Type 2 diabetes (T2D) is caused by interactions between genetic predisposition and environmental exposure. DNA methylation, a key epigenetic mark, is one measure of the molecular impact of environmental factors on DNA. We tested for associations between glycemic traits and peripheral blood cell DNA methylation in 473,669 CpG sites across the genome in individuals from multiple ancestries, and tested for causal relationships at these loci.

We conducted an epigenome-wide fixed effects meta-analysis for 3 glycemic traits-fasting glucose (FG), log fasting insulin (FI), and HbA1c—in 13,543 nondiabetic individuals of African (N=3,746), European (N=9,254), and Hispanic (N=543) ancestry across 16 cohorts. All analyses were adjusted for age, sex, smoking status, cell types, BMI, and technical covariates. We used Mendelian randomization (MR) analyses to determine whether glycemic traits causally influence methylation at the CpG sites identified in the meta-analysis. We performed MR analyses in Framingham Heart Study individuals (N=6,472), using separate genetic risk scores (GRS) as instruments for each of the 3 glycemic traits.

In the epigenome-wide meta-analysis of all ancestries, there were 43 CpG sites associated with FG, 89 with FI, and 54 with HbA1c, using a Bonferroni threshold of P<1.1×10-7. In addition to detecting many novel candidate loci for each trait, we found associations with all 3 glycemic traits at previously reported T2D methylated loci SREBF1, LINC00649, and ABCG1. In MR analyses, we showed that a higher GRS for FI was associated with greater methylation at cg17058475 in CPT1A (P=4×10-4) and higher GRS for HbA1c with greater methylation at cg05779219 near THBD (P=4×10-4).

We identified many novel candidate methylation loci associated with glycemic traits and 2 CpG sites where MR analysis indicated that methylation is causally influenced by glycemic traits. These findings support the hypothesis that variation in glycemia influences peripheral blood DNA methylation.


D.A. DiCorpo: None. S. Lent: None. W. Guan: None. M. Hivert: None. J.S. Pankow: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.