1

0 2 4 6 80

2

4

6

Cholesterol(mM )

PDK

1 m

RN

A(R

elat

ive

expr

essi

on)

r = -0.0765P = 0.6623

0.0 0.5 1.0 1.5 2.00

2

4

6

HDL (mM)

PDK

1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.0521P = 0.7662

0 2 4 60

2

4

6

LDL (mM)

PDK

1 m

RN

A(R

elat

ive

expr

essi

on)

r = -0.0106P = 0.9516

0 5 10 150

1

2

3

4

5

Glucose (mM)

PDK

1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.2827P = 0.0999

0 10 20 30 400

1

2

3

4

5

BMI (kg/m2 )

PDK

1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.0184P = 0.9166

0 40 80 1200

1

2

3

4

5

ALT (U/L)

PDK

1 m

RN

A(R

elat

ive

expr

essi

on)

r = -0.0458P = 0.7939

0 50 100 1500

1

2

3

4

5

AST (U/L)

PDK

1 m

RN

A(R

elat

ive

expr

essi

on)

r = -0.2137P = 0.2178

0 1 2 3 40

1

2

3

4

5

Triglycerides (mM)

PDK

1 m

RN

A(R

elat

ive

expr

essi

on)

r = -0.1272P = 0.4665

0 2 4 6 80

5

10

15

Cholesterol(mM )

DC

YTB

mR

NA

(Rel

ativ

e ex

pres

sion

)

r = 0.4058*P = 0.0141

0.0 0.5 1.0 1.5 2.00

5

10

15

20

HDL (mM)

DC

YTB

mR

NA

(Rel

ativ

e ex

pres

sion

)

r = -0.2204P = 0.2033

0 2 4 60

5

10

15

LDL (mM)

DC

YTB

mR

NA

(Rel

ativ

e ex

pres

sion

)

r = 0.4771**P = 0.0038

0 5 10 150

5

10

15

20

Glucose (mM)D

CYT

B m

RN

A(R

elat

ive

expr

essi

on)

r = 0.1802P = 0.3004

0 10 20 30 400

5

10

15

BMI (kg/m2 )

DC

YTB

mR

NA

(Rel

ativ

e ex

pres

sion

)

r = 0.6558**P < 0.0001

0 40 80 1200

5

10

15

20

ALT (U/L)

DC

YTB

mR

NA

(Rel

ativ

e ex

pres

sion

)

r = 0.7128**P < 0.0001

0 50 100 1500

5

10

15

20

AST (U/L)

DC

YTB

mR

NA

(Rel

ativ

e ex

pres

sion

)

r = 0.5958**P = 0.0002

0 1 2 3 40

5

10

15

20

Triglycerides (mM)

DC

YTB

mR

NA

(Rel

ativ

e ex

pres

sion

)

r = 0.5355**P = 0.0009

0 2 4 6 80

2

4

6

8

10

Cholesterol(mM )

DM

T1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.0074P = 0.9662

0.0 0.5 1.0 1.5 2.00

2

4

6

8

10

HDL (mM)

DM

T1 m

RN

A(R

elat

ive

expr

essi

on)

r = -0.0664P = 0.7046

0 2 4 60

2

4

6

8

10

LDL (mM)

DM

T1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.3274P = 0.0549

0 5 10 150

2

4

6

8

10

Glucose (mM)

DM

T1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.0356P = 0.8392

β-ACTIN

HIF2α

HIF1α

Chow HFDb

Dmt1 Dcytb Epo Fpn Pdk10.0

2.0

4.0

6.0

8.0

Rel

ativ

e ex

pres

sion Chow HFD

*

**

**

**

a

0 10 20 30 400

2

4

6

8

10

BMI (kg/m2 )

DM

T1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.4956**P = 0.0025

0 40 80 1200

2

4

6

8

10

ALT (U/L)

DM

T1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.4513**P = 0.0065

0 50 100 1500

2

4

6

8

10

AST (U/L)

DM

T1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.5499**P = 0.0006

0 1 2 3 40

2

4

6

8

10

Triglycerides (mM)

DM

T1 m

RN

A(R

elat

ive

expr

essi

on)

r = 0.0105P = 0.9522

Nature Medicine: doi:10.1038/nm.4412

2

Supplementary Figure 1. Increased HIF2α signaling in human ileum biopsies is correlated with obesity. (a) Correlative analysis of ileum DMT1, DCYTB, and PDK1 mRNA levels with BMI, ALT, AST, triglycerides, cholesterol, HDL, LDL, and glucose. n = 35. Correlations were assessed by nonparametric Spearman’s test. (b) Western blot analysis of HIF2α and HIF1α protein expression (n = 3/group) and mRNA expression analysis of their target genes in small intestine from chow or HFD-fed mice (1 week). n = 4 for chow and n = 5 for HFD. For box plots, the midline represents the median; box represents the interquartile range (IQR) between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus chow, by two-tailed Student’s t-test.

Nature Medicine: doi:10.1038/nm.4412

3

0 3 6 9 120

20

40

60

Time after HFD (weeks)

Bod

y m

ass(

g)Hif2afl/fl Hif2aΔIE

* ** ** ****** **

*

0 30 60 900

250

500

750

Time (min)

Blo

od g

luco

se (m

g/dl

) Hif2afl/fl Hif2aΔIE

*

***

0 30 60 900

50

100

150

Time (min)

% In

itial

blo

od g

luco

se

Hif2afl/fl Hif2aΔIE

**

*

*

0 3 6 9 120

2020

25

30

35

Time after chow diet (weeks)

Bod

y m

ass(

g)

Hif2afl/fl Hif2aΔIE

Hif2a

fl/fl

Hif2aΔIE

0

40

80

120

ALT

(U/L

)

Hif2a

fl/fl

Hif2aΔIE

0

10

20

30

40

Hep

atic

trig

lyce

rides

(mg/

g of

live

r)

Hif2a

fl/fl

Hif2aΔIE

0

10

20

30

Seru

m tr

igly

cerid

e(m

g/dl

)

Hif2a

fl/fl

Hif2aΔIE

0

1

2

3

Hep

atic

cho

lest

erol

(mg/

g of

live

r)

Hif2a

fl/fl

Hif2aΔIE

0

40

80

120

Ser

um c

hole

ster

ol(m

g/dl

)

Hif2a

fl/fl

Hif2aΔIE

0.0

0.5

1.0

1.5

2.0

Live

r wei

ght (

g)

Hif2a

fl/fl

Hif2aΔIE

0

1

2

3

4

5

Live

r w

eigh

t (%

bod

y m

ass)

Duodenum

Jejunum

IleumColon

Liver

KidneyLung

Heart

SpleenBAT

WATBrai

n0

1

2

3

Hif2α

R

elat

ive

expr

essi

on

Hif2afl/fl Hif2aΔIE

** ** ****

a b

d

Hif2afl/fl

Hif2aΔIE

g hf

k lj m n

i

e

β-ACTIN

HIF2α

HIF1α

Hif2afl/fl Hif2aΔIE

c

Nature Medicine: doi:10.1038/nm.4412

4

Supplementary Figure 2. Lack of intestinal HIF2α prevents HFD-induced obesity and improves metabolic homeostasis. (a) Growth curves of HFD-fed Hif2afl/fl and Hif2a∆IE mice. (b) Glucose tolerance testf HFD-fed Hif2afl/fl and Hif2a∆IE mice. (c) Insulin tolerance testf HFD-fed Hif2afl/fl and Hif2a∆IE mice. (d) Hif2a mRNA expression in different tissues from HFD-fed Hif2afl/fl and Hif2a∆IE mice. (e) Western blot analysis of liver HIF2α and HIF1α from f HFD-fed Hif2afl/fl and Hif2a∆IE mice (n = 3/group). (f) Growth curves of chow-fed Hif2afl/fl and Hif2a∆IE mice. (g) Representative H&E staining of liver sections of chow-fed Hif2afl/fl and Hif2a∆IE mice (n = 15, 3 images/mouse). Scale bars, 100 µm. (h) Liver weights of chow-fed Hif2afl/fl and Hif2a∆IE mice. (i) Liver weight to body weight ratios of chow-fed Hif2afl/fl and Hif2a∆IE mice. (j, k) Liver (j) and serum (k) triglyceride content of chow-fed Hif2afl/fl and Hif2a∆IE mice. (l,m) Liver (l) and serum (m) cholesterol content of chow-fed Hif2afl/fl and Hif2a∆IE mice. (n) Serum ALT levels of chow-fed Hif2afl/fl and Hif2a∆IE mice. n = 5/group. Data are presented as the mean ± sem. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus Hif2afl/fl mice, by two-tailed Student’s t-test.

Nature Medicine: doi:10.1038/nm.4412

5

Hif2a

Dmt1

Dcytb

0.0

0.5

1.0

1.5

2.0

2.5

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

**

**

**

Sptlc1

Sptlc2Kdsr

Cers2Cers4Cers6

Degs1

Degs2

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

*

Smpd1

Smpd2

Smpd3

Smpd4

Enpp7

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

***

**

Sgms1

Sgms2

Acer1

Acer2

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

C16:0

C18:0

C20:0

C22:0

C24:0

C24:1

0.0

0.5

1.0

1.5

2.0

Ser

um g

luco

sylc

eram

ide

rela

tive

abun

danc

e Hif2afl/fl Hif2aΔIE

** ****

*

*

C16:0

C18:0

C18:1

C20:0

C22:0

C24:0

C24:1

0.0

0.5

1.0

1.5

2.0

Ser

um s

phin

gom

yelin

re

lativ

e ab

unda

nce Hif2afl/fl Hif2aΔIE

* *

M5

M4

M3M1

M2

Hif2a∆IE

Hif2afl/fl

a b

c d

e f

g h

Nature Medicine: doi:10.1038/nm.4412

6

Supplementary Figure 3. Loss of HIF2α in the intestine affects ceramide metabolism in HFD-fed Hif2afl/fl and Hif2a∆IE mice. (a) Score scatter plot of a PCA model of the serum metabolites between Hif2afl/fl (circle) and Hif2a∆IE (square) mice. (b) S-plot of an OPLS-DA model of the serum metabolites. (c,d) The relative levels of sphingomyelin (c) and glucosylceramide (d) in serum. (e) Expression of intestinal Hif2a mRNA and HIF2α target gene mRNAs. (f,g) Intestinal expression of mRNAs encoded by ceramide synthesis-related genes, including the de-novo pathway (f) and the sphingomyelinase pathway (g). (h) Intestinal expression of mRNAs encoded by ceramide catabolism-related genes. n = 6/group. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus Hif2afl/fl mice, by two-tailed Student’s t-test.

Nature Medicine: doi:10.1038/nm.4412

7

0

50

100

150

200

EE (k

cal/d

ay/k

g^0.

75) Hif2afl/fl

Hif2aΔIE

8:00

**

16:00 24:00 8:00 0

10

20

30

VO2 (

ml/k

g0.75

/min

) Hif2afl/flHif2aΔIE

**

8:00 16:00 24:00 8:00 0

5

10

15

20

25

VCO

2 (m

l/kg0.

75/m

in) Hif2afl/fl

Hif2aΔIE

**

8:00 16:00 24:00 8:00

0 30 60 900

50

100

150

Time (min)

% In

itial

blo

od g

luco

se Hif2afl/fl

Hif2aΔIE

*

0 30 60 900

100

200

300

400

500

Time (min)B

lood

glu

cose

(mg/

dl)

Hif2afl/fl

Hif2aΔIE

**

Hif2a

fl/fl

Hif2aΔIE

0.0

0.4

0.8

1.2

Live

r wei

ght (

g)

Hif2a

fl/fl

Hif2aΔIE

0

1

2

3

4

5

Live

r w

eigh

t (%

bod

y m

ass)

Hif2a

fl/fl

Hif2aΔIE

0

20

40

60

80

Hep

atic

trig

lyce

rides

(mg/

g of

live

r) *

Hif2a

fl/fl

Hif2aΔIE

0

20

40

60

Seru

m tr

igly

cerid

e(m

g/dl

)

Hif2a

fl/fl

Hif2aΔIE

0

1

2

3

4

Hep

atic

cho

lest

erol

(mg/

g of

live

r)

Hif2a

fl/fl

Hif2aΔIE

0

50

100

150

200

Ser

um c

hole

ster

ol(m

g/dl

)Hif2a

fl/fl

Hif2aΔIE

0

50

100

150

200

250

ALT

(U/L

)

Day Night0

2000

4000

6000

Act

ivity

(Cou

nts)

Hif2afl/fl

Hif2aΔIE

Hif2a

fl/fl

Hif2aΔIE

0

5

10

15

Cum

ulat

ive

food

inta

ke

(kca

l/day

/mou

se)

Hif2a

fl/fl

Hif2aΔIE

0

10

20

30B

ody

mas

s (g

)

f g

a

h

i j k l m n o

b c d

e

Srebp1cCideaCd36

Fabp1

Fabp2

Fabp3

Fabp4Acly

AcacaFasn

Scd1

Elovl6

Dgat1

Dgat2 Pli

n20.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

****

**

** ***

**

*

*

**

Acox1

Acox2Cpt1Cpt2Acsl1Acadl

Ehhadh

Acaa1a

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

Tnfa

Pai1

Ccl2

Ccl3 Il1

b Il6 Il10

0.0

0.5

1.0

1.5

2.0

2.5

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

*

p

rq

Nature Medicine: doi:10.1038/nm.4412

8

Supplementary Figure 4. Inhibition of the intestinal HIF2α substantially increases the metabolic rate and decreases hepatic steatosis independent of body weight changes in HFD-fed Hif2afl/fl and Hif2a∆IE mice. (a) body weight. (b) Glucose tolerance test. (c) Insulin tolerance test. (d) Cumulative food intake. (e) Activity. (f) Energy expenditure. (g) Oxygen consumption rate. (h) Carbon dioxide production rate. (i) Liver weights. (j) Liver weight to body weight ratios. (k,l) Liver (k) and serum (l) triglyceride content. (m,n) Liver (m) and serum (n) cholesterol content. (o) Serum ALT levels (P = 0.08). (p) Hepatic expression of mRNAs encoding fatty acid transport and lipogenesis. (q) Hepatic expression of mRNAs encoding fatty acid oxidation-related enzymes. (r) Hepatic expression of mRNAs encoding inflammatory cytokines and chemokines. n = 6 for Hif2afl/fl group and n = 5 for Hif2a∆IE group. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus Hif2afl/fl mice, by two-tailed Student’s t-test.

Nature Medicine: doi:10.1038/nm.4412

9

Hif2α

Dmt1

Dcytb

0.0

0.5

1.0

1.5

2.0

2.5R

elat

ive

expr

essi

onHif2afl/fl Hif2aΔIE

*

**

**

Sptlc1

Sptlc2Kdsr

Cers2Cers4Cers6

Degs1

Degs2

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

**

Smpd1

Smpd2

Smpd3

Smpd4

Enpp7

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

* * *

Neu1

Neu2

Neu3

Glb1

Gba1

Gba2

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl Hif2aΔIE

****

b

c d

a

Tota

l

C16:0

C18:0

C20:0

C22:0

C24:0

C24:1

0.00.30.6

5

10

15

Ser

um c

eram

ide

(µM

) Hif2afl/fl Hif2aΔIE

*

*

*

*

Ucp1

Ppargc1a

Prdm16

Tmem26

Elovl3

Elovl6

0

1

2

3

4

Rel

ativ

e ex

pres

sion Hif2αfl/fl

Hif2αΔIE**

* * *

Ucp1

Ppargc1a

Prdm16

Elovl3

Cox8b

0

1

2

3

Rel

ativ

e ex

pres

sion Hif2αfl/fl

Hif2αΔIE

Ucp1

Prdm16

Tmem26

Elovl3

Cox8b

0.00.51.01.52.02.5

Rel

ativ

e ex

pres

sion Hif2αfl/fl

Hif2αΔIE

Tota

l

C16:0C18

:0C18

:1C20

:0C22

:0C24

:0

C24:1

0

102030

100

200

300

400

Inte

stin

al c

eram

ide

(nm

ol/g

)

Hif2afl/fl Hif2aΔIE

****

*****

Tota

l

C16:0

C18:0

C20:0

C22:0

C24:0

C24:1

0.00.20.4

5

10

15

Por

tal c

eram

ide

(µM

) Hif2afl/fl Hif2aΔIE

*

** *

* *

*

e f g

Hif2afl/fl

Hif2aΔIE

h l

Hif2afl/fl Hif2a∆IE

eIF5

UCP1

i

j k

Hif2af

l/fl

Hif2aΔ

IE0.00.51.01.52.02.5

UC

P1/

eIF5

(r

el. e

xpre

ssio

n)

*

Nature Medicine: doi:10.1038/nm.4412

10

Supplementary Figure 5. Intestinal HIF2α deficiency reduces ceramide synthesis in the small intestine independent of body weight changes in HFD-fed Hif2afl/fl and Hif2a∆IE mice. (a) Expression of intestinal Hif2a mRNA and HIF2α target gene mRNAs. (b-d) Intestinal expression of mRNAs encoded by ceramide synthesis-related genes, including the de-novo pathway (b), the sphingomyelinase pathway (c), and the salvage pathway (d). (e-g) Ceramide levels in the small intestine (e), systematic serum (f), and portal serum (g). (h-j) Thermogenic gene expression in scWAT (h), BAT (i), and eWAT (j). (k) Western blot analysis of UCP1 protein expression in scWAT (n = 3/group). Data are presented as the mean ± sem. n = 6 for Hif2afl/fl group and n = 5 for Hif2a∆IE group. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus Hif2afl/fl mice, by two-tailed Student’s t-test. (l) Representative UCP1 immunohistochemistry staining of scWAT sections (n = 3 images/mice). Scale, 100 µm.

Nature Medicine: doi:10.1038/nm.4412

11

Supplementary Figure 6. HIF2α regulates the ceramide synthesis in the small intestine. (a) Intestinal expression of Vhl, Hif1a, Hif2a, Dmt1 and Dcytb mRNAs. (b-d) Intestinal expression of mRNAs

Control

siNEU3

0

50

100

150

Eff

icie

ncy

(%)

**

Hif2a

fl/fl

Hif2aΔI

E0.0

0.4

0.8

1.2

Inte

stin

al la

ctos

ylce

ram

ide

C16

:0 r

elat

ive

abun

danc

e

**

DMT1

DCYTB

NEU3

0.0

1.0

2.0

3.0

4.0

5.0

Rel

ativ

e ex

pres

sion

Normoxia + VehicleNormoxia + PT2385Hypoxia (CoCl2) + VehicleHypoxia (CoCl2) + PT2385

**

##

#

**

**

##

Vhlfl/fl

VhlΔIE

Vhl/Hif1afl/fl

Vhl/Hif1aΔ

IE

Vhl/Hif2afl/fl

Vhl/Hif2

aΔIE

0

1

2

3

4

Pdk1

mR

NA

(Rel

ativ

e ex

pres

sion

)

n.s.

n.s.

**** **

Vhlfl/fl

VhlΔIE

Vhl/Hif1afl/f

l

Vhl/Hif1aΔ

IE

Vhl/Hif2afl/f

l

Vhl/Hif2aΔ

IE0

6

12

18

Neu3

mR

NA

(Rel

ativ

e ex

pres

sion

)

n.s.

n.s.

****

**

Vhlfl/fl

VhlΔIE

Vhl/Hif1afl/fl

Vhl/Hif1aΔ

IE

Vhl/Hif2afl/fl

Vhl/Hif2

aΔIE

0.0

0.5

1.0

1.5

Vhl m

RN

A(R

elat

ive

expr

essi

on) **

****

Vhlfl/fl

VhlΔIE

Vhl/Hif1

afl/fl

Vhl/Hif1

aΔIE

Vhl/Hif2

afl/fl

Vhl/Hif2

aΔIE

0.0

0.5

1.0

1.5

2.0

Hif2

a m

RN

A(R

elat

ive

expr

essi

on)

**

n.s. n.s.

Vhlfl/fl

VhlΔIE

Vhl/Hif1

afl/fl

Vhl/Hif1

aΔIE

Vhl/Hif2

afl/fl

Vhl/Hif2

aΔIE

0.0

0.5

1.0

1.5

2.0

Hif1

a m

RN

A(R

elat

ive

expr

essi

on)

n.s.

*

**

Vhlfl/fl

VhlΔIE

Vhl/Hif1afl/fl

Vhl/Hif1aΔ

IE

Vhl/Hif2afl/fl

Vhl/Hif2

aΔIE

0

20

40

60

Dmt1

mR

NA

(Rel

ativ

e ex

pres

sion

) n.s.**

**

n.s.

Vhl

Hif1a

Hif2a

Dmt1

Dcytb

0

10

20

30R

elat

ive

expr

essi

onVhl/Hif1afl/fl

Vhl/Hif1aΔIE

**

**

**

Vhl/Hif1aΔIE + PT2385

##

**

##

Sptlc1

Sptlc2Kdsr

Cers2Cers4Cers6

Degs1

Degs2

0

1

2

3

4

5

Rel

ativ

e ex

pres

sion

Vhl/Hif1afl/fl

Vhl/Hif1aΔIE

**

Vhl/Hif1aΔIE + PT2385

#

Smpd1

Smpd2

Smpd3

Smpd4

Enpp7

0

1

2

3

4

Rel

ativ

e ex

pres

sion

Vhl/Hif1afl/fl

Vhl/Hif1aΔIE

*

Vhl/Hif1aΔIE + PT2385

#*

#

***

##

Neu1

Neu2

Neu3

Glb1

Gba1

Gba2

0

5

10

15

Rel

ativ

e ex

pres

sion

Vhl/Hif1afl/fl

Vhl/Hif1aΔIE

*

Vhl/Hif1aΔIE + PT2385

#

**##

**

##

a

d

e f

b

c

g h

i lj k m

Nature Medicine: doi:10.1038/nm.4412

12

encoded by ceramide synthesis-related genes, including the de-novo pathway (b), the sphingomyelinase pathway (c), and the salvage pathway (d). Male Vhl/Hif1afl/fl and Vhl/ Hif1a∆IE mice fed a chow diet were treated with or without PT2385 (20 mg/kg) for three consecutive days (n = 4 to 6/group). *P < 0.05, **P < 0.01 versus vehicle-treated Vhl/Hif1afl/fl mice, #P < 0.05, ##P < 0.01 versus vehicle-treated Vhl/Hif1a∆IE mice, by one-way ANOVA with Tukey's correction. (e-j) Intestinal mRNA expression levels of Vhl (e), Hif2a (f), Hif1a (g), Dmt1 (h), Pdk1 (i), and Neu3 (j) in Vhlfl/fl, Vhl∆IE, Vhl/Hif1afl/fl, Vhl/Hif1a∆IE, Vhl/Hif2afl/fl, and Vhl/Hif2a∆IE mice fed a chow diet (n = 4 to 6/group). (k) The relative levels of lactosylceramide C16:0 in the small intestine from Hif2afl/fl and Hif2a∆IE mice fed a HFD for 12 weeks (n = 6/group). **P < 0.01 versus Hif2afl/fl mice, by two-tailed Student’s t-test. (l) mRNA expression of DMT1, DCYTB, and NEU3 in HCT116 cells treated with vehicle or PT2385 and exposed to either vehicle or CoCl2 (n = 5/group). **P < 0.01 versus Normoxia + Vehicle treatment, #P < 0.05, ##P < 0.01 versus Hypoxia (CoCl2) + Vehicle treatment, by one-way ANOVA with Tukey's correction. (m) The knockdown efficiency of siNEU3 in HCT116 cells. **P < 0.01, by two-tailed Student’s t-test. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles.

Nature Medicine: doi:10.1038/nm.4412

13

0 1 2 3 40

2020

25

30

35

Time after HFD (weeks)

Bod

y m

ass

(g)

VehicleDANANaringin

*

**

**

**

Srebp1cCideaCd36

Fabp1

Fabp2

Fabp3

Fabp4Acly

AcacaFasnScd1

Elovl6

Dgat1

Dgat2Plin2

0

1

2

3

Rel

ativ

e ex

pres

sion

****

Vehicle DANA Naringin

**** *** *** ****

***

*****

*

**** **

TnfaPai1Ccl2Ccl3 Il1

b Il6 Il10

0

1

2

3

4

Rel

ativ

e ex

pres

sion

**

VehicleDANANaringin

**

* **

Vehicl

eDANA

Naringin

0

50

100

150

Hep

atic

trig

lyce

rides

(mg/

g of

live

r)

****

Vehicl

eDANA

Naringin

0

50

100

150

Seru

m tr

igly

cerid

e(m

g/dl

)

Vehicl

eDANA

Naringin

0

1

2

3

4

Hep

atic

cho

lest

erol

(mg/

g of

live

r)

**

Vehicl

e

DANA

Narin

gin0

50

100

150

200

Ser

um c

hole

ster

ol

(mg/

dl) ****

Vehicl

eDANA

Naringin

0

50

100

150

200A

LT (U

/L) *

**

Vehicl

eDANA

Naringin

0.0

0.5

1.0

1.5

Live

r wei

ght (

g)****

Vehicl

e

DANA

Narin

gin012345

Live

r w

eigh

t (%

bod

y m

ass)

* *

Vehicl

e

DANA

Narin

gin0

1

2

3

4

5In

test

inal

neu

ram

inid

ase

activ

ity (m

U/m

g pr

otei

n

***

Vehicl

eDANA

Naringin

0.0

0.5

1.0

1.5

2.0

2.5

Hep

atic

neu

ram

inid

ase

activ

ity (m

U/m

g pr

otei

n

Vehicl

e

DANA

Narin

gin0.0

0.2

0.4

0.6

0.8

1.0

Whi

te a

dipo

se

neur

amin

idas

e ac

tivity

(m

U/m

g pr

otei

n

Tota

l

C16:0

C18:0

C18:1

C20:0

C22:0

C24:0

C24:1

0

2040

200

400

600

Inte

stin

al c

eram

ide

(nm

ol/g

)

*

**

*

VehicleDANANaringin

** ***

******

Tota

l

C16:0

C18:0

C20:0

C22:0

C24:0

C24:1

0245

10

15

20

Ser

um c

eram

ide

(µM

)

**

**

**

** ** **

**

** ** **

**

VehicleDANANaringin

fe

a b c

Vehicle DANA Naringin

p

g

o

j

h

k l m n

i

d

Nature Medicine: doi:10.1038/nm.4412

14

Supplementary Figure 7. NEU3 inhibitor DANA and naringin treatments protect mice from HFD-induced obesity and hepatic steatosis. (a-c) Neuraminidase activities in intestine (a), liver (b), and white adipose tissue (c). (d,e) Ceramide levels in the small intestine (d) and serum (e). (f) Growth curves. (g) Representative H&E staining of liver sections (n = 3 images/mouse). Scale bars, 100 µm. (h) Liver weights. (i) Liver weight to body weight ratios. (j,k) Liver (j) and serum (k) triglyceride content. (l,m) Liver (l) and serum (m) cholesterol content. (n) Serum ALT levels. (o) Hepatic expression of mRNAs encoding fatty acid transport and lipogenesis. (p) Hepatic expression of mRNAs encoding inflammatory cytokines and chemokines. n = 6/group. Data are presented as the mean ± sem. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus vehicle treatment, by two-tailed Student’s t-test.

Nature Medicine: doi:10.1038/nm.4412

15

Supplementary Figure 8. Administration of ceramide reverses the protective effects of intestinal HIF2α inhibition on the HFD-induced obesity and insulin resistance in Hif2afl/fl and Hif2a∆IE. (a,b) Ceramide levels in the small intestine (a) and serum (b). (c) Growth curves. (d,e) Glucose tolerance test (d) and glucose AUC (e). (f) Insulin tolerance test. n = 5/group. Data are presented as the mean ± sem. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus vehicle-treated Hif2afl/fl mice, #P < 0.05, ##P < 0.01 versus vehicle-treated Hif2a∆IE mice, by one-way ANOVA with Tukey’s correction.

0 25 50 75 1000

40

80

120

Time (min)

% In

itial

blo

od g

luco

se

Hif2afl/fl+ VehicleHif2aΔIE+ VehicleHif2aΔIE+ Ceramide

*

#

* *

##

0 20 40 60 80 1000

100

200

300

400

500

Time (min)

Blo

od g

luco

se (m

g/dl

)

Hif2afl/fl + VehicleHif2aΔIE + VehicleHif2aΔIE + Ceramide

*

#

**

*

#

Tota

l

C16:0

C18:0

C18:1

C20:0

C22:0

C24:0

C24:1

0

30

60100200300400500

Inte

stin

al c

eram

ide

(nm

ol/g

)Hif2afl/fl + VehicleHif2aΔIE + VehicleHif2aΔIE + Ceramide**

##

**##

**## *

## *#

*#

Tota

l

C16:0

C18:0

C20:0

C22:0

C24:0

C24:1

0245

15

25

Ser

um c

eram

ide

(µM

)

*

Hif2afl/fl + VehicleHif2aΔIE + VehicleHif2aΔIE + Ceramide

##

*#

*#

**##

*##

#

*

Hif2afl/fl +

Vehicl

e

Hif2aΔ

IE + Veh

icle

Hif2aΔ

IE + Cera

mide0

1×104

2×104

3×104

AU

C (m

g/dl×m

in)

**#

a

d

b

e f

c

0 2 4 60

2020

30

40

50

Time after HFD (weeks)

Bod

y m

ass

(g)

Hif2afl/fl + VehicleHif2aΔIE + VehicleHif2aΔIE + Ceramide

* ** ** **## #

Nature Medicine: doi:10.1038/nm.4412

16

0 3 6 9 120

2020

30

40

50

Time after HFD (weeks)

Bod

y m

ass

(g)

Hif2afl/fl + Vehicle Hif2afl/fl + PT2385 Hif2aΔIE + PT2385

* or **

Hif2aΔIE + Vehicle

0 20 40 60 80 1000

40

80

120

Time (min)

% In

itial

blo

od g

luco

se

Hif2afl/fl + VehicleHif2afl/fl + PT2385 Hif2aΔIE + PT2385

Hif2aΔIE + Vehicle

* or **

0 20 40 60 80 1000

250

500

750

Time (min)

Blo

od g

luco

se (m

g/dl

)

Hif2afl/fl + VehicleHif2afl/fl + PT2385 Hif2aΔIE + PT2385

Hif2aΔIE + Vehicle

* or **

Tota

l

C16:0

C18:0

C20:0

C22:0

C24:0

C24:1

0

5

10

15

20

25

Ser

um c

eram

ide

(µM

)

Hif2afl/fl + VehicleHif2afl/fl + PT2385

**

**

**

** ****

Hif2aΔIE + VehicleHif2aΔIE + PT2385

**

** ** **

****

Tota

l

C16:0

C18:0

C18:1

C20:0

C22:0

C24:0

C24:1

0

50

100

200

400

600

Inte

stin

al c

eram

ide

(nm

ol/g

) Hif2afl/fl + VehicleHif2afl/fl + PT2385

**

*

**

Hif2aΔIE + VehicleHif2aΔIE + PT2385

**

*

** ********

****

Hif2afl/fl

+ Veh

icle

Hif2afl/fl

+ PT23

85

Hif2aΔ

IE + Veh

icle

Hif2aΔ

IE + PT23

850

100

200

300

Hep

atic

trig

lyce

rides

(mg/

g of

live

r)

**

Hif2afl/fl

+ Veh

icle

Hif2afl/fl

+ PT23

85

Hif2aΔ

IE + Veh

icle

Hif2aΔ

IE + PT23

850

50

100

150

200

250

Seru

m tr

igly

cerid

e(m

g/dl

)

Hif2afl/fl

+ Veh

icle

Hif2afl/fl

+ PT23

85

Hif2aΔ

IE + Veh

icle

Hif2aΔ

IE + PT23

850

1

2

3

4

Hep

atic

cho

lest

erol

(mg/

g of

live

r)

**

Hif2afl/f

l + Veh

icle

Hif2afl/f

l + PT23

85

Hif2aΔ

IE + V

ehicl

e

Hif2aΔ

IE + P

T2385

0

50

100

150

200

Ser

um c

hole

ster

ol

(mg/

dl)

***

Hif2afl/fl

+ Veh

icle

Hif2afl/fl

+ PT23

85

Hif2aΔ

IE + Veh

icle

Hif2aΔ

IE + PT23

850

100

200

300

ALT

(U/L

)

**

Hif2afl/fl

+ Veh

icle

Hif2afl/fl

+ PT23

85

Hif2aΔ

IE + Veh

icle

Hif2aΔ

IE + PT23

850.0

0.6

1.2

1.8

Live

r wei

ght (

g)

** **

Hif2afl/f

l + Veh

icle

Hif2afl/f

l + PT23

85

Hif2aΔ

IE + V

ehicl

e

Hif2aΔ

IE + P

T2385

0

1

2

3

4

Live

r w

eigh

t (%

bod

y m

ass) * *

a c

f

g h i

d e

l m

Hif2afl/fl + Vehicle Hif2aΔIE + Vehicle

Hif2afl/fl + PT2385 Hif2aΔIE + PT2385

j k

b

Nature Medicine: doi:10.1038/nm.4412

17

Supplementary Figure 9. PT2385 prevents mice from HFD-induced obesity and hepatic steatosis through inhibition of the intestinal HIF2α-ceramide axis in Hif2afl/fl and Hif2a∆IE mice. (a) Growth curves. (b) Glucose tolerance test. (c) Insulin tolerance test. (d) Representative H&E staining of liver sections (n = 3 images/mouse). Scale bars: 100 µm. (e) Liver weights. (f) Liver weight to body weight ratios. (g,h) Liver (g) and serum (h) triglyceride content. (i,j) Liver (i) and serum (j) cholesterol content. (k) Serum ALT levels. (l) Quantitation of ceramide concentrations in the intestine. (m) Quantitation of ceramide concentrations in serum. n = 5/group. Data are presented as the mean ± sd. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus vehicle-treated Hif2afl/fl mice, by one-way ANOVA with Tukey’s correction.

Nature Medicine: doi:10.1038/nm.4412

18

Neu1

Neu2

Neu3

Glb1Gba1

Gba20

1

2

3

Rel

ativ

e ex

pres

sion

Hif2afl/fl + VehicleHif2afl/fl + PT2385Hif2aΔIE + VehicleHif2aΔIE + PT2385

**** *** *

TnfaPai1 Ccl2 Ccl3 Il1

b Il6 Il10

0

1

2

3

Rel

ativ

e ex

pres

sion

Hif2afl/fl + VehicleHif2afl/fl + PT2385Hif2aΔIE + VehicleHif2aΔIE + PT2385

* * ** ** * **

Srebp1c

CideaCd36

Fabp1

Fabp2

Fabp3

Fabp4

Acly

Acaca

Fasn

Scd1

Elovl6Dgat1

Dgat2 Plin2

0

1

2

3

Rel

ativ

e ex

pres

sion

Hif2afl/fl + VehicleHif2afl/fl + PT2385

Hif2aΔIE + VehicleHif2aΔIE + PT2385

** #

**

*

**

**** * * *

* *** *

******** **

Sptlc1

Sptlc2

Kdsr

Cers2

Cers4

Cers6

Degs1

Degs2

0

1

2

3

Rel

ativ

e ex

pres

sion

Hif2afl/fl + VehicleHif2afl/fl + PT2385

*

Hif2aΔIE + VehicleHif2aΔIE + PT2385

*

Smpd1

Smpd2

Smpd3

Smpd4

Enpp70.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl + VehicleHif2afl/fl + PT2385

*

Hif2aΔIE + VehicleHif2aΔIE + PT2385

***

***

***

Hif2a

Dmt1Dcy

tb0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Hif2afl/fl + VehicleHif2afl/fl + PT2385

**

*

Hif2aΔIE + VehicleHif2aΔIE + PT2385

*

**

*

a b

c d

e

f

Nature Medicine: doi:10.1038/nm.4412

19

Supplementary Figure 10. PT2385 inhibits ceramide synthesis in the small intestine and alters fatty acid synthesis, metabolism, and inflammation in the liver dependent on intestinal HIF2α in Hif2afl/fl and Hif2a∆IE mice. (a) Expression of Hif2a mRNA and its target gene mRNAs in the intestine. (b-d) Intestinal expression of mRNAs encoded by ceramide synthesis-related genes, including the de-novo pathway (b), the sphingomyelinase pathway (c) and the salvage pathway (d). (e) Hepatic expression of mRNAs encoding fatty acid transport and lipogenesis. (f) Hepatic expression of mRNAs encoding inflammatory cytokines and chemokines. n = 5/group. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus vehicle-treated Hif2afl/fl mice, by one-way ANOVA with Tukey’s correction.

Nature Medicine: doi:10.1038/nm.4412

20

0 1 2 3 40

3030

40

50

60

Time after HFD (weeks)

Bod

y m

ass(

g)Vehicle PT2385

**

Srebp1cCideaCd36

Fabp1

Fabp2

Fabp3

Fabp4Acly

AcacaFasnScd1

Elovl6

Dgat1

Dgat2Plin2

0.0

0.6

1.2

1.8

Rel

ativ

e ex

pres

sion

Vehicle PT2385

**

**

****

****

*

*

TnfaPai1Ccl2Ccl3 Il1

b Il6 Il10

0

1

2

3

Rel

ativ

e ex

pres

sion

Vehicle PT2385

*

*

** **

Sptlc1Sptlc2KdsrCers2Cers4Cers6Degs1Degs2

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Vehicle PT2385

*

Smpd1

Smpd2

Smpd3

Smpd4

Enpp7

0.0

0.4

0.8

1.2

1.6

Rel

ativ

e ex

pres

sion

Vehicle PT2385P=0.055

****

**

Dmt1

Dcytb

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pres

sion

Vehicle PT2385

**

Tota

l

C16:0

C18:0

C20:0

C22:0

C24:0

C24:1

0

5

10

15

20

Ser

um c

eram

ide

(µM

)

Vehicle PT2385

**

*

**

**

a

i

d e

f g

h

0 30 60 900

250

500

750

Time (min)

Blo

od g

luco

se (m

g/dl

) Vehicle PT2385

**

*****

b

0 30 60 900

50

100

150

Time (min)

% In

itial

blo

od g

luco

se Vehicle PT2385

****

**

c

Nature Medicine: doi:10.1038/nm.4412

21

Supplementary Figure 11. PT2385 reverses metabolic dysfunctions in HFD-induced obese mice. (a) Growth curves. (b) glucose tolerance test. (c) Insulin tolerance test. (d) Expression levels of HIF2α target gene mRNAs in the intestine. (e) Quantitation of ceramide concentrations in serum. (f,g) Intestinal expression of mRNAs encoded by ceramide synthesis-related genes, including the de novo pathway (f) and the sphingomyelinase pathway (g). (h) Hepatic expression of mRNA encoding fatty acid transport and lipogenesis-related enzymes. (i) Hepatic expression of mRNAs encoding inflammatory cytokines and chemokines. n = 4 for vehicle group, n = 5 for PT2385 group. Data are presented as the mean ± sem. For box plots, the midline represents the median; box represents the IQR between the first and third quartiles, and whiskers represent the lowest or highest values within 1.5 times IQR from the first or third quartiles. *P < 0.05, **P < 0.01 versus vehicle treatment, by two-tailed Student’s t-test.

Nature Medicine: doi:10.1038/nm.4412

22

a b

50 kD

25 kD

Chow HFD 8 weeksHIF2α

100 kD

150 kD

75 kD

250 kD

50 kD

37 kD

25 kD

75 kD

Ctrl

100 kD

150 kD

75 kD

250 kD

50 kD

37 kD

β-ACTIN

HIF1α

NEU3

37 kD

β-ACTIN

HIF2α

130 kD

70 kD

100 kD

Non-obese Obese

130 kD

70 kD100 kD

HIF1α

35 kD40 kD

55 kD

β-ACTIN

170 kD

25 kD

Nature Medicine: doi:10.1038/nm.4412

23

130 kD

72 kD

95 kD

Chow HFD 1 weekHIF2α

130 kD

72 kD

95 kD

CtrlHIF1α

β-ACTIN55 kD

34 kD

43 kD

c d

β-ACTIN

Hif2afl/fl Hif2aΔIEHIF2α

HIF1α

100 kD

150 kD

75 kD

50 kD

37 kD

100 kD

150 kD

75 kD

50 kD

37 kD

100 kD

150 kD

75 kD

50 kD

37 kD

Nature Medicine: doi:10.1038/nm.4412

24

e f

100 kD

150 kD

75 kD

50 kD

37 kD

250 kD

Hif2afl/fl Hif2aΔIE

100 kD

150 kD

75 kD

50 kD

37 kD

250 kD

100 kD

150 kD

75 kD

50 kD

37 kD

250 kD

Vehicle PT2385

100 kD

150 kD

75 kD

50 kD

37 kD

250 kDβ-ACTIN β-ACTIN

NEU3 NEU3

eIF5

UCP1

55 KD

40 KD

70 KD

Hif2afl/fl Hif2aΔIE

35 KD

25 KD

15 KD

g

Nature Medicine: doi:10.1038/nm.4412

25

Supplementary Figure 12. Full western blot gel panels. (a-d) HIF2α, HIF1α, NEU3 and β-ACTIN from which the data in Figure 1b (a), Figure 1f (b), Supplementary Figure 1b (c), Figure 4g (b) and Supplementary Figure 2e (d) were derived. (e, f) NUE3 and β-ACTIN from which the data in Figure 4h (e) and Figure 6k (f) were derived. (g) UCP1 and EIF5 from which the data in Supplementary Figure 4v were derived.

Nature Medicine: doi:10.1038/nm.4412

26

Supplementary Table 1. Demographic characteristics of the subjects

Characteristics Non-obese Obese

Cohort 1 (n = 12)

Gender, n (%)

Male 4 (66.7) 4 (66.7)

Female 2 (33.3) 2 (33.3)

Age, y

Mean (sem) 49.7 (4.8) 50.2 (6.0)

Range 29.0-61.0 23.0-63.0

Body weight, kg

Mean (sem) 59.5 (2.4) 76.0 (3.2) **

Range 51.0-67.0 70.0-91.0

Body mass index, kg/m2

Mean (sem) 21.5 (0.6) 28.5 (0.8) **

Range 19.8-23.9 26.0-31.1

Cohort 2 (n = 35) Gender, n (%)

Male 9 (50.0) 7 (41.2)

Female 9 (50.0) 10 (58.8)

Age, y

Mean (sem) 57.4 (3.2) 54.7 (2.2)

Range 23.0-73.0 36.0-69.0

Body weight, kg

Mean (sem) 55.7 (1.2) 82.1 (1.5) **

Range 49.0-65.0 72.0-90.0

Body mass index, kg/m2

Mean (sem) 20.4 (0.6) 29.8 (0.3) **

Range 15.8-25.0 28.1-32.8

**P < 0.01 versus Non-obese, by two-tailed Student’s t-test.

Nature Medicine: doi:10.1038/nm.4412

27

Supplementary Table 2. Clinical biochemistry of the subjects

Laboratory analytes Non-obese Obese P value

Cohort 1 (n = 12)

Liver enzymes, U/L

ALT 13.2 (1.3) 25.0 (4.2) 0.0222

AST 17.5 (2.7) 30.3 (4.0) 0.0241

Lipids, mM

Triglycerides 1.57 (0.33) 2.67 (0.44) 0.0755

Cholesterol 4.14 (0.46) 4.79 (0.32) 0.2695

HDL 1.25 (0.14) 0.99 (0.21) 0.3359

LDL 2.07 (0.45) 3.29 (0.36) 0.0615

Glucose, mM 5.35 (0.27) 5.39 (0.37) 0.9352

Cohort 2 (n = 35)

Liver enzymes, U/L

ALT 19.6 (1.4) 38.2 (4.2) 0.0001

AST 19.0 (1.4) 41.4 (4.8) 0.0001

Lipids, mM

Triglycerides 1.45 (0.16) 1.79 (0.12) 0.1045

Cholesterol 3.63 (0.10) 4.22 (0.25) 0.0352

HDL 1.09 (0.04) 1.01 (0.06) 0.2630

LDL 1.99 (0.09) 2.67 (0.19) 0.0027

Glucose, mM 5.42 (0.18) 6.13 (0.36) 0.0798 Data are presented as mean (sem). Two-tailed Student’s t-test.

Nature Medicine: doi:10.1038/nm.4412

28

Supplementary Table 3. Primer list

Mouse primers Sequence β-Actin FWD 5'- GGCTGTATTCCCCTCCATCG -3' β-Actin REV 5'- CCAGTTGGTAACAATGCCATGT -3' Hif1a FWD 5'- ATAGCTTCGCAGAATGCTCAGA -3' Hif1a REV 5'- CAGTCACCTGGTTGCTGCAA -3' Hif2a FWD 5'- TGAGTTGGCTCATGAGTTGC -3' Hif2a REV 5'- TATGTGTCCGAAGGAAGCTG -3’ Dmt1 FWD 5'- TGTTTGATTGCATTGGGTCTG -3’ Dmt1 REV 5'- CGCTCAGCAGGACTTTCGAG -3’ Dcytb FWD 5'- CATCCTCGCCATCATCTC -3’ Dcytb REV 5'- GGCATTGCCTCCATTTAGCTG -3’ Pdk1 FWD 5'- TTACTCAGTGGAACACCGCC -3’ Pdk1 REV 5'- GTTTATCCCCCGATTCAGGT -3’ Vhl FWD 5'- ACATCGTCAGGTCACTCTATGA -3’ Vhl REV 5'- CTCTTGGCTCAGTCGCTGTAT -3’ Sptlc1 FWD 5'- CGAGGGTTCTATGGCACATT-3' Sptlc1 REV 5'- GGTGGAGAAGCCATACGAGT -3' Sptlc2 FWD 5'- TCACCTCCATGAAGTGCATC -3' Sptlc2 REV 5'- CAGGCGTCTCCTGAAATACC -3’ Kdsr FWD 5'- TCCAGTGGCATTGGGAAGTG -3’ Kdsr REV 5'- CTTCTCTTGTGCCTGCTTTATGA -3’ Degs1 FWD 5'- AATGGGTCTACACGGACCAG -3’ Degs1 REV 5'- TGGTCAGGTTTCATCAAGGAC -3’ Degs2 FWD 5'- AAGCCAATGGACCACAAACT -3’ Degs2 REV 5'- TGCTTGGAGAGCCCTTCTAAT -3’ Cers2 FWD 5'- AAGTGGGAAACGGAGTAGCG-3’ Cers2 REV 5'- ACAGGCAGCCATAGTCGTTC -3’ Cers4 FWD 5'- GGATTAGCTGATCTCCGCAC -3’ Cers4 REV 5'- CCAGTATGTCTCCTGCCACA -3’ Cers6 FWD 5'- AAGCCAATGGACCACAAACT -3’ Cers6 REV 5'- TGCTTGGAGAGCCCTTCTAAT -3’ Smpd1 FWD 5'- GTTACCAGCTGATGCCCTTC -3’ Smpd1 REV 5'- AGCAGGATCTGTGGAGTTG -3’ Smpd2 FWD 5'- AGCAGGATCTGTGGAGTTG -3’ Smpd2 REV 5'- CTCCAGCCATGAAGCTCAAC -3’ Smpd3 FWD 5'- CCTGACCAGTGCCATTCTTT -3’ Smpd3 REV 5'- AGAAACCCGGTCCTCGTACT -3’

Nature Medicine: doi:10.1038/nm.4412

29

Smpd4 FWD 5'- ACCTGGCCCTCAATCCATTTG -3’ Smpd4 REV 5'- ATAGGCACAGTCCGAAGTACG -3’ Enpp7 FWD 5'- AAGCCCAGTATATGACTCCTGC -3’ Enpp7 REV 5'- ACCGTGCTGGTGGTATTGTAG -3’ Neu1 FWD 5'- GGACCGCTGAGCTATTGGG -3’ Neu1 REV 5'- CGGGATGCGGAAAGTGTCTA -3’ Neu2 FWD 5'- CACAGGCGTCCATGCTTACA -3’ Neu2 REV 5'- CTGCGTGCTCATCCGTCTT -3’ Neu3 FWD 5'- ATGGAGGCCACATTACCTGG -3’ Neu3 REV 5'- TCTGGCACCTCTCAGTAACAT -3’ Glb1 FWD 5'- GCACGGCATCTATAATGTCACC -3’ Glb1 REV 5'- GTATCGGAATGGCTGTCCATC -3’ Gba1 FWD 5'- GCCAGGCTCATCGGATTCTTC -3’ Gba1 REV 5'- CACGGGGTCAAGAGAGTCAC -3’ Gba2 FWD 5'- GGCTGTGCCGAAAGAGATTC -3’ Gba2 REV 5'- ATCCTGGGGTCCACTATCCTC -3’ Galc FWD 5'- CGCCTACGTGCTAGACGAC -3’ Galc REV 5'- ACGATAGGGCTCTGGGTAATTT -3’ Srebp1c FWD 5'- GGAGCCATGGATTGCACATT-3' Srebp1c REV 5'- GCTTCCAGAGAGGAGGCCAG -3' Cidea FWD 5'- TGACATTCATGGGATTGCAGAC -3' Cidea REV 5'- GGCCAGTTGTGATGACTAAGAC -3' Cd36 FWD 5'-AGATGACGTGGCAAAGAACAG -3’ Cd36 REV 5'-CCTTGGCTAGATAACGAACTCTG -3’ Fabp1 FWD 5'- ATGAACTTCTCCGGCAAGTACC -3’ Fabp1 REV 5'-CTGACACCCCCTTGATGTCC -3’ Fabp2 FWD 5'- GTGGAAAGTAGACCGGAACGA -3’ Fabp2 REV 5'- CCATCCTGTGTGATTGTCAGTT -3’ Fabp3 FWD 5'- GTGGAAAGTAGACCGGAACGA -3’ Fabp3 REV 5'- CCATCCTGTGTGATTGTCAGTT -3’ Fabp4 FWD 5'- AAGGTGAAGAGCATCATAACCCT -3’ Fabp4 REV 5'- TCACGCCTTTCATAACACATTCC -3’ Acly FWD 5'- ACCCTTTCACTGGGGATCACA -3’ Acly REV 5'- GACAGGGATCAGGATTTCCTTG -3’ Acaca FWD 5'- ATGGGCGGAATGGTCTCTTTC -3’ Acaca REV 5'- TGGGGACCTTGTCTTCATCAT -3’ Fasn FWD 5'- AAGTTGCCCGAGTCAGAGAACC -3' Fasn REV 5'- ATCCATAGAGCCCAGCCTTCCATC -3' Scd1 FWD 5'- TTCTTGCGATACACTCTGGTGC -3’

Nature Medicine: doi:10.1038/nm.4412

30

Scd1 REV 5'- CGGGATTGAATGTTCTTGTCGT -3’ Elovl6 FWD 5'- GAAAAGCAGTTCAACGAGAACG -3' Elovl6 REV 5'- AGATGCCGACCACCAAAGATA -3' Dgat1 FWD 5'- GACGGCTACTGGGATCTGA -3' Dgat1 REV 5'- TCACCACACACCAATTCAGG -3' Dgat2 FWD 5'- CGCAGCGAAAACAAGAATAA -3' Dgat2 REV 5'- GAAGATGTCTTGGAGGGCTG -3' Plin2 FWD 5'- GACCTTGTGTCCTCCGCTTAT -3' Plin2 REV 5'- CAACCGCAATTTGTGGCTC -3' Acox1 FWD 5'- GGGCACGGCTATTCTCACAG -3’ Acox1 REV 5'- CATCAAGAACCTGGCCGTCT -3’ Acox2 FWD 5'- ACGGTCCTGAACGCATTTATG -3’ Acox2 REV 5'- TTGGCCCCATTTAGCAATCTG -3’ Cpt1 FWD 5'- GAACACAAATGTGCAAGCAGC -3’ Cpt1 REV 5'- GCCATGACCGGCTTGATCTC -3’ Cpt2 FWD 5'- CAGCACAGCATCGTACCCA -3’ Cpt2 REV 5'- TCCCAATGCCGTTCTCAAAAT -3’ Acsl1 FWD 5'- CGATGGCTGTTGGACTTTGC -3’ Acsl1 REV 5'- CACCCAGGCTCGACTGTATC -3’ Acadl FWD 5'- TCTTTTCCTCGGAGCATGACA -3’ Acadl REV 5'- GACCTCTCTACTCACTTCTCCAG -3’ Ehhadh FWD 5'- CGGTCAATGCCATCAGTCCAA -3’ Ehhadh REV 5'- TGCTCCACAGATCACTATGGC -3’ Acaa1a FWD 5'- AGGCTTCAAGAACACCACCC -3’ Acaa1a REV 5'- GGCTCCTGGCTCAAGAACAT -3’ Tnfa FWD 5'- AGGGTCTGGGCCATAGAACT -3’ Tnfa REV 5'- CCACCACGCTCTTCTGTCTAC -3’ Pai1 FWD 5'- TTCAGCCCTTGCTTGCCTC -3’ Pai1 REV 5'- ACACTTTTACTCCGAAGTCGGT -3’ Ccl2 FWD 5'- TTAAAAACCTGGATCGGAACCAA -3’ Ccl2 REV 5'- GCATTAGCTTCAGATTTACGGGT -3’ Ccl3 FWD 5'- TTCTCTGTACCATGACACTCTGC -3’ Ccl3 REV 5'- CGTGGAATCTTCCGGCTGTAG -3’ Il1b FWD 5'- AAGAGCTTCAGGCAGGCAGTATCA -3’ Il1b REV 5'- TGCAGCTGTCTAGGAACGTCA -3’ Il6 FWD 5'- TAGTCCTTCCTACCCCAATTTCC -3’ Il6 REV 5'- TTGGTCCTTAGCCACTCCTTC -3’ Il10 FWD 5'- GCTCTTACTGACTGGCATGAG -3’ Il10 REV 5'- CGCAGCTCTAGGAGCATGTG -3’

Nature Medicine: doi:10.1038/nm.4412

31

Human primers Sequence GAPDH FWD 5'- GGAGCGAGATCCCTCCAAAAT -3' GAPDH REV 5'- GGCTGTTGTCATACTTCTCATGG -3' HIF1A FWD 5'- GAACGTCGAAAAGAAAAGTCTCG -3' HIF1A REV 3'- CCTTATCAAGATGCGAACTCACA -3' HIF2A FWD 5'- CGGAGGTGTTCTATGAGCTGG -3' HIF2A REV 5'- AGCTTGTGTGTTCGCAGGAA -3' DMT1 FWD 5'- GCTCTCATACCCATCCTCACATT -3' DMT1 REV 5'- TCCATTGGCAAAGTCACTCATT -3' DCYTB FWD 5'- GGTGTTTTCGTAAATACGCTTGG -3' DCYTB REV 5'- ATTGCGGTCTGGTGACTATCC -3' PDK1 FWD 5'- CTGTGATACGGATCAGAAACCG -3’ PDK1 REV 5'- TCCACCAAACAATAAAGAGTGCT -3’ NEU3 FWD 5'- AAGTGACAACATGCTCCTTCAA -3’ NEU3 REV 5'- TCTCCTCGTAGAACGCTTCTC -3’ ChIP primers Sequence NEU3 FWD 5'- TGTCAAGAGGGGCCTTTTCC -3’ NEU3 REV 5'- ACAATCACACGCAAGGACCA -3’ Construct primers Sequence Neu3 promoter FWD 5'- AACTCGAGCACAGACGCTGAAACGAACCC -3’ Neu3 promoter REV 5'- AAGGTACCTGCTGCAGTGTGTCAAGAGGG -3’ Neu3 promoter∆HRE1 FWD 5'- CGCGCAGTCTGTGGAAGGTATCACCGCC -3’ Neu3 promoter∆HRE1 REV 5'- TTCCACAGACTGCGCGGTGAAGGGGC -3’ Neu3 promoter∆HRE2 FWD 5'- GAATCCATATCCCATTCTTCCCCACCTAC -3’ Neu3 promoter∆HRE2 REV 5'- AATGGGATATGGATTCAGTACTTGGTGACAG -3’

Nature Medicine: doi:10.1038/nm.4412