From: Recent developments on the role of epigenetics in obesity and metabolic disease
 | Phenotype | Association with epigenetic marks or changes in epigenetic marksa | Includes validationb | Ref. |
---|---|---|---|---|
Cross-sectional studies | ||||
 BMI, WC | Methylation in 37 CpGs associated with BMI and 1 probe with WC in blood (n = 5465); of those 16 CpGs (e.g., CpGs in CPT1A, ABCG1, and SREBPF1) were also associated with BMI in subcutaneous adipose tissue (n = 648) | Yes, 3 other cohorts | [51] | |
 BMI | DNA methylation of 4979 CpGs (e.g., in FTO, TCF7L2, FASN, PGC1A, CCRL2) in subcutaneous adipose tissue (n = 190). Several BMI-related methylation sites were also associated with age (e.g., ELOVL2), and with gene expression levels | Yes, 2nd cohort | [9] | |
 BMI | HIF3A methylation 3 CpGs in whole blood (n = 2587) and subcutaneous adipose tissue (n = 635), not in skin (n = 395) | Yes, 3 other cohorts | [52] | |
 Obesity | LY86 methylation 1 CpG in blood leukocytes adolescents and adults (n = 1534) | Yes, 4 other cohorts | [53] | |
 Obesity | 249 DMRs in subcutaneous adipose tissue (n = 21), methylation in all these 249 regions also changed with high-fat feeding in mice and includes regions overlapping T2DM loci such as in TCF7L2 | No, only in mice | [43] | |
 WHR, blood pressure LDL cholesterol | Association between WHR and ADRB3 methylation in whole blood and between blood pressure and ADRB3 methylation in visceral adipose tissue of obese (n = 25). Association between ADRB3 methylation and LDL cholesterol in whole blood of men with familial hypercholesterolemia (n = 41) | Not validated in 2nd cohort | [55] | |
 BMI | ADCY3 meQTL in subcutaneous adipose tissue from twins (n = 648) | Yes, technical | [8] | |
 BMI | Methylation differences at 1236 CpGs in leukocytes of monozygotic twins discordant for BMI and liver fat (n = 13) | Yes, technical | [96] | |
 Adiposity phenotypes | Methylation in 101 genes in subcutaneous adipose tissue (n = 106) including AOC3, SOD3, DOCK9, AQP7, ANGPT4, ANGPT2, TIMP4, ADAMST4, HOXA3, and LIPE, not found in blood leukocytes same individuals | No validation performed | [56] | |
 TG, VLDL | CPT1A methylation 1 CpG in CD4+T cells (n = 991) and leukocytes (n = 1261) | Yes, technical and 2nd cohort | [57] | |
 VLDL and LDL | CPT1A methylation 2 CpGs in CD4+T cells (n = 994) | Yes, in the same cohort | [58] | |
 Cholesterol and TG | Methylation in 9 CpGs, including in ABCG1, CPT1A and SREBF1, in whole blood (n = 2747), 5 of these CpGs also showed associations in subcutaneous adipose tissue (n = 634) | Yes, 2 other cohorts | [59] | |
 Insulin and HOMA-IR | ABCG1 1 CpG in CD4+T cells (n = 837) | Yes, in the same cohort | [61] | |
 HbA1C | DNA methylation of 711 CpGs in subcutaneous adipose tissue (n = 96 males). Not validated in female cohort (n = 94) but 30 CpGs validated in T2DM case-control cohort | Yes, validated in 1 of 2 cohorts | [9] | |
 T2DM | MALT1 methylation whole blood (n = 27 twins + n = 263 individuals) and other less significant DMRs (FDR < 0.1) overlapping T2DM GWAS loci | Yes, other cohort | [48] | |
 T2DM | Methylation in 1649 CpG sites, some overlapping T2DM, and obesity GWAS loci such as TCF7L2, FTO, and KCNQ1 in pancreatic islets (n = 49) | Yes, technical | [62] | |
 T2DM | No differentially methylated sites (after FDR correction) in T2DM discordant monozygotic twins (n = 14 pairs). Differential methylation at 15,627 CpGs, including in T2DM GWAS loci PPARG, KCNQ1, TCF7L2, and IRS1 in subcutaneous adipose tissue (n = 56). 1410 of these CpGs were also differentially methylated (P < 0.05) in the T2DM discordant twins | Yes, other cohort | [63] | |
Longitudinal studies | ||||
 Adiposity measured annually age 9–14 years | Increase PGC1A promoter methylation in whole blood children measured annually from 5–7 years (n = 40) | No, but measures at multiple time points | [54] | |
Maternal exposure or phenotype and epigenetic marks in offspring | ||||
 Prenatal famine | 181 DMRs in adult whole blood (n = 48), including in CDH23, SMAD7, INSR, CPT1A, KLF13, RFTN1 (validated in n = 120) | Yes, technical and 2nd cohort | [68] | |
 Variation methyl donor intake | Changes in mean methylation across BOLA3, LOC654433, EXD3, ZFYVE28, RBM46, and ZNF678 in blood leukocytes (n = 126) and hair follicles (n = 82) of 2- to 8-month infants | No validation performed | [69] | |
 Periconceptional BMI | Decreased mean methylation across BOLA3, LOC654433, EXD3, ZFYVE28, RBM46, and ZNF678 in blood leukocytes (n = 126) and hair follicles (n = 82) of 2- to 8-month infants | No validation performed | [69] | |
 Gestational weight gain early pregnancy | Increased methylation 4 CpGs in MMP7, KCNK4, TRPM5, and NFKB1 in newborn cord blood (n = 88), no association in 2nd cohort (n = 170) | Not validated (technical and 2nd cohort) | [71] | |
 Preconceptional BMI | Differential methylation in ZCCHC10 in newborn cord blood (n = 308), other less significant sites were found in WNT16, ACPL2, C18orf8, ANGPTL2, SAPCD2, and ADCY3 | No | [73] | |
 Gestational Diabetes | 42 CpGs in newborn cord blood (n = 136). ~1/3 of CpGs overlapped with sites that were associated with maternal glucose levels (n = 36) or micronutrient supplementation (n = 59) in 2 other child cohorts | Yes, technical and 2 other cohorts | [70] | |
 Gestational Diabetes | No differentially methylated sites (after FDR correction) in cord blood and placenta (n = 44) but enrichment of sites in genes metabolic disease pathway | No | [72] | |
Intervention | ||||
 Weight loss surgery | Change in methylation at 3601 CpGs (195 DMRs) in subcutaneous adipose tissue and 15 CpGs in omental adipose tissue (n = 15), some DMRs overlapping known obesity and T2DM loci | Yes, technical | [64] | |
 Weight loss surgery | 227 DMRs, methylation in these regions also changed with high-fat feeding in mice | No, only in mice | [43] | |
 Weight loss surgery in liver disease | Before surgery 467 differentially methylated CpGs between control (n = 18), healthy obese (n = 18), steatosis (n = 12), and NASH (n = 15) liver samples. After surgery changes in methylation at 113 CpGs, disease-associated methylation was reversible at the HOXB1, PRKCZ, SLC38A10, and SECTM1 loci | No, but for baseline technical and 2nd cohort | [79] | |
 Weight loss | Methylation profiles of RYR1, TUBA3C and BDNF in PBMCs of successful weight loss maintainers (n = 16) more closely resembled lean (n = 16) than obese subjects (n = 16) | No | [77] | |
 Endurance and strength exercise | Changes in DNA methylation in skeletal muscle of obese T2DM subjects (n = 17) after 16 weeks, most pronounced with endurance exercise | No | [83] | |
 High-fat diet (5 days) | PPARGC1A DNA methylation across 4 CpGs increased in subcutaneous adipose tissue of lean adults born with low birth weight (n = 19) but not in controls (n = 26) | No | [76] |