twist1-induced mir-424 reversibly drives...

15
Tumor and Stem Cell Biology TWIST1-Induced miR-424 Reversibly Drives Mesenchymal Programming while Inhibiting Tumor Initiation David J. Drasin 1 , Anna L. Guarnieri 2 , Deepika Neelakantan 1 , Jihye Kim 3 , Joshua H. Cabrera 2 , Chu-An Wang 2 , Vadym Zaberezhnyy 4 , Pierluigi Gasparini 5 , Luciano Cascione 5 , Kay Huebner 5 , Aik-Choon Tan 3 , and Heide L. Ford 1,2,4 Abstract Epithelial-to-mesenchymal transition (EMT) is a dynamic process that relies on cellular plasticity. Recently, the process of an oncogenic EMT, followed by a reverse mesenchymal-to- epithelial transition (MET), has been implicated as critical in the metastatic colonization of carcinomas. Unlike governance of epithelial programming, regulation of mesenchymal pro- gramming is not well understood in EMT. Here, we describe and characterize the rst microRNA that enhances exclusively mesenchymal programming. We demonstrate that miR-424 is upregulated early during a TWIST1 or SNAI1-induced EMT, and that it causes cells to express mesenchymal genes without affecting epithelial genes, resulting in a mixed/intermediate EMT. Furthermore, miR-424 increases motility, decreases adhe- sion, and induces a growth arrest, changes associated with a complete EMT that can be reversed when miR-424 expression is lowered, concomitant with an MET-like process. Breast cancer patient miR-424 levels positively associate with TWIST1/2 and EMT-like gene signatures, and miR-424 is increased in primary tumors versus matched normal breast. However, miR-424 is downregulated in patient metastases versus matched primary tumors. Correspondingly, miR-424 decreases tumor initiation and is posttranscriptionally downregulated in macrometastases in mice, suggesting the need for biphasic expression of miR-424 to transit the EMTMET axis. Next-generation RNA sequencing revealed miR-424 regulates numerous EMT and cancer stem- ness-associated genes, including TGFBR3, whose downregula- tion promotes mesenchymal phenotypes, but not tumor-initi- ating phenotypes. Instead, we demonstrate that increased MAPKERK signaling is critical for miR-424mediated decreases in tumor-initiating phenotypes. These ndings sug- gest miR-424 plays distinct roles in tumor progression, poten- tially facilitating earlier, but repressing later, stages of metas- tasis by regulating an EMTMET axis. Cancer Res; 75(9); 190821. Ó2015 AACR. Introduction Over the last decade, there has been a growing emphasis on understanding the metastatic cascade of carcinomas by examining the epithelial-to-mesenchymal transition (EMT; ref. 1). An onco- genic EMT, similar to a developmental EMT, causes epithelial cells to lose tight cellcell contacts and gain a more motile and invasive phenotype, traits postulated to facilitate dissemination from a primary tumor (2). Carcinoma-associated experimental studies and patient tissue analyses have further demonstrated that acqui- sition of a more mesenchymal phenotype is linked with other prometastatic traits such as tumor-initiating cell (TIC) character- istics (e.g., self-renewal, multipotency) and resistance to conven- tional therapeutics (35). Many of these informative studies have demonstrated that EMT-inducing transcription factors such as SIX1, TWIST1/2, and SNAI1/2 can mediate these prometastatic and TIC phenotypes (1). MicroRNAs (miRNA) have relatively recently been identied as a class of factors that also regulate oncogenic EMTs. Such mechan- isms include, but are not limited to, ability of miRNAs to target the 3 0 untranslated region (UTR) of CDH1 (E-cadherin; ref. 1) or to target the 3 0 UTR of the CDH1 repressors ZEB1/2 (6). Further- more, miRNAs have been reported to alter signaling pathways implicated in EMT, such as the EGFR pathway (1). To date, no miRNAs have been identied that solely upregulate the mes- enchymal arm of an oncogenic EMT without repressing epithelial programming. Identication of potential mesenchymal-specic regulators has likely been hindered by the ability of EMT-inducing factors to inuence one another (7). The ubiquitous loss of functional E-cadherin during an EMT has further confounded analysis of mesenchymal programming in isolation from epithelial pro- gramming, as E-cadherin loss can be sufcient to drive a full EMT (8). Studies manipulating the mesenchymal half of the EMT program are increasingly relevant, as human carcinoma cells are documented to coexpress epithelial and mesenchymal markers (911), thus having not fully lost their epithelial identity. Such an intermediate EMT state, as opposed to a fully mesenchymal state, may more easily undergo a reverse mesenchymal-to-epithelial 1 Program in Molecular Biology, University of Colorado Anschutz Med- ical Campus, Aurora, Colorado. 2 Department of Pharmacology, Uni- versity of Colorado Anschutz Medical Campus, Aurora, Colorado. 3 Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado. 4 Department of Biochemistry and Molec- ular Genetics, University of Colorado Anschutz Medical Campus, Aurora, Colorado. 5 Department of Molecular Virology, Immunology and Molecular Genetics, Ohio State University, Columbus, Ohio. Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). Corresponding Author: Heide L. Ford, University of Colorado School of Med- icine, 12800 East 19th Avenue, Mail Stop 8303, Aurora, CO 80045. Phone: 303- 724-3509; Fax: 303-724-3663; E-mail: [email protected] doi: 10.1158/0008-5472.CAN-14-2394 Ó2015 American Association for Cancer Research. Cancer Research Cancer Res; 75(9) May 1, 2015 1908 on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

Upload: trinhque

Post on 19-Jun-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

Tumor and Stem Cell Biology

TWIST1-Induced miR-424 Reversibly DrivesMesenchymal Programming while InhibitingTumor InitiationDavid J. Drasin1, Anna L.Guarnieri2, DeepikaNeelakantan1, JihyeKim3, JoshuaH.Cabrera2,Chu-An Wang2, Vadym Zaberezhnyy4, Pierluigi Gasparini5, Luciano Cascione5,Kay Huebner5, Aik-Choon Tan3, and Heide L. Ford1,2,4

Abstract

Epithelial-to-mesenchymal transition (EMT) is a dynamicprocess that relies on cellular plasticity. Recently, the processof an oncogenic EMT, followed by a reverse mesenchymal-to-epithelial transition (MET), has been implicated as critical inthe metastatic colonization of carcinomas. Unlike governanceof epithelial programming, regulation of mesenchymal pro-gramming is not well understood in EMT. Here, we describeand characterize the first microRNA that enhances exclusivelymesenchymal programming. We demonstrate that miR-424 isupregulated early during a TWIST1 or SNAI1-induced EMT, andthat it causes cells to express mesenchymal genes withoutaffecting epithelial genes, resulting in a mixed/intermediateEMT. Furthermore, miR-424 increases motility, decreases adhe-sion, and induces a growth arrest, changes associated with acomplete EMT that can be reversed when miR-424 expression islowered, concomitant with an MET-like process. Breast cancerpatient miR-424 levels positively associate with TWIST1/2 and

EMT-like gene signatures, and miR-424 is increased in primarytumors versus matched normal breast. However, miR-424 isdownregulated in patient metastases versus matched primarytumors. Correspondingly, miR-424 decreases tumor initiationand is posttranscriptionally downregulated in macrometastasesin mice, suggesting the need for biphasic expression of miR-424to transit the EMT–MET axis. Next-generation RNA sequencingrevealed miR-424 regulates numerous EMT and cancer stem-ness-associated genes, including TGFBR3, whose downregula-tion promotes mesenchymal phenotypes, but not tumor-initi-ating phenotypes. Instead, we demonstrate that increasedMAPK–ERK signaling is critical for miR-424–mediateddecreases in tumor-initiating phenotypes. These findings sug-gest miR-424 plays distinct roles in tumor progression, poten-tially facilitating earlier, but repressing later, stages of metas-tasis by regulating an EMT–MET axis. Cancer Res; 75(9); 1908–21.�2015 AACR.

IntroductionOver the last decade, there has been a growing emphasis on

understanding themetastatic cascade of carcinomasby examiningthe epithelial-to-mesenchymal transition (EMT; ref. 1). An onco-genic EMT, similar to a developmental EMT, causes epithelial cellsto lose tight cell–cell contacts and gain amoremotile and invasivephenotype, traits postulated to facilitate dissemination from aprimary tumor (2). Carcinoma-associated experimental studiesand patient tissue analyses have further demonstrated that acqui-sition of a more mesenchymal phenotype is linked with otherprometastatic traits such as tumor-initiating cell (TIC) character-

istics (e.g., self-renewal, multipotency) and resistance to conven-tional therapeutics (3–5). Many of these informative studies havedemonstrated that EMT-inducing transcription factors such asSIX1, TWIST1/2, and SNAI1/2 can mediate these prometastaticand TIC phenotypes (1).

MicroRNAs (miRNA) have relatively recently been identified asa class of factors that also regulate oncogenic EMTs. Suchmechan-isms include, but are not limited to, ability of miRNAs to targetthe 30 untranslated region (UTR) of CDH1 (E-cadherin; ref. 1) orto target the 30 UTR of the CDH1 repressors ZEB1/2 (6). Further-more, miRNAs have been reported to alter signaling pathwaysimplicated in EMT, such as the EGFR pathway (1). To date,no miRNAs have been identified that solely upregulate the mes-enchymal arm of an oncogenic EMTwithout repressing epithelialprogramming.

Identification of potential mesenchymal-specific regulators haslikely been hindered by the ability of EMT-inducing factorsto influence one another (7). The ubiquitous loss of functionalE-cadherin during an EMT has further confounded analysisof mesenchymal programming in isolation from epithelial pro-gramming, as E-cadherin loss can be sufficient to drive a full EMT(8). Studies manipulating the mesenchymal half of the EMTprogram are increasingly relevant, as human carcinoma cells aredocumented to coexpress epithelial and mesenchymal markers(9–11), thus having not fully lost their epithelial identity. Such anintermediate EMT state, as opposed to a fully mesenchymal state,may more easily undergo a reverse mesenchymal-to-epithelial

1Program in Molecular Biology, University of Colorado Anschutz Med-ical Campus, Aurora, Colorado. 2Department of Pharmacology, Uni-versity of Colorado Anschutz Medical Campus, Aurora, Colorado.3Department of Medicine, University of Colorado Anschutz MedicalCampus, Aurora, Colorado. 4Department of Biochemistry and Molec-ular Genetics, University of Colorado Anschutz Medical Campus,Aurora, Colorado. 5Department of Molecular Virology, Immunologyand Molecular Genetics, Ohio State University, Columbus, Ohio.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: Heide L. Ford, University of Colorado School of Med-icine, 12800 East 19th Avenue, Mail Stop 8303, Aurora, CO 80045. Phone: 303-724-3509; Fax: 303-724-3663; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-14-2394

�2015 American Association for Cancer Research.

CancerResearch

Cancer Res; 75(9) May 1, 20151908

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

transition (MET), which has been postulated to facilitate meta-static colonization at secondary sites (2), and is compatible withthe epithelial nature of carcinomametastases and the similarity ofa metastatic lesion to its originating primary tumor.

Studies now experimentally support that anMET, following anEMT, facilitates metastatic colonization. Such studies focused onthe EMT inducers TWIST1 and PRRX1, which promoted invasivephenotypes in vitro and in vivo, but needed to be downregulated,concurrent with an MET, for metastatic outgrowth to occur (12,13). Contrary to previous studies that positively link EMT andTIC attributes (3, 4), an MET link to stemness appears just asrelevant to metastatic outgrowth, as metastases are epithelial innature. Oncogenic EMT inducers may thus be regulating a moreplastic state, rather than a full EMT, to enable reversion to anepithelial phenotype along with self-renewal and proliferativecapability at the secondary site. To better understandmediators ofthis plastic state, we set out to identify novel regulators of themesenchymal arm of EMT and to ask whether such regulatorsimpinge on TIC phenotypes, thereby influencing various stages ofthe metastatic cascade. In doing so, we identify miR-424, actingthrough TGFBR3, as the first miRNA regulator of mesenchymalprogramming that simultaneouslymaintains epithelialmorphol-ogy and molecular characteristics. Although miR-424 enhancesearly prometastatic functions, our data demonstrate that thismicroRNA inhibits TIC phenotypes, mediated by ERK signaling,and ultimately becomes downregulated in experimental andclinical macrometastases.

Materials and MethodsCell culture

HMLEderivatives and SUM149PT lineswere generous gifts fromDr.RobertWeinberg (Massachusetts Institute of Technology; 2009)and Dr. Stephen Ethier (Medical University of South Carolina;2012), respectively. MCF12A cells were obtained from The Uni-versity of Colorado Comprehensive Cancer Center (UCCC) TissueCulture Shared Resource and cultured according to the ATCCrecommendations. Cell lines were authenticated on April 3,2013, by short tandem repeat profiling or by methods previouslydescribed (14). TWIST1 and SNAI1 induction were performed aspreviously described (4). MEK inhibitors, 5 mmol/L PD98059(Sigma) or 5 mmol/L AZD6244 (Otava Chemicals), were used for24 hours. Doxycycline (Sigma) inductions were done at 0.5 mg/mLin complete SUM149PTmedia for the indicated amount of time foreach assay. Inductions were always performed on cells that hadnever previously been exposed to doxycycline.

BrdUrd/PI analysisAttached cells were serum-starved, then given completemedium

supplemented with 10 mmol/L bromodeoxyuridine (BrdUrd; Sig-ma) for 2 hours. Cells were fixed with ice-cold 70% ethanol,permeabilized with 0.5% Tween-20/PBS, and denatured with 2NHCl. Cells were incubated with anti-BrdUrd antibody (Roche;1:100), FITC-anti–mouse IgG(Sigma; 1:20), andpropidium iodide(PI)/RNaseA. Cells were analyzed as in flow-cytometry analyses.

microRNA microarray and next-generation RNA sequencingA 10 mg total RNA was isolated using the miRNeasy Mini Kit

(Qiagen) and was submitted in triplicate to LC Sciences formicroRNA microarray services and data analysis using miRBasev15.0. Results are deposited under GSE58560.

A1mg total RNA isolated using themiRNeasyMiniKit (Qiagen)was submitted in triplicate to the University of Colorado AMCGenomics and Microarray Core for construction and sequencingof the RNA-seq library on a HiSeq-2000 (Illumina). On average,we obtained approximately 54 million (range, 44–60 million)single-end 50 bp sequencing reads per sample. RNA-seq analysiswas performed as previously described (15, 16). Results aredeposited under GSE54505.

Cell migration and adhesion assaysMotility was determined by gap closure assay, where a culture

insert (Ibidi) created a500-mmgapbetween4and6�104 cells perchamber. Inserts were removed the next day, and distancemigrat-ed was measured using DP2-BSW software (v2.2; Olympus).Adhesion assays were performed as previously described (17).

Immunoblot analysisWhole-cell extracts were collected in RIPA buffer as previously

described (17) and run on 7%, 8%, or 10% SDS-polyacrylamidegels. Antibodies used for immunoblots: a-catenin, E-cadherin,fibronectin, N-cadherin, plakoglobin (BD Biosciences), actin,b-tubulin, vimentin (Sigma), total ERK1/2, p-ERK1/2 (Cell Sig-naling Technology), TGFBR3 (Santa Cruz Biotechnology),ST3GAL5 (Abgent), and TMPRSS4 (Proteintech).

Flow cytometry analysesALDEFLUOR assays (Stemcell Technologies) to detect ALDH1

activity were performed according to the manufacturer's recom-mendation. Diethylaminobenzaldehyde (DEAB) was used as anALDH1 inhibitor to set ALDH1 gates. Surface TGFBR3 levels weredetermined with a polyclonal goat primary antibody (R&D Sys-tems) and AlexaFluor-594 secondary antibody (Life Technolo-gies). All samples were analyzed on a Gallios Flow Cytometer(Beckman Coulter) at the UCCC Flow Cytometry SharedResource.

ImmunocytochemistryImmunocytochemistry was performed as previously describ-

ed (17). Antibodies used: N-cadherin (BD Biosciences; 1:50),E-cadherin (Cell Signaling Technology; 1:200), AlexaFluor-488(Life Technologies; 1:800), AlexaFluor-594 (Life Technologies;1:800). Coverslips were mounted with Vectashield (Vector Lab-oratories) and DAPI (1 mg/mL).

Mammosphere formationPrimary and secondary mammosphere formation were per-

formed as previously described (18). Mammosphere media:MEBMbasalmedia (500mL), 1X B27, hydrocortisone (1 mg/mL),insulin (5 mg/mL), b-mercaptoethanol (5 mg/mL), and EGF(10 ng/mL).

30 UTR reporterReporter constructs and 5 nmol/L miR-424 or negative

control mimics (Thermo Scientific) were cotransfected intocells with X-tremeGENE siRNA Transfection Reagent (Roche).Lysates were collected and analyzed as previously described(18). See Supplementary Materials and Methods for constructdesign.

An Intermediate EMT, by miR-424, Inhibits Tumor Initiation

www.aacrjournals.org Cancer Res; 75(9) May 1, 2015 1909

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

Figure 1.TWIST1 induces mesenchymal markers beforerepression of epithelial genes, concomitantwith miR-424 upregulation. A, phase contrastimages of HMLE EV and HMLE TWIST1-inducible cells after 4 and 12 days of TWIST1induction; bar, 100 mm. B and C, epithelial andmesenchymal markers, by qPCR, after 4 (B)and 12 (C) days of TWIST1 induction; the two-tailed unpaired t test. D, miRNA microarrayresults (P < 0.1) after 4 and 12 days ofTWIST1 induction compared with HMLEcontrol cells. E, mature and primary (F) miR-424 levels, by qPCR, after 4 and 12 days ofTWIST1 induction; two-way ANOVA, theBonferroni multiple comparison post-test;SEM shown. Representative images/graphsof n � 3; #, P < 0.1; � , P < 0.05; �� , P < 0.01;���, P < 0.001.

Drasin et al.

Cancer Res; 75(9) May 1, 2015 Cancer Research1910

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

Statistical analysesPrism software (v5.0; GraphPad) was used to conduct appro-

priate statistical procedures, as noted in the individual figurelegends. P value <0.05 or FDR <0.25 were considered significantunless noted otherwise. The frequency of stem cells in the tumor-initiation experiment was estimated with 95% confidence usingthe ELDA analysis tool (19).

Mouse modelsTumor-initiation experiments and experimental metastasis

assays are detailed in Supplementary Materials and Methods. Allanimal studies were performed according to protocols reviewedand approved by the Institutional Animal Care and Use Com-mittee at the University of Colorado AMC.

ResultsTWIST1 and SNAI1 induce mesenchymal markers beforerepressing epithelial markers during an EMT

To examine the temporal regulation of mesenchymal andepithelial genes during an EMT, we used inducible models ofEMTpreviously established in the nontumorigenic, humanmam-mary epithelial HMLE cell line (4). In these models, TWIST1 orSNAI1 are fused to the estrogen receptor ligand-binding domain,allowing for posttranslational activation by 4-hydroxytamoxifen(4-OHT). After 12 days of 4-OHT treatment, we observed theexpected morphologic and molecular changes (Fig. 1A and C;Supplementary Fig. S1A and S1C) characteristic of an EMT (4).Interestingly, 4 days of 4-OHT treatmentwas not sufficient to alterthe epithelial morphology (Fig. 1A; Supplementary Fig. S1A,middle) or to decrease levels of the epithelial markers E-cadherinor miR-200c, but was sufficient time for TWIST1 and SNAI1 toinduce expression of the mesenchymal markers fibronectin, N-cadherin and vimentin (Fig. 1B; Supplementary Fig. S1B). Thesefindings suggested the existence of a mesenchymal regulator(s)that operates upstream of E-cadherin repression during an EMT.

TWIST1 and SNAI1 upregulate miR-424 early during an EMTTo identify a potential mechanism by which mesenchymal

markers are upregulated early during an EMT, we screened RNAfrom different time points during a TWIST1-induced EMT usinganmiRNAmicroarray. Asmembers of themiR-8 family (e.g.,miR-200a/b/c) have been well documented to promote an epithelialphenotype (20), we reasoned that another miRNAmay serve as amesenchymal-promoting counterpart. 41 and 77 miRNAs weredifferentially expressed after 4 (early EMT) and 12 (late EMT) daysof TWIST1 induction, respectively (Fig. 1D; Supplementary TablesS1 and S2). Notablemicroarray changes includedmembers of themiR-8 family, miR-146a, miR-31 and the previously identifiedTWIST1-regulated miR-10b, each of which have been implicatedin the metastatic process (20–23). Because the miR-8 family wasnot downregulated until day 12 (Fig. 1D), yet mesenchymalmarkers were upregulated earlier in the EMT process, additionalmechanisms downstream of TWIST1 that mediate the mesenchy-mal phenotype likely exist.

We used quantitative PCR (qPCR) with independent TWIST1inductions to validate the changes of miR-200a/b/c, as well asmiR-429, miR-10a, and miR-152 (data not shown). Additionalvalidation of the screen led to the confirmation that miR-424 wasupregulated in HMLE cells early after TWIST1 or SNAI1 induction(Fig. 1E; Supplementary Fig. S1D), which we further pursued, as

miR-424 had not previously been directly implicated as anexclusive regulator of mesenchymal programming. PrimarymiR-424 (pri–miR-424), the first precursor during miR-424 bio-genesis, was also upregulated by TWIST1 and SNAI1 induction(Fig. 1F; Supplementary Fig. S1E). Together, these data demon-strate that TWIST1 and SNAI1 can regulate miR-424 expression.

miR-424 regulates the mesenchymal arm of an EMTWe next asked whether miR-424 could mediate a change in

expression of mesenchymal markers by creating three humanbreast cell lines that stably overexpressed miR-424. The rationalefor cell line choice was as follows: HMLE, cell line used for theoriginal miRNA screen; MCF12A, nontransformed human mam-mary cell line to verify HMLE findings; SUM149PT, human breastcancer cell line with low endogenous miR-424 useful for in vivostudy of exogenous miR-424. In SUM149PT cells, miR-424(Fig. 2A) induced a dramatic upregulation of N-cadherin withmore modest increases in vimentin and fibronectin (Fig. 2B). Incontrast, miR-424 had little to no effect on the levels of theepithelial markers E-cadherin, a-catenin, plakoglobin, andmiR-200c, nor did miR-424 alter cellular morphology (Fig. 2B;Supplementary Fig. S2A and S2B). Similar results were observedinmiR-424–expressingMCF12A andHMLE cells (SupplementaryFig. S2C–S2F, S2G–S2J, respectively). Notably, immunocyto-chemistry revealed miR-424 expression caused a mixed/interme-diate phenotype, where 52% of SUM149PT cells simultaneouslyexpressed N-cadherin and E-cadherin at cell membranes, ascompared with 10% of empty vector (EV) control cells (Fig. 2CandD); immunocytochemistry also demonstrated localization ofE-cadherin was unaffected by miR-424. Thus, miR-424 is suffi-cient to regulate the mesenchymal arm of an EMT inmultiple celllines, independent of epithelial arm repression, mimicking anearly TWIST1 or SNAI1-induced EMT.

Gains in mesenchymal EMT markers, particularly N-cadherin,have been associated with increased motility (24). We thus testedwhether miR-424 affects cellular migration, focusing on theSUM149PTs and MCF12As to characterize the tumorigenic cellline chosen for in vivo studies and an additional nontransformedcell line that had the more robust intermediate EMT profile(Supplementary Fig. S2), respectively. Using gap closure assays,we found that miR-424 increased migration of SUM149PT cells(Fig. 2E). In addition, miR-424 shifted cells into the G1 phase,akin to an EMT-associated growth arrest (Fig. 2F). Finally, weexamined cell-matrix adhesion, finding that miR-424 decreasedcell adhesion tofibronectin-coated plates (Fig. 2G). Similar effectsof miR-424 on motility, growth, and adhesion were seen inMCF12A cells (Supplementary Fig. S3A–S3C). Together, thesedata demonstrate thatmiR-424 induces EMT-associated functionswhile promoting an intermediate EMT.

Inducible miR-424 reversibly recapitulates EMT-associatedphenotypes

As recent evidence has demonstrated the importance of anMETfollowing an oncogenic EMT, we asked whether the intermediateEMT state was reversible by developing a doxycycline-induciblemiR-424 system (i-miR-424) in SUM149PT cells. We character-ized clones from the EV control (i-EV 1 and i-EV 2) and miR-424(i-miR-424 1 and i-miR-424 2) expressing cells. Of note, 30 UTRluciferase reporters with a perfect miR-424 binding site demon-strated that doxycycline led to repression of luciferase in thei-miR-424 clones, which was alleviated bymutating the reporter's

An Intermediate EMT, by miR-424, Inhibits Tumor Initiation

www.aacrjournals.org Cancer Res; 75(9) May 1, 2015 1911

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

Drasin et al.

Cancer Res; 75(9) May 1, 2015 Cancer Research1912

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

miR-binding site (Fig. 2H). Removal of doxycycline also allevi-ated luciferase repression, demonstrating that i-miR-424 activityis functionally reversible (Supplementary Fig. S4A).

Importantly, induction of miR-424 led to an increase in N-cadherin and fibronectin levels without affecting levels of E-cadherin (Fig. 2I), and did so in a reversible manner (Fig. 2J).This proof-of-principle experiment demonstrates that reductionof miR-424, after it has already driven an intermediate EMT, issufficient to reverse mesenchymal phenotypes similar to an MET.

We also validated that i-miR-424 recapitulated functionsobserved with constitutive miR-424 expression. Doxycyclinetreatment shifted i-miR-424 cells into the G1-phase in a reversiblemanner (Supplementary Fig. S4B, white, gray bars, respectively).In addition, although doxycycline treatment led to an increase inmigration in the control i-EV 2 clone, doxycycline increasedmigration to a significantly greater extent in the i-miR-424 1 clone(Supplementary Fig. S4C).

To address the temporal role of miR-424 in EMT–MET andmetastasis, we decided to characterize our induciblemodel in vivo.By using doxycycline tomanipulatemiR-424 levels before or afterinjections (i.e., high to lowmiR-424, low to highmiR-424, alwayslow miR-424, always high miR-424), we aimed to determinewhich metastatic phenotypes miR-424 regulates and whethermanipulation of miR-424 to allow completion of an EMT–METaxis was required. However, neither orthotopic nor tail vein

injections resulted in metastatic burden after 70 and 90 days,respectively, with either the i-EV or i-miR-424 clones (data notshown). We were also unable to detect circulating tumor cells(CTC) after orthotopic injection. As SUM149PT cells have beenrecently documented to consist of varying subpopulations (25),interclonal cooperativity may be necessary for SUM149PT cells tometastasize, or the clones selected may lack the innate ability tometastasize. Regardless, in vitro reversibility of the intermediateEMT and its associated functions, by manipulating miR-424levels, is consistent with an EMT, followed by anMET, potentiallyfacilitating the various steps of metastasis.

miR-424 is posttranscriptionally downregulated in overtmetastatic lesions

Although we could not evaluate inducible miR-424 effects invivo, the role of an intermediate EMT in vivo remained unknown.This question is important, as tumor cells have been found to existin an intermediate state both in primary tumors and in circulation(11). Thus, we characterized our constitutive miR-424 system inan experimental metastasis mouse model. EV control and miR-424–overexpressing SUM149PT cells were tagged with fireflyluciferase to allow detection of metastatic burden by in vivoimaging, and were verified to express exogenous primary andmature miR-424 before injection (Fig. 3A). Cells were subse-quently injected into the left ventricle of nude female mouse

Figure 3.miR-424 is posttranscriptionallydownregulated in an experimentalmetastasis mouse model. A,preintracardiac injection levels ofmature and primary miR-424, byqPCR, of luciferase-tagged SUM149PTcells overexpressing constitutive miR-424. B, Kaplan–Meier graph ofmetastasis-free survival in nude miceafter intracardiac injection with cells inA. C and D, levels of mature (C) andprimary (D) miR-424, by qPCR, inmetastases frommice injectedwith EVor miR-424 cells profiled in A; SDshown. E, levels of mature and primarymiR-424, by qPCR, in HMLE cellsconstitutively expressing TWIST1past 20 days; SEM shown. Thetwo-tailed unpaired t test; �� , P < 0.01;��� , P < 0.001; N.S., not significant.

Figure 2.miR-424 drives a reversible intermediate EMT without repressing epithelial characteristics. A, level of stable constitutive miR-424, by qPCR; SEM shown.B, immunoblots of mesenchymal (N-cadherin, fibronectin, and vimentin) and epithelial (E-cadherin, a-catenin, and plakoglobin) markers in stable miR-424 cells.Representative loading control (b-tubulin) shown. �, incomplete stripping. C, quantification of E-cadherin and N-cadherin immunocytochemical localization;n� 155 nuclei. D, representative immunocytochemistry of E-cadherin (red) and N-cadherin (green); DAPI, blue; arrows, coexpression of E-cadherin and N-cadherin.E–G, stable miR-424–expressing SUM149PT cells displaying distance migrated over 8 hours (E); cell-cycle analysis by BrdUrd/PI staining (F); and adhesion tofibronectin-coated wells (G). H–J, doxycycline and reversed SUM149PT-inducible clones were treated with 0.5 mg/mL doxycycline for 7 to 10 days, whereasreversed conditions were cultured for an additional 7 to 10 days in the absence of doxycycline: 30 UTR luciferase reporter with perfect or mutant miR-424–bindingsites (H); EMT markers by immunoblots of whole-cell lysates (I) and by qPCR (J). The two-tailed unpaired t test for all except two-way ANOVA (H), theBonferroni multiple comparison post-test; SEM shown, compiled experiments of n � 3; NS, not significant; � , P < 0.05; �� , P < 0.01; ��� , P < 0.001.

An Intermediate EMT, by miR-424, Inhibits Tumor Initiation

www.aacrjournals.org Cancer Res; 75(9) May 1, 2015 1913

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

hearts. Following injection, each mouse was imaged to ensuresystemic luciferase signal indicative of dissemination. Over 15weeks, a trend toward decreased metastasis incidence wasobserved in the miR-424 group, though this difference was notsignificant (Fig. 3B).

To verify that ectopic expression ofmiR-424wasmaintained inmetastases that occurred in the miR-424 group, metastatic tissueswere collected using ex vivo luciferase imaging to confirm thepresence of a metastatic lesion and to discard as much normaltissue as possible, before RNA extraction. Remarkably, levels ofmature miR-424 were relatively similar in metastases from miceinjectedwith either EVormiR-424 SUM149PT cells (Fig. 3C). Thissimilarity contrasted with mature miR-424 expression levels incells preinjection, where there was an approximately 15-folddifference in mature miR-424 levels (Fig. 3A, left). However, wefound that ectopic pri–miR-424 was still expressed at relativelyhigh levels within the miR-424 metastases (Fig. 3D). This expres-sion of pri–miR-424 was similar to preinjection cells (Fig. 3A,

right), demonstrating that miR-424 was posttranscriptionallyregulated in vivo in the miR-424 metastases. Global miRNAprocessing did not appear affected, as multiple other maturemiRNAs were not downregulated in the same metastatic tissue(Supplementary Fig. S5).

A recent study showed that, although ectopic TWIST1 increasedtumor cell dissemination, subsequent downregulation of TWIST1enhanced metastatic colonization (13). These data suggest arequirement for biphasic TWIST1 expression at different stagesof the metastatic cascade, or for biphasic expression of TWIST1-induced genes that regulate various aspects of the metastaticprocess. Thus, we examined how miR-424 levels are regulatedupon long-term TWIST1 and SNAI1 expression, especially in lightof miR-424 posttranscriptional downregulation in metastases.We found that constitutive TWIST1 or SNAI1 expression inHMLEcells resulted in downregulation of mature miR-424 (Fig. 3E andSupplementary Fig. S6A), as opposed to the upregulationobserved early after TWIST1 or SNAI1 induction (Fig. 1E). This

Figure 4.miR-424 represses tumor-initiatingproperties. A, representative ALDH1activity, as measured by ALDEFLUORassay, with and without DEAB (ALDH1inhibitor) treatment in constitutivemiR-424 versus EV control SUM149PTcells. B, quantification of ALDH1activity without DEAB in thesame cell lines as A. C, secondarymammosphere formation ofSUM149PT cells with constitutivemiR-424 or EV; SEM shown, n � 3;� , P < 0.05; ��� , P < 0.001, thetwo-tailed unpaired t test. D, tumorinitiation results from palpatingNOD/SCID mice injected withSUM149PT-424 and EV cells in limitingdilution.

Drasin et al.

Cancer Res; 75(9) May 1, 2015 Cancer Research1914

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

downregulation is controlled posttranscriptionally, as pri–miR-424 is not repressed by either constitutive TWIST1 or SNAI1 (Fig.3E and Supplementary Fig. S6A). Together, these data are con-sistent with tumor cells downregulating mature miR-424 toreverse the intermediate EMT to allow outgrowth at a secondarysite.

miR-424 decreases tumor-initiating characteristicsIncreasing evidence suggests that the EMT state does not always

increase TIC characteristics (12, 26, 27); therefore, we askedwhether miR-424 influences TIC characteristics, as this may helpexplain why miR-424 trends toward decreasing metastasis inci-dence. We first determined ALDH1 activity in response to miR-424, as ALDH1-active populations are enriched for TICs (28).Overexpression of miR-424 in SUM149PT and MCF12A cellsdecreased ALDH1 activity (Fig. 4A and B; Supplementary Fig.S3D), whereas alternative TIC-enriching markers [i.e., CD24,CD44 and EPCAM (ESA)] were not consistently changed. Inaddition, inducible miR-424 also decreased ALDH1 activity,

which was reversible by doxycycline removal (SupplementaryFig. S4D, white, gray bars, respectively). The lack of CD24/CD44changes was unexpected given the ALDH1 changes and the factthat both sets of markers enrich for a self-renewal phenotype.However, CD24�/CD24þ populations are known to have rela-tively little overlapwithALDH1þpopulations (28, 29), thus,miR-424 seems to specifically affect ALDH1þ cells.

We then examined mammosphere formation as a functionalTIC assay (30). Consistent with the effects on ALDH1, miR-424decreased numbers of secondary mammospheres formed (Fig.4C). Thus, miR-424 is sufficient to impart EMT-related func-tions on cells, while simultaneously repressing TIC character-istics, traits previously thought to positively associate with oneanother (1).

miR-424 decreases orthotopic tumor initiationOur in vitro TIC data prompted us to address whether miR-

424 can affect tumor initiation by performing a limiting-dilu-tion analysis in vivo. Between 20 and 300 SUM149PT cells

Figure 5.TGFBR3 KD is necessary and sufficientto recapitulate much of anintermediate EMT. A and B, SUM149PTcells expressing constitutive miR-424assayed for TGFBR3, ST3GAL5, andTMPRSS4 by immunoblots (A) andTGFBR3 mRNA by qPCR (B). C,luciferase reporter using 30 UTR ofTGFBR3 cotransfected with 5 nmol/LmiR-424 mimics. Mutant, mutatedmiR-424 seed binding site in reporterconstruct. Two-way ANOVA with theBonferroni post-test. D and E, qPCRwith the two-tailed unpaired t test (D)and immunoblots (E) for EMT markersin SUM149PT cells with stable TGFBR3KD. F, immunoblots for EMT markersand phospho-ERK1/2 in SUM149PTcells overexpressing miR-424and a TGFBR3 rescue construct.Representative actin show for F. SEMshown, n � 3; � , P < 0.05; �� , P < 0.01;��� , P < 0.001.

An Intermediate EMT, by miR-424, Inhibits Tumor Initiation

www.aacrjournals.org Cancer Res; 75(9) May 1, 2015 1915

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

overexpressing miR-424 or control EV were injected into themammary fat pads of female NOD/SCIDmice, which were thenpalpated weekly for tumor formation. Although no differencesin tumor-initiating capability were observed when larger num-bers of cells were injected, injection of lower numbers of cells(20 cells) led to a significant reduction in TIC frequency due tomiR-424, as assessed (19) up to 3 months (Fig. 4D), supportingmiR-424 as an inhibitor of tumor initiation.

Interestingly, an oncogenic EMT has been shown to increasetumor initiation in RAS-transformed HMLE (HMLER) cellsconstitutively expressing TWIST1 or SNAI1 (4). Because weinitially observed TWIST1 and SNAI1 increasing miR-424levels, which is sufficient to decrease TIC properties, weattempted to reconcile the conflicting TIC data by examiningmiR-424 levels in HMLER cells, finding HMLER cells posttran-scriptionally downregulated miR-424 (Supplementary Fig.S6B). This additional example of miR-424 posttranscriptionaldownregulation is consistent with decreased miR-424 as amechanism to allow tumor initiation or metastatic outgrowth,possibly by removing signals that inhibit tumor initiation andcell proliferation.

miR-424 represses a breast cancer stem cell gene signatureTo investigate how miR-424 controls TIC and EMT-like

phenotypes, we used Gene Set Enrichment Analysis (GSEA)after performing next-generation RNA sequencing (RNA-seq)on MCF12A cells overexpressing miR-424. The nontumorigenicMCF12A cells were used to reduce potentially masking effectsof broader genomic alterations in the transformed SUM149PTcells. GSEA revealed that miR-424 represses several genesthat are normally downregulated in an EMT signature (7),revealing hypothetical targets of miR-424 involved in mesen-chymal programming (Supplementary Fig. S7A). miR-424 alsorepressed genes typically upregulated in a breast cancer stemcell gene signature (3), giving insight into how miR-424represses TIC phenotypes (Supplementary Fig. S7B). Target-Scan (31) analysis of the 30 UTRs of these downregulated genesrevealed that a significant number (20 of 60, P < 10�5; 8 of 23,P ¼ 0.002, respectively) of the most negatively associated coregenes in the two signatures had predicted miR-424–bindingsites (Supplementary Fig. S7, asterisks). Further GSEA analysis,using genome-wide predicted miR-424 targets (TargetScan) asa gene signature, revealed that our RNA-seq results were

Figure 6.ERK signaling is necessary for miR-424–mediated repression of tumor-initiating properties. A, ALDH1 activityof SUM149PT cells with stable KD ofTGFBR3. B, ALDH1 activity inSUM149PT cells expressing miR-424with restoration of TGFBR3. C, totaland phospho-ERK1/2 levels ofSUM149PT cells expressing miR-424with vehicle or MEK inhibitor(5mmol/L) treatment after 24hours. D,ALDH1 activity of cells in C. E and F,secondary mammosphere formationresults (E) and phase contrast imagesof SUM149PT cells expressingmiR-424and treated with PD98059 (F). SEMshown, n � 3; bar, 100 mm. Two-wayANOVA with the Bonferroni post-testfor all except one-way ANOVA (A)with Bonferroni post-test; #, P < 0.1;� , P < 0.05; �� , P < 0.01; ��� , P < 0.001.

Drasin et al.

Cancer Res; 75(9) May 1, 2015 Cancer Research1916

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

enriched with repressed predicted target genes (P < 0.001),demonstrating utility in coupling functional data and in silicopredictions. These data suggest that miR-424 may drive mes-enchymal EMT programming and inhibit TIC phenotypes bytargeting multiple genes simultaneously.

TGFBR3 is a novel target of miR-424To narrow down a mechanism downstream of miR-424, we

validated several RNA-seq hits. Immunoblot analyses ofSUM149PT lysates confirmed that miR-424 downregulatesTGFBR3, likely observed as a number of bands due to knownextensive posttranslational modifications (32), ST3GAL5 andTMPRSS4 (Fig. 5A). TGFBR3 was identified directly from RNA-seq differences, whereas ST3GAL5 and TMPRSS4 came through

GSEA of the breast cancer stem cell and EMT signatures, respec-tively. Each gene also contains a predicted miR-424–binding site(TargetScan) and has not previously been identified as a target ofmiR-424.

Because TGFb signaling has been implicated in both EMTand TIC phenotypes (32–35), we focused on TGFBR3 down-stream of miR-424. QPCR confirmed that miR-424 decreasesTGFBR3 mRNA in SUM149PT, HMLE, and MCF12A cells(Fig. 5B; Supplementary Fig. S7C and S7D), whereas a 30 UTRreporter assay demonstrated that miR-424 can repress theTGFBR3 30 UTR (Fig. 5C). Mutating the putative miR-424seed-binding site in the TGFBR3 30 UTR reporter constructsignificantly alleviated miR-424 repression of the reporter,though to a small degree (Fig. 5C).

Figure 7.Human data support a positive role formiR-424 in EMT, but a negative role inmetastasis. A and B, contingency tableanalysis of above/below median miR-424 levels versus above/belowmedian levels of TWIST1 (A) andTWIST2 (B). C, miR-424 levels inPAM50 subtypes from the TCGAdataset; quartiles shown; n ¼ 291. TheANOVA test with multiplecomparisons corrected by Tukey HSD.D, GSEA using mRNA from breastcancer patients with high or low miR-424 expression in the Buffa dataset(40) compared with upregulatedgenes from three EMT signatures byTaube and colleagues (7), Onder andcolleagues (8), and Blick andcolleagues (44). E, expression of miR-424 and TGFBR3 in individual primarybreast tumors; Pearson correlationanalysis. F and G, expression ofmiR-424 in normal breast versusmatched primary tumor (n ¼ 46; F)and primary tumor versus matchedmetastasis (n ¼ 18; G) in a humantriple-negative breast cancer dataset(45), the two-tailed paired t test;#, P < 0.0001 versus LumA; $,P < 0.05 versus LumB; %, P < 0.01versus LumA; � , P < 0.05.

www.aacrjournals.org Cancer Res; 75(9) May 1, 2015 1917

An Intermediate EMT, by miR-424, Inhibits Tumor Initiation

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

TGFBR3 mediates the intermediate EMT phenotype driven bymiR-424

To determinewhether TGFBR3 inhibition,mimickingmiR-424expression, is sufficient to recapitulate an intermediate EMT, weused two shRNAs to knockdown (KD) TGFBR3 in parentalSUM149PT cells (Fig. 5D). TGFBR3 KD increased vimentin andfibronectin levels, by qPCR and immunoblot, andN-cadherin, byqPCR,without affecting E-cadherin levels (Fig. 5D andE).We alsodetermined whether TGFBR3 loss is necessary for an miR-424–induced intermediate EMT by restoring the coding sequence ofTGFBR3, which cannot be repressed by miR-424 as it lacks a 30

UTR, to miR-424–overexpressing cells (Supplementary Fig. S7E).TGFBR3 restoration reversed the mesenchymal marker levelsoriginally increased by miR-424 while not affecting E-cadherinlevels (Fig. 5F). Together, decreased TGFBR3 is necessary inmediating miR-424–dependent EMT changes and sufficient toregulate, at least in part, mesenchymal EMT programming.

ERK signaling mediates miR-424 repression of TIC phenotypesBecause TGFBR3 is responsible for the majority of mesenchy-

mal changes induced bymiR-424, we askedwhether TGFBR3 alsoplays a role in repressing TIC phenotypes. We found that TGFBR3KD, phenocopying miR-424 expression, did not lower ALDH1activity (Fig. 6A), nor was TGFBR3 restoration able to rescuedecreased ALDH1 activity downstream of miR-424 (Fig. 6B).

Subsequently, we shifted our focus to ERK signaling, as weobserved a marked increase in this noncanonical TGFb signalingarm downstream of miR-424 (Fig. 5F), and as ERK signaling isknown to negatively regulate self-renewal of embryonic stem cells(36), though in other contexts ERK has been found to positivelyregulate self-renewal (37). Interestingly, although TGFBR3 KDincreased phospho-ERK1/2 (Fig. 5E), revealing TGFBR3 caninhibit ERK, restoring TGFBR3 did not reverse miR-4240s abilityto increase phospho-ERK1/2 levels (Fig. 5F), indicating that miR-424 activates the ERK pathway through multiple mechanisms.

To examine ERK signaling functionally, we used two small-molecule MEK inhibitors, AZD6244 (38) and PD98059 (39), toprevent MEK from phosphorylating ERK1/2 (Fig. 6C). ERK sig-naling inhibition reversed miR-4240s ability to decrease ALDH1activity (Fig. 6D). Extending the PD98059 MEK inhibitor tofunctionalmammosphere formation assays,we observed apartialreversal of mammosphere repression due to miR-424 (Fig. 6E).Interestingly, although MEK inhibition downstream of miR-424increased the number of mammospheres, they were smaller insize (Fig. 6F), suggesting divergent effects of the ERK pathway onself-renewal and proliferation in mammosphere-initiating cellsand the progeny that comprise the bulk of the mammosphere,respectively. We thus decided against using MEK inhibitors toexamine tumor initiation, as effects onTICpropertieswould likelybe masked by the antiproliferative properties. In light of TGFBR3KD not altering ALDH1 activity, yet being sufficient to increaseERK signaling, these data together suggest that ERK signaling isnecessary, but not sufficient, for repressing in vitro tumor-initiat-ing phenotypes downstream of miR-424.

miR-424 associates with TWIST1 and is biphasically expressedin human breast cancers

To determine the clinical relevance of miR-424, we assessedthe expression of miR-424 and EMT-inducing transcription fac-tors in human breast cancer datasets containing patient-matched

miRNA and mRNA expression data. In two datasets (40, 41), apositive relationship existed between miR-424 and TWIST1mRNA (Fig. 7A), in addition to TWIST2 mRNA (Fig. 7B), whichhas similar EMT- and TIC-associated properties as TWIST1 (42).

Recently, CTCs from triple-negative breast cancers, whichlargely overlap with basal tumors, and from HER2-enrichedbreast cancers have been shown to consist largely of mesen-chymal-like and intermediate EMT-like cells (11). We thusexamined The Cancer Genome Atlas (TCGA) data (43) todetermine whether miR-424 levels were associated with aparticular breast cancer subtype. Basal tumors had significantlyhigher miR-424 levels than luminal A and B tumors, whereasHER2-enriched tumors expressed significantly more miR-424than luminal A tumors (Fig. 7C). These subtype differences areconsistent with miR-424 inducing an intermediate EMTphenotype.

We then examined patient mRNA profiles from high and lowmiR-424 groups (40). GSEA identified a significant enrichment ofthree different EMT-associated gene signatures (7, 8, 44) inpatienttumors that expressed high miR-424 (Fig. 7D). These data cor-roborate our experimental findings that miR-424 induces anintermediate EMT and suggest that miR-424 may regulate anEMT-like program in human breast cancers.

Finally, to assess miR-424 levels during cancer progression, weinvestigated our own independent cohort of triple-negative breastcancer patients (45) where high expression of miR-424 wasexpected. We found that miR-424 was upregulated in primarytumors compared with matched normal tissue (Fig. 7F), and inthese primary tumors, miR-424 inversely correlated with TGFBR3(Fig. 7E). Furthermore, miR-424 was downregulated in metasta-ses compared with matched primary tumors (Fig. 7G), similar towhat occurred in our experimentalmetastases. Thus, examinationof multiple patient cohorts revealed that miR-424 was upregu-lated in primary tumors in association with EMT drivers and EMTgene signatures, but was downregulated in matched metastaticlesions. These data are consistent with our hypothesis that miR-424 expressionmay promote early stages of metastasis associatedwith an EMT, but that downregulated miR-424 facilitates meta-static outgrowth.

DiscussionThe oncogenic EMT has been implicated as a critical mediator

ofmetastasis. Previous controversy in the fieldmay have been dueto the challenge of identifying human tumor cells that are histo-logically mesenchymal and definitively originate from a carcino-ma (46). However, recent studies using experimental models andhuman tumors indicate that disseminating cells may exist in amore plastic state where cells transition between epithelial andmesenchymal phenotypes, or exist in an intermediate statebetween the two (11, 13). We thus examined an inducible EMTmodel to characterize adynamic, intermediate EMT state. Indoingso, we demonstrated that EMT-inducing factors TWIST1 andSNAI1 can increase expression of mesenchymal genes beforerepression of epithelial genes. Furthermore, we found thatmiR-424, downstream of TWIST1 and SNAI1, is sufficient toinduce multiple mesenchymal EMT characteristics, without alter-ing epithelial attributes, in a reversible manner. To our knowl-edge, this is the first demonstration of anmiRNA that can controlspecifically the mesenchymal arm of an EMT independent ofepithelial changes.

Cancer Res; 75(9) May 1, 2015 Cancer Research1918

Drasin et al.

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

Although miR-424 induced functions associated with a fulloncogenic EMT,miR-424 decreased TIC characteristics in vitro andin vivo, in contrast with other EMT-inducing factors (4, 42). Thenegative growth effects of miR-424, observed by others (47) andus, could be partially responsible for the TIC repression; however,over the last month of the experiment, no new miR-424 tumorswere formed, whereas new control tumors formed each week. Inaddition, although miR-424 overexpression led to smaller mam-mosphere formation, possibly due to growth effects, there wereimportantly fewer mammospheres, possibly indicative of TICinhibition. These data, coupled with ALDH1 repression, suggestthat miR-424 may decrease tumor initiation through TIC effectsand not just through growth inhibition.

Interestingly, although miR-424 did not significantly affectovert metastasis incidence, we observed posttranscriptional reg-ulation of miR-424 in metastatic outgrowths. HMLE and HMLERcells constitutively expressing TWIST1 and SNAIL1 over timedemonstrated posttranscriptional reduction of miR-424.Althoughposttranscriptional regulation ofmiR-424 has also beenobserved in endothelial cells (48), the mechanism remainsunknown. Posttranscriptional regulation of miR-424 in the con-stitutive TWIST1 and SNAI1 HMLER cells, which have increasedTIC frequency (4), may have allowed those cells to overcome theinhibitory tumor-initiation effects ofmiR-424, whilemaintainingother protumorigenic phenotypesmediated by TWIST1 or SNAI1.

Examination of how miR-424 mediates an intermediate EMTled to the identification of TGFBR3. Previous studies demonstrat-ed that TGFBR3 decreases during breast cancer progression (49),as well as during an EMT in prostate cancer before E-cadherindownregulation (33); similarly during anEMT,miR-424 increasesbefore E-cadherin downregulation. Importantly, TGFBR3 KD,while increasing vimentin and fibronectin protein levels, did notincrease N-cadherin protein. These data suggest that, althoughTGFBR3 functions downstream of miR-424 for induction of themesenchymal phenotype, it does not act alone, and may act incombination with other genes identified by GSEA.

To repress TIC attributes, we found that miR-424 requires ERKsignaling. A similar negative TIC role for ERK signaling has beenshown in normal and tumorigenic settings (36, 50), but ERKsignaling can also increase TIC properties (37), implicatingimportant contextual differences. In addition, because ERK sig-naling is necessary, but not sufficient, for repression of TICphenotypes, we speculate that numerous effectors of miR-424are responsible for mediating TIC phenotypes, particularly asmiR-424 represses part of a breast cancer stem cell signature.

Clinically, we found multiple associations with miR-424 thatparallel our experimental data. Analysis of primary breast tumorsfrom four independent cohorts demonstrated a positive linkbetween miR-424 and EMT-associated molecular changes. Ourown cohort revealed increased miR-424 in primary tumors com-pared with matched normal tissue, with an inverse correlationbetween TGFBR3 and miR-424 in the same tumors. Critically,miR-424 was decreased in metastases compared with matchedprimary tumors. These human data are consistent with ourexperimental findings that miR-424 induced a mesenchymal-likephenotype, but that mature miR-424 was downregulated inmetastases in vivo, supporting the possibility that decreasedmiR-424 at a distant site may facilitate metastatic outgrowth(Supplementary Fig. S8).

Recently, miR-424 was shown to induce a full oncogenic EMT(51), though we have not observed alterations in the epithelial

EMT arm in four different models. Comparing techniques mayhelp reconcile differing results, as ectopicmiR-424 expressionwasachieved through very different approaches. Nonetheless, ourfindings contribute to other studies dissociating mesenchymaland TIC phenotypes (12, 26, 27). These studies demonstrate thatan EMT may only be relevant to the metastatic cascade in thecontext of an EMT–MET spectrum. Critically, the intermediateEMT state has been observed in human tumor cells (9–11). Ourfindings that miR-424 drives an intermediate EMT, where themesenchymal N-cadherin protein is coexpressed with the epithe-lial E-cadherin protein, suggest that miR-424 may poise cells formovement along the epithelial–mesenchymal spectrum as ameans to mediate plasticity that facilitates anMET andmetastaticoutgrowth.

As more experimental studies demonstrate roles for both EMTand MET in metastatic progression, we must reevaluate whetherclinically targeting EMT-inducing factors will be uniformly effi-cacious. The position of a tumor cell along the EMT–MET axis, aswell as physical location in a patient, may dictate differentresponses to novel therapeutics. Thus, obtaining a better under-standing of the role of plasticity within themetastatic cascade willhelp identify which patients will benefit from potential EMT-related therapies.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: D.J. Drasin, H.L. FordDevelopment of methodology: D.J. Drasin, J.H. CabreraAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): D.J. Drasin, A.L. Guarnieri, D. Neelakantan,J.H. Cabrera, V. Zaberezhnyy, P. Gasparini, H.L. FordAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): D.J. Drasin, A.L. Guarnieri, J. Kim, P. Gasparini,L. Cascione, K. Huebner, A.-C. Tan, H.L. FordWriting, review, and/or revision of the manuscript: D.J. Drasin, J. Kim,K. Huebner, A.-C. Tan, H.L. FordAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): D.J. Drasin, J.H. Cabrera, C.-A. WangStudy supervision: H.L. Ford

AcknowledgmentsThe authors thank Paul Jedlicka for examining tumor sections, as well as the

UCCC Functional Genomics Facility, the Tissue Culture Shared Resource andLester Acosta, Christine Childs, and Karen Helm of the Flow Cytometry SharedResource for technical expertise and reagents (P30CA046934). RobertWeinberg(Massachusetts Institute of Technology), Stephen Ethier (Medical University ofSouth Carolina), and Gerard Blobe (Duke University) all generously providedcell lines or reagents. The authors thank members of the Ford lab for criticalreading of this article.

Grant SupportThis work was supported by grants from the National Cancer Institute (R01-

CA095277) and from the Metavivor Foundation (H.L. Ford). D.J. Drasin wassupported by an NCI predoctoral fellowship (F31-CA165617), and A.L. Guar-nieri was supported by a DOD Breast Cancer Research Program PredoctoralFellowship (W81XWH-10-1-0296). P. Gasparini, L. Cascione, and K. Huebnerwere supported by a grant from the National Cancer Institute (U01-154200).

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received August 19, 2014; revised November 25, 2014; accepted December19, 2014; published OnlineFirst February 25, 2015.

An Intermediate EMT, by miR-424, Inhibits Tumor Initiation

www.aacrjournals.org Cancer Res; 75(9) May 1, 2015 1919

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

References1. Drasin DJ, Robin TP, Ford HL. Breast cancer epithelial-to-mesenchymal

transition: examining the functional consequences of plasticity. BreastCancer Res 2011;13:226.

2. Micalizzi DS, Farabaugh SM, Ford HL. Epithelial–mesenchymal transitionin cancer: parallels between normal development and tumor progression. JMammary Gland Biol Neoplasia 2010;15:117–34.

3. Creighton CJ, Li X, Landis M, Dixon JM, Neumeister VM, Sjolund A, et al.Residual breast cancers after conventional therapy displaymesenchymal aswell as tumor-initiating features. Proc Natl Acad Sci U S A 2009;106:13820–5.

4. Mani SA, Guo W, Liao MJ, Eaton EN, Ayyanan A, Zhou AY, et al. Theepithelial-mesenchymal transition generates cells with properties of stemcells. Cell 2008;133:704–15.

5. Phillips TM, McBride WH, Pajonk F. The response of CD24(-/low)/CD44þ

breast cancer-initiating cells to radiation. J Natl Cancer Inst 2006;98:1777–85.

6. Gregory PA, Bert AG, Paterson EL, Barry SC, Tsykin A, Farshid G, et al. ThemiR-200 family and miR-205 regulate epithelial-to-mesenchymal transi-tion by targeting ZEB1 and SIP1. Nat Cell Biol 2008;10:593–601.

7. Taube JH, Herschkowitz JI, Komurov K, Zhou AY, Gupta S, Yang J, et al.Core epithelial-to-mesenchymal transition interactome gene-expressionsignature is associated with claudin-low and metaplastic breast cancersubtypes. Proc Natl Acad Sci U S A 2010;107:15449–54.

8. Onder TT, Gupta PB, Mani SA, Yang J, Lander ES, Weinberg RA. Loss of E-cadherin promotes metastasis via multiple downstream transcriptionalpathways. Cancer Res 2008;68:3645–54.

9. Livasy CA, KaracaG,Nanda R, TretiakovaMS,OlopadeOI,MooreDT, et al.Phenotypic evaluation of the basal-like subtype of invasive breast carci-noma. Mod Pathol 2006;19:264–71.

10. Thomas PA, Kirschmann DA, Cerhan JR, Folberg R, Seftor EA, Sellers TA,et al. Association between keratin and vimentin expression, malignantphenotype, and survival in postmenopausal breast cancer patients. ClinCancer Res 1999;5:2698–703.

11. Yu M, Bardia A, Wittner BS, Stott SL, Smas ME, Ting DT, et al. Circulatingbreast tumor cells exhibit dynamic changes in epithelial andmesenchymalcomposition. Science 2013;339:580–4.

12. OcanaOH, Corcoles R, Fabra A,Moreno-BuenoG, AcloqueH, Vega S, et al.Metastatic colonization requires the repression of the epithelial–mesen-chymal transition inducer Prrx1. Cancer Cell 2012;22:709–24.

13. Tsai JH, Donaher JL, Murphy DA, Chau S, Yang J. Spatiotemporal regu-lation of epithelial-mesenchymal transition is essential for squamous cellcarcinoma metastasis. Cancer Cell 2012;22:725–36.

14. ZhangH,Meng F, LiuG, Zhang B, Zhu J,Wu F, et al. Forkhead transcriptionfactor foxq1 promotes epithelial–mesenchymal transition and breast can-cer metastasis. Cancer Res 2011;71:1292–301.

15. Keysar SB, Astling DP, Anderson RT, Vogler BW, Bowles DW, Morton JJ,et al. A patient tumor transplant model of squamous cell cancer identifiesPI3K inhibitors as candidate therapeutics in defined molecular bins. MolOncol 2013;7:776–90.

16. Ware KE, Hinz TK, Kleczko E, Singleton KR, Marek LA, Helfrich BA, et al. Amechanism of resistance to gefitinib mediated by cellular reprogrammingand the acquisition of an FGF2-FGFR1 autocrine growth loop. Oncogen-esis 2013;2:e39.

17. Micalizzi DS, Christensen KL, Jedlicka P, Coletta RD, Baron AE, Harrell JC,et al. The Six1 homeoprotein induces humanmammary carcinoma cells toundergo epithelial–mesenchymal transition and metastasis in micethrough increasing TGF-beta signaling. J Clin Invest 2009;119:2678–90.

18. Smith AL, Iwanaga R, Drasin DJ, Micalizzi DS, Vartuli RL, Tan AC, et al. ThemiR-106b-25 cluster targets Smad7, activates TGF-beta signaling, andinduces EMT and tumor initiating cell characteristics downstream of Six1in human breast cancer. Oncogene 2012;31:5162–71.

19. Hu Y, Smyth GK. ELDA: extreme limiting dilution analysis for comparingdepleted and enriched populations in stem cell and other assays. J Immu-nol Methods 2009;347:70–8.

20. Howe EN, Cochrane DR, Richer JK. The miR-200 and miR-221/222microRNA families: opposing effects on epithelial identity. J MammaryGland Biol Neoplasia 2012;17:65–77.

21. Hurst DR, Edmonds MD, Scott GK, Benz CC, Vaidya KS, Welch DR. Breastcancer metastasis suppressor 1 up-regulates miR-146, which suppressesbreast cancer metastasis. Cancer Res 2009;69:1279–83.

22. Ma L, Teruya-Feldstein J, Weinberg RA. Tumour invasion and metastasisinitiated by microRNA-10b in breast cancer. Nature 2007;449:682–8.

23. Valastyan S, Reinhardt F, BenaichN, Calogrias D, Szasz AM,Wang ZC, et al.A pleiotropically acting microRNA, miR-31, inhibits breast cancer metas-tasis. Cell 2009;137:1032–46.

24. Hazan RB, Phillips GR, Qiao RF, Norton L, Aaronson SA. Exogenousexpression of N-cadherin in breast cancer cells induces cell migration,invasion, and metastasis. J Cell Biol 2000;148:779–90.

25. Prat A, Karginova O, Parker JS, Fan C, He X, Bixby L, et al. Characterizationof cell lines derived from breast cancers and normal mammary tissues forthe study of the intrinsic molecular subtypes. Breast Cancer Res Treat2013;142:237–55.

26. Celia-Terrassa T, Meca-Cortes O, Mateo F, de Paz AM, Rubio N, Arnal-Estape A, et al. Epithelial–mesenchymal transition can suppress majorattributes of human epithelial tumor-initiating cells. J Clin Invest2012;122:1849–68.

27. Stankic M, Pavlovic S, Chin Y, Brogi E, Padua D, Norton L, et al. TGF-beta-Id1 signaling opposes Twist1 and promotes metastatic colonization via amesenchymal-to-epithelial transition. Cell Rep 2013;5:1228–42.

28. Ginestier C, Hur MH, Charafe-Jauffret E, Monville F, Dutcher J, Brown M,et al. ALDH1 is a marker of normal andmalignant humanmammary stemcells and a predictor of poor clinical outcome. Cell Stem Cell 2007;1:555–67.

29. Liu S, Cong Y, Wang D, Sun Y, Deng L, Liu Y, et al. Breast cancer stem cellstransition between epithelial and mesenchymal states reflective of theirnormal counterparts. Stem Cell Rep 2014;2:78–91.

30. Dontu G, Abdallah WM, Foley JM, Jackson KW, Clarke MF, Kawamura MJ,et al. In vitropropagation and transcriptional profiling of humanmammarystem/progenitor cells. Genes Dev 2003;17:1253–70.

31. Lewis BP, Burge CB, Bartel DP. Conserved seed pairing, often flanked byadenosines, indicates that thousands of human genes are microRNAtargets. Cell 2005;120:15–20.

32. Kirkbride KC, Townsend TA, Bruinsma MW, Barnett JV, Blobe GC. Bonemorphogenetic proteins signal through the transforming growth factor-beta type III receptor. J Biol Chem 2008;283:7628–37.

33. Gordon KJ, Dong M, Chislock EM, Fields TA, Blobe GC. Loss of type IIItransforming growth factor beta receptor expression increases motility andinvasiveness associated with epithelial to mesenchymal transition duringpancreatic cancer progression. Carcinogenesis 2008;29:252–62.

34. Sharifi N, Hurt EM, Kawasaki BT, Farrar WL. TGFBR3 loss and conse-quences in prostate cancer. Prostate 2007;67:301–11.

35. Singh A, Settleman J. EMT, cancer stem cells and drug resistance: anemerging axis of evil in the war on cancer. Oncogene 2010;29:4741–51.

36. Burdon T, Stracey C, Chambers I, Nichols J, Smith A. Suppression of SHP-2and ERK signalling promotes self-renewal of mouse embryonic stem cells.Dev Biol 1999;210:30–43.

37. Iwanaga R, Wang CA, Micalizzi DS, Harrell JC, Jedlicka P, Sartorius CA,et al. Expression of Six1 in luminal breast cancers predicts poor prognosisand promotes increases in tumor initiating cells by activation of extracel-lular signal-regulated kinase and transforming growth factor-beta signalingpathways. Breast Cancer Res 2012;14:R100.

38. Pang L, Sawada T, Decker SJ, Saltiel AR. Inhibition of MAP kinase kinaseblocks the differentiation of PC-12 cells induced by nerve growth factor.J Biol Chem 1995;270:13585–8.

39. DudleyDT, Pang L,Decker SJ, Bridges AJ, Saltiel AR. A synthetic inhibitor ofthe mitogen-activated protein kinase cascade. Proc Natl Acad Sci U S A1995;92:7686–9.

40. Buffa FM, Camps C, Winchester L, Snell CE, Gee HE, Sheldon H, et al.microRNA-associated progression pathways and potential therapeutictargets identified by integrated mRNA and microRNA expression profilingin breast cancer. Cancer Res 2011;71:5635–45.

41. Enerly E, Steinfeld I, Kleivi K, Leivonen SK, Aure MR, Russnes HG, et al.miRNA-mRNA integrated analysis reveals roles for miRNAs in primarybreast tumors. PLoS ONE 2011;6:e16915.

42. Fang X, Cai Y, Liu J, Wang Z, Wu Q, Zhang Z, et al. Twist2 contributes tobreast cancer progression by promoting an epithelial–mesenchymal tran-sition and cancer stem-like cell self-renewal. Oncogene 2011;30:4707–20.

43. Cancer Genome Atlas Network. Comprehensive molecular portraits ofhuman breast tumours. Nature 2012;490:61–70.

Drasin et al.

Cancer Res; 75(9) May 1, 2015 Cancer Research1920

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

44. Blick T, Hugo H, Widodo E, Waltham M, Pinto C, Mani SA, et al.Epithelial—mesenchymal transition traits in human breast cancercell lines parallel the CD44(hi/)CD24 (lo/�) stem cell phenotypein human breast cancer. J Mammary Gland Biol Neoplasia 2010;15:235–52.

45. Cascione L, Gasparini P, Lovat F, Carasi S, Pulvirenti A, Ferro A, et al.Integrated microRNA and mRNA signatures associated with survival intriple negative breast cancer. PLoS ONE 2013;8:e55910.

46. Yang J, Weinberg RA. Epithelial-mesenchymal transition: at the cross-roads of development and tumor metastasis. Dev Cell 2008;14:818–29.

47. Forrest AR, Kanamori-Katayama M, Tomaru Y, Lassmann T, Ninomiya N,Takahashi Y, et al. Induction of microRNAs, mir-155, mir-222, mir-424and mir-503, promotes monocytic differentiation through combinatorialregulation. Leukemia 2010;24:460–6.

48. Chamorro-Jorganes A, Araldi E, Penalva LO, Sandhu D, Fernandez-Her-nando C, Suarez Y. MicroRNA-16 and microRNA-424 regulate cell-auton-omous angiogenic functions in endothelial cells via targeting vascularendothelial growth factor receptor-2 and fibroblast growth factor receptor-1. Arterioscler Thromb Vasc Biol 2011;31:2595–606.

49. DongM,HowT,KirkbrideKC,GordonKJ, Lee JD,HempelN, et al. The typeIII TGF-beta receptor suppresses breast cancer progression. J Clin Invest2007;117:206–17.

50. Guo Y, Zhang X, Meng J, Wang ZY. An anticancer agent icaritin inducessustained activation of the extracellular signal-regulated kinase (ERK)pathway and inhibits growth of breast cancer cells. Eur J Pharmacol2011;658:114–22.

51. Banyard J, Chung I, Wilson AM, Vetter G, Le Bechec A, Bielenberg DR, et al.Regulation of epithelial plasticity by miR-424 and miR-200 in a newprostate cancer metastasis model. Sci Rep 2013;3:3151.

www.aacrjournals.org Cancer Res; 75(9) May 1, 2015 1921

An Intermediate EMT, by miR-424, Inhibits Tumor Initiation

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394

2015;75:1908-1921. Published OnlineFirst February 25, 2015.Cancer Res   David J. Drasin, Anna L. Guarnieri, Deepika Neelakantan, et al.   Programming while Inhibiting Tumor InitiationTWIST1-Induced miR-424 Reversibly Drives Mesenchymal

  Updated version

  10.1158/0008-5472.CAN-14-2394doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://cancerres.aacrjournals.org/content/suppl/2015/02/26/0008-5472.CAN-14-2394.DC1

Access the most recent supplemental material at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/75/9/1908.full#ref-list-1

This article cites 51 articles, 14 of which you can access for free at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  [email protected]

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/75/9/1908To request permission to re-use all or part of this article, use this link

on July 10, 2018. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 25, 2015; DOI: 10.1158/0008-5472.CAN-14-2394