【作者】 沈冲；

【导师】 顾东风；

【作者基本信息】 中国协和医科大学 ， 流行病与卫生统计学， 2008， 博士

【摘要】 研究背景和目的当今,心脑血管疾病已经成为世界大多数发达国家乃至发展中国家(如中国)居民的首位死因。而血压升高是脑卒中和冠心病发病的首要独立危险因素,因此,预防和控制高血压是预防心脑血管病的主要措施。近半个世纪来,由于我国经济发展,人民生活改善和生活节奏加快带来的一系列不健康生活方式,我国人群高血压患病率上升很快。尤其值得关注的是,年轻人群和农村人群的高血压患病率迅速增加。目前,高血压已成为危害中国人群健康的严重公共卫生问题。因此,开展高血压发病机制和人群防控的研究具有十分重要的意义。成人高血压中绝大多数是原发性高血压(Essential hypertension,EH)。EH是一种复杂的多基因疾病,环境因素与遗传因素都对疾病的发生起到了重要的作用。遗传因素通过多个微效基因(Minor-effect genes)及其和环境因素共同作用而导致血压升高。目前,广泛采用基于人群的关联研究来探索遗传标志与高血压这一类常见复杂性状(Complex trait)疾病的关系。动脉结构与功能改变是原发性高血压的重要病理基础。血管壁弹性降低/失活、血管内皮损伤相关的周围阻力增加和动脉血管重塑,与血压持续升高及某些器官损害密切相关,但具体作用机制目前尚未明确。为了探讨血管发育和重塑与EH发生的关系,本研究通过对参与血管发育和维持血管形态的生物信息学分析,确定了参与转化生长因子β(Transforminggrowth factor-β,TGF-β)信号通路的几个关键功能候选基因,包括血管弹性微纤维膜——原微纤维蛋白(FBN1)基因、弹性纤维胞外基质组件(EMILIN1)基因、抑制TGF-β前体(proTGF-β)水解的转化酶FURIN基因、血管发育和重塑过程中控制TGF-β释放的凝血酶敏感素1结构域包含体1(THSD1)基因、编码TGF-β1的TGFB1基因和TGF-β1受体(TGFBR1)基因。应用tagSNP选择和功能预测分析相结合的方法,从覆盖功能候选基因区域的人类遗传变异的重要标志——单核苷酸多态性(Single nucleotide polymorphism,SNP)中选择有代表性的SNPs,通过以人群为基础的两阶段关联研究筛选高血压易感SNPs,并分析了可能存在的基因-基因和基因-环境交互作用,研究结果将有助于揭示中国人EH发病机制及其特点,并为EH的药物基因组学研究以及人群防制提供理论依据。研究对象与方法1、所有研究对象的DNA样本和临床资料均来自亚洲国际心血管疾病协作研究(International collaborative study of cardiovascular disease in Asia,InterASIA)的中国部分。本研究从四阶段分层抽样调查获得的中国北方四个省市:北京、吉林、山东、陕西的35-74岁无亲缘关系人群中选择了1317名收缩压(Systolic bloodpressure,SBP)≥150mmHg和/或舒张压(Diastolic blood pressure,DBP)≥95mmHg或正在服用降压药物的高血压患者,和1269名年龄、性别匹配的正常对照(SBP＜140mmHg且DBP＜90mmHg)作为研究对象。我们设计了两阶段关联研究,即第一阶段采用相对小的样本用于筛选,第二阶段对第一阶段筛选出来的阳性关联的位点继续进行研究。为了增加检测出病例和对照人群高血压易感等位基因频率差异的能力,我们选择了503例血压水平为SBP≥160mmHg和/或DBP≥100mmHg的高血压患者和490例年龄、性别匹配的正常对照(SBP＜140mmHg且DBP＜90mmHg)作为第一阶段研究对象。第二阶段样本包括了814名高血压病例和779名年龄和性别匹配的正常对照。为提高研究效能,最后采用两阶段研究人群的联合样本进行验证分析。采用基于连锁不平衡分析的标签SNP(tagging SNP,tagSNP)选择和生物信息学的功能预测分析相结合的策略,对覆盖6个功能候选基因及其上下游区域的23个有代表性的SNPs进行基因分型。首先分析单个SNP位点的等位基因和基因型与高血压的关联,然后分析多位点构建的单体型(Haplotype)和双体型(Diplotype)与高血压之间的关系。2、本研究应用目前较为常用的多元适应性回归样条(Multivariate adaptiveregression splines,MARS)方法和在多因子降维法(Multifactor-dimensionalityreduction)基础上发展的广义多因子降维法(Generalized multifactordimensionality reduction,GMDR),分析潜在的基因-基因及基因-环境交互作用,然后进一步应用Logistic回归模型对分析的结果进行重复分析。研究结果1、第一阶段的关联研究初步筛选出8个与高血压显著相关的SNPs(P＜0.05),分别为EMILIN1基因rs3754734和rs2011616,FURIN基因rs4932178,TGFB1基因rs12980942,TGFBR1基因rs12346650、rs1888223以及FBN1基因rs140598和rs6493333。进一步对这8个SNPs在两阶段研究人群样本中进行联合分析,发现有3个多态位点与高血压的关联得到了重复验证。携带FURIN基因rs4932178的TT(vs.CC+CT)突变基因型个体的高血压患病风险显著增加,OR(95%可信区间)为2.543(1.324-4.886);携带TGFBR1基因rs12346650的A(vs.G)等位基因个体的高血压患病风险显著增加,具有加性作用,OR为1.473(1.315-1.650),而FBN1基因rs140598 G(vs.C)等位基因携带者高血压患病风险相对较低,也具有加性作用,OR值为0.827(0.703-0.973)。单体型与双体型分析结果表明,与TGFBR1基因的G-W单体型作比较,A-G和A-W单体型携带者高血压患病风险显著增加,OR(95%CI)分别为1.269(1.095-1.470)和1.550(1.299-1.856);G-G单体型携带者高血压患病风险明显降低,OR为0.112(0.094-0.134)。与双体型Hap1(G-W)-Hap1(G-w)相比较,其它常见5种双体型Dip2(Hap1-Hap2)、Dip3(Hap1-Hap4)、Dip4(Hap2-Hap2)、Dip5(Hap3-Hap3)和Dip6(Hap2-Hap3)均具有显著的危险作用,经过Bonferroni校正后,P值均有显著性意义(P＜0.003)。与FBN1基因C-C单体型相比较,单体型G-C对于EH具有保护作用,OR=0.823(Bonferroni校正后,P＜0.0125);与Hap(C-C)-Hap(C-C)双体型相比,Hap(G-C)-Hap(G-C)双体型具有保护作用,OR=0.619(0.436-0.880)。2、应用MARS对不同模型的交互作用进行分析发现,Model 1中,TGFBR1基因rs12346650与FBN1基因rs140598及吸烟之间的交互作用有显著意义,并且rs12346650具有显著的主效应;Model 2中,TGFBR1基因rs12346650除了具有主效应外,还与其它位点和环境因素之间存在显著的交互作用,而FURIN基因rs4932178具有独立的主效应;Model 3中,3个微效多态位点TGFBR1基因rs1888223、TGFB1基因rs12980942、FBN1基因rs649333和饮酒之间存在多阶交互作用。GMDR对Model 1和Model 2的分析结果与MARS基本一致,但在Model 3中没有发现有显著意义的交互作用。随后,对于3个模型中MARS分析显示有显著性的3个3阶交互作用rs12346650-rs140598-吸烟、rs12346650-rs140598-rs3754734和rs1888223-rs12980942-饮酒,采用Logistic回归进行了重复分析,结果显示两者具有较好的一致性。结论1、通过以基于人群的两阶段关联研究对TGF-β信号通路关键基因的23个有代表性SNPs进行分析,发现8个SNPs与原发性高血压显著相关;其中FURIN基因的rs4932178、TGFBR1基因的rs12346650和FBN1基因的rs140598在两阶段大样本人群中得到重复验证。结果提示TGF-β信号通路3个基因多态位点rs4932178、rs12346650和rs140598与中国北方汉族人群原发性高血压之间存在关联。2、通过不同分析方法探讨TGF-β信号通路多个基因与环境因素对原发性高血压的影响,结果一致显示:TGFBR1基因多态位点rs12346650对高血压危险性具有显著的主效应,并与TGF-β信号通路中其它关键基因FURIN、FBN1、EMLIN1和TGFB1以及环境因素之间存在着对高血压的交互作用。更多还原

【Abstract】 Background and objectivesCardiovascular disease is the leading cause of death for adults in developed or developing countries, including China. Hypertension has been proved the primary and independent modifiable risk factor for stroke and coronary heart disease (CHD) so that controlling hypertension is one of the most effective ways to prevent cardiovascular disease.With changes in lifestyle (such as smoking, overdrinking and unhealthy diet, etc.) and the increase in life expectancy, the prevalence of hypertension has been increasing significantly in China in the past decades. Particularly, the prevalence of hypertension has also increased rapidly in younger people and the rural areas. These results underscore the urgent need to carry out active researches on the etiology and develop effective strategies to improve prevention, diagnosis, and treatment of hypertension.Hypertension among adults with no identifiable cause is essential hypertension (EH). EH arises as a complex quantitative trait that is affected by varying combinations of genetic and environmental factors. Probably, genetic factor influences the blood pressure through numerous minor-effect genes, and biological process of hypertension involves multiple physiological pathways with gene and environment factors. Recently, two-stage association study based community population is a cost effective approach for identifying complex disease genes in genetic studies and it has received much attention.As the major basis in pathologic process of EH, large artery stiffening, peripheral resistance increasing caused by endothelial dysfunction and small artery remodeling have been proved the primary agent for blood pressure increase or even organ impairment. Researchers have focused on the molecular genetics of the pathology of artery stiffening and remodeling. However, the mechanism of human hypertension is still not fully understood. In animal study, proof indicated that Emilinl modulates TGF-βavailability in the pathogenesis of hypertension and links TGF-βmaturation to blood pressure homeostasis. This led us to further research the development and remodeling of artery in EH.With the purpose of exploring the relationship between development and remodeling of artery and EH, we collected and analyzed some biological data about the development and remodeling of blood vessel and selected several candidate genes, including FBN1, EMLIN1, FURIN, THSD1, TGFB1 and TGFBR1 gene, which proved being involved in the TGF-βsignal pathway. We applied a two-stage case-control study. We selected tagSNPs with consideration of predicted SNP (Single nucleotide polymorphisms) functions and obtained twenty three representative SNPs covered the candidate genes. Furthermore, we examined the relationship between the twenty three polymorphisms and EH in Chinese.Methods and SubjectsSubjectsAll the studied subjects recruited from the International Collaborative Study of Cardiovascular Disease in Asia (InterASIA in China), from which all the DNA samples and clinical data for participants were obtained. The local bioethical committee approved the protocol, and informed consent was obtained from each participant. InterASIA used a four-stage stratified sampling method to select a representative sample of the general population aged 35 to 74 years in China. We enrolled 1,317 unrelated hypertensive patients and 1,269 age and gender-matched unrelated normotensive from four northern field centers of InterASIA, namely Beijing, Jilin, Shandong, and Shanxi province. In this study, hypertensive were selected with systolic BP (SBP)≥150 mmHg, and/or diastolic BP (DBP)≥95 mmHg, or current use of antihypertensive medication, whereas subjects with a clinical history of secondary hypertension, coronary heart disease and diabetes were excluded. At the same time, healthy normotensives with SBP<140 mmHg and DBP<90 mmHg were selected from the same target study population as controls.MethodsPart 1We conducted a two-stage association study and took total 2,586 unrelated subjects as the main study population in this study. 993 subjects were selected as stage 1 subsample containing 503 hypertensive patients (SBP≥160 mmHg and/or DBP≥100 mmHg) and 490 age- and gender-matched normotensive controls (SBP<140 mmHg and DBP<90 mmHg). The criterion for case selection of the subsample based on the hypothesis that individuals with higher BP were likely to be enriched for genetic susceptibility, which might increase the difference in frequency of susceptibility alleles between cases and controls to improve power. At stage 2, subsample included 814 cases with hypertension and 779 age- and gender-matched normotensives.Combinating SNP functional prediction analysis, We selected twenty three representative tagSNP from HapMap database of Han Chinese in Beijing, China (CHB) according to the linkage disequilibrium (LD) analyses of all SNPs, which covered the candidate gene, reasonable length upstream and downstream. All the SNPs were examined the association with EH.Part 2We applied three methods to explore gene-gene and gene-environment interactions for EH. First, we utilized multivariate adaptive regression splines (MARS) and generalized multifactor-dimensionality reduction (GMDR) to identify latent interactions involved in the TGF-βpathway gene polymorphisms associated with EH. Furthermore, we employed Logistic regression to test the interactions identified by MARS or GMDR in three different models.ResultsPart 1 By single locus analysis, rs2011616 and rs3754734 of EMILIN1, rs4932178 of FURIN, rs12980942 of TGFB1, rs12346650 and rs1888223 of TGFBR1, and rs140598 and rs6493333 of FBN1 showed significant associations with EH (P<0.05) in stage 1 and we further genotyped the eight SNPs among remained sample. Joint analysis indicated that three SNPs presented significant association with EH. TT (vs. CC+CT) genotype of rs4932178 and allele A (vs.G) of rs12346650 increased the risk of EH significantly, odds ratios (ORs) were 2.543 (95% CI 1.324-4.886) and 1.473 (95% CI 1.315-1.650) respectively. G (vs.C) of rs140598 showed a protective effect on EH (OR=0.827 95% CI 0.703-0.973) after adjustment for conventional risk factors.Haplotypes of TGFBR1 gene were composed of rs12346650 and rs1888223. With the haplotype G-W used as reference, the adjusted ORs of haplotype A-G and haplotype A-W were 1.269 (95% CI: 1.095 to 1.470), 1.550 (1.299 to 1.856) respectively, and the OR of haplotype G-G was 0.112 (0.094 to 0.134). With the diplotype Hap1 (G-W)-Hap1 (G-W) used as reference, diplotype Dip2 (Hap1-Hap2), Dip3 (Hap1-Hap4), Dip4 (Hap2-Hap2), Dip5 (Hap3-Hap3) and Dip6 (Hap2-Hap3) were at higher risk for EH (after adjusted by Bonferroni correction, P<0.003). The haplotypes of FBN1 were composed of rs140598 and rs6493333. Haplotype C-C of FBN1 gene was used as reference and the adjusted OR of haplotype G-C was 0.823 (95% CI, 0.715 to 0.948). With diplotype Hap (C-C)-Hap (C-C) used as reference, the adjusted OR of diplotype Hap (G-C)-Hap (G-C) was 0.619 (95% CI, 0.436 to 0.880). Part 2First, MARS identified a significant interaction among smoking, rs140598 of FBN1 gene and rs12346650 of TGFBR1 gene, and rs12346650 presented a main effect for EH in model 1. In model 2, rs12346650 main effect showed in six out of seven significant interactions and rs140598 only appeared in five interactions whereas rs4932178 only presented an independent main effect. In model 3, three interactions were detected among drinking and four SNPs with latent minor effect.Subsequently, GMDR verified rs12346650 main effect in four interactions among rs4932178, rs140598, smoking and drinking in model 1. In model 2, Analogously, GMDR verified rs12346650 main effect in four interactions of rs4932178, rs140598, rs12980942, smoking and driking. However, there were no interactions identified by GMDR in model 3.Furthermore, Logistic regression tested three major 3-way interactions of rs12346650-rs140598-smoking, rs12346650-rs140598-rs3754734 and rs1888223 -rs12980942-drinking, which had been identified by MARS in the three models respectively, and achieved an explainable concordance with MARS.ConclusionsOur findings indicated TGF-βpathway genes might harbor some susceptible variations to essential hypertension in north Han Chinese population. Three methods, GMDR, MARS and Logistic regression analyses identified consistently significant gene by gene and gene by environmental interactions that associated with hypertension and the findings provide further support for the possible contributions of TGF-βpathway genes to essential hypertension.更多还原