Longitudinal Association of DNA Methylation with Type 2 Diabetes and Glycemic Traits: A 5-Year Cross-Lagged Twin Study

Xuanming Hong,Zhiyu Wu,Weihua Cao,Jun Lv,Canqing Yu,Tao Huang,Dianjianyi Sun,Chunxiao Liao,Yuanjie Pang,Zengchang Pang,Liming Cong,Hua Wang,Xianping Wu,Yu Liu,Wenjing Gao,Liming Li
DOI: https://doi.org/10.2337/db22-0513
IF: 7.7
2022-01-01
Diabetes
Abstract:Previous cross-sectional Epigenome-Wide Association Studies (EWASs) in adults have reported hundreds of 5′-cytosine-phosphate-guanine-3′ (CpG) sites associated with type 2 diabetes mellitus (T2DM) and glycemic traits. However, the results from EWASs have been inconsistent, and longitudinal observations of these associations are scarce. Furthermore, few studies have investigated whether DNAm could be modified by smoking, drinking and glycemic traits, which have broad impacts on genome-wide DNAm, and results in altering the risk of T2DM. Twin studies provide a valuable tool for epigenetic studies, as they are naturally matched for genetic information. In this study, we conducted a systematic literature search in PubMed and EMBASE for EWASs, and 214, 33, and 117 candidate CpG sites were selected for T2DM, HbA1c and fasting blood glucose (FBG). Based on 1,070 twins from the Chinese National Twin Registry, 67, 17 and 16 CpG sites from previous studies were validated for T2DM, HbA1c and FBG. Longitudinal review and blood sampling for phenotypic information and DNAm were conducted twice in 2013 and 2018 on 308 twins. A cross-lagged analysis was performed to examine the temporal relationship between DNAm and T2DM or glycemic traits in the longitudinal data. 11 significant paths from T2DM to subsequent DNAm and 15 paths from DNAm to subsequent T2DM were detected, suggesting both directions of associations. For glycemic traits, we detected 17 cross-lagged associations from baseline glycemic traits to subsequent DNAm, and none was from the other cross-lagged direction, indicating CpG sites may be the consequences, not the causes, of glycemic traits. Finally, a longitudinal mediation analysis was performed to explore the mediation effects of DNAm on the associations of smoking, drinking and glycemic traits with T2DM. No significant mediations of DNAm in the associations linking smoking and drinking with T2DM were found. In contrast, our study suggested the potential role of DNAm of cg19693031, cg00574958 and cg04816311 in mediating the effect of altered glycemic traits on T2DM.
What problem does this paper attempt to address?