Purpose: An epigenome-wide association study (EWAS) was conducted to detect specific epigenetic variants potentially related to fasting plasma glucose (FPG) in middle-aged and elderly Chinese monozygotic twins.

Methods: Association between DNA methylation (DNAm) of single CpG and FPG was tested by applying generalized estimation equation in 52 twin pairs. Differentially methylated regions (DMRs) were identified by comb-P approach. Inference about Causation through Examination of Familial Confounding (ICE FALCON) was utilized to perform the causal inference. Candidate CpGs were quantified using Sequenom MassARRAY platform in a community population. Weighted gene co-expression network analysis (WGCNA) was conducted using gene expression data.

Results: The relationship between DNAm of 30 top CpGs and FPG reached P-value<1×10^-6 level. Thirty-two DMRs within 24 genes, such as TLCD1, MRPS31P5, CASZ1 and CXADRP3, were identified. Causal relationship of 20 top CpGs within TLCD1, MZF1, PTPRN2, SLC6A18, ASTN2, IQCA1, GRIN1, and PDE2A with FPG were further identified. Pathways related to FPG such as mitogen-activated protein kinase p38 binding were found. Three CpGs mapped to SLC6A18 were validated in a community population, with a hypermethylated direction in type 2 diabetes cases. The gene expression levels of 18 genes where the top CpGs and DMRs located, such as SLC6A18 and TLCD1, were positively correlated with FPG. Besides, common biological pathways, such as dopamine binding, regulation of biosynthetic process, and neuron fate specification, were found in methylation and gene expression analyses.

Conclusion: Our findings showed that multiple methylated CpGs and regions, crucial genes, and biological pathways were underlyingly associated with FPG and diabetes.