0

500

1000

1500

MDA-MB-231

NMXHPX

Mot

ility

(µM

)

*****

B C

0

200

400

600

800

1000

MCF7

NMXHPX

********

Mot

ility

(µM

)0

200

400

600

800

1000

MCF7

NMXHPX

Mot

ility

(µM

)

0

500

1000

1500

2000

MDA-MB-231

********

*

Mot

ility

(µM

)

NMXHPX

A

0 20 40 60 80 1000

500

1000

1500

2000

MCF7

Time (hours)

Vehicle

UNC0642 2 µM

UNC0642 5 µM ****

MDA-MB-231

Time (hours)

Vehicle

UNC0642 2 µM

UNC0642 5 µM*******

Mot

ility

(µM

)

Mot

ility

(µM

)

shNS shG9a Vehicle UNC0642 shNS shG9a Vehicle UNC0642

Figure S1(A) Holomonitor analysis of cellular motility in MCF7 and MDA-MB-231 cell lines following treatment with UNC0642 (2 or 5 μM) for 96 hours. (B) Bar graph representing the total distance covered by MDA-MB-231 and (C) MCF7 tracked through holographic microscopy for 48 hours under normoxic (21% O2) or hypoxic (1% O2) conditions, following G9a knock down or UNC0642 treatment (5 μM). Data are represented as mean ± SEM (One-way ANOVA, * p<0.05, *** p<0.0005, **** p<0.0001).

******

********

0 20 40 60 80 1000

1000

2000

3000

4000

A Vehicle HPX B

UNC0642 HPX

Vehicle NMX

UNC0642 NMX

200μm

shG9a NMX

MCF7shNS NMX shNS HPX

shG9a HPX

0 12 24 36 480

200

400

600

800

1000MCF7 HPX

Time (hours)

shNSshG9a

****

0 12 24 36 480

200

400

600

800

1000MCF7 NMX

Time (hours)

shNSshG9a

****

Mot

ility

(µM

)

Mot

ility

(µM

)

0 12 24 36 480

200

400

600

800

1000

Time (hours)

VehicleUNC0642

****

0 12 24 36 480

200

400

600

800

1000

Time (hours)

VehicleUNC0642

****

MCF7 NMX MCF7 HPX

Mot

ility

(µM

)

Mot

ility

(µM

)

Vehicle

0h 24h 48h 72h

UNC0642

MCF7 NMX

Vehicle

UNC0642

0h 24h 48h 72h

Wou

nd c

losu

re (%

)

MCF7 HPX24h 48h 72h

0

50

100

150 VehicleUNC0642

****

****

24h 48h 72h0

50

100

150 VehicleUNC0642

*****

Wou

nd c

losu

re (%

)

C

MCF7

+- +-+ - +-

D

G9a

Lamin A/C

shG9aHPX

MDA-MB-231Vehicle UNC 2 µM MCF7

0h

96h

E

HIF-1α

UNC 5 µM

Figure S2. (A) Evaluation of the migratory distance covered by MCF7 cells treated with or without 5 μM UNC0642 for 48 hours or (B) following G9a KD. Results were evaluated through real-time imaging using the Holomonitor M4, taking pictures every 10 minutes. An average of 20 cells per condition is shown for the representative displacement images. (C) Western Blot analysis of G9a in MCF7 following G9a KD. (D) Scratch wound assay of MCF7 breast cancer cells treated with UNC0642 (5 μM) under both normoxic (21% O2) and hypoxic (1% O2) conditions. (E) Matrigel invasion assay of MDA-MB-231 and MCF7 cells in the presence of indicated concentrations of UNC0642. Graph summarises percentage of wound closure at every imaged time. Results were evaluated by real-time imaging performed by the IncuCyte Zoom every 24 hours and wound closure was quantified using ImageJ. Data are represented as mean ± SEM (non-parametric, Student t-test),*p<0.05, ** p<0.005, **** p<0.0001.

MCF7

50 100-50

0

50

100

150

Invasion in matrigel

Time (hours)

Wou

nd c

losu

re (%

)

VehicleUNC0642 2 µM UNC0642 5 µMMCF7 **

MDA-MB-231

Vehicle UNC0642 Doxorubicin

0h

72h

0h

72h

Merged

NucGreen

shNS shG9a

0 20 40 60 800

100

200

300

400

NucGreen positive cells

Time (Hours)

Num

ber o

f dea

d ce

lls

VehicleUNC0642Doxorubicin

****

shNSshG9a

*

Figure S3Scratch wound healing assay in MDA-MB-231 cells treated with UNC0642 (5 μM) or following G9a KD in the presence of NucGreen dead cells stain. Top phase-contrast images are merged with Green channel. Bottom images show only the relative green signal. Graph represents NucGreen quantification over time. Doxorubicin (1 μM) was used as positive control. Data are represented as mean ± SEM (non-parametric, Student’s t-test), * p<0.05, **** p<0.0001).

A

0h 9h 18h0

1

2R

elat

ive

mR

NA

leve

l

HPX

*

CDH10 B

0

1

2

3

Fold

Cha

nge

**MCF7

shNS shG9a

- -+ +

MDA-MB-231IGFBP3

** ** **

shNS shG9a

- -+ +0

1

2

3

Fold

Cha

nge

**

MCF7

- -+ +

MDA-MB-231IGFBP3

** ** **

- -+ +

Vehicle UNC0642 Vehicle UNC0642

HPX HPX

MCF7

shG9aHPX

+- +-+ - +-

MCF7

G9a

C

G9a

H3K9me2

H3

Tubulin

CDH10

G9a-/-

G9a WT ++

-G9a ∆SET - -

-

Empty Vec

tor

G9a W

T

0

0.5

1.0

1.5

CDH10

***

D E

F

UNC0642HPX

+- +-+ - +-

MCF7 MDA-MB-231+- +-

+ - +-

CDH10

CDH1

EP-CAM

Tubulin

G9a ∆SET

Rel

ativ

e m

RN

A le

vel

Figure S4. (A) CDH10 expression in MCF7 cells cultured under hypoxic conditions for the indicated times. (B) IGFBP3 expression evaluated in MDA-MB-231 and MCF7 cells as positive control for the hypoxic environment. (C) Western Immunoblotting analysis of CDH10 in MCF7 cells transfected with shG9a and exposed to normoxia or hypoxia for 24 hours. (D) Western Immunoblotting analysis of CDH1, EpCAM and CDH10 in MCF7 and MDA-MB-231following exposure to normoxia or hypoxia for 24 hours in the presence or in the absence of UNC0642 (5 μM). (E) Western immunoblotting analysis of CDH10 and H3K9me2 in G9a-/- MEFs transfected with WT G9a or G9a ΔSET. (F) CDH10 mRNA levels in G9a-/- MEFs transfected as described. Data are represented as mean ± SEM (non-parametric, Student’s t-test), * p<0.05, ** p<0.005, *** p<0.005.

**

**

Tubulin

CDH10

CDH10

Tubulin

shCDH10 +_

Figure S5. Evaluation of the migratory distance covered in 48 hours by MCF7 following CDH10 KD using theHoloMonitor M4. Data are represented as mean ± SEM (non-parametric, Student’s t-test), **** p<0.0001.

0 12 24 36 480

200

400

600

800

1000

MCF7

Time (hours)

shNS

shCDH10

Mot

ility

(µM

) ****

0 100 200 3000

50

100

Time (months)

High=1680Low=1402

HR: 1.12p=0.078

0 50 100 150 200 2500

50

100

Time (months)

High=434Low=435

HR: 0.9p=0.32

0

All breast cancer patients

Time (months)

EHMT2

Rel

apse

-free

sur

viva

l

Rel

apse

-free

sur

viva

l

Rel

apse

-free

sur

viva

l

EHMT2ER+

EHMT2ER-

100 200 300 400

High=1820Low=2131

HR: 1.07p=0.240

50

100

0 100 200 300Time (Months)

High=1083Low=850

HR: 0.87p=0.099

0 100 200 3000

50

100

Time (Months)

High=687Low=462

HR: 1.27p=0.018

Luminal AEHMT2

Luminal BEHMT2

Rel

apse

-free

sur

viva

l

Rel

apse

-free

sur

viva

l

0 50 100 150 200 2500

50

100

Time (Months)

High=127Low=124

HR: 0.73p=0.11

HER2+EHMT2

Rel

apse

-free

sur

viva

l

Rel

apse

-free

sur

viva

l

0 50 100 150 200 2500

50

100

Time (Months)

High=309Low=309

HR: 1.05p=0.71

Basal-likeEHMT2

0

50

100

A B

GFED

C

Figure S6. Kaplan-Meyer relapse-free survival analysis of EHMT2 expression in (A) all breast cancer patients, (B) ER+, (C) ER-, (D) luminal A, (E) luminal B, (F) HER2+ and (G) basal-like.

Alte

ratio

n fre

quen

cy

0.5%

1%

1.5%

2%

BRCA (INSERM 2016)

Breast (TCGA)

The MBC Project

Breast (METABRIC 2016)

CNA dataMutation data

EHMT2 CDH10

1%

2%

3%

4%

5%

6%

BRCA (INSERM 2016)

Breast (METABRIC 2016)

Breast (TCGA)

Breast (BCCRC 2012)

The MBC Project

Breast (Sanger)

Breast (Broad 2012)

CNA dataMutation data

Alte

ratio

n fre

quen

cy

A

2%

4%

6%

8%

Lung adeno (MSKCC)

NSCLC (MSKCC)

CCLE (Novartis/Broad 2012)

Lung adeno (TCGA pub)

Small Cell Lung (UCOLOGNE 2015)

Lung adeno (TCGA)

NSCLC (MSKCC 2018)

Lung adeno (TCGA PanCan)

Lung adeno (Broad)

NSCLC (TCGA 2016)

Lung squ (TCGA pub)

Lung squ (TCGA PanCan)

Lung squ (TCGA)

CNA dataMutation data

Alte

ratio

n fre

quen

cy

EHMT2

5%

10%

15%

20%

25%

30%

Lung squ (TCGA PanCan)

NSCLC (TCGA 2016)

Lung adeno (TCGA pub)

Lung squ (TCGA)

Lung adeno (Broad)

Lung squ (TCGA pub)

Lung adeno (TCGA PanCan)

Small Cell Lung (CLCGP)

NSCLC (MSKCC)

Lung adeno (MSKCC)

Lung adeno (TCGA)

Tracerx (NSCLC 2017)

Small Cell Lung (UCOLOGNE 2015)

NSCLC (MSKCC 2018)

SCLC (GARDNER 2017)

CCLE (Novartis/Broad 2012)

Small Cell Lung (JHU)

CNA dataMutation data

Alte

ratio

n fre

quen

cy

CDH10

Figure S7. (A) Percentage and type of alterations in the EHMT2 and CDH10 genes in breast cancer patients and (B) lung cancer patients, subdivided by cancer study.

Mutation

Amplification

Deep Deletion

Multiple Alterations

Mutation

Amplification

Deep Deletion

Multiple Alterations

B

A

B

Figure S8. (A) TCGA Pan-cancer data comparing the expression levels of EHMT2 in tumor and normal tissue samples. (B) TCGA Pan-cancer data comparing the expression levels of CDH10 in tumor and normal tissue samples.

0 2000 4000 60000.1

1

10

100

1000

10000

CD

H10

Basal-like

EHMT2

(n=171)

0 2000 4000 60000.1

1

10

100

1000

10000 Her2+

EHMT2

CD

H10

(n=78)

0 2000 4000 60000.1

1

10

100

1000

10000 Luminal A

EHMT2

CD

H10

(n=498)

0 2000 4000 60000.1

1

10

100

1000 Luminal B

EHMT2

CD

H10

(n=197)

0 2000 4000 60000.1

1

10

100

1000

10000 Normal

EHMT2

CD

H10

(n=36)

0 2000 4000 60000.1

1

10

100

1000

10000

EHMT2

CD

H10

Other(n=101)

A B C

D E F

Figure S9. Correlation between CDH10 and EHMT2 (G9a) mRNA levels in (A) normal, (B) basal-like, (C) Her2+, (D) Luminal A, (E) Luminal B and (F) others in patient samples. (G) Correlation between G9a and CDH10 protein levels in breast cancer patients using publicly available databases. Blue square identifies patients with no detectable G9a. In the red square are patients with no detectable CDH10. In the black square are patients for which both proteins were detectable.

G

-1.0 -0.5 0.5 1.0 1.5

-1.0

-0.5

0.5

1.0

1.5

G9a (rlog)

(n=21)(n=18)

n=(26)

ER- ER+

G9a and CDH10 protein correlation

CD

H10

(rlo

g)

7

Table S1. Table S1: RT-PCR and ChIP primers

8

Table S2. Pan-cancer analysis of CDH10 and EHMT2.