目的利用孟德尔随机化方法,探索DNA甲基化位点与肥胖的相关性。方法2013年以中国双生子登记系统中山东、江苏、浙江和四川招募的双生子人群中≥18岁的肥胖状况不一致的双生子作为本次调查对象,共469名。通过问卷调查、体格检查及血样采集,收集双生子的一般信息、行为信息及体格信息,并提取外周血DNA进行相关基因型测定及全基因组甲基化检测。并分析了全基因组DNA甲基化与BMI的关联以及CpGs与邻近SNP的关联。利用孟德尔随机化方法,以SNP位点r8748212作为工具,分析肥胖指标BMI与cg15053022甲基化的关系。结果研究共纳人469名调查对象,年龄为(44.8±13.2)岁,BMI为(25.0±3.8)kg/m^2。全基因组DNA甲基化与BMI的关联分析发现,有9个CpG位点甲基化水平与BMI存在关联。在进行CpGs甲基化与邻近SNP的关联分析发现,ATP4A基因的cg15053022与rs748212呈负相关(β=-0.020),rs748212AA、AC、CC基因型的cg15053022甲基化水平分别为0.212±0.025、0.242±0.024、0.264±0.028。孟德尔随机化分析发现,rs748212与BMI存在关联(β=0.04,P=0.007),且与ATP4A基因的cg15053022关联强度较高(F=237.66,P=0.143)。孟德尔随机化结果显示,cg15053022甲基化水平每降低10%,BMI增加0.197kg/m^2(8—1.97,P〈0.001)。结论本研究利用孟德尔随机化的方法发现观察性研究中ATP4A基因上的甲基化位点cg15053022甲基化能够影响BMI,为肥胖甲基化孟德尔随机化相关研究的方法学提供了参考依据。
Objective To explore the association between DNA methylation and body mass index (BMI) using Mendelian randomization analysis. Methods A total of 469 participants were selected from the Chinese National Twin Registry in 2013, who were living in Shandong, Jiangsu, Zhejiang, and Sichuan provinces, and at least 18 years of age. A questionnaire survey and physical examination were conducted to collect demographic, clinical, and behavioral information. Peripheral blood cells were collected to detect genotype and methylation status. Association analyses between DNA methylation and BMI and between CpGs and cis-SNP were conducted. With rs748212 as the instrumental variable, the association between cg15053022 and BMI was explored using the Mendelian randomization method. Results A total of 469 participants were selected. The mean age of participants was (44.8±13.2) years and the BMI was (25.0±3.8) kg/m^2. Nine BMI-related DNA methylation sites were found and DNA methylation site cg15053022 in the ATP4A gene was negatively associated with cis-SNP rs748212 (β=-0.020); the mean methylation level of AA, AC, and CC were 0.212 ± 0.025, 0.242 ± 0.024, and 0.264 ± 0.028, respectively, rs748212 was associated with BMI (β=0.04, P=0.007) and closely related to cg15053022 (F=237.66, P=0.143). Mendelian randomization analysis showed lower methylation levels at cg15053022 were associated with higher BMI (β =- 1.97, P〈0.001). Conclusion This study supported the impact of cg15053022 methylation in the ATP4A gene on BMI using Mendelian randomization analysis and provided the basis for using Mendelian randomization analysis in methylation studies.