Data from: DNA methylation changes associated with type 2 diabetes and diabetic kidney disease in an East Asian population

<p><b>Objective:</b> There is growing body of evidence that epigenetic changes including DNA methylation influence the risk of type 2 diabetes and its microvascular complications. We conducted a methylome-wide association study (MWAS) to identify differentially methylated regions (DMRs) of type 2 diabetes and diabetic kidney disease (DKD) in Korean population.</p> <p><b>Methods:</b> We performed an initial MWAS in 232 participants in a case-control study design with type 2 diabetes and 197 non-diabetic controls with Illumina EPIC bead chip using peripheral blood leukocytes. Type 2 diabetes group was subdivided to 87 DKD cases and 80 non-DKD controls. Additional 819 individuals from two population-based cohorts were used to investigate the association of the identified DMRs with quantitative metabolic traits. We developed a DNA methylation score using identified DMRs to predict the occurrence of type 2 diabetes. To examine the causal relationship between the metabolic traits and differentially methylated status, we performed Mendelian randomization (MR) analyses.</p> <p><b>Results:</b> We identified eight DMRs (each at <i>BMP8A</i>, <i>NBPF20, STX18, ZNF365, CPT1A,</i> and <i>TRIM37</i>, and two at <i>TXNIP</i>) which were significantly associated with risk of type 2 diabetes (P &lt; 9.0×10^-8), including three that were previously known (DMRs in <i>TXNIP</i> and<i> CPT1A</i>), in 429 type 2 diabetes cases and controls. DNA methylation score consisted of these DMRs differentiated the risk of developing type 2 diabetes in an independent prospective cohort with a relative risk of 2.44 (95% confidence interval 1.39–4.28) between the lowest and highest deciles of DNA methylation score. DMRs in <i>CPT1A</i> and <i>TXNIP</i> were associated with quantitative metabolic traits, including fasting glucose, HbA1c, and body mass index. We also identified three DMRs (on <i>COMMD1</i>, <i>TMOD1</i>, and <i>FHOD1</i>) associated with DKD in 167 DKD cases and controls. The DMRs of DKD did not show meaningful overlap with those of type 2 diabetes. In MR analysis, the estimated glomerular filtration rate was causally associated with DNA methylation of these three DMRs.</p> <p><b>Conclusions</b>: In an East Asian population, we identified eight DMRs, including five novel ones, associated with type 2 diabetes and three DMRs associated with DKD at methylome-wide statistical significance. Our findings suggest that epigenetics of DKD may share little with those responsible for the development of type 2 diabetes.</p>