1、中文摘要 中国成年双生子 肥胖与 DNA 甲基化的相关 性 研究 一、 研究背景： 肥胖是全球和中国面临的日趋严重 的 健康问题，它不但是一种严重危害人们健康的疾病，而且是高血压、 2 型糖尿病、心脑血管疾病等慢性病的重要危险因素，从而导致巨大的疾病负担。 WHO 估计目前全球及中国成人肥胖率分别为 13.0%和 7.3%，每年至少280 万成年人死于由超重或肥胖 导致 的疾病。深入了解探讨肥胖的病因及发病机制对开展有针对性的预防与控制有着十分重要的意义。大量研究发现，肥胖是由环境因素、遗传因素以及二者交互作用决定的复杂性疾病。与肥胖相关的 环境因素包括膳食、体力活动、吸烟、饮酒等生活方式以及
2、社会经济状况等。截至 2015 年，全基因组关联研究（ GWAS）及其 Meta 分析研究发现，与肥胖相关单核苷酸多态性位点（ SNPs）有 97 个。所有这些基因多态性累计对肥胖变异的解释程度约为 2.7%。研究表明，遗传对肥胖的作用不局限于基因结构变异，而是表现在基因 -环境的交互作用或基因功能变异。已有队列研究及临床试验发现，膳食和 或体力活动与肥胖相关基因存在交互作用。然而这些基因 -环境交互作用背后的原因或机制尚不明确。表观遗传现象可能是解释基因 -环境交互作用背后原因 的重要因素，也有可能是独立 于 基因 -环境交互作用对肥胖产生影响的基因功能变异。表观遗传学是研究基因核苷酸序列不
3、发生改变的情况下，基因表达发生可遗传的变化。表观遗传变异包括 DNA 甲基化、基因组印记、染色质重塑等方面，其中 DNA 甲基化是目前研究最多的一种表观遗传现象。近年来，有关肥胖的表观遗传学研究受到广泛关注，探讨相关基因 DNA 甲基化水平与 肥胖的 相关性及其规律与特点已成为本领域热中文摘要 点。本研究采用全基因组 DNA 甲基化关联研究策略，以中国双生子人群作为发现人群，以 4 个中国一般人群为验证人群，以期发现在中国人群中与肥胖 相关的 DNA 甲基化位点及其作用机制。 二、 研究目的： 1. 探 索 特异 DNA 甲基化位点与肥胖及其相关指标的相关性，环境因素对肥胖与DNA 甲基化位点
4、关系的影响； 2. 探索 多个 DNA 甲基化位点所在基因 构成的生物 学途径 与肥胖的关系； 3. 探索 肥胖相关 DNA 甲基化位点与基因结构及基因表达的关系。 三、 研究内容与方法： 本研究以中国双生子登记系统的双生子为发现人群，以另外 4个中国一般人群队列为验证人群，在全基因组范围分析与肥胖相关的特异 DNA 甲基化位点、涉及生物 学途径 及其与基因结构和基因表达之间的作用规律。 1. 研究对象 采用双生子人群作为 全基 因组 DNA 甲基化位点发现人群，一般人群作为 全基因组DNA 甲基化位点验证人群 ： （ 1） 双生子人群 中国双生子登记系统是目前中国最大的双生子登记平台，有超过
5、 3万对有效登记的双生子。 本研究纳入符合以下所有标准的双生子： 2013 年居住在山东、江苏、浙江和四川地区，年龄 18 岁 ，在系统中自报 BMI 或腰围满足一定条件 （双生子中一人BMI27kg/m2，另一人 BMI 27kg/m2 and another person BMI 25kg/m2, with a difference at least 3kg/m2; or male same-sex twins with one waist 90cm, another waist 90cm; female same-sex twins with one waist 80cm, anothe
6、r person waist 80cm), not suffering from cancer, coronary heart disease (including angina, myocardial infarction, acute coronary syndrome, etc.) and/or a history of stroke. (2) General populations 英文摘要 There were four general population: Dongfeng tongji cohort (DFTJ), the coke oven workers cohort (C
7、OW), Wuhan-Zhuhai community based cohort (WHZH) and Shiyan health examination data (SY). Include criterias were: aged more than 18 years, not suffering from acute or chronic illness at the time of investigation and the absence of any physical discomfort or infection status in recent two weeks. 2. Da
8、ta Collection We collected twins information about demographic characteristics, lifestyle and disease conditions and other information by questionnaires, including gender, age, smoking, drinking, diet, physical activity, recent drug use and socio-economic status. BMI, waist circumference, waist-hip
9、ratio were collected by physical examination. Lipid and glucose levels were determined by 1ml peripheral blood serum (total cholesterol TC, triglyceride TG, low density lipoprotein cholesterol LDL-C, high-density lipoprotein cholesterol HDL-C, fasting glucose, fasting serum insulin, HOMA- IR). We co
10、llected 2ml whole blood for testing glycated hemoglobin, whole-genome genotyping and DNA methylation detection. Analogous methods were used by questionnaires, physical examination and DNA detection in replication populations. 3. Sample Detection (1) Biochemical parameters In twins population, we use
11、d enzyme colorimetry for TC and TG, direct measurement for LDL-C and HDL-C, improved fasting blood glucose hexokinase enzymatic assay for glucose, ADVIA Centaur Immunoassay System chemiluminescence immunoassay for fasting serum insulin, and TosoH G7 glycated hemoglobin meter with high pressure liqui
12、d chromatography (HPLC) for hemoglobin. HOMA-IR was calculated by function with fasting blood glucose and insulin. These indicators mainly conducted by Swedish company Roche reagents and standard methods. (2) DNA samples tested We used BioTeke whole blood DNA extraction kit to extract blood DNA. 英文摘
13、要 DNA methylation was detected by containing sodium bisulfite kit (Zymo EZ DNA Methylation kit) to convert unmethylated cytosine to thymidine. After conversion, we used Infinium HumanMethylation 450K chip to detect peripheral blood leukocyte genome-wide DNA methylation then scanned and imaged the ch
14、ip information through iScan instrument. Whole blood DNA concentration and purity were tested prior to genome-wide DNA detecting. The Human Omni ZhongHua-8 BeadChip was for genome-wide genotyping and twin zygosity determination. It was used TRIZOL LS solution (Invitrogen) to isolate total RNA from w
15、hole blood white blood cells, then detected by HumanHT-12 v4 Expression BeadChip for gene expression testing. 4. Statistical analysis We used R (3.1.2) language minfi package to read DNA methylation signal and convert it to value ( value is defined as the proportions of methylated signal in all unme
16、thylated and methylated signals), carried out sample and probes quality control then normalized the data by DASEN function. Normalized data was analyzed by surrogate variable analysis by sva function for adjusting potential confounding factors. For continuous variables (BMI, waist circumference, wai
17、st-hip ratio), we used nlme function of mixed effects model, with Manhattan and q-q plots to describe the location of methylation sites and population stratification. Categorical variables (being obese or not) were used ebayes function of empirical Bayes paired moderated t-test, describing the diffe
18、rences in methylation levels in obese and control groups with volcano plots. In addition, the enrichment analysis were conducted by GOrilla software and R codes by Fishers exact test and permutation test to find out obesity-related biological pathways. The asscociation of methylation sites and neigh
19、boring SNPs was analyzed using Locus Zoom plot. False discovery rate FDR (Benjamini & Hochberg method) was used to correct the level of significance. Replication populations were used R (3.1.2) lm function to conduct linear regression of DNA methylation and BMI (or waist circumference, waist-hip ratio) respectively, then metaed by Metafor package (fixed effects model) to summarized the four replication cohorts results
