Table 1 Description and summary of data from 21 studies that investigated associations between DNA methylation and cardiometabolic risk factors

Daniel Castellano-Castillo [9]

Spain 2018

case-control

108

Non MetS (55)

Age:48.4 ± 13.9

BMI:29.8 ± 7.9

M/F: 52/48

MetS (53)

Age:52.7 ± 14.6

BMI:36.4 ± 10.9

M/F: 44/56

Visceral Adipose Tissue

LINE-1

P1-P6^*1

age, sex

Negative correlation between LINE-1 P2 and the MetS index and no correlation at P1, P3, P4, P5, P6.

Negative correlations between LINE-1 P1, P2 and P5 and glucose levels. No correlation between LINE-1 and P3, P4, P6 and glucose levels.

18

Valérie Turcot [7]

Canada 2012

case-control

176

severely obese undergoing a biliopancreatic

diversion with sleeve gastrectomy to treat obesity

Non MetS (98)

Age mean: 34.9 ± 8.1

BMI mean:49.8 ± 8.4

M/F: 14/84

MetS (88)

Age mean: 35.3 ± 7.3

BMI mean:53.8 ± 10.8

M/F: 20/68

Visceral Adipose Tissue

LINE-1

age, sex and smoking

LINE-1%meth levels in VAT were associated negatively with fasting plasma glucose, blood pressure and MetS.

18

Jose Luiz Marques-Rocha [10]

Brazil 2016

cross-sectional

156

M/F:91/65

Age mean: 23.1 ± 3.5

BMI mean: 22 ± 2.9

WBC

LINE-1

calories, sex, age, smoking, regular physical activity

LINE-1 methylation associated with body fat, waist girth and waist-to-hip ratio, total fat mass, blood pressure.

20

Mark S Pearce [8]

UK 2012

Cohort (The Newcastle Thousand Families Study)

228

Age: 49–51 y

BMI mean: 25.70 (22.94–28.93)

M/F: 85/143

peripheral blood samples

LINE-1

Sex

Increased LINE-1 DNA methylation was associated with increasing fasting glucose, total cholesterol, total triglycerides, and LDL cholesterol and with decreasing HDL cholesterol, and HDL:LDL ratio

18

Haley L. Cash [21]

US 2011

case-control

355

American Samoa (198)

Age mean:

M (57): 36.1 ± 5.4

F(141):29.7 ± 6.8

BMI mean:

M (57): 34.8 ± 6.6

F(141): 36.0 ± 9.2

Samoa (157)

Age mean:

M (31): 39.2 ± 5.7

F(126): 29.2 ± 6.2

BMI mean:

M (31): 28.7 ± 5.4

F(126): 31.1 ± 5.8

peripheral lymphocyte

LINE-1

age, sex, BMI

Significant positive association between BMI and HDL with LINE-1 methylation

Significant negative association between LDL and LINE-1 methylation

17

Stacey E Alexeeff [22]

US 2013

longitudinal study

(cohort)

798

M/F:

Age mean: 74

(55–100 y)

BMI mean: 27.5

buffy coat

LINE-1

BMI, age, smoking, T2D, alcohol, race, IHD/MI, Neut count, season, day of week.

LINE-1 methylation Inversely associated with DBP,

LINE-1 methylation association with SBP was weak.

17

Yoshiki Tsuboi [13]

Japan 2018

cross-sectional study

420

M/F:187/233

Age mean: 61.46 y

BMI mean: 24.16

WBC

LINE-1

sex, age, smoking, alcohol, BMI, CRP, anti hyperlipidemic drug use

Significant positive association between LINE-1 DNA methylation and LDL/HDL ratio.

Negative and weak association between LINE-1 DNA methylation and HDL.

19

Carolina Ferreira Nicoletti [14]

Brazil-Spine 2016

cross-sectional study

45

control group (9)

normal weight individuals

M/F: 0/9

Age mean: 31.7 ± 8.6

BMI mean: 22.0 ± 2.0

obese with energy restriction group (22)

M/F: 0/22

Age mean: 52.6 ± 9.9

BMI mean: 38.2 ± 3.7

obese with bariatric surgery group (14)

M/F: 0/14

Age mean: 35.5 ± 10.1

BMI mean: 44.6 ± 6.2

buffy coats

LINE-1

age and BMI

Significant association between LINE-1 methylation and serum glucose levels.

18

Liliane Pfeiffer [11]

Germany 2014

Augsburg cohort

2747

M/F: 1341/1406

Age mean: 61.63 y BMI mean: 27.46

whole blood samples

ABCG1

cg06500161

SREBF1

cg11024682

age,sex, BMI, smoking, alcohol, lipid lowering drugs, physical activity, history of MI, hypertension, HbA1c, CRP, WBC count

Opposite directions ABCG1 methylation association with HDL and triglyceride levels.

Association between triglyceride levels and ABCG1, SREBF1.

20

Alexis C. Frazier-Wood [23]

US 2014

GOLDN cohort

994

Discovery (663)

M/F: 312/351

Age mean: 48.6 ± 16.4

BMI mean:

Replication (331)

M/F: 165/166

Age mean: 47.7 ± 16.2

BMI mean:

CD4 + T cells

ABCG1

cg06500161

age, sex, study site, and the first four principal components generated to estimate T-cell

purity as fixed effects, and pedigree as a random effect using the lmekin function of the kinship package in R

LDL associated with ABCG1 methylation.

18

Tasnim Dayeh [12]

Sweden 2016

Botnia prospective study

258

non-diabetic at

baseline:

Controls (129)

M/F:62/67

Age mean: 51.4 ± 9.1

BMI mean: 27.6 ± 3.0

Converters(129)

M/F:65/64

Age mean: 52.8 ± 12.3

BMI mean: 28.8 ± 4.3

blood

ABCG1

cg06500161

PHOSPHO1

cg02650017

age, gender, fasting glucose, and family relation

Positive correlation between DNA methylation at the ABCG1 locus cg06500161 with BMI, HbA1c, fasting insulin, and triglyceride levels.

Positive correlation between DNA methylation at the PHOSPHO1 locus cg02650017 with HDL levels.

DNA methylation at the ABCG1 locus cg06500161: 9% increased risk for future T2D

DNA methylation at the PHOSPHO1 locus cg02650017: 15% decreased risk for future T2D

19

Eliza Walaszczyk [15]

Netherland 2017

case–control sample

Lifelines cohort

198

Type 2 diabetic (100)

M/F: 52/48

Age mean: 62 (53–69y)

BMI mean: 30.8 ± 4.7

Control individuals (98)

M/F: 44/54

Age mean: 50 (46–63y)

BMI mean: 25.3 ± 3.6

whole blood

ABCG1

SREBF1

age, sex, measured blood cell composition, plate number and

position on the plate as covariates

ABCG1 methylation associated with FBS, TG and TC

SREBF1 methylation associated with FBS, TG, TC and LDL

18

John C Chambers [16]

London 2015

prospective nested case-control (LOLIPOP)

13,535

M/F: 8175/5360

Age mean: 49.1 ± 10.9

BMI mean: 27.0 ± 4.4

peripheral blood leucocytes

SREBF1

PHOSPHO1

ABCG1

Age, sex

Methylation at SREBF1, PHOSPHO1, and ABCG1association with quantitative measures of total and regional body fat distribution

18

Jennifer Kriebel [25]

Germany 2016

KORA F4 Study

1448

non-diabetic individuals

M/F: 682/766

Age mean: 59 (32-81y)

BMI mean: 27.1

whole blood

SREBF1

cg11024682

ABCG1

cg06500161

age, sex, estimated white blood cell proportions, smoking, BMI

Significant associations between cg06500161 (ABCG1) methylation and waist circumference, triglycerides, fasting glucose, and 2-hour glucose, fasting insulin, CD8 + T cells, and monocytes

Significant associations between cg09694782 (SREBF1) methylation and age, fasting insulin, and HOMA-IR.

19

Kim V. E. Braun [24]

Netherland 2017

Rotterdam Study

1485

Discovery (725)

M/F: 336/389

Age mean: 59.9 ± 8.2

BMI mean: 27.6 ± 4.6

Replication (760)

M/F: 334/426

Age mean: 67.7 ± 5.9

BMI mean: 27.8 ± 4.2

whole blood

SREBF1

cg11024682

ABCG1

cg06500161

age, gender, current smoking, leukocyte proportions, array number, and position on array

Association between ABCG1 methylation and HDL

Association between ABCG1, and SREBF1 methylation with triglycerides

21

Ping Peng [26]

China 2014

case-control

139

CHD patients (85)

M/F: 58/67.4

Age mean: 61.33 ± 9.22

control group (54)

M/F: 31/57.4

Age mean: 56.35 ± 9.0

peripheral blood

ABCG1

age, gender, smoking, lipid level, hypertension, and diabetes

Significant statistical association of the promoter Hyper-methylation of the ABCG1 gene with CHD risk

ABCG1 and GALNT2 gene promoter regions are positively associated with CHD both in the male group

20

S. Sayols-Baixeras [27]

Spain 2016

REGICOR and

Framingham study

cross-sectional

2858

REGICOR discovery sample (645)

M/F: 316/329

Age mean: 63.2 ± 11.7

BMI mean: 26.9 ± 4.1

Framingham (2542)

M/F: 1164/1378

Age mean: 66.3 ± 8.9

BMI mean: 28.2 ± 5.4

whole peripheral blood

SREBF1

PHOSPHO1

ABCG1

sex, age, smoking status,batch effect and estimated cell count

Positive association between SREBF2 methylation and TC, in the same direction as the association between SREBF1 and TG.

Significant association between methylation levels of SREBF1 and HDL in the opposite direction to that observed with TG.

Direct association between PHOSPHO1 methylation and HDLholesterol levels.

21

Simon-Pierre Guay [28]

Canada 2014

Case-controle study

61

severely obese non-FH (30)

BMI > 40

familial hypercholesterolemia (61)

M/F:61/0

Whole blood

ADRB3

age, waist circumference, fasting

triglyceridemia

Higher ADRB3DNA methylation levels were significantly associated with lower low-density lipoprotein cholesterol levels in FH, and with a lower waist-to-hip ratio and higher blood pressure in severely obese men.

17

Raquel PatríciaAtaíde Lima [31]

Brazil 2019

cross-sectional representative study

265

M/F: 79/186

Age mean: 40.3 ± 14.3

(20–59)

BMI mean: 27.08 ± 5.88

leukocytes

ADRB3

–

LDL above the median had a 164% higher chance of ADRB3 hyper-methylation, whereas individuals with triglyceride values above the median had a higher chance of hyper-methylation.

16

Andrée-Anne Houde [29]

Canada 2015

Cross-sectional

73

men and premenopausal women (BMI > 40 kg/m2) undergoing bioliopancreatic diversion with duodenal switch to treat obesity

(severely obese)

M/F: 33/40

whole blood

SAT

VAT

LEP

Age, sex and waist circumference

LEP DNA methylation levels in blood cells were negatively associated with body mass index (BMI).

Fasting LDL levels positively correlated with DNA methylation levels at LEP-CpG11 and -CpG17 in blood and SAT and with ADIPOQ -CpGE1 and - CpGE3 DNA methylation levels in SAT and CpGE1 in VAT.

Associations between LDL levels and both LEP and ADIPOQ DNA methylation levels.

18

Jonathan Y Huang 2017 [30]

Israel

Sub-cohort

589

M/F: 0/589

Age mean:32

BMI mean: 27.08 ± 5.88

maternal pre-pregnancy BMI ≥27 kg/m2 and offspring birth weight ≤ 2500 g or ≥ 4000 g

peripheral blood (buffy coat)

LEP

ethnic origin, offspring age at blood draw, maternal characteristics (pre-pregnancy BMI, gestational weight gain, age, parity, education), paternal characteristics (education and smoking status), offspring variables (childhood overweight, education, parity, current smoking status)

ABCA1 methylation appeared to be directly related to both maternal gestational weight Gain and some markers of glucose homeostasis.

LEP methylation associated with waist-to-hip ratio

19