EN1 Is a Transcriptional Dependency in Triple- …Tumor Biology and Immunology EN1 Is a...
Transcript of EN1 Is a Transcriptional Dependency in Triple- …Tumor Biology and Immunology EN1 Is a...
Tumor Biology and Immunology
EN1 Is a Transcriptional Dependency in Triple-Negative Breast Cancer Associated with BrainMetastasisGuillermo Peluffo1,2, Ashim Subedee1,3, Nicholas W. Harper1, Natalie Kingston1,Bojana Jovanovi�c1,2, Felipe Flores1,4, Laura E. Stevens1,2, Francisco Beca5,6, Anne Trinh1,2,Chandra Sekhar Reddy Chilamakuri7, Evangelia K. Papachristou7, Katherine Murphy1,Ying Su1,2, Andriy Marusyk1,2, Clive S. D'Santos7, Oscar M. Rueda7, Andrew H. Beck5,6,Carlos Caldas7, Jason S. Carroll7, and Kornelia Polyak1,2,3
Abstract
To define transcriptional dependencies of triple-negativebreast cancer (TNBC), we identified transcription factors high-ly and specifically expressed in primary TNBCs and tested theirrequirement for cell growth in a panel of breast cancer celllines. We found that EN1 (engrailed 1) is overexpressed inTNBCs and its downregulation preferentially and significantlyreduced viability and tumorigenicity in TNBC cell lines. Byintegrating gene expression changes after EN1downregulationwith EN1 chromatin binding patterns, we identified genesinvolved in WNT and Hedgehog signaling, neurogenesis, andaxonal guidance as direct EN1 transcriptional targets. Quan-
titative proteomic analyses of EN1-bound chromatin com-plexes revealed association with transcriptional repressors andcoactivators including TLE3, TRIM24, TRIM28, and TRIM33.High expression of EN1 correlated with short overall survivaland increased risk of developing brain metastases in patientswith TNBC. Thus, EN1 is a prognostic marker and a potentialtherapeutic target in TNBC.
Significance: These findings show that the EN1 transcrip-tion factor regulates neurogenesis-related genes and is associ-ated with brain metastasis in triple-negative breast cancer.
IntroductionBreast cancer is aheterogeneous groupof diseaseswithdifferent
biological and clinical characteristics. On the basis of the presenceof estrogen and progesterone receptors (ER and PR), andHER2, tumors are classified into ERþ, HER2þ, and ER�PR�HER2�
(triple-negative) subtypes, whereas gene expression and epige-netic profiles divide breast tumors into luminal and basal
groups (1). Knowledge of the molecular properties of luminalERþ andHER2þ subtypes has led to thedevelopment of endocrineand HER2-targeted therapies. However, currently there is noeffective targeted therapy for triple-negative breast tumors.
Triple-negative breast cancer (TNBC) constitutes 10%–20% ofbreast cancer cases in the United States and more commonlyaffects younger and African-American women (2, 3). TNBCs havehigher risk of developing distant metastases and in general havepoor clinical outcome. TNBCs are also heterogeneous and havebeen grouped into luminal, basal, and mesenchymal subtypesbased on gene expression patterns (2–4). However, these sub-types, besides luminal androgen receptor (AR)-positive tumors,have not impacted the clinical management of patients withTNBC (3) highlighting the need for additional molecular markersto guide treatment decisions. TNBC genome sequencing studieshave so far failed to identify novel recurrent mutations besidesTP53, PIK3CA, andPTEN (2, 3), suggesting that TNBCphenotypesmay be driven by nongenetic alterations such as perturbed epi-genetic and transcriptional programs.
The luminal phenotype is defined by a set of lineage-specifictranscription factors including ESR1, FOXA1, GATA3, and SPDEFthat also represent transcriptional dependencies in luminal breasttumors (2, 3). We hypothesized that transcription factors (TF)specifically expressed in TNBCs may also exemplify such depen-dencies that can be exploited therapeutically and may divideTNBC into clinically relevant subsets. To test this hypothesis, firstwe selected TFs that are highly and specifically expressed inprimary TNBCs followed by a targeted cellular viability screenfor these TFs in a panel of breast cancer cell lines of differentsubtypes. Using this approach, we have identified several TFsspecifically required for the survival of TNBCs and among these
1Department of Medical Oncology, Dana-Farber Cancer Institute Boston,Massachusetts. 2Department of Medicine, Harvard Medical School, Boston,Massachusetts. 3BBS Program, Harvard Medical School, Boston, Massachusetts.4Harvard University, Cambridge, Massachusetts. 5Department of Pathology,Harvard Medical School, Boston, Massachusetts. 6Department of Pathology,Beth Israel Deaconess Medical Center, Boston, Massachusetts. 7CambridgeResearch Institute, University of Cambridge, Cambridge, United Kingdom.
Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).
G. Peluffo and A. Subedee contributed equally to this article.
Current address for A. Subedee: NIH, Rockville, Maryland; current address forY. Su: Deciphera Pharmaceuticals, Waltham, Massachusetts; current address forF. Beca: Stanford University, Stanford, California; current address for A. Marusyk:Moffitt Cancer Center, Tampa, Florida; andcurrent address forA. H. Beck: PathAI,Cambridge, Massachusetts.
Corresponding Author: Kornelia Polyak, Dana-Farber Cancer Institute, 450Brookline Ave, Boston, MA 02215. Phone: 617-632-2106, Fax: 617-582-8490,E-mail: [email protected]
Cancer Res 2019;79:4173–83
doi: 10.1158/0008-5472.CAN-18-3264
�2019 American Association for Cancer Research.
CancerResearch
www.aacrjournals.org 4173
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
further characterized EN1, a TF with known roles in brain (5, 6)and dermomyotome (7) development.
Materials and MethodsCell lines
Breast and colon cancer cell lines were obtained from ATCC orgenerously provided by Steve Ethier (SUM149 and SUM159 celllines, University of Michigan, Ann Arbor, MI). Cells were culturedin media recommended by the provider, except SUM149 andSUM159 cells were cultured in DMEM/F12 supplementedwith 5% FBS, 10mmol/L HEPES pH7.4, 1 mg/mL hydrocortisone,5mg/mL insulin, 50U/mLpenicillin, and 50 mg/mL streptomycin.MCF7 cells were cultured in 4.5mg/L glucose DMEM supplemen-ted with 10% FBS, 10 mg/mL insulin, 50 U/mL penicillin, and50mg/mL streptomycin. Cellswere cultured at 37�Cwith 5%CO2.The identities of the cell lines were confirmed by short tandemrepeat analysis; and they were regularly tested for Mycoplasma.
Generation of cell line derivatesSUM149 and SUM159 cells expressing TET-inducible short
hairpin RNAs (shRNA) targeting EN1, CTNNB1, or NLGN4X inpLKO lentiviral vector were generated by selecting with 5 mg/mLpuromycin for 5 days after lentiviral infection. Entry cDNA openreading frame (ORF) for EN1 in pENTR221 was obtained fromhuman ORFeome collection v5.1. pCDNA3-CTNNB1 andpCDNA3-CTNNB1S33Y were obtained from Addgene. Lentiviralexpression constructs were generated by Gateway swap intopLenti6.3/V5-Dest Vector (Life Technologies) and sequence ver-ified. MCF7-lacZ andMCF7-EN1 cells were selected with 5 mg/mLblasticidin.
Colony growth assaysA total of 500–1,500 cells expressing TET-inducible EN1- or
CTNNB1-targeting shRNAs were plated into each well of a 6-wellplate. Next day, regular media (no doxycycline control) ormedia containing 500 ng/mL doxycycline were added to induceshRNA-mediated downregulation. The media was changed every2 days and colonies stained with 0.5% crystal violet solution after10–15 days.
Xenograft assaysAnimal experiments were approved by DFCI Institutional
Animal Care and Use Committee under protocol #11-023.SUM149 or SUM159 cells (1 � 106) expressing TET-inducibleluciferase or EN1-targeting shRNAs were resuspended in 50%Matrigel (BD Biosciences) and injected orthotopically into themammary fat pads of 6-week-old female NOG mice (Taconic).When the tumors became palpable, shRNAs were induced in thetreatment group by administering a doxycycline diet (625 ppm).Animals were euthanized and tumors were harvested whentumors in the control group reached approximately 1.5 cm size.For histologic analyses, 5-mm sections of formalin-fixed, paraffin-embedded tissue slides were stained with hematoxylin and eosinusing standard protocols.
Cell-cycle analysesThirty-six hours after induction of EN1 shRNA with doxycy-
cline, these cells and the control cells were treated with nocoda-zole (200 ng/mL) for 12 hours. Plates were tappedmultiple timesto detach cells arrested in G2–M-phase. After washing twice with
PBS, cells were plated in fresh medium into collagen-coatedplates. Cells were collected at 0, 3, 9, 12, and 24-hour timepointsfor FACS and immunoblot analyses. For cell-cycle analysis, cellswere harvested, washed in PBS, and fixed in ice-cold 70% ethanolat �20�C overnight. Fixed cells were resuspended in a solutioncontaining 100 mg/mL RNase and incubated for 30 minutes at37�Cwith agitation. The cells were then resuspended in a solutioncontaining 40 mg/mL propidium iodide (Sigma), and the analysiswas performed on a FACS Aria II Cytometer (BD Biosciences).
METABRIC expression and survival analysisExpression analysis was performed on METABRIC dataset.
Using the EN1 mRNA expression distribution we separated EN1expression into four clusters using a univariate Gaussian mixturemodel–based clustering (mclust version 5.4.2 package for R;refs. 8–10). Using the EN1-defined clusters, Kaplan–Meier sur-vival curves were plotted and a log-rank P value was computedusing the function km.coxph.plot in the R package survcomp.Next, the survival in the C4 cluster was compared with the otherclusters. A univariable Cox proportional hazards analysis wasperformed using the coxph function in R to assess the associationof EN1 mRNA expression with overall survival. A multivariableCox proportional hazard analysis was performed using the coxphfunction in R, and the age, the Nottingham Prognostic Index, andcontinuous EN1 expression were considered. The EN1 expressionz-scores of 243 triple-negative tumors from theMETABRIC cohortwere computed and the mean levels in the primary tumor forpatients that had a brainmetastasis (n¼ 17) and patients that didnot (n¼ 226) were compared using a two-sided t test. Differentialgene expression was performed using two-class unpaired signif-icance of microarray analysis (SAM 2.0 package, R 3.2.2) in basalcarcinomas in the C1, C2, and C3 cluster and the C4 cluster. Next,we performed a preranked GSEA (gene set enrichment analysis)using software providedby theBroad Institute (http://www.broadinstitute.org/gsea/msigdb/annotate.jsp) on a ranked gene listranked (after exclusion of EN1 gene) based on the d-statisticcomputed from the differential expression analysis, and weassessed enrichment using the Broad Institute's Molecular Signa-tures Database (mSigDB) Hallmark gene sets collection (n¼ 50).
Statistical analysisUnpaired two-tailed Student t tests were used except otherwise
stated. P < 0.05 was defined as statistically significant.
Accession codesGene Expression Omnibus (GEO): RNA-seq and chromatin
immunoprecipitation–sequencing (ChIP-seq) datasets have beendeposited to GEO with accession number GSE120957.
ResultsTNBC-specific transcription factors
We previously analyzed the inheritance pattern of basal/mesenchymal and luminal phenotypes by generating and char-acterizing somatic cell fusions between luminal and basal/mesenchymal breast cancer cell lines (11). We determined thatthe basal/mesenchymal phenotype is dominant, and it is definedby epigenetic factors. We also identified 17 TFs, the expressionof which strongly correlated with the inheritance of basal/mesenchymal phenotype and was also significantly higher inTNBCs compared with other subtypes in the The Cancer Genome
Peluffo et al.
Cancer Res; 79(16) August 15, 2019 Cancer Research4174
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
B
A C
D
E
F
% Viability
HC
C38
HC
C70
AIM2
CEBPBCEBPD
DEPDC7EN1
ETV5FOSL1
HMGA1IFI16JAG1
MED30MYBL1PPARDTBX18
TRIP13ZNF639
ATF3
BT-
549
HC
C19
54H
S57
8TM
DA
MB
231
MD
AM
B46
8S
UM
159
MC
F-10
AS
UM
149
MC
F-12
A
BT4
74
MC
F7ZR
75-1
T47D
MD
AM
B45
3S
UM
190
SK
BR
3
1000 1000 1000
0 1 2 3 4 5Days
shLucshLuc + doxshEN1-1shEN1-1 + doxshEN1-2shEN1-2 + dox P < 0.0001
P < 0.0001
0 1 2 3 4 5Days
P < 0.0001P < 0.0001
0 1 2 3 4 50
10
20
30
40
50
60
Days
Cel
l num
ber (
×103 )
P < 0.0001P < 0.0003
SUM149
SUM159
MCF7
0102030405060
Cel
l num
ber (
×103 )
0102030405060
Cel
l num
ber (
×103 ) 70
708090
0 h 3 h 9 h 24 hTime after release from G/2M arrest
-+
shE
N1-
2D
oxG
METABRIC dataset
EN1
mR
NA
Expr
essi
on (z
-sco
re)
0
2
4
6
Basal HER2 LumA LumB Norm UnclassClau199 220 679 461 140 6199
Non-TNBC (n = 364)TNBC (n = 42)
Row maxRow min
AIM2
CEBPBFOSL1ETV5
HMGA1
EN1
MYBL1ZNF639
MED30PPARD
DEPDC7
CEBPD
TRIP13
TBX18
IFI16JAG1ATF3
0
200
400
600
800
1,000
1,200
+- - -+ +
P = 0.0077 P < 0.0001
Doxycycline
5 10 15 20 25 300
200
400
600
800
1,000
1,200
Days
0200400600800
1,0001,2001,400
+- - -+ +
SUM149
5 10 15 20 25 300
200400600800
1,0001,2001,400
Days
SUM159
Doxycycline
P < 0.0001 P < 0.0001
shlucshluc +doxshEN1-1shEN1-1 +doxshEN1-2shEN1-2 +dox
shlucshluc +doxshEN1-1shEN1-1 +doxshEN1-2shEN1-2 +dox
shluc shEN1-1 shEN1-2 shluc shEN1-1 shEN1-2
Tum
or w
eigh
t (g)
Tum
or w
eigh
t (g)
Tum
or w
eigh
t (g)
Tum
or w
eigh
t (g)
shEN1-1 shEN1-2
SUM
149
SUM
159
- -+ +DoxshLuc
- +
SMADAPI
SMADAPI
Figure 1.
TFs in nonluminal breast cancers. A, Expression of selected TFs in TNBC and non-TNBC breast tumors in the TCGA dataset. Colors reflect z-scores normalized todepict relative values for each factor. B, Heatmap depicting cell viability siRNA screen for 17 TFs in 18 breast cancer cell lines: 9 nonluminal (red), 2 immortalizedmammary epithelial cell lines (orange), and 7 luminal (blue). Shading indicates percentage (0%–100%) of viable cells after siRNA-mediated downregulation ofthe selected factors compared with nontargeting siRNA control. C, Boxplot depicting the expression of EN1 in breast tumor subtypes (basal, claudin-low, HER2,luminal A, luminal B, normal-like, and unclassified) in the METABRIC dataset. Numbers indicate number of tumors in each subtype. D, Proliferation of SUM149,SUM159, and MCF7 cell lines following downregulation of EN1 by TET-inducible shRNAs. shlucwas used as control. E, Plots depicting tumor volume and weightsat 30 days after injection of TET-inducible shluc- or shEN1-expressing SUM149 and SUM159 cells, respectively, in the presence (þDox) and absence (�Dox) ofdoxycycline. F, Immunofluorescence staining of SMA expression in SUM149 and SUM159 xenografts expressing TET-inducible shluc or shEN1 grown in presence(þ) or absence (�) of doxycycline. Scale bars, 50 mm. G, Cell-cycle profile of SUM159 cells expressing TET-inducible shEN1 grown in presence (þ) or absence (�)of doxycycline, synchronized in G2–M-phase, and analyzed at different timepoints (0, 3, 9, and 24 hours) after release.
EN1 and Breast Cancer Brain Metastasis
www.aacrjournals.org Cancer Res; 79(16) August 15, 2019 4175
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
Atlas (TCGA) dataset (Fig. 1A; ref. 12) and in basal comparedwithother breast tumors in the Gene Expression-Based Outcome forBreast Cancer Online (GOBO) dataset (Supplementary Fig. S1A;ref. 13). To determine whether these TFs are specifically requiredfor the survival or proliferation of TNBC cells, we performed atargeted siRNA screen of all 17 TFs in 18 breast cancer cell lines ofdifferent subtypes. Downregulation of multiple factors (CEBPB,EN1, JAG1,MED30, PPARD, and TRIP13) led to amore than 50%decrease in viable cell numbers in multiple basal/mesenchymalcell lines with only modest reduction of viability in a few luminallines (Fig. 1B; Supplementary Fig. S1B). This subtype specificity ofdifference in cellular viability was statistically significant (P <0.05) for EN1, PPARD, and TRIP13. EN1 displayed the mostTNBC-specific expression pattern in primary tumors and themostsignificant and specific effect on viability in basal/mesenchymalversus luminal breast cancer cell lines; thus, we investigated it infurther detail.
High expression and essential role of EN1 in TNBCWe further explored the expression of EN1 in primary breast
tumors by analyzing the METABRIC (14), TCGA, and GOBOdatasets, and confirmed its specific and high expression in basalbreast tumors compared with other subtypes (Fig. 1C; Supple-mentary Fig. S1C). Similarly, EN1 was highly expressed in non-luminal compared with luminal breast cancer cell lines (Supple-mentary Fig. S1D) and there was a highly significant correlation(R2¼ 0.85; P < 0.0001) between EN1mRNA levels and the degreeof loss of cellular viability following its downregulationby siRNAs(Supplementary Fig. S1E). We could detect neither endogenousEN1 protein levels nor the effect of shEN1 on EN1 protein due toour inability to reliably detect EN1 with any of the 13 commer-cially available EN1 antibodies tested (see Supplementary Mate-rials and Methods).
We validated the results of siRNA-mediated downregulation ofEN1by TET-inducible EN1-targeting shRNAs (Supplementary Fig.S1F). shRNA mediated downregulation of EN1 in SUM149 basaland SUM159 mesenchymal TNBC lines, but not in non-TNBCMCF7 luminal cells showed significant reduction in cellularviability (Fig. 1D). Overexpression of a constitutively expressedEN1 in SUM159 cells was able to partially rescue this growthinhibition in a dose-dependentmanner confirming the specificityof shEN1 (Supplementary Fig. S1G). Unfortunately, we could notidentify an shRNA targeting the noncoding region of EN1, thus,the overexpressed EN1was still sensitive to the shRNA, leading topartial rescue likely due to this.Downregulation of EN1 also led tosignificant reduction of colony forming ability of SUM149 andSUM159 cells, but not in theMDA-MB-231TNBCcell line that hasno detectable levels of EN1 mRNA further supporting that thedecrease in cell viability after shEN1 expression is due to down-regulation of EN1 (Supplementary Figs. S1D and S2A).
To determinewhether EN1 is essential for tumor growth in vivo,we performed xenograft studies using SUM149 and SUM159cells expressing two independent TET-inducible EN1-targetingshRNAs. Xenografts were allowed to grow to palpable size beforeinducing shRNA expression (Supplementary Fig. S2B). Down-regulation of EN1 significantly reduced tumor weight, volume,and cellularity in xenografts derived from both SUM149 andSUM159 cell lines (Fig. 1E; Supplementary Fig. S2C). Immuno-fluorescence analysis of smooth muscle actin (SMA), a marker ofstromal myofibroblasts and myoepithelial cells, demonstrated asignificant increase in SMAþ cells within tumors following EN1
downregulation (Fig. 1F). To determine whether this is due toincreased recruitment of mouse stroma, we performed immuno-FISH using mouse and human-specific DNA probes combinedwith SMA immunofluorescence and confirmed that SMAþ cellsare mousemyofibroblasts (Supplementary Fig. S2D). In line withthis, IHC analysis of p63, a basal cell–specific TF, showed nosignificant change in expression following EN1 downregulation(Supplementary Fig. S2E).
To investigate mechanisms underlying EN1 loss–induceddecrease in cellular viability, we analyzed the cell-cycle profile ofsynchronized SUM159 cells following downregulation of EN1.SUM159 cells expressing TET-inducible EN1-targeting shRNAswere synchronized in G2–M-phase with nocodazole 36 hoursafter induction of shEN1. Upon release from G2–M-phase block-ade by washing off nocodazole and replating the cells in freshmedium, cells with downregulation of EN1 arrested in G1, where-as themajority of control cells progressed to S andG2–M-phase by9 and 24 hours, respectively (Fig. 1G). The relatively low levels ofphospho-histone H3 (phosphorylated in metaphase of mitosis)in control cells at 24 hours (Supplementary Fig. S2F) despite thehigh fraction of cells with 4n DNA content based on FACS couldbe due to the cells being more in G2/early–M-phase than in laterphases ofmitosis when histoneH3 is phosphorylated. The down-regulation of EN1 in synchronized cells led to a notable reductionof CCND1 and c-MYC protein levels that was themost significantat later (9 and 24 hour) time-points after release when controlcells entered S/G2–phase of the cell cycle (Supplementary Fig.S2F). This decrease in cyclin D1 and c-MYC could directly play arole in the EN1 loss–mediated growth suppression. CCND1 andc-MYC are targets of the WNT signaling pathway and EN1 is aknown WNT target during brain development (15). Thus, thedecrease in CCND1 and c-MYC levels following EN1 loss inSUM159 cells could potentially indicate downregulation of WNTsignaling. At later timepoints following EN1 downregulation, weobserved an increase in apoptosis based on an increase in sub-G1
population by FACS (Supplementary Fig. S2G and S2H) andincrease in cleaved caspase-3 levels (Supplementary Fig. S2I).These results correlate with the antiapoptotic function of EN1 indopaminergic neurons and suggest similar roles for EN1 inTNBC (16).
Transcriptional and genomic targets of EN1 in TNBCsNext, we examined changes in gene expression profiles of
SUM149 and SUM159 cells expressing TET-inducible shEN1 at3 days and 5 days after doxycycline treatment. A large fraction ofgenes displayed similar expression changes in both cell linesanalyzed (Fig. 2A; Supplementary Fig. S3A and S3B; Supplemen-tary Table S1). We also performed RNA-seq on SUM149 andSUM159 xenografts from control and doxycycline-treated mice.The number of differentially expressed genes was significantlyhigher in cell culture at relatively early timepoints after EN1downregulation comparedwith the xenografts that were collected30 days after injection into mice. Furthermore, residual tumorsobserved in doxycycline-treated mice apparently escaped theshRNA effect based on the lack of significant differences in EN1mRNA levels between control and doxycycline conditions (Sup-plementary Table S1). Functional analysis of genes differentiallyexpressed after EN1 knockdown in vitro or in vivo using Meta-core (17) revealed enrichment in development, inflammation,and cell adhesion/migration-related pathways including WNTand Hh signaling, and neurogenesis (Fig. 2B). We analyzed the
Peluffo et al.
Cancer Res; 79(16) August 15, 2019 Cancer Research4176
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
A B
C
E
SUM149 log2 fold change SUM149 log2 fold change
−4
−2
0
2
4
6
−5.0 −2.5 0.0 2.5 5.0 7.5
SU
M15
9 lo
g 2 fo
ld c
hang
e
CASP14
DMGDH
ESM1
MMP1
MMP3
OR7A5PIPOX
RPSAP52
SERPINB2
SLC30A2
−4
0
4
8
−5 0 5
SU
M15
9 lo
g 2 fo
ld c
hang
e
Day 3 Day 5
CRYM
GBP6
GGT8PMMP1
MMP3
HLHSCEL
CDKN3
CYP4B1STAC2
FGene distance (kbp)
SUM149 SUM159
−2.0 2.0 −2.0 2.0
0
1
2 123
1 (2
,971
)2
(840
)3
(383
)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
−2.0 2.0 − 2.0 2.0
SUM149
WISP1
SUM159
44 kb
92 kbSUM149
SUM159
TCF7L1
D
GnRH signaling pathway
Transmission of nerve impulseMIF signaling
Attractive and repulsive receptors
WNT signalingProgesterone signalingGonadotropin regulationFSH−beta signaling pathway
Connective tissue degradation
Regulation of angiogenesisNeurogenesis_synaptogenesisNeurogenesis_axonal guidanceBlood vessel morphogenesis
Synaptic contactCell−matrix interactions
0 2−log10 (FDR)
015
30
Color key
Cou
nt
SU
M14
9S
UM
159
Down
SU
M14
9S
UM
159
Up
Develop
Cell adhesion
proteolysisR
eproductionsignalingneurop hysiology
SUM159
SUM149
0 500 1,000 1,500 2,000 2,500 3,000 3,500
0
20
40
60
80
100
Rank of genes based on Regulatory Potential Score
Cum
ulat
ive
fract
ion
of g
enes
% Static (background)Upregulate (0.964)Downregulate (0.307)
0 1,000 2,000 3,000 4,000 5,000 6,000
0
20
40
60
80
100
Rank of genes based on Regulatory Potential Score
Cum
ulat
ive
frac
tion
of g
enes
% Static (background)Upregulate (1.1e−05)Downregulate (0.000276)
Cyt
oske
leto
n_ac
tin fi
lam
ents
Ske
leta
l mus
cle
deve
lopm
ent
Cel
l−m
atrix
inte
ract
ions
EC
M re
mod
elin
g
Blo
od v
esse
l mor
phog
enes
is
Con
nect
ive
tissu
e de
grad
atio
n
Inte
rfero
n si
gnal
ing
IL−6
sig
nalin
gC
ompl
emen
t sys
tem
Spi
ndle
mic
rotu
bule
s
Th17
−der
ived
cyt
okin
esA
mph
oter
in s
igna
ling
Pla
tele
t−en
doth
eliu
m−l
euco
cyte
inte
ract
ions
Reg
ulat
ion
of a
ngio
gene
sis
Neu
roge
nesi
s_ax
onal
gui
danc
e
Pro
tein
C s
igna
ling
Neu
troph
il ac
tivat
ion
His
tam
ine
sign
alin
gIn
nate
infla
mm
ator
y re
spon
seK
allik
rein
−kin
in s
yste
m
Hed
geho
g si
gnal
ing
WN
T si
gnal
ing
FGF_
Erb
B s
igna
ling
Cyt
opla
smic
mic
rotu
bule
s
Inna
te im
mun
e re
spon
se to
RN
A v
iral i
nfec
tion
IL−1
0 an
ti−in
flam
mat
ory
resp
onse
MIF
sig
nalin
g
Neu
roge
nesi
s_sy
napt
ogen
esis
Car
tilag
e de
velo
pmen
t
Syn
aptic
con
tact
Am
yloi
d pr
otei
ns
Neu
roge
nesi
s in
gen
eral
Reg
ulat
ion
of e
pith
elia
l−to
−mes
ench
ymal
tran
sitio
n
Reg
ulat
ion
of c
ytos
kele
ton
rear
rang
emen
tIn
term
edia
te fi
lam
ents
Cel
l jun
ctio
nsA
ttrac
tive
and
repu
lsiv
e re
cept
ors
Inte
grin
−med
iate
d ce
ll−m
atrix
adh
esio
n
0 8
030
0C
ount
DevelopmentCell adhesioncytoskeleton,proteolysis
Inflammationimmune resp.
shE
N1-
2
SUM149
SUM159
Day3Day5Day3Day5
shE
N1-
2
SUM149
SUM159
Day3Day5Day3Day5
in v
ivo
shEN1-2 SUM149shLuc
shEN1-2 SUM159shLuc
in v
ivo
shEN1-2 SUM149shLuc
shEN1-2 SUM159shLuc
Dow
nin
vitr
oin
vitr
oU
p
−log10 (FDR)
Figure 2.
Transcriptional and genomic targets of EN1 in TNBC cells. A, Gene expression changes in SUM149 (x-axis) and SUM159 (y-axis) cells expressing TET-inducibleshEN1 grown in presence of doxycycline for 3 or 5 days. Red and blue indicates significantly up- and downregulated genes, respectively, in both cell lines.B, Top process networks significantly enriched in genes differentially expressed following EN1 downregulation in SUM149 and SUM159 cell lines grown in cellculture (in vitro) or as xenografts (in vivo). C, Plot of ChIP-seq signal of the EN1 peaks in SUM149 and SUM159 cells. Numbers indicate cell line unique andoverlapping peaks. D, BETA factor function prediction analysis to predict EN1 direct transcriptional function (activating/repressing) in SUM149 and SUM159 cells.Red and blue lines, up- and downregulated genes following EN1 knockdown. Black dashed line, static genes. Genes are ranked by regulatory potential score, andsignificance relative to static genes was determined by Kolmogorov–Smirnov test. E, Top process networks significantly enriched in genes that were directgenomic targets of EN1 and were differentially expressed 3 days following EN1 downregulation in SUM149 and SUM159 cell lines. Each ChIP-seq peak is assignedto the gene whose transcription start site is closest to the peak, defining the factors direct genomic targets. F, EN1 ChIP-seq signal at selected genes in SUM149(red) and SUM159 (green) cells. Thick black lines below the signal tracks denote peaks.
EN1 and Breast Cancer Brain Metastasis
www.aacrjournals.org Cancer Res; 79(16) August 15, 2019 4177
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
expression of genes involved in WNT signaling in further detailand found that the majority showed decreased levels after shEN1expression in both cell lines and timepoints highlighting theimpact of EN1 expression on WNT pathway activation (Supple-mentary Fig. S3C). We validated the expression of WISP1 andWISP2, direct transcriptional targets of b-catenin (3), by immu-noblot analysis of cell lysates fromcell cultures and xenografts andconfirmed their decline after shEN1 expression (SupplementaryFig. S3D and S3E).
To identify the direct genomic targets of EN1, we performedChIP-seq in SUM149 and SUM159 cells for EN1 and histone H3lysine 27 acetyl (H3K27ac) associated with transcriptionallyactive regions to define super enhancers (SE) (18). The validityof the EN1 ChIP-seq data was confirmed based on (i) the datapassing all quality control in theChiLin 2.0 pipeline, (ii) ChIP-sequsing the antibody against endogenous EN1 and V5-tagged EN-1showed high correlation in all three cell lines tested (Supplemen-tary Fig. S3F), (iii) the number of peaks correlated with endog-enous EN1 levels; for example, SUM149 cells have higher endog-enous EN1 than SUM159 cells and have more EN1 ChIP-seqpeaks than SUM159 cells (3,354 vs.1,223), and (iv) the peaksshowed enrichment in EN1 sequence motif (Supplementary Fig.S3G). The majority of EN1 peaks were located in nonpromoterregions in both cell lines (Supplementary Fig. S3H) and a signif-icant fraction overlapped between the two cell lines with SUM149cells havingmore unique peaks consistent with the higher expres-sion level of EN1 in this cell line (Fig. 2C). EN1 and H3K27acpeaks also demonstrated significant overlap in both cell lines(Supplementary Fig. S3I). Analysis of SEs for core regulatorycircuits (19) revealedZNF217, POU3F3, ELF5, andKLF4asmasterTF key for the establishment of the SE landscape in SUM159 cells,whereas in SUM149 cells it uncovered a larger set of TFs includingEN1 that formed an interconnected network centered aroundMYC based on STRING protein interaction network analysis(Supplementary Fig. S3I; ref. 20).
Next, we explored whether the up- or downregulated geneswere more likely to be direct EN1 targets at different timepointsafter EN1 downregulation. On the basis of the fraction of directEN1 targets up- or downregulated (Supplementary Fig. S3K) andon BETA analysis (Fig. 2D; Supplementary Fig. S3L), EN1 appearsto both positively and negatively regulate gene expression in bothTNBC cell lines with more genes being up than downregulated inSUM149 cells. The top process networks enriched for direct EN1targets that were up- or downregulated after shEN1 expression,were essentially the same as observed for differentially expressedgenes with WNT signaling and neurogenesis-related functionsshowing the most significant enrichment (Fig. 2E). Examples forEN1 targets include TCF7L1 transcription factor and WISP1-secreted growth factor mediating WNT signaling (Fig. 2F; Sup-plementary Table S2). Thus, similar to brain anddermomyotome,EN1 may modulate WNT signaling and neural-related functionsin TNBC (15).
To investigate the EN1–WNT pathway link in more detail, wetested the effects of wild-type or a constitutively active S33Ymutant b-catenin overexpression or CTNNB1 downregulation inSUM159 and SUM149 cells. Endogenousb-catenin andphospho-b-catenin levels were higher in SUM149 compared with SUM159cells, but downregulation of EN1 had no effect on these in eithercell line (Supplementary Fig. S4A). Overexpression of CTNNB1and CTNNB1S33Y in SUM159 cells had no effect on cellularviability and did not rescue the EN1 loss–induced growth arrest
(Supplementary Fig. S4A–S4C). Similarly, downregulation ofCTNNB1 in SUM159 cells did not affect cellular viability andcolony growth (Supplementary Fig. S4D–S4F). Finally, wetested the effect of several compounds targeting different stepsof the canonical WNT signaling pathway (i.e., porcupine,tankyrase, and WNT inhibitors, and Axin2 activator) on thegrowth of TNBC (SUM149 and SUM159, high EN1 expression)and luminal ERþ (MCF7 and T47D, no EN1 expression) breastcancer cells, but did not find any significant subtype-specificdifferences or associations with the degree of growth inhibitionand endogenous b-catenin or EN1 levels (Supplementary Fig.S4G and S4H). Thus, EN1 appears to modulate WNT signalingin TNBC cells via alternative mechanisms not directly involvingCTNNB1 itself.
The effect of EN1 expression in luminal breast cancer cellsOur prior data demonstrated that EN1 was one of the few TFs
that were able to switch MCF7 luminal breast cancer cells to amorebasal/mesenchymal phenotype (11). To investigatewhetherthis is due to the reprogramming of the epigenetic landscape byEN1,we expressed EN1 inMCF7 cells and confirmed the switch toamore mesenchymal phenotype based onmorphologic changes,as well as increased migration and invasion (Fig. 3A), althoughthe cells maintained their estrogen dependence for growth (Sup-plementary Fig. S5A). Interestingly, EN1 expression did not havethe same effect in the T47D luminal breast cancer cell linehighlighting the importance of cellular context (SupplementaryFig. S5B). Similar context-dependentobservationswere alsomadefor several other EMT-inducing TFs including ZEB1 andTWIST (21). However, genes differentially expressed after EN1downregulation in SUM159 and SUM149 cells that are also directEN1 targets did classify primary breast tumors into basal andnonbasal subsets highlighting the role of EN1 and its targets inregulating luminal and basal subtypes (Supplementary Fig. S5C).
To examine whether EN1 binds to and regulates the sameset of genes in MCF7 cells as it does in TNBC cell lines, weperformed RNA-seq and EN1 ChIP-seq comparing MCF7-lacZand MCF7-EN1 cells. We found that a subset of EN1 peaksoverlapped among the three cell lines, but there were numerousMCF7-specific EN1 peaks aswell. (Fig. 3B). Integrating EN1ChIP-seq and RNA-seq data demonstrated that EN1 targets in MCF7cells include genes both up- and downregulated after EN1 expres-sion (Supplementary Fig. S5D and S5E; Supplementary Table S3).Functional analysis of differentially expressed genes that are EN1targets showed higher number of process networks enriched indownregulated genes after EN1 overexpression including manydevelopment- and neurogenesis-related pathways such as axonalguidance, WNT, and NOTCH signaling (Fig. 3C).
To determine whether the luminal-to-mesenchymal switchobserved in EN1-expressing MCF7 cells is due to the EN1'smodulation of chromatin binding of luminal TFs such as FOXA1,a lineage determining TF, we performed FOXA1 ChIP-seq onLacZ- and EN1-expressing MCF7 cells. Overall there was a sub-stantial overlap between FOXA1 and EN1 peaks, and whilewe identified some FOXA1 peaks that were only detected inMCF7-lacZ or in MCF7-EN1 cells (Supplementary Fig. S5F andS5G; Supplementary Table S4), themajority of FOXA1 peakswerethe same in the two cells. These data imply that EN1 expressiondoes not significantly affect FOXA1 chromatin binding, althoughit might still modulate FOXA1 transcriptional activity for a subsetof its target genes.
Peluffo et al.
Cancer Res; 79(16) August 15, 2019 Cancer Research4178
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
A C D
LacZ EN10
100
200
300
400
500N
umbe
r of m
igra
ting
cells
MigrationP < 0.0001
LacZ EN10
20
40
60
80
100
120
140
Num
ber o
f inv
adin
g ce
lls
InvasionP = 0.0362
E
F
G
SUM159
FANCI
SIN3ATRIM33
CHD4
MSH6
MNAT1
PRPF31
GNL1
FAM192A
NCOR1RSRC2EN1
PSMD5
LONP1ATG2ASEC13
TLE3
HDAC2
PREBUSP7
ANP32EFKBP5
IRX2
CALCOCO2UBAP2
XRCC6
MORC3XPO5
PPM1GPXN
CC2D1BTRIM28
PUS1
NOSIPPSME3WDHD1
HTATSF1
ZYX
TBL1XR1KIF4A
CEBPB
RPA1
ZMYND8
MAGEC2
RNH1
TRIM24
HCFC1
MTA2
AKAP13
PUF60
MCF7
SMAD3TLE3EN1TLE1
CEBPB
CSE1L
MSH6
USP7GNL1TRIM28TRIM33
PUS1
TBL1XR1RTFDC1
ZNF217PPM1G
NOSIPDPF2
CHD4
NELFB
MORC3
FEN1
NCOR2
SART3RNH1
WIZ
FANCIXPO5
NAB2
GTF2I
GTF3C1
MSH2
CBX3
PPP1R8
WDHD1OGT
PSMD5SIN3A
HAT1
SUDS3
TRPS1
ZMYND8
CBX1
RXRAANP32E
GRHL2
HTATSF1
TRIM24
ZMYM2CBX5
0 1 2 3
Signal transductionCholecystokinin signaling
Cell adhesionIntegrin priming
Muscle contraction
ReproductionProgesterone signaling
InflammationHistamine signaling
Signal transductionWNT signaling
Chemotaxis
InflammationNK cell cytotoxicity
DevelopmentBlood vessel morphogenesis
Development_neurogenesisAxonal guidance
InflammationInterferon signaling
Cell adhesionAttractive and repulsive receptors
-log(FDR)
DownUp
Peak distance (kbp)
MCF7 SUM149 SUM159
−2.0 2.0 −2.0 2.0 −2.0 2.0Peak distance (kbp) Peak distance (kbp)
0
21 (2,070)2 (2,843)3 (823)4 (126)5 (17)6 (365)7 (18)
12
34
56
7 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
-2.0 2.0 −2.0 2.0 −2.0 2.0
B
0 1 2−log10(FDR)
04
8C
ount
FSH−beta signaling pathway
TGF−beta, GDF and Activin signaling
Cartilage developmentNeurogenesis_axonal guidance
Cholecystokinin signalingNOTCH signaling
Regulation of angiogenesisG1−S Interleukin regulation
Keratinocyte differentiation
Negative regulation of cell proliferation
Blood vessel morphogenesis
G1−S Growth factor regulationG0−G1
WNT signaling
Down Up
H
Signal transduction
TranscriptionD
NA dam
ageC
ell cycleA
poptosis
Transcription by RNA polymerase II
Chromatin modification
ESR2 pathway
ESR1−Nuclear pathway
Androgen receptor nuclear signaling
MMR Repair
DBS Repair
Core
Checkpoint
BER−NER repair
S-phase
Mitosis
Meiosis
Apoptotic nucleus
0 4 8
04
8
Color key
Cou
nt
−log10(FDR)
MC
F7
SU
M15
9
Peak distance (kbp)−2.0 2.0 −2.0 2.0 −2.0 2.0
0
2
12
34
765
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Control shEN1 EN1 peaks
-2.0 2.0 −2.0 2.0 −2.0 2.0
1 (11,428)2 (7,150)3 (616)4 (16,604)5 (24)6 (152)7 (428)
TLE3 peaks
SUM149 SUM159
Dow
nU
pD
own
Up
Dow
nU
pD
own
Up
+Dox-Dox +Dox-Dox
Blood vessel morphogenesis
Connective tissue degradation
Histamine signaling
Cell−matrix interactions
Progesterone signalingFSH−beta signaling pathway
Kallikrein−kinin system
GnRH signaling pathwayGonadotropin regulation
Regulation of cytoskeleton rearrangementCadherins
Regulation of angiogenesisOssification and bone remodeling
Th17−derived cytokinesIL−10 anti−inflammatory response
Anti−apoptosis by ext signals via MAPK and JAK/STATAnti−apoptosis by ext signals via PI3K/AKT
Neurogenesis_axonal guidanceCartilage development
Platelet aggregationECM remodeling
Negative regulation of cell proliferation
Feeding and neurohormone signaling
MIF signaling
Reproduction
Developm
entC
ell adhesionC
ytoskeleton,P
roteolysisInflam
mation
Imm
une resp. Proliferat.
Apoptosis
0 2
040
100
Cou
nt
−log10(FDR)
Figure 3.
EN1-associated chromatin complexes in breast cancer cells. A, Cell migration and invasion of MCF7 cells expressing lacZ or EN1. P value of difference indicated(t test). B, Plot of ChIP-seq signal of the EN1 peaks in SUM149, SUM159, and MCF7 cells. Numbers indicate cell line unique and overlapping peaks. C, Top processnetworks significantly enriched in genes that were direct genomic targets of EN1 and were differentially expressed following EN1 expression in MCF7 cells. D,Word clouds illustrating top EN1-associated proteins in MCF7 and SUM159 cells. E, Top process networks enriched in EN1-interacting proteins in MCF7 andSUM159 cells. F, Heatmap and clustering of TLE3 and EN1 peaks in SUM159 cells. G, Top process networks enriched in differentially expressed genes after EN1knockdown associated with TLE3-EN1 overlapping peaks in either control (-dox) or shEN1 (þdox) conditions. H, Top process networks enriched in b-cateninpeaks in SUM149 cells associated with genes differentially expressed after EN1 knockdown.
EN1 and Breast Cancer Brain Metastasis
www.aacrjournals.org Cancer Res; 79(16) August 15, 2019 4179
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
EN1 interacting proteins in breast cancer cellsDuring brain development, EN1 function is modulated by its
interaction with coactivators and corepressors such as PITX3 (22)and TLE3 (23). To identify EN1-associated factors in MCF7and SUM159 breast cancer cells that exogenously express EN1with a V5 tag, we performed qPLEX-RIME (quantitative multi-plexed rapid immunoprecipitation mass spectrometry of endog-enous proteins; ref. 24) using the anti-V5 antibody. We identifiednumerous EN1-associated proteins in both cell lines, which wealso confirmed by coimmunoprecipitation (Fig. 3D; Supplemen-tary Fig. S5H). EN1-associated proteins showed significant enrich-ment in chromatin modification, AR signaling, and DNA damagefunctional categories (Fig. 3E). Several of the EN1-interactingproteins common between the two cell lines represent transcrip-tional cofactors TLE3 and TRIM28 (25), and nuclear hormonereceptor coactivators TRIM33 and TRIM24. Interestingly TRIM28was recently identified as a critical regulator of TRIM24, anoncogene in prostate cancer (26) and TRIM24–TRIM28–TRIM33complex has previously been identified as a suppressor of murinehepatocellular carcinoma growth (27). Because of prior dataimplying a functional role for the TLE3–EN1 interaction duringdevelopment (23), we further explored the relevance of thisassociation in breast cancer cells.
To confirm that EN1 and TLE3 bind to the same genomicregions and to test whether downregulating EN1 changes TLE3chromatin binding patterns, we performed TLE3 ChIP-seq inSUM149 and SUM159 cells �/þ induction of shEN1. Wedetected a significant overlap between TLE3 and EN1 peaksand a subset of the TLE3 peaks showed decreased (11,428) orincreased (7,150) signal in shEN1-expressing cells (Fig. 3F;Supplementary Fig. S5I; Supplementary Table S5). To investi-gate the functional relevance of TLE3–EN1 interaction, weintegrated TLE3 ChIP-seq and differential gene expression aftershEN1 induction and performed Metacore analyses. We foundthat the most significant network process enrichment in bothSUM149 and SUM159 cells was in progesterone and GnRHsignaling, angiogenesis and ECM remodeling–related pathwaysby genes associated with TLE3 peaks present only in shEN1 cellsand upregulated after shEN1 (Fig. 3G). These data suggest thatEN1 may suppress the expression of these genes by preventingTLE3 binding.
To investigate whether the EN1-mediated changes in theexpression of genes involved in WNT signaling is due to directeffects on b-catenin transcriptional activity, we also performedb-catenin ChIP-seq in cells with or without EN1 expression andintegratedb-catenin chromatinbindingpatternswith gene expres-sion changes. We observed a significant enrichment in WNTsignaling pathway among genes that are direct b-catenin targetsare downregulated after shEN1, implying that EN1 lossmay affectb-catenin transcriptional activity (Fig. 3H). However, EN1 expres-sion did not significantly impact b-catenin chromatin binding,thus, EN1 may regulate the WNT signaling pathway via otherindirect mechanisms.
High EN1 is associated with poor prognosis and brainmetastasis in TNBC
To investigate the clinical relevance of EN1 in breast cancer, weanalyzed associations between EN1 expression and clinical out-come using the METABRIC and TCGA datasets. EN1 is highlyexpressed in TNBCs compared with non-TNBC and also in basalbreast cancers compared with other subtypes (Fig. 1A and C).
However, even within basal breast cancers there was a significantheterogeneity for EN1 expression and tumors could be dividedinto four groups basedonEN1 expression distribution usingfinitenormal mixture modelling (mclust package for R, version 5.4.2;Fig. 4A; Supplementary Fig. S6A; refs. 8, 9). Tumors with thehighest EN1 expression were enriched for WNT signaling andMYC target genes (Fig. 4B) and showed worse overall survival(Fig. 4C). A Cox multiple regression model including tumorgrade, size, and numbers of positive lymph nodes demonstratedstatistically significant (P ¼ 0.0085) association between metas-tasis-free survival and EN1 expression. Similarly, in the TCGAdataset, basal breast tumors with high EN1 expression had worseoverall survival than those with low EN1 levels, but this did notreach statistical significance likely due to the relatively smaller sizeof this cohort and more limited numbers of basal breast cancers(Supplementary Fig. S6B).EN1 expression significantly associatedwith overall survival in a Cox simple regressionmodel. Amultipleregression model adjusted for the Nottingham Prognosis Indexand age also showed significant association between overallsurvival and EN1 expression.
The short survival of patients with TNBC with high EN1expression suggested that these patients might be more likely todevelop brain metastases commonly associated with shortersurvival in patientswithbreast cancer. Indeed, patientswith TNBCwho developed brain metastases had primary tumors with sig-nificantly higher EN1 levels than those who did not (Fig. 4D). Thesignificant association with brain metastases was still observedwhen compared with patients with metastases in otherorgans (28). We also analyzed gene expression data of matchedprimary tumors and brain metastases and found EN1 transcrip-tional targets were significantly enriched in genes differentiallyexpressed between primary breast tumors and brain metastasis(Fig. 4E). These results suggest that high expression of EN1 inprimary TNBCs may increase the risk of brain metastases poten-tially due to its regulation of WNT signaling and neurogenesis-related pathways including the expression of several neuroliginsin these breast tumors. Neural activity–induced NLGN3 wasshown to promote the growth of high-grade brain tumors suchas glioblastoma and DIPG (29), thus, their upregulation inprimary TNBCs may select for tumors with preferential growthin the brain.
Interestingly, in our prior breast cancer somatic cell fusionstudies the only SNP that was significantly associated withthe inheritance of SUM159mesenchymal cell features was withintheNLGN4X gene (11), implying that NLGN4X plays an essentialrole in these cells. In both SUM149 and SUM159 cell lines theexpression of NLGN4X, and other neuroligins, decreased follow-ing EN1 knockdown (Supplementary Fig. S6C). To test whetherNLGN4X expression in SUM159 breast cancer cells may reflect adependency for this neural growth factor, we downregulated itsexpression using TET-inducible shRNAs (Supplementary Fig.S6D) and analyzed changes in cell viability and proliferation.We detected a significant decrease in cell number 5 days followinginduction of two independent shNLGN4X confirming thatNLGN4X positively affects SUM159 cell growth (Fig. 4F). Toinvestigate potential mechanisms bywhich loss of NLGN4X leadsto decreased cell growth, we analyzed gene expression changesafter shNLGN4X induction. We found that downregulated genesshowed significant enrichment in muscle contraction (e.g.,CAPN3, CALR3, and VIPR1), whereas upregulated genes wereenriched in cell adhesion (e.g., NRXN1, TNC) and neurogenesis
Peluffo et al.
Cancer Res; 79(16) August 15, 2019 Cancer Research4180
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
A
B
No brain metastasis(n = 226)
Brain metastasis(17)
0
2
4
6P = 0.006
EN
1 m
RN
A e
xpre
ssio
n z−
scor
e
P = 0.00180.00
0.25
0.50
0.75
1.00
0 2 4 6 8 10 12 14Time (years)
Ove
rall
surv
ival
pro
babi
lity
non−C4C4
251 207 164 127 102 82 59 2427 19 12 9 8 6 4 30 2 4 6 8 10 12 14
Time (years)
Number at risknon−C4
C4
C
1 2 3 4
EN
1 m
RN
A e
xpre
ssio
n
n = 114
n = 52
n = 135
n = 30
0
2,000
4,000
6,000
Expression clusters
Cel
l num
ber
8×106
6×106
4×106
2×106
0
shNLGN4X-1 shNLGN4X-2- + - + Dox 5 µg/mL
P < 0.001P < 0.001
D
E
Enr
ichm
ent s
core
Rank in ordered datasetRan
ked
list m
etric
(Pre
Ran
ked)
0.000
0.175
0.0250.0500.075
0.1250.150
0.100
−0.025−0.050
0 5,000 10,000 15,000 20,000 25,000
Zero cross at 12,412
Enrichment profile Hits Ranking metric scores
F
DNA_REPAIRAPICAL_JUNCTION
CHOLESTEROL_HOMEOSTASISMYC_TARGETS_V1
GLYCOLYSISTGF_BETA_SIGNALING
ESTROGEN_RESPONSE_EARLYAPICAL_SURFACE
HEME_METABOLISMESTROGEN_RESPONSE_LATEPI3K_AKT_MTOR_SIGNALING
PANCREAS_BETA_CELLSPEROXISOME
HYPOXIAUV_RESPONSE_DN
PROTEIN_SECRETIONIL6_JAK_STAT3_SIGNALING
P53_PATHWAYANGIOGENESIS
IL2_STAT5_SIGNALINGBILE_ACID_METABOLISM
REACTIVE_OXIGEN_SPECIES_PATHWAYXENOBIOTIC_METABOLISM
ANDROGEN_RESPONSEINTERFERON_ALPHA_RESPONSE
APOPTOSISMTORC1_SIGNALING
FATTY_ACID_METABOLISMOXIDATIVE_PHOSPHORYLATION
INFLAMMATORY_RESPONSETNFA_SIGNALING_VIA_NFKB
ADIPOGENESISCOAGULATION
KRAS_SIGNALING_UPALLOGRAFT_REJECTION
INTERFERON_GAMMA_RESPONSEEPITHELIAL_MESENCHYMAL_TRANSITION
COMPLEMENTMYC_TARGETS_V2G2M_CHECKPOINTMITOTIC_SPINDLE
E2F_TARGETSWNT_BETA_CATENIN_SIGNALING
KRAS_SIGNALING_DNSPERMATOGENESIS
0 1 2Normalized enrichment score
Low in cluster 4
High in cluster 4
TNBC with high EN1
Activation of WNT and Hh signalinghigh expression of neuroligins andneural-related proteins
Increased risk of brain metastasis Shorter survival
G
Figure 4.
EN1 expression in breast cancer and clinical outcome. A, Basal breast cancers in METABRIC dataset divided into four different clusters based on EN1mRNAexpression. The number of tumor samples in each cluster is indicated. B,GSEA of genes differentially expressed between tumors in cluster 4 versus clusters 1–3combined. C, Kaplan–Meier survival plot showing overall survival of patients with basal breast cancer in the METABRIC cohort classified into cluster 4 (C4) and allother clusters (cluster 1–3) combined based on EN1 expression. P value indicated the statistical significance of the observed difference.D, The expression of EN1in primary TNBCs with and without brain metastasis in the METABRIC cohort. E, GSEA analysis depicting enrichment of matched primary versus brain metastaticTNBC samples in genes that go up after EN1 knockdown in SUM149 and SUM159 cells (P¼ 0.01). F, Cellular viability after NLGN4X downregulation in SUM159cells. G, Schematic summary of our EN1 data.
EN1 and Breast Cancer Brain Metastasis
www.aacrjournals.org Cancer Res; 79(16) August 15, 2019 4181
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
axonal guidance (e.g., SEMA3B and SEMA7A) proteins (Supple-mentary Fig. S6E; Supplementary Table S6). These results supportour hypothesis that neural survival–related pathways play animportant role in a subset of TNBCs with high EN1 expression.
DiscussionIn a targeted cellular viability screen for TFs selected on the
basis of their high expression in TNBCs we identified EN1, aneural-specific homeo-domain TF (30) with functional andclinical relevance in TNBCs. During embryonic development,EN1 plays essential roles in mid-hindbrain pattern formation,axonal guidance, and neuron specification (5). EN1 is highlyexpressed in dopaminergic neurons both during developmentand in adulthood, and it is essential for their differentiation,maintenance, and survival (16). High expression of EN2 wasreported in prostate, breast, and ovarian carcinomas (31), andEN2 has an oncogenic function in breast cancer (32). Inagreement with our findings, a recent study also reported thatthe downregulation of EN1 reduces cellular viability in theSUM149 TNBC cell line (33).
We demonstrated an oncogenic role for EN1 in TNBCs basedon the observation that its downregulation leads to G1 arrestand apoptosis, and reduced tumor growth. These findingscorrelate with the antiapoptotic function of EN1 in dopami-nergic neurons, where EN1 modulates mitochondrial signalsand upregulates cell survival pathways (34). Functional enrich-ment analysis of genes differentially expressed following EN1downregulation for process networks and pathways identifiedWNT and Hh signaling as top enriched networks. In the brain,EN1 is a downstream target of WNT, and EN1 and WNTpathway genes cooperate in mid-hindbrain development (15).EN1 was also shown to negatively regulate b-catenin transcrip-tional activity in a cell culture model (23). The identification ofWNT as top enriched process network positively regulated byEN1 in TNBC suggests similar interactions between EN1 andWNT signaling in TNBCs. Indeed, our b-catenin ChIP-seqanalysis showed that a subset of genes differentially expressedafter EN1 downregulation are direct b-catenin targets.
We defined the genomic targets of EN1 by ChIP-seq for endog-enous EN1 in SUM159 and SUM149 TNBC and for exogenouslyexpressed EN1 in MCF7 luminal breast cancer cells and deter-mined that EN1 can both positively and negatively affect genetranscription depending on the cell line and its association withother transcriptional regulators. Direct EN1 targets that showexpression changes after EN1 knockdown also showed highestenrichment for WNT and Hh signaling, and neural-related func-tions. Among others, neuroligins and their receptors are directtargets and regulated by EN1 in breast cancer cells. On the basis ofqPLEX-RIME we identified TLE3, TRIM24, TRIM28, and TRIM33as EN1-interacting proteins in breast cancer cells. ChIP-seq forTLE3 revealed that the TLE3–EN1 complex may have a negativeeffect on the expression of genes involved in angiogenesis andECM remodeling, because TLE3 peaks present in shEN1-expres-sing cells were preferentially associated with genes showing upre-gulation after EN1 knockdown.
Besides brain and neural development, EN1 is essentialfor the specification of dorsal dermal fate via transmittingWNT signaling (7) and it is implicated as a regulator ofdermal-derived fibroblasts involved in wound healing andcancer-associated stroma formation (35). During early embry-
onic development, EN1 is expressed in the dermal precursorsand later on its expression is detected in dermal progenitors.Because the mammary gland is derived from the dermis, it istempting to speculate that there could be rare EN1-expressingmammary progenitors and that a subset of TNBCs may orig-inate from these cells.
We showed that EN1 was specifically and highly expressed intriple-negative and basal breast cancers. However, we also foundsignificant heterogeneity in EN1 expression and patients with thehighest expression of EN1 had worse overall survival in bothMETABRIC and TCGA datasets. Furthermore, patients with TNBCwith high expression of EN1 were more likely to develop brainmetastases potentially due to the expression and dependency ofthese tumors for neural survival factors such as NLGN4X andother neuroligins. Besides neural cells, neurexins and neuroliginsare also expressed in the vasculature and they promote thematuration and formation of blood vessels (36). Thus, the upre-gulation of neuroligins by EN1 in breast tumors could promotethe growth of the primary tumor, enhance their ability to dis-seminate, and also promote the growth of metastatic lesions,especially in the brain.
In summary, we identified EN1 as a TNBC-specific TF, theexpression of which is associated with neural features, angiogen-esis, and brain metastasis of breast cancer leading to poor clinicaloutcome (Fig. 4G). Thus, targeting EN1 and neuroligins might beexplored as a potential therapeutic strategy for the prevention andtreatment of brain metastases of TNBC.
Disclosure of Potential Conflicts of InterestA.H. Beck has ownership interest (including stock, patents, etc.) in PathAI.
C. Caldas reports receiving a commercial research grant from AstraZeneca,Genentech, Roche, and Servier and is a consultant/advisory board member forAstraZeneca and Illumina. J.S. Carroll is director of biology at AstraZeneca and isa consultant/advisory board member for Azeria Therapeutics. K. Polyak is aconsultant/advisory boardmember forMitra Biotech and Acrivon Therapeutics.No potential conflicts of interest were disclosed.
Authors' ContributionsConception and design: G. Peluffo, A. Subedee, K. PolyakDevelopment of methodology: G. Peluffo, A. Subedee, Y. SuAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): G. Peluffo, A. Subedee, N. Kingston, B. Jovanovi�c,L.E. Stevens, K. Murphy, Y. Su, C. Caldas, J.S. CarrollAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): G. Peluffo, A. Subedee, N.W. Harper, B. Jovanovi�c,F. Flores, L.E. Stevens, F. Beca, A. Trinh, C.S.R. Chilamakuri, O.M. Rueda,A.H. Beck, C. Caldas, J.S. CarrollWriting, review, and/or revision of the manuscript: G. Peluffo, A. Subedee,B. Jovanovi�c, F. Beca, A. Trinh, C.S.R. Chilamakuri, C. Caldas, J.S. Carroll,K. PolyakAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): G. Peluffo, A. Subedee, C. CaldasStudy supervision: A. Subedee, C. Caldas, K. PolyakOther (mass spectrometry analysis): E.K. Papachristou, C.S. D'SantosOther (was involved in co-mentoring the co-first author of this study(A. Subedee), advising on experimental design and analysis, and providedassistance with some of the earlier experiments for this study.): A. Marusyk
AcknowledgmentsWe thank Ramesh Shivdasani and members of our laboratory for
their critical reading of the article and for useful discussions. This workwas supported by the NCI R35CA197623 (to K. Polyak), CDRMP W81XWH-09-1-0131 (to K. Polyak), W81XWH-14-1-0212 (to K. Polyak), W81XWH-09-1-0561 (to A. Marusyk), W81XWH-18-1-0027 (to B. Jovanovi�c), the
Peluffo et al.
Cancer Res; 79(16) August 15, 2019 Cancer Research4182
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
Susan G. Komen Foundation (to Y. Su and F. Beca), and CRUK core fundingand an ERC Consolidator award (to J.S. Carroll).
The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.
Received October 15, 2018; revised March 28, 2019; accepted June 14, 2019;published first June 25, 2019.
References1. Russnes HG, Lingjaerde OC, Borresen-Dale AL, Caldas C. Breast cancer
molecular stratification: from intrinsic subtypes to integrative clusters. Am JPathol 2017;187:2152–62.
2. Bianchini G, Balko JM, Mayer IA, Sanders ME, Gianni L. Triple-negativebreast cancer: challenges and opportunities of a heterogeneous disease.Nat Rev Clin Oncol 2016;13:674–90.
3. Garrido-Castro AC, Lin NU, Polyak K. Insights into molecular classifica-tions of triple-negative breast cancer: improving patient selection fortreatment. Cancer Discov 2019;9:176–98.
4. Lehmann BD, Jovanovic B, Chen X, Estrada MV, Johnson KN, Shyr Y, et al.Refinement of triple-negative breast cancer molecular subtypes: implica-tions for neoadjuvant chemotherapy selection. PLoS One 2016;11:e0157368.
5. Fuchs J, Stettler O, Alvarez-Fischer D, Prochiantz A, Moya KL, Joshi RL.Engrailed signaling in axon guidance and neuron survival. Eur J Neurosci2012;35:1837–45.
6. Yang J, Brown A, Ellisor D, Paul E, Hagan N, Zervas M. Dynamic temporalrequirement of Wnt1 in midbrain dopamine neuron development. Devel-opment 2013;140:1342–52.
7. Atit R, Sgaier SK, Mohamed OA, Taketo MM, Dufort D, Joyner AL, et al.Beta-catenin activation is necessary and sufficient to specify the dorsaldermal fate in the mouse. Dev Biol 2006;296:164–76.
8. Yeung KY, Fraley C, Murua A, Raftery AE, Ruzzo WL. Model-based clus-tering and data transformations for gene expression data. Bioinformatics2001;17:977–87.
9. Baudry JP, Raftery AE, Celeux G, Lo K, Gottardo R. Combiningmixture components for clustering. J Comput Graph Stat 2010;9:332–53.
10. Scrucca L, FopM,Murphy TB, Raftery AE.mclust 5: clustering, classificationand density estimation using Gaussian finite mixture models. R J 2016;8:289–317.
11. Su Y, Subedee A, Bloushtain-QimronN, Savova V, KrzystanekM, Li L, et al.Somatic cell fusions reveal extensive heterogeneity in basal-like breastcancer. Cell Rep 2015;11:1549–63.
12. Cancer Genome Atlas Network. Comprehensive molecular portraits ofhuman breast tumours. Nature 2012;490:61–70.
13. Ringner M, Fredlund E, Hakkinen J, Borg A, Staaf J. GOBO: geneexpression-based outcome for breast cancer online. PLoS One 2011;6:e17911.
14. Curtis C, Shah SP, Chin SF, Turashvili G, RuedaOM,DunningMJ, et al. Thegenomic and transcriptomic architecture of 2,000 breast tumours revealsnovel subgroups. Nature 2012;486:346–52.
15. Alves dos Santos MT, Smidt MP. En1 and Wnt signaling in midbraindopaminergic neuronal development. Neural Dev 2011;6:23.
16. Alvarez-Fischer D, Fuchs J, Castagner F, Stettler O, Massiani-BeaudoinO, Moya KL, et al. Engrailed protects mouse midbrain dopaminergicneurons against mitochondrial complex I insults. Nat Neurosci 2011;14:1260–6.
17. Ekins S, Nikolsky Y, Bugrim A, Kirillov E, Nikolskaya T. Pathway mappingtools for analysis of high content data. Methods Mol Biol 2007;356:319–50.
18. Hnisz D, Abraham BJ, Lee TI, Lau A, Saint-Andre V, Sigova AA, et al. Super-enhancers in the control of cell identity and disease. Cell 2013;155:934–47.
19. Saint-Andre V, Federation AJ, Lin CY, Abraham BJ, Reddy J, Lee TI, et al.Models of human core transcriptional regulatory circuitries. Genome Res2016;26:385–96.
20. Snel B, Lehmann G, Bork P, Huynen MA. STRING: a web-server to retrieveand display the repeatedly occurring neighbourhood of a gene.Nucleic Acids Res 2000;28:3442–4.
21. Stemmler MP, Eccles RL, Brabletz S, Brabletz T. Non-redundant functionsof EMT transcription factors. Nat Cell Biol 2019;21:102–12.
22. Veenvliet JV, Dos Santos MT, Kouwenhoven WM, von Oerthel L, Lim JL,van der Linden AJ, et al. Specification of dopaminergic subsets involvesinterplay of En1 and Pitx3. Development 2013;140:3373–84.
23. Bachar-Dahan L, Goltzmann J, Yaniv A, Gazit A. Engrailed-1 negativelyregulates beta-catenin transcriptional activity by destabilizing beta-cateninvia a glycogen synthase kinase-3beta-independent pathway. Mol Biol Cell2006;17:2572–80.
24. Papachristou EK, Kishore K, Holding AN, Harvey K, Roumeliotis TI,Chilamakuri CSR, et al. A quantitative mass spectrometry-based approachto monitor the dynamics of endogenous chromatin-associated proteincomplexes. Nat Commun 2018;9:2311.
25. Agarwal M, Kumar P, Mathew SJ. The Groucho/Transducin-like enhancerof split protein family in animal development. IUBMB Life 2015;67:472–81.
26. Fong KW, Zhao JC, Song B, Zheng B, Yu J. TRIM28 protects TRIM24 fromSPOP-mediated degradation and promotes prostate cancer progression.Nat Commun 2018;9:5007.
27. Herquel B, Ouararhni K, Khetchoumian K, Ignat M, Teletin M, Mark M,et al. Transcription cofactors TRIM24, TRIM28, and TRIM33 associate toform regulatory complexes that suppress murine hepatocellular carcino-ma. Proc Natl Acad Sci U S A 2011;108:8212–7.
28. Rueda OM, Sammut SJ, Seoane JA, Chin SF, Caswell-Jin JL, Callari M, et al.Dynamics of breast-cancer relapse reveal late-recurring ER-positive geno-mic subgroups. Nature 2019;567:399–404.
29. Venkatesh HS, Johung TB, Caretti V, Noll A, Tang Y, Nagaraja S, et al.Neuronal activity promotes glioma growth through neuroligin-3 secretion.Cell 2015;161:803–16.
30. Danielian PS, McMahon AP. Engrailed-1 as a target of theWnt-1 signallingpathway in vertebrate midbrain development. Nature 1996;383:332–4.
31. McGrath SE, Michael A, Morgan R, Pandha H. EN2: a novel prostate cancerbiomarker. Biomark Med 2013;7:893–901.
32. Martin TA, Goyal A, Watkins G, Jiang WG. Expression of the transcriptionfactors snail, slug, and twist and their clinical significance in human breastcancer. Ann Surg Oncol 2005;12:488–96.
33. Beltran AS, Graves LM, Blancafort P. Novel role of Engrailed 1 as aprosurvival transcription factor in basal-like breast cancer and engineeringof interference peptides block its oncogenic function. Oncogene 2014;33:4767–77.
34. Alberi L, Sgado P, Simon HH. Engrailed genes are cell-autonomouslyrequired to prevent apoptosis in mesencephalic dopaminergic neurons.Development 2004;131:3229–36.
35. Rinkevich Y, Walmsley GG, Hu MS, Maan ZN, Newman AM, Drukker M,et al. Skin fibrosis. Identification and isolation of a dermal lineage withintrinsic fibrogenic potential. Science 2015;348:aaa2151.
36. Arese M, Serini G, Bussolino F. Nervous vascular parallels: axon guidanceand beyond. Int J Dev Biol 2011;55:439–45.
www.aacrjournals.org Cancer Res; 79(16) August 15, 2019 4183
EN1 and Breast Cancer Brain Metastasis
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264
2019;79:4173-4183. Published OnlineFirst June 25, 2019.Cancer Res Guillermo Peluffo, Ashim Subedee, Nicholas W. Harper, et al. Cancer Associated with Brain Metastasis
Is a Transcriptional Dependency in Triple-Negative BreastEN1
Updated version
10.1158/0008-5472.CAN-18-3264doi:
Access the most recent version of this article at:
Material
Supplementary
http://cancerres.aacrjournals.org/content/suppl/2019/06/25/0008-5472.CAN-18-3264.DC1
Access the most recent supplemental material at:
Cited articles
http://cancerres.aacrjournals.org/content/79/16/4173.full#ref-list-1
This article cites 36 articles, 8 of which you can access for free at:
Citing articles
http://cancerres.aacrjournals.org/content/79/16/4173.full#related-urls
This article has been cited by 1 HighWire-hosted articles. Access the articles at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://cancerres.aacrjournals.org/content/79/16/4173To request permission to re-use all or part of this article, use this link
on March 7, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from
Published OnlineFirst June 25, 2019; DOI: 10.1158/0008-5472.CAN-18-3264