ONLINE SUPPLEMENTARY MATERIALS AND METHODS

Histological analysis

After fixed in 4% formalin and embedded in paraffin, tissues were cut into 4-μm

sections. H&E staining were performed following the conventional protocol [1].

Immunohistochemistry (IHC) staining was performed using rabbit antibodies for DLC3

(Proteintech, Wuhan, China), MACC1 (Abnova, Taipei, China) and Ki-67 (Cell

signaling technology, Danvers, MA) and stained by Dako Envision System (Dako,

Glostrup, Denmark). The protein expressions were analyzed by a semi-quantitative

method [1]. The intensities of the protein were scored as 0 (negative), 1 (weak), 2

(medium) or 3 (intense), while the staining extent were scored as 0 (0%), 1 (1%-25%),

2 (26%-50%), 3 (51%-75%) or 4 (76%-100%) according to the stained cells in

percentage. Multiplying the staining intensity and extent scores together gets the final

protein expression score. Western blotting and Immunofluorescence

For western blotting, Goat Anti-Rabbit antibodies of DLC1 (Abcam, Cambridge, MA,

US), DLC3 (Proteintech), MACC1 (Abnova), JNK (Abcam), phosphorylated-

JNK1/2/3 at sites Y185/Y185/Y223 (Abcam), c-JUN (Cell Signaling Tech., Danvers,

MA, US), phosphorylated-c-JUN at site S63 (Cell Signaling Tech.), E-cadherin (Cell

Signaling Tech.), vimentin (Cell Signaling Tech.) and GAPDH (Cell Signaling Tech.)

were used. For immunofluorescence, antibodies of DLC3 (Proteintech) and MACC1

(Abcam), E-cadherin (Cell Signaling) and vimentin (Cell Signaling) were used.

RhoA activity assay

RhoA activity was measured using a RhoA activation assay kit (cytoskeleton, Denver,

OH). Briefly, freshly prepared cell lysates of 800 μg total protein were incubated with

50 μg rhotekin-RBD beads at 4℃ on a rotator for 1h. After centrifugation, supernatant

removal and carefully washing, the bead pellet was suspended in 2× Laemli buffer and

boiled for 2 min. Then, the prepared samples were subjected for Western blotting using

the anti-RhoA antibody included in the kit. Total RhoA expressions were also blotted

as reference.

Protein array assay

To find out DLC3 downstream signaling, protein array assay was performed using

Human Phospho-kinase array kit (R&D Systems, Minneapolis, MN) and following the

manufacturer's protocol. qRT-PCR

Trizol kit (Life science, Carlsbad, CA) for cells and tissues total RNA extraction and a

reverse transcriptase (Roche, Penzberg, Germany) were used according to the protocols

as recommended. qRT-PCR was performed using SYBR Green I Master kit (Roche) on

a LightCycler 480 system. Primer sequences involved in present study are listed in

online supplementary Table 1.

Dual-luciferase reporter assay

According to the documented functional binding site for AP-1 in the MACC1

promoter [2], mutated MACC1 reporter fragments was generated (see online

supplementary Table 2). To detect the influence of DLC3 on the MACC1 transcription,

MACC1 and its corresponding mutated reporter fragments were cloned into the

luciferase pGL3 vector. Luciferase activities of each groups was detected using Dual-

Luciferase Reporter Assay System (Promega, Madison, WI, US) according to the

manufacturer’s instructions. ROS detection and cell apoptosis assays

ROS was measured by fluorescent DCFH-DA dye (Jiancheng, Nanjing, China) and

observed under fluorescent microscope after 6 hours glucose deprivation. Cell

apoptosis was analyzed by staining Annexin V-FITC and Propidium Iodide after 12

hours glucose deprivation, and detected by flow cytometry analysis using the

FACSCanto II system (BD Biosciences, San José, CA).

Cell viability, proliferation and colony formation assays

Cells viability was measured by methyl-thiazolyl-tetrazolium (MTT) assay as

previously described [3]. Cell proliferation ability was reflected by EdU incorporation

assay using the EdU Assay Kits (Life Technologies, Carlsbad, CA, USA) according to

the manufacturer's instruction and observed under fluorescent microscope. For colony

formation, trypsinized cells were suspended in 2 mL of complete medium (Sigma, St.

Louis, MO, US) in complete medium. After 14 days culturing, cells were observed by

crystal violet. Migration, invasion and glucose chemotaxis assays

For scratch healing assay, trypsinized cells were seeded on 24-well culture plates

(5×104/well). Upon confluence, cells were scratched with a sterile 200 μL pipette tip.

Wound closure was observed at indicated time point. For Boyden chamber transwell

assay, cells were planted in serum free medium in the upper chamber, which consisted

of 8 μm membrane filter inserts (5×104/ml, 200 μL/well) with Matrigel. The lower

chamber was complete medium with serum as chemoattractant. After 48 hours culturing,

cells at upper chamber were fixed by fixed in 4% paraformaldehyde and stained by

crystal violet. Cells were photographed under microscope. For glucose chemotaxis

assay, cells were seeded on the upper chamber coated with Matrigel (BD Biosciences)

and incubated with RPMI 1640 glucose free medium, while the lower chamber was

supplemented with RPMI 1640 high glucose medium (2 g/L) as attractant. Cells were

cultured for 48 hours before crystal violets staining.

Glucose uptake and pH evaluation

Cells were seeded on 6-well plates (5×105/well) and cultured in 2 ml complete medium.

The 24-hour conditioned medium and the original medium (baseline control) were

collected. Glucose uptake was determined using a glucose assay kit by conventional

enzymatic methods (Randox, Antrim, UK). Total protein of cells was detected by

conventional BCA method (Beyotime, Haimen, China). The pH value was detected by

a pH meter.

Reference 1. Wang L, Wu Y, Lin L, Liu P, Huang H, Liao W, et al. Metastasis-associated in colon cancer-1 upregulation predicts a poor prognosis of gastric cancer, and promotes tumor cell proliferation and invasion. International Journal of Cancer. 2013; 133: 1419-30. 2. Juneja M, Ilm K, Schlag PM, Stein U. Promoter identification and transcriptional regulation of the metastasis gene MACC1 in colorectal cancer. Mol Oncol. 2013; 7: 929-43. 3. Wang L, Lin L, Chen X, Sun L, Liao Y, Huang N, et al. Metastasis-associated in colon cancer-1 promotes vasculogenic mimicry in gastric cancer by upregulating TWIST1/2. Oncotarget. 2015; 6: 11492-506.

OnlineSupplementary

Figures and Tables

Supplementary Figure S1 DLC3 is associated with perfusion status in tumor microenvironment, andDLC3 low expression indicates poor cancer prognosis.(A) Quantified of the RhoA activity changes upon glucose deprivation by three independent repeats. (B)Using TCGA GC tissue data, the mRNA expressions of DLC family members in correlation with CD31, andthree representative GEF members (TRIO, ARHGEF1 and MCF2L) were analyzed. Angiogenic factorVEGFA in correlation with DLC3 was also analyzed. (C and D) Using online Kaplan-Meier plotter forbioinformatic analysis, DLC3 downregulation was associated with poor survival in (C) breast cancer(GSE42568) and (D) lung cancer (CaArray) respectively.

0 40 80 1200

20

40

60

80

100 DLC3 lowDLC3 high

Time (month)

Recu

rren

ce F

ree

(%)

n=52

n=52

0 40 80 1200

20

40

60

80

100 DLC3 lowDLC3 high

Time (month)

Ove

rall

Surv

ival

(%)

n=52

n=52

Breast Cancer (GSE42568)

0 60 120 180 2400

20

40

60

80

100 DLC3 lowDLC3 high

n=246

n=122

Time (month)Free

of p

rimar

y pr

ogre

ssio

n (%

)

0 60 120 180 2400

20

40

60

80

100 DLC3 lowDLC3 high

n=302

n=302

Time (month)

Ove

rall S

urviv

al (%

)Lung Cancer (CaArray)

C D

0 1 2 3 4 50

2

4

6

8 r2=0.4387p<0.001

CD31

DLC

1

0 1 2 3 4 50

2

4

6

8

10 r2=0.1796p<0.001

CD31

DLC

2

0 1 2 3 4 50

1

2

3

4

5 r2=0.6203p<0.001

CD31

DLC

3

A

0 1 2 3 4 50

2

4

6 r2=0.0333p<0.001

CD31

TRIO

0 1 2 3 4 50

2

4

6 r2=0.0044p=0.199

CD31

ARHG

EF1

0 1 2 3 4 50

2

4

6 r2<0.0001p=0.907

CD31

MC

F2L

0 1 2 3 4 50

2

4

6

8 r2=0.0013p=0.4933

DLC3

VEG

FAB

+ -0

1

2

3

*

Glucose

Activ

e Rh

oA/T

otal

Rho

A

Supplementary Figure S2 Constructing DLC3-overexpression and -silenced GC cell models.

DLC3 was successfully (A) overexpressed and (B) silenced in MKN45 cells. (C) Correspondingly, RhoAactivities were altered by DLC3. (D) Glucose deprivation did not influence DLC3 expression inoeDLC3 cells.

BCtrl oeDLC3

GAPDH

DLC3

A

120kDa

36kDa

C

GAPDH

DLC3shN -#1 -#2

shDLC3

120kDa

36kDa

RhoA-GTP 21kDa

Total RhoA 21kDa

shN -#1 -#2shDLC3

D

RhoA-GTP 21kDa

Total RhoA 21kDa

Ctrl oeDLC3

0 1 3 6GAPDH 36kDa

DLC3 120kDa

Glucose deprivation time (h)12

shDLC3■-#1 ■-#2■shN

BGC

823

0

100

200

300

400 +

#

Col

ony

num

ber

0

50

100

150

+

Col

ony

num

ber

□oeDLC3■Ctrl

BGC

823

BGC823

■Ctrl

□oeD

LC3

EdU Hoechst Merge

0

10

20

30

40

+

EdU/

Hoec

hst (

%)

A B

C

0.0

0.4

0.8

1.2

+

Cel

l via

bility

(fol

d)

■Ctrl□oeDLC3

BGC823

0

500

1000

1500 +

#

Inva

ded

cells

BGC823

shDLC3■-#1 ■-#2■shN

0

500

1000

1500

*

Inva

ded

cells

□oeDLC3

BGC823

■Ctrl

D

shDLC3

shN

BGC823

0h 6h 12h 24h

-#1

-#2

BGC823

0h 6h 12h 24h

Ctrl

oeDLC3

E

Supplementary Figure S3 DLC3 suppressed proliferation and migration in BGC823 cells.

(A) The viability and (B) DNA replication of BGC823 were suppressed by DLC3 overexpression. (C)DLC3 influenced colony formation of BGC823. (D) The migration and (E) invasiveness was inhibitedby DLC3 overexpression and enhanced by its silencing. *P<0.05, #P<0.01, +P<0.001.

Supplementary Figure S4 DLC3 inhibited MACC1 expression via RhoA/JNK/AP-1 signaling inMKN45 cells.(A) OeMACC1 did not influence DLC3 expression, but oeDLC3 significantly downregulated MACC1expression. OeMACC1 reversed the epithelial mesenchymal transition marker alterations inducedby oeDLC3 in MKN45 cells. Right panel: semi-quantified MACC1 relative gray values (n=3). (B and C)Glucose deprivation upregulated (B) MACC1 protein and (C) mRNA expressions via DLC3 in MKN45cells. (B) Right panel: semi-quantified MACC1 relative gray values (n=3). (D) In protein array assay,shDLC3 enhanced the phosphorylation of JNK and several other protein kinases. (E) Semi-quantifiedMACC1 relative gray values of shDLC3 and corresponding control (n=3). (F and G) RhoA, ROCK orJNK inhibitors suppressed shDLC3-promoted MKN45 (F) colony formation and (G) invasion.Quantified results related to Figure 6G and H.

A

B

JNK1/2/3EGFR

TORSrcFynHck

p53AMPKα1

HSP27Yes

MSK1/2

Lyn

shN

shDLC3-#2

oeDLC3 - - + +oeMACC1 - + - +

DLC3MACC1

E-cadherinvimentinGAPDH

120kDa

36kDa

97kDa135kDa57kDa

Glucose + - + -oeDLC3 - - + +

DLC3

MACC1GAPDH

120kDa

36kDa

97kDa

D

0

1

2

3 *

GlucoseoeDLC3

+ - + -- - + +

*

*

MAC

C1

mRN

A(fo

ld to

Ctrl

)

0

12

34

5#

#

oeDLC3oeMACC1

- - + +- + - +

*ns

MAC

C1

gray

val

ue

0.0

0.5

1.0

1.5

2.0 *

GlucoseoeDLC3

+ - + -- - + +

*

*

MAC

C1

gray

val

ue

F

G

E

0100

200

300

400

500

shDLC3RhoAi

- + - +- - + +

- +- -

ROCKi - - - - + +

P valuesCtrl vs.. shDLC3Ctrl vs. RhoAiCtrl vs. ROCKiCtrl vs. JNKiCtrl vs. shDLC3+RhoAiCtrl vs. shDLC3+ROCKiCtrl vs. shDLC3+JNKishDLC3 vs. RhoAishDLC3 vs. shDLC3+RhoAishDLC3 vs. ROCKishDLC3 vs. shDLC3+ROCKishDLC3 vs. JNKishDLC3 vs. shDLC3+JNKiRhoAi vs. shDLC3+RhoAiROCKi vs. shDLC3+ROCKiJNKi vs. shDLC3+JNKi

<0.01<0.001<0.001<0.001<0.001<0.01<0.01<0.001<0.001<0.001<0.001<0.001<0.001<0.05<0.05<0.05ROCKi - - - - - -

- +- -- -+ +

Col

ony

num

ber

0

100

200300

400

500

shDLC3RhoAi

- + - +- - + +

- +- -

ROCKi - - - - + +

P valuesCtrl vs.. shDLC3Ctrl vs. RhoAiCtrl vs. ROCKiCtrl vs. JNKiCtrl vs. shDLC3+RhoAiCtrl vs. shDLC3+ROCKiCtrl vs. shDLC3+JNKishDLC3 vs. RhoAishDLC3 vs. shDLC3+RhoAishDLC3 vs. ROCKishDLC3 vs. shDLC3+ROCKishDLC3 vs. JNKishDLC3 vs. shDLC3+JNKiRhoAi vs. shDLC3+RhoAiROCKi vs. shDLC3+ROCKiJNKi vs. shDLC3+JNKi

<0.001<0.001<0.001<0.001<0.01<0.001<0.05<0.001<0.001<0.001<0.001<0.001<0.001<0.05<0.05<0.01ROCKi - - - - - -

- +- -- -+ +

Inva

ded

cells

C

shN shDLC30.0

0.5

1.0

1.5

2.0 #

MAC

C1

gray

val

ue

Supplementary Figure S5 The DLC3/MACC1 axis modulated glycolysis remodeling.(A) Indicated by EdU staining, siMACC1 significantly reversed shDLC3 accelerated DNA-replication inMKN45 cells. (B) The DLC3/MACC1 axis regulated cellular 18F-FDG uptake in culturing plates. (C)Silencing MACC1 reduced glucose uptake in shDLC3 MKN45 cells. (D) The DLC3/MACC1 axisinfluenced acidosis in the conditioned medium. (E) SiMACC1 further increased glucose deprivation-promoted ROS generation. #P<0.01, +P<0.001.

- +shDLC3

siM

ACC

1 -

+

EdU/Hoechst0

10

20

30

40

50+

shDLC3siMACC1

- + - +- - + +

+ +E

dU/H

oech

st (%

)

A

7.0

7.2

7.4

7.6

pH v

alue

shDLC3 +- - +siMACC1 -- + +

siMACC1shDLC3 - -+ +

- +- +

Low High

B

C D E

0

100

200

300

400 +

shDLC3siMACC1

- + - +- - + +

+

+

Glu

cose

/Pro

tein

(mm

ol/g

)

- +siMACC1

Glu

cose

+

-

DCFH-DA/Bright field

Supplementary Figure S6 The DLC3/MACC1 axis modulated GC cell invasion and glucosechemotaxis.(A and B) The DLC3/MACC1 axis modulation on MKN45 cell motility was confirmed by (A) scratchhealing migration assay and (B) transwell invasion assays. (C) DLC3/MACC1 axis modulation onepithelial and mesenchymal marker expressions were confirmed by immunofluorescence. (D and E)Silencing MACC1 (D) reversed the glucose-deprivation-induced cytoskeleton changes and (E)inhibited glucose chemotaxis in MKN45 cells. *P<0.05, #P<0.01, +P<0.001.

- +shDLC3

siM

ACC

1 -

+

siMACC1shDLC3 - - + +

- + - +

E-ca

dher

inD

LC3

MAC

C1

vim

entin

50μm

050

100150

200

250 ++

shDLC3siMACC1

- + - +- - + +

+

+

Inva

ded

cells

A

siMACC1shDLC3

0h

++1

-+1

+-1

--1

48h

0.660.730.440.58

B

D

E

C

0 100 200 300 400

#

++

GlucosesiMACC1

+ -

- -

+ +

- +

Invaded cells

Ctrl Glc(-)Glc(+)

Ctrl Glc(+)Glc(+)

siMACC1

siMACC1

+

-

-siMACC1

+

20μm

Glu

cose

CTTCAGCTCTGAATCACCGAAAGAGAATCTCTTCAGCTCTGAATTGCCGAAAGAGAATCT

MACC1mutated

AP-1

Supplementary Table S1 The qRT-PCR primer sequences used in this study

Target SequenceMACC1 F GAGTTAGTCGCACGTCTCATCC

R AGTGAGCACTCCAGGTATACAGDLC3 F GACGGAGCAGTCCCTCCT

R CTTTTGGCCTCGGCTTCTβ-actin F TGGCACCCAGCACAATGAA

R CTAAGTCATAGTCCGCCTAGAAGCA

Supplementary Table S2 The functional binding site for AP-1 in MACC1 promoter