Hong Xuanming, Miao Ke, Cao Weihua, Lv Jun, Yu Canqing, Huang Tao, Sun Dianjianyi, Liao Chunxiao, Pang Yuanjie, Pang Zengchang, Yu Min, Wang Hua, Wu Xianping, Liu Yu, Gao Wenjing, Li Liming
Department of Epidemiology and Biostatistics, School of Public Health, Peking University, China (X.H., K.M., W.C., J.L., C.Y., T.H., D.S., C.L., Y.P., W.G., L.L.).
Qingdao Center for Disease Control and Prevention, China (Z.P.).
Hypertension. 2023 Jan;80(1):169-181. doi: 10.1161/HYPERTENSIONAHA.122.19953. Epub 2022 Nov 8.
Previous EWASs (Epigenome-Wide Association Studies) have reported hundreds of blood pressure (BP) associated 5'-cytosine-phosphate-guanine-3' (CpG) sites. However, their results were inconsistent. Longitudinal observations on the temporal relationship between DNA methylation and BP are lacking.
A candidate CpG site association study for BP was conducted on 1072 twins in the Chinese National Twin Registry. PubMed and EMBASE were searched for candidate CpG sites. Cross-lagged models were used to assess the temporal relationship between BP and DNA methylation in 308 twins who completed 2 surveys in 2013 and 2018. Then, the significant cross-lagged associations were validated by adopting the Inference About Causation From Examination of Familial Confounding approach. Finally, to evaluate the cumulative effects of DNA methylation on the progression of hypertension, we established methylation risk scores based on BP-associated CpG sites and performed Markov multistate models.
16 and 20 CpG sites were validated to be associated with systolic BP and diastolic BP, respectively. In the cross-lagged analysis, we detected that methylation of 2 CpG sites could predict subsequent systolic BP, and systolic BP predicted methylation at another 3 CpG sites. For diastolic BP, methylation at 3 CpG sites had significant cross-lagged effects for predicting diastolic BP levels, while the prediction from the opposite direction was observed at one site. Among these, 3 associations were validated in the Inference About Causation From Examination of Familial Confounding analysis. Using the Markov multistate model, we observed that methylation risk scores were associated with the development of hypertension.
Our findings suggest the significance of DNA methylation in the development of hypertension.
既往的全基因组关联研究（EWASs）已报道了数百个与血压（BP）相关的5'-胞嘧啶-磷酸-鸟嘌呤-3'（CpG）位点。然而，其结果并不一致。目前缺乏关于DNA甲基化与血压之间时间关系的纵向观察。
在中国国家双胞胎登记处对1072对双胞胎进行了一项关于BP的候选CpG位点关联研究。通过检索PubMed和EMBASE获取候选CpG位点。采用交叉滞后模型评估了在2013年和2018年完成两次调查的308对双胞胎中BP与DNA甲基化之间的时间关系。然后，采用基于家族混杂因素检验的因果推断方法对显著的交叉滞后关联进行验证。最后，为了评估DNA甲基化对高血压进展的累积影响，我们基于与BP相关的CpG位点建立了甲基化风险评分，并进行了马尔可夫多状态模型分析。
分别验证了16个和20个CpG位点与收缩压和舒张压相关。在交叉滞后分析中，我们检测到2个CpG位点的甲基化可预测随后的收缩压，而收缩压可预测另外3个CpG位点的甲基化。对于舒张压，3个CpG位点的甲基化对预测舒张压水平具有显著的交叉滞后效应，而在一个位点观察到了相反方向的预测。其中，3个关联在基于家族混杂因素检验的因果推断分析中得到验证。使用马尔可夫多状态模型，我们观察到甲基化风险评分与高血压的发生发展相关。
我们的研究结果提示了DNA甲基化在高血压发生发展中的重要性。
