Lipid Rafts and Caveolin-1 Are Required for Invadopodia Formation and Extracellular Matrix Degradation by Human Breast Cancer Cells

Home , Caveolin 1, Invadopodia, Lipid raft

Published OnlineFirst November 3, 2009; DOI: 10.1158/0008-5472.CAN-09-2305

Cell, Tumor, and Stem Cell Biology

Hideki Yamaguchi,1,2 Yukiko Takeo,1 Shuhei Yoshida,1 Zen Kouchi,1 Yoshikazu Nakamura,1 and Kiyoko Fukami1

1Laboratory of Genome and Biosignal, Tokyo University of Pharmacy and Life Sciences, Hachioji, Tokyo, Japan and 2PRESTO, Japan Science and Technology Agency, Kawaguchi, Saitama, Japan

Abstract for cancer invasion and metastasis (3). Invadopodia are enriched Invadopodia are ventral membrane protrusions through with actin filaments (F-actin) and proteins involved in the actin which invasive cancer cells degrade the extracellular matrix. cytoskeleton, cell signaling, cell-ECM adhesion, and membrane re- They are thought to function in the migration of cancer modeling (4, 5). The matrix degradation activity of invadopodia is cells through tissue barriers, which is necessary for cancer mainly mediated by the focal concentration of membrane type 1 invasion and metastasis. Although many protein compo- matrix metalloproteinase (MT1-MMP) at the surface of invadopo- – nents of invadopodia have been identified, the organization dia (6 8). However, the molecular mechanisms underlying the re- and the role of membrane lipids in invadopodia are not cruitment of invadopodia components are not well understood. well understood. In this study, the role of lipid rafts, which Many plasma membrane proteins are compartmentalized into are cholesterol-enriched membrane microdomains, in the sphingolipid- and cholesterol-rich microdomains called lipid rafts assembly and function of invadopodia in human breast can- (9). Lipid rafts are enriched with a variety of proteins, including cer cells was investigated. Lipid rafts are enriched, internal- caveolins and flotillins (10), and thought to function as platforms ized, and dynamically trafficked at invadopodia sites. for localizing signaling proteins and eliciting spatially controlled Perturbation of lipid raft formation due to depleting or se- signal transduction (11). Because lipid raft membranes are actively questering membrane cholesterol blocked the invadopodia- internalized and trafficked in the cell, lipid rafts are also involved mediated degradation of the gelatin matrix. Caveolin-1 in the transport of lipid raft-associated molecules (12). Recent (Cav-1), a resident protein of lipid rafts and caveolae, studies have confirmed the importance of lipid rafts and associat- accumulates at invadopodia and colocalizes with the inter- ed proteins in the pathogenesis of several diseases, including can- nalized lipid raft membranes. Membrane type 1 matrix cer progression (13, 14). metalloproteinase (MT1-MMP), a matrix proteinase associ- Caveolae are flask-shaped membrane invaginations and major ated with invadopodia, is localized at lipid raft-enriched subtypes of lipid rafts (15). Caveolins are integral membrane pro- membrane fractions and cotrafficked and colocalized with teins and major protein components of caveolae and lipid rafts Cav-1 at invadopodia. The small interfering RNA–mediated (16). The caveolin family consists of three isoforms in mammals: silencing of Cav-1 inhibited the invadopodia-mediated and caveolin (Cav)-1, -2, and -3 (17). Cav-1 is coexpressed with Cav-2 MT1-MMP–dependent degradation of the gelatin matrix. in a variety of tissues, whereas Cav-3 expression is restricted to Furthermore, Cav-1 and MT1-MMPare coexpressed in inva- muscle tissues (17). Genetic studies with knockout mice showed sive human breast cancer cell lines that have an ability to that Cav-1, but not Cav-2, is essential for caveolae formation form invadopodia. These results indicate that invadopodia (15). Cav-1 regulates diverse cellular processes, including raft- are the sites where enrichment and trafficking of lipid mediated endocytosis, vesicular transport, cell migration, and – rafts occur and that Cav-1 is an essential regulator of signal transduction (16 19). Clinical studies revealed that Cav-1 MT1-MMPfunction and invadopodia-mediated breast upregulation is associated with a poor prognosis and the occur- cancer cell invasion. [Cancer Res 2009;69(22):8594–602] rence of metastatic regions in several human cancers (20, 21). It was recently reported that Cav-1 regulates cell migration and invasion in human breast cancer cells (22). However, the molecular Introduction mechanisms by which Cav-1 regulates the ECM remodeling asso- Invadopodia are ventral membrane protrusions that can de- ciated with cancer cell invasion are poorly understood. grade the extracellular matrix (ECM; ref. 1). A variety of invasive In this study, the roles of lipid rafts and associated molecules in tumor cells, including mammary adenocarcinoma, are reported invadopodia formation are investigated in human breast cancer to form invadopodia when they are cultured on physiologic sub- cells. It is shown that invadopodia are the sites where lipid raft strates (2). Invadopodia are thought to function in cancer cell mi- formation and trafficking occur and that Cav-1 is an essential reg- gration in vivo through the physical barriers of the dense ECM ulator of invadopodia-mediated breast cancer cell invasion. existing in the tumor microenvironment, which is a prerequisite

Materials and Methods Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). Cell culture. Human breast cancer cell lines MDA-MB-231, T47D, Requests for reprints: Hideki Yamaguchi, Laboratory of Genome and Biosignal, BT549, and Hs578T were obtained from the American Type Culture Col- Tokyo University of Pharmacy and Life Sciences, 1432-1 Horinouchi, Hachioji, Tokyo lection. MDA-MB-453 and SK-BR-3 were obtained from RIKEN BRC Cell 192-0392, Japan. Phone: 81-42-676-7232; Fax: 81-42-676-7249; E-mail: hyamaguc@ Bank and Cell Resource Center for Biomedical Research Tohoku Univer- toyaku.ac.jp. ©2009 American Association for Cancer Research. sity, respectively. MDA-MB-231 cells were maintained in a 1:1 mixture of doi:10.1158/0008-5472.CAN-09-2305 high-glucose DMEM and RPMI 1640 supplemented with 10% fetal bovine

Cancer Res 2009; 69: (22). November 15, 2009 8594 www.aacrjournals.org

Figure 1. Localization, internalization, and trafficking of lipid rafts at invadopodia. A, MDA-MB-231 cells were cultured on fluorescent gelatin-coated coverslips for 4 h and the lipid rafts of the plasma membrane were labeled with Alexa 488-CTxB. Cells were then stained with phalloidin to visualize F-actin–richinvadopodia. Top and bottom images are the XY and XZ sections, respectively, taken by confocal microscopy. Insets, magnified images of the boxed regions. Arrows, invadopodia and associated gelatin degradation sites. B, MDA-MB-231 cells cultured on gelatin-coated coverslips for 4 hwere labeled withAlexa 555-CTxB and incubated at 37°C for 10 min. Cells were fixed and the surface-bound CTxB was stained with anti-CTxB antibody before cell permeabilization and phalloidin staining. To visualize the internalized CTxB, the signals for the surface CTxB were subtracted from the Alexa 555 (total)-CTxB image. Arrows, position of invadopodia associated withinternalized CTxB signals. C, MDA-MB-231 cells stably expressed GFP-actin were incubated withAlexa 555-CTxB for 10 min and thenimmediately observed using time-lapse fluorescent microscopy. Right, time-lapse image sequences of the boxed regions. Arrows, raft membranes transported from (a) and to (b) invadopodia. Numbers in the images denote the time points in seconds. serum, 10 units/mL penicillin, and 10 μg/mL streptomycin. Other cell Antibodies. Anti–flotillin-1 (Flot-1), Cav-1, Cav-2, and Cdc42 antibodies lines were maintained according to the cell banks' instructions. were purchased from BD Biosciences. Anti–MT1-MMP antibodies were Constructs. For Cav-1 constructs, cDNAs encoding human and mouse from Abcam (ab38971) and Daiichi Fine Chemical (114-6G6). Anti-cortactin Cav-1 were amplified by reverse transcription-PCR with total RNA prepared and β-actin antibodies were from Millipore. Anti–cholera toxin B subunit from MDA-MB-231 cells and mouse intestine, respectively. The cDNAs were (CTxB),Src,andArf6antibodieswerefromCalbiochem,CellSignaling subcloned into pEGFP-C1, pECFP-C1, and pmCherry-C1 vectors (Clontech). Technology, and Santa Cruz Biotechnology, respectively. Anti–N-WASP For MT1-MMP constructs, the cDNA clone of human MT1-MMP was ob- antibody was described previously (23). tained from Kazusa DNA Research Institute and subcloned into pEGFP-N1 Invadopodia assay. Fluorescent gelatin-coated coverslips were pre- and pmCherry-N1 vectors (Clontech). pared as described previously (24). Breast cancer cells were cultured on www.aacrjournals.org 8595 Cancer Res 2009; 69: (22). November 15, 2009

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/0008-5472.CAN-09-2305 Cancer Research the coverslips for indicated periods. In general, MDA-MB-231 cells begin ing to the manufacturer's instructions. For retroviral infection, cDNAs were to degrade the gelatin matrix ∼3 h after plating. In the case of cells over- inserted into the pMXs-IP vector and recombinant retroviruses were pro- expressing MT1-MMP, they were cultured on the gelatin matrix for 2 h. duced with the Plat-A packaging cell line as described previously (25). Cells To quantitate the degradation activity of invadopodia, 10 randomly se- were then infected with the recombinant retroviruses and selected with 0.5 lected fields, usually containing 30 to 50 cells, were imaged with a ×60 μg/mL puromycin. objective for each determination. The degradation area was determined Reverse transcription-PCR. Total RNA was isolated with a RNeasy by using ImageJ 1.41 software and normalized for the number of cells. Mini Kit (Qiagen). Template cDNAs were synthesized with SuperScript III In each analysis, the mean value of the control cells was set at 100% (Invitrogen). PCR amplification was done with KOD plus DNA polymerase and the relative values of the cells treated with small interfering RNAs (Toyobo). The primer pairs used were as follows: human Cav-1 (5′- (siRNA) or inhibitors were then calculated. Data are mean ± SE of at least ATGTCTGGGGGCAAATACG-3′ and 5′-TTATATTTCTTTCTGCAAGTT- four independent determinations. GATGC-3′), mouse Cav-1 (5′-ATGTCTGGGGGCAAATACG-3′ and 5′-TCA- Immunofluorescence and time-lapse microscopy. Cells were stained TATCTCTTTCTGCGTGCTG-3′), and human cyclophilin B2 (5′- with the indicated antibodies and phalloidin as described previously (23). GCACAGGAGGAAAGAGCATC-3′ and 5′-CTTCTCCACCTCGATCTTGC-3′). The cells were observed with an Olympus IX81-ZDC-DSU confocal micro- Labeling cells with CTxB. Cells cultured on gelatin-coated coverslips scope equipped with a cooled CCD camera (ORCA-ER; Hamamatsu) or were washed with ice-cold PBS and incubated with 10 μg/mL fluorescent with an Olympus confocal laser scanning microscope FV-1000. Time-lapse CTxB (Invitrogen) in PBS for 20 min on ice. For the internalization experi- series of the cells were taken at 37°C using an Olympus IX81-ZDC-DSU mi- ments, cells were labeled with CTxB in the growth medium for 20 min on croscope equipped with a cooled CCD camera, humidified CO2 chamber, ice and then incubated for 10 min at 37°C. After fixing in 4% formaldehyde and autofocus system and operated by MetaMorph software (Universal Im- for 20 min, the cells were labeled with anti-CTxB antibody in 1% bovine aging Corp.). serum albumin and 1% fetal bovine serum in PBS for 1 h at room temper- RNA interference. The stealth RNA interference molecules (Invitrogen) ature. For time-lapse imaging, cells were incubated with CTxB in the used were Negative Control Medium GC Duplex #2 and Stealth Select RNA growth medium for 10 min at 37°C, washed with growth medium, and then interference for Cav-1 (HSS141466 and HSS141467), Cav-2 (HSS101398), and immediately imaged. Flot-1 (HSS115568). Cells were transfected with 30 nmol/L siRNA using Cholesterol depletion, sequestering, replenishment, and quantita- Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's in- tion. To deplete cellular cholesterol, cells were pretreated with 5 mmol/L structions. The cells were cultured for 48 h and subjected to invadopodia methyl-β-cyclodextrin (MβCD; Sigma) for 30 min at 37°C and subjected to assay or immunoblotting. invadopodia assay in the presence of 1 mmol/L MβCD or water-soluble Immunoblotting. Immunoblotting was done as described previously cholesterol balanced with MβCD (137 μmol/L cholesterol/1 mmol/L (23). Equal amounts of protein (10-40 μg) were loaded per lane for each MβCD; Sigma) for cholesterol replenishment. To sequester membrane cho- analysis. Protein concentration was determined by using BCA Protein As- lesterol, cells were pretreated with 50 μg/mL nystatin (Sigma) for 30 min at say kit (Pierce). 37°C and tested for invadopodia activity in the continuous presence of nys- Plasmid transfection and retroviral infection. Cells were transfected tatin. The relative amount of cellular cholesterol was determined with cho- with the indicated plasmids using Lipofectamine 2000 (Invitrogen) accord- lesterol/cholesterol ester quantitation kit (Calbiochem).

Figure 2. Lipid raft formation is necessary for invadopodia formation. A, MDA-MB-231 cells were pretreated with5 mmol/L M βCD for 30 min at 37°C and cultured on fluorescent gelatin-coated coverslips in the presence of 1 mmol/L MβCD or 137 μmol/L cholesterol/1 mmol/L MβCD complex for 7 h. Cells were then stained withphalloidinto visualize invadopodia. B, degradation area of the gelatin was quantified, as described in Materials and Methods, and shown as a percentage of the control cells. C and D, MDA-MB-231 cells were pretreated with DMSO or 50 μg/mL nystatin for 30 min at 37°C and cultured on fluorescent gelatin-coated coverslips for 7 h. Cells were then stained with phalloidin (C) and the degradation area of the gelatin was quantified (D). Columns, mean of at least four independent determinations; bars, SE. *, P < 0.02; **, P < 0.0004; ***, P < 0.00001, Student's t test.

Cancer Res 2009; 69: (22). November 15, 2009 8596 www.aacrjournals.org

Figure 3. Cav-1 is an essential regulator of invadopodia formation. A, MDA-MB-231 cells (MDA) and cells retrovirally transduced withmouse Cav-1 ( MDA msCav-1) were transfected withthe indicated siRNAs for 48 hand subjected to immunoblot analysis. Representative of three independent experiments. B, MDA-MB-231 and MDA msCav-1 cells transfected withtheindicated siRNAs were cultured on fluorescent gelatin-coated coverslips for 7 hand stained with phalloidin, and the degradation area of the gelatin was quantified. C, representative images of MDA-MB-231 and MDA msCav-1 cells transfected withcontrol or Cav-1 siRNA. D, MDA-MB-231 cells transfected withtheindicated siRNAs were replenished with 137 μmol/L cholesterol/1 mmol/L MβCD and tested for gelatin degradation activity for 7 h. Columns, mean of at least four independent determinations; bars, SE. *, P < 0.03; **, P < 0.0003, Student's t test.

Cell fractionation. Cells were plated onto 10 cm gelatin-coated culture tured for 10 min to allow them to internalize lipid raft membranes. 6 dishes at 1 × 10 per dish, and the cytosolic fraction (1 mL) was extracted CTxB remaining on the cell surface was then stained with anti- with the ProteoExtract Subcellular Proteome Extraction kit (Calbiochem). CTxB antibody before cell permeabilization. Most of the signals μ Triton X-100 soluble materials were extracted with 500 L TNE buffer [25 for total and surface CTxB were colocalized on the plasma mem- mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 5 mmol/L EDTA, and prote- brane (Fig. 1B). Internalized CTxB was then visualized by subtract- ase inhibitors] containing 1% Triton X-100. Insoluble materials were further extracted with 250 μL TNE buffer containing 1% SDS. Equal amounts of ing the surface CTxB signals from the total CTxB image. The protein from each fraction were analyzed by immunoblotting. resulting image showed that the signals for internalized CTxB Further details for all of these methods are available upon request. are localized at F-actin–rich invadopodia (Fig. 1B). Time-lapse fluorescent microscopy was also carried out with MDA-MB-231 cells that stably expressed GFP-actin. The cells were labeled with Results Alexa 555-CTxB for 10 min and immediately imaged. CTxB signals Lipid rafts are enriched, internalized, and trafficked at in- were observed as small vesicles and some of them were localized vadopodia in human breast cancer cells. We first confirmed around invadopodia (Fig. 1C; Supplementary Video 1). These vesi- the localization of lipid rafts at invadopodia by using a lipid raft cles were dynamically transported to and from invadopodia and marker, CTxB; CTxB binds to lipid raft-enriched GM1 ganglio- actively fragmented into small vesicles or fused into large clusters side and has been widely exploited to visualize lipid rafts (26). (Fig. 1C). Taken together, these results indicate that invadopodia MDA-MB-231 invasive human breast cancer cells were plated are enriched with lipid raft membranes where active raft mem- onto a fluorescently labeled gelatin matrix and the lipid rafts brane internalization and transport occur. of the plasma membrane were labeled with fluorescent CTxB. Lipid raft formation is necessary for invadopodia-mediated At the ventral surface of the cell, strong CTxB signals were ob- ECM degradation. The role of lipid raft formation in invadopodia- served at invadopodia, which were defined by the dot-like accu- mediated gelatin degradation was then assessed. MDA-MB-231 mulation of F-actin associated with the degradation sites of the cells were pretreated with MβCD, which depletes cellular choles- gelatin matrix (Fig. 1A). The green fluorescent protein (GFP)– terol and therefore disassembles lipid rafts (28); the cells were then tagged NH2-terminal sequence of Lyn (GFP-LynN), which was plated onto a gelatin matrix and tested for invadopodia formation. shown to target lipid rafts (27), was also accumulated at the MβCD treatment inhibited invadopodia formation and degrada- gelatin degradation sites (Supplementary Fig. S1). Nevertheless, tion of the gelatin matrix (Fig. 2A and B). CTxB staining of the GFP-LynN signals were more broadly distributed throughout MβCD-treated cells showed that patched signals of CTxB at the the plasma membrane than CTxB signals, probably because this ventral membrane were decreased in these cells (Supplementary probe is less specific than CTxB. Fig. S2). The invadopodia formation and matrix degradation activ- To follow the course of lipid raft membranes at invadopodia, ity of MβCD-treated cells were restored by cholesterol replenish- MDA-MB-231 cells were first labeled with Alexa 555-CTxB and cul- ment (Fig. 2A and B). In this experimental condition, the amounts www.aacrjournals.org 8597 Cancer Res 2009; 69: (22). November 15, 2009

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/0008-5472.CAN-09-2305 Cancer Research of cellular cholesterol in cells treated with MβCD and replenished siRNA targeting a different region of the Cav-1 gene (Supplemen- with cholesterol were 41% and 175% of that of untreated cells, re- tary Fig. S4A and B). The Cav-1 knockdown phenotype was rescued spectively (Supplementary Fig. S3A). by retroviral reintroduction of mouse Cav-1, which has mis- We also used the more specific lipid raft-inhibiting reagent matches in the siRNA targeting sequence (Fig. 3A-C). In contrast, nystatin, which binds and sequesters membrane cholesterol the knockdown of Cav-2 or Flot-1, a lipid raft-associated protein (29). Treatment of cells with 50 μg/mL nystatin significantly sup- involved in endocytosis (31), had no significant effect on invadopo- pressed invadopodia-mediated gelatin degradation (Fig. 2C and dia formation (Fig. 3A and B). Although it was reported that knock- D). At this concentration, nystatin had no significant effect on down of Flot-1 leads to downregulation of Cav-1 via lysosomal cell viability during the assay period (data not shown). Nystatin degradation (32), we could not detect any change in the amount treatment had little effect on the amount of cellular choles- of Cav-1 in Flot-1 knockdown cells in this experimental condition terol (Supplementary Fig. S3B), consistent with its cholesterol- (Fig. 3A). It was recently reported that invadopodia formation in sequestering but not extracting effect. These observations indicate human melanoma cells is regulated by Cav-1–mediated control that lipid raft formation is necessary for the assembly and function- of membrane cholesterol levels (33). However, Cav-1 knockdown ing of invadopodia. caused little change in the amount of cellular cholesterol, and Cav-1 is an essential regulator of invadopodia formation. the invadopodia formation in Cav-1 knockdown cells could not Caveolin is an essential component of caveolae, which are subtypes be rescued with provision of cholesterol in this cell type (Fig. 3D; of lipid rafts, and is known to regulate the organization and dy- Supplementary Fig. S4C). Similar results were obtained when lower namics of lipid rafts (17, 30). Therefore, the role of caveolin in in- concentrations of cholesterol were used for the provision (data not vadopodia formation was investigated. MDA-MB-231 cells were shown). These results suggest that Cav-1 is specifically involved in transfected with siRNA targeting Cav-1 or Cav-2 and then tested invadopodia formation by breast cancer cells independently of its for invadopodia formation. Immunoblot analysis confirmed the se- role in the regulation of membrane cholesterol levels. lective knockdown of the expression of each form of caveolin Next, the localization of endogenous Cav-1 at invadopodia was (Fig. 3A). Cells with a reduced amount of Cav-1 showed a signifi- examined using immunocytochemistry. Cav-1 signals were de- cant decrease in invadopodia formation and gelatin degradation tected as patched clusters distributed throughout the plasma activity (Fig. 3B and C). Similar results were obtained with another membrane and vesicle-like pattern at the submembrane region

Figure 4. Localization of endogenous Cav-1 at invadopodia. A, MDA-MB-231 cells were cultured on fluorescent gelatin-coated coverslips for 4 hand stained withanti-Cav-1 antibody. Top images are theXY sections of the ventral cell surface and bottom images are the XZ sections of the cell taken by confocal microscopy. Arrows, localization of Cav-1 at the degradation sites of the gelatin matrix. B, MDA-MB-231 cells were cultured on gelatin-coated coverslips for 4 h, incubated with CTxB for 10 min, and stained with phalloidin. The submembrane regions of the ventral cell membrane were imaged by confocal microscopy. Insets, magnified images of the boxed regions. Arrows, Cav-1 colocalizing withthe internalized CTxB around invadopodia.

Cancer Res 2009; 69: (22). November 15, 2009 8598 www.aacrjournals.org

Figure 5. Cav-1 is colocalized and cotrafficked with MT1-MMP and necessary for MT1-MMP–dependent gelatin degradation. A, MDA-MB-231 cells were separated into the cytosolic and Triton X-100 soluble and insoluble fractions. The presence of invadopodia components in eachfraction was determined by immunoblotting. Representative of three independent experiments. B, MDA-MB-231 cells cotransfected withmCherry-Cav-1 and GFP-MT1-MMP were cultured on gelatin-coated coverslips for 2 h and stained with phalloidin. The submembrane region of the ventral cell membrane was imaged by confocal microscopy. Insets, magnified images of the boxed regions. C, MDA-MB-231 cells cotransfected withCFP-Cav-1 and mCherry-MT1-MMP were plated onto gelatin-coated coverslips for 2 hand analyzed using time-lapse microscopy. Note that the slight shift of the CFP and mCherry signals is due to consecutive imaging. D, MDA-MB-231 cells (MDA) and cells that stably expressed mCherry-MT1-MMP (mCherry-MT1) were transfected withtheindicated siRNAs for 48 hand tested for gelatin degradation activity for 2h.Columns, mean of at least four independent determinations; bars, SE. *, P < 0.0001, Student's t test.

(Fig. 4A). Accumulation of the signals was observed at invadopodia- observed as small vesicles concentrated around invadopodia and mediated degradation sites of the gelatin matrix (Fig. 4A). Localiza- were colocalized with MT1-MMP signals (Fig. 5B). When analyzed tion of Cav-1 was observed at some but not all invadopodia, using live-cell imaging, these small vesicles showed highly dynamic suggesting that Cav-1 is not a structural component of invadopodia behavior, and CFP-Cav-1 and mCherry-MT1-MMP were cotraffick- but rather has modifying functions required for proper invadopodia ing with them in the cells (Fig. 5C; Supplementary Video 2). formation. Cav-1 was also colocalized with the internalized CTxB To examine the involvement of Cav-1 in MT1-MMP function, signals existing around invadopodia at the submembrane region MDA-MB-231 cells stably expressing mCherry-MT1-MMP were (Fig. 4B). Thus, Cav-1 may be involved in the trafficking of lipid raft generated by retroviral transfection. When these cells were cul- membranes at invadopodia. tured on a fluorescent gelatin matrix, extensive gelatin degradation Cav-1 regulates MT1-MMP–dependent matrix degradation was detected throughout the cell body within 2 h (Fig. 5D; Supple- activity. The invadopodia component that is coenriched with mentary Fig. S5A). At this time point, little gelatin degradation ac- Cav-1 at lipid rafts was then determined. MDA-MB-231 cells cul- tivity was detected in the parental MDA-MB-231 cells (Fig. 5D). tured on gelatin-coated dishes were separated into cytosolic, Tri- Cav-1 knockdown significantly reduced the gelatin degradation ac- ton X-100 soluble, and Triton X-100 insoluble fractions. Lipid rafts tivity of MT1-MMP–expressing cells (Fig. 5D). Interestingly, choles- are known enriched in the Triton X-100 insoluble fraction (34). The terol depletion by MβCD treatment did not significantly affect the presence of major invadopodia components, including MT1-MMP, gelatin degradation by MT1-MMP cells, although this treatment N-WASP, Arf6, cortactin, Cdc42, and Src, in each fraction was ex- efficiently reduced the amount of cellular cholesterol (Supplemen- amined using immunoblotting (Fig. 5A). Among the invadopodia tary Fig. S5B and C). This observation implies that overexpression components examined, MT1-MMP was primarily detected in the of MT1-MMP can overcome the effect of raft disruption. Altogeth- Triton X-100 insoluble fraction, which is enriched with lipid raft er, these results indicate that Cav-1 regulates proper functioning of proteins, Cav-1 and Flot-1. MT1-MMP. Next, the localization of Cav-1 and MT1-MMP in association Cav-1 expression is correlated with the invadopodia activity with invadopodia was examined. The mCherry-Cav-1 and GFP- in human breast cancer cell lines. To further confirm that Cav-1 MT1-MMP constructs were cotransfected into cells and the local- is an essential regulator of invadopodia formation, the correlation ization of these proteins was examined. Strong Cav-1 signals were between Cav-1 expression and invadopodia activity in human www.aacrjournals.org 8599 Cancer Res 2009; 69: (22). November 15, 2009

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/0008-5472.CAN-09-2305 Cancer Research breast cancer cell lines with different invasive phenotypes was de- suppressed the invadopodia-mediated gelatin degradation activity termined. Cell lines tested were tumorigenic but noninvasive of Hs578T and BT549 cells (Fig. 6D). breast cancer cell lines T47D, SK-BR-3, and MDA-MB-453 and invasive breast cancer cell lines, Hs578T, BT549, and MDA-MB-231. When cultured on fluorescent gelatin-coated coverslips, the nonin- Discussion vasive cells showed only minimal formation of invadopodia and This study showed that lipid raft membranes are enriched, inter- degradation of the gelatin matrix (Fig. 6A). In contrast, the invasive nalized, and actively trafficked at invadopodia and the disruption BT549, Hs578T, and MDA-MB-231 cells showed invadopodia for- of lipid rafts suppresses invadopodia formation by human breast mation associated with gelatin degradation (Fig. 6A). The expres- cancer cells. We reported previously that local actin polymeriza- sion of Cav-1 was readily detected only in cells possessing tion mediated by N-WASP and the Arp2/3 complex is required invadopodia activity (Fig. 6B). In contrast, no such correlation for the assembly of invadopodia core structures (23). Two related was observed in Flot-1 (Fig. 6B). The presence of processed active reports showed that lipid rafts are the platforms for localized actin form of MT1-MMP (∼60 kDa) was also detected only in invasive polymerization through the WASP/N-WASP-Arp2/3 complex path- cell lines and is enriched in Triton X-100 insoluble lipid raft frac- way (35, 36). Therefore, lipid rafts may function in the recruitment tions (Fig. 6C). On the other hand, inactive pro-form of MT1-MMP of these actin nucleators, and possibly other invadopodia compo- (∼63 kDa) was detected in all cell lines regardless of their invasive nents, to the sites of invadopodia assembly. Because several inva- capacity (Fig. 6C). Finally, Cav-1 knockdown by siRNA significantly dopodia components, such as dynamin-2 and Arf6, are involved in

Figure 6. Cav-1 expression correlates withinvadopodia formation in humanbreast cancer cell lines. A, human breast cancer cells were cultured on fluorescent gelatin-coated coverslips for 7 hand stained withphalloidin. Arrows, invadopodia-mediated degradation sites of the gelatin. B, expression of Cav-1 and Flot-1 in indicated cell lines was analyzed using reverse transcription-PCR and/or immunoblotting (IB). Cyclophilin (Cycl) and actin were used as internal controls. C, total cell lysates and the cytosolic and Triton X-100 soluble and insoluble fractions were prepared from indicated cell lines. The presence of pro–MT1-MMP (white arrows) and active MT1-MMP (black arrows) in eachfraction was determined by immunoblotting. D, Hs578T and BT549 cells were transfected withtheindicated siRNAs for 48 hand subjected to immunoblot analysis ( left) and invadopodia assay for 7 h( right). Columns, mean of six independent determinations; bars, SE. *, P < 0.004; **, P < 0.002, Student's t test. Representative of at least three independent experiments.

Cancer Res 2009; 69: (22). November 15, 2009 8600 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/0008-5472.CAN-09-2305 Lipid Rafts and Caveolin-1 in Invadopodia Formation raft endocytosis and trafficking (12, 37), these proteins may medi- the intracellular pool of lipid rafts. Recently, two studies showed ate the dynamic behavior of lipid rafts at invadopodia. that MT1-MMP is delivered to invadopodia via targeted exocytosis It has been shown that MT1-MMP resides in lipid rafts (38, 39) by the exocyst complex and via targeted transport regulated by and increased localization of MT1-MMP in lipid rafts enhances in- VAMP7 (46). As both the exocyst complex and VAMP7 are associ- vasion of prostate cancer cells (40). Consistent with these reports, ated with lipid rafts (47, 48), Cav-1 and these molecules may coor- the present study found that MT1-MMP is preferentially localized dinately regulate the MT1-MMP transport in the same lipid raft in lipid raft fractions among major invadopodia components. We compartment. Contradicting our conclusion, it was reported that observed that lipid raft disruption blocked the gelatin degradation overexpression of Cav-1 inhibits the activity of coexpressed MT1- activity of parental MDA-MB-231 cells, whereas it had little effect MMP (49). Because overexpression of Cav-1 compromises lipid on that of the cells overexpressing MT1-MMP. This observation raft-mediated endocytic pathway (50), the strict control of the raises the possibility that endogenous MT1-MMP must be concen- expression level of caveoin-1 may be required for the proper trated at invadopodia by lipid rafts to mediate focal matrix degra- trafficking of lipid raft-associated MT1-MMP. dation, whereas overexpressed MT1-MMP can exert its activity Caldieri and colleagues reported that Cav-1 regulates invadopo- without lipid raft-mediated concentration. This idea is supported dia formation by modulating the balance of membrane cholesterol by the fact that the cells overexpressing MT1-MMP degrade the levels (33). In this study, we showed that the inhibition of invado- gelatin matrix throughout the cell body irrespective of the pres- podia formation by Cav-1 knockdown could not be canceled ence of invadopodia. by addition of cholesterol in breast cancer cells. Additionally, In the present study, we showed that Cav-1 accumulates at in- Cav-1 knockdown, but not MβCD treatment, blocked MT1-MMP– vadopodia and its knockdown inhibits invadopodia formation. It dependent gelatin degradation activity. These results indicate that was also shown that Cav-1 is expressed in invasive human breast Cav-1 regulates the MT1-MMP function and invadopodia formation cancer cell lines that possess an ability to form invadopodia. Sup- independently of the balance of membrane cholesterol. Therefore, porting our observation, Cav-1 knockdown was reported to inhibit the present results imply that Cav-1 plays multiple roles in the traf- invadopodia formation in human melanoma cells (33). In breast ficking of invadopodia components and that the roles of Cav-1 in cancer, Cav-1 is expressed in invasive and metastatic tumors and cellular invasion differ to some extent among cancer cell types. associated with poor patient prognosis (41–43). Furthermore, Joshi In conclusion, the results of this study provide evidence that lip- and colleagues recently reported that Cav-1 regulates invasion of id raft formation and Cav-1 are necessary for invadopodia func- human breast cancer cells (22). Therefore, our findings may pro- tioning in human breast cancer cells. These findings provide new vide a novel insight into the molecular mechanisms by which insight into the molecular mechanisms of invadopodia formation, Cav-1 regulates the ECM remodeling associated with breast cancer which will contribute to the development of new therapeutic treat- cell invasion. ments for cancer invasion and metastasis. Several studies have reported that MT1-MMP is internalized, trafficked, and recycled to the cell surface through the caveolin- dependent pathway (38, 39). Additionally, Cav-1 has been shown Disclosure of Potential Conflicts of Interest to associate with MT1-MMP (44). We showed, for the first time, No potential conflicts of interest were disclosed. that MT1-MMP–containing vesicles are cotrafficked with Cav-1. Moreover, Cav-1 was coexpressed with active MT1-MMP in invasive breast cancer cell lines. These results suggest that Cav-1 reg- Acknowledgments ulates transport of lipid raft-associated MT1-MMP for functioning Received 6/26/09; revised 8/24/09; accepted 9/9/09; published OnlineFirst 11/3/09. of invadopodia. We also found that Cav-1 knockdown, but not Grant support: Grant-in-Aid for Scientific Research (B) and for Young Scientists MβCD treatment, affects gelatin degradation activity of the cells (B), The Uehara Memorial Foundation, and The Kao Foundation for Arts and Sciences. overexpressing MT1-MMP. Consistent with this observation, it The costs of publication of this article were defrayed in part by the payment of page was reported that MβCD treatment does not disrupt the associa- charges. This article must therefore be hereby marked advertisement in accordance tion of Cav-1 with MT1-MMP (44). As MβCD preferentially de- with 18 U.S.C. Section 1734 solely to indicate this fact. We thank Dr. Toshio Kitamura for providing the Plat-A cells and pMXs vectors and pletes cholesterol on the plasma membrane (45), it is possible Masahiro Kawamura, Yumiko Konko, Keiko Takayama, and Emi Muroi for technical that the regulation of MT1-MMP by Cav-1 occurs primarily in assistance.

References 5. Yamaguchi H, Condeelis J. Regulation of the actin cy- 9. Jacobson K, Mouritsen OG, Anderson RG. Lipid rafts: toskeleton in cancer cell migration and invasion. Bio- at a crossroad between cell biology and physics. Nat 1. Buccione R, Caldieri G, Ayala I. Invadopodia: special- chim Biophys Acta 2007;1773:642–52. Cell Biol 2007;9:7–14. ized tumor cell structures for the focal degradation of 6. Artym VV, Zhang Y, Seillier-Moiseiwitsch F, Yamada 10. Lajoie P, Goetz JG, Dennis JW, Nabi IR. Lattices, the extracellular matrix. Cancer Metastasis Rev 2009;28: KM, Mueller SC. Dynamic interactions of cortactin rafts, and scaffolds: domain regulation of receptor 137–49. and membrane type 1 matrix metalloproteinase at inva- signaling at the plasma membrane. J Cell Biol 2009; 2. Weaver AM. Invadopodia: specialized cell structures dopodia: defining the stages of invadopodia formation 185:381–5. – for cancer invasion. Clin ExpMetastasis 2006;23:97 105. and function. Cancer Res 2006;66:3034–43. 11. Simons K, Toomre D. Lipid rafts and signal transduc- 3. Yamaguchi H, Wyckoff J, Condeelis J. Cell migration in 7. Itoh Y, Seiki M. MT1-MMP: a potent modifier of peri- tion. Nat Rev Mol Cell Biol 2000;1:31–9. tumors. Curr Opin Cell Biol 2005;17:559–64. cellular microenvironment. J Cell Physiol 2006;206:1–8. 12. Hanzal-Bayer MF, Hancock JF. Lipid rafts and mem- 4. Gimona M, Buccione R, Courtneidge SA, Linder S. As- 8. Linder S. The matrix corroded: podosomes and inva- brane traffic. FEBS Lett 2007;581:2098–104. sembly and biological role of podosomes and invadopo- dopodia in extracellular matrix degradation. Trends Cell 13. Michel V, Bakovic M. Lipid rafts in health and dis- dia. Curr Opin Cell Biol 2008;20:235–41. Biol 2007;17:107–17. ease. Biol Cell 2007;99:129–40. www.aacrjournals.org 8601 Cancer Res 2009; 69: (22). November 15, 2009

14. Patra SK. Dissecting lipid raft facilitated cell signal- 27. Zacharias DA, Violin JD, Newton AC, Tsien RY. Parti- loproteinase traffic in human endothelial cells. Mol Biol ing pathways in cancer. Biochim Biophys Acta 2008; tioning of lipid-modified monomeric GFPs into mem- Cell 2004;15:678–87. 1785:182–206. brane microdomains of live cells. Science 2002;296:913–6. 40. Wang X, Wilson M, Slaton JW, Sinha AA, Ewing SW, 15. Parton RG, Hanzal-Bayer M, Hancock JF. Biogenesis 28. Zidovetzki R, Levitan I. Use of cyclodextrins to ma- Pei D. Increased aggressiveness of human prostate PC-3 of caveolae: a structural model for caveolin-induced do- nipulate plasma membrane cholesterol content: evi- tumor cells expressing cell surface localized membrane main formation. J Cell Sci 2006;119:787–96. dence, misconceptions and control strategies. Biochim type-1 matrix metalloproteinase (MT1-MMP). J Androl 16. Quest AF, Leyton L, Parraga M. Caveolins, caveolae, Biophys Acta 2007;1768:1311–24. 2009;30:259–74. and lipid rafts in cellular transport, signaling, and dis- 29. Rothberg KG, Heuser JE, Donzell WC, Ying YS, 41. Savage K, Lambros MB, Robertson D, et al. Caveolin 1 ease. Biochem Cell Biol 2004;82:129–44. Glenney JR, Anderson RG. Caveolin, a protein compo- is overexpressed and amplified in a subset of basal-like 17. Williams TM, Lisanti MP. The caveolin proteins. Ge- nent of caveolae membrane coats. Cell 1992;68:673–82. and metaplastic breast carcinomas: a morphologic, ul- nome Biol 2004;5:214. 30. Nabi IR, Le PU. Caveolae/raft-dependent endocytosis. trastructural, immunohistochemical, and in situ hybrid- – 18. Grande-Garcia A, del Pozo MA. Caveolin-1 in cell po- J Cell Biol 2003;161:673–7. ization analysis. Clin Cancer Res 2007;13:90 101. larization and directional migration. Eur J Cell Biol 31. Babuke T, Tikkanen R. Dissecting the molecular 42. Yang G, Truong LD, Timme TL, et al. Elevated ex- 2008;87:641–7. function of reggie/flotillin proteins. Eur J Cell Biol pressionofcaveolinisassociatedwithprostateand – 19. Salanueva IJ, Cerezo A, Guadamillas MC, del Pozo 2007;86:525–32. breast cancer. Clin Cancer Res 1998;4:1873 80. MA. Integrin regulation of caveolin function. J Cell 32. Vassilieva EV, Ivanov AI, Nusrat A. Flotillin-1 stabi- 43. Van den Eynden GG, Van Laere SJ, Van der Auwera I, Mol Med 2007;11:969–80. lizes caveolin-1 in intestinal epithelial cells. Biochem et al. Overexpression of caveolin-1 and -2 in cell lines 20. BurgermeisterE,LiscovitchM,RockenC,Schmid Biophys Res Commun 2009;379:460–5. and in human samples of inflammatory breast cancer. – RM, Ebert MP. Caveats of caveolin-1 in cancer progres- 33. Caldieri G, Giacchetti G, Beznoussenko G, Attanasio Breast Cancer Res Treat 2006;95:219 28. sion. Cancer Lett 2008;268:187–201. F, Ayala I, Buccione R. Invadopodia biogenesis is regu- 44. LabrecqueL,NyalendoC,LangloisS,etal.Src- 21. Goetz JG, Lajoie P, Wiseman SM, Nabi IR. Caveolin-1 lated by caveolin-mediated modulation of membrane mediated tyrosine phosphorylation of caveolin-1 in- in tumor progression: the good, the bad and the ugly. cholesterol levels. J Cell Mol Med 2008;10.1111/j.1582– duces its association with membrane type 1 matrix – Cancer Metastasis Rev 2008;27:715–35. 4934.2008.00568.x. metalloproteinase. J Biol Chem 2004;279:52132 40. 22. Joshi B, Strugnell SS, Goetz JG, et al. Phosphorylated 34. Pike LJ. Lipid rafts: heterogeneity on the high seas. 45. Keller P, Simons K. Cholesterol is required for surface caveolin-1 regulates Rho/ROCK-dependent focal adhe- Biochem J 2004;378:281–92. transport of influenza virus hemagglutinin. J Cell Biol 1998;140:1357–67. sion dynamics and tumor cell migration and invasion. 35. Golub T, Caroni P. PI(4,5)P2-dependent microdomain Cancer Res 2008;68:8210–20. assemblies capture microtubules to promote and con- 46. Poincloux R, Lizarraga F, Chavrier P. Matrix invasion 23. Yamaguchi H, Lorenz M, Kempiak S, et al. Molecular trol leading edge motility. J Cell Biol 2005;169:151–65. by tumour cells: a focus on MT1-MMP trafficking to in- – mechanisms of invadopodium formation: the role of the 36. Rozelle AL, Machesky LM, Yamamoto M, et al. Phos- vadopodia. J Cell Sci 2009;122:3015 24. N-WASP-Arp2/3 complex pathway and cofilin. J Cell phatidylinositol 4,5-bisphosphate induces actin-based 47. Inoue M, Chiang SH, Chang L, Chen XW, Saltiel AR. Biol 2005;168:441–52. movement of raft-enriched vesicles through WASP- Compartmentalization of the exocyst complex in lipid 24. Bowden ET, Coopman PJ, Mueller SC. Invadopodia: Arp2/3. Curr Biol 2000;10:311–20. rafts controls Glut4 vesicle tethering. Mol Biol Cell – unique methods for measurement of extracellular ma- 37. Balasubramanian N, Scott DW, Castle JD, Casanova 2006;17:2303 11. trix degradation in vitro. Methods Cell Biol 2001;63: JE, Schwartz MA. Arf6 and microtubules in adhesion- 48. Lafont F, Verkade P, Galli T, Wimmer C, Louvard D, 613–27. dependent trafficking of lipid rafts. Nat Cell Biol 2007; Simons K. Raft association of SNAP receptors acting in 25. Kitamura T, Koshino Y, Shibata F, et al. Retrovirus- 9:1381–91. apical trafficking in Madin-Darby canine kidney cells. – mediated gene transfer and expression cloning: power- 38. Annabi B, Lachambre M, Bousquet-Gagnon N, Page Proc Natl Acad Sci U S A 1999;96:3734 8. ful tools in functional genomics. ExpHematol 2003;31: M, Gingras D, Beliveau R. Localization of membrane- 49. Kim HN, Chung HS. Caveolin-1 inhibits membrane- 1007–14. type 1 matrix metalloproteinase in caveolae membrane type 1 matrix metalloproteinase activity. BMB Rep 2008; – 26. Chinnapen DJ, Chinnapen H, Saslowsky D, Lencer domains. Biochem J 2001;353:547–53. 41:858 62. WI. Rafting with cholera toxin: endocytosis and traffick- 39. Galvez BG, Matias-Roman S, Yanez-Mo M, Vicente- 50. Le PU, Guay G, Altschuler Y, Nabi IR. Caveolin-1 is a ing from plasma membrane to ER. FEMS Microbiol Lett Manzanares M, Sanchez-Madrid F, Arroyo AG. Caveolae negative regulator of caveolae-mediated endocytosis to 2007;266:129–37. are a novel pathway for membrane-type 1 matrix metal- the endoplasmic reticulum. J Biol Chem 2002;277:3371–9.

Cancer Res 2009; 69: (22). November 15, 2009 8602 www.aacrjournals.org

Lipid Rafts and Caveolin-1 Are Required for Invadopodia Formation and Extracellular Matrix Degradation by Human Breast Cancer Cells

Hideki Yamaguchi, Yukiko Takeo, Shuhei Yoshida, et al.

Cancer Res 2009;69:8594-8602. Published OnlineFirst November 3, 2009.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-09-2305

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2009/11/03/0008-5472.CAN-09-2305.DC1

Cited articles This article cites 49 articles, 17 of which you can access for free at: http://cancerres.aacrjournals.org/content/69/22/8594.full#ref-list-1

Citing articles This article has been cited by 25 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/69/22/8594.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/69/22/8594. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.