Supplemental Information Table of Contents Supplemental ...genesdev.cshlp.org/.../Supplemental_Data.pdf · Qiao_Supplemental Figure 2, related to Figure 2. (A) Loss of REDD1 confers

Supplemental Information Table of Contents

Supplemental Figure 1, related to Figure 1






Supplemental Methods

Supplemental References

Qiao_Supplemental Materials

C

pancreas

shRedd1

shCtrl xx

p48-Cre;

LSL-KrasG12D

primary

epithelial cells

(KPECs)x

B

A

D

F

E

e.

* *

Pancreas

Liver Lung

PDAC

Qiao_Supplemental Figure 1, related to Figure 1.

Lung KrasG12D;

Redd1+/+

KrasG12D;

Redd1-/-

Metastasis 0/16 2/23

p48-C

re; K

ras

G12D;

Redd1

+/+p

48-C

re; K

ras

G12D;

Redd1

-/-

Pancreas Lung

Intratracheal

Ad-Cre

LoxP LoxP

Kras *G12D

p48 Cre

STOP 1 2*

Redd1 Neoβ-gal(exon 1)

LoxP LoxP

Kras *G12D

Redd1 Neoβ-gal(exon 1)

STOP 1 2*

G

H


(A) Schematic of genetic alleles employed. The REDD1-/- allele replaces the entire Redd1 coding

region with the bgal/neo fusion cDNA.

(B) Tumor-free survival of pancreas cohort. Kaplan-Meier analysis of p48-Cre;KrasG12D;Redd-/-

(p48KR) mice (n = 6) and p48-Cre;KrasG12D;Redd1+/+ (p48K) mice (n = 12) based on the

presence of invasive tumors (PDAC). Mice reaching euthanasia endpoint without PDAC were

censored (tic marks). P value by log-rank test.

(C) Representative photomicrographs of H&E stained tissue sections showing metastases to liver

and lung (indicated by *) from primary PDAC arising in p48KR mice. Scale bar, 100m.

(D) Schematic representation of orthotopic pancreatic tumor model. Primary KRAS-activated

pancreatic epithelial cells (KPECs) expressing control (shCtrl) or REDD1-deficient shRNA

(shREDD1) are re-implanted into the pancreas of host mice.

(E) Representative H&E stained tissue sections of pancreas following orthotopic implantation,

showing histology of poorly differentiated PDAC.

(F) Wet weight of mouse pancreas following orthotopic implantation of REDD1 knockdown or

control primary KPECs. Horizontal lines indicate mean SD. * p = 0.038 by two tailed t-test.

(G) Analysis of the subset of AdK and AdKR mice whose lungs were prepped by perfusion at

necropsy for whole-mount sectioning. Graph at left shows the percentage of lobes with invasive

disease in the respective genotypes, as assessed with H+E (representative sections shown at

right) and with elastic stain (not shown) by an expert lung pathologist (MMK). Pre-invasive

lesions (AdK) are characterized by papillary architecture and condensed nuclei, while invasive

lesions (AdKR) lack papillary morphology and have nuclei with open chromatin and prominent

nucleoli. Scale bar, 250m.

(H) Summary of metastases identified in the lung cancer cohort, by genotype.

0 2 4 6 8

Steroid biosynthesis

Cell adhesion molecules (CAMs)

ECM-receptor interaction

Pathways in cancer

PPAR signaling pathway

Fatty Acid Biosynthesis

Biosynthesis of unsaturated fatty acids

-log 10 (P-Value)

-log 10 (P-Value)

(Gene Number)

KEGG Pathway

PPAR signaling pathway

Fatty acids biosynthesis

Pathways in cancer

ECM-receptor interaction

Cell adhesion molecules (CAMs)

Steroids biosynthesis

7

5

68

11

18

37

11

-log10 (p value)

Biosynthesis of unsaturated fatty acids

KEGG Pathway Gene number

0 2 4 6 8

A


B

D E

F G

J

H

+/++/+

FA

bio

synth

esis

FA

transport

KrasG12D

min max

normoxia hypoxia

Redd1

Fabp3Fabp4Fabp7Slc27a1FasnScd1AcacaHmgcs1Acly

–/– –/–

β-actin

REDD1

shCtrl

shREDD1

1 2 N H

293TA549

K

Fluorescence intensity

Redd1 +/+

norm

oxia

Redd1 -/- n

ormoxia

0.0

0.5

1.0

1.5

2.0

lipid

dro

plet

s le

vel [

LD54

0]

Copy of normoxia

Redd1 +/+ normoxiaRedd1 -/- normoxia*

LD

540

[RF

U]

1.5

2.0

1.0

0.5

0.0

–/–Redd1 +/+

Redd1+/+ Redd1-/-

Counts

104 105 106

400

200

LD540 L

I

shCrl

shRedd1 (E3)

shRedd1 (F9)

0.0

0.5

1.0

1.5

re-plot A549 LPC

*

**1.5

1.0

0.5

0.0LP

C u

pta

ke [R

FU

] A549

shRedd1_1

shRedd1_2

shCtrl ++

--

-

-- +-

M

C


(A) Loss of REDD1 confers tumorigenesis in the setting of mutant RAS expression. Primary Redd1-/-

or Redd1+/+ MEFs were co-infected with lentiviral constructs expressing the adenoviral E1A

protein and KRAS G12V, then injected into immunodeficient (nude) mice. Primary Redd1-/-

MEFs are non-tumorigenic with or without E1A expression alone (not shown). N=8 mice per

genotype. Error bars denote SD. p-value by multiple measures ANOVA.

(B) Schematic of genetic alleles employed in Figure 2.

(C) GSEA plots of RNA-seq data showing suppression of “Horton SREBF targets” signature

(systematic signature M3009) in paired primary KrasG12D;Redd1-/- versus KrasG12D;Redd1+/+

primary cells, cultured under normoxia.

(D) DAVID analysis (https://david.ncifcrf.gov/) of KEGG pathway signatures, showing an

unselected ranking by p-value of the top canonical pathways altered in RNA-seq data of

KrasG12D;Redd1-/- versus KrasG12D;Redd1+/+ primary MEFs (KRMEFs and KMEFs,

respectively). Differential expression was defined as mean > two-fold difference in triplicate

samples of each genotype when normalized by DeSeq. Cells from three mice per genotype were

analyzed.

(E) Enrichment of lipid transport genes and suppression of de novo lipogenesis genes in KRMEFs as

compared to KMEFs cultured under normoxia or hypoxia (1% O2, 18 hr). RNA-seq data from

three independent pairs of KMEFs and KRMEFs.

(F) Increased phospholipids in KRMEFs compared to KMEFs during growth under normoxia as

detected by UHPLC-MS lipidomics analysis.

(G) Uptake of Top-Fluor-LPC in immortalized Redd1-/- as compared to Redd1+/+ MEFs cultured

under normoxia and hypoxia (1% O2, 18 hr). Bar graph (right) represents quantification from

three independent experiments. Error bars indicate SD.

(H) Knockdown of REDD1 in KRAS-activated primary pancreatic epithelial cells (KPECs) by

lentiviral shRNA (shRedd1) or control vector (shCtrl), assessed by qRT-PCR analysis.

(I) TAGs enriched in KRMEFs under normoxia.

(J) TAGs enriched in KRMEFs under hypoxia remain unchanged in the corresponding culture

medium as detected by lipidomics analysis. For (I) and (J), box-and-whisker plots denote

normalized abundance of metabolite. Whiskers denote range and boxes denote SD. Triplicate

samples from each of two mice per genotype were analyzed.

(K) Increased LD540 staining of neutral lipid droplets in in immortalized Redd1-/- MEFs as

measured by flow cytometry analysis. Bar graph at right represents summary of three

independent experiments from paired Redd1+/+ and Redd1-/- MEFs.

(L) Immunoblot for REDD1 confirming shRNA knockdown of REDD1 in A549 cells. Whole cell

lysates from 293T cells cultured under normoxia (N) or hypoxia (H, 1% O2, 4h) were used as

positive control.

(M) Uptake of Top-Fluor-LPC in A549 cells with shCtrl or shREDD1 cultured under normoxia. Bar

graph represents quantification from three independent experiments. Error bars indicate SD.

For all panels, * p< 0.05, ** p < 0.01.

A

00.

10.

20.

40.

60.

8

0

50

100

150

H2O2 [mM]

Copy of Copy of Data 4

AH375 pLKOAH375 D11 shRedd1

H2O2 (mM)

0.0 0.1 0.2 0.4 0.6 0.8

Cell

via

bili

ty (

%) shRedd1

shCtrl

50

100

p < 0.001, ****

Lamin B164

98Nrf2

KrasG12D

+/+ -/-Redd1

C


Vector Redd1-HA

Counts

DCF Fluorescence intensity

300

200

100

KPEC shRedd1

DC

F [

RF

U]

2.5

2.0

0.0

0.5

1.5

1.0

Kra

sG

12D ; R

edd1

+/+

Kra

sG12

D ; Red

d1 -/-

0.0

0.5

1.0

1.5

2.0

2.5

DCF RD1-HA

n.s.

103 104 105

DKra

s G12

D ; Red

d1 +/

+

Kra

s G12

D ; Red

d1 -/-

0.0

0.5

1.0

1.5

DC

F [R

FU

]

KRAS PMEFs DCF Ratio

DC

F[R

FU

]

1.0

0.5

0.0

*

+/+ -/-

KrasG12D

Redd1


102 103 105 106

KrasG12D; Redd1 +/+

KrasG12D; Redd1 -/-

Counts

100

50

DCF

104

B

(A) Decreased ROS in KRMEFS as compared to KMEFs, assessed by staining with CM-H2DCFDA.

Right: Summary from four independent experiments/mice measured in triplicate.

(B) Transfection of REDD1 induces H2O2 as compared to vector control in KPECs with stable

knockdown of endogenous REDD1. Errors bars denote SD. Graph at right shows mean of two

experiments. Error bars denote SD.

(C) Immunoblot showing equal NRF2 protein levels in nuclear extracts of KMEFs and KRMEFs.

TBHQ (Tert-butylhydroquinone) treatment serves as a positive control for NRF2 induction.

(D) Ablation of REDD1 in primary KPECs induces resistance to oxidative stress (H2O2 treatment).

Bars indicate mean of two experiments performed in duplicate. p-value by repeated

measurements of ANOVA.

Unless otherwise noted, for all panels error bars denote SD.


Vehic

leB

SO

KrasG12D;Redd -/- (AdKR) visible orthotopic tumors were harvested from mice treated with either

vehicle or glutathione synthase inhibitor buthionine sulfoximine (BSO) at the end of the experiment.


ED

F H

D11

norm

oxia

D11

hyp

oxia

24hr

D11

hyp

oxia

24hr +

LW

6 20

µM

D11

hyp

oxia

24hr +

PX47

8 10

µM

0.0

0.5

1.0

1.5

Copy of only high dose

normoxiahypoxia 24hrhypoxia 24hr + LW6 hypoxia 24hr + PX478

1.5

1.0

0.5

0.0

Rela

tive m

RN

A

**

***PparƔ

+Hypoxia 18hr

-Normoxia

+LW6

--+

+

-

---

- --+

PX478

+

G

KrasG

12D ; Redd1 +

/+ norm

oxia

KrasG

12D ; Redd1 -

/- norm

oxia

KrasG

12D ; Redd1 +

/+ hypoxia

KrasG

12D ; Redd1 -

/- hypoxia

0

2

4

6

Nor

mal

ized

[log

2]

*Normalized PPARg_061918

Normoxia Hypoxia

**

**6

4

2

0

PparƔ

Rela

tive m

RN

A

KrasG12D; Redd1+/+

shRed

d1_1

shRed

d1_2

0.0

0.5

1.0

1.5

Copy of LPC SSO_summary

VechicleSSO KrasG12D; Redd1-/-

I

4

2

0

Normoxia Hypoxia

CD36

Rela

tive m

RN

A

6

Kra

sG12

D ; Red

d1 +/

+ norm

oxia

Kra

sG

12D ; R

edd1

-/- norm

oxia

Kra

sG

12D ; R

edd1

+/+ hyp

oxia

Kra

sG12

D ; Red

d1 -/- h

ypoxi

a

0

2

4

6

Norm

aliz

ed [lo

g 2

]

*Normalized CD36_061918

*

*

shRed

d1_1

shRed

d1_2

0.0

0.5

1.0

1.5

Ah375 replot GW9662

Vechicle

GW9662

LP

C u

pta

ke [R

FU

]

1.0

0.5

0.0

VehicleGW9662

shRedd1_2 +-

shRedd1_1 + -

**

shRed

d1_1

shRed

d1_2

0.0

0.5

1.0

1.5

Copy of LPC SSO_summary

VechicleSSO

1.5


Counts

GW9662Vehicle

TopFluor-LPC

150

50

100

104 105

148HIF1α

+/+ -/- Redd1

KrasG12D

Lamin B1

Nuclear extract

B C

Kra

s;Red

d1 +/+

Kra

s;Red

d1 -/-

0

1

2

3

mR

NA

fo

ld in

du

cti

on

[vs. A

CT

B]

qRT-PCR of GLUT1

KrasG12D;Redd1 +/+KrasG12D;Redd1 -/-

Rela

tive m

RN

A

3

2

1

0

**

Glut1

-/-Redd1 +/+

KrasG12D

J

A

K


(A) Knockdown of REDD1 in primary KPECs increases protein level of HIFα as detected by

IP/Western analysis under hypoxic conditions (1% O2). IgG serves as a control for IP.

(B) Immunoblot detection of HIF1 protein level in nuclear extracts from paired primary KRMEFs

and KMEFs (normoxia). Two independent experiments were performed.

(C) Expression of Glut1 in KMEFs and KRMEFs, assessed by qRT-PCR analysis. Graph shows

mean of three experiments. Error bars denote SD.

(D) Decreased basal O2 consumption rate (OCR), mitochondrial ATP synthesis (ATP) and maximal

respiratory capacity (MRC) in KRMEFs as measured via Seahorse XFe96. Shown is a

representative experiment, performed three times.

(E) Heatmap depicting significantly altered glycolytic, PPP and TCA metabolite levels in KRMEFs

versus KMEFs as analyzed by UHPLC-MS. Steady-state metabolomics data were normalized to

sample median by MetaboAnalyst. PPP: Pentose phosphate pathway, TCA: tricarboxylic acid

cycle. Triplicate samples from two mice per genotype were analyzed.

(F) Quantification of PPARγ expression levels in KRMEFs and KMEFs cultured under normoxia

and hypoxia (1% O2, 18 hr) as analyzed by RNA-seq analysis. n = 4 independent cultures for

normoxia; n = 3 for hypoxia. Error bars denote SEM.

(G) Induction of PPARγ mRNA under hypoxia (1% O2, 18 hr) in KPECs expressing a REDD1

shRNA can be suppressed by co-treatment with HIF1 inhibitors including LW6 (20M) and

PX478 (10M), as assessed by qRT-PCR analysis. Data represent two experiments performed in

duplicate.

(H) Topfluor-LPC uptake is inhibited by the PPARγ antagonist GW9662 treatment (20M, 12 hr) in

REDD1-ablated KPECs.

(I) Quantification of three independent experiments showing Topfluor-LPC uptake inhibited by

GW9662 treatment as described in (H).

(J) Quantification of CD36 expression levels in KMEFs and KRMEFs cultured under normoxia and

hypoxia (1% O2, 18 hr), as analyzed by RNAseq analysis. n=4 independent cultures for

normoxia; n=3 for hypoxia.

(K) CD36 expression is negatively correlated with REDD1 expression in Patient-Derived Xenograft

(PDX) models of pancreas carcinoma. Values reflect tumor-intrinsic REDD1 levels, as non-

tumor elements are largely murine-derived. Data were derived from the Mouse Models of

Human Cancer Database (MMHC, formerly MTB), Mouse Genome Informatics, The Jackson

Laboratory, Bar Harbor, Maine (http://tumor.informatics.jax.org/). Pearson r and p value (two-

tailed) are shown.

Unless otherwise specified, error bars denote SD.

For all panels, * p < 0.05, ** p < 0.01, *** p < 0.001 by two-tailed t-test.

http://tumor.informatics.jax.org/


B

A

(A) REDD1 expression in normal human cells/tissues. Data were ranked from the highest (left) to

the lowest (right) according to REDD1 expression. Mean fold-difference in expression versus

normal lung or pancreas is shown at top. Data were obtained from the Genotype-Tissue

Expression (GTEx) Project database (https://gtexportal.org/home/).

(B) Proportion of smokers with RAS-mutant lung adenocarcinoma is not significantly different

based on REDD1 gene expression signature. Shown is the TCGA RAS-MUT LUAD population,

stratified into top (high) and bottom (low) quartiles of the signature metagene values as in Fig.

6A. P value by chi-square test.

Analysis of REDD1 expression in human cells/tissues

Data for Supplemental Figure S6A were derived from The Genotype-Tissue Expression (GTEx)

Project. The GTEx Project was supported by the Common Fund of the Office of the Director of

the National Institutes of Health, and by NCI, NHGRI, NHLBI, NIDA, NIMH, and NINDS.

The data used for the analyses were obtained from the GTEx Portal

(https://gtexportal.org/home/ ) on 01/17/2020.

Generation of the REDD1-associated gene expression signature

Gene expression values were derived from RNA-Seq data for KRMEFs compared to KMEFs (4

vs. 4 samples). Reads were aligned to the mm10 reference genome with BWA by the MGH

sequencing core. Bam to Sam file conversion, sorting, indexing, and file merging was done

with SamTools (Li et al., 2009). FPKM values (Fragments per Kilobase of transcript Per

Million mapped reads) were calculated by Cufflinks (Trapnell et al., 2010) version 2.2.1 suit of

tools. First, cufflinks was used to normalize the results for the assembled isoforms using flags

to normalize using only compatible hits that map to the transcriptome and only those in the

upper quartile (and also masked chrM and rRNA). Second, cuffquant was run to quantize reads

with rRNA masking. Finally, cuffnorm was run to produce normalized FPKM. Cufflinks

FPKM was loaded into a matrix in R, quantile normalized, and then a variation filter was

applied to remove genes with less than 1.5 fold minimum variation and 2 minimum absolute

variation (leaving 4124 out of 23235 genes). A t-test was then performed to find genes

significantly varying between KRMEF and KMEF and corrected for multiple hypothesis testing

Qiao_Supplemental Methods.

using the Benjamini-Hochberg (Benjamini & Hochberg, 1995) step-up FDR-controlling

procedure. Genes from the KRMEFs vs. KMEFs with a p-value less than 0.05 were selected

leaving 415 genes (197 up in KRAS null and 218 down in KRAS null), then X and Y

chromosome genes were removed from the signatures leaving 187 up and 210 down genes,

which were used to make a meta-gene to analyze the signature in patient tumor samples. These

genes were mapped to genes in TCGA RNASeqV2 data leaving 159 genes up and 180 genes

down. The meta-gene was made from mean of the log2 of KRAS null up genes thresholded to a

minimum of 0.1 minus the mean of the log2 of KRAS null down up genes thresholded to a

minimum of 0.1.

Qiao_Supplemental Methods.

Benjamini Y, Hochberg Y. 1995. Controlling the False Discovery Rate: a Practical and Powerful

Approach to Multiple Testing. Journal of the Royal Statistical Society 57: 11.

Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R,

Genome Project Data Processing S. 2009. The Sequence Alignment/Map format and

SAMtools. Bioinformatics 25: 2078-2079.

Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold

BJ, Pachter L. 2010. Transcript assembly and quantification by RNA-Seq reveals

unannotated transcripts and isoform switching during cell differentiation. Nat

Biotechnol 28: 511-515.

Qiao_Supplemental References.

Supplemental Information Table of Contents Supplemental ...genesdev.cshlp.org/.../Supplemental_Data.pdf · Qiao_Supplemental Figure 2, related to Figure 2. (A) Loss of REDD1 confers

Documents