Association Between VEGF Polymorphisms and Homocysteine Levels in Patients With Ischemic Stroke and Silent Brain Infarction

Introduction

Stroke is the third most common cause of death in many developed countries. Approximately 80% of strokes are ischemic in origin.^1–3 In South Korea, stroke is the most frequent cause of death after cancer and is more frequent than heart disease.⁴ Multiple factors, including hypertension, diabetes, smoking, hyperlipidemia, and hyperhomocysteinemia, are associated with a higher risk of stroke.³

Silent brain infarction (SBI) is defined as a cerebral infarction evident by brain imaging but without a clinical syndrome. SBI is characterized by the rapid development of clinical syndromes and signs of focal and, at times, global loss of brain function. SBIs are common with advanced age.^5,6 Although the clinical significance of SBI remains controversial, its presence can predict widespread vascular damage such as a clinically overt stroke.^7–9 Based on several lines of evidence, hyperhomocysteinemia is thought to be an independent risk factor for SBI.^10,11 Other metabolic syndromes, including hypertension and impaired fasting glucose, are also common risk factors for SBI.^5,12–14

Vascular endothelial growth factor (VEGF) is a major angiogenic factor and a prime regulator of endothelial cell proliferation.¹⁵ The gene that encodes VEGF is comprised of a 14-kb coding region with 8 exons and 7 introns located on chromosome 6.¹⁶VEGF undergoes transcriptional and posttranscriptional induction by hypoxia in the vicinity of tumor necrosis and in various models of ischemia.^17,18 Moreover, VEGF couples hypoxia to angiogenesis in diverse tissues, including the brain.^18,19 Because ischemia stimulates VEGF expression in the brain, VEGF may be important for the vascular response to cerebral ischemia.^20–22 Several single nucleotide polymorphisms (SNPs) have been described in the VEGF gene (National Center for Biotechnology Information, Gene association no: NT 007592). The VEGF gene includes at least 4 relatively common polymorphisms, -2578C>A (rs699947), -1154G>A (rs1570360), -634G>C (rs2010963), and 936C>T (rs3025039) SNPs that may influence VEGF expression.^23–25 Three of these polymorphisms are located in the promoter region at -2578C>A, -1154G>A, and -634G>C relative to the translation start site. The -2578A, -1154A, and -634G alleles are all associated with decreased VEGF expression.^23,24 In addition to promoter region polymorphisms, the T allele of the common 936C>T polymorphism in the 3′-untranslated region is associated with significantly decreased serum VEGF levels.²⁵ These VEGF polymorphisms are associated with various diseases, including recurrent abortion, pre-eclampsia, Alzheimer disease, colon cancer, breast cancer, and gastric cancer.^26–32 Recently, Yang et al³³ reported therapeutic effects of different doses of intranasal VEGF on angiogenesis and functional recovery of ischemic brains in adult rats. The VEGF -1154G>A (rs1570360) and -634G>C (rs2010963) polymorphisms in patients with ischemic stroke alone have been studied in the Chinese population,³⁴ although no significant findings were demonstrated. However, to our knowledge, the effects of the VEGF -2578C>A (rs699947) and 936C>T (rs3025039) polymorphisms on the risk of stroke and SBI have not been evaluated previously.

Hyperhomocysteinemia has been implicated as a risk factor for a number of vascular diseases, including ischemic stroke and SBI. Increased homocysteine (Hcy) concentrations are found in 40% of patients with coronary, cerebral, or peripheral artery diseases and only in 15% of healthy individuals.³⁵ Hcy inhibits endothelial cell proliferation and migration resulting in decreased angiogenesis.³⁶ However, the mechanism by which hyperhomocysteinemia induces the development of vascular lesions is still obscure.³⁷ The aims of this study were to evaluate the frequencies of VEGF -2578C>A, -1154G>A, -634G>C, and 936C>T polymorphisms in Korean patients with ischemic stroke and SBI and to determine the relationship between VEGF polymorphisms and plasma total Hcy (tHcy) levels.

Materials and Methods

Study Population

The study subjects were recruited between 2000 and 2008 from the Seoul and Kyeonggi-do provinces of South Korea. The Institutional Review Board of CHA Bundang Medical Center approved this genetic study in June 2000. We studied 615 consecutive patients with ischemic stroke referred from the Department of Neurology at CHA Bundang Medical Center, CHA University. Ischemic stroke was defined as a stroke (a clinical syndrome characterized by rapidly developing clinical symptoms and signs of focal and/or global loss of brain function) with evidence of a cerebral infarction in clinically relevant areas of the brain according to brain imaging by MRI scan and electrocardiography. Based on the clinical manifestations and neuroimaging data, 2 neurologists classified all ischemic strokes into 4 etiologic subtypes using the criteria from the Trial of Org 10172 in Acute Stroke Treatment (TOAST)³⁸ as follows: (1) large-artery disease, an infarction lesion ≥15 mm in diameter by MRI, and significant (>50%) stenosis of a major brain artery or a branch cortical artery by cerebral angiography with symptoms associated with that arterial territory; (2) small-artery occlusion (SAO), an infarction lesion <15 mm and ≥5 mm in diameter by MRI, and classic lacunar syndrome without evidence of a cerebral cortical dysfunction or potentially detectable cardiac sources for embolism; (3) cardioembolism, arterial occlusions presumably due to an embolus arising in the heart, as detected by cardiac evaluation; and (4) undetermined etiology, the cause of stroke could not be determined with any degree of confidence or involved >2 etiologies. The frequencies of stroke subtype were 40% (n=247) large-artery disease, 35% (n=214) SAO, 8% (n=47) cardioembolism, and 17% (n=107) undetermined etiology, respectively. These proportions are similar to reported values for the Korean population.³⁹ Our research focused on patients with small-artery disease. Single and multiple (≥2 lesions) SAOs were distinguished by brain MRI scan. The size and site of cerebral infarction were documented only by MRI.

We selected 376 patients with SBI (170 men and 206 women; age range, 40 to 88 years) who visited the CHA Bundang Medical Center. The diagnoses were made by MRI examination and by agreement between 2 independent experienced neurologists. All patients underwent a brain MRI scan and electrocardiography. The criteria for SBI were as follows: (1) spotted areas ≥3 mm in diameter in areas supplied by deep perforating arteries showing high intensity in the T2 and fluid-attenuated inversion recovery images and low intensity in the T1 image; (2) the absence of neurological signs and symptoms that could be explained by lesions observed by MRI; and (3) no history of clinical stroke, including transient ischemic attacks. Small punctate hyperintensities (1 to 2 mm in diameter) were likely to represent dilated perivascular spaces and were not considered in the present study. A SBI was excluded when an agreement could not be reached and patients with cerebral hemorrhage were excluded in advance. All examinations were performed according to the methods described previously.^27,40

We selected 494 control subjects matched for gender and age within 5 years from patients presenting at our hospitals for health examinations, which included biochemistry testing, an electrocardiogram, and brain MRI during the same period and who were free from a recent history of cerebrovascular disease or myocardial infarction. Exclusion criteria were the same as those used in the patient group, as mentioned previously. Hypertension was defined as a systolic pressure >140 mm Hg and/or a diastolic pressure >90 mm Hg on >1 occasion and includes patients currently taking hypertensive medications. Diabetes was defined as a fasting plasma glucose >126 mg/dL (7.0 mmol/L) and includes patients taking diabetic medications. Smoking refers to current smoking. Hyperlipidemia was defined as a high fasting serum total cholesterol level (≥240 mg/dL) or an antihyperlipidemic agent treatment history. Some of the study subjects (n=89 [6%]) were found to have new vascular risk factors such as hypertension or diabetes mellitus at the time of examination. The demographic and laboratory data are summarized in Table 1. Significant differences were detected in patients with hypertension, diabetes mellitus, and hyperlipidemia between the stroke and control groups. Although patients with SBI had a significantly higher prevalence of all conventional vascular risk factors relative to control subjects, patients with SBI also had significantly higher plasma tHcy levels and a greater prevalence of hyperlipidemia than control subjects (Table 1).

Genotype Determination of VEGF Polymorphisms

DNA was extracted using the G-DEX blood extraction kit (iNtRON Biotechnology, Inc, Seongnam, South Korea). The 4 best-studied SNPs in the VEGF gene were determined by a documentary search, which included 3 5′-untranslated region SNPs (-2578C>A, rs699947; -1154G>A, rs1570360; and -634G>C, rs2010963) and 1 3′-untranslated region SNP (936C>T, rs3025039). All SNP sequences were obtained from the HapMap database (www.hapmap.org). The VEGF -2578C>A and 936C>T polymorphisms were analyzed by the polymerase chain reaction–restriction fragment length polymorphism method. Real-time polymerase chain reaction was used to analyze the VEGF -1154G>A and -634G>C polymorphisms. The primers and polymerase chain reaction conditions for VEGF polymorphism analyses have been previously described.^27,28

Estimation of Hcy and Folate Levels

Plasma samples were collected to measure the levels of tHcy and folate within 48 hours of the onset of a stroke or SBI. Blood was collected in a tube containing anticoagulant 12 hours after a patient's previous meal. The tube was centrifuged for 15 minutes at 1000 g, and the plasma was separated. The concentration of Hcy in the plasma was measured by fluorescent polarizing immunoassay with IMx (Abbott Laboratories, Chicago, IL). The plasma concentration of folate was determined using a radioassay kit (ACS 180; Bayer, Tarrytown, NY).

Statistical Analysis

The associations among ischemic stroke, SBI, and VEGF genotypes were estimated by computing the ORs and 95% CIs from Fisher exact test. The adjusted ORs for VEGF polymorphisms were determined from multiple logistic regression analysis using gender, age, diabetes mellitus, hypertension, hyperlipidemia, and smoking. Stratification analysis was used to stroke subgroups according to the size of the occluded vessel. One-way analysis of variance was performed to compare the mean levels of Hcy concentrations among different genotypes. We carried out multiple hypotheses testing using the Benjamini-Hochberg method to control for false discovery rate (FDR) in the unconditional logistic regression analysis.⁴¹ Calculation of the FDR is a way to address the problems associated with multiple comparisons and provides a measure of the expected proportion of false-positives among data. Statistical significance was accepted at the P<0.05 level. StatsDirect Statistical Software (Version 2.4.4; StatsDirect Ltd, Altrincham, UK) was used to calculate the adjusted AOR and 95% CI. The linkage disequilibrium between loci was measured using the absolute value of Lewontin D′.⁴² Haplotype frequencies for multiple loci were estimated using the expectation-maximization algorithm with SNPAlyze (Version 5.1; DYNACOM Co, Ltd, Yokohama, Japan).

Results

A comparison of genotype frequencies of the VEGF -2578C>A, -1154G>A, -634G>C, and 936C>T polymorphisms between the patients with stroke and those with SBI and control groups is shown in Table 2. Genotype distributions of each polymorphism did not deviate from those expected based on the Hardy-Weinberg equilibrium in the 3 groups. The linkage disequilibrium of the VEGF polymorphisms at loci -2578(rs699947)/-1154(rs1570360)/ -634(rs2010963)/936(rs3025039) in patients with ischemic stroke and those with SBI is shown in the Figure. There was strong linkage disequilibrium between loci -1154 and -634 (D′=0.819) and -2578 and -634 (D′=0.807) in patients with ischemic stroke (Figure 1A). Polymorphisms -2578C>A and -634G>C were in strong linkage disequilibrium in the patients with SBI (D′=0.880; Figure 1B).

• Download figure
• Download PowerPoint

The CT+TT (T allele-bearing) genotype frequency of the 936C>T polymorphism (FDR-adjusted OR, 1.44; 95% CI, 1.09 to 1.90; P=0.044) and the AA genotype and CC+CA (C allele-bearing) genotype frequencies of the -2578C>A polymorphism (FDR-adjusted OR, 2.13; 95% CI, 1.27 to 3.59; P=0.016 and FDR-adjusted OR, 1.93; 95% CI, 1.17 to 3.20; P=0.044, respectively; Table 2) were significantly different between the stroke and control groups. The -634 polymorphism in SBI showed a significantly lower frequency of variant genotypes (GC and CC) compared with control groups (adjusted OR, 0.71; 95% CI, 0.52 to 0.97 for GC and adjusted OR, 0.61; 95% CI, 0.40 to 0.94 for CC, and adjusted OR, 0.69; 95% CI, 0.51 to 0.93 for GC+CC compared with the GG genotype; Table 2). However, after measurement by multiple hypothesis testing, the frequency difference for the GC+CC (C allele-bearing) genotype of the -634G>C polymorphism was marginally significant between the SBI and control groups (FDR-adjusted P=0.056). Interestingly, when the data were stratified by the size of the occluded vessel, patients with SAOs, especially multiple SAOs, were associated with the -2578C>A, -1154G>A, and 936C>T polymorphisms (Table 3). Furthermore, the adjusted OR values of multiple SAOs were much higher than those of SAOs. The -634CC genotype had a lower adjusted OR value (2.27-fold) than the GG genotype in the patients with multiple SAOs. We did not find any significant association between VEGF polymorphisms and patients with single SAOs.

We also constructed haplotypes of the -2578C>A, -1154G>A, -634G>C, and 936C>T polymorphisms of the VEGF gene (Table 4). Several haplotype frequencies were significantly different between the control subjects and patients with stroke with SAOs, single SAOs, and multiple SAOs. The A-A-G-T, C-A-C-C (-2578/-1154/-634/936), C-A-C (-2578/-1154/-634), and A-C-C (-1154/-634/936) haplotypes had significant differences in the 3 SAO groups. Several haplotype frequencies were also significantly different between the SBI patient and control groups. The OR values of VEGF haplotypes between the patients with stroke and those with SBI and the control subjects are shown in Supplemental Table I (http://stroke.ahajournals.org). The OR values of the VEGF haplotypes among the stroke subtypes and control subjects are presented in Supplemental Table II. A comparison of genotype frequency and adjusted OR of VEGF polymorphisms between the stroke subtypes and patients with SBI is shown in Supplemental Table III. The frequency differences for the -2578C>A, -1154G>A, and 936C>T polymorphisms were marginally significant between the multiple SAO subtypes and the patients with SBI.

We sought to determine whether VEGF polymorphisms were associated with plasma tHcy levels (Table 5). The -634G>C and 936C>T polymorphisms were associated with tHcy levels in patients with multiple and single SAOs, respectively. However, the VEGF polymorphisms did not show any association with tHcy levels in the SBI group. Plasma tHcy levels of the VEGF -2578C>A, -1154G>A, -634G>C, and 936C>T haplotypes among the patients with stroke, patients with SBI, stroke subtypes, and control subjects are shown in Supplemental Table IV. The plasma tHcy levels of the VEGF haplotypes between the ischemic stroke subtypes and control subjects are presented in Supplemental Table V. Some haplotypes showed significant differences between the groups.

A total of 1485 individuals (615 patients with stroke, 376 patients with SBI, and 494 control subjects) from 2 different case–control samples, Sample 1 and Sample 2, were analyzed according to recruitment duration (Supplemental Table VI). The genotype frequencies of the VEGF polymorphisms were significantly different between the control, ischemic stroke, and SBI groups in Samples 1 and 2 (Supplemental Tables VII to XI). These results suggest that VEGF is also a candidate susceptibility gene of ischemic stroke, although the association of VEGF -2578C>A, 936C>T, and ischemic stroke was not replicated in subjects from Sample 1 and Sample 2 by multivariable logistic regression analysis.

Discussion

Angiogenesis is critical to the progression of atherogenesis, collateral vessel development in ischemia, and plaque instability.^43,44 VEGF is believed to be important for initiating angiogenesis and is a major mediator of the progression of atherothrombotic vascular disease, including ischemic stroke. The effects of VEGF on the risk of stroke have been suggested in a number of biological and pathological studies. For example, it has been suggested that the VEGF/VEGF receptor system, which is induced by hypoxia, leads to growth of new vessels after cerebral ischemia. Exogenous support of this natural protective mechanism might lead to enhanced survival after stroke.⁴⁵ Sun et al⁴⁶ have suggested that, in the ischemic brain, VEGF exerts an acute neuroprotective effect as well as longer-lasting effects on the survival of new neurons and on angiogenesis and that these effects may operate independently. Verheul et al⁴⁷ reported that VEGF-stimulated human umbilical endothelial cells promote the adhesion and activation of platelets. They also found that activated platelets are present in the microvessels of VEGF-producing soft tissue sarcomas.⁴⁸ There is strong evidence to support a close relationship between VEGF and ischemic stroke. Intranasal administration of VEGF may induce angiogenesis in the ischemic boundary and improve behavioral recovery after cerebral ischemia in rats.⁴⁹ Astrocytes, which morphologically resemble injury-induced VEGF-positive cells, also react to injury by increasing VEGF expression, indicating that VEGF might participate in the central nervous system response to injury.⁵⁰ Therefore, VEGF may improve the histological and functional outcomes of stroke through multiple mechanisms.

SBI is a kind of cerebral infarction event. Despite the functional studies of VEGF described here, the effect of VEGF polymorphisms on the risk of stroke and SBI has not been reported. Recently, several investigators performed genomewide association studies and meta-analysis of the genetic susceptibility to stroke in Asian populations^51–54; however, they did not find an association with VEGF in Asian populations. Based on the known biological and pathological significance of VEGF, it is reasonable to hypothesize that VEGF is a good candidate for determining the risk of developing a stroke and SBI. In the present study, although only the VEGF -2578C>A and 936C>T polymorphisms were associated with the risk of ischemic stroke, the -2578C>A, -1154G>A, and 936C>T polymorphisms were associated with SAOs when the data were stratified by the size of the occluded vessel. Moreover, when patients with SAOs were divided into single and multiple SAOs by brain MRI, variant alleles of the -2578C>A, -1154G>A, and 936C>T polymorphisms were only significantly different in patients with multiple SAOs compared with control subjects. Although we do not know the exact causes of stroke, research on the etiologic heterogeneity and subtypes of stroke has been performed.^55–57 Some evidence suggests that there are different pathophysiological mechanisms for single and multiple SAOs.^58,59 Therefore, despite the heterogeneity of ischemic strokes, our data suggest that VEGF polymorphisms are an independent risk factor for multiple SAOs. Several articles have shown that the MTHFR 677C>T polymorphisms are associated with multiple infarctions.^60–62

Haplotype analysis in this study revealed that the frequencies of several haplotypes were significantly different between the control subjects and patients with subtypes of stroke (SAOs, single SAOs, and multiple SAOs) and SBI. Thus, the haplotypes of VEGF polymorphisms provide data for susceptibility to stroke and SBI.

SBI shares a close similarity with pathophysiological aspects of single SAOs. Thus, we can expect that their genetic compositions are also similar. As shown in Supplemental Table III, there were no significant differences in the -2578C>A, -1154G>A, -634G>C, and 936C>T polymorphisms between SBI and SAOs and single SAO subtypes. However, the genotype frequencies of VEGF polymorphisms, except for -634G>C in SBI, were only marginally different from ones in multiple SAOs, suggesting a difference in the pathophysiological mechanisms of SBI and multiple SAOs.

Angiogenesis is regulated by a balance of various cytokines and biological molecules. The final outcome does not occur through the independent actions of these factors, but rather depends on the relative input of each factor. Hcy and VEGF have been implicated in angiogenesis and in the development and progression of atherothrombotic vascular disease.^63,64 Ischemic stroke is a multifactorial disorder in which genetic and environmental factors, including defects in Hcy metabolism, play a major role. Hcy inhibits angiogenesis in vitro and in vivo.⁶⁵ Hcy increases the expression of VEGF through a mechanism involving endoplasmic reticulum stress and the transcription factor ATF4⁶⁶ and in differentiated THP-1 macrophages.⁶⁷ Shastry et al⁶⁸ have reported that Hcy inhibits angiogenesis, partly by decreasing VEGF, as shown in an experiment using mouse brain microvascular endothelial cells. Furthermore, Atta et al⁶⁹ reported that lowering Hcy levels with vitamin B and folic acid results in a substantial reduction of VEGF plasma levels in patients with peripheral arterial disease or diabetes mellitus. Thus, it is possible that VEGF genotypes are associated with circulating tHcy levels.

In this study, we found that the VEGF -634G>C and 936C>T genotypes were only associated with tHcy levels in patients with multiple and single SAOs, respectively, suggesting that VEGF polymorphisms may weakly influence tHcy levels. However, the -2578C>A and -1154G>A polymorphisms were not associated with tHcy levels. Therefore, VEGF polymorphisms showed a weak association with tHcy levels in Koreans. However, there have been no prior reports of an association between VEGF polymorphisms and tHcy levels in any population. One possible explanation for this is that differences in tHcy levels and the prevalence of ischemic stroke may exist among various ethnic populations due to environmental factors such as dietary habits, daily folate intake, and lifestyle. Thus, we cannot exclude strong associations between VEGF polymorphisms and tHcy levels in other ethnic populations.

Hypoxia is a potent stimulus for VEGF expression in vivo and in vitro. Hypoxia-induced proteins bind to the 3′-untranslated region of the VEGF mRNA, resulting in a significantly increased half-life of the mRNA.⁶⁶ One explanation for our results is that the 936C>T polymorphism, located in the 3′-untranslated region, leads to the loss of a potential binding site for AP-4.^31,70 Posttranscriptional regulation could also affect not only the VEGF gene, but also other hypoxia-inducible genes such as erythropoietin or tyrosine hydroxylase.⁷¹

There are several limitations of the present study: (1) as noted, it is not yet clear which genetic polymorphisms predict the phenotypes associated with ischemic stroke and SBI. The study population comprised of only Korean individuals, and our findings will need to be validated in other ethnic groups; (2) this was a hospital-based case–control study, which had a relatively small sample size of individual stroke subtypes. However, we think that the recruitment of >1000 individuals from an ethnically homogeneous population (Koreans have a low degree of interracial marriage) is enough to give reliable data; (3) control subjects were not recruited from a completely “healthy” population, because some of them were seeking medical attention. Those who agreed to diagnostic evaluation differed from those who did not agree. Therefore, it would not be easy to identify the casual effects of vascular risk factors in these subjects. Our results may underestimate the true impact of individual risk factors based on a selection bias. However, in our experience, recruitment of healthy participants markedly reduces recruitment rates because of the refusal of laboratory and imaging studies. Only interview-dependent risk factor assessment without laboratory and imaging studies may fail to find covert risk factors and asymptomatic lesions (for example, SBI), leading to another potential for bias. In the present study, approximately 6% of the study subjects were found to have a new vascular risk factor through laboratory tests at the time of examination. All of our study subjects had brain imaging and were confirmed by negative symptomatic brain lesion, thereby enhancing the diagnostic accuracy and exact case–control grouping. Although a population-based study may be optimal to reduce the referral or selection bias, it is difficult to obtain sufficient numbers of stroke incidents among the cohort because the estimated annual incidence of stroke is known to be low in the general population; and (4) our results cannot be extrapolated to other races because interethnic variability in frequencies of stroke subtypes and genotypes may produce different results.

The genotype and allele frequencies of the VEGF polymorphisms may vary among different populations. For example, Park et al²⁹ reported a comparison of VEGF polymorphism data in healthy populations obtained from various studies, finding that the frequency of the VEGF -2578A allele was 0.378 to 0.504 in whites and 0.276 to 0.280 in Asians. For the VEGF-1154A allele, the frequencies were 0.30 to 0.32 in whites and 0.18 in Koreans,²⁶ suggesting that there is a racial difference in the allele frequencies of the -2578C>A and -1154G>A polymorphisms. Therefore, additional studies involving different racial or ethnic groups or samples of populations of homogeneous origin are needed to confirm our results.

In conclusion, the VEGF polymorphisms were associated with the risk of ischemic stroke, particularly in patients with multiple SAOs. The adjusted OR values of the -2578A, -1154A, -634G, and 936T alleles, which are related to abnormal VEGF expression levels, were much higher in multiple SAO patients than in the single SAO group. The VEGF polymorphisms were significantly different between patients with SBI and multiple SAOs. These findings suggest that VEGF polymorphisms are a genetic determinant for the risk of multiple SAOs in the Korean population. Further studies of other racial or ethnic populations and of the biological functions of VEGF are needed to fully understand the role of VEGF polymorphisms in the risk of multiple SAOs in patients with ischemic stroke and patients with SBI.

Sources of Funding

This work was partly supported by National Research Foundation of Korea Grant funded by the Korean Government (2009-0070341) and partly supported by Priority Centers Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2009-0093821).

Disclosures

None.

Footnotes