (4), 2113–2121.

Ganguly, P., Alam, S.F., 2015. Role of homocysteine in the development of cardiovascular disease. Nutr. J. 14, 6.

Gong, Y., Beitelshees, A.L., Cooper-DeHoff, R.M., et al., 2011. Chromosome 9p21 hap- lotypes and prognosis in white and black patients with coronary artery disease. Circ. Cardiovasc. Genet. 4 (2), 169–178.

Hara, M., Sakata, Y., Nakatani, D., et al., 2014. Reduced risk of recurrent myocardial infarction in homozygous carriers of the chromosome 9p21 rs1333049 C risk allele in the contemporary percutaneous coronary intervention era: a prospective observa- tional study. BMJ Open 4 (8), e005438.

Hixson, J.E., Vernier, D.T., 1990. Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with HhaI. J. Lipid Res. 31 (3), 545–548.

Hughes, M.F., Saarela, O., Stritzke, J., et al., 2012. Genetic markers enhance coronary risk prediction in men: the MORGAM prospective cohorts. PLoS One 7 (7), e40922.

Jannes, J., Hamilton-Bruce, M.A., Pilotto, L., et al., 2004. Tissue plasminogen activator -7351C/T enhancer polymorphism is a risk factor for lacunar stroke. Stroke 35 (5), 1090–1094.

Jarinova, O., Stewart, A.F., Roberts, R., et al., 2009. Functional analysis of the chromo- some 9p21.3 coronary artery disease risk locus. Arterioscler. Thromb. Vasc. Biol. 29 (10), 1671–1677.

Kosokabe, T., Okumura, K., Sone, T., et al., 2001. Relation of a common methylenete- trahydrofolate reductase mutation and plasma homocysteine with intimal hyper- plasia after coronary stenting. Circulation 103 (16), 2048–2054.

Labos, C., Martinez, S.C., Leo Wang, R.H., et al., 2015. Utility of a genetic risk score to predict recurrent cardiovascular events 1 year after an acute coronary syndrome: a pooled analysis of the RISCA, PRAXY, and TRIUMPH cohorts. Atherosclerosis 242 (1), 261–267.

Larson, M.G., Atwood, L.D., Benjamin, E.J., et al., 2007. Framingham heart study 100K project: genome-wide associations for cardiovascular disease outcomes. BMC Med Genet. 8 (Suppl 1:S5).

Liu, Y., Sanoff, H.K., Cho, H., et al., 2009. INK4/ARF transcript expression is associated with chromosome 9p21 variants linked to atherosclerosis. PLoS One 4 (4), e5027.

Máchal, J., Pávková-Goldbergová, M., Hlinomaz, O., Groch, L., Vašků, A., 2014. Patients with chronic three-vessel disease in a 15-year follow-up study: genetic and non-ge- netic predictors of survival. Medicine (Baltimore) 93 (28), e278.

Makeeva, O.A., Zykov, M.V., Golubenko, M.V., et al., 2013. The role of genetic factors in the prediction of myocardial infarction complications within one year follow up. Kardiologiia 53 (10), 16–23 (in Russian).

McFadden, D., 1973. Conditional logit analysis of qualitative choice behavior. In: Zarembka, P. (Ed.), Frontiers in Econometrics. Academic Press, New York, pp. 105–142.

Mega, J.L., Stitziel, N.O., Smith, J.G., et al., 2015. Genetic risk, coronary heart disease events, and the clinical benefit of statin therapy: an analysis of primary and sec- ondary prevention trials. Lancet 385 (9984), 2264–2271.

Mulders, T.A., Meyer, Z., van der Donk, C., Kroon, A.A., Ferreira, I., Stehouwer, C.D., Pinto-Sietsma, S.J., 2011. Patients with premature cardiovascular disease and a po- sitive family history for cardiovascular disease are prone to recurrent events. Int. J. Cardiol. 153, 64–67.

Nelson, A.A., Tsao, H., 2009. Melanoma and genetics. Clin. Dermatol. 27, 46–52. Nowbar, A.N., Howard, J.P., Finegold, J.A., Asaria, P., Francis, D.P., 2014. 2014 global

geographic analysis of mortality from ischaemic heart disease by country, age and income: statistics from world health organisation and United Nations. Int. J. Cardiol. 174, 293–298.

Pang, X., Liu, J., Zhao, J., et al., 2014. Homocysteine induces the expression of C-reactive protein via NMDAr-ROS-MAPK-NF-κB signal pathway in rat vascular smooth muscle cells. Atherosclerosis 236 (1), 73–81.

Patel, R.S., Asselbergs, F.W., Quyyumi, A.A., et al., 2014. Genetic variants at chromosome 9p21 and risk of first versus subsequent coronary heart disease events: a systematic review and meta-analysis. J. Am. Coll. Cardiol. 63, 2234–2245.

Pavkova Goldbergova, M., Jarkovsky, J., Lipkova, J., et al., 2017. Relationship of long- term prognosis to MMP and TIMP polymorphisms in patients after ST elevation myocardial infarction. J. Appl. Genet. 58 (3), 331–341.

Peng, W.H., Lu, L., Zhang, Q., et al., 2009. Chromosome 9p21 polymorphism is associated with myocardial infarction but not with clinical outcome in Han Chinese. Clin. Chem. Lab. Med. 47 (8), 917–922.

Pereira, A.C., Miyakawa, A.A., Lopes, N.H., et al., 2007. Dynamic regulation of MTHFR mRNA expression and C677T genotype modulate mortality in coronary artery disease patients after revascularization. Thromb. Res. 121 (1), 25–32.

Roberts, R., 2014. Genetics of coronary artery disease. Circ. Res. 114, 1890–1903. Russo, G.T., Friso, S., Jacques, P.F., et al., 2003. Age and gender affect the relation be-

tween methylenetetrahydrofolate reductase C677T genotype and fasting plasma homocysteine concentrations in the Framingham offspring study cohort. J. Nutr. 133 (11), 3416–3421.

Shimo-Nakanishi, Y., Urabe, T., Hattori, N., et al., 2001. Polymorphism of the lipoprotein lipase gene and risk of atherothrombotic cerebral infarction in the Japanese. Stroke 32 (7), 1481–1486.

Simino, J., Shi, G., Bis, J.C., et al., 2014. Gene-age interactions in blood pressure reg- ulation: a large-scale investigation with the CHARGE, global BPgen, and ICBP con- sortia. Am. J. Hum. Genet. 95 (1), 24–38.

Sukhinina, T.S., Shakhnovich, R.M., Barsova, R.M., et al., 2012. Value of allele gene polymorphism of the inflammation system for prognosis of patients with myocardial infarction. Kardiologiia 52 (3), 15–21 (in Russian).

Szpakowicz, A., Kiliszek, M., Pepinski, W., et al., 2014. Polymorphism of 9p21.3 locus is associated with 5-year survival in high-risk patients with myocardial infarction. PLoS One 9 (8), e104635.

Tajbakhsh, A., Khorrami, M.S., Hassanian, S.M., et al., 2016. The 9p21 locus and its potential role in atherosclerosis susceptibility; molecular mechanisms and clinical implications. Curr. Pharm. 22 (37), 5730–5737.

The Comprehensive R Archive Network Available at. http://cran.r-project.org, Accessed date: 15 July 2017.

Thygesen, K., Alpert, J.S., Jaffe, A.S., Simoons, M.L., Chaitman, B.R., White, H.D., et al., 2012. Third universal definition of myocardial infarction. Eur. Heart J. 33 (20), 2551–2567.

Ulumbekova, G.E., 2010. Population health in the Russian Federation: risk factors and role of healthy nutrition. Vopr. Pitan. 79 (2), 33–38 (in Russian).

Weijmans, M., de Bakker, P.I., van der Graaf, Y., et al., 2015. Incremental value of a genetic risk score for the prediction of new vascular events in patients with clinically manifest vascular disease. Atherosclerosis 239 (2), 451–458.

Writing Group Members, Mozaffarian, D., Benjamin, E.J., et al., 2016. Executive sum- mary: heart disease and stroke statistics—2016 update: a report from the American Heart Association. Circulation 133 (4), 447–454.

Zeng, Q., Yuan, Y., Wang, S., Sun, J., Zhang, T., Qi, M., 2013. Polymorphisms on chro- mosome 9p21 confer a risk for acute coronary syndrome in a Chinese Han popula- tion. Can. J. Cardiol. 29 (8), 940–944.

G.J. Osmak et al. Gene 647 (2018) 283–288

288