Endophilin A2 Promotes TNBC Cell Invasion and Tumor ......Signal Transduction Endophilin A2 Promotes TNBC Cell Invasion and Tumor Metastasis Tomas Baldassarre1,2, Kathleen Watt1,2,

Signal Transduction

Endophilin A2 Promotes TNBC Cell Invasion andTumor MetastasisTomas Baldassarre1,2, Kathleen Watt1,2, Peter Truesdell1,2, Jalna Meens1,2,Mark M. Schneider3, Sandip K. Sengupta3, and Andrew W. Craig1,2

Abstract

Triple-negative breast cancers (TNBCs) are highly aggressivecancers that lack targeted therapies. However, EGFR is frequentlyactivated in a subset of TNBCs and represents a viable clinicaltarget. Because the endocytic adaptor protein Endophilin A2(SH3GL1/Endo II) has been implicated in EGFR internalization,we investigated Endo II expression and function in humanTNBCs. Endo II expression was high in several TNBC cells com-pared with normal breast epithelial cells. Stable knockdown (KD)of Endo II was achieved in two TNBC cell lines, and althoughcell viability was unaffected, defects in receptor-mediated endo-cytosis were observed. EGFR signaling to Erk and Akt kinases wasimpaired in Endo II KD cells, and this correlated with reducedrates of EGFR internalization and cell motility. Endo II KD cellsalso displayed defects in three dimensional (3D) cell invasion,

and this correlatedwith impaired extracellularmatrix degradationand internalization of MT1-MMP. Endo II silencing also caused asignificant reduction in TNBC tumor growth and lung metastasisinmammary orthotopic tumor xenograft assays. In human breasttumor specimens, Endo II expressionwas highest in TNBC tumorscomparedwith other subtypes, and at the level of gene expression,high Endo II was associated with reduced relapse-free survival inpatients with basal-like breast cancers. Together, these resultsidentify a positive role for Endo II in TNBC tumor metastasisand a potential link with poor prognosis.

Implications: Endophilin A2 and related adaptor proteins rep-resent important signaling hubs to target inmetastatic cancers.MolCancer Res; 13(6); 1044–55. �2015 AACR.

IntroductionBreast cancer continues to be the most prevalent and second

most fatal cancer for women within the developed world (1).Triple-negative breast cancers (TNBC) lack expression of estrogenreceptor (ER), progesterone receptor (PgR), and HER2, and cur-rently lack targeted therapies. TNBCs are often aggressive withfrequent invasionof lymphnodes anddevelopment ofmetastases(2). These properties of TNBCs are often due to activation ofEGFR, which leads to downstream activation of signaling path-ways promoting cell growth, survival, motility, and invasionthrough basement membranes (3, 4). This invasive phenotypeis dependent on downstream activation of Erk and Akt kinases (5,6), which signal to key regulators of cell motility, proliferation,and survival (7). Sustained signaling to Erk and Akt pathways hasbeen linked to internalization of EGFR via clathrin-mediatedendocytosis (CME) and signaling from endosomes (8, 9). Thus,amore complete understanding of EGFR internalizationmechan-isms in EGFR-driven cancers may yield important new targetsrelevant to TNBC and other cancers. One such study used an

RNAi-based screen to identify Annexin A2 as a novel regulator ofEGFR trafficking, signaling, and metastasis in TNBC cells (10).

Another protein family implicated in EGFR internalization isthe Bin1/Amphiphysin/Rvs (BAR) domain-containing proteinfamily (11). BAR proteins both sense and promote membranecurvature via their BAR domains, and recruit key regulators ofendocytosis and actin polymerization to these membranes viatheir SH3 domains (12, 13). Indeed, several BAR proteins havebeen implicated in regulating cell motility, invasion, and metas-tasis in EGFR-driven cancermodels (14–19). Another BARproteinof potential relevance to metastatic TNBC is Endophilin A2(hereafter called Endo II). Endo II has been implicated in pro-moting internalization of EGFR (20) and membrane type-1matrixmetalloproteinase-14 (MT1-MMP; ref. 21). Tyrosine phos-phorylation of Endo II within its SH3 domain (Y315) ismediatedby Src kinase in complex with focal adhesion kinase (FAK). EndoII phosphorylation by Src prevents MT1-MMP internalizationin fibroblasts by disrupting binding to Dynamin (21). Recently,disruption of this FAK/Src/Endo II regulatory loop was shownto affect tumor growth andmetastasis inmousemammary tumorsderived from the Polyoma virus middle T antigen model (22).Silencing of Endo II caused a partial reversal of the epithelial–mesenchymal transition (EMT) phenotype and reduced numbersof tumor-initiating cells (TIC; ref. 22). Although these studiesin mouse models identify Endo II as a candidate regulator ofbreast cancer metastasis, further studies in human breastcancers are required. This is especially true since receptor cargosof Endo II, such as MT1-MMP and EGFR, are key drivers of TNBCmetastasis (23).

In this study,we report on the role of Endo II in regulating EGFRinternalization, signaling, and motility of human TNBC cellmodels. Endo II expression was high in TNBC compared withcell lines representative of luminal subtypes or normal mammary

1Department of Biomedical and Molecular Sciences, Queen's Univer-sity, Kingston, Ontario, Canada. 2Cancer Biology and Genetics Divi-sion, Queen's Cancer Research Institute, Kingston, Ontario, Canada.3Department of Pathology and Molecular Medicine, Queen's Univer-sity, Kingston, Ontario, Canada.

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

Corresponding Author: Andrew W. Craig, Queen's University, 18 Stuart St.,Kingston, ON K7L 3N6, Canada. Phone: 613-533-2496; Fax: 613-533-6830;E-mail: [email protected]

doi: 10.1158/1541-7786.MCR-14-0573

�2015 American Association for Cancer Research.

MolecularCancerResearch

Mol Cancer Res; 13(6) June 20151044

on January 4, 2021. © 2015 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst March 17, 2015; DOI: 10.1158/1541-7786.MCR-14-0573

http://mcr.aacrjournals.org/

epithelial cells. Lentiviral shRNA–mediated silencing of Endo II intwo TNBC cell lines (HCC 1806, MDA-MB-231), resulted inreduced internalization of EGFR, and diminished downstreamsignaling to Erk andAkt kinases. These signaling defects correlatedwith impaired cell motility, invasion, MT1-MMP internalization,and extracellular matrix (ECM) degradation in Endo II knock-down (KD) cells. In mammary orthotopic xenograft assays, weobserved reduced tumor growth and lung metastases for Endo IIKD TNBC cells compared with control cells. Lastly, profiling ofhuman TNBCs revealed high Endo II expression in this subtype,and a correlation with worse relapse-free survival in patients withbasal-like breast cancer, which is a major subset of TNBCs.

Materials and MethodsCell lines and antibodies

Normal-like MCF-10A and all breast cancer cell lines (MCF-7,T-47D, ZR-75-1, BT-474, MDA-MB-453, SK-BR-3, MDA-MB-231,HCC 1806, Hs578T, MDA-MB-468, Du4475, BT-20, HCC 38)were obtained and cultured according to the ATCC guidelines.Media and supplements were from Sigma-Adrich. The followingantibodies were used in this study: anti-Endophilin II [Santa CruzBiotech (SCBT); H-60; rabbit], anti-RasGAP (rabbit) (24), anti–b-Actin (SCBT; C4;mouse), anti-EGFR (SCBT; 1005; rabbit), anti-pEGFR [Cell Signaling Technologies (CST); Y1068; 1H12;mouse], anti-ERK1 (SCBT; K-23; rabbit), anti-pERK (SCBT; E-4;mouse), anti-Akt (CST; C67E7; rabbit); anti-pAkt (CST; 193H12;rabbit), anti-MMP14 (Abcam; ab78738; mouse), and anti-Ki67(Abcam; ab15580; rabbit).

Generation of stable Endo II KD cell linesLentiviral shRNA constructs in pGIPZ vector were obtained

(Open Biosystems) for a nontargeting (NT) control shRNA andtwo Endo II shRNAs. Lentiviral transduction of HCC 1806 andMDA-MB-231 cells with NT or Endo II shRNAs and selection withpuromycin was performed as previously described (15). Endo IIKD1 cells correspond with clone V3LHS_345456, containingtarget sequence 50AGAACTGCTTCTTCAGCCC30. Endo II KD2cells correspond with clone V3LHS_345454, containing targetsequence 50TGACCTCGATGTCCAGGGA30.

Cell lysis and immunoblottingBreast cancer cell lines were lysed with NP-40 lysis buffer (20

mmol/L Tris-HCl, pH 7.5; 150 mmol/L NaCl; 1 mmol/L EDTA;1% Nonidet P-40; 10 mg/mL aprotinin; 10 mg/mL leupeptin; 1mmol/L Na3VO4; 100 mmol/L phenylmethylsulfonyl fluoride).The following antibodies and concentrations were used forimmunoblot (IB) analysis: Endo II (1:200); RasGAP (1:2,000);b-actin (1:2,000); EGFR (1:200); pEGFR (1:1,000); Erk (1:200);pErk (1:200); Akt (1:1,000); pAkt (1:1,000); MMP14 (1:500).Horseradish peroxidase–conjugated sheep anti-mouse immuno-globulin or goat anti-rabbit immunoglobulin (1:10,000; GEHealthcare) and enhanced chemiluminescence (ECL WesternBlotting Substrate; Thermo Scientific) were used for detection.Densitometry data were calculated using ImageJ software (RSB).

EGF internalization and stimulation assaysFor internalization assays, NT and Endo II KD cells were seeded

onto coverslips inside six-well plates and allowed to adhereovernight. Wells were then incubated with serum-free mediacontaining 50 ng/mL Texas Red–conjugated EGF (Molecular

Probes) for 10 minutes or 25 mg/mL Texas Red–conjugatedTransferrin (Invitrogen) for 3 minutes. Cells were fixed andcounterstained with DAPI (Life Technologies), then washed withacid wash buffer (0.2 mol/L NaOAc; 0.2 mol/L NaCl; pH 5.3) toquench extracellular background fluorescence, followed by threewashes in PBS. The coverslips were then mounted on glass slidesusing Mowiol mounting medium (Sigma) and imaged on Quo-rum WaveFX-X1 spinning disc confocal microscope (QuorumTechnologies Inc.). The images were analyzed for the presence offluorescent vesicles using Image-Pro Plus 6 (Media Cybernetics).For stimulation assays, HCC 1806 NT and Endo II KD cells wereserum starved overnight and treated with RPMI 1640 mediasupplemented with EGF (50 ng/mL) for 0 to 16 minutes.

Cell growth and viability assaysCell viability assays were done as previously described (25).

Briefly, theHCC1806 parental cells, NT cells, and Endo II KD cellswere seeded onto 96-well plates (2,500 cells in 100 mL/well), andafter 18 hours AlamarBlue (Life Technologies) was added to thegrowth medium (10 mL/well). After 6 hours, the spectral absor-bance at 595 nmwasmeasured using aMultiskan Spectrum platereader (Thermo). Cell growth curves were generated by seedingHCC 1806 and MDA-MB-231 cells onto 12-well plates (5,000cells/well) in triplicate. Once a day for 4 days following seeding,cells were trypsinized, diluted to 1mL in PBS, and counted using aCoulter Counter (Beckman-Coulter).

Cell migration assaysHCC 1806 and MDA-MB-231 cells (50,000) were seeded in

transwell inserts (8 mm pore; BD Falcon), and allowed to migratetoward either media supplemented with either serum (MDA-MB-231 cells) or 50 ng/mL EGF in the lower chamber. Cells on thefilter were fixed, and nuclei were stained with DAPI. Cells wereremoved from the upper surface of the filters before mounting inMowiol (Sigma), and imaging by epifluorescence microscopy.Migrating cells were scored using Image-Pro Plus 6 software(Media Cybernetics).

Spheroid invasion assaysHCC 1806 cell invasion was investigated using the Cultrex 96

Well 3D Spheroid BME Cell Invasion Assay, as per the manufac-turer's instructions (cat #3500-096-K; Trevigen). Briefly, 3,000cells were resuspended in spheroid formationmatrix solution andpelleted in a 96-well round bottom plate and images acquired onday 3. Serum-supplemented invasion matrix was added to eachwell, and images were acquired after 7 days. The area of eachspheroid was measured on day 3 (preinvasion) and day 10(postinvasion) using Image-Pro Plus 6 software (Media Cyber-netics), and the difference was used to calculate total area of cellinvasion.

ECM degradation assaysCoverslips are coated with a mixture of 0.2% gelatin and

TRITC ester (8:1 ratio) for HCC 1806 cells as previouslydescribed (26). For MDA-MB-231, coverslips with TRITC-fibro-nectin were prepared, as previously described (15). For bothcell lines, cells were seeded onto coverslips with completemedia and incubated overnight, then fixed and counterstainedwith AlexaFluor350 Phalloidin (Life Technologies). Coverslipswere mounted onto slides with Mowiol and imaged at 100� on

Endophilin Promotes Cancer Metastasis

www.aacrjournals.org Mol Cancer Res; 13(6) June 2015 1045




a BX51 epifluorescence microscope (Olympus). Degradationarea and total cell areas were scored using Image-Pro Plus 6software (Media Cybernetics).

Surface biotinylation assaysSurface biotinylation assays were performed as previously

described (18). Briefly, MDA-MB-231 cells were seeded on six-well plates, and the next day, cells were chilled on ice beforeaddition of biotin. Samples were either kept on ice (total surface)or incubated at 37�C for 15 minutes, followed by once againchilling on ice. All cells except those in the surface level groupwereincubated with a biotin stripping buffer. All cells were then lysedand their lysates incubated with streptavidin-coated beads(Roche). Both lysates and streptavidin pull down fractions wereanalyzed by IB as indicated in figure legend.

Tumor xenograft assaysMammary orthotopic xenograft assays were performed using

MDA-MB-231 andHCC1806 cellmodels andRag2�/�:IL2Rgc�/�

(BALB/c)mice, as described previously (15). Briefly,mammary fatpads were injected with 1.5� 106 NT or Endo II KD cells in a 50%Matrigel (BD Biosciences) mixture. Mammary tumors wereallowed to grow for 5 weeks before animal sacrifice and tissueharvesting. Primary tumor homogenates were prepared by sus-

pension of primary tumor in NP-40 lysis buffer followed byhomogenization. Whole lungs were imaged by epifluorescencemicroscopy to detect GFP-positive metastatic nodules beforeformalin fixation and paraffin embedding. Sections (5 mm)were prepared and stained with hematoxylin/eosin (H&E). Fortail vein injections, 2 � 105 MDA-MB-231 NT or Endo II KDcells in 200 mL serum-free medium were injected into the tailvein of Rag2�/�:IL2Rgc�/� mice. At 2 weeks after injection,animals were sacrificed and tissues harvested and analyzed asabove. All animals were housed in a specific pathogen-freefacility (Queen's University Animal Care Services), and proce-dures were approved by the Queen's University Animal CareCommittee in accordance with the Canadian Council on Ani-mal Care guidelines.

Immunohistochemistry staining and scoringHuman breast cancer tissue microarrays (BR10010b, T087; US

Biomax) were stained using the Discovery XT Staining System(Ventana Medical Systems, Inc.). Antigens were retrieved with anEDTA pH 8.0 solution and incubated with rabbit anti-Endo II(1:100) or rabbit anti-Ki67 (1:1000) antibody. Endo II immu-nohistochemistry (IHC) staining was visualized using DAB and ahematoxylin counterstain. Microarrays were scanned using theAperio CS digital slide scanner (Queen's Laboratory forMolecularPathology) andanalyzedwith ImageScope software (Aperio). As a

Figure 1.Endo II expression and stable KD inTNBC cells. A, a panel consisting ofhuman normal-like breast epithelialMCF-10Acells and the indicatedbreastcancer cell lines spanning the majorsubtypeswere subjected to IB analysisfor Endo II expression. Densitometrywas performed to determine relativeEndo II expression between cell lines(RasGAP served as a loading control).B and C, stable silencing of Endo II inHCC 1806 (B) and MDA-MB-231 cells(C) was confirmed by IB for cellstransducedwith either anNTshRNAorone of two Endo II shRNA lentiviruses(KD1; KD2; b-actin served as a loadingcontrol). Unless otherwise specified,experiments were performed on KD1cells which showed the most efficientsilencing of Endo II. D, cell growthcurves for HCC 1806 and MDA-MB-231NT and KD1 cells. Cells (5,000) wereseeded in culture wells, and a triplicateset of wells was counted every dayafter seeding for 4 days. Results arerepresentative of 3 independentexperiments.

Baldassarre et al.

Mol Cancer Res; 13(6) June 2015 Molecular Cancer Research1046




control, Endo II antibody was precleared with beads coated withGST-Endo II linker-SH3 construct. The depleted antibody andmock-depleted antibodywere tested by IP/IB usingMDA-MB-231cell lysates. IHC staining of human tumors using precleared EndoII antibody was used to calculate the minimum threshold forsubsequent staining with Endo II antibody. Core sections selectedfor scoring were of similar areas, and included only epithelial-likecells. H-scores were calculated based on positive pixel intensityaccording to the formula: (%weak positive) þ (%positive � 2)þ(%strong positive � 3).

BioinformaticsFor analysis of Endo II transcripts in breast cancer cell lines, we

used openly available Affymetrix microarray datasets (www.biogps.org). For analysis of Endo II transcripts in breast cancermicroarray studies, a Kaplan–Meier curve for relapse-free survivalwas created using Kaplan–Meier Plotter (www.kmplot.com) for199basal-like breast cancer patients (ER�/PgR�/HER2�) groupedas above or below median expression levels of the sh3gl1 gene.Hazard ratio (with95%confidence interval) and log-rankP valueswere calculated.

Statistical analysisUnless otherwise specified, all experiments were performed in

triplicate and presented as themeans� standard error (SEM). TheStudent two-tailed t testwasused to compareNT andKDcell lines,with significant differences defined by P < 0.05, unless otherwisestated in figure legends.

ResultsEndo II is highly expressed in TNBC cell lines

Endo II has been implicated in a FAK/Src kinase signalingaxis in fibroblasts and mouse mammary tumor cell lines thatregulates MT1-MMP internalization (21, 22). To begin testingthe role of Endo II in human breast cancer, we profiled Endo IIexpression in a panel of human breast–derived cell lines. Thispanel included an immortalized normal-like breast epithelialcell line (MCF-10A), luminal A (MCF-7, T-47D), luminal B (ZR-75-1, BT-474), HER2 (MDA-MB-453, SK-BR-3), and TNBC celllines (MDA-MB-231, HCC 1806, Hs578T, MDA-MB-468,Du4475, BT-20, HCC 38). IB analyses revealed that Endo IIwas expressed at high levels in HER2 and TNBC cell lines,compared with MCF-10A or luminal cell lines (Fig. 1A;p120RasGAP served as a loading control). Similar results wereobserved at the level of Endo II gene expression, based onmicroarray results for 51 normal-like and breast cancer celllines (27, 28), with a trend toward higher transcript levels inTNBC and HER2 cell lines (Supplementary Fig. S1). To studythe function of Endo II in TNBC, we generated stable KDs ofEndo II in two TNBC cell lines using a lentiviral system. It isworth noting that HCC 1806 cells represent basal-like, andMDA-MB-231 cells are basal B/claudin low (29). For each cellline, two separate shRNAs were tested (KD1, KD2), and thesewere compared with a control line expressing an NT shRNA.Endo II expression was greatly reduced in KD derivatives fromboth HCC 1806 and MDA-MB-231 cells based on IB analysis(Fig. 1B and C). For both cell lines, shRNA1 delivered a morecomplete silencing of Endo II compared with shRNA2 (Fig. 1Band C; compare KD1 and KD2), and was prioritized for use inmost assays. Next, we determined whether Endo II KD had any

effect on cell viability and cell growth. No significant differ-ences were observed between NT and Endo KD TNBC cellgrowth curves (Fig. 1D), or metabolic activity measured usingalamarBlue (data not shown).

Endo II promotes EGFR internalization in TNBC cellsPrevious studies have implicated Endo II in promoting early

and late events in CME, such as vesicle scission and clathrin-coated vesicle uncoating (20, 30–32). To test the potential effectsof Endo II KDonCME inHCC1806 cell lines (NT, KD1, KD2), weincubated them with Texas Red–labeled transferrin (TR-Tfn).After 3minutes, extracellular TR-Tfn was removed, and cell nucleiwere stained with DAPI before imaging by confocal microscopy.Although NT and Endo II KD cell lines both showed internali-zation of TR-Tfn (Fig. 2A), we observed a significant reduction inTfn-positive vesicles in Endo KD1 and KD2 cell lines comparedwith NT cells (Fig. 2B). The magnitude of these defects in CMEwas proportional to levels of Endo II remaining in these cell lines(Fig. 1B and C), and therefore supports that Endo II silencingcauses these defects rather than potential off-target effects ofindividual shRNAs.

EGFR is frequently activated in TNBC, and internalization ofactivated EGFR regulates signaling and degradation (10, 33). Totest whether Endo II regulates EGFR internalization, we

Figure 2.Endo II promotes transferrin uptake via CME in TNBC cells. A, representativeimages of Texas Red–labeled transferrin (TR-Tfn) internalization in HCC 1806NT and Endo II KD cells at 3 minutes after treatment (cell nuclei werecounterstained with DAPI). B, numbers of transferrin-containing vesicles ineach cell line were quantified for Endo II NT, KD1, and KD2 cells (N¼ 25, 41, 30cells for NT, KD1, KD2, respectively). Results are representative of 3independent experiments (�� , P < 0.001).






analyzed the effects of Endo II KD on uptake of Texas Red–labeled EGF (TR-EGF). Although EGF uptake was observed inboth NT and Endo II KD TNBC cell lines, fewer EGF-positivevesicles were observed in Endo II KD cells (Supplementary Fig.S2A and S2C). Quantification of these results revealed a sig-nificant reduction in EGF uptake in Endo II KD cells for bothHCC 1806 and MDA-MB-231 cell lines, compared with NTcontrols (Supplementary Fig. S2B and S2D). Together, theseresults support a role for Endo II in promoting early events inreceptor endocytosis in TNBC cells.

Endo II regulates EGFR signaling in TNBCBecause internalization of EGFR has been linked to sustained

activation of Erk and Akt signaling pathways (10, 33), we testedthe potential effects of Endo II on EGFR signaling in TNBC cells. Atime course of EGF treatment was performed on serum-starvedHCC 1806 NT or KD1 cells, and EGFR signaling was analyzed byIB. To test the extent of EGFR activation, autophosphorylation ofEGFR at Y1068 (pEGFR) was analyzed and showed similarkinetics of activation in NT and KD cells (Fig. 3A). Densitometricanalyses revealed a small increase in relative pEGFR levels in

Figure 3.Endo II promotes EGFR signaling toErk and Akt kinases in TNBC cells.A, HCC 1806 NT and KD cells weretreated with EGF (50 ng/mL) for theindicated times (minutes). Lysateswere subjected to IB analysis for theindicated proteins and theirphosphorylated forms. B, graphdepicts relative levels of pEGFR, pErk,and pAkt compared with totallevels observed in 3 independentexperiments. (� , P < 0.05; �� , P < 0.01;�� , P < 0.001)

Baldassarre et al.





Endo II KD cells, compared with control, but not at all timepoints(Fig. 3B). Next, we investigated EGFR signaling to Erk and Aktkinases, which regulate myriad responses to EGF, including cellproliferation, survival, motility, and invasion (34, 35). Interest-ingly, Endo II KD cells displayed a dramatic reduction in EGF-induced phosphorylation of Erk (pErk) and Akt (pAkt) kinasescompared with NT control cells (Fig. 3A). Densitometric analysesfrom multiple experiments revealed significant reductions inpErk and pAkt levels at each time point of EGF treatment inEndo II KD cells (Fig. 3B).

Endo II promotes TNBC cell motility and invasion of ECMTo test whether defects in EGFR internalization and signaling

in Endo II KD cells alter the highly invasive phenotypes of theseTNBC cell lines, we tested them in chemotaxis and invasionassays. Using transwell migration chambers, and EGF as che-moattractant, we observed a significant reduction in chemo-taxis of HCC 1806 cells upon silencing of Endo II (Fig. 4A).Similar assays were performed in MDA-MB-231 NT, KD1, andKD2 cells using serum as the chemoattractant, which alsorevealed a role for Endo II in promoting TNBC cell motility(data not shown). To test the potential effects of Endo II KD onTNBC cell invasion, we compared HCC 1806 NT and KD cellsfor their ability to form spheroids in 3D culture and invade thesurrounding ECM. We observed similar growth of NT and EndoII KD cells as spheroids at day 3, but following overlay withECM permissive of cell invasion for 7 days, we observed adramatic reduction in 3D cell invasion (Fig. 4B and C). Togeth-er, these results demonstrate a key role for Endo II in promotingTNBC cell motility and invasion of ECM.

Endo II promotes invadopodia formation and ECMdegradation in TNBC cells

Previous studies in Src-transformed fibroblasts and mousemammary tumor cells have implicated Endo II in promotinginternalization of MT1-MMP and ECM degradation by inva-dopodia (21, 22). Endo II inactivation following phosphory-lation by FAK/Src complexes was shown to promote theinvasive phenotypes in these models. To test effects of EndoII KD on invadopodia formation in human TNBC cells,we plated NT and KD cells on coverslips coated with a thinlayer of TRITC-labeled ECM proteins, such as gelatin (forHCC 1806 cells) or fibronectin (for MDA-MB-231 cells). After24 hours, cells were fixed and filamentous actin (F-actin) wasstained before imaging by epifluorescence microscopy. In theseassays, invadopodia are visualized as F-actin dots over areas ofECM degradation, and these F-actin dots were observed in theperinuclear region of NT and KD MDA-MB-231 cells (Fig. 5A).Although no significant differences were observed in F-actindots (Fig. 5B), we observed a significant defect in ECM deg-radation in Endo II KD cells (Fig. 5C). We observed verysimilar results in parallel studies with HCC 1806 NT and KDcells (Supplementary Fig. S3). Together, these results implicateEndo II in promoting invadopodia maturation and ECMdegradation in TNBC cells.

In fibroblasts, overexpression of a gain-of-function mutant inEndo II (Y315F), lacking the Src phosphorylation site, wasshown to increase MT1-MMP internalization and reduce ECMdegradation (21). To address whether Endo II KD alters MT1-MMP internalization in TNBC cells, surface biotin assays were

performed to compare levels of cell surface and internalizedpools of MT1-MMP. In the highly invasive MDA-MB-231 cells,Endo II KD cells exhibited increased levels of MT1-MMP on thesurface of Endo II KD cells in these assays (Fig. 5D; 4�Ccondition). Upon shifting the cells to 37�C to allow endocy-tosis to proceed, we observed defects in internalization of MT1-MMP in Endo II KD cells (Fig. 5D). As a positive control, weanalyzed EGFR internalization and observed similar defects inEndo II KD cells (Fig. 5D). These results are consistent withEndo II promoting MT1-MMP internalization in human TNBCcells, and the loss of endosomal pools of MT1-MMP and EGFRmay prevent mobilization to invadopodia precursors in thesecancer cells.

Figure 4.Endo II promotes TNBC cell motility and invasion. A, Transwell migrationassay using EGF-induced migration of HCC 1806 NT and KD cells. Results arerepresentative of 3 independent experiments (� , P < 0.05). B, NT andEndo II KD cells are imaged upon formation of spheroids (preinvasion) and aweek after addition of a Matrigel layer (postinvasion). C, quantification ofinvasion area was determined by subtracting preinvasion from postinvasionarea using imaging software. Results are representative of 3 independentexperiments (� , P < 0.05).






Endo II promotes tumor growth and metastases in miceTo better characterize the contributions of Endo II to disease

progression in TNBC, we performed mammary orthotopic xeno-graft assays using Endo II KD TNBC cells and their controlsinjected into Rag2�/�:IL-2Rgc�/�mice. At 5 weeks after injection,micewere sacrificed and several tissues alongwithprimary tumorswere harvested. For both MDA-MB-231 and HCC 1806 cellmodels, Endo II silencing caused a significant reduction in tumorgrowth (Fig. 6A; Supplementary Fig. S4A). This defect in tumorgrowth was not due to differences in cell growth rates in vitro (Fig.1D), or with expression of the Ki67marker of proliferating tumorcells in vivo (Fig. 6B; Supplementary Fig. S4B). To establishwhether the silencing of Endo II remained intact in vivo, weprepared homogenates from primary tumors for IB analysis. Thisshowed a substantial reduction in Endo II within the primarytumors for KD cells compared with NT controls (Fig. 6C; Sup-plementary Fig. S4C). Lungs from the mice were harvested and

imaged by fluorescence microscopy to detect GFP-positive TNBCmetastases. In bothmodels, Endo II KD resulted in less GFPþ lungmetastases (Fig. 6D; Supplementary Fig. S4D). To quantify theseresults, lung tissue sectionswere stainedwithH&E, andmetastaticnodules were quantified (Fig. 6D and E; Supplementary Fig. S4Dand S4E). We observed a significant reduction in lung metastasesin both TNBC cell models upon Endo II silencing (Fig. 6E;Supplementary Fig. S4E). These results were likely not explainedby differences in tumor size for Endo II KD group, because weobserved no strong correlation between number of metastasesand mass of the primary tumors in these TNBC models (Supple-mentary Fig. S5). To test whether Endo II functions in early or lateevents in metastasis, we compared the lung seeding efficiency ofMDA-MB-231NTandKDcells.Weobserved a significant defect inEndo II KD cells in lung seeding efficiency compared with NTcontrol cells (Supplementary Fig. S6). These results implicateEndo II in promoting TNBCmetastasis during early events within

Figure 5.Endo II promotes invadopodia maturation andMT1-MMP internalization in TNBC cells. A, MDA-MB-231 cells are plated on TRITC-fibronectincoated coverslips for 24 hours before fixationand F-actin staining. Representativeepifluorescence images are shown for F-actin(pseudocolored green), ECM (red), and anoverlay (merge). B and C, graphs depictquantification of F-actin dots (B) and ECMdegradation area (C) by MDA-MB-231 NT and KDcells (N ¼ 25 cells; n.s.: not significant; �� , P <0.01). D, surface biotin assayswere performed onMDA-MB-231 NT and Endo II KD cells asdescribed in Materials and Methods. The totalsurface pool (4�C) and internalized fractions[37�C 150 (min); 300 (min)] were captured withstreptavidin beads and subjected to IBwith EGFRand MT1-MMP antibodies (PD: pull-down; SCL:soluble cell lysate; Strip: extent of biotin strippingof a separate 4�C sample with MESNA; dashedlines indicate that a lower exposure of the sameblot was shown for 4�C samples compared withother conditions). Results are representative of 3independent experiments.

Baldassarre et al.





the mammary gland (orthotopic model) and following dissem-ination in the peripheral blood (tail vein model).

High Endo II expression in TNBC correlates with poor clinicaloutcome

Since our functional studies of Endo II in TNBC cell and tumormodels support a key role for Endo II in TNBC, we extended ourstudy of Endo II to examine its expression in human breast tumorsamples. Endo II protein levels were analyzed by IHC staining oftissue microarrays containing 50 invasive ductal carcinomas. Totest the specificity of the Endo II IHC signal, we depleted theantisera with aGST–Endo II fusion protein, or withGST alone as acontrol. The GST-Endo II depletion resulted in a nearly completeloss of recovery of Endo II by immunoprecipitation from MDA-MB-231 cells (Supplementary Fig. S7A). This depleted antibodyalso showed very little signal in IHC staining of human breasttumor samples compared with control Endo II antibody (Sup-plementary Fig. S7B). We went on to profile Endo II expression inhuman breast tumors and grouped them according to molecularsubtypes of luminal (ERþ/PgRþ), HER2 (HER2þ), and triple-negative (ER�/PgR�/HER2�). Although all tumors showeddetectable Endo II expression, the highest levels were observedin triple-negative tumors (Fig. 7A). To quantify and analyze theseresults, we converted our IHC signals to H-Scores using imagingsoftware.Whenorganized by subtype,we observed a trend towardhigher Endo II expression in TNBC tumors compared with lumi-

nal or HER2 cases (Fig. 7B). It is worth noting that this result isconsistent with our profiling of Endo II levels in breast cancer celllines isolated from TNBCs (Fig. 1A; Supplementary Fig. S1). Toextend this study of Endo II expression to samples with knownoutcomes, we queried results of gene expression data for �200basal-like breast cancer patients using KM Plotter (36), and thisrevealed significantly higher risk of relapse for patients withtumors characterized by high levels of Endo II mRNA expression(Fig. 7C). Similar results were not observed in patients with otherbreast cancer subtypes (data not shown). Together, these resultssuggest that high levels of Endo II expression may contribute toTNBC progression and metastasis, which is supported by ourfunctional studies of Endo II in TNBC xenograft models.

DiscussionDynamic changes in membrane shape and composition are

central to many biologic processes (37). Invasive cancer cells takeadvantage of these processes during cancer cell invasion andtumor metastasis (38). This study adds to a growing list ofmembrane shaping BAR proteins that have been implicated incancermetastasis (14, 15, 19, 22, 39, 40). Through stable silencingof Endo II in TNBC cell models, we provide evidence that Endo IIpromotes internalization of EGFR and downstream signaling toErk and Akt pathways. Endo II KD cells display defects in cellmotility and invasion that correlatedwith a striking defect in ECM

Figure 6.Endo II regulates TNBC tumor growthand metastasis in mice. A, graphdepicts tumormass following injectionof MDA-MB-231 NT or Endo II KD cellsinto the mammary fat pad at 5 weeks(N ¼ 12/group from 3 independentexperiments; � , P < 0.05). B,representative images of Ki67 IHCstaining in NT and Endo II KD primarytumor sections. C, Endo II silencing invivo was assessed by IB of primarytumor homogenates. D,representative images of lungmetastases in mice harboring MDA-MB-231 NT or Endo II KD mammarytumors at 5 weeks. Representativeepifluorescence images ofwhole lungsto visualize GFPþ lungmets are shown(4� objective). Also shown arebrightfield images of H&E-stained lungsections (10� objective). E, graphdepicts numbers of metastasesobserved in H&E lung sections (N¼ 12/group, �� , P < 0.001).






degradation activity in two TNBC cell models. The lack ofmature invadopodia in Endo II KD cells may be due to defectsin MT1-MMP internalization and subsequent trafficking in theabsence of Endo II. Consistent with role of invadopodia inpromoting tumor progression and metastasis in TNBC (23, 41),we also show that Endo II promotes metastasis in TNBC tumorxenograft models. Lastly, we identify a link between elevatedEndo II expression and an increased risk of relapse in basal-likebreast cancer patients.

Endo II has recently been studied in mammary carcinoma cellsisolated from theMMTV-PyMTmodel crossed with FAK knock-inmice withmutations to the SH3 domain–binding motif for EndoII (þ/PA-MT and PA/PA-MT; ref. 22). Endo II association withFAK/Src kinases is retained in þ/PA-MT cells, which allowsphosphorylation (pY315)-mediated inactivation of Endo II andreduced MT1-MMP internalization (21, 22). The þ/PA-MT cellshave high surface levels of MT1-MMP and low E-cadherin levels

comparedwith PA/PA-MT cells. Endo II silencing had no effect onsurface MT1-MMP inþ/PA-MT cells, but did promote MT1-MMPsurface expression in PA/PA-MT cells (22). Although it is difficultto make direct comparisons between our human TNBC modelsand those in the Fan and colleagues study, we have also observedincreased surface levels of MT1-MMP with Endo II silencing.However, this did not correlate with increased ECM degradationactivity in Endo II KD TNBC cells, as reported in Src-transformedfibroblasts (21). In fact, we observe impaired ECMdegradation inEndo II KD TNBC cells. This likely relates to the complex traf-ficking of MT1-MMP from the cell surface to late endosomes andregulated delivery to invadopodia precursors (42, 43). Becauseour TNBC cell models are not engineered to disrupt the FAKinteraction with Endo II, we hypothesize that Endo II can largelyescape inhibition by Src, and thereby promote endocytosis ofMT1-MMP at membrane sites lacking activation of the FAK/Srccomplex. Here, we show that internalization of MT1-MMP is

Figure 7.High levels of Endo II in human TNBCsand link with poor prognosis. A, threerepresentative images are shown forEndo II IHC staining in 50 humanbreast carcinomas organizedaccording to molecular subtypes.B, graph depicts results of H-scoreanalysis of Endo II IHC signal in tissuemicroarray (TMA) cohortscorresponding to luminal, HER2, andtriple-negative subtypes. C, Kaplan–Meier plot comparing relapse-freesurvival of approximately 200 basal-like breast cancer patientswith high orlow Endo II (sh3gl1) mRNA expression.

Baldassarre et al.





impaired in Endo II KD MDA-MB-231 cells subjected to surfacebiotin assays. We hypothesize that lack of an endosomal poolof MT1-MMP may limit subsequent delivery to invadopodiaprecursors in Endo II KD TNBC cells. It is worth noting that theEndo II binding partner Dynamin has recently been implicatedin the formation and fission of tubulovesicular carriers fromthis intracellular pool of MT1-MMP (44). Like our findingshere for Endo II, Dynamin was also identified as a positiveregulator of invadopodia formation and cell invasion in MDA-MB-231 cells (45). In future, the Endo II KD TNBC cell modelsthat we have developed can be used to define the cellularcompartments of MT1-MMP that are most affected by Endo IIsilencing.

In this study, we identified a role for Endo II in promotinginternalization of EGFR, which is a key pathway linked to TNBCtumormetastasis (46, 47). Our results are consistent with a recentstudy showing that Endo II associates with Lamellipodin, an actinfilament elongation protein, to promote internalization of EGFRin HeLa cells (20). Recent studies also implicate Endo II andLamellipodin in mediating a distinct, non–clathrin-dependent,fast endocytosis pathway for ligand-activated receptors (48, 49).Although we have not been able to confirm Lamellipodin inter-action with Endo II in TNBC cells (data not shown), it will beinteresting to test for potential defects in membrane recruitmentof Lamellipodin in Endo II KD cells. It is worth noting thatLamellipodin was shown to localize to the leading edge ofMDA-MB-231 cells in a phosphoinositide (PI(3,4)P2)-dependentmanner (50). Since PI(3,4)P2 is enriched in invadopodia pre-cursors (51), it will be interesting to test whether Lamellipodinalso localizes to invadopodia to regulate actin polymerization,and whether this depends on Endo II. Endo II was also recentlyidentified as an interacting partner for the proline-rich region ofTks4 adaptor protein (52). It is worth noting that lung metastasisof tumor cells requires Tks4-induced invadopodia formation andmaturation (53). Future studies of this Tks4–Endo II axis will berequired to understand their roles in formation or function ofinvadopodia in TNBC cells (52). Another recently describedbinding partner of Endo II is the exchange factor for Arf6 (EFA6)whose activity is enhanced by Endo II leading to increased Arf6GTP

signaling at cell protrusions (54). Interestingly, Arf6 was previ-ously shown to promote Erk activation, invadopodia formation,and cell invasion inmelanoma andbreast cancermodels (55, 56).Given the defects in EGFR signaling to Erk that we have observedin Endo II KD cells, it will be interesting to test whether this is dueto altered Arf6 activation, or altered transit time of EGFR withinsignaling endosomes. A recent study of proteins involved in EGFRinternalization provides further support for compartmentalizedsignaling of EGFR from endosomes (10). In their study, AnnexinA2 silencing enhanced EGF uptake, Akt activation, and lungmetastasis in xenograft models. In our study, Endo II silencingled to defects in EGFR internalization, Akt activation, and lungmetastasis in xenograft assays. Because both Erk andAkt pathwaysplay key roles in tumor progression, it will be interesting to definethe gene expression signature of TNBCs in Endo II KD tumors.

Our analysis of Endo II expression in human breast tumorsrevealed a trend toward high Endo II expression in TNBC cases,compared with luminal and HER2 tumors. These results areconsistent with our results in surveying breast cancer cell lines.Considering the importance of receptor cargos (MT1-MMP,EGFR) and signaling pathways (Akt, Erk) that are regulated byEndo II in TNBC cell models, it will be interesting to perform

parallel studies of these signaling proteins in tumors with varyinglevels of Endo II. Since our tumor xenograft studies identify EndoII as a metastasis-promoting factor, it will be important to profileEndo II expression and the above pathways in paired samples ofprimary tumors and lymph node or distant metastases. Given theevidence that FAK/Src kinases phosphorylate Endo II at Y315leading to loss of binding to SH3 domain partners (21), it wouldbe very interesting to profile Endo II pY315 expression andlocalization within TNBC cells and tumors. However, this willrequire preparation of new phospho-specific antibodies for EndoII. For a widely expressed gene, it is interesting that expression ofthe sh3gl1 gene encoding Endo II at levels above or below themedian (within the primary tumor) was associated with risk ofrelapse in basal-like breast cancer patients. These results will spurfurther testing of this association at the Endo II protein level inpatient cohorts with known clinical outcomes. Given that someofthe pathways regulatedby Endo II are being considered in targetedtherapies, the relative levels of Endo II expression in these patientsamplesmaybe relevant to the response to therapy. It is alsoworthnoting that in the rapidly emerging field of antibody–drug con-jugates for cancer therapy, biomarkers related to the endocyticpathways may be useful to predict response, since receptor inter-nalization is a necessary step for release of these anticancer drugs(57).

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

DisclaimerThe funders had no role in study design, data collection or analysis, decision

to publish, or preparation of the article.

Authors' ContributionsConception and design: T. Baldassarre, A.W. CraigDevelopment of methodology: T. Baldassarre, P. Truesdell, A.W. CraigAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): T. Baldassarre, J. MeensAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): T. Baldassarre, K. Watt, M.M. Schneider, S.K. Sen-gupta, A.W. CraigWriting, review, and/or revision of the manuscript: T. Baldassarre, K. Watt,M.M. Schneider, S.K. Sengupta, A.W. CraigAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): J. MeensStudy supervision: S.K. Sengupta, A.W. Craig

AcknowledgmentsThe authors thank Andrea Cowan for generating Endo II KD MDA-MB-231

cells and Stephanie Everingham for technical assistance.

Grant SupportThis research was supported by operating grants from Canadian Institutes

of Health Research (MOP 119562) and Cancer Research Society to AWBC.T. Baldassarre and K. Watt are supported by the Terry Fox FoundationTraining Program in Transdisciplinary Cancer Research at the Queen'sUniversity.

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.

ReceivedOctober 22, 2014; revised February 9, 2015; acceptedMarch 8, 2015;published OnlineFirst March 17, 2015.






References1. Ferlay J, Forman D, Mathers CD, Bray F. Breast and cervical cancer in 187

countries between 1980 and 2010. Lancet 2012;379:1390–1.2. Hudis CA, Gianni L. Triple-negative breast cancer: an unmet medical need.

Oncol 2011;16 Suppl 1:1–11.3. Park HS, Jang MH, Kim EJ, Kim HJ, Lee HJ, Kim YJ, et al. High EGFR gene

copy number predicts poor outcome in triple-negative breast cancer.Modern Pathol 2014;27:1212–22.

4. Koletsa T, Kotoula V, Karayannopoulou G, Nenopoulou E, Karkavelas G,Papadimitriou CS, et al. EGFR expression and activation are common inHER2 positive and triple-negative breast tumours. Histol Histopathol2010;25:1171–9.

5. Park S, JungHH, Park YH, Ahn JS, ImYH. ERK/MAPKpathways play criticalroles in EGFR ligands-induced MMP1 expression. Biochem Biophys ResComm 2011;407:680–6.

6. Ajiro M, Nishidate T, Katagiri T, Nakamura Y. Critical involvement ofRQCD1 in the EGFR-Akt pathway inmammary carcinogenesis. Intl JOncol2010;37:1085–93.

7. Nunes-Xavier CE, Elson A, Pulido R. Epidermal growth factor receptor(EGFR)-mediated positive feedback of protein-tyrosine phosphatase epsi-lon (PTPepsilon) on ERK1/2 and AKT protein pathways is required forsurvival of human breast cancer cells. J Biol Chem 2012;287:3433–44.

8. Sigismund S, Argenzio E, Tosoni D, Cavallaro E, Polo S, Di Fiore PP.Clathrin-mediated internalization is essential for sustained EGFR signalingbut dispensable for degradation. Dev Cell 2008;15:209–19.

9. Murphy JE, Padilla BE, Hasdemir B, Cottrell GS, Bunnett NW. Endosomes:a legitimate platform for the signaling train. Proc Natl Acad Sci U S A2009;106:17615–22.

10. deGraauwM,Cao L,Winkel L, vanMiltenburgMH, leDevedec SE, KlopM,et al. Annexin A2 depletion delays EGFR endocytic trafficking via cofilinactivation and enhances EGFR signaling and metastasis formation. Onco-gene 2014;33:2610–9.

11. Itoh T, De Camilli P. BAR, F-BAR (EFC) and ENTH/ANTH domains in theregulation of membrane-cytosol interfaces and membrane curvature.Biochim Biophys Acta 2006;1761:897–912.

12. Rao Y, Haucke V. Membrane shaping by the Bin/amphiphysin/Rvs(BAR) domain protein superfamily. Cell Mol Life Sci 2011;68:3983–93.

13. Suetsugu S, Gautreau A. Synergistic BAR-NPF interactions in actin-drivenmembrane remodeling. Trends Cell Biol 2012;22:141–50.

14. Ahn J, Truesdell P, Meens J, Kadish C, Yang X, Boag AH, et al. Fer protein-tyrosine kinase promotes lung adenocarcinoma cell invasion and tumormetastasis. Mol Cancer Res 2013;11:952–63.

15. Chander H, Truesdell P, Meens J, Craig AW. Transducer of Cdc42-depen-dent actin assembly promotes breast cancer invasion and metastasis.Oncogene 2013;32:3080–90.

16. Pichot CS, Arvanitis C, Hartig SM, Jensen SA, Bechill J, Marzouk S, et al.Cdc42-interacting protein 4 promotes breast cancer cell invasion andformation of invadopodia through activation of N-WASp. Cancer Res2010;70:8347–56.

17. Hu J, Mukhopadhyay A, Craig AW. Transducer of Cdc42-dependent actinassembly promotes epidermal growth factor-induced cell motility andinvasiveness. J Biol Chem 2011;286:2261–72.

18. Hu J, Mukhopadhyay A, Truesdell P, Chander H, Mukhopadhyay UK, MakAS, et al. Cdc42-interacting protein 4 is a Src substrate that regulatesinvadopodia and invasiveness of breast tumors by promoting MT1-MMPendocytosis. J Cell Sci 2011;124:1739–51.

19. Truesdell P, Ahn J, ChanderH,Meens J,Watt K, YangX, et al. CIP4promoteslung adenocarcinoma metastasis and is associated with poor prognosis.Oncogene 2014 Sep 1. [Epub ahead of print].

20. Vehlow A, Soong D, Vizcay-Barrena G, Bodo C, Law AL, Perera U, et al.Endophilin, Lamellipodin, andMena cooperate to regulate F-actin-depen-dent EGF-receptor endocytosis. EMBO J 2013;32:2722–34.

21. Wu X, Gan B, Yoo Y, Guan JL. FAK-mediated src phosphorylation ofendophilin A2 inhibits endocytosis of MT1-MMP and promotes ECMdegradation. Dev Cell 2005;9:185–96.

22. Fan H, Zhao X, Sun S, Luo M, Guan JL. Function of focal adhesion kinasescaffolding to mediate endophilin A2 phosphorylation promotes epithe-lial-mesenchymal transition and mammary cancer stem cell activities invivo. J Biol Chem 2013;288:3322–33.

23. Perentes JY, Kirkpatrick ND,Nagano S, Smith EY, Shaver CM, Sgroi D, et al.Cancer cell-associated MT1-MMP promotes blood vessel invasion anddistant metastasis in triple-negative mammary tumors. Cancer Res2011;71:4527–38.

24. Henkemeyer M, Rossi DJ, Holmyard DP, Puri MC, Mbamalu G, Harpal K,et al. Vascular system defects and neuronal apoptosis in mice lacking rasGTPase-activating protein. Nature 1995;377:695–701.

25. Sharma N, Everingham S, Zeng LF, Zhang ZY, Kapur R, Craig AW. Onco-genic KIT-induced aggressive systemic mastocytosis requires SHP2/PTPN11 phosphatase for disease progression in mice. Oncotarget 2014;5:6130–41.

26. MagalhaesMA, LarsonDR,Mader CC, Bravo-Cordero JJ, Gil-HennH, OserM, et al. Cortactin phosphorylation regulates cell invasion through a pH-dependent pathway. J Cell Biol 2011;195:903–20.

27. Hoeflich KP, O'Brien C, Boyd Z, Cavet G, Guerrero S, Jung K, et al. In vivoantitumor activity of MEK and phosphatidylinositol 3-kinase inhibitors inbasal-like breast cancer models. Clin Cancer Res 2009;15:4649–64.

28. Stinson S, Lackner MR, Adai AT, Yu N, Kim HJ, O'Brien C, et al. TRPS1targeting by miR-221/222 promotes the epithelial-to-mesenchymal tran-sition in breast cancer. Sci Signal 2011;4:ra41.

29. 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.

30. Hill E, van Der Kaay J, Downes CP, Smythe E. The role of dynamin and itsbinding partners in coated pit invagination and scission. J Cell Biol2001;152:309–23.

31. Milosevic I, Giovedi S, Lou X, Raimondi A, Collesi C, Shen H, et al.Recruitment of endophilin to clathrin-coated pit necks is required forefficient vesicle uncoating after fission. Neuron 2011;72:587–601.

32. Neumann S, Schmid SL. Dual role of BAR domain-containing proteins inregulating vesicle release catalyzed by the GTPase, dynamin-2. J Biol Chem2013;288:25119–28.

33. Wang HQ, Altomare DA, Skele KL, Poulikakos PI, Kuhajda FP, Di Cristo-fano A, et al. Positive feedback regulation between AKT activation and fattyacid synthase expression in ovarian carcinoma cells. Oncogene 2005;24:3574–82.

34. KohnoM, Tanimura S,Ozaki K. Targeting the extracellular signal-regulatedkinase pathway in cancer therapy. Biol Pharm Bull 2011;34:1781–4.

35. Altomare DA, Testa JR. Perturbations of the AKT signaling pathway inhuman cancer. Oncogene 2005;24:7455–64.

36. Gyorffy B, Lanczky A, Eklund AC, Denkert C, Budczies J, Li Q, et al. Anonline survival analysis tool to rapidly assess the effect of 22,277 genes onbreast cancer prognosis using microarray data of 1,809 patients. BreastCancer Res Treat 2010;123:725–31.

37. Mim C, Unger VM. Membrane curvature and its generation by BARproteins. Trends Biochem Sci 2012;37:526–33.

38. Scita G, Di Fiore PP. The endocytic matrix. Nature 2010;463:464–73.

39. Ivanova IA, Vermeulen JF, Ercan C, Houthuijzen JM, Saig FA, Vlug EJ, et al.FER kinase promotes breast cancer metastasis by regulating alpha6- andbeta1-integrin-dependent cell adhesion and anoikis resistance. Oncogene2013;32:5582–92.

40. Koshkina NV, Yang G, Kleinerman ES. Inhibition of Cdc42-interactingprotein 4 (CIP4) impairs osteosarcoma tumor progression. Current CancerDrug Targets 2013;13:48–56.

41. Jiang WG, Davies G, Martin TA, Parr C, Watkins G, Mason MD, et al.Expression of membrane type-1 matrix metalloproteinase, MT1-MMP inhuman breast cancer and its impact on invasiveness of breast cancer cells.Intl J Mol Med 2006;17:583–90.

42. Remacle A, Murphy G, Roghi C. Membrane type I-matrix metalloprotei-nase (MT1-MMP) is internalised by two different pathways and is recycledto the cell surface. J Cell Sci 2003;116:3905–16.

43. Poincloux R, Lizarraga F, Chavrier P. Matrix invasion by tumour cells: afocus on MT1-MMP trafficking to invadopodia. J Cell Sci 2009;122:3015–24.

44. Rosse C, Lodillinsky C, Fuhrmann L, Nourieh M, Monteiro P, Irondelle M,et al. Control of MT1-MMP transport by atypical PKC during breast-cancerprogression. Proc Natl Acad Sci U S A 2014;111:E1872–9.

Baldassarre et al.





45. Baldassarre M, Pompeo A, Beznoussenko G, Castaldi C, Cortellino S,McNiven MA, et al. Dynamin participates in focal extracellular matrixdegradation by invasive cells. Mol Biol Cell 2003;14:1074–84.

46. UenoNT, ZhangD. Targeting EGFR in triple negative breast cancer. J Cancer2011;2:324–8.

47. de Ruijter TC, Veeck J, de Hoon JP, van Engeland M, Tjan-Heijnen VC.Characteristics of triple-negative breast cancer. J Cancer Res Clin Oncol2011;137:183–92.

48. Renard HF, Simunovic M, Lemiere J, Boucrot E, Garcia-Castillo MD,Arumugam S, et al. Endophilin-A2 functions in membrane scission inclathrin-independent endocytosis. Nature 2015;517:493–6.

49. Boucrot E, Ferreira AP, Almeida-Souza L, Debard S, Vallis Y, Howard G,et al. Endophilin marks and controls a clathrin-independent endocyticpathway. Nature 2015;517:460–5.

50. Zou L, Ding Z, Roy P. Profilin-1 overexpression inhibits proliferation ofMDA-MB-231 breast cancer cells partly through p27kip1 upregulation. JCell Physiol 2010;223:623–9.

51. Oikawa T, Itoh T, Takenawa T. Sequential signals toward podosomeformation in NIH-src cells. J Cell Biol 2008;182:157–69.

52. Havrylov S, Park M. MS/MS-based strategies for proteomic profiling ofinvasive cell structures. Proteomics 2015;15:272–86.

53. Leong HS, Robertson AE, Stoletov K, Leith SJ, Chin CA, Chien AE, et al.Invadopodia are required for cancer cell extravasation and are a therapeutictarget for metastasis. Cell Rep 2014;8:1558–70.

54. Boulakirba S, Macia E, Partisani M, Lacas-Gervais S, Brau F, Luton F, et al.Arf6 exchange factor EFA6 and endophilin directly interact at the plasmamembrane to control clathrin-mediated endocytosis. Proc Natl Acad SciU S A 2014;111:9473–8.

55. Tague SE, Muralidharan V, D'Souza-Schorey C. ADP-ribosylationfactor 6 regulates tumor cell invasion through the activation of theMEK/ERK signaling pathway. Proc Natl Acad Sci U S A 2004;101:9671–6.

56. Hashimoto S, Onodera Y, Hashimoto A, Tanaka M, Hamaguchi M,Yamada A, et al. Requirement for Arf6 in breast cancer invasive activities.Proc Natl Acad Sci U S A 2004;101:6647–52.

57. Ritchie M, Tchistiakova L, Scott N. Implications of receptor-mediatedendocytosis and intracellular trafficking dynamics in the development ofantibody drug conjugates. mAbs 2013;5:13–21.






2015;13:1044-1055. Published OnlineFirst March 17, 2015.Mol Cancer Res Tomas Baldassarre, Kathleen Watt, Peter Truesdell, et al. Endophilin A2 Promotes TNBC Cell Invasion and Tumor Metastasis

Updated version

10.1158/1541-7786.MCR-14-0573doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2015/03/18/1541-7786.MCR-14-0573.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/13/6/1044.full#ref-list-1

This article cites 56 articles, 21 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/13/6/1044.full#related-urls

This article has been cited by 4 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

[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://mcr.aacrjournals.org/content/13/6/1044To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-14-0573

http://mcr.aacrjournals.org/content/suppl/2015/03/18/1541-7786.MCR-14-0573.DC1

http://mcr.aacrjournals.org/content/13/6/1044.full#ref-list-1

http://mcr.aacrjournals.org/content/13/6/1044.full#related-urls

http://mcr.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mcr.aacrjournals.org/content/13/6/1044


Documents

Endophilin A2 Promotes TNBC Cell Invasion and Tumor ......Signal Transduction Endophilin A2 Promotes TNBC Cell Invasion and Tumor Metastasis Tomas Baldassarre1,2, Kathleen Watt1,2,