Circulating triglycerides are associated with human adipose tissue DNA methylation of genes linked to metabolic disease

Tina Rönn; Alexander Perfilyev; Josefine Jönsson; Karl-Fredrik Eriksson; Sine W Jørgensen; Charlotte Brøns; Linn Gillberg; Allan Vaag; Elisabet Stener-Victorin; Charlotte Ling

doi:10.1093/hmg/ddad024

. 2023 Feb 8;32(11):1875–1887. doi: 10.1093/hmg/ddad024

Circulating triglycerides are associated with human adipose tissue DNA methylation of genes linked to metabolic disease

Tina Rönn ¹, Alexander Perfilyev ², Josefine Jönsson ³, Karl-Fredrik Eriksson ⁴, Sine W Jørgensen ⁵, Charlotte Brøns ^6,^✉, Linn Gillberg ⁷, Allan Vaag ⁸, Elisabet Stener-Victorin ⁹, Charlotte Ling ^10,^✉

PMCID: PMC10196668 PMID: 36752523

Abstract

Dysregulation of circulating lipids is a central element for the metabolic syndrome. However, it is not well established whether human subcutaneous adipose tissue is affected by or affect circulating lipids through epigenetic mechanisms. Hence, our aim was to investigate the association between circulating lipids and DNA methylation levels in human adipose tissue. DNA methylation and gene expression were analysed genome-wide in subcutaneous adipose tissue from two different cohorts, including 85 men and 93 women, respectively. Associations between DNA methylation and circulating levels of triglycerides, low-density lipoprotein, high-density lipoprotein and total cholesterol were analysed. Causal mediation analyses tested if adipose tissue DNA methylation mediates the effects of triglycerides on gene expression or insulin resistance. We found 115 novel associations between triglycerides and adipose tissue DNA methylation, e.g. in the promoter of RFS1, ARID2 and HOXA5 in the male cohort (P ≤ 1.1 × 10⁻⁷), and 63 associations, e.g. within the gene body of PTPRN2 and COL6A3 in the female cohort. We further connected these findings to altered mRNA expression levels in adipose tissue (e.g. HOXA5, IL11 and FAM45B). Interestingly, there was no overlap between methylation sites associated with triglycerides in men and the sites found in women, which points towards sex-specific effects of triglycerides on the epigenome. Finally, a causal mediation analysis provided support for adipose tissue DNA methylation as a partial mediating factor between circulating triglycerides and insulin resistance. This study identified novel epigenetic alterations in adipose tissue associated with circulating lipids. Identified epigenetic changes seem to mediate effects of triglycerides on insulin resistance.

Introduction

Metabolic diseases, including type 2 diabetes, obesity, insulin resistance, hypertension and abnormal lipid metabolism, increase because of environmental factors and are a growing threat to human health. Blood lipid levels influence the risk for cardiovascular disease, but can be modified by lifestyle changes, such as improved diet, physical activity, alcohol consumption and smoking habits, or drug therapy (1,2). Although lipid levels are considered heritable traits, genome-wide association studies have only explained 12% of the variance for blood lipids (total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL) and triglycerides) (3). As a second layer of genomic information that may be partly heritable, the epigenome, comprises inducible changes to the genome, altering gene regulation and expression, without altering the genetic code. Epigenetic mechanisms can link environmental factors to disease and may be important in phenotype transmission (4). Thereby, epigenetic factors, including DNA methylation, could be a link to explain the missing heritability for lipid traits and a mechanism for the rapid adaptation to a changing environment.

We have previously characterized the human DNA methylome in adipose tissue from healthy individuals, as well as epigenetic adaptation in response to exercise (5). Also, age, BMI and HbA1c were associated with epigenetic variation of candidate genes for obesity, type 2 diabetes and cancer in human adipose tissue (6). Genome-wide studies of adipose tissue DNA methylation and genetic variation showed numerous associations between genotype and methylation, some also linked to gene expression and metabolic traits (7). Together, these studies support altered adipose tissue DNA methylation in the interplay between environmental factors and metabolism.

Some studies have also explored the relation between circulating lipids and genome-wide DNA methylation levels, however, mainly in blood cells (8–13). The nucleated blood cells have a short life span (days), in contrast to cells of metabolically active tissues, e.g. the adipocytes which in adults have an average age of almost 10 years (14). Pfeiffer et al. (13) reported associations between lipid levels and methylation in blood of genes involved in cholesterol and fatty acid metabolism. Using mendelian randomization, Dekkers et al. (9) showed that differential methylation in circulating blood cells is a consequence rather than cause of altered lipid levels. Although large cohorts, these studies only identified a small number of CpG sites with genome-wide significance. Hedman et al. (11) reported methylation of numerous CpG sites in blood to be associated with lipid traits, some also connected to cis-expression and future risk of coronary heart disease. More recently, Jhun et al. (12) reported 26 CpG sites associated with triglycerides, based on DNA methylation in blood from three ethnic groups comprising more than 16 000 samples. Six CpGs were also associated with expression of the corresponding gene (PHGDH, SLC7A11, CPT1A and ABCG1). Gomez-Alonso et al. (10) performed an epigenome-wide association study of lipid-related metabolic measures, connecting blood DNA methylation with lipid composition and concentration. Here, a site annotated to SREBF1 was also associated with SREBF1 expression in adipose tissue.

Hence, based on genome-wide data from blood cells, there is evidence for differential DNA methylation associated with lipid traits, but similar genome-wide studies connecting circulating lipids with the methylome of human adipose tissue are missing. Interestingly, we have previously shown that epigenetic biomarkers in blood can mirror epigenetic signatures in target tissues for metabolic diseases, i.e. adipose tissue (6,15). Based on its central function for metabolism, we hypothesize that adipose tissue is a likely organ to be affected by or affect circulating lipids through epigenetic mechanisms. The main aim of this study was therefore to investigate the association between circulating triglycerides and the genome-wide DNA methylation pattern in human adipose tissue from two different cohorts, including 85 men and 93 women, respectively. Because of known sex-specific effects on lipid metabolism (16) and DNA methylation in several tissues (17,18), men and women were studied separately. Triglyceride levels and differential DNA methylation were further connected to gene expression in adipose tissue. Next, causal mediation analyses provided examples where adipose tissue DNA methylation may act as a mediator of triglyceride levels on peripheral insulin resistance. Finally, we also investigated the association between adipose tissue DNA methylation and circulating levels of LDL, HDL and total cholesterol, and compared these results with the triglyceride associations.

Results

Circulating triglycerides are associated with altered DNA methylation in adipose tissue

The first cohort of this study includes 85 men without known diseases, with an age span from 23 to 46 years and with a wide range in BMI (17.5–34.0 kg/m²) (Table 1). Here, we investigated whether circulating triglycerides are associated with DNA methylation in adipose tissue. Using a random effect mixed model adjusted for age and BMI, we found triglyceride levels associated with DNA methylation of 14 943 individual CpG sites based on false discovery rate (FDR) below 5% (q < 0.05; Supplementary Material, Table S1). These were equally distributed between positive and negative correlations and annotated to 8 536 genes (each CpG sites can be annotated to several genes). As the large number of CpG sites associated with triglycerides based on q < 0.05 may suggest statistical inflation, we performed Bonferroni correction for the continued analyses, applying a threshold for epigenome-wide significance of P = 1.1 × 10⁻⁷ (Fig. 1A). Amongst the 115 CpG sites with P ≤ 1.1 × 10⁻⁷, 95 were annotated to coding genes, 15 to non-coding RNA and 13 were intergenic, and 108 of these associations remained after adjustment for cellular composition (P = 1.1 × 10⁻⁷; Supplementary Material, Table S1) (19). Furthermore, methylation of one of these 115 sites was also associated with age, whereas none were associated with BMI based on Bonferroni correction (Supplementary Material, Table S1). The most significant associations between triglycerides and adipose tissue DNA methylation of sites annotated to coding genes were found in the RSF1 promoter region, the promoter of ANKS1A and TAF11, and within the gene body of RPH3AL (Fig. 1A–D).

Table 1.

Clinical characteristics of study participants

Phenotype	Male cohort (n = 85)	Female cohort (n = 93)	P-value
Age (years)	29.0 ± 6.7 (23–46)	29.2 ± 4.3 (21–37)	n.s.
BMI (kg/m²)	25.2 ± 3.4 (17.5–34.0)	27.2 ± 6.7 (18.2–45.0)	0.015
Triglycerides (mmol/l)	1.11 ± 0.56 (0.3–3.5)	0.90 ± 0.50 (0.3–3.4)	0.009
LDL (mmol/l)	2.89 ± 0.81 (1.0–4.7)	2.39 ± 0.82 (0.8–5.0)	6 × 10⁻⁵
HDL (mmol/l)	1.19 ± 0.27 (0.5–1.8)	1.55 ± 0.40 (0.9–2.8)	6 × 10⁻¹¹
Total cholesterol (mmol/l)	4.57 ± 0.83 (2.1–6.5)	4.28 ± 0.88 (2.7–7.5)	0.024
HOMA-IR	1.24 ± 0.56 (0.5–3.3)	1.97 ± 1.67 (0.4–10.4)	2 × 10⁻⁴

Open in a new tab

Data are expressed as the mean ± SD (range). P-values for differences between the two cohorts are based on two-sample t-tests (two-tailed). n.s.: not significant (P > 0.05).

Associations between circulating triglycerides and adipose tissue DNA methylation. (A) Manhattan plot showing chromosomal distribution and significance level for all 448 785 CpG sites analysed in the male cohort (unadjusted P-values without Houseman correction; genes of interest being labelled). Associations between DNA methylation and triglycerides were analysed using a random effect mixed model, where the cohort was considered the random effect variable and age, BMI and triglycerides were included as fixed factors. Threshold for epigenome-wide significance (Bonferroni) P < 1.1 × 10⁻⁷. Top associations between triglycerides and DNA methylation were identified for (B) cg16516888 in the *RSF1* promoter, (C) cg17149859 in the promoters of *ANKS1A* and *TAF11* and (D) cg26352439 in the gene body of *RPH3AL* in the male cohort. (E) Manhattan plot showing chromosomal distribution and significance level for all 453 039 CpG sites analysed in the female cohort, using linear regression including triglycerides, PCOS status (case or control), age and BMI in the model (unadjusted P-values without Houseman correction; genes of interest being labelled). The threshold was q < 0.05 (P ≤ 6.7 × 10⁻⁶). Top associations between triglycerides and DNA methylation were identified for (F) cg20194947 and (G) cg21386952 in the gene bodies of *PTPRN2* and *COL6A3*, respectively, in the female cohort. The plotted data are not adjusted for any confounding factors, to visualize the spread in triglycerides and methylations levels. The presented p-values are produced based on the full statistical model, including age, BMI and batch effects, and not corrected for multiple testing.

The second cohort of this study consisted of 93 women, part of a polycystic ovary syndrome (PCOS) case control cohort. Compared with the males, the female cohort had a slightly smaller age span (21–37 years), but a wider range in BMI (18.2–45.0 kg/m²) (Table 1). A linear regression including triglycerides, PCOS status (case or control), age and BMI showed that circulating triglycerides are associated with adipose tissue DNA methylation of 63 CpG sites after FDR analysis to control for multiple testing (q < 0.05, P ≤ 6.7 × 10⁻⁶; Fig. 1E and Supplementary Material, Table S2). Of these sites, 47 were annotated to coding genes, one to non-coding RNA and 16 were intergenic. These associations remained after adjustment for cellular composition (q < 0.05; Supplementary Material, Table S2) (19). Eight of these 63 CpG sites additionally associated with BMI (q < 0.05), whereas none associated with age (Supplementary Material, Table S2). In the female cohort, associations between triglycerides and DNA methylation were identified, e.g. within the gene bodies of PTPRN2 and COL6A3 (Fig. 1F and G). Interestingly, there was no overlap between the 63 CpG sites associated with triglycerides in women and the 115 sites found in men.

Circulating triglyceride levels are associated with both DNA methylation and cis-gene expression of several genes in adipose tissue

To further translate the differential methylation data into a phenotype, we extracted mRNA expression data for all genes within 10 kb distance, i.e. in cis, from the 115 CpG sites associated with triglycerides (epigenome-wide significance; P ≤ 1.1 × 10⁻⁷) in the male cohort. We found that circulating triglycerides were nominally associated with mRNA expression for 36 of those genes (P < 0.05; Table 2). Additionally, 19 of these genes, where both DNA methylation and gene expression associated with triglyceride levels, displayed correlations between methylation and expression levels (P < 0.05; Table 2). These include IL11, FAM45B and HOXA5 (Fig. 2).

Table 2.

Genes where circulating triglyceride levels associated with both DNA methylation and cis-gene expression in male adipose tissue

Gene name	Association triglycerides—DNA methylation						Association triglycerides—mRNA expression			Correlation DNA methylation—mRNA expression
Gene name	ILMNID	Gene region	P-value	Bonf. P	Coeff	sem	p-value	Coeff	sem	Corr coeff (95% CI)	p-value
ANXA5	cg08715877	5'UTR;1stExon	2.9E−08	0.0132	0.29	0.05	0.011	0.078	0.030	0.29 (0.08–0.48)	0.008
AP5S1	cg23561371	Body	7.0E−08	0.031	−0.25	0.04	0.005	0.080	0.028	0.06 (−0.15 to 0.27)	n.s.
CTTNBP2	cg05951860	Body	6.1E−08	0.0275	−0.22	0.04	0.012	−0.108	0.042	0.01 (−0.20 to 0.23)	n.s.
DDX42	cg25365221	TSS1500	1.3E−08	0.0056	0.14	0.02	0.015	−0.056	0.023	0.03 (−0.19 to 0.24)	n.s.
FAM45B	cg20227259	Body	2.7E−08	0.0119	0.16	0.03	0.011	0.084	0.033	0.46 (0.27–0.62)	<0.0001
GON4L	cg07222421	Downstream	2.3E−09	0.0010	−0.35	0.05	1.6E−04	−0.064	0.016	0.28 (0.07–0.47)	0.011
HIST1H2AE	cg05480042	TSS200	6.5E−10	2.9E−04	−0.29	0.04	0.000	0.138	0.036	−0.20 (−0.39 to 0.02)	n.s.
HOXA5	cg09880291	TSS1500	2.8E−09	0.0013	−0.53	0.08	0.009	−0.102	0.038	−0.85 (−0.90 to −0.77)	<0.0001
HOXA5	cg05835726	TSS1500	9.6E−08	0.0431	−0.41	0.07	0.009	−0.102	0.038	−0.83 (−0.89 to −0.75)	<0.0001
HOXA6	cg09880291	Downstream	2.8E−09	0.0013	−0.53	0.08	0.013	−0.107	0.042	−0.87 (−0.91 to −0.80)	<0.0001
HOXA6	cg05835726	Downstream	9.6E−08	0.0431	−0.41	0.07	0.013	−0.107	0.042	−0.86 (−0.90 to −0.79)	<0.0001
IL11	cg13508283	TSS200	1.9E−08	0.0086	−0.41	0.07	0.016	0.057	0.023	−0.32 (−0.50 to −0.11)	0.004
KIF21A	cg15418826	Body	1.2E−09	5.4E−04	0.27	0.04	0.048	−0.074	0.037	0.49 (0.31–0.64)	<0.0001
LRRC63	cg03436178	TSS1500	3.1E−08	0.014	−0.29	0.05	0.046	0.084	0.041	−0.04 (−0.25 to 0.18)	n.s.
NAGS	cg16182148	Downstream	5.3E−08	0.024	0.16	0.03	0.011	0.071	0.027	0.18 (−0.03 to 0.38)	n.s.
NCAM1	cg11994052	Body	5.3E−08	0.0237	−0.32	0.05	0.043	−0.105	0.051	−0.06 (−0.27 to 0.15)	n.s.
NFIA	cg11531787	Body	9.7E−08	0.0434	−0.33	0.06	0.046	−0.054	0.026	−0.07 (−0.28 to 0.14)	n.s.
NOL10	cg00029053	1stExon;5'UTR	1.9E−08	0.0086	0.16	0.03	0.024	−0.058	0.025	0.50 (0.31–0.64)	<0.0001
PPP1R36	cg13210403	9.9 kb upstream	1.5E−08	0.0069	−0.33	0.05	0.029	0.108	0.049	−0.18 (−0.38 to 0.04)	n.s.
PRAC1	cg17462165	6.7 kb upstream	6.3E−08	0.028	0.24	0.04	0.009	0.055	0.021	0.30 (0.09–0.49)	0.0054
PSMD14	cg21949512	Downstream	5.0E−10	2.2E−04	0.29	0.04	0.021	0.107	0.046	0.19 (−0.02 to 0.39)	n.s.
PYY	cg16182148	10 kb upstream	5.3E−08	0.0236	0.16	0.03	0.013	0.078	0.031	−0.23 (−0.43 to −0.02)	0.0339
RSF1	cg16516888	TSS1500	2.4E−11	1.1E−05	−0.47	0.06	0.035	−0.067	0.031	0.005 (−0.21 to 0.22)	n.s.
SEC22B	cg18788524	1stExon	2.1E−08	0.0092	0.22	0.04	0.035	0.068	0.032	0.54 (0.37–0.68)	<0.0001
SETD9	cg14703610	Body	5.2E−10	2.3E−04	−0.29	0.04	0.009	−0.095	0.035	0.27 (0.06–0.46)	0.0129
SLC25A51	cg13986249	Body	1.1E−08	0.0049	0.22	0.03	0.009	0.138	0.51	0.10 (−0.11 to 0.31)	n.s.
SNORD78	cg04705273	2.5 kb upstream	3.0E−08	0.0136	0.19	0.03	0.048	−0.145	0.072	0.29 (0.08–0.47)	0.0085
TACSTD2	cg13786801	1stExon	9.7E−08	0.0436	0.23	0.04	0.037	0.106	0.050	0.26 (0.05–0.45)	0.016
TFAP2A-AS1	cg13486532	Downstream	1.0E−07	0.0449	0.28	0.05	0.040	0.082	0.040	0.08 (−0.14 to 0.29)	n.s.
TMEM101	cg16182148	Body	5.3E−08	0.0236	0.16	0.03	0.002	0.069	0.022	−0.25 (−0.45 to −0.41)	0.02
TMEM190	cg13508283	6.3 kb upstream	1.9E−08	0.0086	−0.41	0.07	0.016	0.058	0.023	−0.38 (−0.55 to −0.18)	0.0004
UHRF1BP1	cg17149859	Downstream	7.1E−11	3.2E−05	−0.37	0.05	0.026	0.085	0.037	−0.08 (−0.29 to 0.13)	n.s.
VASH1	cg03535239	Body	4.3E−10	1.9E−04	−0.43	0.06	0.043	0.087	0.042	0.08 (−0.14 to 0.29)	n.s.
VSIG10	cg23354014	2.7 kb upstream	2.7E−08	0.0119	0.18	0.03	0.009	−0.066	0.024	−0.46 (−0.61 to −0.27)	<0.0001
WDR17	cg24141863	TSS200	4.4E−08	0.0196	−0.21	0.04	0.005	−0.172	0.060	−0.08 (−0.29 to 0.14)	n.s.
WSB2	cg23354014	Body	2.7E−08	0.0119	0.18	0.03	0.019	0.074	0.031	−0.3571 (−0.53 to −0.15)	0.0009
YAE1D1	cg27071152	Body	1.4E−11	6.1E−06	0.50	0.06	4.6E−05	0.122	0.028	0.61 (0.45–0.73)	<0.0001
YLPM1	cg00878827	TSS200	2.1E−09	9.2E−04	0.19	0.03	0.020	−0.053	0.022	−0.02 (−0.24–0.19)	n.s.

Open in a new tab

Bonf. P: Bonferroni corrected P-values for epigenome-wide significance (i.e. P × 448 785), Coeff: coefficient, sem: standard error of mean, Corr: Pearson correlation, CI: confidence interval, n.s.: not significant (P > 0.05).

Genes where triglycerides are associated with both adipose tissue DNA methylation and gene expression in the male cohort, and with a correlation between methylation and expression. Triglycerides display a negative association with cg13508283 close to *IL11* transcription start site (A), and a positive association with *IL11* mRNA expression (B). Also, there is a correlation between *IL11* DNA methylation and gene expression (C). For *FAM45B*, triglyceride levels are positively associated with both DNA methylation (cg20227259) (D) and mRNA expression (E), and this gene body methylation correlates with expression (F). *HOXA5* gene expression correlates with two CpG sites within a CpG island upstream of *HOXA5* transcription start site (G, H). The plotted data are not adjusted for confounding factors, to visualize the spread in triglycerides and methylation or expression levels. P-values in a–b, d–e are produced based on the full statistical model, including age, BMI and batch effects. In c, f–h, P-values are based on Pearson correlation. The presented P-values are not corrected for multiple testing.

Furthermore, amongst the 63 CpG sites found associated with triglycerides in the female cohort (q < 0.05), seven were within 10 kb distance from six genes that also displayed association between triglyceride levels and adipose tissue mRNA expression (Table 3). These include, e.g. MYO9B and PPM1A (Fig. 3).

Table 3.

Genes where circulating triglyceride levels associated with both DNA methylation and gene expression in female adipose tissue

	Association triglycerides—DNA methylation						Association triglycerides—mRNA expression			Correlation DNA methylation—mRNA expression
Gene name	ILMNID	Gene region	P-value	q-value	Coeff	sem	P-value	Coeff	sem	Corr coeff (95% CI)	P-value
MYO9B	cg18874907	Body	2.0E−06	0.043	−0.26	0.05	0.027	0.14	0.06	−0.07 (−0.27 to 0.14)	n.s.
PPM1A	cg22786756	5'UTR;Body	3.4E−06	0.043	−0.25	0.05	0.003	−0.25	0.08	0.19 (−0.02 to 0.38)	0.07
RRAS	cg06613203	TSS200	2.8E−06	0.043	−0.18	0.04	0.033	0.10	0.05	−0.19 (−0.38 to 0.01)	0.07
TRAPPC9	cg10726725	Body	3.9E−06	0.044	−0.24	0.05	0.020	0.16	0.07	0.03 (−0.18 to 0.23)	n.s.
TRAPPC9	cg12195808	Body	6.5E−06	0.048	−0.21	0.04	0.020	0.16	0.07	0.05 (−0.16 to 0.25)	n.s.
TRIM26	cg16446015	Body	4.0E−06	0.044	−0.28	0.06	0.005	0.10	0.03	0.10 (−0.10 to 0.30)	n.s.
UBE2L3	cg20703614	Downstream	1.7E−06	0.040	−0.24	0.05	0.016	0.14	0.06	−0.37 (−0.53 to −0.18)	0.0003

Open in a new tab

Coeff: Coefficient, sem: standard error of mean, Corr: Pearson correlation, CI: confidence interval, n.s.: not significant (P > 0.05).

Genes where triglycerides are associated with both adipose tissue DNA methylation and gene expression in the female cohort, and with a correlation between methylation and expression. Triglycerides display a negative association with DNA methylation of cg18874907 within the gene body of *MYO9B* (A), and a positive association with *MYO9B* mRNA expression (B). For *PPM1A*, both DNA methylation of cg22786756 in the 5′UTR (C) and mRNA expression (D) was negatively associated with circulating triglycerides. Triglycerides display a negative association with cg20703614 (E), located 9.5 kb downstream from *UBE2L3*, and a positive association with *UBE2L3* expression (F). Also, cg20703614 methylation correlates with *UBE2L3* expression (G). The plotted data are not adjusted for confounding factors, to visualize the spread in triglycerides and methylation or expression levels. The presented P-values are produced based on the full statistical model, including age, BMI and batch effects, and not corrected for multiple testing.

Adipose tissue DNA methylation as a partial causal mediator of the effect of circulating triglyceride levels on insulin resistance

It is known that elevated triglyceride levels are associated with overall insulin resistance based on HOMA-IR and locally in adipose tissue based on Adipo-IR (20). Using casual mediation analysis (21), we investigated whether DNA methylation of the CpG sites found to be associated with circulating triglycerides (n = 115 in the male and n = 63 in the female cohort) may mediate the effect of triglycerides on insulin resistance (HOMA-IR). The causal mediation analysis separates the total effect of triglyceride levels (independent variable) on HOMA-IR (dependent variable) into (i) average causal mediation effects (ACME), the mediator of interest in our case DNA methylation, and (ii) average direct effects (ADE), the effect acting directly or via a mediator other than what is under study (Fig. 4). First, we could confirm that circulating triglycerides have an overall significant association with HOMA-IR in our study as well (total effect, β = 0.23, P = 0.014 in the male and β = 0.89, P < 0.001 in the female cohort; Table 4). Second, in the male cohort, we found methylation of eight CpG sites suggested as partial causal mediators of the effect that circulating triglycerides have on HOMA-IR (Table 4), e.g. DNA methylation in the promoter region of TMEM95 and FAIM, and within the gene body of SPTBN1 and SLC25A17. In addition, adipose tissue mRNA expression of two genes, ANXA5 and CTTNBP2, was also found to partially mediate the effect of triglyceride levels on HOMA-IR (Table 4). In the female cohort, adipose tissue DNA methylation of three CpG sites was, according to the casual mediation analysis, found to partially mediate the effect of circulating triglycerides on HOMA-IR, namely, cg26583088 in KCNH4, cg17953300 in SIPA1 and cg00409636 in FAM176A (Table 4).

Schematic representation of the causal mediation analysis to test the potential causative effect of subcutaneous adipose tissue DNA methylation between circulating triglycerides and insulin resistance. The causal mediation analysis separates the total effect of triglyceride levels (independent variable) on insulin resistance (HOMA-IR; dependent variable) into (i) ACME, the mediator of interest in this case DNA methylation, and (ii) ADE, i.e. the effect acting directly or via a mediator other than what is under study. Secondary, we tested the mediating effect of mRNA expression between triglycerides and HOMA-IR, and if DNA methylation may mediate the effect of triglycerides on adipose tissue mRNA expression levels.

Table 4.

Model-based causal mediation analyses displaying adipose tissue DNA methylation and mRNA expression as a partial mediator of the effect of circulating triglyceride levels (independent variable) on insulin resistance (HOMA-IR, dependent variable)

Mediating CpG site	Cohort	ACME est (95% CI)	P-value	Total effect (95% CI)	Proportion mediated by CpG (95% CI)
cg23561371	Male	−0.20 (−0.34; −0.05)	0.004	0.23 (0.03; 0.57)	−0.89 (−5.69; −0.10)
cg14626525 (MIR602)	Male	−0.19 (−0.36; −0.03)	0.006	0.23 (0.03; 0.57)	−0.83 (−5.28; −0.05)
cg14984878 (SPTBN1)	Male	−0.2 (−0.46; −0.02)	0.008	0.23 (0.03; 0.57)	−0.87 (−6.47; −0.04)
cg06038180 (TMEM95)	Male	−0.14 (−0.3; −0.02)	0.016	0.23 (0.03; 0.57)	−0.62 (−3.65; −0.05)
cg09429345 (FAIM)	Male	−0.16 (−0.34; −0.03)	0.022	0.23 (0.03; 0.57)	−0.72 (−3.73; −0.06)
cg26548134 (SLC25A17)	Male	−0.15 (−0.33; −0.01)	0.028	0.23 (0.03; 0.57)	−0.66 (−3.55; −0.02)
cg17760881	Male	−0.12 (−0.28; −0.01)	0.030	0.23 (0.03; 0.57)	−0.53 (−3.58; −0.01)
cg14940420 (IQCF2)	Male	−0.13 (−0.24; 0.00)	0.046	0.23 (0.03; 0.57)	−0.57 (−3.28; 0.05)
cg26583088 (KCNH4)	Female	0.21 (0.03; 0.47)	0.024	0.89 (0.45; 1.43)	0.24 (0.04; 0.58)
cg17953300 (SIPA1)	Female	0.26 (0.03; 0.55)	0.028	0.89 (0.45; 1.43)	0.29 (0.03; 0.72)
cg00409636 (FAM176A)	Female	−0.31 (−0.71; −0.02)	0.032	0.89 (0.45; 1.43)	−0.35 (−0.97; −0.03)
Mediating mRNA	Cohort	ACME est (95% CI)	P-value	Total effect (95% CI)	Proportion mediated by mRNA (95% CI)
ANXA5	Male	0.079 (0.012; 0.186)	0.020	0.22 (0.03; 0.58)	0.35 (0.03; 2.09)
CTTNBP2	Male	0.086 (0.002; 0.207)	0.044	0.22 (0.03; 0.58)	0.38 (−0.03; 1.79)

Open in a new tab

Models adjusted for age, BMI and sub-cohort. ACME est: average causal mediation effect, estimate of mediator.

Since the causal effect of DNA methylation on mRNA levels is debated, we also tested if adipose tissue DNA methylation may mediate the effect of circulating triglycerides on mRNA levels. Amongst the 18 genes where DNA methylation and mRNA expression individually associated with triglyceride levels in the male cohort, three CpG sites were suggested as mediators, i.e. cg09880291 and cg05835726 as partial causal mediators of HOXA5 and cg24141863 of WDR17 mRNA expression (Supplementary Material, Table S3). In the female cohort, cg18874907 mediated a part of the effect circulating triglycerides exerts on adipose tissue MYO9B expression (Supplementary Material, Table S3).

Circulating LDL, HDL and total cholesterol levels are associated with altered DNA methylation in adipose tissue

We finally studied the association between circulating LDL, HDL and total cholesterol levels and DNA methylation in adipose tissue of the male and female cohorts. Of note, one needs to take into account that the different lipid measurements are not independent variables. Nevertheless, in the male cohort, we found that LDL, HDL and total cholesterol levels were associated with DNA methylation of 34, 15 and 5 individual CpG sites, respectively, based on the same linear model as the main analysis (Bonferroni corrected P-values < 0.05; Supplementary Material, Table S4). After adjustment for cellular composition (19) and multiple testing (Bonferroni correction), 15, 14 and 5 of these associations were still identified (Supplementary Material, Table S4). Moreover, some CpG sites showed association with more than one of the lipid traits (Fig. 5A). DNA methylation of a CpG site in the gene body of NEUROG3 showed a positive association with triglycerides, and a negative association with HDL, and sites in HRNBP3 and MEF2B showed positive associations with both LDL and total cholesterol (Supplementary Material, Table S4 and Fig. 5A). Together, the associations between circulating LDL, HDL and total cholesterol levels and DNA methylation in adipose tissue support that there may be metabolic cross-talk between the liver and adipose tissue, affecting the epigenome (22). Furthermore, in the female cohort, we found no associations between LDL, HDL or total cholesterol and adipose tissue DNA methylation, based on q < 0.05 and the same linear model as the main analysis.

(A) Venn diagram showing the overlap of associations between adipose tissue DNA methylation and the different lipid traits studied. The figure visualizes the number of CpG sites (n) associated with one or more lipid traits, i.e. circulating triglycerides, LDL, HDL or total cholesterol, in the male cohort. Number of associations are based on Bonferroni corrected P-values < 0.05. Gene names are based on Illumina annotation; for triglycerides, genes are shown for the 50 most significant associations. (B) Proposed model for the impact of circulating triglyceride levels on the adipose tissue epigenome and potential functional outcome. Triglycerides are associated with adipose tissue DNA methylation of several genes, and in some cases also mRNA expression of the corresponding gene. Our results point to genes involved in insulin resistance, lipid metabolism and inflammation, but also less well-known genes and intergenic methylation sites were identified. Causal mediation analyses further connected triglyceride levels to insulin resistance through differential DNA methylation of several genes in adipose tissue.

Discussion

Our current study presents associations between circulating triglycerides and the adipose tissue epigenome. These findings can be extended to differential gene expression for several genes and causal mediation of triglycerides on insulin resistance. Hence, the results from this study suggest mechanisms for how altered lipid levels in circulation potentially may affect the adipose tissue function, linked to peripheral insulin resistance and metabolic diseases (Fig. 5B).

We focus our study on the adipose tissue, an endocrine organ with a central function for metabolic diseases. Adipose tissue biology differs between men and women, because of body fat distribution, sex hormones and genetic factors, hence sex needs to be considered when studying the mechanisms behind metabolic dysfunction (22,23). Studying men and women separately will likely be important for future precision medicine. Looking at the overlap of our epigenetic findings in the male versus female cohort, no individual CpG site was found associated with triglycerides in both cohorts, and after the stringent Bonferroni correction in the male cohort, the gene body of CADM2 was the only region displaying altered DNA methylation in both cohorts, however, at CpG sites 169 kb apart. The lack of overlap between the two cohorts may be because of sex-specific effects of triglycerides on the adipose tissue epigenome but could also be because of different study design and choice of statistical analysis; optimally the male and female cohorts should be sampled from the same population.

When FDR was used to adjust for multiple testing in the male cohort, we identified a large number of methylation sites which may be the result of an inflated type I error rate, also supported by a lambda of 2.2. It should, however, be noted that lambda is not suitable to measure inflation in epigenome-wide associations studies (24). We mitigated the potential issue of statistical inflation by performing Bonferroni correction in the male cohort. By associating circulating triglycerides with adipose tissue DNA methylation in the male cohort, we detected 115 CpG sites, with the most significant annotated to RSF1, ANKS1A/TAF11 and RPH3AL. RSF1 encodes a chromatin-remodelling factor that has been linked to DNA damage repair and it controls p53-mediated transcription (25). Our data suggest that the high levels of triglycerides decrease promoter methylation of RSF1, which may affect chromatin structure and genomic stability. cg17149859 is located within 1500 bp from transcription start site of both ANKS1A and TAF11 (transcribed on opposite DNA strands), hence the main downstream target could be either or even both. A SNP in ANKS1A is associated with cardiovascular disease (26) and familial hypercholesterolemia (27), hence a link between ANKS1A and lipid levels is already established (Fig. 5B). Interesting, RPH3AL (also known as NOC2) is involved in insulin-dependent translocation of GLUT4 in adipocytes (28). Our results of triglyceride-association with RPH3AL DNA methylation hence suggest a mechanism for how circulating triglycerides may influence adipose tissue metabolism, although follow-up studies are needed to demonstrate this (Fig. 5B). In the female cohort, associations between triglyceride levels and adipose tissue DNA methylation were highlighted for PTPRN2 and COL6A3. Altered DNA methylation of PTPRN2 in human blood and pancreatic islets has been linked to type 2 diabetes (29), and PTPRN2 methylation was also associated with coronary heart disease (30).

For several genes, we discovered that triglycerides were associated with both DNA methylation and expression, e.g. IL11, FAM45B, HOXA5 and ANXA5 in the male cohort, and MYO9B, PPM1A and RRAS in the female cohort. In an adipocyte cell line, IL11 inhibited adipogenesis and lipoprotein lipase activity (31). Differential methylation of FAM45B has previously been described in Parkinson’s disease and keloid scar formation (32,33), but its biological role is not clear. Another gene where triglycerides associated with both adipose tissue methylation and expression was ANXA5, which has been suggested to affect triglyceride metabolism and obesity (34). Moreover, HOXA5 expression is upregulated in adipose tissue after weight loss (35). This is in agreement with our result showing a negative association between triglyceride levels and HOXA5 expression, as weight loss in response to bariatric surgery is likely to reduce circulating triglycerides. Also in agreement, mice fed a high fat diet responded with decreased HOXA5 expression, and altered DNA methylation, in adipose tissue (36). Our causal mediation analysis further suggested methylation of two HOXA5 CpG sites to mediate the effect of circulating triglycerides on expression of this gene (Fig. 5B).

With regards to the female cohort, epigenetic alteration of MYO9B has to our knowledge not previously been described. Nevertheless, MYO9B has rho-GTPase activity and is involved in cell signalling and motility, and has been associated with autoimmune and intestinal bowel diseases (37). In our study, PPM1A was also differentially methylated and displayed reduced expression with increasing triglycerides. The protein phosphatase PPM1A regulates PPARG through dephosphorylation, and its expression was decreased in mouse models of insulin resistance (38). Thereby, our study adds a link of how high triglyceride levels may decrease this phosphatase, leading to reduced PPARG activity and risk of metabolic disorders (Fig. 5B). In addition, we found triglyceride levels to associate with decreased promoter methylation and increased expression of RRAS, encoding a small GTPase. A genetic variant in RRAS was associated with hypertension (39), and in breast cancer cell lines, RRAS activity suppresses insulin and IGF-1 signalling (40).

Causal mediation analyses are used to find mechanisms that may transmit effects between the exposure and the outcome. In our case, we investigated possible mediators between circulating triglycerides and insulin resistance. Indeed, methylation levels of several sites were suggested to partially mediate this effect, e.g. in SPTBN1 in the male and SIPA1 in the female cohort (Fig. 5B). SPTBN1 was reported to be epigenetically silenced in Beckwith–Wiedemann syndrome (41). SIPA1 methylation in blood was linked to smoking, in a study connecting smoking to risk of coronary artery disease (42).

As triglyceride levels are strongly associated with non-alcoholic fatty liver disease (43), we also compared our results in adipose tissue to epigenetic alterations observed in liver from non-alcoholic steatohepatitis (NASH) patients of the Kuopio Obesity Bariatric Study (44). Seven of the sites where DNA methylation in adipose tissue was associated with triglycerides were also associated with NASH in human liver. In adipose tissue from the male cohort, three sites, annotated to ZNF83, WDR17 and DIP2C, overlapped, whereas four sites from the female cohort overlapped the methylation changes observed in NASH liver. Mutations in ZNF83 were recently associated with cancer, and found to increase NF-κB activation (45). Altered activation of NF-κB could contribute to the inflammatory process observed in both metabolic disease and NASH, as it regulates transcription of genes affecting the immune system (Fig. 5B).

Previous genome-wide studies have mainly investigated the link between circulating lipid levels and DNA methylation pattern in blood (8–13). Although important, because of the short life span of nucleated blood cells, the epigenome of these cells is less likely to have a regulatory role on whole-body metabolism, than the epigenome of long-lived adipocytes. However, there may still be similarities between DNA methylation in adipose tissue and blood. Potentially, alterations in the blood epigenome may mimic the epigenome of other tissues as suggested before (6,15,46,47), or methylation in both adipose tissue and blood could be a reflection of the metabolic state, such as plasma lipid levels, that influences the activity of epigenetic enzymes and thereby the epigenome (48,49). Indeed, triglyceride levels were associated with DNA methylation of a site annotated to SARS in both blood (12) and in adipose tissue in our male cohort when applying q < 0.05. The life span of cells is important to consider when directing drugs for one tissue, where future drugs directed to the epigenome of blood cells may require daily administration and the risk of broad side effects, whereas drugs directed to adipocytes may be administrated by few injections, likely causing less side effects.

A study by Castellano-Castillo et al. (50) found a relation between DNA methylation in visceral adipose tissue and the metabolic syndrome. They investigated genes related to adipogenesis, lipid metabolism and inflammation, and found that methylation of LPL and TNF was associated with metabolic syndrome and triglyceride levels. These genes were not detected in our study; however, the functions of subcutaneous versus visceral adipose tissue differ, the study design was different (candidate gene versus genome-wide methylation analysis) and the very stringent correction for multiple testing in our study may hide minor differences compared with studies based on nominal P-values. Another study associated subcutaneous adipose tissue DNA methylation to the amount of visceral fat and identified 1181 CpG sites, which were further related to blood lipids (51). Methylation of almost half of these sites was also associated with triglycerides and 10% with HDL levels. Amongst them, one intergenic site (cg16472834) was also identified in our study, with adipose tissue DNA methylation associated with triglycerides in the female cohort. Also, Keller et al. (52) studied the relationship between adipose tissue DNA methylation and blood lipids; however, this analysis was restricted to 24 genes, and no significant associations were observed. For the future, whole-genome bisulfite sequencing of adipose tissue will provide a more detailed map of the methylome and allow for detection of differentially methylated regions associated with metabolic traits like circulating lipids.

In conclusion, our study demonstrates novel associations between circulating triglycerides and the adipose tissue epigenome in males and females, respectively, thereby proposing a model where epigenetic factors mediate external influences on cell function and eventually phenotypes related to metabolic diseases such as insulin resistance.

Materials and Methods

Study participants

The male cohort consists of 85 Swedish and Danish men without any known disease, from three different studies as previously described (5,53,54). In these studies, a criterion for inclusion was to be without medication. The subcutaneous adipose tissue biopsies, anthropometrics and blood samples included in the present study were collected in the fasting state in the control arm of each study, before initiation of any intervention. Subcutaneous adipose tissue biopsies were abdominal (n = 57 (53,54)) or taken from the thigh (n = 28 (5)) under local anaesthesia (Lidocaine or Xylocain, AstraZeneca, Cambridge, UK) using a Bergström needle (Stille AB, Sweden). The female cohort consists of 93 Swedish women, part of a PCOS case control cohort as previously described (55,56). To avoid the interference of medication, the female participants were off medication for at least three months prior to baseline measurements. All individuals included in the present study had available blood lipid measurements (total cholesterol, HDL, LDL and triglycerides) as well as abdominal subcutaneous adipose tissue biopsies taken after an overnight fast under local anaesthesia (Table 1). The research protocols have been accepted by local human research ethics authorities, and all participants provided written informed consent. The investigation was performed in accordance with the Declaration of Helsinki. All biopsies were snap frozen in liquid nitrogen and stored at −80°C.

DNA methylation analysis

Bisulfite conversion of 500 ng genomic DNA extracted from the adipose tissue biopsies was performed using EZ DNA methylation Kit (Zymo Research, Irvine, CA, USA). Genome-wide DNA methylation was analysed using the Infinium HumanMethylation450 BeadChip assay (Illumina, San Diego, CA, USA), with details as previously described (6,55). For each probe, raw methylation scores, represented as methylation β-values, was calculated using GenomeStudio Methylation module software (Illumina). β-values were then converted to M-values (M = log²(β/(1 − β))), a more statistically valid method for differential analysis of methylation levels, hence further association analyses are based on the M-value (57). After quality control (QC), removal of probes on the Y chromosome (female cohort only), with SNPs in the target CpG (58), reported to be cross-reactive (59) or with an average detection of P-value > 0.01, followed by quantile normalization using BMIQ (60) and batch correction using ComBat (61), DNA methylation data for 448 785 CpG sites in the male and 453 039 CpG sites in the female cohort were included in subsequent analyses. To facilitate interpretation, after statistical analyses, data were reconverted to β-values, which are presented in tables and figures.

mRNA expression analysis

RNA was extracted from adipose tissue of the same individuals as included in the DNA methylation analyses. For the male cohort, expression data were generated using the GeneChip Human Gene 1.0 ST array (Affymetrix, Santa Clara, CA, USA). After QC, we successfully obtained expression data covering 26 720 transcripts in 83 male samples. Global expression data for all samples from the female cohort were generated using HumanHT-12 v4 BeadChip (Illumina). Here, data for 14 168 transcripts were included after QC. The complete RNA extraction, QC and analysis pipelines have been described in detail before (6,55).

Statistical analysis

For the male cohort, associations between DNA methylation and the studied phenotypes were analysed using a random effect mixed model, where the cohort was considered the random effect variable and age, BMI and triglycerides (or HDL, LDL or total cholesterol levels) were included as fixed factors. This model was selected as the age span from the different sub-cohorts did not overlap and the arrays were analysed at different occasions. ComBat was further applied to each sub-cohort separately to handle batch effects. The same model was applied to analyse the mRNA expression data. The female cohort was analysed using linear regression including triglycerides, PCOS status (case or control), age and BMI in the model. We applied an FDR analysis to account for multiple testing (62). For the analysis of DNA methylation data in the male cohort, we also performed the more stringent Bonferroni correction to account for potential inflation and reduce number of false positives. Adjustment for cellular composition was done using the reference-free method from Houseman et al. (19). All P-values are two tailed. Genome positions are based on genome build 37/hg19.

Causal mediation analysis

Non-parametric causal mediation analyses were performed using the mediation R package (R version 4.1.1) and default settings (21), to test the potential causative effect of subcutaneous adipose tissue DNA methylation (mediator of interest) between circulating triglycerides (independent variable) and insulin resistance (HOMA-IR, dependent variable). Normalized CpG methylation β-value was modelled as the mediator, including age, BMI and sub-cohort (male cohort) or PCOS status (female cohort) as co-variates. Secondary, we also tested the mediating effect of mRNA expression between triglycerides and HOMA-IR, and if DNA methylation may partially mediate the effect of triglycerides on adipose tissue mRNA expression levels, including the same co-variates as above.

DNA methylation datasets generated for this study and the data code used to generate the results are available from the corresponding author on request.

Supplementary Material

Supplementary_table_1_ddad024

Click here for additional data file.^{(4.4MB, zip)}

Supplementary_table_2_ddad024

Click here for additional data file.^{(47.4KB, xlsx)}

Supplementary_table_3_ddad024

Click here for additional data file.^{(9.5KB, xlsx)}

Supplementary_table_4_ddad024

Click here for additional data file.^{(3.5MB, zip)}

SUPPLEMENTARY_Legends_ddad024

Click here for additional data file.^{(13.1KB, docx)}

Acknowledgements

We thank the participating volunteers as well as clinicians and laboratory technicians, including Leif Groop, Targ Elgzyri and Ylva Wessman at Scania University Hospital, Malmö, Sweden, and Marianne Modest at Steno Diabetes Center, Denmark. DNA methylation and mRNA expression arrays were analysed with help from Swegene Center for Integrative Biology at Lund University (SCIBLU).

Conflict of Interest statement:None declared.

Contributor Information

Tina Rönn, Epigenetics and Diabetes Unit, Department of Clinical Sciences, Lund University Diabetes Centre, Lund University, Scania University Hospital, 205 02 Malmö, Sweden.

Alexander Perfilyev, Epigenetics and Diabetes Unit, Department of Clinical Sciences, Lund University Diabetes Centre, Lund University, Scania University Hospital, 205 02 Malmö, Sweden.

Josefine Jönsson, Epigenetics and Diabetes Unit, Department of Clinical Sciences, Lund University Diabetes Centre, Lund University, Scania University Hospital, 205 02 Malmö, Sweden.

Karl-Fredrik Eriksson, Department of Clinical Sciences, Vascular Diseases, Lund University, 205 02 Malmö, Sweden.

Sine W Jørgensen, Department of Endocrinology, Rigshospitalet, Copenhagen, Denmark.

Charlotte Brøns, Department of Endocrinology, Rigshospitalet, Copenhagen, Denmark.

Linn Gillberg, Department of Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark.

Allan Vaag, Steno Diabetes Center Copenhagen, DK-2820, Gentofte, Denmark.

Elisabet Stener-Victorin, Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden.

Charlotte Ling, Epigenetics and Diabetes Unit, Department of Clinical Sciences, Lund University Diabetes Centre, Lund University, Scania University Hospital, 205 02 Malmö, Sweden.

Funding

Swedish Research Council (Grants Dnr 2016-02486, 2018-02567, 2019-01406, 2014-2775, 2018-02435); Region Skåne ALF; Strategic Research Area Exodiab (Dnr 2009-1039); the Novo Nordisk Foundation; the Swedish Foundation for Strategic Research (Dnr IRC15-0067); Syskonen Svensson Foundation; the Diabetes Foundation; Kungliga Fysiografiska Sällskapet i Lund; Magnus Bergvall Foundation; Åke Wiberg Foundation; the Påhlsson Foundation.

Data availability

DNA methylation datasets generated for this study and the data code used to generate the results are available from the corresponding author on request.

References

1. American Heart Association Nutrition Committee, Lichtenstein, A.H., Appel, L.J., Brands, M., Carnethon, M., Daniels, S., Franch, H.A., Franklin, B., Kris-Etherton, P., Harris, W.S.et al. (2006) Diet and lifestyle recommendations revision 2006: a scientific statement from the American Heart Association Nutrition Committee. Circulation, 114, 82–96. [DOI] [PubMed] [Google Scholar]
2. Elder, S.J., Lichtenstein, A.H., Pittas, A.G., Roberts, S.B., Fuss, P.J., Greenberg, A.S., McCrory, M.A., Bouchard, T.J., Jr., Saltzman, E. and Neale, M.C. (2009) Genetic and environmental influences on factors associated with cardiovascular disease and the metabolic syndrome. J. Lipid Res., 50, 1917–1926. [DOI] [PMC free article] [PubMed] [Google Scholar]
3. Global Lipids Genetics Consortium, Willer, C.J., Schmidt, E.M., Sengupta, S., Peloso, G.M., Gustafsson, S., Kanoni, S., Ganna, A., Chen, J., Buchkovich, M.L.et al. (2013) Discovery and refinement of loci associated with lipid levels. Nat. Genet., 45, 1274–1283. [DOI] [PMC free article] [PubMed] [Google Scholar]
4. Ling, C. and Rönn, T. (2019) Epigenetics in human obesity and type 2 diabetes. Cell Metab., 29, 1028–1044. [DOI] [PMC free article] [PubMed] [Google Scholar]
5. Rönn, T., Volkov, P., Davegardh, C., Dayeh, T., Hall, E., Olsson, A.H., Nilsson, E., Tornberg, A., Dekker Nitert, M., Eriksson, K.F.et al. (2013) A six months exercise intervention influences the genome-wide DNA methylation pattern in human adipose tissue. PLoS Genet., 9, e1003572. [DOI] [PMC free article] [PubMed] [Google Scholar]
6. Rönn, T., Volkov, P., Gillberg, L., Kokosar, M., Perfilyev, A., Jacobsen, A.L., Jorgensen, S.W., Brons, C., Jansson, P.A., Eriksson, K.F.et al. (2015) Impact of age, BMI and HbA1c levels on the genome-wide DNA methylation and mRNA expression patterns in human adipose tissue and identification of epigenetic biomarkers in blood. Hum. Mol. Genet., 24, 3792–3813. [DOI] [PubMed] [Google Scholar]
7. Volkov, P., Olsson, A.H., Gillberg, L., Jorgensen, S.W., Brons, C., Eriksson, K.F., Groop, L., Jansson, P.A., Nilsson, E., Ronn, T.et al. (2016) A genome-wide mQTL analysis in human adipose tissue identifies genetic variants associated with DNA methylation, gene expression and metabolic traits. PLoS One, 11, e0157776. [DOI] [PMC free article] [PubMed] [Google Scholar]
8. Braun, K.V.E., Dhana, K., deVries, P.S., Voortman, T., vanMeurs, J.B.J., Uitterlinden, A.G., Consortium, B., Hofman, A., Hu, F.B., Franco, O.H.et al. (2017) Epigenome-wide association study (EWAS) on lipids: the Rotterdam study. Clin. Epigenetics, 9, 15. [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Dekkers, K.F., vanIterson, M., Slieker, R.C., Moed, M.H., Bonder, M.J., vanGalen, M., Mei, H., Zhernakova, D.V., van denBerg, L.H., Deelen, J.et al. (2016) Blood lipids influence DNA methylation in circulating cells. Genome Biol., 17, 138. [DOI] [PMC free article] [PubMed] [Google Scholar]
10. Gomez-Alonso, M.D.C., Kretschmer, A., Wilson, R., Pfeiffer, L., Karhunen, V., Seppala, I., Zhang, W., Mittelstrass, K., Wahl, S., Matias-Garcia, P.R.et al. (2021) DNA methylation and lipid metabolism: an EWAS of 226 metabolic measures. Clin. Epigenetics, 13, 7. [DOI] [PMC free article] [PubMed] [Google Scholar]
11. Hedman, A.K., Mendelson, M.M., Marioni, R.E., Gustafsson, S., Joehanes, R., Irvin, M.R., Zhi, D., Sandling, J.K., Yao, C., Liu, C.et al. (2017) Epigenetic patterns in blood associated with lipid traits predict incident coronary heart disease events and are enriched for results from genome-wide association studies. Circ. Cardiovasc. Genet., 10, e001487. [DOI] [PMC free article] [PubMed] [Google Scholar]
12. Jhun, M.A., Mendelson, M., Wilson, R., Gondalia, R., Joehanes, R., Salfati, E., Zhao, X., Braun, K.V.E., Do, A.N., Hedman, A.K.et al. (2021) A multi-ethnic epigenome-wide association study of leukocyte DNA methylation and blood lipids. Nat. Commun., 12, 3987. [DOI] [PMC free article] [PubMed] [Google Scholar]
13. Pfeiffer, L., Wahl, S., Pilling, L.C., Reischl, E., Sandling, J.K., Kunze, S., Holdt, L.M., Kretschmer, A., Schramm, K., Adamski, J.et al. (2015) DNA methylation of lipid-related genes affects blood lipid levels. Circ. Cardiovasc. Genet., 8, 334–342. [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Spalding, K.L., Arner, E., Westermark, P.O., Bernard, S., Buchholz, B.A., Bergmann, O., Blomqvist, L., Hoffstedt, J., Naslund, E., Britton, T.et al. (2008) Dynamics of fat cell turnover in humans. Nature, 453, 783–787. [DOI] [PubMed] [Google Scholar]
15. Garcia-Calzon, S., Perfilyev, A., Martinell, M., Ustinova, M., Kalamajski, S., Franks, P.W., Bacos, K., Elbere, I., Pihlajamaki, J., Volkov, P.et al. (2020) Epigenetic markers associated with metformin response and intolerance in drug-naive patients with type 2 diabetes. Sci. Transl. Med., 12, eaaz1803. [DOI] [PubMed] [Google Scholar]
16. Sugiyama, M.G. and Agellon, L.B. (2012) Sex differences in lipid metabolism and metabolic disease risk. Biochem. Cell Biol., 90, 124–141. [DOI] [PubMed] [Google Scholar]
17. Garcia-Calzon, S., Perfilyev, A., deMello, V.D., Pihlajamaki, J. and Ling, C. (2018) Sex differences in the methylome and transcriptome of the human liver and circulating HDL-cholesterol levels. J. Clin. Endocrinol. Metab., 103, 4395–4408. [DOI] [PMC free article] [PubMed] [Google Scholar]
18. Hall, E., Volkov, P., Dayeh, T., Esguerra, J.L., Salo, S., Eliasson, L., Rönn, T., Bacos, K. and Ling, C. (2014) Sex differences in the genome-wide DNA methylation pattern and impact on gene expression, microRNA levels and insulin secretion in human pancreatic islets. Genome Biol., 15, 522. [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Houseman, E.A., Kelsey, K.T., Wiencke, J.K. and Marsit, C.J. (2015) Cell-composition effects in the analysis of DNA methylation array data: a mathematical perspective. BMC Bioinformatics, 16, 95. [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Semnani-Azad, Z., Connelly, P.W., Bazinet, R.P., Retnakaran, R., Jenkins, D.J.A., Harris, S.B., Zinman, B. and Hanley, A.J. (2021) Adipose tissue insulin resistance is longitudinally associated with adipose tissue dysfunction, circulating lipids, and dysglycemia: the PROMISE cohort. Diabetes Care, 44, 1682–1691. [DOI] [PubMed] [Google Scholar]
21. Tingley, D., Yamamoto, T., Hirose, K., Keele, L. and Imai, K. (2014) Mediation: R package for causal mediation analysis. J. Stat. Softw., 59. [Google Scholar]
22. Goossens, G.H., Jocken, J.W.E. and Blaak, E.E. (2021) Sexual dimorphism in cardiometabolic health: the role of adipose tissue, muscle and liver. Nat. Rev. Endocrinol., 17, 47–66. [DOI] [PubMed] [Google Scholar]
23. Chang, E., Varghese, M. and Singer, K. (2018) Gender and sex differences in adipose tissue. Curr. Diab. Rep., 18, 69. [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Van Iterson, M., Van Zwet, E.W., Consortium, B. and Heijmans, B.T. (2017) Controlling bias and inflation in epigenome- and transcriptome-wide association studies using the empirical null distribution. Genome Biol., 18, 19. [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Min, S., Kim, K., Kim, S.G., Cho, H. and Lee, Y. (2018) Chromatin-remodeling factor, RSF1, controls p53-mediated transcription in apoptosis upon DNA strand breaks. Cell Death Dis., 9, 1079. [DOI] [PMC free article] [PubMed] [Google Scholar]
26. Charmet, R., Duffy, S., Keshavarzi, S., Gyorgy, B., Marre, M., Rossing, P., McKnight, A.J., Maxwell, A.P., Ahluwalia, T.V.S., Paterson, A.D.et al. (2018) Novel risk genes identified in a genome-wide association study for coronary artery disease in patients with type 1 diabetes. Cardiovasc. Diabetol., 17, 61. [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Fantino, M., Paquette, M., Bernard, S. and Baass, A. (2021) ANKS1A genotype predicts cardiovascular events in patients with familial hypercholesterolemia. J. Clin. Lipidol., 15, 602–607. [DOI] [PubMed] [Google Scholar]
28. Koumanov, F., Pereira, V.J., Richardson, J.D., Sargent, S.L., Fazakerley, D.J. and Holman, G.D. (2015) Insulin regulates Rab3-Noc2 complex dissociation to promote GLUT4 translocation in rat adipocytes. Diabetologia, 58, 1877–1886. [DOI] [PMC free article] [PubMed] [Google Scholar]
29. Ouni, M., Saussenthaler, S., Eichelmann, F., Jahnert, M., Stadion, M., Wittenbecher, C., Ronn, T., Zellner, L., Gottmann, P., Ling, C.et al. (2020) Epigenetic changes in islets of Langerhans preceding the onset of diabetes. Diabetes, 69, 2503–2517. [DOI] [PMC free article] [PubMed] [Google Scholar]
30. Agha, G., Mendelson, M.M., Ward-Caviness, C.K., Joehanes, R., Huan, T., Gondalia, R., Salfati, E., Brody, J.A., Fiorito, G., Bressler, J.et al. (2019) Blood leukocyte DNA methylation predicts risk of future myocardial infarction and coronary heart disease. Circulation, 140, 645–657. [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Yin, T., Miyazawa, K. and Yang, Y.C. (1992) Characterization of interleukin-11 receptor and protein tyrosine phosphorylation induced by interleukin-11 in mouse 3T3-L1 cells. J. Biol. Chem., 267, 8347–8351. [PubMed] [Google Scholar]
32. Alghamdi, M.A., Wallace, H.J., Melton, P.E., Moses, E.K., Stevenson, A., Al-Eitan, L.N., Rea, S., Duke, J.M., Danielsen, P.L., Prele, C.M.et al. (2020) Identification of differentially methylated CpG sites in fibroblasts from keloid scars. Biomedicine, 8, 181. [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Moore, K., McKnight, A.J., Craig, D. and O'Neill, F. (2014) Epigenome-wide association study for Parkinson’s disease. NeuroMolecular Med., 16, 845–855. [DOI] [PubMed] [Google Scholar]
34. Seok, H., Park, H.J., Lee, B.W., Kim, J.W., Jung, M., Lee, S.R., Park, K.H., Park, Y.G., Baik, H.H. and Chung, J.H. (2013) Association of annexin A5 polymorphisms with obesity. Biomed. Rep., 1, 654–658. [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Dankel, S.N., Fadnes, D.J., Stavrum, A.K., Stansberg, C., Holdhus, R., Hoang, T., Veum, V.L., Christensen, B.J., Vage, V., Sagen, J.V.et al. (2010) Switch from stress response to homeobox transcription factors in adipose tissue after profound fat loss. PLoS One, 5, e11033. [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Parrillo, L., Costa, V., Raciti, G.A., Longo, M., Spinelli, R., Esposito, R., Nigro, C., Vastolo, V., Desiderio, A., Zatterale, F.et al. (2016) Hoxa5 undergoes dynamic DNA methylation and transcriptional repression in the adipose tissue of mice exposed to high-fat diet. Int. J. Obes., 40, 929–937. [DOI] [PubMed] [Google Scholar]
37. Hanley, P.J., Vollmer, V. and Bahler, M. (2020) Class IX myosins: motorized RhoGAP signaling molecules. Adv. Exp. Med. Biol., 1239, 381–389. [DOI] [PubMed] [Google Scholar]
38. Khim, K.W., Choi, S.S., Jang, H.J., Lee, Y.H., Lee, E., Hyun, J.M., Eom, H.J., Yoon, S., Choi, J.W., Park, T.E.et al. (2020) PPM1A controls diabetic gene programming through directly dephosphorylating PPARgamma at Ser273. Cell, 9, 343. [DOI] [PMC free article] [PubMed] [Google Scholar]
39. Surendran, P., Drenos, F., Young, R., Warren, H., Cook, J.P., Manning, A.K., Grarup, N., Sim, X., Barnes, D.R., Witkowska, K.et al. (2016) Trans-ancestry meta-analyses identify rare and common variants associated with blood pressure and hypertension. Nat. Genet., 48, 1151–1161. [DOI] [PMC free article] [PubMed] [Google Scholar]
40. Yu, Y., Hao, Y. and Feig, L.A. (2006) The R-Ras GTPase mediates cross talk between estrogen and insulin signaling in breast cancer cells. Mol. Cell. Biol., 26, 6372–6380. [DOI] [PMC free article] [PubMed] [Google Scholar]
41. Yao, Z.X., Jogunoori, W., Choufani, S., Rashid, A., Blake, T., Yao, W., Kreishman, P., Amin, R., Sidawy, A.A., Evans, S.R.et al. (2010) Epigenetic silencing of beta-spectrin, a TGF-beta signaling/scaffolding protein in a human cancer stem cell disorder: Beckwith-Wiedemann syndrome. J. Biol. Chem., 285, 36112–36120. [DOI] [PMC free article] [PubMed] [Google Scholar]
42. Steenaard, R.V., Ligthart, S., Stolk, L., Peters, M.J., vanMeurs, J.B., Uitterlinden, A.G., Hofman, A., Franco, O.H. and Dehghan, A. (2015) Tobacco smoking is associated with methylation of genes related to coronary artery disease. Clin. Epigenetics, 7, 54. [DOI] [PMC free article] [PubMed] [Google Scholar]
43. Tomizawa, M., Kawanabe, Y., Shinozaki, F., Sato, S., Motoyoshi, Y., Sugiyama, T., Yamamoto, S. and Sueishi, M. (2014) Triglyceride is strongly associated with nonalcoholic fatty liver disease among markers of hyperlipidemia and diabetes. Biomed. Rep., 2, 633–636. [DOI] [PMC free article] [PubMed] [Google Scholar]
44. de Mello, V.D., Matte, A., Perfilyev, A., Mannisto, V., Ronn, T., Nilsson, E., Kakela, P., Ling, C. and Pihlajamaki, J. (2017) Human liver epigenetic alterations in non-alcoholic steatohepatitis are related to insulin action. Epigenetics, 12, 287–295. [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Lyu, Z.J., Wang, Y., Huang, J.L., Chen, M., Wu, S.Y., Yan, Q., Zhang, Y., Tang, Y., Jiang, C., Li, L.et al. (2021) Recurrent ZNF83-E293V mutation promotes bladder cancer progression through the NF-kappaB pathway via transcriptional dysregulation of S100A8. Mol. Ther., 29, 275–290. [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Bacos, K., Gillberg, L., Volkov, P., Olsson, A.H., Hansen, T., Pedersen, O., Gjesing, A.P., Eiberg, H., Tuomi, T., Almgren, P.et al. (2016) Blood-based biomarkers of age-associated epigenetic changes in human islets associate with insulin secretion and diabetes. Nat. Commun., 7, 11089. [DOI] [PMC free article] [PubMed] [Google Scholar]
47. Dayeh, T., Tuomi, T., Almgren, P., Perfilyev, A., Jansson, P.A., deMello, V.D., Pihlajamaki, J., Vaag, A., Groop, L., Nilsson, E.et al. (2016) DNA methylation of loci within ABCG1 and PHOSPHO1 in blood DNA is associated with future type 2 diabetes risk. Epigenetics, 11, 482–488. [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Davegardh, C., Sall, J., Benrick, A., Broholm, C., Volkov, P., Perfilyev, A., Henriksen, T.I., Wu, Y., Hjort, L., Brons, C.et al. (2021) VPS39-deficiency observed in type 2 diabetes impairs muscle stem cell differentiation via altered autophagy and epigenetics. Nat. Commun., 12, 2431. [DOI] [PMC free article] [PubMed] [Google Scholar]
49. Malmgren, S., Spegel, P., Danielsson, A.P., Nagorny, C.L., Andersson, L.E., Nitert, M.D., Ridderstrale, M., Mulder, H. and Ling, C. (2013) Coordinate changes in histone modifications, mRNA levels, and metabolite profiles in clonal INS-1 832/13 beta-cells accompany functional adaptations to lipotoxicity. J. Biol. Chem., 288, 11973–11987. [DOI] [PMC free article] [PubMed] [Google Scholar]
50. Castellano-Castillo, D., Moreno-Indias, I., Sanchez-Alcoholado, L., Ramos-Molina, B., Alcaide-Torres, J., Morcillo, S., Ocana-Wilhelmi, L., Tinahones, F., Queipo-Ortuno, M.I. and Cardona, F. (2019) Altered adipose tissue DNA methylation status in metabolic syndrome: relationships between global DNA methylation and specific methylation at adipogenic, lipid metabolism and inflammatory candidate genes and metabolic variables. J. Clin. Med., 8, 87. [DOI] [PMC free article] [PubMed] [Google Scholar]
51. Christiansen, C., Tomlinson, M., Eliot, M., Nilsson, E., Costeira, R., Xia, Y., Villicana, S., Mompeo, O., Wells, P., Castillo-Fernandez, J.et al. (2022) Adipose methylome integrative-omic analyses reveal genetic and dietary metabolic health drivers and insulin resistance classifiers. Genome Med., 14, 75. [DOI] [PMC free article] [PubMed] [Google Scholar]
52. Keller, M., Hopp, L., Liu, X., Wohland, T., Rohde, K., Cancello, R., Klos, M., Bacos, K., Kern, M., Eichelmann, F.et al. (2017) Genome-wide DNA promoter methylation and transcriptome analysis in human adipose tissue unravels novel candidate genes for obesity. Mol. Metab., 6, 86–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
53. Gillberg, L., Perfilyev, A., Brons, C., Thomasen, M., Grunnet, L.G., Volkov, P., Rosqvist, F., Iggman, D., Dahlman, I., Riserus, U.et al. (2016) Adipose tissue transcriptomics and epigenomics in low birthweight men and controls: role of high-fat overfeeding. Diabetologia, 59, 799–812. [DOI] [PubMed] [Google Scholar]
54. Gillberg, L., Ronn, T., Jorgensen, S.W., Perfilyev, A., Hjort, L., Nilsson, E., Brons, C., Vaag, A. and Ling, C. (2019) Fasting unmasks differential fat and muscle transcriptional regulation of metabolic gene sets in low versus normal birth weight men. EBioMedicine, 47, 341–351. [DOI] [PMC free article] [PubMed] [Google Scholar]
55. Kokosar, M., Benrick, A., Perfilyev, A., Fornes, R., Nilsson, E., Maliqueo, M., Behre, C.J., Sazonova, A., Ohlsson, C., Ling, C.et al. (2016) Epigenetic and transcriptional alterations in human adipose tissue of polycystic ovary syndrome. Sci. Rep., 6, 22883. [DOI] [PMC free article] [PubMed] [Google Scholar]
56. Stener-Victorin, E., Holm, G., Labrie, F., Nilsson, L., Janson, P.O. and Ohlsson, C. (2010) Are there any sensitive and specific sex steroid markers for polycystic ovary syndrome? J. Clin. Endocrinol. Metab., 95, 810–819. [DOI] [PubMed] [Google Scholar]
57. Du, P., Zhang, X., Huang, C.C., Jafari, N., Kibbe, W.A., Hou, L. and Lin, S.M. (2010) Comparison of beta-value and M-value methods for quantifying methylation levels by microarray analysis. BMC Bioinformatics, 11, 587. [DOI] [PMC free article] [PubMed] [Google Scholar]
58. Price, M.E., Cotton, A.M., Lam, L.L., Farre, P., Emberly, E., Brown, C.J., Robinson, W.P. and Kobor, M.S. (2013) Additional annotation enhances potential for biologically-relevant analysis of the Illumina Infinium HumanMethylation450 BeadChip array. Epigenetics Chromatin, 6, 4. [DOI] [PMC free article] [PubMed] [Google Scholar]
59. Chen, Y.A., Lemire, M., Choufani, S., Butcher, D.T., Grafodatskaya, D., Zanke, B.W., Gallinger, S., Hudson, T.J. and Weksberg, R. (2013) Discovery of cross-reactive probes and polymorphic CpGs in the Illumina Infinium HumanMethylation450 microarray. Epigenetics, 8, 203–209. [DOI] [PMC free article] [PubMed] [Google Scholar]
60. Teschendorff, A.E., Marabita, F., Lechner, M., Bartlett, T., Tegner, J., Gomez-Cabrero, D. and Beck, S. (2013) A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450 k DNA methylation data. Bioinformatics, 29, 189–196. [DOI] [PMC free article] [PubMed] [Google Scholar]
61. Johnson, W.E., Li, C. and Rabinovic, A. (2007) Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics, 8, 118–127. [DOI] [PubMed] [Google Scholar]
62. Benjamini, Y. and Hochberg, Y. (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B, 57, 289–300. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary_table_1_ddad024

Click here for additional data file.^{(4.4MB, zip)}

Supplementary_table_2_ddad024

Click here for additional data file.^{(47.4KB, xlsx)}

Supplementary_table_3_ddad024

Click here for additional data file.^{(9.5KB, xlsx)}

Supplementary_table_4_ddad024

Click here for additional data file.^{(3.5MB, zip)}

SUPPLEMENTARY_Legends_ddad024

Click here for additional data file.^{(13.1KB, docx)}

Data Availability Statement

DNA methylation datasets generated for this study and the data code used to generate the results are available from the corresponding author on request.

[ref1] 1. American Heart Association Nutrition Committee, Lichtenstein, A.H., Appel, L.J., Brands, M., Carnethon, M., Daniels, S., Franch, H.A., Franklin, B., Kris-Etherton, P., Harris, W.S.et al. (2006) Diet and lifestyle recommendations revision 2006: a scientific statement from the American Heart Association Nutrition Committee. Circulation, 114, 82–96. [DOI] [PubMed] [Google Scholar]

[ref2] 2. Elder, S.J., Lichtenstein, A.H., Pittas, A.G., Roberts, S.B., Fuss, P.J., Greenberg, A.S., McCrory, M.A., Bouchard, T.J., Jr., Saltzman, E. and Neale, M.C. (2009) Genetic and environmental influences on factors associated with cardiovascular disease and the metabolic syndrome. J. Lipid Res., 50, 1917–1926. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref3] 3. Global Lipids Genetics Consortium, Willer, C.J., Schmidt, E.M., Sengupta, S., Peloso, G.M., Gustafsson, S., Kanoni, S., Ganna, A., Chen, J., Buchkovich, M.L.et al. (2013) Discovery and refinement of loci associated with lipid levels. Nat. Genet., 45, 1274–1283. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref4] 4. Ling, C. and Rönn, T. (2019) Epigenetics in human obesity and type 2 diabetes. Cell Metab., 29, 1028–1044. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref5] 5. Rönn, T., Volkov, P., Davegardh, C., Dayeh, T., Hall, E., Olsson, A.H., Nilsson, E., Tornberg, A., Dekker Nitert, M., Eriksson, K.F.et al. (2013) A six months exercise intervention influences the genome-wide DNA methylation pattern in human adipose tissue. PLoS Genet., 9, e1003572. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref6] 6. Rönn, T., Volkov, P., Gillberg, L., Kokosar, M., Perfilyev, A., Jacobsen, A.L., Jorgensen, S.W., Brons, C., Jansson, P.A., Eriksson, K.F.et al. (2015) Impact of age, BMI and HbA1c levels on the genome-wide DNA methylation and mRNA expression patterns in human adipose tissue and identification of epigenetic biomarkers in blood. Hum. Mol. Genet., 24, 3792–3813. [DOI] [PubMed] [Google Scholar]

[ref7] 7. Volkov, P., Olsson, A.H., Gillberg, L., Jorgensen, S.W., Brons, C., Eriksson, K.F., Groop, L., Jansson, P.A., Nilsson, E., Ronn, T.et al. (2016) A genome-wide mQTL analysis in human adipose tissue identifies genetic variants associated with DNA methylation, gene expression and metabolic traits. PLoS One, 11, e0157776. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref8] 8. Braun, K.V.E., Dhana, K., deVries, P.S., Voortman, T., vanMeurs, J.B.J., Uitterlinden, A.G., Consortium, B., Hofman, A., Hu, F.B., Franco, O.H.et al. (2017) Epigenome-wide association study (EWAS) on lipids: the Rotterdam study. Clin. Epigenetics, 9, 15. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref9] 9. Dekkers, K.F., vanIterson, M., Slieker, R.C., Moed, M.H., Bonder, M.J., vanGalen, M., Mei, H., Zhernakova, D.V., van denBerg, L.H., Deelen, J.et al. (2016) Blood lipids influence DNA methylation in circulating cells. Genome Biol., 17, 138. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref10] 10. Gomez-Alonso, M.D.C., Kretschmer, A., Wilson, R., Pfeiffer, L., Karhunen, V., Seppala, I., Zhang, W., Mittelstrass, K., Wahl, S., Matias-Garcia, P.R.et al. (2021) DNA methylation and lipid metabolism: an EWAS of 226 metabolic measures. Clin. Epigenetics, 13, 7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref11] 11. Hedman, A.K., Mendelson, M.M., Marioni, R.E., Gustafsson, S., Joehanes, R., Irvin, M.R., Zhi, D., Sandling, J.K., Yao, C., Liu, C.et al. (2017) Epigenetic patterns in blood associated with lipid traits predict incident coronary heart disease events and are enriched for results from genome-wide association studies. Circ. Cardiovasc. Genet., 10, e001487. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref12] 12. Jhun, M.A., Mendelson, M., Wilson, R., Gondalia, R., Joehanes, R., Salfati, E., Zhao, X., Braun, K.V.E., Do, A.N., Hedman, A.K.et al. (2021) A multi-ethnic epigenome-wide association study of leukocyte DNA methylation and blood lipids. Nat. Commun., 12, 3987. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref13] 13. Pfeiffer, L., Wahl, S., Pilling, L.C., Reischl, E., Sandling, J.K., Kunze, S., Holdt, L.M., Kretschmer, A., Schramm, K., Adamski, J.et al. (2015) DNA methylation of lipid-related genes affects blood lipid levels. Circ. Cardiovasc. Genet., 8, 334–342. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref14] 14. Spalding, K.L., Arner, E., Westermark, P.O., Bernard, S., Buchholz, B.A., Bergmann, O., Blomqvist, L., Hoffstedt, J., Naslund, E., Britton, T.et al. (2008) Dynamics of fat cell turnover in humans. Nature, 453, 783–787. [DOI] [PubMed] [Google Scholar]

[ref15] 15. Garcia-Calzon, S., Perfilyev, A., Martinell, M., Ustinova, M., Kalamajski, S., Franks, P.W., Bacos, K., Elbere, I., Pihlajamaki, J., Volkov, P.et al. (2020) Epigenetic markers associated with metformin response and intolerance in drug-naive patients with type 2 diabetes. Sci. Transl. Med., 12, eaaz1803. [DOI] [PubMed] [Google Scholar]

[ref16] 16. Sugiyama, M.G. and Agellon, L.B. (2012) Sex differences in lipid metabolism and metabolic disease risk. Biochem. Cell Biol., 90, 124–141. [DOI] [PubMed] [Google Scholar]

[ref17] 17. Garcia-Calzon, S., Perfilyev, A., deMello, V.D., Pihlajamaki, J. and Ling, C. (2018) Sex differences in the methylome and transcriptome of the human liver and circulating HDL-cholesterol levels. J. Clin. Endocrinol. Metab., 103, 4395–4408. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref18] 18. Hall, E., Volkov, P., Dayeh, T., Esguerra, J.L., Salo, S., Eliasson, L., Rönn, T., Bacos, K. and Ling, C. (2014) Sex differences in the genome-wide DNA methylation pattern and impact on gene expression, microRNA levels and insulin secretion in human pancreatic islets. Genome Biol., 15, 522. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref19] 19. Houseman, E.A., Kelsey, K.T., Wiencke, J.K. and Marsit, C.J. (2015) Cell-composition effects in the analysis of DNA methylation array data: a mathematical perspective. BMC Bioinformatics, 16, 95. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref20] 20. Semnani-Azad, Z., Connelly, P.W., Bazinet, R.P., Retnakaran, R., Jenkins, D.J.A., Harris, S.B., Zinman, B. and Hanley, A.J. (2021) Adipose tissue insulin resistance is longitudinally associated with adipose tissue dysfunction, circulating lipids, and dysglycemia: the PROMISE cohort. Diabetes Care, 44, 1682–1691. [DOI] [PubMed] [Google Scholar]

[ref21] 21. Tingley, D., Yamamoto, T., Hirose, K., Keele, L. and Imai, K. (2014) Mediation: R package for causal mediation analysis. J. Stat. Softw., 59. [Google Scholar]

[ref22] 22. Goossens, G.H., Jocken, J.W.E. and Blaak, E.E. (2021) Sexual dimorphism in cardiometabolic health: the role of adipose tissue, muscle and liver. Nat. Rev. Endocrinol., 17, 47–66. [DOI] [PubMed] [Google Scholar]

[ref23] 23. Chang, E., Varghese, M. and Singer, K. (2018) Gender and sex differences in adipose tissue. Curr. Diab. Rep., 18, 69. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref24] 24. Van Iterson, M., Van Zwet, E.W., Consortium, B. and Heijmans, B.T. (2017) Controlling bias and inflation in epigenome- and transcriptome-wide association studies using the empirical null distribution. Genome Biol., 18, 19. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref25] 25. Min, S., Kim, K., Kim, S.G., Cho, H. and Lee, Y. (2018) Chromatin-remodeling factor, RSF1, controls p53-mediated transcription in apoptosis upon DNA strand breaks. Cell Death Dis., 9, 1079. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref26] 26. Charmet, R., Duffy, S., Keshavarzi, S., Gyorgy, B., Marre, M., Rossing, P., McKnight, A.J., Maxwell, A.P., Ahluwalia, T.V.S., Paterson, A.D.et al. (2018) Novel risk genes identified in a genome-wide association study for coronary artery disease in patients with type 1 diabetes. Cardiovasc. Diabetol., 17, 61. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref27] 27. Fantino, M., Paquette, M., Bernard, S. and Baass, A. (2021) ANKS1A genotype predicts cardiovascular events in patients with familial hypercholesterolemia. J. Clin. Lipidol., 15, 602–607. [DOI] [PubMed] [Google Scholar]

[ref28] 28. Koumanov, F., Pereira, V.J., Richardson, J.D., Sargent, S.L., Fazakerley, D.J. and Holman, G.D. (2015) Insulin regulates Rab3-Noc2 complex dissociation to promote GLUT4 translocation in rat adipocytes. Diabetologia, 58, 1877–1886. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref29] 29. Ouni, M., Saussenthaler, S., Eichelmann, F., Jahnert, M., Stadion, M., Wittenbecher, C., Ronn, T., Zellner, L., Gottmann, P., Ling, C.et al. (2020) Epigenetic changes in islets of Langerhans preceding the onset of diabetes. Diabetes, 69, 2503–2517. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref30] 30. Agha, G., Mendelson, M.M., Ward-Caviness, C.K., Joehanes, R., Huan, T., Gondalia, R., Salfati, E., Brody, J.A., Fiorito, G., Bressler, J.et al. (2019) Blood leukocyte DNA methylation predicts risk of future myocardial infarction and coronary heart disease. Circulation, 140, 645–657. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref31] 31. Yin, T., Miyazawa, K. and Yang, Y.C. (1992) Characterization of interleukin-11 receptor and protein tyrosine phosphorylation induced by interleukin-11 in mouse 3T3-L1 cells. J. Biol. Chem., 267, 8347–8351. [PubMed] [Google Scholar]

[ref32] 32. Alghamdi, M.A., Wallace, H.J., Melton, P.E., Moses, E.K., Stevenson, A., Al-Eitan, L.N., Rea, S., Duke, J.M., Danielsen, P.L., Prele, C.M.et al. (2020) Identification of differentially methylated CpG sites in fibroblasts from keloid scars. Biomedicine, 8, 181. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref33] 33. Moore, K., McKnight, A.J., Craig, D. and O'Neill, F. (2014) Epigenome-wide association study for Parkinson’s disease. NeuroMolecular Med., 16, 845–855. [DOI] [PubMed] [Google Scholar]

[ref34] 34. Seok, H., Park, H.J., Lee, B.W., Kim, J.W., Jung, M., Lee, S.R., Park, K.H., Park, Y.G., Baik, H.H. and Chung, J.H. (2013) Association of annexin A5 polymorphisms with obesity. Biomed. Rep., 1, 654–658. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref35] 35. Dankel, S.N., Fadnes, D.J., Stavrum, A.K., Stansberg, C., Holdhus, R., Hoang, T., Veum, V.L., Christensen, B.J., Vage, V., Sagen, J.V.et al. (2010) Switch from stress response to homeobox transcription factors in adipose tissue after profound fat loss. PLoS One, 5, e11033. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref36] 36. Parrillo, L., Costa, V., Raciti, G.A., Longo, M., Spinelli, R., Esposito, R., Nigro, C., Vastolo, V., Desiderio, A., Zatterale, F.et al. (2016) Hoxa5 undergoes dynamic DNA methylation and transcriptional repression in the adipose tissue of mice exposed to high-fat diet. Int. J. Obes., 40, 929–937. [DOI] [PubMed] [Google Scholar]

[ref37] 37. Hanley, P.J., Vollmer, V. and Bahler, M. (2020) Class IX myosins: motorized RhoGAP signaling molecules. Adv. Exp. Med. Biol., 1239, 381–389. [DOI] [PubMed] [Google Scholar]

[ref38] 38. Khim, K.W., Choi, S.S., Jang, H.J., Lee, Y.H., Lee, E., Hyun, J.M., Eom, H.J., Yoon, S., Choi, J.W., Park, T.E.et al. (2020) PPM1A controls diabetic gene programming through directly dephosphorylating PPARgamma at Ser273. Cell, 9, 343. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref39] 39. Surendran, P., Drenos, F., Young, R., Warren, H., Cook, J.P., Manning, A.K., Grarup, N., Sim, X., Barnes, D.R., Witkowska, K.et al. (2016) Trans-ancestry meta-analyses identify rare and common variants associated with blood pressure and hypertension. Nat. Genet., 48, 1151–1161. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref40] 40. Yu, Y., Hao, Y. and Feig, L.A. (2006) The R-Ras GTPase mediates cross talk between estrogen and insulin signaling in breast cancer cells. Mol. Cell. Biol., 26, 6372–6380. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref41] 41. Yao, Z.X., Jogunoori, W., Choufani, S., Rashid, A., Blake, T., Yao, W., Kreishman, P., Amin, R., Sidawy, A.A., Evans, S.R.et al. (2010) Epigenetic silencing of beta-spectrin, a TGF-beta signaling/scaffolding protein in a human cancer stem cell disorder: Beckwith-Wiedemann syndrome. J. Biol. Chem., 285, 36112–36120. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref42] 42. Steenaard, R.V., Ligthart, S., Stolk, L., Peters, M.J., vanMeurs, J.B., Uitterlinden, A.G., Hofman, A., Franco, O.H. and Dehghan, A. (2015) Tobacco smoking is associated with methylation of genes related to coronary artery disease. Clin. Epigenetics, 7, 54. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref43] 43. Tomizawa, M., Kawanabe, Y., Shinozaki, F., Sato, S., Motoyoshi, Y., Sugiyama, T., Yamamoto, S. and Sueishi, M. (2014) Triglyceride is strongly associated with nonalcoholic fatty liver disease among markers of hyperlipidemia and diabetes. Biomed. Rep., 2, 633–636. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref44] 44. de Mello, V.D., Matte, A., Perfilyev, A., Mannisto, V., Ronn, T., Nilsson, E., Kakela, P., Ling, C. and Pihlajamaki, J. (2017) Human liver epigenetic alterations in non-alcoholic steatohepatitis are related to insulin action. Epigenetics, 12, 287–295. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref45] 45. Lyu, Z.J., Wang, Y., Huang, J.L., Chen, M., Wu, S.Y., Yan, Q., Zhang, Y., Tang, Y., Jiang, C., Li, L.et al. (2021) Recurrent ZNF83-E293V mutation promotes bladder cancer progression through the NF-kappaB pathway via transcriptional dysregulation of S100A8. Mol. Ther., 29, 275–290. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref46] 46. Bacos, K., Gillberg, L., Volkov, P., Olsson, A.H., Hansen, T., Pedersen, O., Gjesing, A.P., Eiberg, H., Tuomi, T., Almgren, P.et al. (2016) Blood-based biomarkers of age-associated epigenetic changes in human islets associate with insulin secretion and diabetes. Nat. Commun., 7, 11089. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref47] 47. Dayeh, T., Tuomi, T., Almgren, P., Perfilyev, A., Jansson, P.A., deMello, V.D., Pihlajamaki, J., Vaag, A., Groop, L., Nilsson, E.et al. (2016) DNA methylation of loci within ABCG1 and PHOSPHO1 in blood DNA is associated with future type 2 diabetes risk. Epigenetics, 11, 482–488. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref48] 48. Davegardh, C., Sall, J., Benrick, A., Broholm, C., Volkov, P., Perfilyev, A., Henriksen, T.I., Wu, Y., Hjort, L., Brons, C.et al. (2021) VPS39-deficiency observed in type 2 diabetes impairs muscle stem cell differentiation via altered autophagy and epigenetics. Nat. Commun., 12, 2431. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref49] 49. Malmgren, S., Spegel, P., Danielsson, A.P., Nagorny, C.L., Andersson, L.E., Nitert, M.D., Ridderstrale, M., Mulder, H. and Ling, C. (2013) Coordinate changes in histone modifications, mRNA levels, and metabolite profiles in clonal INS-1 832/13 beta-cells accompany functional adaptations to lipotoxicity. J. Biol. Chem., 288, 11973–11987. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref50] 50. Castellano-Castillo, D., Moreno-Indias, I., Sanchez-Alcoholado, L., Ramos-Molina, B., Alcaide-Torres, J., Morcillo, S., Ocana-Wilhelmi, L., Tinahones, F., Queipo-Ortuno, M.I. and Cardona, F. (2019) Altered adipose tissue DNA methylation status in metabolic syndrome: relationships between global DNA methylation and specific methylation at adipogenic, lipid metabolism and inflammatory candidate genes and metabolic variables. J. Clin. Med., 8, 87. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref51] 51. Christiansen, C., Tomlinson, M., Eliot, M., Nilsson, E., Costeira, R., Xia, Y., Villicana, S., Mompeo, O., Wells, P., Castillo-Fernandez, J.et al. (2022) Adipose methylome integrative-omic analyses reveal genetic and dietary metabolic health drivers and insulin resistance classifiers. Genome Med., 14, 75. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref52] 52. Keller, M., Hopp, L., Liu, X., Wohland, T., Rohde, K., Cancello, R., Klos, M., Bacos, K., Kern, M., Eichelmann, F.et al. (2017) Genome-wide DNA promoter methylation and transcriptome analysis in human adipose tissue unravels novel candidate genes for obesity. Mol. Metab., 6, 86–100. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref53] 53. Gillberg, L., Perfilyev, A., Brons, C., Thomasen, M., Grunnet, L.G., Volkov, P., Rosqvist, F., Iggman, D., Dahlman, I., Riserus, U.et al. (2016) Adipose tissue transcriptomics and epigenomics in low birthweight men and controls: role of high-fat overfeeding. Diabetologia, 59, 799–812. [DOI] [PubMed] [Google Scholar]

[ref54] 54. Gillberg, L., Ronn, T., Jorgensen, S.W., Perfilyev, A., Hjort, L., Nilsson, E., Brons, C., Vaag, A. and Ling, C. (2019) Fasting unmasks differential fat and muscle transcriptional regulation of metabolic gene sets in low versus normal birth weight men. EBioMedicine, 47, 341–351. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref55] 55. Kokosar, M., Benrick, A., Perfilyev, A., Fornes, R., Nilsson, E., Maliqueo, M., Behre, C.J., Sazonova, A., Ohlsson, C., Ling, C.et al. (2016) Epigenetic and transcriptional alterations in human adipose tissue of polycystic ovary syndrome. Sci. Rep., 6, 22883. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref56] 56. Stener-Victorin, E., Holm, G., Labrie, F., Nilsson, L., Janson, P.O. and Ohlsson, C. (2010) Are there any sensitive and specific sex steroid markers for polycystic ovary syndrome? J. Clin. Endocrinol. Metab., 95, 810–819. [DOI] [PubMed] [Google Scholar]

[ref57] 57. Du, P., Zhang, X., Huang, C.C., Jafari, N., Kibbe, W.A., Hou, L. and Lin, S.M. (2010) Comparison of beta-value and M-value methods for quantifying methylation levels by microarray analysis. BMC Bioinformatics, 11, 587. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref58] 58. Price, M.E., Cotton, A.M., Lam, L.L., Farre, P., Emberly, E., Brown, C.J., Robinson, W.P. and Kobor, M.S. (2013) Additional annotation enhances potential for biologically-relevant analysis of the Illumina Infinium HumanMethylation450 BeadChip array. Epigenetics Chromatin, 6, 4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref59] 59. Chen, Y.A., Lemire, M., Choufani, S., Butcher, D.T., Grafodatskaya, D., Zanke, B.W., Gallinger, S., Hudson, T.J. and Weksberg, R. (2013) Discovery of cross-reactive probes and polymorphic CpGs in the Illumina Infinium HumanMethylation450 microarray. Epigenetics, 8, 203–209. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref60] 60. Teschendorff, A.E., Marabita, F., Lechner, M., Bartlett, T., Tegner, J., Gomez-Cabrero, D. and Beck, S. (2013) A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450 k DNA methylation data. Bioinformatics, 29, 189–196. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref61] 61. Johnson, W.E., Li, C. and Rabinovic, A. (2007) Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics, 8, 118–127. [DOI] [PubMed] [Google Scholar]

[ref62] 62. Benjamini, Y. and Hochberg, Y. (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B, 57, 289–300. [Google Scholar]

PERMALINK

Circulating triglycerides are associated with human adipose tissue DNA methylation of genes linked to metabolic disease

Tina Rönn

Alexander Perfilyev

Josefine Jönsson

Karl-Fredrik Eriksson

Sine W Jørgensen

Charlotte Brøns

Linn Gillberg

Allan Vaag

Elisabet Stener-Victorin

Charlotte Ling

Abstract

Introduction

Results

Circulating triglycerides are associated with altered DNA methylation in adipose tissue

Table 1.

Figure 1.

Circulating triglyceride levels are associated with both DNA methylation and cis-gene expression of several genes in adipose tissue

Table 2.

Figure 2.

Table 3.

Figure 3.

Adipose tissue DNA methylation as a partial causal mediator of the effect of circulating triglyceride levels on insulin resistance

Figure 4.

Table 4.

Circulating LDL, HDL and total cholesterol levels are associated with altered DNA methylation in adipose tissue

Figure 5.

Discussion

Materials and Methods

Study participants

DNA methylation analysis

mRNA expression analysis

Statistical analysis

Causal mediation analysis

Supplementary Material

Acknowledgements

Contributor Information

Funding

Data availability

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases