Interplay between 3′-UTR polymorphisms in the methylenetetrahydrofolate reductase (MTHFR) gene and the risk of ischemic stroke

Stroke incidence is a multifactorial disease and especially hyperhomocysteinemia is associated with a higher risk of stroke. Previous studies have reported a folate metabolism disorder associated with the MTHFR gene. We investigated four single nucleotide polymorphisms in the MTHFR 3′-UTR [2572 C > A (rs4846049), 4869 C > G (rs1537514), 5488 C > T (rs3737967), and 6685 T > C (rs4846048)] to elucidate associations between ischemic stroke prevalence and prognosis. We examined 511 consecutive patients with ischemic stroke. Additionally, we selected 411 sex-/age-matched control subjects from patients presenting at our hospitals during the same period. The MTHFR 2572 C > A and 6685 T > C were significantly associated with ischemic stroke prevalence in the cardioembolism subgroup (MTHFR 2572CC vs. CA + AA: AOR, 2.145; 95% CI, 1.203–3.827; P = 0.010; MTHFR 6685TT vs. CC: AOR, 10.146; 95% CI, 1.297–79.336; P = 0.027). The gene-environment combined effect was significant, with MTHFR 2572CA + AA and folate levels ≤3.45 ng/mL correlating with ischemic stroke incidence. In addition, the total homocysteine (tHcy) levels in subjects with MTHFR 2572AA were elevated compared to tHcy levels in subjects with MTHFR 2572CC. Therefore, we suggest that MTHFR 2572 C > A and 6685 T > C are associated with ischemic stroke pathogenesis. The combined effects of the MTHFR 3′-UTR polymorphisms and tHcy/folate levels may contribute to stroke prevalence.

involved in Hcy and folate metabolism also play an important role in determining the susceptibility of an individual to disease 16 . Lower MTHFR enzyme activity, which can increase total plasma homocysteine (tHcy) levels and decrease plasma folate levels, contributes to stroke development [16][17][18] . Folate concentrations inversely correlate with tHcy levels 19 . The role of hyperhomocysteinemia in vascular and thromboembolic disease has been extensively studied. Previous studies reported significant vascular disease in patients with markedly elevated tHcy levels [20][21][22] . The tHcy is hypothesized to increase thrombotic risk by inducing endothelial injury in venous and arterial vasculatures 21 . Abnormal folate concentrations have also been implicated in the development of diseases, such as cardiovascular diseases, neural tube defects, cleft lip and palate, late pregnancy complications, and neurodegenerative and psychiatric disorders 23,24 . Recent studies have shown the clinical impacts of polymorphisms in the 3′-UTR of certain genes, which may potentially bind to specific microRNAs (miRNAs) in various diseases [25][26][27] . However, variants in the MTHFR 3′-UTR have not been extensively examined.
In the present study, we selected four single nucleotide polymorphisms (SNPs) in the MTHFR 3′-UTR region: MTHFR 2572 C > A (rs4846049), 4869 C > G (rs1537514), 5488 C > T (rs3737967), and 6685 T > C (rs4846048). We then determined their associations with ischemic stroke prevalence and prognosis. The minor allele frequencies of the studied polymorphisms are higher than 5% in the Asian population, and little is known about their genetic associations with ischemic stroke. Therefore, we investigated whether MTHFR 3′-UTR polymorphisms correlate with ischemic stroke susceptibility in Korean subjects.

Combined effects of MTHFR gene polymorphisms and clinical parameters on disease prevalence.
To determine additional clinical significance, we evaluated the combined effects of the gene environment. The MTHFR gene polymorphism elevated stroke prevalence in several conditions, including hypertension, diabetes mellitus, hyperlipidemia, smoking, high density lipoprotein-cholesterol levels, triglyceride levels, folate levels ≤3.45 nmol/mL, and tHcy levels ≥11.22 μmol/L ( Table 2). The MTHFR 2572 CA + AA genotype was shown to have synergic effects with ischemic stroke prevalence. In particular, low folate levels were the most predictive, with MTHFR 2572CA + AA and folate ≤3.45 nmol/L (AOR, 6.532; 95% CI, 2.592-16.46) shown to significantly increase ischemic stroke incidence. In addition, the MTHFR 4869CG + GG genotype had a combinatorial effect with hypertension (AOR, 3.217; 95% CI, 1.763-5.872) and smoking (AOR, 5.067; 95% CI, 1.788-14.360), whereas the MTHFR 5488CT + TT genotype was significant only in the smoking group (AOR, 2.740; 95% CI, 1.196-6.278). In addition, we performed stratified analyses for clinical factors including sex, age, hypertension, diabetes mellitus, hyperlipidemia, smoking status, folate levels, and homocysteine levels (Supplemental Table 2).

Differences of blood coagulation factors according to MTHFR polymorphisms. Analyses of var-
iance were used to show differences in blood coagulant factors (fibrinogen, antithrombin, platelet, activated partial thromboplastin time, and prothrombin time), folate, and tHcy levels according to genotype (Table 4). The tHcy levels in subjects with the MTHFR 2572AA polymorphism were elevated compared to the tHcy levels in subjects with the MTHFR 2572CC (P = 0.011). Additionally, the MTHFR 5488 and 6685 mutant genotypes had higher tHcy levels than that of the wild-type genotypes (MTHFR 5488, P = 0.020; MTHFR 6685, P = 0.005). Supplemental Table 4 shows the combination models with MTHFR 677 C > T, which measured the Hcy levels. The MTHFR 677-2572 combination group (677CC-2572AA: 12.14 ± 5.56; P = 0.010) and MTHFR 677-6685 combination group (677CC-6685CC: 14.61 ± 6.53; P < 0.0001) were significantly different compare to each wild type genotype.

Survival analysis of MTHFR polymorphisms and ischemic stroke mortality. During a mean
follow-up period of 7 years, 98 patients (17.5%) died. We investigated whether the MTHFR 3′-UTR genotypes were associated with long-term overall survival (OS) after ischemic stroke using Kaplan-Meier analyses. However, we did not find significant associations with individual MTHFR 3′-UTR genotypes (Supplemental Figure 2).

Discussion
Hcy is a well-known thrombotic factor in vascular diseases including coronary artery disease 28 , heart disease 29 , arteriosclerosis 3,15 , myocardial infarction 30 , venous thrombosis 31 , chronic kidney disease 32 , and ischemic stroke [33][34][35] . There is increasing evidence that Hcy may affect the coagulation system and the resistance of the endothelium to thrombosis 36 . Moreover, Hcy may interfere with the vasodilator and antithrombotic functions of nitric oxide 37 . Notably, vascular complications reported in patients with homocystinuria are related to thrombosis rather than atherosclerosis 38 , and a relationship between tHcy levels and the incidence of thrombotic events has recently been reported in patients with systemic lupus erythematosus 30 . Previous studies have identified the MTHFR gene as being associated with ischemic stroke prevalence [33][34][35][36][37] . Numerous studies reported that MTHFR 677 C > T was associated with increased stroke risk 34,35 , likely because the MTHFR 677 T allele decreased MTHFR gene activity 35 . Other studies reported that the methylation pattern of CpG island regions in the MTHFR gene had decreased MTHFR activity, causing an abnormality for tHcy and serum folate levels that were associated with ischemic stroke occurrence 36,37 . Moreover, we have shown an association with ischemic stroke risk based on computational epigenetic profiling of CpG islands in the MTHFR gene 37 . However, epigenetic regulation of the MTHFR gene could occur via another mechanism that included RNA interference with miRNA binding. In addition, a previous study showed differential mRNA expression levels of the MTHFR gene according to 3′-UTR polymorphisms 38,39 . Therefore, understanding the genesis of ischemic events due to decreased MTHFR activity might explain why 3′-UTR polymorphisms could affect ischemic events, stroke occurrence, and patient prognosis. Polymorphisms in the 3′-UTR region could affect mRNA stability and translation, which may significantly impact gene expression by abolishing, weakening, or creating miRNA binding sites. Currently, there are not sufficient data to indicate that miRNA binding activity is modulated depending on MTHFR 3′-UTR polymorphisms. One study reported that miR-149 binding activity was affected by the MTHFR 2572 C > A polymorphism in coronary heart disease risk 38 . Therefore, we investigated MTHFR 3′-UTR polymorphisms that have potential miRNA binding sites. We found that these regions were capable of binding with miRNA. Despite the lack of data, these miRNAs may be important genetic factors for the prevalence and progression of ischemic stroke because their expression is altered in some genotypes 38 .
In conclusion, we investigated the relationship of the MTHFR 2572 C > A and 6685 T > C polymorphisms with ischemic stroke incidence and progression. The MTHFR 2572 C > A and 6685 T > C polymorphisms were shown to increase the risk of embolisms of cardiac origin and ischemic stroke occurrence, and the prevalence of large-artery-origin ischemic stroke had a decreased odds ratio (OR) for the MTHFR 6685 T > C polymorphism. Moreover, the combination of MTHFR 2572CA + AA genotypes and serum folate levels or tHcy levels were synergic for ischemic stroke susceptibility, whereas other MTHFR 3′-UTR polymorphisms were not associated with serum folate and tHcy for ischemic stroke risk. Therefore, we hypothesized that the MTHFR 3′-UTR regulates one-carbon metabolism through miRNA binding. We found that these genotypes and haplotypes positively correlated with the occurrence and unfavorable prognosis of stroke, according to vascular disease risk factors, including hypertension, diabetes mellitus, HDL-C levels, tHcy levels, and folate levels.
This study has several limitations. First, the mechanisms by which 3′-UTR polymorphisms in the MTHFR gene affect stroke development remain unclear. Second, the controls in our study were not completely healthy   Table 3. Combined genotype and allele frequencies of the MTHFR 3′-UTR polymorphisms for ischemic stroke patients and control subjects. † The P-value was calculated by multiple logistic regression on the basis of risk factors such as age, gender, hypertension, hyperlipidemia, and diabetes mellitus. Abbreviations are defined in Table 1. ‡ The false discovery rate-adjusted P value for multiple hypothesis testing using the Benjamini-Hochberg method.
because some of them were seeking medical attention. However, the recruitment of only healthy participants for imaging and laboratory tests would markedly reduce the enrollment number, and enrollment of participants without imaging and laboratory tests may produce other biases in vascular risk factor assessment. Third, information regarding additional environmental risk factors in stroke patients remains to be investigated. Finally, the population of this study was restricted to patients of Korean ethnicity. Although the results from this study  provide the first evidence for 3′-UTR variants in MTHFR as potential biomarkers of stroke prevention and prognosis, a prospective study using a larger cohort of patients is warranted to validate these findings.

Methods
Ethics statement. All study protocols of participants were reviewed and approved by The Institutional function) with evidence of cerebral infarction in clinically relevant areas of the brain according to brain imaging using magnetic resonance imaging (MRI). The date and cause of death were identified using death certificates from the Korean National Statistical Office. Patients who were alive on Dec 31, 2012 were censored at that point. The death statistics of the Korean National Statistical Office have been previously reported to be reliable 40 . Based on clinical manifestations and neuroimaging data, two neurologists classified all ischemic strokes into four causative subtypes using the TOAST criteria as follows: (1) large-artery disease (LAD), significant (≥50%) stenosis of a relevant cerebral artery confirmed by cerebral angiography; (2) small-vessel disease (SVD), an infarction lesion <15 mm in diameter, and classic lacunar syndrome without evidence of a cerebral cortical dysfunction or potentially detectable cardiac sources for embolism; (3) cardioembolism (CE), presumably due to an embolus arising in the heart, as detected by cardiac evaluation; and (4) undetermined pathogenesis, in which the cause of stroke could not be determined or patients with two or more potential causes 41 . The frequencies of the stroke subtypes were 40% LAD (n = 205), 29% SVD (n = 149), 11% CE (n = 55), and 20% undetermined pathogenesis (n = 102). These proportions are similar to previously reported values for the Korean population 42 .
We selected 411 sex-and age-matched (±5 years) control subjects from patients presenting at our hospitals during the same period for health examinations, including biochemical testing, electrocardiogram analyses, and brain MRI. Control subjects did not have a recent history of cerebrovascular disease or myocardial infarction. Exclusion criteria were the same as those used in the patient group. The demographic and laboratory data of patients with ischemic stroke, subtype patients [LAD, SVD, and CE] and control subjects are summarized in Table 5. In our sample, 43.1% and 42.1% of stroke patients and control subjects were male, respectively. The mean ages of stroke patients and the control population were 62.96 ± 10.90 years and 62.82 ± 10.61 years, respectively. There were few significant differences between the two groups. Ischemic stroke patients were significantly more likely to have metabolic syndrome, as well as DM, hypertension, fibrinogen, increased tHcy levels, and decreased folate levels (P < 0.05).

Estimation of tHcy and folate levels.
Within 48 hours of stroke onset, we collected plasma samples to measure tHcy and folate levels. Twelve hours after the patient's previous meal, we collected whole blood in a tube containing anticoagulant. Tubes were centrifuged for 15 minutes at 1000 × g to separate the plasma. The tHcy concentrations were measured using a fluorescent polarizing immunoassay with the IMx system (Abbott Laboratories, Chicago, IL, USA), and folate concentrations were measured using a radioimmunoassay kit (ACS 180; Bayer, Tarrytown, NY, USA).
Genotyping. DNA was extracted using the G-DEX blood extraction kit (iNtRON Biotechnology, Inc., Seongnam, Republic of Korea). The four best-studied SNPs in the MTHFR gene were determined by a documentary search, which included four 3′-UTR SNPs (2572 C > A, rs4846049; 4869 C > G, rs1537514; 5488 C > T, rs3737967; and 6685 T > C rs4846048). All SNP sequences were obtained from the HapMap database (http:// www.hapmap.org). The MTHFR 2572 C > A and 4869 C > G polymorphisms were analyzed by the polymerase chain reaction-restriction fragment length polymorphism method. Real-time polymerase chain reaction (PCR) was used to analyze the MTHFR 5488 C > T and 6685 T > C polymorphisms. For each polymorphism, 30% of the PCR assay samples were randomly selected and repeated, and followed by DNA sequencing, to validate the RFLP findings. Sequencing was performed using an ABI 3730xl DNA Analyzer (Applied Biosystems, Foster City, CA, USA). The concordance of quality control samples was 100%.