ArticlePDF Available

Epigenome-wide association study of body mass index, and the adverse outcomes of adiposity

December 2016
Nature 541(7635)

DOI:10.1038/nature20784

Authors:

Simone Wahl

Roche Diagnostics, Penzberg, Germany

Benjamin Lehne

Imperial College London

Marie Loh

Imperial College London

Show all 107 authorsHide

Approximately 1.5 billion people worldwide are overweight or affected by obesity, and are at risk of developing type 2 diabetes, cardiovascular disease and related metabolic and inflammatory disturbances. Although the mechanisms linking adiposity to associated clinical conditions are poorly understood, recent studies suggest that adiposity may influence DNA methylation, a key regulator of gene expression and molecular phenotype. Here we use epigenome-wide association to show that body mass index (BMI; a key measure of adiposity) is associated with widespread changes in DNA methylation (187 genetic loci with P < 1 × 10−7, range P = 9.2 × 10−8 to 6.0 × 10−46; n = 10,261 samples). Genetic association analyses demonstrate that the alterations in DNA methylation are predominantly the consequence of adiposity, rather than the cause. We find that methylation loci are enriched for functional genomic features in multiple tissues (P < 0.05), and show that sentinel methylation markers identify gene expression signatures at 38 loci (P < 9.0 × 10−6, range P = 5.5 × 10−6 to 6.1 × 10−35, n = 1,785 samples). The methylation loci identify genes involved in lipid and lipoprotein metabolism, substrate transport and inflammatory pathways. Finally, we show that the disturbances in DNA methylation predict future development of type 2 diabetes (relative risk per 1 standard deviation increase in methylation risk score: 2.3 (2.07–2.56); P = 1.1 × 10−54). Our results provide new insights into the biologic pathways influenced by adiposity, and may enable development of new strategies for prediction and prevention of type 2 diabetes and other adverse clinical consequences of obesity.

Content uploaded by Marian Beekman

Content may be subject to copyright.

5 JANUARY 2017 | VOL 541 | NATURE | 81

LETTER doi:10.1038/nature20784

Epigenome-wide association study of body mass

index, and the adverse outcomes of adiposity

A list of authors and affiliations appears at the end of the paper.

Approximately 1.5 billion people worldwide are overweight or

affected by obesity, and are at risk of developing type 2 diabetes,

cardiovascular disease and related metabolic and inflammatory

disturbances1,2. Although the mechanisms linking adiposity to

associated clinical conditions are poorly understood, recent studies

suggest that adiposity may influence DNA methylation3–6, a key

regulator of gene expression and molecular phenotype

. Here we

use epigenome-wide association to show that body mass index

(BMI; a key measure of adiposity) is associated with widespread

changes in DNA methylation (187genetic loci with P < 1 × 10

−7

range P = 9.2 × 10−8 to 6.0 × 10−46; n = 10,261 samples). Genetic

association analyses demonstrate that the alterations in DNA

methylation are predominantly the consequence of adiposity,

rather than the cause. We find that methylation loci are enriched

for functional genomic features in multiple tissues (P < 0.05), and

show that sentinel methylation markers identify gene expression

signatures at 38loci (P < 9.0 × 10−6, range P = 5.5 × 10−6 to

6.1 × 10−35, n = 1,785 samples). The methylation loci identify genes

involved in lipid and lipoprotein metabolism, substrate transport

and inflammatory pathways. Finally, we show that the disturbances

in DNA methylation predict future development of type 2 diabetes

(relative risk per 1standard deviation increase in methylation risk

score: 2.3 (2.07–2.56); P = 1.1 × 10−54). Our results provide new

insights into the biologic pathways influenced by adiposity, and may

enable development of new strategies for prediction and prevention

of type 2 diabetes and other adverse clinical consequences of obesity.

Our study design is summarized in Extended Data Fig. 1. We carried

out epigenome-wide association in 5,387 individuals from the EPICOR

(n = 514), KORA (n = 2,193) and LOLIPOP (n = 2,680) population

studies (Supplementary Tables 1, 2, and Supplementary Information).

We studied individuals of European (EPICOR, KORA) and Indian

Asian (LOLIPOP) ancestry, both populations are known to be at high

risk of obesity and related metabolic disturbances2,8. DNA methylation

in genomic DNA from blood was quantified using an Illumina Infinium

450K human methylation array. Blood was chosen for the analysis as

it is a metabolically active tissue, with an important role in the adverse

inflammatory and vascular consequences of adiposity, and is widely

used for clinical diagnostic purposes.

Epigenome-wide association identified 278 CpG sites associated

with BMI at P < 1 × 10−7, distributed between 207 genetic loci

(Supplementary Tables 3, 4). At each locus, we identified the sentinel

marker (CpG site with lowest P value for association with BMI), and

carried out replication testing in separate samples of whole blood from

European and Indian Asian men and women in population-based

studies (n = 4,874; Supplementary Table 1). The association between

DNA methylation and BMI was replicated at 187 out of 207 markers

(associated with BMI at P < 0.05 in replication samples with directional

consistency, and at epigenome-wide significance in combined analysis

of discovery and replication data (Fig. 1 and Supplementary Table 3)).

Regional plots for the 187identified loci are shown in Supplementary

Figs 1 and 2. Effect sizes range from 6.3 ± 0.9 to 40.2 ± 3.1 kgm−2

change in BMI per unit increase in DNA methylation in blood (scale

for methylation 0–1, in which 1 represents 100% methylation), with

little evidence for heterogeneity between Europeans and Indian Asians

(Supplementary Table 3). At sevenloci, the associations between DNA

methylation and BMI are stronger amongst Indian Asians or Europeans

(heterogeneity P < 1.0 × 10−7), raising the possibility that some effects

may be population specific.

Sensitivity analyses show that our findings are robust to the choice of

analytic strategy. The associations of DNA methylation in blood with

BMI are not explained by population stratification caused by DNA

sequence variation, or by genetic confounding of single-nucleotide

polymorphisms (SNPs) in the probe sequence (Supplementary Table 5

and Supplementary Figs 3, 4). In addition, to address the possibility of

confounding by technical factors, we further replicated the associations

of DNA methylation in blood with BMI at 4loci, in 990 Europeans and

1,720 Indian Asians (LOLIPOP study), using pyrosequencing as an

alternative approach to quantification of methylation (P = 1.2 × 10

−7

to 2.1 × 10−12 for association of methylation with BMI; Supplementary

Table 6).

The 187identified methylation markers are strongly enriched for

CpG sites with intermediate levels of methylation, consistent with

the presence of mosaicism, that is, epigenetic heterogeneity, at these

loci (P = 1.4 × 10−22, Fisher’s test; Extended Data Fig. 2). To better

understand the underlying cellular events, and exclude changes in

cell subset composition as the basis for our findings, we carried out

replication testing of the sentinel loci in isolated white-blood-cell

subsets (monocytes, neutrophils, CD4+T cells and CD8+T cells,

n = 60; Supplementary Table 7). Epigenetic heterogeneity is present at

the majority of loci, in each of the cell subsets studied (Extended Data

Fig. 3 and Supplementary Table 8). The sentinel markers are enriched

for association with adiposity in each of the isolated cell subsets

(Extended Data Fig. 4 and Supplementary Table 8), and the relation-

ships between methylation and obesity are directionally consistent with

the discovery epigenome-wide association study at 130loci (CD4

P = 1.2 × 10−9, sign test) and 166loci (neutrophils, P = 5.6 × 10−35,

sign test) (Supplementary Table 9). Furthermore, effect sizes are direc-

tionally consistent and of similar magnitude between the isolated cell

subsets (Extended Data Fig. 5). The association of DNA methylation

with BMI therefore reflects epigenetic heterogeneity at the identified

loci, is independent of changes in cell subset distribution, and com-

prises an effect of adiposity on methylation that is shared across the

cell subsets studied.

To assess the relevance of our observations in blood to other meta-

bolically relevant tissues, we first compared methylation levels at

the 187loci in blood, subcutaneous and omental fat, liver, muscle,

spleen and pancreas9. Mean methylation levels at the 187loci

correlate moderately to strongly between the tissues (R = 0.37− 0.93,

P = 8.9 × 10−8 to 1.9 × 10−82 for the 21tissue pairs; Extended Data

Fig. 6 and Supplementary Fig. 5), supporting the view that methylation

levels in blood are related to methylation patterns in other tissues at the

CpG sites examined.

Inflammatory and hormonal disturbances in adipocytes of obese

people contribute to the development of insulin resistance and other

metabolic consequences of adiposity10. To better understand how our

findings in the blood might reflect processes in the adipose tissue,

Letter

reSeArCH

82 | NATURE | VOL 541 | 5 JANUARY 2017

we therefore quantified the relationship between DNA methylation

and BMI in adipose tissue. We found that 120 of the CpG sites show

directional consistency for association with BMI in both adipose

tissue and blood (P = 1.3 × 10

−4

, binomial test), whereas 91sites are

associated with BMI in adipose tissue (P < 2.7 × 10−4, that is, P < 0.05

after Bonferroni correction for 187tests; Supplementary Table 10). The

associations of DNA methylation with BMI in adipose tissue are also

unlikely to be the result of differences in the composition of canonical

cell types. First we used principal component analysis to assess the

cryptic structure arising from variation in cell subset composition in

the methylation data. Including principal components as covariates

in regression models did not materially influence the association of

DNA methylation with BMI in adipose tissue (Supplementary Fig. 6).

In separate studies, we quantified DNA methylation in isolated

adipocytes from subcutaneous adipose tissue collected from morbidly

obese (BMI > 40 kg m−2, n = 24) and normal weight (n = 24)

individuals. Despite the small sample size, 6 out of 187sentinel markers

were associated with obesity at P < 2.7 × 10

−4

(P < 0.05 after Bonferroni

correction; Supplementary Table 11), and 108markers show relation-

ships with obesity that are directionally consistent with those observed

in the discovery epigenome-wide association study (P = 0.04). We sepa-

rately tested the association of our sentinel methylation markers with

BMI in samples of the liver (n = 55), as this is also a metabolically rele-

vant tissue. We found that 114CpG sites showed consistent direction

of association with BMI compared to our findings in blood (P = 0.001,

sign test; Supplementary Table 10). Our findings indicate that many of

the relationships between methylation and BMI in blood are shared by

adipose and liver cells, but also identify effects that are tissue specific.

Next, we used genetic association and the concept of Mendelian ran-

domization to investigate the potential causal relationships between

DNA methylation in blood and BMI11. We first identified SNPs influ-

encing DNA methylation in blood in cis (1 Mb, n = 4,034 people). We

then tested whether SNPs that influence methylation in blood also

influence BMI, and whether the predicted effects of SNPs on BMI

through methylation are consistent with the observed association.

We identified a single CpG (cg26663590: NFATC2IP) that showed

evidence of a genetic association for a causal role of methylation on

BMI (P = 9.6 × 10−7 for association of SNP rs11150675 near NFATC2IP

with BMI; Fig. 2a and Supplementary Table 12). In keeping with a

causal role for methylation at NFATC2IP underlying adiposity, baseline

levels of methylation at cg26663590 predict weight gain in longitudinal

population studies (P = 0.03, Supplementary Table 13). The NFATC2IP

locus contains the gene that encodes SH2B1, which is known to be

involved in energy and glucose homeostasis and has previously been

linked with obesity, including in genome-wide association studies12,13.

To investigate whether DNA methylation in blood is the consequence

of adiposity, we used a weighted genetic risk score that combines effects

across SNPs known to influence BMI (Fig. 2b and Supplementary Table 14).

We observed a strong correlation between predicted (through

BMI) and observed effects of BMI genetic risk score on methylation

(R2 = 0.65; P = 4.7 × 10−44) at the CpG sites evaluated. In particular,

genetic risk score was associated with DNA methylation at the ABCG1,

KLHL18 and FTH1P20 loci with P < 2.7 × 10−4 (corresponding to

P < 0.05 after Bonferroni correction for 187tests). An effect of BMI

on ABCG1 methylation is consistent with observations that weight

loss influences both ABCG1 expression in adipose tissue and ABCG1

SDF4

TIE1

PDE4DIP

SLC41A1

TGFA

CRYGFP

ACOX2

GOLIM4

SLC7A11

FLT4

ARID1B

RFC2

RPL5P24

NACC2

SMC3

PRDM11

PLBD1

LINC00598

ZC3H14

ACSBG1

NFATC2IP

SLC7A5

SREBF1

PHOSPHO1

P4HB

DEDD2

ABCG1

WRAP73

DHCR24

S100A2

TMEM63A

SULT1C2

SLC11A1

PRICKLE2

SOX2−OT

ROPN1L

EXOC2

SYNJ2

TRIP6

LINC00964

BEND7

FAM53B

SLC43A1

CNTN1

SMIM2−AS1

GPR68

FSD2

84FNZ

CBFA2T3

AKAP10

VMP1

SBNO2

BCL3

PIK3IP1

C1orf127

PHBP3

LMNA

LINC00184

MAP3K2

HDAC4

ZPLD1

ABCC5

TRIO

JARID2

RPS6KA2

SH2B2

SLC45A4

ANAPC16

PTPRE

BSCL2

IGFBP6

RNASE10

IGHA2

MAN2A2

NOD2

ANKRD11

USP22

CPSF4L

CREB3L3

SLC1A5

CARD10

SPATA21

NFIA

BCAN

LPIN1

UGGT1

IL5RA

EAF2

BCL6

TMEM173

GABBR1

MAD1L1

ST7

ZC3H3

C10orf99

CTSD

CPT1A

DCN

CPNE6

ATP10A

SLCO3A1

BBS2

CLUH

RARA

SLC9A3R1

ATG4D

ANGPT4

TAB1

AHDC1

S1PR1

DARC

RNA5SP89

ZEB2

DLEC1

EEFSEC

MUC4

RNF145

LY6G6F

ICA1

SND1

AQP3

PCGF5

KCNQ1

BCO2

DRAM1

RPL10L

APBA2

AXIN1

DUS2L

ZZEF1

VPS25

RHBDF2

PKN1

PPIAP3

CRELD2

KIAA0319L

SARS

CD247

LBH

NFE2L2

KLHL18

1CCFE

LINC00504

NPM1

COX6A1P2

HOXA−AS2

RN7SL142P

ANXA1

LINC00263

PRR5L

BCL9L

UBE2L5P

HSPA2

MYO5C

SLC9A3R2

MAF

ACADVL

MEOX1

SOCS3

MYO9B

TOX2

STK40

PHGDH

RPS10P7

CDC42EP3

FTH1P20

NT5DC2

LINC00880

CEP135

SLC34A1

C6orf223

GARS

LOXL2

PRRC2B

ELOVL3

CD82

CLMP

UFM1

ELMSAN1

DAPK2

RMI2

GSE1

CHD3

NFE2L1

LGALS3BP

UPF1

LINC00649

–log10(P)

Figure 1 | Circos plot of the epigenome-wide association of DNA

methylation in blood with BMI. Results are presented as CpG-specific

association test results (− log10(P)) ordered by genomic position.

Green and blue symbols, CpG sites at loci reaching epigenome-wide

significance (P < 1 × 10−7); grey symbols, CpG sites at loci not reaching

epigenome-wide significance. Chromosome numbers are shown on

the inner ring. Tick marks on the outer ring identify the genomic loci

reaching epigenome-wide significance. The genes nearest to the sentinel

methylation markers at each of the 187loci are listed around the circos

plot.

Letter reSeArCH

5 JANUARY 2017 | VOL 541 | NATU RE | 83

activity14,15, and by the close relationship between change in BMI

and change in methylation during longitudinal follow-up analyses

of participants in our population studies (Fig. 3 and Supplementary

Table 13). Although further studies are needed to consider mechanisms,

our findings suggest that adiposity determines the alterations in

methylation at the majority of the identified CpG sites.

We used genetic association separately to test the causal relation-

ships between BMI and DNA methylation in adipose tissue. The results

further confirm that in adipose tissue, as in blood, the differences

in methylation observed are primarily the consequence of adiposity

(R = 0.73, P = 1.6 × 10−32; Extended Data Fig. 7).

We carried out functional genomic analyses to investigate the

potential mechanisms linking the 187sentinel CpGs sites with

adiposity. The CpG sites are strongly enriched in active chromatin

sites, including at DNase hypersensitivity sites and the activating

histone marks H3K4me1 and H3K27ac in a wide range of cell lines

(P < 0.05; Supplementary Fig. 7), suggesting that the adiposity-related

methylation changes we identified occur at constitutive cis-regulatory

regions that operate across tissues. In keeping with a regulatory role,

DNA methylation at the 187CpG sites is enriched for association with

expression of cis genes (500 kb) in blood (Supplementary Tables 15,

16 and Extended Data Fig. 8). We find 44transcripts of 38annotated

genes that are associated with DNA methylation at P < 9.0 × 10−6

(that is, P < 0.05 after Bonferroni correction; Supplementary Table 16);

an approximate threefold enrichment compared to expectations under

the null hypothesis (P = 3.0 × 10

−4

; Extended Data Fig. 8). Sensitivity

analyses, by limiting assessment of the relationship between methy-

lation and gene expression to the nearest gene or Illumina-annotated

gene, identified five additional loci potentially associated with gene

expression (Supplementary Table 17). The strongest cis signals observed

were for cg09315878 with TNFRSF4 transcription (P = 7.2 × 10

−86

cg14476101 with PHGDH transcription (P = 1.0 × 10−64) and

cg09152259 with MAP3K2 transcription (P = 1.6 × 10

−67

). On average,

a 5% absolute change in methylation was associated with a 7% change

in gene expression across the 44transcripts identified (range 1.8%

for AKAP to 19% for SPNS3; Supplementary Table 16). Among

the 38methylation–gene expression associations observed in blood,

3were also found in adipose tissue (HOXA5, BBS2 and SELM (also

known as SELENOM) and 3 in the liver (ANXA1, LGALS3BP and

PHGDH) at P < 1.3 × 10

−3

(that is, P < 0.05 after Bonferroni correction

for 38tests), all with consistent direction of effect (Supplementary

Table 18), suggesting that the relationships between methylation and

6.0

4.0

2.0

0.0

–2.0

–4.0

–6.0

5.4

3.6

1.8

0.0

–1.8

–3.6

–5.4

–6.0 –4.0 –2.0 0 2.0 4.0 6.0 –5.4 –3.6 –1.8 01.8 3.6 5.4

Observed effect (×10–2)

Observed effect (×10–3)

Methylation consequence of BMIMethylation cause of BMI

CpG

SNPBMI

BMI

GRSCpG

FTH1P20

ABCG1

KLHL18

Predicted effectPredicted effect

Figure 2 | Genetic association studies to investigate the potential

relationships between BMI and DNA methylation in blood. a, Results

for the causality analysis investigating whether DNA methylation in blood

at the sentinel CpG sites influences BMI. Units are change in BMI per copy

of effect allele. For each sentinel CpG site, we identified the cis SNP (1 Mb)

most closely associated with DNA methylation levels. For each SNP, we

then determined the effect of SNP on BMI predicted via methylation

(x axis) and the directly observed effect of the SNP on BMI (y axis). Grey,

CpGs not significantly associated with a SNP; blue, CpGs significantly

associated with a SNP. For a single CpG (NFATC2IP) the associated SNP

is also associated with BMI and 95% CI error bars are shown. At the other

loci, there was little relationship between the effects of the SNPs on BMI

predicted through methylation and directly observed BMI (R2 = 0.00,

P = 0.86). b, Consequential analysis for a causality analysis investigating

whether DNA methylation in blood at the sentinel CpG sites is the

consequence of BMI. Units are change in methylation per unit change in

weighted genetic risk score. We identified the SNPs reported to influence

BMI in a genome-wide association meta-analysis12, and calculated a

weighted genetic risk score (see Methods). For each sentinel CpG site, we

then determined the effect of genetic risk score on methylation predicted

through BMI (x axis) and the directly observed effect of genetic risk

score on CpG (y axis). Three CpGs (ABCG1, KLHL18 and FTH1P20)

are associated with the genetic risk score at P < 2.7 × 10−4 (P < 0.05

after Bonferroni correction for 187 tests; 95% CI error bars shown). The

overall correlation between observed and predicted effects (R2 = 0.81;

P = 4.7 × 10−44) suggests that methylation in blood at the majority of CpG

sites is consequential to BMI.

–15

–10

–5

–50 –30 –10 10 30

Cross-sectional study

Longitudinal study

178 out of 187 markers with

consistent direction of

effect (P = 6.8 × 10

–42

)

Figure 3 | Relationship between DNA methylation in blood and

BMI amongst 1,435 participants of the KORA S4/F4 population

cohort. Cross-sectional results (x axis) are for the relationship between

methylation in blood and BMI at each of the 187sentinel CpG sites

in the baseline samples; longitudinal results are for the relationship

between change in methylation (in blood) and change in BMI after 7year

follow-up. Units for both axes are kgm−2 change in BMI per unit increase

in methylation (scale 0–1, in which 1 represents 100% methylation).

Letter

reSeArCH

84 | NATURE | VOL 541 | 5 JANUARY 2017

gene expression are in part shared between blood, adipose and liver

tissue.

We prioritized genes as potential candidate genes involved in the

association between BMI and DNA methylation at the 187loci on the

basis of two criteria: (1) proximity, the gene nearest to the sentinel

methylation marker and (2) functional genomics, genes within

500 kb of the sentinel methylation marker showing association of

gene expression with methylation (Supplementary Table 19). These

criteria identified 210unique genes, many with established roles in

adipose tissue biology and insulin resistance (for example, ABCG1,

LPIN1, HOXA5, LMNA, CPT1A, SOCS3, SREBF1 and PHGDH;

Supplementary Tables 19, 20). Gene-set enrichment analyses show that

the 210candidate genes are enriched for genes involved in lipid and

lipoprotein metabolism, amino-acid and small-molecule transport, and

inflammatory pathways involving NF-κ B, MAPK, TAK1 (also known

as MAP3K7), IRAK2 and TRAF6 (Supplementary Table 21).

To investigate the potential clinical significance of the DNA

methylation changes, we first tested the cross-sectional relationship

of DNA methylation in blood with fasting glucose, insulin, HDL

(high-density lipoprotein) cholesterol, triglycerides, Haemoglobin

A1c (HbA1c) and other clinical traits. We found that 879 methylation–

clinical-trait pairs tested were significant at P < 2.1 × 10−5 (that is,

P < 0.05 after Bonferroni correction for the 2,431tests performed;

Supplementary Fig. 8 and Supplementary Table 22), consistent with

recent studies reporting close relationships of DNA methylation with

blood lipids and glucose traits16,17. We used genetic association to

investigate the potential causal relationships between DNA methylation

and the identified clinical traits. SNPs influencing methylation markers

in blood showed little evidence for association with the respective

clinical traits (Extended Data Fig. 9). By contrast, the predicted effect

of genetic risk score on DNA methylation through clinical trait is

correlated with the directly observed effect of genetic risk score on

methylation for HbA1c, HDL cholesterol, triglycerides and insulin

(P = 1 × 10

−3

to 2 × 10

−14

; Extended Data Fig. 9). Our findings suggest

that the methylation changes in blood may in part be a consequence

of the changes in lipid and glucose metabolism associated with BMI.

Finally, we tested whether DNA methylation levels in blood at the

187sentinel CpG sites predict new onset, incident type 2 diabetes, a

major clinical consequence associated with obesity, in parti cipants

of the LOLIPOP study (n = 2,664). In single-marker tests, 62 of

the 187methylation markers were associated with incident type 2

diabetes at P < 2.7 × 10

−4

(that is, P < 0.05 after Bonferroni correction;

Supplementary Table 23). The strongest association was observed for

the ABCG1 locus, a gene known to be involved in insulin secretion

and pancreatic β -cell function14,15. To integrate information across

CpG sites, we calculated a weighted methylation risk score as the sum

of methylation values at each of the markers associated with type 2

diabetes, weighted by marker-specific effect size. Methylation risk

score is strongly predictive of incident type 2 diabetes (relative risk

2.29 (95% confidence intervals (CI), 2.06–2.55) per 1s.d. change in

methylation risk score; P = 4.2 × 10−52). The association of methylation

risk score with incident type 2 diabetes was also found in Europeans

from the KORA study (relative risk, 2.51; (95% CI, 1.49–4.23) per 1s.d.

change in methylation risk score; P = 5.7 × 10

−4

), with no evidence for

heterogeneity of effect (P = 0.74). Methylation risk score predicts type

2 diabetes beyond traditional risk factors including BMI and waist–hip

ratio (Supplementary Table 24), and in particular identifies obese and

overweight individuals with a high risk of developing type 2 diabetes

in the future(relative risk for type 2 diabetes in obese subjects: 7.3

(4.1–12.9), P = 8.2 × 10−12 in the top compared to the lowest quar-

tile; Fig. 4). This risk of type 2 diabetes associated with DNA methyl-

ation markers, which we have estimated in our study is numerically

similar to, or greater than, the estimated risk conferred by traditional

risk factors, including overweight, obesity, central obesity, impaired

fasting glucose and hyperinsulinaemia (Extended Data Fig. 10).

Furthermore, DNA methylation remains strongly and independently

associated with risk of future type 2 diabetes even after adjustment for

adiposity and glycaemic measures. By contrast, emergent risk factors

such as C-reactive protein and amino acid concentrations showed little

evidence for an independent association with type 2 diabetes. Our

findings therefore raise the possibility that DNA-methylation markers

may help to identify individuals with metabolically unfavourable

adiposity that are at increased risk of developing type 2 diabetes.

Our large-scale epigenome-wide association study identifies and

replicates changes in DNA methylation associated with BMI in blood

and adipose tissue. The associations of methylation with BMI are inde

pendent of variation in cell subset composition and replicate in both

isolated white blood cells and isolated adipocytes. Genetic association

in both blood and adipose tissue supports the view that the changes

in DNA methylation are a consequence and not the cause of adiposity,

at the majority of the identified CpG sites. The presence of epigenetic

heterogeneity at the identified loci, even within isolated canonical cell

subsets, together with a graded relationship between methylation and

BMI, suggest epigenetic reprogramming within committed cell subsets

in response to adiposity, as recently described in other tissues18. In

keeping with this, the methylation loci are enriched for sites of open

chromatin in multiple tissues, consistent with the presence of consti-

tutive cis enhancers.

The candidate genes at these loci include genes with annotated

roles in lipid metabolism, amino acid and small molecule transport,

inflammation, as well as metabolic, cardiovascular, respiratory and

neoplastic disease. For example, TNFRSF4 and MAP3K2 encode

proteins involved in activation of NF-κ B

, whereas IL5RA is involved

in development and activation of eosinophil and other immune cells,

and is causally linked to asthma, eczema and cardiovascular disease20.

ABCG1 is involved in cholesterol and phospholipid transport, and

regulates insulin secretion

17,21

. Our observations thus provide insight

into the regulatory pathways that may link adiposity to metabolic and

cardiovascular disease, asthma and a wide range of cancers, although

our study is limited in the tissues examined, and further studies

are needed to include additional biologically relevant tissues. Our

prospective population studies show that DNA methylation iden-

tifies people at high risk of incident type 2 diabetes, independent of

conventional risk factors. Further studies are needed to examine

whether DNA-methylation markers may be useful in distinguishing

metabolically unhealthy obesity. This may prove useful in risk stratifi-

cation and personalized medicine, to help to tackle the current global

epidemic of obesity and its associated cardiovascular and metabolic

disturbances.

024681012141

618

Odds ratio

Controls/cases PP trend

144/29 3.85 × 10–7

141/32 4.97 × 10–1

130/43 2.61 × 10–2

106/69 1.89 × 10–7

129/27 9.50 × 10–1 5.66 × 10–19

185/78 7.70 × 10–4

169/115 9.00 × 10–8

301/321 4.00 × 10–16

27/17 2.50 × 10–3 4.19 × 10–7

50/28 5.20 × 10–4

59/61 5.10 × 10–10

149/255 7.90 × 10–22

Normal

Overweight

Obese

P interaction = 0.56

Figure 4 | Relative risk of incident type 2 diabetes by quartile

of methylation risk score amongst Indian Asians. Subjects were

normoglycaemic (HbA1c < 6% and fasting glucose < 6 mmoll−1)

and normal weight (BMI 18.5–24.9 kgm−2), overweight (BMI 25.0–

29.9 kgm−2) or obese (BMI ≥ 30.0 kgm−2). The interaction P value is for

the interaction between adiposity and DNA methylation in risk of type 2

diabetes.

Letter reSeArCH

5 JANUARY 2017 | VOL 541 | NATU RE | 85

Online Content Methods, along with any additional Extended Data display items and

Source Data, are available in the online version of the paper; references unique to

these sections appear only in the online paper.

Received 15 March 2015; accepted 10 November 2016.

Published online 21 December 2016.

1. Wang, Y. C., McPherson, K., Marsh, T., Gortmaker, S. L. & Brown, M. Health and

economic burden of the projected obesity trends in the USA and the UK.

Lancet 378, 815–825 (2011).

2. Ng, M. et al. Global, regional, and national prevalence of overweight and obesity

in children and adults during 1980–2013: a systematic analysis for the Global

Burden of Disease Study 2013. Lancet 384, 766–781 (2014).

3. Dick, K. J. et al. DNA methylation and body-mass index: a genome-wide

analysis. Lancet 383, 1990–1998 (2014).

4. Feinberg, A. P. et al. Personalized epigenomic signatures that are stable

over time and covary with body mass index. Sci. Transl. Med. 2, 49ra67

(2010).

5. Xu, X. et al. A genome-wide methylation study on obesity: dierential variability

and dierential methylation. Epigenetics 8, 522–533 (2013).

6. Demerath, E. W. et al. Epigenome-wide association study (EWAS) of BMI, BMI

change and waist circumference in African American adults identies multiple

replicated loci. Hum. Mol. Genet. 24, 4464–4479 (2015).

7. Portela, A. & Esteller, M. Epigenetic modications and human disease.

Nat. Biotechnol. 28, 1057–1068 (2010).

8. Danaei, G. et al. National, regional, and global trends in fasting plasma glucose

and diabetes prevalence since 1980: systematic analysis of health

examination surveys and epidemiological studies with 370 country–years and

2·7 million participants. Lancet 378, 31–40 (2011).

9. Slieker, R. C. et al. Identication and systematic annotation of tissue-specic

dierentially methylated regions using the Illumina 450k array. Epigenetics

Chromatin 6, 26 (2013).

10. Rosen, E. D. & Spiegelman, B. M. What we talk about when we talk about fat.

Cell 156, 20–44 (2014).

11. Relton, C. L. & Davey Smith, G. Two-step epigenetic Mendelian randomization:

a strategy for establishing the causal role of epigenetic processes in pathways

to disease. Int. J. Epidemiol. 41, 161–176 (2012).

12. Speliotes, E. K. et al. Association analyses of 249,796 individuals reveal 18 new

loci associated with body mass index. Nat. Genet. 42, 937–948 (2010).

13. Bochukova, E. G. et al. Large, rare chromosomal deletions associated with

severe early-onset obesity. Nature 463, 666–670 (2010).

14. Johansson, L. E. et al. Dierential gene expression in adipose tissue from obese

human subjects during weight loss and weight maintenance. Am. J. Clin. Nutr.

96, 196–207 (2012).

15. Aron-Wisnewsky, J. et al. Eect of bariatric surgery-induced weight loss on

SR-BI-, ABCG1-, and ABCA1-mediated cellular cholesterol eux in obese

women. J. Clin. Endocrinol. Metab. 96, 1151–1159 (2011).

16. Pfeierm, L. et al. DNA methylation of lipid-related genes aects blood lipid

levels. Circ Cardiovasc Genet 8, 334–342 (2015).

17. Hidalgo, B. et al. Epigenome-wide association study of fasting measures of

glucose, insulin, and HOMA-IR in the Genetics of Lipid Lowering Drugs and

Diet Network study. Diabetes 63, 801–807 (2014).

18. Donkin, I. et al. Obesity and Bariatric surgery drive epigenetic variation of

spermatozoa in humans. Cell Metab. 23, 369–378 (2016).

19. Karin, M. & Ben-Neriah, Y. Phosphorylation meets ubiquitination: the control of

NF-κ B activity. Annu. Rev. Immunol. 18, 621–663 (2000).

20. Brightling, C. E. et al. Benralizumab for chronic obstructive pulmonary disease

and sputum eosinophilia: a randomised, double-blind, placebo-controlled,

phase 2a study. Lancet Respir. Med. 2, 891–901 (2014).

21. Chambers, J. C. et al. Epigenome-wide association of DNA methylation markers

in peripheral blood from Indian Asians and Europeans with incident type 2

diabetes: a nested case–control study. Lancet Diabetes Endocrinol. 3, 526–534

(2015).

SimoneWahl1,2,3*, AlexanderDrong4*, BenjaminLehne5*, MarieLoh5,6,7*,

WilliamR.Scott5,8*, SonjaKunze1,2, Pei-ChienTsai9, JaninaS.Ried10,

WeihuaZhang5,11, YouwenYang5, SiliTan12, GiovanniFiorito13,14,

LudeFranke15, SimonettaGuarrera13,14, SilvaKasela16,17, JenniferKriebel1,2,3,

RebeccaC.Richmond18, MarcoAdamo19, UzmaAfzal5,11,

MikaAla-Korpela20,21,22, BenedettaAlbetti23, OleAmmerpohl24,

JaneF.Apperley25, MarianBeekman26, PierAlbertoBertazzi23,

S.LucasBlack27, ChristineBlancher28, Marc-JanBonder15, MarioBrosch29,

MarenCarstensen-Kirberg3,30, AntonJ.M.de Craen31‡, SimondeLusignan32,

AbbasDehghan33, MohamedElkalaawy19,34, KristaFischer16,

OscarH.Franco33, TomR.Gaunt18, JochenHampe29, MajidHashemi19,

AaronIsaacs33, AndrewJenkinson19, SujeetJha35, NorihiroKato36,

VittorioKrogh37, MichaelLaffan25, ChristaMeisinger2, ThomasMeitinger38,39,40,

ZuanYuMok12, ValeriaMotta23, HongKiatNg12, ZacharoulaNikolakopoulou41,

GeorgiosNteliopoulos25, SalvatorePanico42, NataliaPervjakova16,17,

HolgerProkisch38,39, WolfgangRathmann43, MichaelRoden3,30,44,

FedericaRota23, MichelleAnnRozario12, JohannaK.Sandling45,46,

ClemensSchafmayer47, KatharinaSchramm38,39, ReinerSiebert24,48,

P.ElineSlagboom26, PasiSoininen20,21, LisetteStolk49, KonstantinStrauch10,50,

E-ShyongTai51,52,53, LetiziaTarantini23, BarbaraThorand2,3,

EttjeF.Tigchelaar15, RosarioTumino54, AndreG.Uitterlinden55,

Corneliavan Duijn33, JoyceB.J.van Meurs49, PaoloVineis13,56,

AnandaRajithaWickremasinghe57, CiscaWijmenga15, Tsun-PoYang45,

WeiYuan9,58, AlexandraZhernakova15, RachelL.Batterham19,59,

GeorgeDaveySmith18, PanosDeloukas45,60,61, BastiaanT.Heijmans26,

ChristianHerder3,30, AlbertHofman33, CeciliaM.Lindgren4,62, LiliMilani16,

Pimvan der Harst15,63,64, AnnettePeters2,3,40, ThomasIllig1,2,65,66,

CarolineL.Relton18, MelanieWaldenberger1,2, Marjo-RiittaJärvelin67,68,69,70,

ValentinaBollati23, RichieSoong12,71, TimD.Spector9, JamesScott8,

MarkI.McCarthy4,72,73, PaulElliott5,74, JordanaT.Bell9§,

GiuseppeMatullo13,14§, ChristianGieger1,2§, JaspalS.Kooner8,11,74§,

HaraldGrallert1,2,3§ & JohnC.Chambers5,11,74,75§

Supplementary Information is available in the online version of the paper.

Acknowledgements Detailed acknowledgments are provided in the

Supplementary Information.

Author Contributions Data collection and analysis in the contributing

population studies, ALSPAC study: T.R.G., C.L.R., R.C.R., G.D.S.; EGCUT study:

K.F., S.Ka., L.M., N.P.; EPICOR study: G.F., S.G., V.K., G.M., S.P., R.T., P.V.; KORA

study: M.C.K., C.G., H.G., C.H., T.I., J.K., S.Ku., C.M., T.M., A.P., H.P., J.S.R., M.R.,

W.R., K.Sc., K.St., B.T., M.W., S.W.; Leiden Longevity Study: M.B., A.J.M.d.C.,

B.T.H., P.E.S.; LIFELINES study: M.J.B., L.F., P.v.d.H., E.F.T., C.W., A.Z.; LOLIPOP

study: B.A., U.A., C.B, P.A.B., V.B., J.C.C., A.Dr., P.E., M.R.J., S.J., J.S.K., M.A.K., N.K.,

B.L., C.M.L., M.Lo., S.d.L., M.I.M., V.M., Z.Y.M., H.K.N., F.R., M.A.R., J.S., P.S., R.So.,

W.R.S., E.S.T., L.T., S.T., A.R.W., W.Z.; Rotterdam Study: A.De., C.v.D., O.H.F., A.H.,

A.I., J.B.J.v.M., L.S., A.G.U.; TwinsUK study: J.T.B., P.D., J.K.S., T.D.S., P.C.T., T.P.Y.,

WY. Data collection and molecular analyses in isolated cell subsets; adipocytes:

M.A., R.L.B., J.C.C., M.E., M.H., A.J., J.S.K., Z.Y.M., H.K.N., M.A.R., J.S., R.So., W.R.S.,

S.T.; hepatocytes: O.A., M.Br., J.H., C.S., R.Si.; leucocytes: J.F.A., S.L.B., J.C.C.,

J.S.K., M.La., Z.Y.M., H.K.N., G.N., Z.N., M.A.R., R.So., W.R.S., S.T., Y.Y. Data analysis

and writing group; J.C.C., A.Dr., P.E., J.S.K., C.G., H.G., B.L., M.Lo., G.M., M.I.M., J.S.,

W.R.S., S.W.

Author Information Reprints and permissions information is available at

www.nature.com/reprints. The authors declare no competing financial

interests. Readers are welcome to comment on the online version of the

paper. Correspondence and requests for materials should be addressed to

J.C.C. (john.chambers@ic.ac.uk) , H.G. (harald.grallert@helmholtz-muenchen.de)

or J.S.K. (j.kooner@ic.ac.uk).

Reviewer Information Nature thanks M. Boehnke, J. M. Greally, B. Voight and

the other anonymous reviewer(s) for their contribution to the peer review of this

work.

1Research Unit of Molecular Epidemiology, Helmholtz Zentrum München, German Research

Centre for Environmental Health, Neuherberg, Germany. 2Institute of Epidemiology II,

Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg,

Germany. 3German Center for Diabetes Research (DZD), München-Neuherberg, Germany.

4Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford OX3

7BN, UK. 5Department of Epidemiology and Biostatistics, MRC-PHE Centre for Environment

and Health, School of Public Health, Imperial College London, London W2 1PG, UK. 6Institute

of Health Sciences, P.O. Box 5000, FI-90014 University of Oulu, Finland. 7Translational

Laboratory in Genetic Medicine (TLGM), Agency for Science, Technology and Research

(A* STAR), 8A Biomedical Grove, Immunos, Level 5, Singapore 138648, Singapore. 8National

Heart and Lung Institute, Imperial College London, London W12 0NN, UK. 9Department of Twin

Research and Genetic Epidemiology, King’s College London, London SE1 7EH, UK. 10Institute

of Genetic Epidemiology, Helmholtz Zentrum München, German Research Center for

Environmental Health, Neuherberg, Germany. 11Ealing Hospital NHS Trust, Middlesex UB1

3HW, UK. 12Cancer Science Institute of Singapore, National University of Singapore, Singapore.

13Human Genetics Foundation—Torino, Torino, Italy. 14Medical Sciences Department, University

of Torino, Torino, Italy. 15University of Groningen, University Medical Center Groningen,

Department of Genetics, 9700 RB Groningen, The Netherlands. 16Estonian Genome Center,

University of Tartu, Riia 23b, 51010 Tartu, Estonia. 17Department of Biotechnology, Institute of

Molecular and Cell Biology, University of Tartu, Riia 23, 51010 Tartu, Estonia. 18MRC Integrative

Epidemiology Unit (IEU), School of Social and Community Medicine, University of Bristol,

Bristol BS8 2BN, UK. 19UCLH Bariatric Centre for Weight Loss, Weight Management and

Metabolic and Endocrine Surgery, University College London Hospitals, Ground Floor West

Wing, 250 Euston Road, London NW1 2PG, UK. 20Computational Medicine, Faculty of Medicine,

University of Oulu and Biocenter Oulu, Oulu, Finland. 21NMR Metabolomics Laboratory, School

of Pharmacy, University of Eastern Finland, Kuopio, Finland. 22Computational Medicine, School

of Social and Community Medicine, University of Bristol and Medical Research Council

Integrative Epidemiology Unit, University of Bristol, Bristol, UK. 23EPIGET Lab, Department of

Clinical Sciences and Community Health, Università degli Studi di Milano and Fondazione

IRCCS Ca’Granda Ospedale Maggiore Policlinico, Milan, Italy. 24Institute of Human Genetics,

University Hospital Schleswig-Holstein, Kiel Campus, Kiel, Germany. 25Centre for Haematology,

Department of Medicine, Faculty of Medicine, Imperial College London, Hammersmith Campus,

London W12 0NN, UK. 26Molecular Epidemiology, Leiden University Medical Center, Leiden,

2333 ZC, The Netherlands. 27Section of Infectious Diseases and Immunity, Department of

Medicine, Imperial College London, London W12 0NN, UK. 28High Throughput Genomics—

Oxford Genomic Centre, Wellcome Trust Centre for Human Genetics, University of Oxford,

Oxford OX3 7BN, UK. 29Medical Department 1, University Hospital of the Technical University

Dresden, Dresden, Germany. 30Institute for Clinical Diabetology, German Diabetes Center,

Leibniz Center for Diabetes Research at Heinrich Heine University Düsseldorf, Düsseldorf,

Letter

reSeArCH

86 | NATURE | VOL 541 | 5 JANUARY 2017

Histopathology Unit, ‘Civile—M.P. Arezzo’ Hospital, ASP 7, Ragusa, Italy. 55Departments of

Internal Medicine and Epidemiology, Erasmus Medical Centre, Rotterdam, The Netherlands.

56Epidemiology and Public Health, Imperial College London, London, UK. 57Department of

Public Health, Faculty of Medicine, University of Kelaniya, PO Box 6, Thalagolla Road, Ragama

11010, Sri Lanka. 58The Institute of Cancer Research, Surrey SM2 5NG, UK. 59Centre for

Obesity Research, Rayne Institute, Department of Medicine, University College London, London

WC1E 6JJ, UK. 60William Harvey Research Institute, Barts and The London School of Medicine

and Dentistry, Queen Mary University of London, London EC1M 6BQ, UK. 61Princess Al-Jawhara

Al-Brahim Centre of Excellence in Research of Hereditary Disorders (PACER-HD), King

Abdulaziz University, Jeddah 21589, Saudi Arabia. 62Broad Institute of the Massachusetts

Institute of Technology and Harvard University, Cambridge, Massachusetts 02142, USA.

63University of Groningen, University Medical Center Groningen, Department of Cardiology,

9700 RB Groningen, The Netherlands. 64Durrer Center for Cardiogenetic Research,

ICIN—Netherlands Heart Institute, 3511 GC Utrecht, The Netherlands. 65Hannover Unified

Biobank, Hannover Medical School, Feodor-Lynen-Strasse 15, D-30625 Hanover, Germany.

66Institute of Human Genetics, Hannover Medical School, Carl-Neuberg-Strasse 1, D-30625

Hanover, Germany. 67Department of Epidemiology and Biostatistics, MRC Health Protection

Agency (HPE) Centre for Environment and Health, School of Public Health, Imperial College

London, London, UK. 68Biocenter Oulu, P.O. Box 5000, Aapistie 5A, FI-90014 University of Oulu,

Finland. 69Center for Life Course Epidemiology, Faculty of Medicine, P.O. Box 5000, FI-90014

University of Oulu, Finland. 70Unit of Primary Care, Oulu University Hospital, Kajaanintie 50,

PO Box 20, FI-90220 Oulu, 90029 OYS, Finland. 71Department of Pathology, National University

Hospital, Singapore. 72Oxford Centre for Diabetes Endocrinology and Metabolism, University of

Oxford, Oxford, UK. 73Oxford NIHR Biomedical Research Centre, Churchill Hospital, Oxford OX3

7LJ, UK. 74Imperial College Healthcare NHS Trust, London W12 0HS, UK. 75Lee Kong Chian

School of Medicine, Nanyang Technological University, Singapore.

*These authors contributed equally to this work.

‡Deceased.

§These authors jointly supervised this work.

Germany. 31Gerontology and Geriatrics, Leiden University Medical Center, Leiden 2300 RC,

The Netherlands. 32Department of Clinical and Experimental Medicine, University of Surrey,

Guildford GU2 7PX, UK. 33Department of Epidemiology, Erasmus Medical Centre, Rotterdam,

The Netherlands. 34Clinical and Experimental Surgery Department, Medical Research Institute,

University of Alexandria, Hadara, Alexandria 21561, Egypt. 35Department of Endocrinology,

Diabetes and Obesity, Max Healthcare, New Delhi 110 017, India. 36Department of Gene

Diagnostics and Therapeutics, Research Institute, National Center for Global Health and

Medicine, Tokyo 1628655, Japan. 37Epidemiology and Prevention Unit, Fondazione IRCSS

Istituto Nazionale Tumori, Milano, Italy. 38Institute of Human Genetics, Helmholtz Zentrum

München, German Research Center for Environmental Health, Neuherberg, Germany.

39Institute of Human Genetics, Technical University Munich, München, Germany. 40DZHK

(German Centre for Cardiovascular Research), partner site Munich Heart Alliance, Munich,

Germany. 41Vascular Biology Section, National Heart and Lung Institute, Faculty of Medicine,

Imperial College London, London SW3 6LY, UK. 42Dipartmento Di Medicina Clinica E Chirurgia

Federio II University, Naples, Italy. 43Institute for Biometrics and Epidemiology, German

Diabetes Center, Leibniz Center for Diabetes Research at Heinrich Heine University Düsseldorf,

Düsseldorf, Germany. 44Department of Endocrinology and Diabetology, Medical Faculty,

Heinrich Heine University Hospital Düsseldorf, Düsseldorf, Germany. 45Wellcome Trust Sanger

Institute, Wellcome Trust Genome Campus, Hinxton, UK. 46Department of Medical Sciences,

Molecular Medicine and Science for Life Laboratory, Uppsala University, 751 44 Uppsala,

Sweden. 47Department of Visceral and Thoracic Surgery, University Hospital Schleswig-

Holstein, Kiel Campus, Kiel, Germany. 48Institute of Human Genetics, University Hospital of Ulm,

Albert-Einstein-Allee 11, D-89081 Ulm, Germany. 49Department of Internal Medicine, Erasmus

Medical Centre, Rotterdam, The Netherlands. 50Institute of Medical Informatics, Biometry and

Epidemiology, Chair of Genetic Epidemiology, Ludwig-Maximilians-Universität, Munich,

Germany. 51Department of Medicine, Yong Loo Lin School of Medicine, National University of

Singapore, Singapore 119228, Singapore. 52Saw Swee Hock School of Public Health, National

University of Singapore, Singapore 117597, Singapore. 53Duke-National University of

Singapore Graduate Medical School, Singapore 169857, Singapore. 54Cancer Registry and

Letter reSeArCH

METHODS

Population samples. Details of the population samples for discovery and replica-

tion are provided in the Supplementary Information. No statistical methods were

used to predetermine sample size. The experiments were not randomized. All

molecular assays were carried out blind to clinical status.

Quantification of DNA methylation. DNA methylation was quantified in

bisulfite-converted genomic DNA from whole blood, using the Illumina Infinium

HumanMethylation450 array for all samples. Cohort specific methods are

summarized in Supplementary Table 2. DNA methylation was quantified on a scale

of 0–1, in which 1 represents 100% methylation. Preprocessing and quality control

criteria are summarized in Supplementary Table 2.

The association of DNA methylation with BMI (a measure of adiposity) was

tested in each cohort separately by linear regression using an established analytic

strategy to reduce batch effects and other technical confounding effects in

quantification of DNA methylation, and to take the potential confounding effects

arising from cryptic alterations in the white blood cell composition of blood into

account. In brief, in the LOLIPOP and KORA studies, raw signal intensities were

retrieved using the function readIDAT of the R package minfi v.1.6.0 from the

Bioconductor open source software (http://www.bioconductor.org/), followed

by background correction with the function bgcorrect.illumina from the same

R package. Detection P values were derived using the function detectionP as the

probability of the total signal (methylation+ unmethylated) being detected above

the background signal level, as estimated from negative-control probes. Signals

with detection P ≥ 0.01 were removed. Similarly, signals summarized from less

than three functional beads on the chip were removed. Observations with less than

95% CpG sites providing a signal were subsequently excluded from the dataset.

To reduce non-biological variability between observations, data were quantile

normalized with the function normalizeQuantiles of the R package limma v.2.12.0

from Bioconductor, separately in six probe categories based on probe type and

colour channel. If not stated otherwise, this preprocessing pipeline was used for

all data used in downstream analyses.

In order to account for technical effects during the experiment, we performed

principal component analysis on the signal intensities for the 235positive control

probes on the 450k array, which assess multiple steps in the laboratory processing.

The resulting principal components are thought to capture technical variability in

the experiment and the first 20control-probe principal components were included

as covariates in the model to remove technical biases.

To estimate proportions of white-blood-cell types, we used a previously

described method

. They provide 500CpG sites showing the most pronounced

cell-type-specific methylation levels in an experiment based on purified cells.

Of these, 473CpGs were available on the 450k array. Following the proposed

procedure and using the R code provided with the manuscript (R function

projectWBC), we used these 473CpG sites to infer white-blood-cell proportions

(that is, the proportion of granulocytes, monocytes, B cells, CD4+ T cells, CD8+

T cells and natural killer cells) in our samples. These proportions were subsequently

used as covariates in the model to avoid cell-type confounding.

Epigenome-wide association. We per formed single-marker tests separately in each

cohort using linear regression to analyse the association of each autosomal CpG site

with BMI; association results are presented, as the change in BMI per unit change

in methylation (0–1 scale, corresponding to 0–100% change in methylation).

We adjusted for age, gender, smoking status, physical activity index and alcohol

consumption, as well as for the first 20control-probe principal components and

for the estimated white-blood-cell proportions; this set of covariates is henceforth

referred to as ‘discovery covariates’. We corrected the association results for the

genomic control inflation factor (GCin), in order to account for population stratifi-

cation and other forms of cryptic structure in the data, which can arise for instance

from unobserved confounding. Markers on the sex chromosomes were tested

similarly for association with BMI, but separately in men and women. Results

were combined across cohorts by inverse variance meta-analysis using METAL

v.2011-03-25 (http://www.sph.umich.edu/csg/abecasis/Metal/). The resulting

P values where then corrected for in a second round of genomic control (GCout).

There were 466,186autosomal markers for analysis after quality control.

We set the threshold for epigenome-wide significance as P < 1 × 10−7, to provide a

conservative Bonferroni correction for the number of markers tested23.

As additional analyses we also investigated the relationship between BMI and

DNA methylation amongst the 11,233X-chromosomal and 417Y-chromosomal

CpG sites assayed. Our sample size (n = 5,387 individuals) provides 80% power

to identify a change of 8.4 kgm

−2

in BMI per unit increase in methylation (that is,

0–1, in which 1 is 100% methylation) at P < 1.0 × 10−7.

To assess the stability of discovery results towards the analytic choices made, we

performed sensitivity analyses to determine the impact of control-probe principal

components, methylation principal components, and genetic principal components

as covariates. Specifically, we compared results from the discovery meta-analysis

when the first 10, 20, 30 and 40control-probe principal components were included

as covariates; 10 or 20principal components derived from a principal component

analysis on the matrix of methylation β values, 10 or 20principal components

derived from a principal component analysis on the matrix of methylation values

adjusted for the discovery covariates and BMI, or 5principal components der ived

from a principal component analysis on SNP data were included as covariates.

Principal component analysis of the methylation data was performed separately

for each cohort based on quantile normalized β values of autosomal probes without

missing data. Genetic principal components (SNP principal components) were

generated separately for each cohort and genotyping platform (Supplementary

Table 25). The correlation between SNP principal components and methylation

principal components was assessed using linear regression (Supplementary Fig. 9).

Discovery results are very stable towards the considered variations in covariates,

with correlations of effect sizes between the models varying between 0.99–1.0

(Supplementary Figs 3, 10). In addition, SNPs in the probe sequences did not mate-

rially affect the observed associations (Supplementary Fig. 4 and Supplementary

Table 5).

Replication testing. Markers associated with BMI at P < 1 × 10−7 in the discovery

experiment within± 500 kb of each other were considered a single genetic region.

At each locus we identified the CpG sites with lowest P value for association with

BMI (sentinel marker). Our choice of 1 Mb to define a genetic locus was made to

account for long-range enhancers.

At each locus we identified the sentinel marker, and carried out replication

testing in separate samples of whole blood from European and Indian Asian men

and women in population-based studies (n = 4,874, Supplementary Table 1). The

207sentinel CpG sites were assayed using the Illumina 40K methylation array;

cohort-specific details of analysis pipelines are described in Supplement ary Table 2.

Results were combined across discovery and replication by weighted z meta-

analysis. Epigenome-wide significance was set at P < 1 × 10

−7

providing Bonferroni

correction for the 466,186 autosomal markers tested. Our choice of threshold is

supported by the results of permutation testing23. 20 out of 207 markers did not

reach P < 0.05 in replication testing. However, all 20 showed consistent direction

of effect between discovery and replication stages (P = 1.9 × 10−6, binomial test;

Supplementary Table 3), suggesting that the majority are unlikely to be false

positive associations.

To assess whether the 187sentinel CpGs were enriched for intermediately

methylated CpGs (sites with 20–80% average methylation), we randomly generated

100,000sets of 187CpGs and determined the number of intermediately methy lated

CpGs for each of them in order to derive an expected distribution under the null

hypothesis of no enrichment. We then compared the observed number of inter-

mediately methylated CpGs for the 187sentinel CpGs against the nu ll distribution

to calculate an empirical P value.

An exact binomial test (R function binom.test) was used to test whether a

direction of effect between discovery and replication was observed more often

than expected by chance amongst the 20non-replicating CpG sites.

Replication by pyrosequencing. As a technical validation, we used pyrosequencing

to carry out replication testing of the relationship between DNA methylation and

BMI at 4loci, using samples of whole blood from 990Europ eans and 1,720Indian

Asians participating in the LOLIPOP study. Pyrosequencing was carried out using

biotinylated primers to amplify bisulfite-treated DNA (Supplementary Table 26).

The biotinylated PCR products were then immobilized on streptavidin-coated

Sepharose beads (GE Healthcare). Pyrosequencing was performed with the

PyroMark Q96 MGMT kit (Qiagen) on a PSQTM96 MA system (Biotage).

Samples for isolated white blood cell studies. 30obese (BMI > 35 kgm−2) and

30normal weight (BMI < 25 kgm−2) individuals were recruited at random from the

outpatient departments at Ealing and University College London Hospitals. All parti-

cipants gave written informed consent for inclusion in the study (research ethics com-

mittee references: 07/H0712/150, 13/LO/0477 and ID#09/H0715/65). Obese subjects

and normal weight controls were matched by age (within 5 years), sex and ethnicity.

Fluorescence activated cell sorting of white blood cells. For each participant,

we collected 12 ml whole blood (EDTA). Samples were processed immediately

to isolate white blood cell subsets (monocytes, neutrophils, CD4 and CD8

lymphocytes) through red blood cell lysis according to the manufacturer’s instruc-

tions (BioLegend) and staining of unlysed white-blood-cell subsets (> 20 min

in 50 μ l) Ca2+-free PBS with 5 mM EDTA and 1% human albumin; containing

1 μ l anti-CD14 PE-Cy7 (Clone-M5E2, BD), anti-CD16 BV510 (Clone-3G8,

BioLegend), anti-CD45 BV605 (Clone-HI30, BioLegend), anti-CD8 APC

(Clone-SK1, BioLegend); 2 μ l anti-CD3 PE (Clone-Leu-4, BD), anti-CD4 FITC

(Clone-RPA-T4, BioLegend). Stained samples were filtered to remove clumped

cells (30 μ m mesh, Miltenyi Biotec) and dead cells were stained (1 μ l Sytox Blue,

Life Technologies)24,25.

Letter

reSeArCH

Lysed, stained samples were sorted on a FACSAria II SORP cell sorter at a

flow rate of 6,000–9,000 events per second. Data was collected with FACSDiva8

and analysed with FlowJo V10. Fluorescence minus one negative controls (that

is, without the primary labelled antibody of interest) were used to determine

positive and negative boundaries for each gate in the experimental set up. Daily

Cytometer Set-up and Tracking quality control beads were run to ensure alignment

and parameterization of the cell sorter (Anti-Mouse Igκ and Negative Control,

BSA; Compensation Plus Particles, BD). Sytox Blue (450/50V nm) negative events

were considered to be live cells. FCS-A and SSC-A were then used to separate

granulocytes from monocyte and lymphocyte populations. Neutrophils (CD14−,

CD16

) were separated from other granulocytes. Monocytes were then separated

from lymphocytes in a two-stage process as CD14+, CD45+ and CD16− cells.

Finally, CD4

and CD8

cells were separated from other lymphocytes based on the

following staining patterns; CD4

cells: CD3

, CD4

, CD8

−

, CD14

−

and CD45

;

and CD8

cells: CD3

, CD4

−

, CD8

, CD14

−

and CD45

. Sorted cell subsets were

assessed for purity, then pelleted and snap-frozen for storage at − 80 °C. Average

purities were: neutrophils 98.3% (± 1.2% (s.d.)); monocytes 99.2% (± 0.7%); CD4+

lymphocytes 99.6% (± 0.4%); CD8+ lymphocytes 97.9% (± 2.0%).

Genomic DNA was isolated (Qiagen QIAshredder; Allprep DNA/RNA Micro)

according to manufacturer’s instructions. Isolated genomic DNA was quantified

(Qubit double-stranded DNA broad range assay) then stored at − 80 °C for

genome-wide DNA-methylation assays.

Quantification of DNA methylation and data processing. Genomic DNA (0.2–

1.0 μ g) was bisulfite converted using an EZ DNA Methylation-Direct Kit (Zymo

Research). In brief, DNA samples were bisulfite converted by incubation with

the CT conversion reagent for 8 min at 98 °C, 3.5 h at 64 °C, followed by 18 h at

4 °C in a thermocycler. The treated DNA was added to a Zymo-Spin IC Column,

desulfonated using M-desulphonation buffer, and then eluted from the column

in 12 μ l of M-elution buffer.

Methylation analysis of the bisulfite-treated DNA was performed using an

Illumina Infinium MethylationEPIC Beadchip (Illumina) according to the standard

protocol. In brief, 4 μ l of bisulfite-treated DNA was denatured, neutralized and

amplified with an overnight whole-genome amplification reaction. The amplified

DNA was then enzymatically fragmented, precipitated and resuspended in

hybridi zation buffer before being dispensed onto the MethylationEPIC beadchips

for hybridization. After hybridization, the beadchips were processed through a

primer-extension protocol and subsequently stained. Finally, the beadchips were

coated and imaged using the HiScan System (Illumina).

All samples passed quality control and principal component analysis showed

clear separation of cell-types. Methylation values for 179 (out of 187) sentinel

CpGs were retrieved, as described above for epigenome-wide association in blood,

and the difference in DNA methylation between obese cases and normal weight

controls was tested using linear regression, adjusted for age, gender and ethnicity.

Genetic association studies. We used genetic association and the concept of

Mendelian randomization to investigate for potent ial causal relationships between

DNA methylation and adiposity

. In brief, Mendelian randomizat ion goes back to

the more general instrumental variable concept. As an instrumental variable, it uses

a genetic variant (or a combination of genetic variants) Z associated with a variable

X in order to show a causal relation between X and another variable Y. It relies on

the fact that the alleles of a genetic variant are inherited randomly from parents

to offspring, so that the relation of a genetic variant with a phenotype should not

be confounded (with exceptions including population stratification). Thus, if the

effect of X on Y is causal and the study has enough power, Z should also associate

with Y. Specifically, the predicted association of Z with Y can be calculated as

follows, assuming linear relationships and assuming that Z is unrelated to Y given

X and unrelated to any unobserved confounders U (where α is the intercept, and β

and γ are the beta-coefficients for the relationships of X with Z and U, respectively):

αβ γ=+ +XZU

111

where γ1U plays the role of the error term that is per assumption unrelated to Z

αβ γ

αβαβ γγ

αβαβββγγ

αβ γ

=+ +

=+ ++ +

=+ +++

=+ +

YXU

ZU U

()

222

22111 2

22121 21 2

333

Therefore the predicted effect of Z on Y is β3 = β2β1

Unbiased estimation and formal inference on the causal effect β1 of X on Y

(where X and Y represent a CpG–phenotype pair) relies on strong genetic effects

and typically requires tens of thousands of samples for adequate power26. As these

sample sizes are currently not available for epigenomic datasets, we explored the

consistency of the predicted effect of Z on Y compared to the observed effect

instead, thereby obtaining some indication on the plausibility of a causal effect of

X on Y. This was done in two directions, studying causality of the effect of DNA

methylation (X) on BMI (Y) and of BMI (X) on DNA methylation (Y).

DNA methylation as determinant of BMI (causal analysis). To investigate

whether DNA methylation is a determinant of BMI (whereby X = D NA

methylation, Y = BMI) we used data on genetic variants from 4,034participants

of the KORA and LOLIPOP studies (Supplementary Table 25) to identify cis (1 Mb)

SNPs (Z) influencing methylation in blood at the 187sentinel CpG sites. The

associations between SNPs and methylation were tested in each dataset separately

using linear models with methylation as response and SNP as independent

variable, adjusting for the discovery covariates, and then combined by inverse

variance meta-analysis using METAL v.2011-03-25. Our sample size (nmax = 4,034

individuals) provides 80% power to identify a change in methylation of 0.5% (in

absolute terms) per allele copy at P < 5.0 × 10−8 (that is, genome-wide signifi-

cance). Results for all 173,367pairs reaching P < 5 × 10

−8

(conventional genome-

wide significance) are provided in Supplementary Table 27. We excluded 3CpGs

that shared no cis SNPs across all datasets, and a further 9CpGs because they had

SNPs within their probe-binding sequence. For the remaining 175CpG sites, the

single SNP with the lowest P value for association with methylation was chosen

as an instrumental variable (Supplement ary Table 28). As mentioned above, to be

an appropriate instrument, a SNP must not be directly associated with BMI (Y),

only through the respective CpG (X). For this purpose, we removed six CpG–SNP

pairs from the analysis because the corresponding SNPs remained associated with

BMI after adjustment for the sentinel CpG (cg07136133, cg08548559, cg09152259,

cg12484113, cg18120259 and cg26403843). Statistical significance was inferred at

P < 2.9 × 10−4 (corresponding to P < 0.05 after Bonferroni correction for 175tests).

To enable comparison with the observed effect of SNPs on BMI obtained

from published data, we assessed the relationship between DNA methylation and

adiposity in linear models, using an inverse-normal transformation of BMI as the

outcome variable to be consistent with the GIANT genome-wide association study

(GWAS)

. The associations between DNA methylation and inverse-normal trans-

formed BMI were quantified in LOLIPOP and KORA cohorts separately, followed

by inverse variance meta-analysis using METAL v.2011-03-25. We then calcu lated

the predicted effect sizes and standard errors (βpred and SEpred, respectively) as

follows (where the term A~ B refers to the magnitude of the relationship between

A and B in regression analysis):

ββ β=×

pred CpGSNP BMICpG

ββ

=×+×+×

~~~~ ~~

SE SE SE SE

pred CpGSNP

2BMICpG

2CpGSNP

2BMICpG

2CpGSNP

The predicted effect sizes were compared to the observed effects of SNPs on BMI,

whereby the latter were obtained from a large published GWAS to increase power

Statistical significance for individual SNPs was inferred at P < 2.9 × 10

−4

. We used

correlation analysis to investigate the global relationship between predicted and

observed effects on BMI for the SNPs influencing DNA methylation across the

sentinel CpG sites.

DNA methylation as consequence of BMI (consequential analysis). To test the

hypothesis of DNA methylation being a consequence of BMI (whereby X = BMI,

Y = DNA methylation), we followed a similar procedure as described above for the

opposite direction with minor differences.

First, instead of using a single SNP as instrumental variable, we calculated a

weighted genetic risk score (GRS) comprising SNPs reported to influence BMI12.

Again, for the GRS to provide a valid instrument, the included SNPs must not

show direct association with the CpG (Y) but only through BMI (X). For this

purpose we removed three SNPs (rs12444979, rs10968576 and rs7359397) which

remained significantly associated at P < 8.4 × 10−6 (corresponding to P < 0.05 after

Bonferroni correction for the 187 × 32 tests performed) with at least one of the

sentinel CpGs after adjusting for BMI. The final GRS was calculated as the sum of

risk allele dosage of the remaining 29SNPs previously reported to associate with

BMI, weighted by the reported effect sizes12.

Second, the observed effects of GRS on DNA methylation were quantified using

linear models as described above adjusted for the discovery covariates amongst

participants of the KORA and LOLIPOP studies. Regression analysis was carried

out in the KORA and LOLIPOP cohorts separately and results combined by inverse

variance meta-analysis using METAL v.2011-03-25.

DNA methylation in blood and adiposity in prospective population studies.

We used data from the KORA (n = 1,435 Europeans) and LOLIPOP (n = 1513

Indian Asians) studies to examine the prospective, longitudinal association between

DNA methylation at baseline and subsequent change in BMI during follow-up. We

carried out linear regression with change in BMI during follow-up as response

variable, and technically adjusted baseline methylation as the predictor variable,

Letter reSeArCH

with age, sex, physical activity, smoking, alcohol intake, estimated white-blood-cell

proportions and BMI at baseline, as well as follow-up time as additional covariates.

Data were analysed in KORA and LOLIPOP separately, followed by inverse

variance meta-analysis using METAL v.2011-03-25.

We studied the longitudinal relationship between change in BMI and change

in DNA methylation amongst 1,435 participants of the KORA S4/F4 cohort

with methylation data available both at baseline and at the 7-year follow-up

time point. To ensure comparability of methylation measurements from the two

time points measured in two batches, methylation β-values were jointly adjusted

for the first 20principal components obtained from a principal component

analysis on the positive-control probes, and residuals were subsequently used as

adjusted methylation values. Linear models were used with change in BMI during

follow-up as response variable, and change in technically adjusted methylation

as independent variable, including age, sex, physical activity, smoking, alcohol

intake and estimated white-blood-cell proportions both at baseline and follow-up.

DNA methylation in adipose tissue. We investigated whether the observed

methylation markers in blood are representative of BMI-associated methylation

changes in adipose tissue. We used a dataset of 542adipose tissue samples from

the TwinsUK study to test association of the 187identified methylation markers

with BMI. The association of BMI with methylation was quantified using a

linear mixed-effects model adjusting for chip, for bisulfite conversion level and

bisulfite conversion efficiency, smoking state (3categories: current, former and

never smokers), alcohol intake (in g per day) and age, with zygosity and family

as random effects.

We carried out sensitivity analyses to assess the potential contribution of

cryptic structure arising from differences in cell composition of the adipose tissue

samples. In the absence of validated approaches for imputation and adjustment for

adipose-tissue cell-subset composition, and the potential limitations of published

reference-free approaches for separation of true and confounded signals

, we used

principal component analysis to quantify latent structure in the adipose-tissue

methylation data, and included the top 5 components as covariates in the regression

model.

We separately compared DNA methylation between paired samples of blood

and subcutaneous adipose tissue (available for the same n = 201 individuals,

TwinsUK). Blood methylation values were first adjusted for age, chip and chip

position, smoking state, alcohol intake, and estimated white-blood-cell subsets

by taking the residuals from a linear model with these as covariates. Similarly,

adipose tissue methylation values were adjusted for age, chip, bisulfite conversion

level, bisulfite conversion efficiency, smoking state, alcohol intake and the top 5

principal components from the adipose methylation data. Pearson’s correlation

was then determined between the adjusted methylation values.

Finally, we used genetic association to carry out causality analyses on the

association between BMI and DNA methylation in adipose tissue, as described

above for blood. We studied a subset of 325adipose tissue samples from the

TwinsUK cohort with genotype data available. Regression analyses in adipose

tissue between BMI, SNPs/genetic risk score and CpGs were carried out using

the R package lme4, and with smoking, alcohol intake, age, zygosity (random

effect), family (random- effect), beadchip, bisulfite conversion batch and bisulfite

conversion efficiency as covariates.

DNA methylation in isolated adipocytes. Subcutaneous adipose tissue

samples were obtained intraoperatively in 24morbidly obese individuals

(BMI > 40 kgm−2) undergoing laparoscopic bariatric surgery and 24healthy

controls (BMI < 30 kgm

−2

) undergoing non-bariatric laparoscopic abdominal

surgery. Participants were unrelated, between 18–60 years of age, from a multi-

ethnic background, and free from type 2 diabetes. Controls were matched to cases

by age, sex and ethnicity. All participants gave informed consent (Ethics committee

reference 13/LO/0477).

Adipose samples were processed immediately to isolate populations of primary

human adipocyte cells using established protocols

. Polypropylene plastic ware

was used to minimise adipocyte cell lysis. Adipose tissue samples were minced

into 1–2 mm

pieces and washed in Hank’s buffered salt solution (HBSS), before

digestion using type 1 collagenase (1 mgml

−1

, Worthington) in a water bath at

37 °C shaking at 100r.p.m. for approximately 45 min. Digested samples were

filtered through a 300-μ m nylon mesh to remove debris, and the filtered solution

centrifuged at low speed (500g; 5 min; 4 °C), to leave four layers, from top to

bottom: (1) oil, (2) mature adipocytes, (3) supernatant and (4) stromovascular

pellet. After removal of the oil layer, the mature adipocyte layer was collected by

pipette, washed in approximately 5× volume of HBSS and recentrifuged. After

3washes the adipocyte cell suspension was collected for snap-freezing and storage

at − 80C.

Genomic DNA and RNA were extracted from the isolated adipocytes using the

Qiagen AllPrep DNA/RNA/miRNA Universal Kit according to manufacturer’s

protocol for lipid-rich samples. Methylation of genomic DNA was quantified using

the Illumina HumanMethylation450 array in a single batch according to manu-

facturer’s specifications. Raw methylation data were preprocessed using R v.2.15.

Bead intensity was retrieved using the R package minfi v.1.6.0. Marker intensities

were quantile normalized for analysis. Principal component analysis of control

probe intensities was performed to quantify cryptic structure in the data arising

from technical factors. Logistic regression was used to examine the association of

each CpG site with morbid obesity compared to normal weight, adjusting for age,

sex and ethnicity, and the first five control probe principal components.

DNA methylation in liver tissue. Liver samples were obtained percutaneously

for patients undergoing liver biopsy for suspected NAFLD or intraoperatively for

assessment of liver histology. Normal control samples were recruited from samples

obtained for exclusion of liver malignancy during major oncological surgery. None

of the normal control individuals underwent pre-operative chemotherapy and

liver histology demonstrated absence of both cirrhosis and malignancy. Study

design, sampling method and data collection have been described in detail

elsewhere28. For methylation analysis, bisulfite conversion was performed using

the Zymo EZ DNA Methylation Kit (Zymo Research), and hybridization of the

Illumina HumanMethylation450 array (Illumina). mRNA expression analysis was

performed using the HuGene 1.1 ST gene (Affymetrix) according to the manu

facturers’ protocols. Hybridization signals were analysed using GenomeStudio

software (default settings; GenomeStudio v.2011.1, Methylation Analysis Module

v.1.9.0; Illumina Inc) and internal controls for normalization.

Cross-tissue methylation. For extended cross-tissue correlation analyses, publicly

available data (GSE48472) were downloaded from the Gene Expression Omnibus

(GEO) database

. Briefly, the dataset consists of 41samples from six individuals of

blood, liver, muscle, pancreas, subcutaneous fat, omentum and spleen analysed on

the 450k methylation array. Data from the 187CpG sites of interest were extracted

and plotted using the heatmap.2 func tion in the R package gplots (v.2.17.0). Mean

methylation levels for each CpG site across all samples within each tissue type were

used to test for pairwise correlation between tissue types.

Genomic annotation analyses. To test for functional enrichment of the 187CpG

sites associated with BMI, we used annotations of genomic context provided by

Illumina, and of histone modification ChIP peaks (H3K4me1, H3K4me3 and

H3K27Ac, marks of open chromatin) and DNaseI Hypersensitivity Sites in 127

different cell types in the Roadmap and ENCODE (Release 9, UCSC) datasets.

We mapped each probe on the Illumina 450k array background to the annotation

categories and recorded overlap at each probe as a binary variable. To determine

whether enrichment occurred more often than expected by chance, we generated

10,000sets of 187CpGs, each matched with the BMI sentinel CpGs for methylation

mean (± 2%) and standard deviation (± 0.2%), but otherwise selected at random.

For each epigenetic mark, we then calculated the number of overlapping sites

amongst the 187replicating markers (observed) and 10,000permuted sets of

187markers (expected). We calculated the fold enrichment as observed/mean

(expected) and obtained an empirical P value from the distribution of expected.

Gene expression studies. Transcriptome-wide measurements of gene expression

in blood along with measurements of DNA methylation from the same blood

sample were available for participants of both the KORA F4 (n = 703) and

LOLIPOP (n = 1,082, 907 Indian Asians, 175 Europeans) studies (Supplement ary

Table 15). KORA samples were analysed with the Illumina HumanHT-12 v3

BeadChip array. Blood sample collection and RNA isolation and preparation

have been described in detail

29,30

. Gene expression data were quantile normalized

and log2 transformed using the R package lumi, v.2.8.0, from Bioconductor in

R, v.2.14.2. In LOLIPOP, gene expression analysis was performed with the

Illumina HumanHT-12 v4 B eadChip array according to manufacturer’s protocol.

Background correction (using negative controls), quantile normalization and log2

transformation was performed using the R package limma (function neqc).

To examine associations of DNA methylation with gene expression we carried

out linear regression with log2 transformed gene expression as the response

variable and methylation β values as independent variable. In KORA, the model

was adjusted for the discovery covariates and technical covariates related to the

expression measurement (RNA integrity number, RNA amplification plate, sample

storage time). In LOLIPOP, the model was adjusted for age, sex, methylation

control probe principal components and technical covariates related to the expres-

sion measurement (RNA integrity number, RNA extraction batch, RNA conversion

batch, scanning batch, array and array position). Results were analysed in KORA,

LOLIPOP Indian Asians and LOLIPOP Europeans separately, then combined by

inverse-variance meta-analysis using METAL v.2011-03.25. Statistical significance

was inferred at P < 9.0 × 10−6 (that is, P < 0.05 after Bonferroni correction for

5,551CpG–expression pairs).

To assess whether the 187sentinel CpGs were enriched for association with

gene expression, we used the same testing concept as described above based on

Letter

reSeArCH

constructing a null distribution from 10,000randomly selected, matched sets of

187CpGs. For each permuted set we determined the number of significantly asso-

ciated expression probes in cis (P < 9.0 × 10−6) as described above and compare the

resulting distribution with the observed number of gene expression associations

for the 187sentinel CpG sites to calculate an empirical P value.

Finally, we examined the association between DNA methylation and gene

expression in TwinsUK adipose tissue samples (n = 499) for the 44methylation–

expression pairs that were significant in blood. Expression values were adjusted for

age and chip using a linear model. The association of methylation and expression

was then determined in linear mixed-effects models with adjusted expression as

response and methylation as the independent variable, adjusting for age, chip,

bisulfite conversion level and bisulfite conversion efficiency, with zygosity and

family as random effects. After QC filtering of methylation and expression data,

results were available for 36methylation–expression pairs.

Candidate genes and gene-set-enrichment analyses. The standard Illumina

annotation does not identify genes for all CpG sites on the 450K microarray.

We therefore identified candidate genes based on the following criteria: (1) proximity,

gene nearest to the CpG site (n = 187 genes), and (2) gene expression, all local genes

(up to ± 500 kb) with expression associated with the marker at P < 0.05 after

Bonferroni correction for 5,551tests (n = 38genes). This resulted in a list of

210unique genes (Supplementar y Table 19).

Gene annotations were downloaded from Ensembl (https://grch37.ensembl.

org/index.html) using the R package biomaRt v.2.18.0 from Bioconductor, and

overlapped with the cg positions as annotated in the Illumina annotation using the

R package GenomicRanges v.1.14.4 from Bioconductor. We downloaded curated

pathway information (c2.all.v5.0.symbols.gmt) from the GSEA MSigDB plat-

form (http://www.broadinstitute.org/gsea/msigdb), resulting in 1,135pathways,

to investigate enrichment of the set of candidate genes against curated pathway

sets (BIOCARTA, KEGG, REACTOME). An enrichment P value was calculated

empirically on the basis of permutation testing, using the Benjamini–Hochberg

(false-discovery rate) procedure. For the sensitivity analysis, the gene-set-

enrichment analysis was repeated using the genes annotated by Illumina, and using

more permissive proximity criteria (Supplementary Table 29). Results become less

statistically significant when candidate gene selection based on proximity alone

was extended to include all genes over distances up to 500 kb.

DNA methylation and metabolic traits. We investigated the association between

the 187sentinel methylation markers and metabolic disturbances associated with

adiposity amongst participants of the KORA (n = 1,697) and LOLIPOP (n = 2,462)

studies with available measurements of the following BMI-related clinical traits:

LDL (low-density lipoprotein) cholesterol, HDL cholesterol, total cholesterol,

fasting triglycerides, fasting glucose, fasting insulin, HbA1c, systolic and diastolic

blood pressure, C-reactive protein, weight, height and waist–hip ratio. Linear

models were used with trait as response and methylation as independent variable,

adjusting for the discovery covariates. Results from KORA and LOLIPOP studies

were analysed separately, then combined by inverse variance meta-analysis using

METAL v.2011-03-25. Associations were considered significant at P < 2.1 × 10−5

(corresponding to P < 0.05 after Bonferroni correction for 187 × 13tests).

To investigate potential causal relationships between the methylation markers

and BMI-related clinical traits, we performed causality analyses as described above

for the primary phenotype (BMI). For each clinical trait, GWAS datasets of the

most comprehensive meta-analyses published to date with access to genome-wide

association results were retrieved (Supplementary Table 30), to provide SNPs influ-

encing trait. SNPs associated with multiple traits were assigned to the most strongly

associated trait (lowest P value). Clinical traits were transformed as described in the

respective GWAS. Genetic risk scores were calculated as described above for BMI,

after removal of SNPs with direct genomic effects (SNPs that remain associated

with the sentinel CpG after adjustment for the trait). Regression analyses were

carried out in the KORA F4 and LOLIPOP cohorts separately and results were

combined by inverse variance meta-analysis using METAL v.2011-03-25.

Association with incident type 2 diabetes. We tested the association of DNA

methylation at the 187identif ied CpG sites with incident type 2 diabetes amongst

participants of the LOLIPOP study. All participants (n = 2,664) were free from

type 2 diabetes at the time of measurement of DNA methylation; incident

type 2 diabetes (n = 1,074) was defined as either new physician diagnosis, or

HbA1c ≥ 6.5%. Associations with type 2 diabetes were evaluated by logistic

regression adjusted for the discovery covariates. We initially tested the association

in single-marker tests, then in a fully saturated model comprising all 187markers

to identify independent effects.

To combine information across loci, we calculated a weighted methylation risk

score as the sum of the standardised methylation values at each marker that reached

nominal significance (P < 0.05) in the fully saturated multivariate model, weighted

by marker-specific effect size. We then tested the association of the methylation

risk score with incident type 2 diabetes using logistic regression, before and after

adjustment for traditional type 2 diabetes risk factors (BMI, waist–hip ratio,

glucose and HbA1c).

Replication testing of the association of methylation risk score with type 2

diabetes was carried out in a nested case–control study within the KORA S3/S4

comprising 200subjects with newly diagnosed type 2 diabetes and 200controls

matched for age (± 2years), sex, cohort and observation time until diagnosis

of diabetes. Data were analysed using conditional logistic regression using the

function clogit of the R package survival v.2.37.4.

Software. Unless stated otherwise, all calculations were performed using R v.3.0.1.

For all meta-analyses, METAL v.2011-03-25 was used. Custom R code for the

respective analyses is available at: http://metabolomics.helmholtz-muenchen.de/

bmi_methylation/.

Data availability. Summary statistics from the epigenome-wide association

study can be accessed from the European Genome–phenome Archive

(EGAS00001001922; https://www.ebi.ac.uk/ega/studies/EGAS00001001922).

KORA methylation data are available upon request through the application tool

KORA.PASST (http:/epi.helmholtz-muenchen.de); LOLIPOP data are available

from the Gene Expression Omnibus (GSE55763); EPICOR data are deposited in

the HuGeF repository (http://www.hugef-torino.org) and are available on request.

22. Houseman, E. A. et al. DNA methylation arrays as surrogate measures of cell

mixture distribution. BMC Bioinformatics 13, 86 (2012).

23. Lehne, B. et al. A coherent approach for analysis of the Illumina

HumanMethylation450 BeadChip improves data quality and performance in

epigenome-wide association studies. Genome Biol. 16, 37 (2015).

24. Lyons, A. B. & Parish, C. R. Determination of lymphocyte division by ow

cytometry. J. Immunol. Methods 171, 131–137 (1994).

25. Park, D. et al. Noninvasive imaging of cell death using an Hsp90 ligand. J. Am.

Chem. Soc. 133, 2832–2835 (2011).

26. Burgess, S. Sample size and power calculations in Mendelian randomization

with a single instrumental variable and a binary outcome. Int. J. Epidemiol. 43,

922–929 (2014).

27. Spalding, K. L. et al. Dynamics of fat cell turnover in humans. Nature 453,

783–787 (2008).

28. Ahrens, M. et al. DNA methylation analysis in nonalcoholic fatty liver disease

suggests distinct disease-specic and remodeling signatures after bariatric

surgery. Cell Metab. 18, 296–302 (2013).

29. Schurmann, C. et al. Analyzing illumina gene expression microarray data from

dierent tissues: methodological aspects of data analysis in the metaxpress

consortium. PLoS One 7, e50938 (2012).

30. Döring, A. et al. SLC2A9 inuences uric acid concentrations with pronounced

sex-specic eects. Nat. Genet. 40, 430–436 (2008).

Letter reSeArCH

Extended Data Figure 1 | Study design. Epigenome-wide association and

replication testing was performed in order to identify methylation sites

associated with adiposity. In the discovery step, four large cohorts were

included with Illumina 450k DNA methylation data available, which were

preprocessed and quality controlled according to a harmonized protocol.

Epigenome-wide association was performed in every single study with

BMI as response variable and methylation β -value as independent variable,

adjusting for covariates as described in the Methods. At a genome-wide

significance level of P < 1 × 10−7, 278methylation sites from 207regions

were identified. In the replication step, 187 of these were replicated in

independent samples. Genetic association and causality analyses were

used in order to investigate whether the identified methylation signals

underlie the development of adiposity or are the consequence of adiposity.

The findings were supported by longitudinal analyses. The cross-tissue

analyses represent a first step towards extending our observations in

blood to metabolically relevant tissues. The functional genomics and gene

expression analyses help to link the observed methylation associations to

transcriptional outcomes, while the gene-set enrichment analysis provides

a way to summarize the potentially affected metabolic pathways. Finally,

we studied the relationships between methylation and adiposity-related

metabolic traits and type 2 diabetes to address the clinical relevance of our

findings.

Letter

reSeArCH

Extended Data Figure 2 | Distribution of methylation values at the

187sentinel CpG sites compared to the approximately 473,000 CpG

sites assayed by the Illumina Infinium 450K Human Methylation array.

The 187identified methylation–BMI associations are strongly enriched

for CpG sites with intermediate levels of methylation, consistent with the

presence of epigenetic heterogeneity at these loci in blood (157 out of

187 sites with 20–80% methylation, a 3.0-fold enrichment compared to

microarray background, P = 1.4 × 10−22 Fisher’s test).

Letter reSeArCH

Extended Data Figure 3 | DNA methylation at the sentinel CpG sites

in whole blood and in 4isolated cell subsets (monocytes, neutrophils,

CD4+ and CD8+ T cells) from 60individuals (30obese individuals

and 30normal weight controls) by Illumina MethylationEPIC array,

which quantifies 179 of the 187 sentinel markers. Results are shown

as a heatmap, coded by methylation value (hypomethylation < 0.2;

intermediate methylation 0.2–0.8; hypermethylation > 0.8). Results

show the presence of intermediate methylation (and hence epigenetic

heterogeneity) at the majority of loci, and in the majority of cell types, in

both cases and controls.

Letter

reSeArCH

Extended Data Figure 4 | Association of DNA methylation with obesity in the 4cell subsets studied, based on quantification of methylation

at 179sentinel methylation markers in 30obese individuals and 30normal weight controls. Results are presented as QQ plots of the observed

association test statistics in each of the isolated cell subsets. λ, the genomic control inflation factor.

Letter reSeArCH

Extended Data Figure 5 | Comparison of effect sizes between isolated white cell subsets. Results are presented as the difference in methylation

between obese cases and normal weight controls (methylation in cases− methylation in controls, in absolute terms on percentage scale) in the respective

isolated white-blood-cell subset (y axis) compared to the average case− control difference across all 4cell subsets studied (x axis).

Letter

reSeArCH

Extended Data Figure 6 | Mean methylation levels at the 187 sentinel methylation markers associated with BMI, across 7 tissue types. Bottom,

pairwise scatterplots (trend line in red). Top, the Pearson correlation coefficient and P values. Blood, n = 6; liver, n = 5; muscle, n = 6; omentum, n = 6;

pancreas, n = 4; subcutaneous (SC) fat, n = 6; spleen, n = 3.

Letter reSeArCH

Extended Data Figure 7 | Causality analysis in adipose tissue to

investigate the potential relationships between BMI and DNA

methylation. Left, causality analysis in adipose tissue investigating

whether DNA methylation at sentinel CpG sites influences BMI. Units

are change in BMI per copy of effect allele. For each sentinel CpG site, we

determined the effect of a previously identified cis SNP on BMI predicted

through methylation (x axis) and the directly observed effect of a SNP

on BMI (y axis). No CpG passed multiple testing corrections for all three

comparisons. Overall there was little relationship between the effects of

SNPs on BMI predicted through methylation and the directly observed

effect (R = − 0.04, P = 0.58). Right, causality analysis in adipose tissue

investigating whether DNA methylation at sentinel CpG sites is the

consequence of BMI. Units are change in methylation per unit change in

weighted genetic risk score. We identified SNPs reported to influence BMI

in GWAS meta-analysis, and calculated a weighted genetic risk score. For

each sentinel CpG site we then determined the effect of genetic risk score

on methylation predicted via BMI (x axis) and the directly observed effect

of genetic risk score on methylation (y axis). No CpG passed multiple

testing corrections for all three comparisons. The overall correlation

between observed and predicted effects (R = 0.73, P = 1.6 × 10−32)

replicates our findings in blood that methylation at the majority of CpG

sites is consequential to BMI.

Letter

reSeArCH

Extended Data Figure 8 | The 187 sentinel CpGs are enriched for

association with gene-expression in cis in blood. a–d, To derive an

expectation under the null-hypothesis we generated 10,000 sets of

matched CpGs (matched for mean methylation and for s.d. of methylation,

see Methods), and tested their association with expression of the nearest

gene (a), the gene allocated to the CpG by the Illumina annotation

(b), all genes within a 500 kb distance (c) and all genes within a 500 kb

distance excluding the nearest gene (d). We observed significantly more

expression-probes associated with the sentinel markers (red arrow) in

blood compared to the 10,000 permuted sets (green bars).

Letter reSeArCH

Extended Data Figure 9 | Summary statistics for the causality analyses

investigating the relationship between DNA methylation in blood and

metabolic disturbances. a, DNA methylation in blood as a potential

determinant of the metabolic disturbances associated with adiposity

(causal analysis). For each of the sentinel CpG sites, we identified the cis

SNP (1 Mb) most closely associated with DNA methylation levels. For each

of the SNPs, we then determined the effect of SNP on phenotype predicted

via methylation and the directly observed effect of SNP on phenotype.

Results are presented as the R2 between phenotype-specific observed

and predicted effects across the 187CpG sites, calculated using linear

regression. b, DNA methylation in blood as a potential consequence of the

metabolic disturbances associated with adiposity (consequential analysis).

We identified the SNPs reported to influence each phenotypic trait (using

the most recent GWAS meta-analysis; Supplementary Table 24), and

calculated phenotype-specific weighted genetic risk scores. For each of the

CpG sites, and each of the phenotypes, we then determined the effect of

genetic risk score on methylation predicted through phenotype, with the

directly observed effect of genetic risk score on methylation. Results are

presented as the R2 between phenotype-specific observed and predicted

effects across the 187CpG sites, calculated using linear regression.

P values are shown for correlations between observed and predicted

effects that reach P < 0.05.

Letter

reSeArCH

Extended Data Figure 10 | Association of established and emergent

biomarkers with type 2 diabetes. Results are presented as risk of type 2

diabetes associated with the specified biomarkers in three models: model 1,

adjusted for age and sex; model 2, as for model 1, but additionally

adjusted for body mass index and impaired fasting glucose; and model 3,

as for model 2, but additionally adjusted for central obesity and insulin

concentrations. CRP, C-reactive protein; MRS, methylation risk score.

Results for quantitative traits (amino acids, C-reactive protein, insulin

and methylation risk score) are presented as risk of type 2 diabetes in Q4

compared to Q1.

Supplementary Tables

Data

January 2017

Simone Wahl · Alexander Drong · Benjamin Lehne · Marie Loh · John C Chambers

Supplementary Information

Data

January 2017

Simone Wahl · Alexander Drong · Benjamin Lehne · Marie Loh · John C Chambers

Supplementary Information Figure 2

Data

January 2017

Simone Wahl · Alexander Drong · Benjamin Lehne · Marie Loh · John C Chambers

Supplementary Information Figure 1

Data

January 2017

Simone Wahl · Alexander Drong · Benjamin Lehne · Marie Loh · John C Chambers

DNA methylation at the suppressor of cytokine signaling 3 (SOCS3) gene influences height in childhood

Article

Full-text available

Aug 2023

Human height is strongly influenced by genetics but the contribution of modifiable epigenetic factors is under-explored, particularly in low and middle-income countries (LMIC). We investigate links between blood DNA methylation and child height in four LMIC cohorts ( n = 1927) and identify a robust association at three CpGs in the suppressor of cytokine signaling 3 ( SOCS3 ) gene which replicates in a high-income country cohort ( n = 879). SOCS3 methylation ( SOCS3m )—height associations are independent of genetic effects. Mendelian randomization analysis confirms a causal effect of SOCS3m on height. In longitudinal analysis, SOCS3m explains a maximum 9.5% of height variance in mid-childhood while the variance explained by height polygenic risk score increases from birth to 21 years. Children’s SOCS3m is associated with prenatal maternal folate and socio-economic status. In-vitro characterization confirms a regulatory effect of SOCS3m on gene expression. Our findings suggest epigenetic modifications may play an important role in driving child height in LMIC.

Toward Precision Weight-Loss Dietary Interventions: Findings from the POUNDS Lost Trial

Article

Full-text available

Aug 2023

The POUNDS Lost trial is a 2-year clinical trial testing the effects of dietary interventions on weight loss. This study included 811 adults with overweight or obesity who were randomized to one of four diets that contained either 15% or 25% protein and 20% or 40% fat in a 2 × 2 factorial design. By 2 years, participants on average lost from 2.9 to 3.6 kg in body weight in the four intervention arms, while no significant difference was observed across the intervention arms. In POUNDS Lost, we performed a series of ancillary studies to detect intrinsic factors particular to genomic, epigenomic, and metabolomic markers that may modulate changes in weight and other cardiometabolic traits in response to the weight-loss dietary interventions. Genomic variants identified from genome-wide association studies (GWASs) on obesity, type 2 diabetes, glucose and lipid metabolisms, gut microbiome, and dietary intakes have been found to interact with dietary macronutrients (fat, protein, and carbohydrates) in relation to weight loss and changes of body composition and cardiometabolic traits. In addition, we recently investigated epigenomic modifications, particularly blood DNA methylation and circulating microRNAs (miRNAs). We reported DNA methylation levels at NFATC2IP, CPT1A, TXNIP, and LINC00319 were related to weight loss or changes of glucose, lipids, and blood pressure; we also reported thrifty miRNA expression as a significant epigenomic marker related to changes in insulin sensitivity and adiposity. Our studies have also highlighted the importance of temporal changes in novel metabolomic signatures for gut microbiota, bile acids, and amino acids as predictors for achievement of successful weight loss outcomes. Moreover, our studies indicate that biochemical, behavioral, and psychosocial factors such as physical activity, sleep disturbance, and appetite may also modulate metabolic changes during dietary interventions. This review summarized our major findings in the POUNDS Lost trial, which provided preliminary evidence supporting the development of precision diet interventions for obesity management.

Implication of DNA methylation during lifestyle mediated weight loss

Article

Full-text available

Aug 2023

Over the past 50 years, the number of overweight/obese people increased significantly, making obesity a global public health challenge. Apart from rare monogenic forms, obesity is a multifactorial disease, most likely resulting from a concerted interaction of genetic, epigenetic and environmental factors. Although recent studies opened new avenues in elucidating the complex genetics behind obesity, the biological mechanisms contributing to individual’s risk to become obese are not yet fully understood. Non-genetic factors such as eating behaviour or physical activity are strong contributing factors for the onset of obesity. These factors may interact with genetic predispositions most likely via epigenetic mechanisms. Epigenome-wide association studies or methylome-wide association studies are measuring DNA methylation at single CpGs across thousands of genes and capture associations to obesity phenotypes such as BMI. However, they only represent a snapshot in the complex biological network and cannot distinguish between causes and consequences. Intervention studies are therefore a suitable method to control for confounding factors and to avoid possible sources of bias. In particular, intervention studies documenting changes in obesity-associated epigenetic markers during lifestyle driven weight loss, make an important contribution to a better understanding of epigenetic reprogramming in obesity. To investigate the impact of lifestyle in obesity state specific DNA methylation, especially concerning the development of new strategies for prevention and individual therapy, we reviewed 19 most recent human intervention studies. In summary, this review highlights the huge potential of targeted interventions to alter disease-associated epigenetic patterns. However, there is an urgent need for further robust and larger studies to identify the specific DNA methylation biomarkers which influence obesity.

Ribosomal DNA copy number is associated with body mass in humans and other mammals

Preprint

Full-text available

Aug 2023

Body mass results from a complex interplay between genetics and environment. The contribution of genetic variation to body mass has been extensively studied, but due to the technical limitations of platforms used for population scale studies, repetitive parts of the genome have not previously been considered. Here we applied genome-wide approaches to identify an association between adult body mass and the copy number (CN) of 45S-ribosomal DNA (rDNA). rDNA codes for the rRNA components of the ribosome and exists in hundreds of copies/cell in mammals. Inter-individual variation in rDNA CN has previously been documented but has not been associated with a mammalian phenotype. Here, we show that germ-line inherited rDNA CN is associated with post-pubertal growth rate in rats and BMI in adult humans. rDNA CN variation is not associated with rRNA transcription rates in adult tissues, suggesting the mechanistic link occurs earlier in development. This is supported by our observation that it is growth rate and body mass, rather than obesity that are associated with copy number and that these phenotypes emerge prior to adulthood. We present hypotheses for future investigation into the mechanistic basis of this association.

Single-nucleus RNA-sequencing reveals NRF1/NFE2L1 as a key factor determining the thermogenesis and cellular heterogeneity and dynamics of brown adipose tissues in mice

Article

Sep 2023

A Negative Feedback Loop Between TET2 and Leptin in Adipocyte Regulates Body Weight

Preprint

Full-text available

Jul 2023

Ten-eleven translocation (TET) 2 is an enzyme that catalyzes DNA demethylation to regulate gene expression by oxidizing 5-methylcytosine to 5-hydroxymethylcytosine, functioning as an essential epigenetic regulator in various biological processes. However, the regulation and function of TET2 in adipocytes during obesity are poorly understood. In this study, we demonstrate that leptin, a key adipokine in mammalian energy homeostasis regulation, suppresses adipocyte TET2 levels via JAK2-STAT3 signaling. Adipocyte Tet2 deficiency protects against high-fat diet-induced weight gain by reducing leptin levels and further improving leptin sensitivity. By interacting with C/EBPα, adipocyte TET2 increases the hydroxymethylcytosine levels of the leptin gene promoter, thereby promoting leptin gene expression. A decrease in adipose TET2 is associated with obesity-related hyperleptinemia in humans. Inhibition of TET2 suppresses the production of leptin in mature human adipocytes. Our findings support the existence of a negative feedback loop between TET2 and leptin in adipocytes and reveal a novel compensatory mechanism for the body to counteract the metabolic dysfunction caused by obesity.

Salivary Epigenetic Measures of Body Mass Index and Social Determinants of Health Across Childhood and Adolescence

Article

Sep 2023

Importance: Children who are socioeconomically disadvantaged are at increased risk for high body mass index (BMI) and multiple diseases in adulthood. The developmental origins of health and disease hypothesis proposes that early life conditions affect later-life health in a manner that is only partially modifiable by later-life experiences. Objective: To examine whether epigenetic measures of BMI developed in adults are valid biomarkers of childhood BMI and if they are sensitive to early life social determinants of health. Design, setting, and participants: This population-based study of over 3200 children and adolescents aged 8 to 18 years included data from 2 demographically diverse US pediatric cohort studies that combine longitudinal and twin study designs. Analyses were conducted from 2021 to 2022. Exposures: Socioeconomic status, marginalized groups. Main outcome and measure: Salivary epigenetic BMI, BMI. Analyses were conducted to validate the use of saliva epigenetic BMI as a potential biomarker of child BMI and to examine associations between epigenetic BMI and social determinants of health. Results: Salivary epigenetic BMI was calculated from 2 cohorts: (1) 1183 individuals aged 8 to 18 years (609 female [51%]; mean age, 13.4 years) from the Texas Twin Project and (2) 2020 children (1011 female [50%]) measured at 9 years of age and 15 years of age from the Future of Families and Child Well-Being Study. Salivary epigenetic BMI was associated with children's BMI (r = 0.36; 95% CI, 0.31-0.40 to r = 0.50; 95% CI, 0.42-0.59). Longitudinal analysis found that epigenetic BMI was highly stable across adolescence but remained both a leading and lagging indicator of BMI change. Twin analyses showed that epigenetic BMI captured differences in BMI between monozygotic twins. Moreover, children from more disadvantaged socioeconomic status (b = -0.13 to -0.15 across samples) and marginalized racial and ethnic groups (b = 0.08-0.34 across samples) had higher epigenetic BMI, even when controlling for concurrent BMI, pubertal development, and tobacco exposure. Socioeconomic status at birth relative to concurrent socioeconomic status best predicted epigenetic BMI in childhood and adolescence (b = -0.15; 95% CI, -0.20 to -0.09). Conclusion and relevance: This study demonstrated that epigenetic measures of BMI calculated from pediatric saliva samples were valid biomarkers of childhood BMI and may be associated with early-life social inequalities. The findings are in line with the hypothesis that early-life conditions are especially important factors in epigenetic regulation of later-life health. Research showing that health later in life is linked to early-life conditions has important implications for the development of early-life interventions that could significantly extend healthy life span.

Blood-based genome-wide DNA methylation correlations across body-fat- and adiposity-related biochemical traits

Article

Aug 2023
AM J HUM GENET

The recent increase in obesity levels across many countries is likely to be driven by nongenetic factors. The epigenetic modification DNA methylation (DNAm) may help to explore this, as it is sensitive to both genetic and environmental exposures. While the relationship between DNAm and body-fat traits has been extensively studied, there is limited literature on the shared associations of DNAm variation across such traits. Akin to genetic correlation estimates, here, we introduce an approach to evaluate the similarities in DNAm associations between traits: DNAm correlations. As DNAm can be both a cause and consequence of complex traits, DNAm correlations have the potential to provide insights into trait relationships above that currently obtained from genetic and phenotypic correlations. Utilizing 7,519 unrelated individuals from Generation Scotland with DNAm from the EPIC array, we calculated DNAm correlations between body-fat- and adiposity-related traits by using the bivariate OREML framework in the OSCA software. For each trait, we also estimated the shared contribution of DNAm between sexes. We identified strong, positive DNAm correlations between each of the body-fat traits (BMI, body-fat percentage, and waist-to-hip ratio, ranging from 0.96 to 1.00), finding larger associations than those identified by genetic and phenotypic correlations. We identified a significant deviation from 1 in the DNAm correlations for BMI between males and females, with sex-specific DNAm changes associated with BMI identified at eight DNAm probes. Employing genome-wide DNAm correlations to evaluate the similarities in the associations of DNAm with complex traits has provided insight into obesity-related traits beyond that provided by genetic correlations.

Layer‐Specific BTX‐A Delivery to the Gastric Muscularis Achieves Effective Weight Control and Metabolic Improvement

Article

Full-text available

Aug 2023

The rising incidence of health‐endangering obesity constantly calls for more effective treatments. Gastric intramural injection of botulinum neurotoxin A (BTX‐A) as a new modality carries great promise yet inconsistent therapeutic efficacy. A layer‐specific delivery strategy enabled by dissolving microneedles is hence pioneered to investigate the working site of BTX‐A and the resulting therapeutic effects. The drug‐loaded tips of the layer‐specific gastric paralysis microneedles (LGP‐MN) rapidly release and achieve uniform distribution of BTX‐A within the designated gastric wall layers. In an obesity rat model, the LGP‐MNs not only prove safer than conventional injection, but also demonstrate consistently better therapeutic effects with muscular layer delivery, including 16.23% weight loss (3.06‐fold enhancement from conventional injection), 55.20% slower gastric emptying rate, improved liver steatosis, lowered blood lipids, and healthier gut microbiota. Further hormonal study reveals that the elevated production of stomach‐derived glucagon‐like peptide‐1 due to the muscularis‐targeting LGP‐MN treatment is an important contributor to its unique glucose tolerance‐improving effect. This study provides clear indication of the gastric muscularis as the most favorable working site of BTX‐A for weight loss and metabolic improvement purposes, and meanwhile suggests that the LGP‐MNs could serve as a novel clinical approach to treat obesity and metabolic syndromes.

Associations of gene expression in blood with BMI and weight changes among women in the Norwegian Women and Cancer postgenome cohort

Article

Aug 2023
OBESITY

Objective: This study aimed to evaluate associations between blood gene expression profiles and (1) current BMI and (2) past weight changes (WCs) among women who had never been diagnosed with cancer in the Norwegian Women and Cancer (NOWAC) postgenome cohort. Methods: This cross-sectional study (N = 1694) used gene expression profiles and information from three questionnaires: Q1 (baseline), Q2 (follow-up), and Q3 (blood collection). The authors performed gene-wise linear regression models to identify differentially expressed genes (DEGs) and functional enrichment analyses to identify their biological functions. Results: When assessing BMIQ3 , the study observed 2394, 769, and 768 DEGs for the obesity-versus-normal weight, obesity-versus-overweight, and overweight-versus-normal weight comparisons, respectively. Up to 169 DEGs were observed when investigating WCQ3-Q1 (mean = 7 years, range = 5.5-14 years) and WCQ3-Q2 (mean = 1 year, range = <1 month-9 years) in interaction models with BMI categories, of which 1 to 169 genes were associated with WCs and 0 to 9 were associated with interaction effects of BMI and WCs. Biological functions of BMI-associated DEGs were linked to metabolism, erythrocytes, oxidative stress, and immune processes, whereas WC-associated DEGs were linked to signal transduction. Conclusions: Many BMI-associated but few WC-associated DEGs were identified in the blood of women in Norway. The biological functions of BMI-associated DEGs likely reflect systemic impacts of obesity, especially blood reticulocyte-erythrocyte ratio shifts.

Epigenome-wide Association Study (EWAS) of BMI, BMI Change, and Waist Circumference in African American Adults Identifies Multiple Replicated Loci

Article

Full-text available

May 2015
HUM MOL GENET

Obesity is an important component of the pathophysiology of chronic diseases. Identifying epigenetic modifications associated with elevated adiposity, including DNA methylation variation, may point to genomic pathways that are dysregulated in numerous conditions. The Illumina 450K Bead Chip array was used to assay DNA methylation in leukocyte DNA obtained from 2,097 African American adults in the Atherosclerosis Risk in Communities (ARIC) study. Mixed effects regression models were used to test the association of methylation beta value with concurrent BMI and waist circumference (WC), and BMI change, adjusting for batch effects and potential confounders. Replication using whole blood DNA from 2,377 White adults in the Framingham Heart Study and CD4+ T cell DNA from 991 Whites in the GOLDN Study was followed by testing using adipose tissue DNA from 648 women in the MuTHER cohort. Seventy-six (76) BMI-related probes, 164 WC-related probes, and 8 BMI change-related probes passed the threshold for significance in ARIC (p<1 x 10(-7); Bonferroni), including probes in the recently reported HIF3A, CPT1A, and ABCG1 regions. Replication using blood DNA was achieved for 37 BMI probes and 1 additional WC probe. Sixteen (16) of these also replicated in adipose tissue, including 15 novel methylation findings near genes involved in lipid metabolism, immune response/cytokine signaling, and other diverse pathways, including LGALS3BP, KDM2B, PBX1, and BBS2, among others. Adiposity traits are associated with DNA methylation at numerous CpG sites that replicate across studies despite variation in tissue type, ethnicity, and analytic approaches. © The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.

A coherent approach for analysis of the Illumina HumanMethylation450 BeadChip improves data quality and performance in epigenome-wide association studies

Article

Full-text available

Dec 2015
Genome Biol

DNA methylation plays a fundamental role in the regulation of the genome, but the optimal strategy for analysis of genome-wide DNA methylation data remains to be determined. We developed a comprehensive analysis pipeline for epigenome-wide association studies (EWAS) using the Illumina Infinium HumanMethylation450 BeadChip, based on 2,687 individuals, with 36 samples measured in duplicate. We propose new approaches to quality control, data normalisation and batch correction through control-probe adjustment and establish a null hypothesis for EWAS using permutation testing. Our analysis pipeline outperforms existing approaches, enabling accurate identification of methylation quantitative trait loci for hypothesis driven follow-up experiments. Electronic supplementary material The online version of this article (doi:10.1186/s13059-015-0600-x) contains supplementary material, which is available to authorized users.

Benralizumab for chronic obstructive pulmonary disease and sputum eosinophilia: A randomised, double-blind, placebo-controlled, phase 2a study

Article

Full-text available

Sep 2014

Background Chronic obstructive pulmonary disease (COPD) is associated with eosinophilic airway inflammation in 10–20% of patients. Benralizumab, an anti-interleukin-5 receptor α monoclonal antibody, depletes blood and sputum eosinophils. We aimed to establish whether benralizumab reduces acute exacerbations of COPD in patients with eosinophilia and COPD. Methods We did this randomised, double-blind, placebo-controlled, phase 2a study between Nov 18, 2010, and July 13, 2013, at 26 sites in the UK, Poland, Germany, Canada, the USA, Denmark, and Spain. Adults aged 40–85 years, with moderate-to-severe COPD, at least one acute exacerbation of COPD, and a sputum eosinophil count of 3·0% or more within the previous year, were randomly assigned (1:1) via computer-generated permuted block randomisation (block size of four), with an interactive voice or web-response system, to receive placebo or 100 mg benralizumab subcutaneously, every 4 weeks (three doses), then every 8 weeks (five doses) over 48 weeks. Study site personnel included in study assessments, participants, and data analysts, were masked to treatment allocation. The primary endpoint was the annualised rate of acute exacerbations of COPD at week 56, defined as the number of acute exacerbations divided by total duration of person-year follow-up. Secondary and exploratory endpoints included COPD-specific Saint George's Respiratory Questionnaire (SGRQ-C), Chronic Respiratory Questionnaire self-administered standardised format (CRQ-SAS), pre-bronchodilator forced expiratory volume in 1 second (FEV1), and safety. We did a prespecified subgroup analysis by baseline blood eosinophil count. Analyses were by intention to treat and per-protocol. This trial is registered with ClinicalTrials.gov, number NCT01227278. Findings We randomly assigned 101 patients to receive placebo (n=50) or benralizumab (n=51), of whom 88 (87%) patients completed the study. Six patients who completed the study were excluded from the per-protocol population because of major protocol violations; the per-protocol population thus included 82 patients. Benralizumab did not reduce the annualised rate of acute exacerbations of COPD compared with placebo in the per-protocol population, with rates of 0·95 (0·68–1·29; n=40) versus 0·92 (0·67–1·25; n=42). Mean pre-bronchodilator FEV1 change from baseline to week 56 was −0·06 L (SD 0·24) with placebo, and 0·13 L (0·41) with benralizumab (p=0·014). Numerical, albeit non-significant, improvement in acute exacerbations of COPD, SGRQ-C, CRQ-SAS, and FEV1 were greater in benralizumab-treated patients with baseline blood eosinophil concentrations of 200 cells per μL or more or 300 cells per μL or more. Incidence of treatment-emergent adverse events was similar between the two groups, with the most common events being respiratory disorders (31 [62%] of 50 patients given placebo vs 32 [63%] of 51 given benralizumab) and infections (28 [56%] vs 27 [53%]). A higher incidence of serious treatment-emergent adverse events were recorded in patients in the benralizumab group than in those in the placebo group (14 vs nine patients), although none of these events were considered by the investigator to be benralizumab related. Interpretation Compared with placebo, benralizumab did not reduce the rate of acute exacerbations of COPD. However, the results of prespecified subgroup analysis support further investigation of benralizumab in patients with COPD and eosinophilia. Funding MedImmune.

Global, Regional, and National Prevalence of Overweight and Obesity in Children and Adults during 1980–2013: A Systematic Analysis for the Global Burden of Disease Study 2013

Article

Full-text available

Aug 2014

Background: In 2010, overweight and obesity were estimated to cause 3·4 million deaths, 3·9% of years of life lost, and 3·8% of disability-adjusted life-years (DALYs) worldwide. The rise in obesity has led to widespread calls for regular monitoring of changes in overweight and obesity prevalence in all populations. Comparable, up-to-date information about levels and trends is essential to quantify population health effects and to prompt decision makers to prioritise action. We estimate the global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013. Methods: We systematically identified surveys, reports, and published studies (n=1769) that included data for height and weight, both through physical measurements and self-reports. We used mixed effects linear regression to correct for bias in self-reports. We obtained data for prevalence of obesity and overweight by age, sex, country, and year (n=19,244) with a spatiotemporal Gaussian process regression model to estimate prevalence with 95% uncertainty intervals (UIs). Findings: Worldwide, the proportion of adults with a body-mass index (BMI) of 25 kg/m(2) or greater increased between 1980 and 2013 from 28·8% (95% UI 28·4-29·3) to 36·9% (36·3-37·4) in men, and from 29·8% (29·3-30·2) to 38·0% (37·5-38·5) in women. Prevalence has increased substantially in children and adolescents in developed countries; 23·8% (22·9-24·7) of boys and 22·6% (21·7-23·6) of girls were overweight or obese in 2013. The prevalence of overweight and obesity has also increased in children and adolescents in developing countries, from 8·1% (7·7-8·6) to 12·9% (12·3-13·5) in 2013 for boys and from 8·4% (8·1-8·8) to 13·4% (13·0-13·9) in girls. In adults, estimated prevalence of obesity exceeded 50% in men in Tonga and in women in Kuwait, Kiribati, Federated States of Micronesia, Libya, Qatar, Tonga, and Samoa. Since 2006, the increase in adult obesity in developed countries has slowed down. Interpretation: Because of the established health risks and substantial increases in prevalence, obesity has become a major global health challenge. Not only is obesity increasing, but no national success stories have been reported in the past 33 years. Urgent global action and leadership is needed to help countries to more effectively intervene. Funding: Bill & Melinda Gates Foundation.

DNA methylation arrays as surrogate measures of cell mixture distribution

Article

Jan 2012
BMC BIOINFORMATICS

E.A. Houseman

Obesity and Bariatric Surgery Drive Epigenetic Variation of Spermatozoa in Humans

Article

Dec 2015

Obesity is a heritable disorder, with children of obese fathers at higher risk of developing obesity. Environmental factors epigenetically influence somatic tissues, but the contribution of these factors to the establishment of epigenetic patterns in human gametes is unknown. Here, we hypothesized that weight loss remodels the epigenetic signature of spermatozoa in human obesity. Comprehensive profiling of the epigenome of sperm from lean and obese men showed similar histone positioning, but small non-coding RNA expression and DNA methylation patterns were markedly different. In a separate cohort of morbidly obese men, surgery-induced weight loss was associated with a dramatic remodeling of sperm DNA methylation, notably at genetic locations implicated in the central control of appetite. Our data provide evidence that the epigenome of human spermatozoa dynamically changes under environmental pressure and offers insight into how obesity may propagate metabolic dysfunction to the next generation.

Epigenome-wide association of DNA methylation markers in peripheral blood from Indian Asians and Europeans with incident type 2 diabetes: A nested case-control study

Article

Jun 2015

Indian Asians, who make up a quarter of the world's population, are at high risk of developing type 2 diabetes. We investigated whether DNA methylation is associated with future type 2 diabetes incidence in Indian Asians and whether differences in methylation patterns between Indian Asians and Europeans are associated with, and could be used to predict, differences in the magnitude of risk of developing type 2 diabetes. We did a nested case-control study of DNA methylation in Indian Asians and Europeans with incident type 2 diabetes who were identified from the 8-year follow-up of 25 372 participants in the London Life Sciences Prospective Population (LOLIPOP) study. Patients were recruited between May 1, 2002, and Sept 12, 2008. We did epigenome-wide association analysis using samples from Indian Asians with incident type 2 diabetes and age-matched and sex-matched Indian Asian controls, followed by replication testing of top-ranking signals in Europeans. For both discovery and replication, DNA methylation was measured in the baseline blood sample, which was collected before the onset of type 2 diabetes. Epigenome-wide significance was set at p<1 × 10(-7). We compared methylation levels between Indian Asian and European controls without type 2 diabetes at baseline to estimate the potential contribution of DNA methylation to increased risk of future type 2 diabetes incidence among Indian Asians. 1608 (11·9%) of 13 535 Indian Asians and 306 (4·3%) of 7066 Europeans developed type 2 diabetes over a mean of 8·5 years (SD 1·8) of follow-up. The age-adjusted and sex-adjusted incidence of type 2 diabetes was 3·1 times (95% CI 2·8-3·6; p<0·0001) higher among Indian Asians than among Europeans, and remained 2·5 times (2·1-2·9; p<0·0001) higher after adjustment for adiposity, physical activity, family history of type 2 diabetes, and baseline glycaemic measures. The mean absolute difference in methylation level between type 2 diabetes cases and controls ranged from 0·5% (SD 0·1) to 1·1% (0·2). Methylation markers at five loci were associated with future type 2 diabetes incidence; the relative risk per 1% increase in methylation was 1·09 (95% CI 1·07-1·11; p=1·3 × 10(-17)) for ABCG1, 0·94 (0·92-0·95; p=4·2 × 10(-11)) for PHOSPHO1, 0·94 (0·92-0·96; p=1·4 × 10(-9)) for SOCS3, 1·07 (1·04-1·09; p=2·1 × 10(-10)) for SREBF1, and 0·92 (0·90-0·94; p=1·2 × 10(-17)) for TXNIP. A methylation score combining results for the five loci was associated with future type 2 diabetes incidence (relative risk quartile 4 vs quartile 1 3·51, 95% CI 2·79-4·42; p=1·3 × 10(-26)), and was independent of established risk factors. Methylation score was higher among Indian Asians than Europeans (p=1 × 10(-34)). DNA methylation might provide new insights into the pathways underlying type 2 diabetes and offer new opportunities for risk stratification and prevention of type 2 diabetes among Indian Asians. The European Union, the UK National Institute for Health Research, the Wellcome Trust, the UK Medical Research Council, Action on Hearing Loss, the UK Biotechnology and Biological Sciences Research Council, the Oak Foundation, the Economic and Social Research Council, Helmholtz Zentrum Munchen, the German Research Center for Environmental Health, the German Federal Ministry of Education and Research, the German Center for Diabetes Research, the Munich Center for Health Sciences, the Ministry of Science and Research of the State of North Rhine-Westphalia, and the German Federal Ministry of Health. Copyright © 2015 Elsevier Ltd. All rights reserved.

Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index

Article

Jan 2010

DNA Methylation of Lipid-Related Genes Affects Blood Lipid Levels

Article

Epigenome-wide association study of body mass index, and the adverse outcomes of adiposity

Abstract

Supplementary resources (4)