DNA methylation markers associated with type 2 diabetes, fasting glucose and HbA1c levels: a systematic review and replication in a case–control sample of the Lifelines study

Aims/hypothesis Epigenetic mechanisms may play an important role in the aetiology of type 2 diabetes. Recent epigenome-wide association studies (EWASs) identified several DNA methylation markers associated with type 2 diabetes, fasting glucose and HbA1c levels. Here we present a systematic review of these studies and attempt to replicate the CpG sites (CpGs) with the most significant associations from these EWASs in a case–control sample of the Lifelines study. Methods We performed a systematic literature search in PubMed and EMBASE for EWASs to test the association between DNA methylation and type 2 diabetes and/or glycaemic traits and reviewed the search results. For replication purposes we selected 100 unique CpGs identified in peripheral blood, pancreas, adipose tissue and liver from 15 EWASs, using study-specific Bonferroni-corrected significance thresholds. Methylation data (Illumina 450K array) in whole blood from 100 type 2 diabetic individuals and 100 control individuals from the Lifelines study were available. Multivariate linear models were used to examine the associations of the specific CpGs with type 2 diabetes and glycaemic traits. Results From the 52 CpGs identified in blood and selected for replication, 15 CpGs showed nominally significant associations with type 2 diabetes in the Lifelines sample (p < 0.05). The results for five CpGs (in ABCG1, LOXL2, TXNIP, SLC1A5 and SREBF1) remained significant after a stringent multiple-testing correction (changes in methylation from −3% up to 3.6%, p < 0.0009). All associations were directionally consistent with the original EWAS results. None of the selected CpGs from the tissue-specific EWASs were replicated in our methylation data from whole blood. We were also unable to replicate any of the CpGs associated with HbA1c levels in the healthy control individuals of our sample, while two CpGs (in ABCG1 and CCDC57) for fasting glucose were replicated at a nominal significance level (p < 0.05). Conclusions/interpretation A number of differentially methylated CpGs reported to be associated with type 2 diabetes in the EWAS literature were replicated in blood and show promise for clinical use as disease biomarkers. However, more prospective studies are needed to support the robustness of these findings. Electronic supplementary material The online version of this article (10.1007/s00125-017-4497-7) contains peer-reviewed but unedited supplementary material, which is available to authorised users.


Search terms
Pubmed database was searched using the following search terms: (  Both searches were performed until April 26, 2017. Of each publication, both title and abstract were examined to evaluate its usefulness for review by two independent reviewers (EW; JvVO). We identified 580 relevant publications in PubMed and Embase. EWASs investigating DNA methylation patterns associated with T2D or glycaemic traits were included. We excluded candidate-gene approach studies, studies that focused on epigenetic mechanisms other than DNA methylation, studies determining only global methylation, studies not related to T2D or glucose metabolism, intervention studies, animal studies, studies in neonates, case reports, promoter methylation, EWAS replication studies, co-morbidities (e.g. dementia in elderly with T2D, Parkinson, Alzheimer), T2D complications (e.g. nephropathy, patient-derived diabetic foot ulcer fibroblasts) and articles published in other languages than English. Disagreements between the two reviewers were solved through discussion and with help of an arbitrator (HS). Full texts of included studies were retrieved.

Quality control and normalization
Methylumi R package was used to extract the data from the raw IDAT files. We performed quality control checks on the probes and samples. Two samples were excluded from the control group for sex-mismatch. After having performed a principal component analysis (PCA), we checked for outliers using the first two principal components (PCs). Next, we performed background correction and probe type normalization based on the Touleimat and Tost pipeline [1], using the dasen normalization strategy [2]. We remapped the 450K probes to the human genome reference (HG19). Details on this procedure can be found in Bonder et al [3]. Next, we removed probes with a known Single Nucleotide Polymorphism (SNP) from "Genome of the Netherlands" (GoNL), minor allele frequency (MAF) > 0.01) at the single base extension (SBE) site or CpG site. Lastly, we removed all probes on the sex chromosomes, resulting in 423,289 high quality methylation sites for subsequent analyses.

Assessing the quality of included publications
We evaluated bias and quality of included publications using the Newcastle-Ottawa Scale, a scale designed to evaluate the quality of case-control and cohort studies [4]. Two independent reviewers (EW; JvVO) assessed the study quality based on the three categories: selection of individuals, comparability of cases and controls and outcome assessment. A maximum of nine stars could be granted. Studies with a score of eight or nine stars were judged to be at low risk of bias, studies with a score of six to seven stars were considered to be at medium risk of bias, whereas studies with a score of five or less stars were considered to be at high risk of bias.
ESM Box 1 Functions of 5 top CpG sites from our replication study in Lifelines T2D EWAS sub study. ABCG1 is involved in lipid homeostasis, especially promoting cholesterol efflux into HDL [5].
Cholesterol homeostasis is known to be essential modulator of insulin secretion, which is reflected in the phenotype of the ABCG1-/-mice, characterized by impaired glucose tolerance and insulin secretion [6].
LOXL2 gene, a member of lysyl oxidase family and an important regulator of tumour progression [7]. It has been shown that LOXL2 expression is glucose-sensitive in eye and can be related to diabetic retinopathy [8]. In mice, LOXL2 inhibitor has a profitable effect on glomerular structure in kidney, implicating that LOXL2 can be a potential therapeutic target for diabetic nephropathy [9].
TXNIP (Thioredoxin-interacting protein) expression is induced by glucose in β-cells as well as in other tissues and also inhibits glucose uptake via enhanced endocytosis of glucosetransporter 1 (GLUT1) [10]. SLC1A5 gene-its protein belongs to different solute carriers (SLC) family of amino-acids transporters. Overexpression of SLC1A5 increases glutamine uptake especially in cancer cells, therefore represents a potential target in cancer as well as in pathologies such as obesity, type 2 diabetes and neurodegeneration [12] ESM Table 1 List of CpG sites associated with T2D and glycaemic traits used for replication analysis (100 unique CpGs). ESM Table 2 Baseline characteristics of the study sample of T2D patients and healthy controls from Lifelines, divided into four groups.   Table 5 Overlap between CpGs that are significantly associated with T2D and fasting glucose (from Table S1) and CpGs significantly associated with lipids and BMI.
Fasting glucose

✓
The comparison was made based on published studies [13][14][15]. Green check marks indicate a directionally consistent overlap.

ESM Figure 1
Correlation between effect sizes based on baseline model 1 and model 2 (Model 1 +BMI).