soluble mmp-14 produced by bone marrow-derived

© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Carcinogenesis vol.35 no.8 pp.1770–1779, 2014doi:10.1093/carcin/bgu061Advance Access publication March 11, 2014

Soluble MMP-14 produced by bone marrow-derived stromal cells sheds epithelial endoglin modulating the migratory properties of human breast cancer cells

Nicolás Tobar, M.Celeste Avalos, Nicolás Méndez, Patricio C.Smith1, Carmelo Bernabeu2,3, Miguel Quintanilla4 and Jorge Martínez*

Laboratorio de Biología Celular, INTA, Universidad de Chile, Santiago 7830490, Chile, 1Laboratorio de Fisiología Periodontal, Facultad de Medicina, Pontificia Universidad Católica de Chile, Santiago 8331150, Chile, 2Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Científicas (CSIC), Madrid, Spain, 3Centro de Investigación Biomédica en Red de Enfermedades Raras (CIBERER), 28040 Madrid, Spain and 4Instituto de Investigaciones Biomédicas Alberto Sols, CSIC, 28029 Madrid, Spain

*To whom correspondence should be addressed. Tel: +56 2 9781419; Fax: +56 2 2214030; Email: [email protected]

It has been proposed that epithelial cells can acquire invasive properties through exposure to paracrine signals originated from mesenchymal cells within the tumor microenvironment. Transforming growth factor-β (TGF-β) has been revealed as an active factor that mediates the epithelial–stroma cross-talk that facilitates cell invasion and metastasis. TGF-β signaling is mod-ulated by the coreceptor Endoglin (Eng), which shows a tumor suppressor activity in epithelial cells and regulates the ALK1-Smad1,5,8 as well as the ALK5-Smad2,3 signaling pathways. In the current work, we present evidence showing that cell surface Eng abundance in epithelial MCF-7 breast cancer cells is inversely related with cell motility. Shedding of Eng in MCF-7 cell surface by soluble matrix metalloproteinase-14 (MMP-14) derived from the HS-5 bone-marrow-derived cell line induces a motile epithe-lial phenotype. On the other hand, restoration of full-length Eng expression blocks the stromal stimulus on migration. Processing of surface Eng by stromal factors was demonstrated by biotin-neutravidin labeling of cell surface proteins and this processing generated a shift in TGF-β signaling through the activation of Smad2,3 pathway. Stromal MMP-14 abundance was stimulated by TGF-β secreted by MCF-7 cells acting in a paracrine manner. In turn, the stromal proteolytic activity of soluble MMP-14, by inducing Eng shedding, promoted malignant progression. From these data, and due to the capacity of TGF-β to regulate malig-nancy in epithelial cancer, we propose that stromal-dependent epithelial Eng shedding constitutes a putative mechanism that exerts an environmental control of cell malignancy.

Introduction

There is strong evidence supporting the participation of stromal cells in carcinoma progression (1). Particularly, it has been demonstrated that cellular stromal components accomplish a crucial role in breast cancer growth and development (2). It is well known that the stroma and epithelium maintain an active dialog that produces functional reciprocal consequences. Thus, stromal soluble factors can modify the invasive potential of carcinoma cells and epithelial factors are respon-sible for the stromal desmoplastic response that characterizes some type of tumors (3). Carcinoma-associated fibroblasts have been iden-tified as an established source of tumor and angiogenesis-promoting factors with a yet unraveled role in tumorigenesis (4). The origin of these carcinoma-associated fibroblasts is still a matter of debate, but

there is no doubt that they exert an important influence in the invasive properties of carcinoma cells in either the primary or the metastatic niches (5). Homing and further proliferation of tumoral cells in the metastatic site is also strongly affected by the secondary stroma and its ability to provide signals that favor metastatic colonization as it occurs in breast metastasis to bone (6).

Endoglin (Eng; CD105) is an integral membrane-bound glycopro-tein, which acts as a transforming growth factor-β (TGF-β) auxiliary receptor (7). Eng is highly expressed in endothelial cells and it has been proposed to play a key role during tumor neoangiogenesis (8). Besides the well-recognized proangiogenic role of endothelial Eng, it has also been demonstrated that epithelial Eng is a crucial player in the carcinogenic phenomenon (9). Current evidence suggests that the main role of Eng in epithelial carcinogenesis is acting as a tumor sup-pressor molecule. Thus, a suppressor role for Eng has been postulated in esophageal, prostate and breast carcinomas (10–12), although in the latter type of cancer some controversy still remains, since a report shows that overexpression of Eng in MDA MB-231 breast cell line enhances TGF-β-induced chemotaxis/invasion through matrigel, sug-gesting a proinvasive role for Eng in breast cancer (13). Conversely, we have shown previously that knockdown of Eng in murine-trans-formed keratinocytes stimulates basal and TGF-β-mediated ALK5/Smad2/3 signaling activity, a route that has been proposed to play a central role in epithelial–mesenchymal transition and the stimulation of invasiveness (14).

To date, data linking the development of metastatic disease with the expression of Eng are mostly limited to the role of Eng in tumor angi-ogenesis (15). However, an interesting and complementary research field has emerged with the identification in breast and colorectal can-cer of elevated levels of a soluble form of Eng associated to a poor prognosis (16,17). In search of the mechanisms by which Eng modu-lates cell motility, it has been demonstrated that overexpression of Eng in mouse fibroblasts lead to decreased migration in chemotactic and wound-healing assays, as well as changes in the cellular morphol-ogy (18). Using a genetic model of prostate cancer that combined Eng haploinsufficiency with the TRAMP (transgenic adenocarcinoma mouse prostate) mouse, Romero et al. (19) have suggested that Eng is required for multiple aspects of carcinoma-associated fibroblast function including viability, endothelial cell recruitment and tumor-induced migration.

The existence of a soluble form of Eng in plasma of breast cancer patients has been associated with the development of a metastatic dis-ease (17). Also, in another study that evaluated serum Eng in patients with a variety of tumors, increased serum levels of Eng were associ-ated with enhanced neovascularization of metastatic tumors (20). In these cases, it is possible to assume that the higher concentration of serum Eng was the consequence of a proteolytic processing of mem-brane-associated Eng, which seems to suggest that membrane shed-ding is a plausible possibility to control Eng protein levels in tumor cells. In fact, using a colorectal carcinoma model, it has been dem-onstrated that membrane-bound matrix metalloproteinase-14 (MMP-14; MT1-MMP) present in the surface of colorectal carcinoma cells cleaves Eng in close proximity to the transmembrane domain (21).

Membrane-bound receptor shedding has been proposed as a func-tional mechanism to regulate growth factor signaling and usually it involves the proteolytic cleavage of the receptors’ extracellular domain (22). A similar phenomenon occurs with betaglycan, also a coreceptor for TGF-β, which after the proteolytic processing by MMP-14, generates a soluble form that competes with membrane-bound betaglycan for TGF-β binding, and in this manner, inhibits the cellular effect of the factor (23).

The aim of this study was to analyze mechanisms that mediate the stromal-induced shedding of Eng in a human breast cancer cell line.

Abbreviations: CM, conditioned medium; DMEM, Dulbecco’s modified Eagle’s medium; Eng, endoglin; MMP, matrix metalloproteinase; mRNA, messenger RNA; PBS, phosphate-buffered saline; SDS, sodium dodecyl sul-fate; TGF-β, transforming growth factor-β.

1770Downloaded from https://academic.oup.com/carcin/article-abstract/35/8/1770/318046by gueston 30 March 2018

mailto:[email protected]?subject=

Soluble MMP-14 sheds epithelial endoglin

We describe for the first time that a stromal proteolytic activity, identi-fied as a soluble form of MMP-14 present in the conditioned medium (CM) derived from a bone marrow-derived cell line (HS-5) (24) cleaves membrane-bound Eng present in the weakly invasive MCF-7 cell line. We found that this paracrine Eng shedding process gives rise to an increase in the migratory potential of tumor cells that is reversed by the overexpression of membrane-bound Eng. In addition, we also studied the influence of epithelial-derived TGF-β1 on stromal MMP-14 expression. Our data suggest that stromal-derived MMP-14 is responsible for epithelial Eng shedding that, in turn, favors the expression of a more motile epithelial phenotype.

Materials and methods

Cell culture, cell lines and chemicalsThe human cell lines mammary breast MCF-7 and bone marrow-derived HS-5, obtained from the ATCC (Manassas, VA) were cultured in Dulbecco’s modified Eagle’s medium (DMEM)/F12 and DMEM (Invitrogen Carlsbad, CA), respec-tively, supplemented with 10% fetal bovine serum (Hyclone, Logan, UT) and maintained in a humidified atmosphere of 37°C, 5% CO2. Human recombinant TGFβ1 and BMP-2 were purchased from R&D Systems (Minneapolis, MN). SB431542 inhibitor was from TOCRIS Bioscience (Business Park Ellisville, MO). GM6001 (Ilomastat) was purchased from Chemicon (Temecula, CA). Protein A Sepharose CL-4B was from Sigma (St Louis, MO). Sulfo-NHS-LC-Biotin and NeutraAvidin Sepharose were from Thermo (Waltham, MA). CM were obtained by culturing confluent HS-5 and MCF-7 cells for 48 h, in serum-free media (1 ml per 10 cm2 of culture area). CM was clarified by cen-trifugation and used freshly.

Western blot and antibodiesAfter treatments, cells were resuspended in lysis buffer (50 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, pH 7.4, 150 mM NaCl, 2 mM MgCl2, 2 mM ethyleneglycol-bis(aminoethylether)-tetraacetic acid, 1% Triton X-100 and 10% glycerol) supplemented with Complete protease inhibitors (Roche, Mannheim, Germany). Protein concentration of cell lysates was determined by a Bradford method (Fermentas, Maryland, MD). Protein extracts were heat denatured in sodium dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis loading buffer 4X (240 mM Tris–HCl, pH 6.8, 8% SDS, 40% glycerol and 20% 2-mercaptoethanol). Equal amounts of protein from different treatments were resolved by SDS–polyacrylamide gel electrophore-sis in 8% acrylamide gels and electrotransferred to polyvinylidene difluoride membranes using a buffer containing 24 mM Tris, 194 mM glycine and 20% methanol. Proteins were further analyzed using the ECL chemiluminescence detection kit (Amersham, Arlington Heights, FL). The immunoreactions were achieved by incubation of the membranes, previously blocked with a solution containing 5% bovine serum albumin in Tris-buffered saline and 0.05% Tween 20 (Sigma), with the following antibodies: rabbit anti human Eng (ab49228) and rabbit anti phospho Smad3 (ab52903) (Abcam, Cambridge, MA), mouse anti human Eng (M3527; DAKO, Glostrup, Denmark), rabbit anti phospho Smad2,3 (sc-11769), goat anti Smad2,3 (sc-6032), mouse anti Smad1 (sc-7965) and rabbit anti c-Myc (sc-789) were from Santa Cruz (Santa Cruz, CA), rabbit anti p-Smad 1,5,8 (#9511; Cell Signaling, Danvers, MA), mouse anti EGF-R (25), mouse anti MMP-14 (MAB918; R&D Systems), mouse anti MMP-2 (Ab-3; Calbiochem, Cambridge MA), rabbit anti human Smad1 (PA5-19748) and mouse anti Smad3 (MA5-15663) were from Thermo scien-tific (Rockford, IL) and mouse anti beta actin (A5441; Sigma). Identifications of proteins in soluble fractions were performed as above in concentrated (10×) CM using Vivaspin concentrators tubes (GE Healthcare, Hatfield, UK). Densitometric analysis of western blot bands was performed using Molecular Imaging Software, version 4.0 of Kodak (Rochester, NY).

Cell migration assayMCF-7 cell migration was studied using a 6.5 mm Transwell chamber with a pore size of 8 μm (Corning, Corning, NY). The transwell membranes were coated with 10% fetal bovine serum in culture media for 2 h at 37°C only on the lower chamber. Media conditioned by HS-5 cells (CM-HS-5) was prepared by incubating 5 × 104 cells per cm2 in a total volume of 8 ml of serum-free cul-ture media for 48 h. When required, MCF-7 cells were pretreated with a 50% (vol/vol in DMEM) solution of CM-HS-5 for 24 h. Control and MC-HS-5-pretreated MCF-7 cells (6 × 104) were resuspended in serum-free medium and seeded on the upper compartment of the chamber. In migration experiments in which MCF-7 cells were stimulated with 10 ng/ml of TGF-β1 and/or 20 ng/ml of BMP-2, growth factors were added in the upper compartment of the tran-swell and cells were allowed to migrate against 5% fetal bovine serum in the lower compartment. Migration was allowed to occur for 24 h, after which cells

of the upper membrane surface were removed by a cotton swab and repeat-edly washed with phosphate-buffered saline (PBS). For migration experiments in which MCF-7 Eng-overexpressing cells were used, control and CM-HS-5 pretreated cells were transiently transfected with a pSuper-C vector containing a wild-type version of the human Eng gene tagged with myc (26) using the TransIT 2020 reagent (Mirus, Madison, WI) according to the manufacturer’s instructions. A group of cells were pretreated with 10 μM GM6001 (Ilomastat), a broad-range inhibitor of metalloproteinases (MMPs), prior to the migration assay, as indicated. In experiments in which the TGF-β-dependent migratory capacity was evaluated, MCF-7 cells were transfected with expression vec-tors encoding HA-tagged wild-type, constitutively active ALK5 (T204D) and kinase-deficient ALK5 (K232R) kindly provided by Dr L.Attisano (University of Toronto, Toronto, Canada) (27). We also used an expression vector encoding constitutively active ALK1 (ALK1QD) kindly provided by Dr K.Miyazono (School of Medicine, The University of Tokyo, Tokyo, Japan). Experiments in which cell migration was evaluated after knocking down Smad3 and Smad1 were performed using a commercial pool of small interfering RNAs directed to both molecules (Santa Cruz, sc-37238 and sc-63289, respectively, and sc-37007 as a control) using the TransIT-siQUEST transfection reagent (Mirus) according to the manufacturer’s instructions. Migration values were determined by counting five fields (20×) per chamber after fixing and staining the migratory cells on the lower side of the membrane with 0.2% crystal violet in 10% methanol for 5 min (28).

Immunofluorescence microscopyMCF-7 cells growing on coverslips were incubated with 50% CM HS-5 for 48 h. Then, cells were treated with 4% paraformaldehyde for 30 min at room temperature. Fixed cells were blocked with gelatin 0.2% in PBS for 1 h at room temperature and then incubated with mouse anti-human Eng antibody (M3527; DAKO) at 1:50 dilution for 16 h at 4°C. Once washed, cells were incubated with Alexa Fluor 594 goat anti-mouse (A-11005; Life Technologies, Carlsbad, CA) at 1:2000 dilution as a secondary antibody, washed and mounted with fluorescence mounting medium (DAKO). Confocal images of cells were acquired using the Confocal Laser Scanning Microscope-510 META (Zeiss, Oberkochen, Germany) and processed using LSM Image Browser software (Zeiss) and Adobe Photoshop 7.0. All images were taken using objective Plan-Apochromat 63×/1.4 Oil at 1024 pixel resolution.

Reporter assaysReporter assays using TGF-β-responsive promoters were performed as described previously (29), using p(CAGA)12-luc (30) and pId-lux (31) con-structs, in the presence or the absence of TGF-β1, as indicated. All tran-sient transfection experiments were carried out using TransIT 2020 reagent (Mirus) as vehicle for plasmids. Relative luciferase units were determined in a Berthold FB12 luminometer. The pRL-tk Renilla-normalized luciferase vector (Promega, Madison, WI) served as internal control to correct for transfection efficiency and for normalization.

Biotinylation of cell surface EngMCF-7 cells (108 per point) were incubated with increasing concentrations of MC-HS-5 for 48 h. After treatment, cells were detached with ethylenediami-netetraacetic acid (0.53 mM) and the cell suspension was washed two times with ice-cold PBS pH 7.4 and each time centrifuged at 2000 r.p.m. at 4°C. Sulfo-NHS-LC-Biotin (Thermo) was dissolved in PBS pH 8.0 (final concen-tration 2 mg/ml) and incubated with cell suspension for 5 h at 4°C with rota-tion. Unreacted biotinylating reagent was blocked two times for 15 min with an equal volume of 10 mM of glycine in PBS. Cells pellet was lysed in a buffer containing 50 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid pH 7.4, 150 mM NaCl, 2 mM MgCl2, 2 mM ethyleneglycol-bis(aminoethylether)-tetraacetic acid, 1% Triton X-100 and 10% glycerol, supplemented with com-plete protease inhibitors (Roche), at 4°C. Next, the solution was sonicated twice for 5 s and cleared by centrifugation at 13 000 for 10 min. Cleared supernatants were incubated overnight with 250 μl of NeutraAvidin Sepharose (Thermo) at 4°C. NeutraAvidin-bound proteins were separated by centrifugation at 2500 r.p.m. for 5 min and pellets were washed four times with PBS. Then, pellets were resuspended with 100 μl of sample buffer 4× (240 mM Tris–HCl, pH 6.8, 8% SDS, 40% glycerol and 20% 2-mercaptoethanol) boiled for 10 min and the supernatant was analyzed by SDS–polyacrylamide gel electrophoresis.

MMP-14 immunoprecipitationTo assess the role of stromal MMP-14 in the proteolytic capacity of CM-HS-5, MMP-14 was immunoprecipitated out from the HS-5 CM. Briefly, 8 μg of anti-human MMP-14 (MAB 918; R&D Systems) were incubated overnight with 200 μl of previously hydrated Protein A Sepharose CL-4B (Sigma). Then, 1.0 ml of a solution of 50% CM (in DMEM) was subjected to two rounds of immunoprecipitation with the matrix-bound antibody. The resulting superna-tant was used to stimulate MCF-7 migration as described above.


N.Tobar et al.

Quantitative PCRTotal RNA was isolated with Trizol (GIBCO) from CM MCF-7 and TGF-β1-treated HS-5 cells according to the manufacturer’s instructions. Reverse transcription to complementary DNA was performed using 1 μg of RNA from each sample using M-MLV reverse transcriptase and oligo (dTs) (Promega) as primer, according to the manufacturer’s protocol. MMP-14 messenger RNA (mRNA) expression was assessed by real-time PCR using a Light Cycler instrument (Roche, Mannheim, Germany). The reaction was performed using 100 ng of complementary DNA and LightCycler® FastStart DNA Master SYBR Green I kit (Roche) in a final volume of 20 μl. All the reac-tions were performed in duplicate and negative controls were included. The primers used were: MMP14F (forward), 5′ CCTG CGTCCATC AACACT 3′, MMP14R (reverse), 5′ GCCCATGAATGACCCTCT 3′; GPDHF (for-ward), 5′ CAAAATCA AGTGGGGCGATGC TG 3′, GPDHR (reverse), 5′ TGTGGTCA TGAGTCCTTCCACGAT 3′. The level of MMP-14 expression was normalized using glyceraldehyde 3-phosphate dehydrogenase as loading control.

Statistical analysisExperimental data are presented as mean ± SD of experiments performed in triplicate. Data were subjected to a variance analysis, and comparison among groups was performed by Kruskal–Wallis/Dunn’s multiple comparison test and Mann–Whitney unpaired t-test. In both cases the GraphPad Prism 5 soft-ware (GraphPad Software, San Diego, CA) was used. A value of P < 0.05 was considered as a statistically significant difference.

Results

Media conditioned by HS-5 cell line stimulates the migratory capacity of MCF-7 cells by diminishing the abundance of EngBone marrow constitutes one of the main metastatic sites for breast cancer cells (32). Therefore, we hypothesized that soluble factors pro-duced by a bone marrow-derived cell line could modulate Eng expres-sion and by this means, the migratory capacity of a breast cancer cell line. To evaluate this hypothesis, breast MCF-7 cells were pretreated with media conditioned by the bone marrow-derived HS-5 stromal cell line (CM-HS-5) and their migratory capacity was measured using a transwell assay. In these experiments, CM-HS-5 was not present during the migration assays. We also measured the expression of Eng in control and pretreated MCF-7 cells by western blot. Figure 1A shows that MCF-7 cells previously exposed to CM-HS-5 exhibit an enhanced migratory behavior. These cells also expressed a dimin-ished amount of Eng. To verify that indeed the increase in migra-tion was a consequence of a decrease in Eng expression, we subjected MCF-7 cells overexpressing a Myc-tagged Eng to the stimulus of CM-HS-5. Figure 1B shows that the forced expression of Eng causes the reversion of the CM HS-5-dependent enhancement of migration, highlighting the role of Eng in epithelial migration. Although migra-tory values of Eng-overexpressing cells are apparently stimulated by CM, we found no statistically significant differences between basal and CM-stimulated migration in the presence of Eng overexpression (Figure 1B).

To confirm that CM HS-5 treatment diminished the abundance of Eng in MCF-7 cells, the surface-associated Eng was analyzed in con-trol and CM HS-5-treated cells by confocal microscopy. Figure 1C shows that treated cells display a significantly reduced amount of immunoreactive Eng compared with control cells. To reinforce the hypothesis that decrease of Eng constitutes a more general mecha-nism to control cell motility, we performed the same migration experiment using a weakly invasive MCF-10 cell line and an inva-sion experiment using the strongly invasive MDA MB-231 cell line. In this case, cells migrate through a filter coated with 10 μg/ml of matrigel. In both cases, irrespective of the basal Eng expression, CM HS-5 increased migratory (or invasive) behavior and decreased the abundance of Eng (Figure 1D and E). With the purpose to evaluate whether CM HS-5 does not affect Eng transcriptional expression, we subjected MCF-7 to the stimulus of increasing concentrations of CM and evaluated mRNA expression by quantitative PCR. We found that the stromal stimulus showed a discrete effect on Eng expres-sion at the transcriptional level that does not follow a dose–response pattern. These results rule out the possibility of a transcriptional

regulation of Eng by the CM-HS-5 (Supplementary Figure S1, avail-able at Carcinogenesis Online).

Pretreatment of MCF-7 cells with CM HS-5 enhanced Smad2/3 and inhibited Smad1,5,8 signalingTo investigate whether the decrease of Eng by CM HS-5 pretreat-ment has a consequence on TGF-β-dependent Smad signaling, we analyzed the phosphorylation pattern of Smad2,3 and Smad 1,5,8 in intact and pretreated cells. Figure 2A shows that CM HS-5-pretreated MCF-7 cells displayed a diminished activation pattern of Smad1,5,8 that is more evident during the first 20 min of activation. In contrast, the same cells showed increased phosphorylation levels of Smad2,3, even at time zero and maintained this activation for 30 min after stim-ulation. These results suggest that the decreased amount of Eng in CM-HS-5-pretreated cells provoke a shift in TGF-β signaling enhanc-ing the activation of the canonical Smad2,3 pathway. To confirm the inhibition of ALK1/Smad1,5,8 pathway by CM-HS-5 pretreatment, we performed the same experiment using BMP2, the primary ligand of this route (33). As Figure 2B shows, pretreatment of MCF-7 cells with CM-HS-5 caused an evident inhibition of BMP2-induced phos-phorylation of pSmad1,5,8.

To assess whether this shift was consistent with the TGF-β-1-dependent transcriptional activity, intact and CM-HS-5-pretreated cells were used to activate pCAGA (30) and pId-lux (31) reporter constructs which are specific for Smad3 and Smad1,5,8, respectively. We found that both basal and TGF-β1-stimulated pCAGA reporter activation was significantly stimulated in CM-HS-5-pretreated cells (Figure 2C). On the other hand, pId-lux reporter responded poorly to TGF-β-1 stimulus in both experimental conditions (Figure 2D). However, when cells were stimulated with BMP-2, we observed a robust reporter activity which was significantly decreased in pretreated cells, supporting the view that the Smad1,5,8 pathway is inhibited by CM-HS-5. To test the differential involvement of both signaling routes in the acquisition of migratory capacity, we performed TGF-β1-stimulated migration experiments using MCF-7 cells that over-express either a wild-type version of ALK5 or a dominant-negative mutant of the receptor. Also, to assess the influence of Smad1,5,8 signaling pathway, we expressed a constitutively active mutant of ALK1. As shown in Figure 2E, overexpression of ALK5 results in an enhanced migratory activity that is strongly counteracted by the expression of the dominant-negative version of the same receptor. In contrast, overexpression of constitutively active ALK1 (and its conse-quent Smad1,5,8 activation, see Figure 2E inset) reduced the TGF-β1-stimulated migration. Additionally, we also assayed MCF-7 migration upon Smad3 and Smad1 knockdown using a small interfering RNA strategy. As expected, knocking down of Smad3 provoked an inhibi-tion of cell motility. However, knocking down Smad1 by a similar procedure does not provoke changes in cell migration reinforcing the notion that cell motility is under the control of ALK5-Smad2/3 path-way (Figure 2F and G).

To investigate whether the decreased expression of Eng is indeed responsible for the increased migratory capacity and the changes observed in TGF-β1 signaling, we performed migration and Smads activation experiments. We used MCF-7 cells whose basal level of Eng was downmodulated with a short hairpin RNA approach. As expected, shEng MCF-7 cells displayed an enhanced capacity to migrate and showed both a decreased activity of Smad1,5,8 pathway and an stimulation of the Smad2,3 route (Supplementary Figure S2, available at Carcinogenesis Online). When we compared the effect of either CM-treated or shEng MCF-7 cells on Smad2/3 activa-tion, we found a similar result. Supplementary Figure S3, available at Carcinogenesis Online shows a quantification of the TGF-β1-dependent activation of Smad2/3 pathway in both experimental con-ditions that display a similar activation pattern. To assess whether the differential activation of both signaling pathways by their ligands also produces a differential migratory response, we evaluated MCF-7 migratory potential using TGF-β1 and BMP-2, which activate ALK5/Smad2,3 and ALK1/Smad1,5,8, respectively. As expected, TGF-β1


http://carcin.oxfordjournals.org/lookup/suppl/doi:10.1093/carcin/bgu061/-/DC1




behaved as a potent stimulator of MCF-7 migration, while BMP-2 inhibited cell movement. Migration rate reached after the stimulus of both agents suggest that, under these experimental conditions, the inhibitory stimulus of BMP-2 was not able to counteract the TGF-β1 stimulus (Supplementary Figure S4, available at Carcinogenesis Online).

Decreased MCF-7 Eng level after CM-HS-5 pretreatment is partially due to a metalloproteinase-dependent shedding processIt has been demonstrated that a shedding process may occur during Eng posttranslational regulation (21). To test if a proteolytic pro-cessing mechanism is involved in the CM-HS-5-dependent reduc-tion of Eng and the associated migratory enhancement, MCF-7 cells were pretreated with CM-HS-5 in the presence or absence of 10 μM GM6001 (Ilomastat), a broad-range inhibitor of metalloproteinase (MMP) activity. Figure 3A shows that inhibition of MMP activity

reduced the CM-HS-5-induced migratory enhancement of MCF-7 cells. These results are in agreement with the expression of Eng after Ilomastat treatment showing that Eng recovery after metalloproteinase inhibition was associated with a lower migration (Figure 3A, inset).

To confirm this hypothesis, MCF-7 cells were labeled at their surface with biotin (Biotin-NeutrAvidin procedure; see Materials and methods) and the avidin-bound Eng was identified by western blot analysis. Figure 3B shows that the abundance of cell surface Eng decreases, whereas soluble levels of Eng increase, in a dose-dependent manner after treatment with media conditioned by HS-5 cell. These results suggest that the decrease of membrane Eng is due to a proteolytic processing that leads to an increase in the solu-ble levels of Eng (sol-Eng). To verify that the soluble form of Eng found after CM HS-5 treatment is of epithelial origin, we also ana-lyzed the soluble Eng content of CM HS-5. We found that HS-5 cells also produced immunoreactive soluble Eng, but at much lower

Fig. 1. The medium conditioned by HS-5 cells reduces Eng levels and stimulates the migratory capacity of human breast cancer cell lines. (A) MCF-7 cells were pretreated (24 h) with CM-HS-5 and thereafter assayed for their migratory capacities in the absence of CM-HS-5 using the transwell method. The upper panel shows the expression levels of membrane Eng protein (m-Eng), as determined by western blotting. (B) Transwell migration assay of MCF-7 cells overexpressing a Myc-tagged version of Eng (Eng+) after preincubation or not with CM-HS-5. As a control, cells transfected with an empty vector (EV) were included. Inset shows expression of Myc-Eng protein. Bars in A and B indicate the average of migrated cells per field. (C) Confocal image of control (upper image) and CM-HS-5-treated MCF-7 cells (lower image) labeled with an anti-Eng antibody. Bar = 20 μm. Arrowheads indicate membrane-bound Eng. (D) Control and CM HS-5-treated MCF-10 cells were assayed for migration as in A. Inset shows the expression levels of Eng protein. (E) The migration abilities of MDA MB-231 cells preincubated or not with CM-HS-5 as in A were assayed using a transwell insert whose filter was coated with 10 μm of matrigel. Inset shows the Eng protein levels in treated and non-treated cells as determined by western blotting. (**) and (***) indicates statistically significant differences with P < 0.01 and P < 0.001, respectively.



N.Tobar et al.

Fig. 2. Pretreatment of MCF-7 cells with CM HS-5 enhances Smad 2,3 and inhibits Smad 1,5,8 signaling. (A) Western blot analysis of Smad1,5,8 and Smad2,3 phosphorylation relative to the total protein expression levels in MCF-7 cells preincubated or not (control) with CM-HS-5 and stimulated with TGF-β1 (5 ng/ml) at the indicated times. Values at the bottom of each lane are indicated relative to unstimulated control cells, in which an arbitrary value of 1 was given. (B) Western blot analysis of Smad1,5,8 phosphorylation relative to the total protein expression levels in MCF-7 cells preincubated or not (control) with CM-HS-5 and stimulated with 25 ng/ml of BMP2 at the indicated times. (C and D) TGF-β1 and BMP2-induced transcriptional activation of pCAGA (C) and pId1-Lux (D) luciferase reporter genes. Control and CM-HS-5-treated MCF-7 cells were transiently transfected with the reporter constructs and treated or not with TGF-β1 or BMP2, as indicated. Figures are representatives of three separate experiments. (E) ALK5 activation stimulates MCF-7 cell migration, whereas activation of ALK1 inhibits it. Cells were transiently transfected with the wild-type form of ALK5 (ALK5wt), a dominant-negative version of ALK5 (dnALK5) and an activated



levels than the epithelial soluble (Figure 3B). Therefore, HS-5 solu-ble Eng levels were negligible compared with those of epithelial origin.

TGF-β1 secreted by MCF-7 cells stimulates the production of MMP-14 by HS-5 cellsThe main hypothesis underlying our work is that stromal and epi-thelial compartments interchange soluble signals that modulate the course of tumoral progression. A previous work has shown that Eng shedding by a MMP-14 activity was a phenomenon that resulted of the close proximity of both molecules (Eng and membrane-bound MMP-14) in endothelial cell membrane (21). To assess if a similar process occurs in our system, the production of soluble and mem-brane-bound MMP-14 and MMP-2 (the most abundant stromal metal-loproteinase) was measured in the epithelial and stromal cells used in our system. Figure 4A shows that HS-5 cells were the unique source of MMPs in their soluble and insoluble form, thus ruling out the pos-sibility that Eng on MCF-7 cells was processed by endogenous MMPs of these cancer cells. Moreover, both soluble and insoluble forms of stromal MMP-14 show slightly increased levels when HS-5 cells were preincubated (16 h) with medium conditioned by MCF-7 cells. No change was observed in the production of stromal MMP-2 upon treat-ment with MCF-7-derived soluble factors (Figure 4A, left panel). In a previous work we have demonstrated that MCF-7 cells produce and secrete TGF-β (34). To test whether this soluble factor is involved in the production of stromal MMP-14, HS-5 cells were treated for 48 h with medium conditioned by MCF-7 (CM-MCF-7) in the pres-ence of increasing concentrations of SB431542, an inhibitor of the TGF-β type I receptor kinase. Figure 4B shows that CM-MCF-7 induces a robust stimulation of the expression of both soluble and membrane-bound MMP-14 and this enhancement is abolished in a dose-dependent manner when TGF-β type I receptor kinase was added to the culture. This result strongly suggests the involvement of epithelial TGF-β1 on stromal MMP-14 production and the existence of an active paracrine relationship between epithelial signals and stro-mal response. This interpretation was further confirmed by the results

presented in Figure 4C, showing that the production of the soluble form of MMP-14 is increased in a dose-dependent manner when HS-5 cells are cultured in the presence of TGF-β1.

To assess if epithelial TGF-β1 exerts its activity on stromal MMP-14 production through a transcriptional mechanism, we replicated the experiments of Figure 4B and 4C and measured the MMP-14 mRNA expression by quantitative PCR. Figure 4D shows that the expression of MMP-14 mRNA in HS-5 cells treated with CM-MCF-7 or TGF-β1 followed a similar response as the MMP-14 protein, suggesting a transcriptional regulation of MMP-14 under these conditions.

Depletion of MMP-14 from the medium conditioned by HS-5 cells reverts its promigratory effect on MCF-7 cellsTo address whether MMP-14 present in CM-HS-5 was directly involved in the acquisition of motile properties by MCF-7 cells, we performed a subset of migration experiments in which MMP-14 pre-sent in CM was previously depleted by a serial immunoprecipitation procedure. Figure 5A shows that pretreatment of MCF-7 cells with increasing concentrations of CM-HS-5 gave rise to a dose-dependent enhancement of their migratory capacity. When the most concentrated solution (50% CM HS-5) was subjected to immunoprecipitation using an anti MMP-14 antibody, the stromal-induced migratory capac-ity decreased in direct proportion with the decline of residual MMP abundance in CM-HS-5, as demonstrated by western blot analysis (Figure 5B). The diminution of MMP-14 content in CM HS-5 has also a consequence in the abundance of MCF-7 Eng. As Figure 5C shows, pretreatment of MCF-7 cells with increasing concentrations of CM HS-5 stimulate their migratory capacity in a linear relation with declination of cellular Eng abundance. Serial immunoprecipitation of MMP-14 allows the recovery of Eng abundance and the inhibition of the migratory potential. These results allow us to propose that MMP-14 present in the medium conditioned by HS-5 cells is responsible for shedding of epithelial membrane Eng and, in consequence, of the increased MCF-7 cells migration observed after preincubation of MCF-7 cells with CM-HS-5.

Fig. 3. A metalloproteinase is responsible for decreased Eng levels in MCF-7 cells preincubated with CM-HS-5. (A) Transwell migration assay of MCF-7 cells pretreated or not with CM-HS-5 in the presence or absence of GM6001 (Ilomastat), as indicated. Inset shows the abundance of Eng protein in the different conditions, as determined by western blotting. (B) Western blot analysis of soluble (sol-Eng) and membrane-bound Eng (m-Eng) after treatment of MCF-7 cells with increasing concentrations of CM-HS-5. Membrane-bound Eng was labeled using a Biotin-NeutrAvidin method. The epidermal growth factor receptor was used as loading protein control of the biotinylated membrane fraction. The level of soluble Eng in the CM of HS-5 cells was also evaluated. Figures are representatives of three independent experiments. (***) indicates statistically significant differences with P < 0.01.

version of ALK1 [ALK1(+)] and then tested for cell migration as in Figure 1. Inset shows that ALK1(+) expressing cells display an activated Smad1 pathway. MCF-7 cells transfected with either a small interfering RNA for Smad3 (F) or Smad1 (G) and non-related small interfering RNA controls, were assayed for their migratory capacities in the presence or absence of 5 ng/ml of TGF-β1 using the transwell method. Cells were allowed to migrate for 24 h and migration rates were evaluated as described in Materials and methods. The expressions of total and phosphorylated Smad3 (F) and Smad1 (G) proteins treated or not with TGF-β1 for 30 min were determined by western blotting. (*), (**) and (***) indicates statistically significant differences with P < 0.05, P < 0.01 and P < 0.001, respectively.

→


N.Tobar et al.

Fig. 4. TGF-β1 secreted by MCF-7 cells stimulates the production of MMP-14 by HS-5 cells. (A) Stromal HS-5 and epithelial MCF-7 cells were preincubated or not with 50% CM-MCF-7 or CM-HS-5, respectively, for 16 h. Afterward, cells were washed and incubated for 24 h in serum-free medium. Then, media were collected and cells lysed. The expression of MMP-14 and MMP-2 was analyzed by western blotting in the collected media and lysates. (B) HS-5 cells were incubated with medium conditioned by MCF-7 cells (50%) for 48 h in the presence or absence of increasing concentrations of SB431542, and MMP-14 protein levels were analyzed both in the collected media and in the cell lysate by western blotting. (C) HS-5 cells were incubated in the presence of increasing



Discussion

Interactions with stromal fibroblasts at the primary site of tumors cre-ate a favorable microenvironment supporting cancer growth, survival, evasion of immune surveillance and resistance to therapy (2). Results in the present work support the hypothesis that a stromal soluble form of the MMP-14 enzyme, acting in a paracrine manner, can modulate the migratory behavior of MCF-7 breast cancer cells by proteolytically processing the membrane-bound Eng present in these epithelial cells. This stromal stimulus on mammary carcinoma cell migration/invasion seems to be a more general response, because very similar responses to that of MCF-7 were obtained using a less migratory breast cell line, such as MCF-10, or a highly invasive MDA MB-231 cell line.

The production of proinvasive proteinases (as MMPs) is a common feature in numerous carcinomas, including breast, stomach, lung, colon and head-and-neck tumors (35). MMP-14 represents a type I transmembrane MMP that is expressed by both stromal fibroblasts and carcinoma cells (36,37). Membrane-bound MMP-14 undergoes a complex process of shedding producing a series of soluble active and inactive fragments. Among them, the 50–53 kDa fragments cor-respond to the entire ectodomain of the enzyme and are active spe-cies. Moreover, the 44 kDa fragment is derived from a self-regulatory mechanism that reduces pericellular proteolytic activity and corre-sponds to an inactive form (38).

Increased levels of serum MMP-14 have been identified in breast cancer patients without clinically apparent metastases, suggesting that this MMP is produced through a shedding process attributable to an enhancement of proteolytic environment that characterizes tumor progression (39). In line with these data, our results show that a solu-ble form of MMP-14, generated by HS-5 bone marrow-derived cells, sheds the membrane-bound Eng on MCF-7 cell surface by a mecha-nism that involves a MCF7-derived stimulus mediated by TGF-β1. The inductive role of bone marrow stromal cells on the acquisition of the malignant properties of breast cancer cells has been evaluated in a variety of experimental models (40). Thus, the paracrine modulation of Eng expression in tumor cells by CM-HS-5, demonstrated here, may contribute to the growth and malignant transformation of tumors. In this regard, using different cellular models, silencing of epithelial

Eng has been proposed as a relevant mechanism to control tumor cell growth, as well as cell migration and invasion in vitro (10,14). In a human mammary cells model, it has been hypothesized that Eng expression attenuates promigratory and proinvasive effects of TGF-β1 in cells containing active oncogenes; in this case, the expression level of Eng does not affect TGF-β1 signaling (11). Experiments performed in a mouse model of skin carcinogenesis have also demonstrated that shedding of surface-bound Eng, which in turn disrupts its TGF-β-signaling regulatory function is an event associated with malignant progression in vivo (14). Other authors provided evidences that Eng shedding constitutes a plausible regulatory mechanism of angiogene-sis (21). These authors showed evidence that a surface-located MMP-14 was able to cleave Eng in endothelial cells, a phenomenon that may regulate the angiogenic potential in a tumor microenvironment.

Our experiments were performed using mainly a weakly invasive tumor carcinoma cell (MCF-7) that express Eng at low levels and does not express detectable amounts of MMPs either at the pericellular and soluble levels. Conversely, the stromal cellular component, repre-sented here by the HS-5 cell line, is an active producer of MMPs, a common fact in many human peritumoral fibroblasts (37). O’Connor et al. (12) also demonstrated that this stromal cell line expresses Eng and that its level is under the control of prostatic soluble factors.

From our data we propose that stromal HS-5 cells produce and secrete a soluble form of MMP-14 that sheds epithelial Eng through a proteolytic process triggered by TGF-β1 secreted by the same breast epithelial cells (Figure 6).

Thereby, overexpression of epithelial TGF-β1, acting in a parac-rine manner, could serve as a signal to produce a stromal proteolytic activity (MMP-14) that, in turn, stimulates the shedding of epithelial Eng, reinforcing the establishment of a malignant state. This potential mechanism is obviously restricted to weakly invasive carcinoma cells, which express detectable levels of Eng, and might represent an early step in the onset of tumor progression. In fact, in a study that evalu-ated by in situ reverse transcription–polymerase chain reaction the expression of MMPs in breast tumor samples, MMP-14 appears to have the highest levels of gene expression in the preinvasive tumors examined, suggesting a role for MMP-14 in the tumor invasion pro-cess (41). Similarly, in pancreatic ductal adenocarcinoma, a deadly

Fig. 5. Depletion of MMP-14 from CM-HS-5 impairs its promigratory effect on MCF-7 cells. (A) Transwell migration assays of MCF-7 cells preincubated (for 24 h) with increasing concentrations of intact CM-HS-5 and CM (50%) serially depleted from MMP-14 by immunoprecipitation (two cycles IP1 and IP2), using an anti MMP-14 antibody bound to protein A sepharose. (B) Western blot analysis of residual MMP-14 levels in CM-HS-5 after IP1 and IP2. (C) Western blot analysis of MCF-7 membrane-bound Eng levels after incubation with increasing concentrations of CM HS-5 before and after serial immunoprecipitation (IP1 and IP2) of MMP-14.

concentrations of TGF-β1 for 16 h, and MMP-14 protein levels were analyzed by western blotting. (D) HS-5 cells were subjected to the same experimental conditions as in B and C. At the end of the incubation period, cells were lysed with Trizol and mRNA levels of MMP-14 were measured by reverse transcription–quantitative PCR.

→


N.Tobar et al.

human malignancy characterized by its desmoplastic nature, an increased expression of MMP-14 is correlated with an upregulation of TGF-β1, a factor that promotes MMP-14 expression (42).

In mechanistic terms, we propose that changes in migratory behav-ior resulting from Eng shedding may be associated with changes in TGF-β signaling. Previous work has proposed that, as a result of the extracellular and cytoplasmic interactions of Eng with ALK5 and ALK1, the latter route is potentiated (9). Our data shows that, after pre-incubation of MCF-7 with CM-HS-5, the expression of surface-bound Eng declines and consequently, the Smad1,5,8 pathway is inhibited, whereas Smad2/3 phosphorylation is stimulated. In many cellular systems it has been demonstrated that TGF-β1-induced activation of Smad3 constitutes a critical step for cellular events that play a key role in the induction of cell migration (43,44). Our data, showing that cell migration is blocked by both overexpression of a dominant-negative version of ALK5 or knocking down Smad3, reinforce the importance of ALK5/Smad2/3 activation in cell motility (Figure 2E and Fig.S2A in Supplementary Figure S2A, available at Carcinogenesis Online). Prior studies performed in prostatic cell lines have shown that the loss of Eng protein expression is associated with decreased cell adhesion and concomitant increased cell migration (45). Moreover, in an in vivo model of prostate cancer, Lakshman et al. (46) reinforces the suppressor role for Eng providing evidence that epithelial Eng loss leads to progressive increases in the number of circulating metastatic tumor cells as well as the formation of distant soft tissue metastasis.

Taken together, this study supports the idea that the fate of a malig-nant epithelial cell depends largely on its interaction with an active stroma (2). In this case, we propose (Figure 6) that the endogenous autocrine production of TGF-β, a characteristic trait of epithelial malignancy, may act in a paracrine manner, stimulating the produc-tion of a soluble form of stromal MMP-14 that, in turn, cleaves the membrane-bound epithelial Eng. It has been clearly demonstrated that disruption of membrane Eng is an important event in tumor progres-sion. Therefore, it is highly possible that Eng-processed cells acquire a whole set of malignant traits (autocrine TGF-β production, among

others) that perpetuates the expression of a more motile malignant phenotype.

Supplementary material

Supplementary Figures S1–S4 can be found at http://carcin.oxford-journals.org/

Funding

Fondo Nacional de Investigación Científica y Tecnológica of Chile (1120187 to J.M.); Ministerio de Economia y Competitividad of Spain (SAF2010-19222 to C.B., SAF2010-19152 to M.Q.); Comunidad de Madrid (S2010/BMD-2359, SkinModel, to M.Q.).

Acknowledgement

We thank CIBERER that is an initiative of the Instituto de Salud Carlos III (ISCIII) of Spain.

Conflict of Interest Statement: None declared.

References

1. De Wever,O. et al. (2003) Role of tissue stroma in cancer cell invasion. J. Pathol., 200, 429–447.

2. Radisky,E.S. et al. (2007) Stromal induction of breast cancer: inflammation and invasion. Rev. Endocr. Metab. Disord., 8, 279–287.

3. Elenbaas,B. et al. (2001) Heterotypic signaling between epithelial tumor cells and fibroblasts in carcinoma formation. Exp. Cell Res., 264, 169–184.

4. Egeblad,M. et al. (2005) The fibroblastic coconspirator in cancer progres-sion. Cold Spring Harb. Symp. Quant. Biol., 70, 383–388.

5. Ostman,A. et al. (2009) Cancer-associated fibroblasts and tumor growth–bystanders turning into key players. Curr. Opin. Genet. Dev., 19, 67–73.

6. DiMeo,T.A. et al. (2006) The evolving paradigm of tissue-specific metas-tasis. Breast Cancer Res., 8, 301.

Fig. 6. Hypothetical model for epithelial–stromal cross-talk modulating Eng-dependent breast cancer cell migration. TGF-β1 secreted by tumor cells stimulates the expression of membrane and soluble MMP-14 in stromal cells. In turn, MMP-14 sheds membrane-bound Eng present on the surface of tumor cells, leading to the stimulation of basal and TGF-β1-induced ALK5/Smad2,3 signaling activity and increased tumor cell migration/invasion.




http://carcin.oxfordjournals.org/

http://carcin.oxfordjournals.org/


7. Cheifetz,S. et al. (1992) Endoglin is a component of the transforming growth factor-beta receptor system in human endothelial cells. J. Biol. Chem., 267, 19027–19030.

8. Fonsatti,E. et al. (2003) Endoglin (CD105): a powerful therapeutic target on tumor-associated angiogenetic blood vessels. Oncogene, 22, 6557–6563.

9. Bernabeu,C. et al. (2009) The emerging role of TGF-beta superfamily coreceptors in cancer. Biochim. Biophys. Acta, 1792, 954–973.

10. Craft,C.S. et al. (2007) Endoglin inhibits prostate cancer motility via acti-vation of the ALK2-Smad1 pathway. Oncogene, 26, 7240–7250.

11. Henry,L.A. et al. (2011) Endoglin expression in breast tumor cells sup-presses invasion and metastasis and correlates with improved clinical out-come. Oncogene, 30, 1046–1058.

12. O’Connor,J.C. et al. (2007) Coculture with prostate cancer cells alters endoglin expression and attenuates transforming growth factor-beta signal-ing in reactive bone marrow stromal cells. Mol. Cancer Res., 5, 585–603.

13. Oxmann,D. et al. (2008) Endoglin expression in metastatic breast cancer cells enhances their invasive phenotype. Oncogene, 27, 3567–3575.

14. Pérez-Gómez,E. et al. (2007) A role for endoglin as a suppressor of malig-nancy during mouse skin carcinogenesis. Cancer Res., 67, 10268–10277.

15. Romani,A.A. et al. (2006) The risk of developing metastatic disease in colorectal cancer is related to CD105-positive vessel count. J. Surg. Oncol., 93, 446–455.

16. Li,C. et al. (2003) Both high intratumoral microvessel density determined using CD105 antibody and elevated plasma levels of CD105 in colorectal cancer patients correlate with poor prognosis. Br. J. Cancer, 88, 1424–1431.

17. Li,C. et al. (2000) Plasma levels of soluble CD105 correlate with metasta-sis in patients with breast cancer. Int. J. Cancer, 89, 122–126.

18. Guerrero-Esteo,M. et al. (1999) Endoglin overexpression modulates cellu-lar morphology, migration, and adhesion of mouse fibroblasts. Eur. J. Cell Biol., 78, 614–623.

19. Romero,D. et al. (2011) Endoglin regulates cancer-stromal cell interactions in prostate tumors. Cancer Res., 71, 3482–3493.

20. Takahashi,N. et al. (2001) Association of serum endoglin with metastasis in patients with colorectal, breast, and other solid tumors, and suppressive effect of chemotherapy on the serum endoglin. Clin. Cancer Res., 7, 524–532.

21. Hawinkels,L.J. et al. (2010) Matrix metalloproteinase-14 (MT1-MMP)-mediated endoglin shedding inhibits tumor angiogenesis. Cancer Res., 70, 4141–4150.

22. Liu,C. et al. (2009) TACE-mediated ectodomain shedding of the type I TGF-beta receptor downregulates TGF-beta signaling. Mol. Cell, 35, 26–36.

23. Velasco-Loyden,G. et al. (2004) The shedding of betaglycan is regulated by pervanadate and mediated by membrane type matrix metalloprotease-1. J. Biol. Chem., 279, 7721–7733.

24. Roecklein,B.A. et al. (1995) Functionally distinct human marrow stromal cell lines immortalized by transduction with the human papilloma virus E6/E7 genes. Blood, 85, 997–1005.

25. Faúndez,V. et al. (1992) Epidermal growth factor receptor in synaptic frac-tions of the rat central nervous system. J. Biol. Chem., 267, 20363–20370.

26. Lux,A. et al. (2000) Expression analysis of endoglin missense and trun-cation mutations: insights into protein structure and disease mechanisms. Hum. Mol. Genet., 9, 745–755.

27. Santibanez,J.F. et al. (2010) The TGF-beta co-receptor endoglin modulates the expression and transforming potential of H-Ras. Carcinogenesis, 31, 2145–2154.

28. Tobar,N. et al. (2008) RAC1 activity and intracellular ROS modulate the migratory potential of MCF-7 cells through a NADPH oxidase and NFkappaB-dependent mechanism. Cancer Lett., 267, 125–132.

29. Blanco,F.J. et al. (2005) Interaction and functional interplay between endoglin and ALK-1, two components of the endothelial transforming growth factor-beta receptor complex. J. Cell. Physiol., 204, 574–584.

30. Dennler,S. et al. (1998) Direct binding of Smad3 and Smad4 to critical TGF beta-inducible elements in the promoter of human plasminogen acti-vator inhibitor-type 1 gene. EMBO J., 17, 3091–3100.

31. López-Rovira,T. et al. (2002) Direct binding of Smad1 and Smad4 to two distinct motifs mediates bone morphogenetic protein-specific transcrip-tional activation of Id1 gene. J. Biol. Chem., 277, 3176–3185.

32. Rose,A.A. et al. (2006) Breast cancer-derived factors facilitate osteolytic bone metastasis. Bull. Cancer, 93, 931–943.

33. Clement,J.H. et al. (2000) Bone morphogenetic protein 2 (BMP-2) induces sequential changes of Id gene expression in the breast cancer cell line MCF-7. J. Cancer Res. Clin. Oncol., 126, 271–279.

34. Guerrero,J. et al. (2010) Soluble factors derived from tumor mammary cell lines induce a stromal mammary adipose reversion in human and mice adi-pose cells. Possible role of TGF-beta1 and TNF-alpha. Breast Cancer Res. Treat., 119, 497–508.

35. Coussens,L.M. et al. (1996) Matrix metalloproteinases and the develop-ment of cancer. Chem. Biol., 3, 895–904.

36. Okada,A. et al. (1995) Membrane-type matrix metalloprotein-ase (MT-MMP) gene is expressed in stromal cells of human colon, breast, and head and neck carcinomas. Proc. Natl Acad. Sci. USA, 92, 2730–2734.

37. Ueno,H. et al. (1997) Expression and tissue localization of membrane-types 1, 2, and 3 matrix metalloproteinases in human invasive breast carci-nomas. Cancer Res., 57, 2055–2060.

38. Osenkowski,P. et al. (2004) Processing, shedding, and endocytosis of membrane type 1-matrix metalloproteinase (MT1-MMP). J. Cell. Physiol., 200, 2–10.

39. Laudański,P. et al. (2010) Increased serum level of membrane type 1-matrix metalloproteinase (MT1-MMP/MMP-14) in patients with breast cancer. Folia Histochem. Cytobiol., 48, 101–103.

40. Karnoub,A.E. et al. (2007) Mesenchymal stem cells within tumour stroma promote breast cancer metastasis. Nature, 449, 557–563.

41. Haupt,L.M. et al. (2008) Matrix metalloproteinase localisation by in situ-RT-PCR in archival human breast biopsy material. Mol. Cell. Probes, 22, 83–89.

42. Shields,M.A. et al. (2012) Biochemical role of the collagen-rich tumour microenvironment in pancreatic cancer progression. Biochem. J., 441, 541–552.

43. Kocic,J. et al. (2012) SMAD3 is essential for transforming growth factor-β1-induced urokinase type plasminogen activator expres-sion and migration in transformed keratinocytes. Eur. J. Cancer, 48, 1550–1557.

44. Olieslagers,S. et al. (2011) TGF-β1/ALK5-induced monocyte migration involves PI3K and p38 pathways and is not negatively affected by diabetes mellitus. Cardiovasc. Res., 91, 510–518.

45. Liu,Y. et al. (2002) Over expression of endoglin in human prostate can-cer suppresses cell detachment, migration and invasion. Oncogene, 21, 8272–8281.

46. Lakshman,M. et al. (2011) Endoglin suppresses human prostate cancer metastasis. Clin. Exp. Metastasis, 28, 39–53.

Received January 31, 2013; revised February 3, 2014; accepted February 4, 2014


soluble mmp-14 produced by bone marrow-derived

Documents