【作者】 刘建平；

【导师】 聂绍发； 程锦泉；

【作者基本信息】 华中科技大学 ， 流行病与卫生统计学， 2008， 博士

【摘要】 研究背景缺血性脑卒中是受遗传与环境因素共同作用的多因子疾病,基因-基因、基因-环境交互作用在缺血性脑卒中的发病中具有重要影响。随着人类基因组计划的完成,从基因水平阐明缺血性脑卒中的发病机制成为医学界关注的热点。近年来研究表明eNOS、GH1、IGF-1R基因多态性能显著影响基因的表达,这可能与缺血性脑卒中、冠心病的发生密切相关。但是关于eNOS、GH1、IGF-1R多态性与缺血性脑卒中之间的关系,以及该多态性与环境因素之间交互作用对缺血性脑卒中的影响,国内外罕见报道。研究目的1.探讨eNOS基因-922A/G、T-786C及G894T基因多态性与缺血性脑卒中的关系。2.探讨eNOS基因多态性与环境因素之间的交互作用在缺血性脑卒中发病中的地位与作用。3.探讨GH1基因T1663A多态性与缺血性脑卒中的关系,同时研究其与相关环境暴露因素之间的交互作用在缺血性脑卒中发病中的作用。4.探讨IGF-1R基因G/A(rs2229765)、A/G(rs951715)、A/G(rs2593053)多态性与缺血性脑卒中之间的关系。5.探讨IGF-1R基因多态性与相关环境暴露因素之间交互作用与缺血性脑卒中的关系。6.采用分类树统计分析方法初步构建缺血性脑卒中发病风险的预测模型。研究方法本研究采用候选基因和病例对照的研究方法,收集深圳市两家大型综合性医院的309名缺血性脑卒中新发病例,按年龄相差小于5岁、性别、民族相同的匹配条件选取对照,开展以医院为基础的1︰1配对病例-对照研究。采用统一的调查问卷对病例和对照进行调查,并按相同的条件和标准采集病例和对照的血液样本。用Taqman MGB荧光定量PCR技术分析基因多态性的基因型。运用单因素及多因素logistic回归分析基因多态性与缺血性脑卒中之间的关系,运用PHASE2.0软件进行单体型分析,用相加模型分析基因与环境相关危险因素之间的潜在交互作用,最后运用分类树分析方法初步构建缺血性脑卒中发病风险的预测模型。主要研究结果1.单因素Logistic回归分析结果表明:文化程度、体质指数、腰臀比、吸烟、高血压、糖尿病、负性生活事件、甘油三酯等是缺血性脑卒中发病的危险因素。在多变量Logistic回归模型中,吸烟(OR=5.42;95%CI:2.00~14.63)和高血压(OR=3.51;95%CI:1.83~6.71)是缺血性脑卒中发病的正关联因素;而体育锻炼(OR=0.10;95%CI:0.05~0.22)、饮茶史(OR=0.25;95%CI:0.12~0.55)是缺血性脑卒中发病的负关联因素。2. eNOS基因T-786C多态性基因型分布差异没有统计学意义(P=0.132),按性别进行分层后,男性病例组与对照组基因型分布差异处于临界值水平(P=0.053);在未调整混杂因素时,携带CC基因型的个体患缺血性脑卒中风险为TT 3.819倍,P=0.029;在调整上述因素的影响后,携带CC基因型的个体患缺血性脑卒中风险为TT的4.533倍,P=0.047。eNOS基因A-922G多态性病例组的G等位基因频率高于对照组,差异有统计学意义(P=0.018),基因型分布进行线性趋势检验,χ2=4.886, P=0.027,可见随着等位基因G的增加,发生缺血性脑卒中的危险性也增高;在调整上述因素影响后,A-922G多态性仍是缺血性脑卒中发病的危险因素, OR值为2.156,P=0.029。3. GH1基因T1663A多态性等位基因频率及基因型频率分布差异无统计学意义,P值分别为0.124和0.358;多因素Logistic回归分析发现,GH1基因多态性与缺血性脑卒中发病无关联。4. IGF-1R基因G→A多态性,病例组A等位基因频率(50.00%)高于对照组(31.93%),等位基因频率分布差异有统计学意义(P=0.001);以GG基因型为参考基因型,在调整其他混杂因素的影响后,携带AA基因型的个体缺血性脑卒中发病风险增加,OR=1.992,P=0.015。对于A/G(rs951715)多态性,病例组G等位基因频率(56.15%)高于对照组(43.85%),P=0.001,以AA基因型为参照,携带AG基因型的个体缺血性脑卒中的发病风险增加,OR值为2.201,P=0.000;当调整其他因素影响后,AG基因型仍是缺血性脑卒中发病的危险因素,OR值为2.381,P=0.000。5. T-786C多态性与高血压家族史、糖尿病、糖尿病家族史及吸烟存在正相加交互作用,S为3.76、3.10、4.22和1.63;A-922G基因多态性与饮酒存在负相加交互作用,S为0.43,与高血压家族史,糖尿病家族史和吸烟存在正相加交互效应,S分别为2.46、3.24、1.99;调整混杂因素后,GH1 AT基因型与超重存在交互效应,P为0.025,OR值为4.06。未发现IGF-1R基因多态性与环境因素存在明显的交互作用。6.采用分类树构建脑卒中发病风险模型,分类树模型共包括4层,共筛选出6个解释变量。采用筛检试验评价指标对模型的灵敏度和特异度进行评价,结果发现灵敏度为76.70%,特异度为81.88%,约登指数为58.58%。研究结论1.传统的危险因素仍是目前深圳市汉族人群中缺血性脑卒中发生的主要原因,因此在人群中培养健康的生活方式,早期、及时地控制高血压、糖尿病、血脂和体重是预防缺血性脑卒中的主要措施。2. eNOS基因T-786C、A-922G多态性和IGF-1R基因G/A(rs2229765)、A/G(rs951715)多态性均于缺血性脑卒中遗传易感性显著相关。3. eNOS基因T-786C、A-922G、GH1基因T1663A多态性与环境因素如吸烟、高血压、家族史等之间在缺血性脑卒中患病中存在不同程度的交互作用。4.分类树模型能够较好地拟合缺血性脑卒中发病风险的预测模型。5.缺血性脑卒中是由许多微效基因协调作用并与环境因素共同作用的结果,研究它们之间的相互关系对阐明缺血性脑卒中的病因及发病机理有重要意义。更多还原

【Abstract】 BackgroundIschemic stroke (IS) is a multi-factorial disease, which is related to both the genetic and environmental factors. Gene-gene and gene-environmental interaction makes great contribution to the risk of IS. With the advance of Human Genome Project, it has become more and more popular in medical fields to clarify the pathogenesis of IS at genetic level, especially IS. In recent years, studies have suggested that endothelial nitric oxide synthase (eNOS), growth hormone(GH) and insulin-like growth factor-I receptor (IGF-1R) gene, which can significantly affect gene expression, have been associated with IS and coronary heart disease (CHD). However, reports are extremely rare presently on the association between the above-mentioned genes and IS in the Han nationality of China, and the interaction between the gene polymorphisms and environmental factors.Objectives1. To explore the association between the T-786C, A-922G and G894T polymorphism in eNOS gene and IS.2. To explore and assess the possible interaction effects between eNOS gene polymorphisms and environmental factors.3. To explore the association between GH1 T1663A gene polymorphism and IS, in addition, to assess the interaction between GH1 T1663A gene polymorphism and IS.4. To explore the possible association between IGF-1R gene G/A (rs2229765), A/G (rs951715) and A/G (rs2593053) polymorphisms and IS.5. To explore and assess the possible interaction effects between the IGF-1R gene polymorphisms and environmental factors.6. Classification tree model was applied to build the risk model for IS.MethodsCandidate genes and case-control study were used to determine the possible the association between gene and IS. A 1:1 matched case-control study was performed. 309 cases were those onset IS patients registered in two general hospitals in Shenzhen. The controls were selected by the same gender and ethnic group, and each pair’s ages were permitted to differ within 5 years. The cases and controls were interviewed using the same questionnaire, and the blood samples were drawn in terms of the same conditions and standards. Gene polymorphisms were determined by using Taqman MGB genotyping assay. Univariate test and multiple logistic regression models were used to explore the association between the above-mentioned genes polymorphisms and IS. Haplotype analyses of these polymorphisms were performed using PHASE2.0 software. Additionally, the interaction between genes and environmental risk factors were assessed by multivariate logistic regression model. The odds ratio values (OR) was calculated by using regression model to determine the addition effects among different factors and measure the interaction. Finally, Classification tree model was applied to build up the risk model for IS.Results1. Univariate logistic regression demonstrated that risk factors of IS included income, education, body mass index (BMI), WHR, smoking, hypertension, diabetes mellitus (DM), negative events and TG levels (triglyceride). In multivariate logistic model, smoking and hypertension were positively associated with IS, with OR=5.42 (95%CI: 2.00~14.63) and OR=3.51 (95%CI: 1.83~6.71) respectively, while moderate physical training and history of tea-drinking were inversely associated with IS, with OR=0.10 (95%CI: 0.05~0.22) and OR=0.25 (95%CI: 0.12~0.55), respectively.2. For eNOS T-786c polymorphism, there were no significant difference in the distributions of genotypes between two groups (P=0.132). Stratified by sex, the p values for the genotypes of the above mentioned polymorphism was 0.053 in male. Conditional logistic regression revealed that the CC genotype of eNOS was associated with IS (OR=3.819, P=0.029). After adjustment for confounding factors, eNOS CC genotype was still significant associated with IS (OR=4.533, P=0.047). For eNOS A-922G polymorphism, the frequency of eNOS -922 G allele was significant higher in the patients than the controls (12.14% vs 8.09%, P=0.018). Linear tendency test showed the risk for development of IS raised with increasing G allele (chi-square value=4.886, P=0.027). After adjustment for confounding factors, eNOS A-922G polymorphism was significant associated with IS (OR=2.156, P=0.029).3. There were no significant difference in the distributions of allele and genotypes in GH1 gene T1663A polymorphism between two groups (P=0.124 and 0.358, respectively.). Multiple logistic regressions revealed that the GH T1663A polymorphism may not be an additional risk factor for the development of IS.4. For IGF-R G/A (rs2229765) polymorphism, the frequency of IGF-1R A allele was significant higher in the patients than the controls (50.00% vs 31.93%, P=0.001). After adjustment for confounding factors, AA genotype was significant associated with an increased risk of developing IS with the GG genotype as reference genotype (OR=1.992, P=0.015). For IGF-R A/G (rs951715) polymorphism, the frequency of IGF-1R G allele was significant higher in the patients than the controls (56.15% vs 43.85%, P=0.000). Compared with AA genotype, AG genotype was significant associated with an increased risk of developing IS without adjusting for other confounding factors (OR=2.201, P=0.000). After adjustment for confounding factors, AG genotype was significant associated with an increased risk of developing IS (OR=2.381, P=0.000).5. The positive additive interactions were found between eNOS gene T-786C polymorphism and family history of hypertension, DM, family history of DM and smoking, synergy Index (S) were 3.76, 3.10, 4.22 and 1.63, respectively. The interactions analysis between A-922G polymorphism and environmental factors indicated that a negative additive interaction was found between A-922G and alcohol drinking (S=0.43), and the S for family history of hypertension, family history of DM and smoking were 2.46, 3.24 and 1.99, respectively. No interactions between IGF-1R and environmental factors were found.6. Classification tree model was applied to build up the risk model for IS, and the model had four stratum. Six explanatory variations were screened out in our model. The indexes of the screening test were used to evaluate the fitness of the model. The results revealed that sensitivity and specificity and Youden index were 76.70%, 81.88% and 58.58%, respectively.Conclusions: 1. The classical risk factors are still main reasons of patients with IS in the Han population of Shenzhen city. Therefore it is an important measure to prevent IS in community population to propose healthy life style including proper exercise, control of high blood pressure, high blood fat and weight.2. The T-786C, A-922G polymorphism in eNOS gene and G/A (rs2229765) and A/G (rs951715) polymorphisms in IGF-1R are all significant associated with the hereditary susceptibility of IS in the Han population of Shenzhen city.3. There are obvious interactions between T-786C and A-922G polymorphisms in eNOS gene and environmental factors such as smoking, alcohol drinking, hypertension, diabetes and family history vascular diseases.4. Classification tree model can properly predict the occurrences of IS.5. IS is caused by the interactions between many minor genes and environmental risk factors. It is very important to study their correlation to classify the cause and pathogenesis of IS.更多还原