The aspartyl-protease cathepsin D (cath-D) is overexpressed and hypersecreted by epithelial breast cancer cells and stimulates their proliferation. As tumor epithelial–fibroblast cell interactions are important events in cancer progression, we investigated whether cath-D overexpression affects also fibroblast behavior. We demonstrate a requirement of cath-D for fibroblast invasive growth using a three-dimensional (3D) coculture assay with cancer cells secreting or not pro-cath-D. Ectopic expression of cath-D in cath-D–deficient fibroblasts stimulates 3D outgrowth that is associated with a significant increase in fibroblast proliferation, survival, motility, and invasive capacity, accompanied by activation of the ras–MAPK pathway. Interestingly, all these stimulatory effects on fibroblasts are independent of cath-D proteolytic activity. Finally, we show that pro-cath-D secreted by cancer cells is captured by fibroblasts and partially mimics effects of transfected cath-D. We conclude that cath-D is crucial for fibroblast invasive outgrowth and could act as a key paracrine communicator between cancer and stromal cells, independently of its catalytic activity.

Introduction

Cathepsin D (cath-D) (EC 3.4.23.5) is a ubiquitous lysosomal aspartyl endoproteinase extensively reported as being overexpressed and hypersecreted by human breast cancer cells (Rochefort et al., 1987). Many independent clinical studies have associated cath-D overexpression with increased risk of clinical metastasis and shorter survival in breast cancer patients (Rochefort, 1992; Ferrandina et al., 1997; Foekens et al., 1999; Westley and May, 1999). During transport to lysosomes, 52-kD human inactive pro-cath-D is proteolytically processed into a single-chain intermediate active enzyme of 48-kD located in endosomes. Further proteolytic processing yields the mature active lysosomal protease which is composed of heavy (34 kD) and light (14 kD) chains (Richo and Conner, 1991). Human cath-D catalytic site includes two critical aspartyl residues (amino acids 33 and 231) located on the 14- and 34-kD chains, respectively (Metcalf and Fusek, 1993). Mannose-6-phosphate (Man-6-P) receptors are involved in cath-D lysosomal routing and in the cellular uptake of the secreted pro-cath-D (Von Figura and Hasilik, 1986) although cath-D may also be targeted to the lysosome and endocytosed independently of Man-6-P receptors (Capony et al., 1994, Laurent-Matha et al., 1998). Recently, cath-D was shown to be a rate limiting factor for outgrowth, tumorigenicity, and lung colonization of MDA-MB-231 breast cancer cells using an RNA antisense strategy (Glondu et al., 2002). Several reports have indicated that cath-D stimulates cancer cell proliferation (Vignon et al., 1986; Vetvicka et al., 1994; Liaudet et al., 1995) and increases metastatic potential in vivo (Garcia et al., 1990; Liaudet et al., 1994). Moreover, we reported that D231N/cath-D mutated in its catalytic site, and hence devoid of proteolytic activity is still mitogenic for cancer cells both in vitro in three-dimensional (3D) matrices and in vivo in athymic nude mice (Rochefort and Liaudet-Coopman, 1999; Glondu et al., 2001). Immunohistochemical studies indicated that cath-D, independently of its proteolytic activity, stimulates not only cancer cell proliferation by an autocrine and/or intracrine mechanism, but also tumor angiogenesis by a paracrine mechanism (Berchem et al., 2002). This revealed for the first time the potential paracrine action of cath-D in the context of a tumor.

Tumor progression requires a continually evolving network of interactions between neoplastic cells and their microenvironment consisting of an ECM, a stroma composed of fibroblasts, adipose, vasculature and resident immune cells, and a mixture of cytokines and growth factors (Pupa et al., 2002). Current observations increasingly point to the contribution of stromal components to oncogenic signals that mediate both phenotypic and genomic changes in epithelial cells (Tlsty and Hein, 2001). Carcinoma formation from its earliest stages seems to depend on the ability of transformed epithelial cells to first recruit and then subvert a variety of stromal cells originating from adjacent normal tissue to become infiltrating or invading (Elenbaas and Weinberg, 2001). In many common carcinomas including breast, colon, stomach, and pancreas, stroma comprises the majority of tumor mass, in some cases accounting for >90%. In various experimental tumor models, the microenvironment affects efficiency of tumor formation, rate of tumor growth, extent of invasiveness, and ability of tumor cells to metastasize. In breast carcinomas, the influences of microenvironment are mediated, in large part, by paracrine signaling between epithelial tumor cells and neighboring stromal fibroblasts (Elenbaas and Weinberg, 2001). In addition to receiving signals from epithelial cells, the stromal fibroblasts stimulate tumorigenesis by releasing factors that act on epithelial tumor cells. Stromal and tumor cells exchange enzymes, growth factors and cytokines that modify local ECM, stimulate migration and invasion, and promote proliferation and survival of tumor cells (Masson et al., 1998; Liotta and Kohn, 2001; Bisson et al., 2003). A stromal reaction immediately adjacent to transformed epithelial cells has been documented in several tumor systems (Basset et al., 1990; Olumi et al., 1999; Shekhar et al., 2001).

In the present study, we highlight the possibility that cath-D may participate in such paracrine signaling interactions between epithelial cancer cells and associated fibroblasts. The paracrine function of cath-D hypersecreted by cancer cells was first examined both in a 3D coculture assay and in an aortic ring assay. We then went on to study the effects of the ectopic human cath-D expression in cath-D–deficient fibroblasts on their behavior profile, e.g., outgrowth, proliferation, survival, migration, invasion, and signalization.

Results

Cancer cells secrete paracrine factors crucial for fibroblast invasive growth

The communication between epithelial cancer cells and fibroblasts and in particular the role of pro-cath-D secreted by cancer cells on stromal compartments was explored using a 3D assay of cath-D–deficient mouse CD55−/−SV40, human skin CCD45K, or human mammary fibroblasts (HMFs) cocultured with cancer cells. Human breast cancer MCF-7 cells induced outgrowth of all fibroblast cell lines tested indicating that epithelial cancer cells were secreting paracrine factors capable of promoting fibroblast outgrowth in 3D matrices (Fig. 1 A). To assess whether secreted pro-cath-D might be one of these crucial paracrine factors, 3D coculture assays were performed with cath-D–transfected rat 3Y1-Ad12 cancer cell lines secreting no cath-D (control), or hypersecreting human wild-type cath-D (cath-D), or D231N mutated cath-D devoid of proteolytic activity (D231N). Only cancer cells secreting wild-type or D231N pro-cath-D stimulated CD55−/−SV40 fibroblast outgrowth, whereas control cells had no effect (Fig. 1 B) suggesting that secreted pro-cath-D might promote 3D outgrowth of fibroblasts in a paracrine manner. Because the catalytic activity was clearly not involved, it might occur by binding to potential receptors present on fibroblast cell surfaces or to nonreceptor proteins. We excluded the Man-6-P/IGF2 receptor as being the mediator of the cath-D induction of fibroblast invasive growth since addition of Man-6-P in excess sufficient to prevent the binding of cath-D to this receptor (Capony et al., 1987) had no effect in our 3D coculture assays (unpublished data). To strengthen the possibility of a direct paracrine action of secreted pro-cath-D, 3D coculture assays were performed with MCF-7 cells whose endogenous pro-cath-D secretion was inhibited by 75% with small RNA-mediated gene silencing (cath-D siRNA) for 2 d before the coculture assay (Fig. 1 C, b). MCF-7 cells transfected with cath-D siRNA were less effective in triggering CD55−/−SV40 fibroblast outgrowth when compared with MCF-7 cells transfected with control luciferase siRNA (luc siRNA; Fig. 1 C, a). A 75% inhibition of pro-cath-D secretion by cath-D siRNA still occurred during our coculture assay even after 4 d of coculture (Fig. 1 C, b). Therefore, we conclude that inhibition of pro-cath-D secretion rapidly lead to a decrease in fibroblast invasive growth. Results similar to those depicted in Fig. 1 (B and C) were obtained with HMF fibroblasts indicating the possible relevance of the paracrine role of secreted pro-cath-D in the pathological context (not depicted). Moreover, secreted wild-type and D231N pro-cath-D also induced fibroblast outgrowth in an aortic ring assay (Fig. S1, available at http://www.jcb.org/cgi/content/full/jcb.200403078/DC1). Altogether, our results support the concept that pro-cath-D secreted by epithelial cancer cells might exhibit a crucial paracrine function on stromal cells. Interestingly, all these stimulatory effects on fibroblasts were independent of the catalytic activity of cath-D.

Characterization of mouse fibroblasts engineered to express catalytically active or inactive human cath-D

To dissect the effect of cath-D on fibroblast behavior, we stably transfected a cath-D–deficient fibroblast CD55−/− cell line with expression vectors for wild-type, D231N cath-D, or with an empty vector, generating CD55−/−cath-D, CD55−/−D231N, and CD55−/−SV40 cell lines, respectively. A control cell line (CD53+/+) expressing endogenous mouse cath-D was stably transfected with an empty vector, generating the CD53+/+SV40 cell line. The cath-D concentration detected in cell extracts for CD55−/−cath-D and CD55−/−D231N cells were 7.9 ± 1 ng/μg DNA and 1.6 ± 0.1 ng/μg DNA corresponding to 15 and 3 pmoles/mg cytosolic protein. The amount of pro-cath-D secreted in 24 h by CD55−/−cath-D and CD55−/−D231N cells was respectively 2.6 ± 0.4 ng/μg DNA and 0.3 ± 0.1 ng/μg DNA. Expression of wild-type cath-D was significantly greater than that of D231N cath-D, suggesting that the D231N mutation might affect the protein stability. However, cellular processing of 52-kD cath-D–D231N into 48-kD intermediate and 34-kD mature forms was not prevented by the D231N mutation in fibroblasts (Fig. S2 A, available at http://www.jcb.org/cgi/content/full/jcb.200403078/DC1) and the D231N mutant was devoid of catalytic function in our cellular model (Fig. S2 B). The level of human wild-type cath-D expressed by transfected CD55−/− fibroblasts was comparable to that of endogenous mouse cath-D produced by CD53+/+SV40 fibroblasts (Fig. S2 B, top). However, mouse cath-D was not secreted (Fig. S2 B, bottom).

Human cath-D triggers 3D outgrowth of fibroblasts independently of its proteolytic activity

We next studied the effect of wild-type or D231N mutated cath-D expression on fibroblast outgrowth in 3D matrices (Fig. 2). As shown in Fig. 2 A, human cath-D promoted outgrowth of mouse fibroblasts embedded into Matrigel. After 8 d of culture CD55−/−cath-D, CD55−/−D231N and CD53+/+SV40 cells had adopted a stellate morphology of growing and invasive colonies with protrusions sprouting into the surrounding matrix (Fig. 2 A). In contrast, CD55−/−SV40 fibroblasts presented a well-delineated spherical appearance of quiescent and/or dying cells and grew poorly, neither invading nor forming protrusions (Fig. 2 A). Similar results were obtained when fibroblasts were embedded into collagen I gels (unpublished data). Moreover, when plated on Matrigel gels, CD55−/−cath-D and CD55−/−D231N cells grew as efficiently as CD53+/+SV40 cells (Fig. 2 B). This growth was shown by cell staining (Fig. 2 B, top) and by phase-contrast microscopy (Fig. 2 B, bottom). In the latter case, higher magnification showed a network of interacting cells (Fig. 2 B, bottom). Cath-D–deficient CD55−/−SV40 fibroblasts grew poorly and displayed a stressed morphology of well-delineated rounded cells (Fig. 2 B, bottom). We finally compared migration of transfected fibroblasts in wound healing experiments performed on collagen I. As shown in Fig. 3, CD55−/−cath-D and CD55−/−D231N fibroblasts migrated more efficiently into a wound area as compared with CD55−/−SV40 fibroblasts. Similar results were obtained with cells directly seeded on plastic (unpublished data). Altogether, these results demonstrate that cath-D triggers fibroblast invasive outgrowth in matrices. Moreover, the mechanism is independent of its proteolytic activity because catalytically inactive D231N cath-D is as effective in stimulating cell outgrowth.

Catalytically inactive cath-D promotes proliferation and survival, as well as stimulates migration and invasion of fibroblasts

We investigated whether stimulation of fibroblast outgrowth induced by cath-D was accompanied by increased proliferation, survival, migration, or invasiveness. As shown in Fig. 4 A, stable expression of wild-type or mutated D231N cath-D in cath-D–deficient CD55−/− cells lead to a significant stimulation of proliferation relative to that of CD55−/−SV40 cells. Similar proliferation was observed in CD53+/+SV40 fibroblasts expressing endogenous mouse cath-D (Fig. 4 A). Cell cycle analyses by flow cytometry revealed that the stimulation of cell proliferation induced by transfected wild-type and D231N cath-D was due to both a significant increase in the number of proliferating cells in S phase, 52.5 ± 2.5%, 61.7 ± 2.5%, and 65.9 ± 2% for CD55−/−SV40, CD55−/−cath-D, and CD55−/−D231N cells, respectively, and to a significant decrease in the number of apoptotic cells, 7.5 ± 2.9%, 0.79 ± 0.47%, and 0.3 ± 0.25% for CD55−/−SV40, CD55−/−cath-D, and CD55−/−D231N cells, respectively (Fig. 4 B, a). BrdUrd-incorporated S phase cells confirmed the increased S phase of CD55−/−cath-D (64%) and CD55−/−D231N cells (63%) relative to that of CD55−/−SV40 cells (50%; Fig. 4 B, b). These results indicate that cath-D stimulates fibroblast proliferation in a manner independent of its catalytic activity. In addition to its positive effect on proliferation, cath-D also seems to rescue fibroblasts from apoptosis and promote their survival. The apoptotic status was further analyzed by flow cytometry with transfected fibroblasts grown for 3 d on Matrigel. A significantly higher proportion of CD55−/−SV40 fibroblasts appeared in a sub-G0/G1 (apoptotic) peak (9,8 ± 2.7%, n = 5) as compared with CD55−/−cath-D (0.39 ± 0.55%, n = 3), CD55−/−D231N (0.59−; 2.16%, n = 2) and CD53+/+SV40 (0 − 0%, n = 2) fibroblasts (Fig. 4 C). Additionally, CD55−/−SV40 fibroblasts embedded in Matrigel and stained with cell permeable Hoechst 33342 presented an apoptotic phenotype characterized by the presence of apoptotic bodies, which could not be detected in CD55−/−cath-D fibroblasts (Fig. 4 D). These results were confirmed by transmission electron microscopy (Fig. 4 E). CD55−/−SV40 fibroblasts exhibited features of apoptotic cells characterized by the presence of a well-delineated cytoplasm and a complete chromatin condensation (Fig. 4 E, a and b). It is noteworthy that this primary apoptosis was followed by secondary necrosis under our culture conditions in Matrigel (Fig. 4 E, b). By contrast, CD55−/−cath-D cells never presented such an apoptotic phenotype (Fig. 4 E, c). These results demonstrate that cath-D–deficient CD55−/−SV40 fibroblasts were dying in matrices and that wild-type cath-D could promote their survival. We next evaluated whether cath-D could also play a role in fibroblast migration or invasiveness, important steps for tumor invasion and metastatic process. Indeed, results from Figs. 2 and 3 already revealed that CD55−/−cath-D, CD55−/−D231N, and CD53+/+SV40 cells migrated and invaded the surrounding matrices, suggesting a possible role of cath-D in migration and invasiveness. Chemotactic migration and invasiveness of fibroblasts were measured in parallel Boyden Chamber assays (Fig. 4, F and G). The basal level of migration of CD55−/−SV40 fibroblasts was relatively high as expected for mesenchymal cells, e.g. 30 ± 5% of cells crossed filters coated with collagen I (Fig. 4 F). Cath-D expressing cells (CD55−/−cath-D, CD55−/−D231N, and CD53+/+SV40 cells) showed a small but significant increase of 1.4-fold in migrative capacity, when compared with cath-D–deficient CD55−/−SV40 cells (Fig. 4 F). In addition, invasive assays performed with a Matrigel barrier highlighted the fact that CD55−/−cath-D, CD55−/−D231N and CD53+/+SV40 cells were significantly more invasive than CD55−/−SV40 cells by a factor of 2.7-fold (Fig. 4 G). A similar induction of invasiveness was also obtained in lower serum concentrations (1% FCS + 0.1% BSA) in the top chamber (unpublished data). Collectively, these results indicate that cath-D could contribute to fibroblast invasive growth by promoting not only proliferation and survival, but also by stimulating migration and invasiveness in a manner independent of its catalytic function.

Activation of MAPK in fibroblasts

The ras–MAPK signaling cascade has been reported as being involved in proliferation, survival, motility, and invasion (Bar-Sagi and Hall, 2000; Schlessinger, 2000; Iwabu et al., 2004). Therefore, we examined whether cath-D could activate the MAPK cascade in fibroblasts. Specific anti-pERK mAbs enabled us to follow levels of ERK1/2 kinase phosphorylation, reflecting activation of the MAPK pathway (Fig. 5 A, a). Results from Fig. 5 A (b) revealed that ERK1 and ERK2 phosphorylation levels were significantly induced by 2.4-, 3.5-, and 5.4-fold in CD55−/−cath-D, CD55−/−D231N, and CD53+/+SV40 fibroblasts, respectively, relative to CD55−/−SV40 fibroblasts. This effect was observed by culturing cells in 2% FCS, which corresponded to the experimental conditions used in our proliferation assays (Fig. 4 A). We then investigated whether fibroblast invasive growth induced in 3D matrices by transfected cath-D could be linked to a stimulation of the MAPK pathway. CD55−/−cath-D cells were either pretreated or not with a MEK1/2 inhibitor, U0126, and were then embedded in Matrigel. U0126 efficiently inhibited ERK1/2 kinase phosphorylation (Fig. 5 B, a) and thereby prevented invasive growth of CD55−/−cath-D fibroblasts (Fig. 5 B, b). Together, these findings suggest that cath-D may be responsible for positive regulation of proliferation, survival, motility, and/or invasion of fibroblasts by triggering activation of ras/MAPK/ERKs.

Paracrine function of cath-D on fibroblasts

To provide further support for a direct paracrine action of pro-cath-D secreted by breast cancer cells on fibroblasts, we next studied the binding and endocytosis of secreted [35S]methionine-labeled pro-cath-D by recipient CD55−/− fibroblasts (Fig. 6 A). The labeled 52-kD pro-enzyme was totally transformed into 48-kD intermediate and 34-kD mature enzymes after its binding and endocytosis after 24 h (Fig. 6 A, b, 0). The addition of Man-6-P in excess, described as preventing cath-D binding to Man-6-P receptors (Capony et al., 1987), partly inhibited pro-cath-D internalization by 65% (Fig. 6 A, b, + Man-6-P). These results support the hypothesis of a paracrine action of pro-cath-D secreted by breast cancer cells being bound and captured by fibroblasts via Man-6-P receptors and another putative cell surface receptor. To further characterize the cath-D paracrine action, we next studied the effect of conditioned medium from mock-transfected (control), wild-type (cath-D), or D231N cath-D (D231N) transfected 3Y1-Ad12 cancer cell lines on CD55−/−SV40 fibroblast outgrowth, proliferation and apoptosis, invasion, and activation of the MAPK signaling pathway. Fig. 6 B shows that conditioned media containing wild-type or D231N secreted pro-cath-D stimulated fibroblast outgrowth, whereas conditioned medium from control cells was inefficient in this respect. The paracrine effect of secreted pro-cath-D on CD55−/−SV40 fibroblast outgrowth was associated with a significant 10% increase of proliferating cells in S phase (Fig. 6 C, a and b), a significant 1.4-fold increase in their invasive capacity (Fig. 6 D) and a threefold induction of ERK1 and ERK2 phosphorylation (Fig. 6 E). However, no significant effect on apoptosis was observed after addition of pro-cath-D (Fig. 6 C, a). Overall, these results show that secreted pro-cath-D partially mimics the effects of transfected cath-D and strongly suggest that pro-cath-D is acting extracellularly via a paracrine loop in a manner independent of its catalytic activity.

Discussion

Interactions between stromal and epithelial cells are important in many steps of tumor progression (Elenbaas and Weinberg, 2001). Stromal and tumor cells interchange numerous growth factors and proteases to activate the adjacent ECM and, in turn, induce the selection and expansion of neoplastic cells (Liotta and Kohn, 2001). Fibroblasts are a major cell type of the stromal compartment, and are intimately involved in orchestrating the stromal part of the dialogue in tissue homeostasis (Grinnell, 1994). Whereas the role of matrix metalloproteinases and urokinase plasminogen activator in the stromal compartment has been documented in various studies (Sieuwerts et al., 1999; Liotta and Kohn, 2001; Singer et al., 2002; Kataoka et al., 2003; Tang et al., 2004), the potential role of cath-D in fibroblasts had not yet been fully determined. It has been proposed that cath-D localized at the surface of breast fibroblasts might be mitogenic (Koblinski et al., 2002) or that intracellular cath-D in fibroblasts might assist cancer cells in the digestion of ECM during tissue invasion (Heylen et al., 2002). Here, we show for the first time that the lysosomal aspartic protease cath-D induces fibroblast invasive growth in a manner independent of its catalytic activity.

The paracrine effect of cath-D secreted by epithelial cancer cells on promoting invasive 3D outgrowth of fibroblasts may have important implications with regard to carcinoma progression. Fibroblasts in stromal tissue surrounding the primary tumor site may become invasive and proliferative in response to secretion of pro-cath-D by tumor cells. Many epithelial tumors need to undergo a distinct epithelial to mesenchymal transition to become fully aggressive and invasive (Vincent-Salomon and Thiery, 2003). A part of this transition may involve activation of stromal fibroblasts (Basset et al., 1990; Olumi et al., 1999; Shekhar et al., 2001). In this respect, in addition to being a marker of poor prognosis in breast cancers (Rochefort, 1992; Ferrandina et al., 1997; Foekens et al., 1999; Westley and May, 1999), cath-D may be a potential target for therapeutic inhibitors of tumor progression.

Invasive growth is generally associated with epithelial cells but had also been attributed to other cell types (Comoglio et al., 1999) and involves proliferation, survival, motility, invasion of extracellular matrices and induction of polarity (Tamagnone and Comoglio, 1997). We demonstrate here that both catalytically active and inactive cath-D stimulate proliferation, survival, motility, and invasion of fibroblasts. The mitogenic activity of cath-D had been widely described for cancer cells (Vignon et al., 1986; Vetvicka et al., 1994; Liaudet et al., 1995; Glondu et al., 2001, 2002) and has been suggested for endothelial cells (Berchem et al., 2002). Our data also clearly highlight the facts that cath-D–deficient fibroblasts undergo apoptosis in matrices and that the expression of catalytically active or inactive cath-D in fibroblasts induced their survival. The anti-apoptotic function of cath-D had been previously suggested from the apoptotic and necrotic phenotypes observed after invalidation of its gene (Saftig et al., 1995; Nakanishi et al., 2001; Koike et al., 2003) and in tumor xenografts (Berchem et al., 2002). We also observed a positive effect of cath-D in the migration and invasive potential of fibroblasts. The involvement of cath-D proteolytic activity in degradation of the ECM, an important step for invasion, is unlikely because cath-D requires an acidic pH to be proteolytically active (Briozzo et al., 1988). Because D231N cath-D was as potent as wild-type cath-D in stimulating invasion, the mechanism responsible seems more likely to implicate either an interaction of cath-D with other factors involved in this process, or an indirect regulation of gene expression. We did not detect any significant change in the mRNA expression levels of a variety of proteases and their inhibitors in transfected fibroblasts (unpublished data), but our preliminary data suggest that cath-D could bind to an inhibitor of one specific class of proteases and therefore increase protease activity. Invasive growth also results from the concomitant activation of ras–MAPK, PI3 kinase, and signal transducer and activator of transcription (Comoglio et al., 1999). Our data suggest a possible link between the biological effects of cath-D in fibroblasts and the MAPK pathway, which is well known for its role in promoting cell proliferation (Bar-Sagi and Hall, 2000). In addition, activation of ras/MEK is frequently associated with up-regulation of PKC, particularly PKCΔ which is known to be associated with changes in cellular motility (Iwabu et al., 2004), another effect seen after cath-D transfection.

The question remains as to how cath-D induces such a major change in fibroblast phenotype. Because fibroblasts can become activated by catalytically active or inactive cath-D or by the pro-cath-D precursor, its proteolytic activity is clearly not directly involved in the stimulatory effect. At present, the only receptor known to interact with secreted pro-cath-D is Man-6-P/IGF2, which has a well-defined function in the transport of various ligands via the endosomal pathway (Clague, 1998). However, the ability of this receptor to stimulate cellular responses via signaling pathways remains controversial. Although a recent study indicates that it may transduce IGF2 mitogenic activity via the MAPK pathway (McKinnon et al., 2001), our results have ruled out the Man-6-P/IGF2 receptor as being the mediator of cath-D induction of fibroblast invasive growth in 3D coculture assays (unpublished data).

Another possibility is that secreted pro-cath-D binds to a yet unidentified cell surface receptor coupled to the MAPK transduction pathway. Indeed it has been proposed that such a receptor exists at the cell surface of cancer and endothelial cells to mediate cath-D mitogenic activity (Vetvicka et al., 1994; Laurent-Matha et al., 1998; Glondu et al., 2001; Berchem et al., 2002).

In conclusion, this study demonstrates that cath-D plays a crucial role for fibroblast outgrowth in 3D matrices. Therefore, we propose that cath-D may favor tumor progression not only by affecting the epithelial compartment but also by promoting fibroblast outgrowth via a paracrine loop. Under these circumstances, cath-D overexpressed and hypersecreted by cancer cells is captured in vivo by stromal cells and may not only promote proliferation and survival, but also stimulate motility and invasion of fibroblasts, and consequently enhance tumor-host homeostasis. If we view the cancer state as a product of its micro-environment, the identification of factors such as cath-D that participate in the tumor-host communication interface activating the host micro-environment is crucial for the development of new stromal therapy.

Materials And Methods

Cell lines, transfections, siRNAs, and growth assays

Cath-D–deficient CD55−/− and CD53+/+ immortalized mouse fibroblasts were provided by C. Peters (University of Freiburg, Freiburg, Germany). Human normal skin CCD45K fibroblasts were purchased from American Type Culture Collection. HMFs, provided by J. Loncarek and J. Piette (CRLC Val d'Aurelle-Paul Lamarque, Montpellier, France), were obtained from reduction mammoplasty tissues from a patient without cancer. Cells were cultured in DME medium with 10% FCS (GIBCO BRL). Cells were cotransfected with a 10-fold excess of wild-type, D231N cath-D, or empty pSG5-neo vector and a pTK-Hyg hygromycin-resistance expression vector (CLONTECH Laboratories, Inc.) using Lipofectamine (GIBCO BRL). Colonies growing in the presence of hygromycin B (100 μg/ml; Invitrogen) were pooled and human cath-D was quantified using an immunoradiometric assay (Glondu et al., 2001). Duplexes of 21-nt human cath-D siRNA (target sequence AAGCUGGUGGACCAGAACAUC, residues 666–684; Bidere et al., 2003) or firefly luc siRNA (target sequence AACGUACGCGGAAUACUUCGA residues 515–534) were synthesized by MWG Biotech S.A.15,000 MCF-7 cells were plated in 24-well plates and after 3 d cells were transiently transfected with 1 μg siRNA using 6 μl Oligofectamine (Invitrogen). The expression (33 μg of protein) and secretion (30-μl medium) of cath-D were monitored by Western blot using 1 μg/ml of anti–human monoclonal cath-D antibody (610801; BD Biosciences). For proliferation assays, 3,000 cells were plated in 24-well plates and then covered with culture medium with 2% FCS. Cells were fixed with methanol and DNA content determined by a diaminobenzoic acid fluorescence assay (Vignon et al., 1986). For outgrowth assay, 100,000 cells were resuspended at 4°C in Matrigel (0.2 ml, 10 mg/ml; Becton and Dickinson), and quickly added to a preset layer of Matrigel in 24-well plates as described previously (Glondu et al., 2001). The top Matrigel layer was gelled at 37°C for 30 min and covered with culture medium containing 10% FCS (0.5 ml). In some experiments, cells were pretreated for 6 h with the MEK1/2 inhibitor, U0126 (10 μM; Cell Signaling Technology, Inc.) and embedded into Matrigel. For treatment of cells embedded into Matrigel, conditioned medium (0.5 ml) supplemented with 5% FCS was directly added after the top Matrigel layer had gelled. To prepare the conditioned media, mock-transfected, cath-D, and D231N transfected 3Y1-Ad12 cancer cells were plated at a 70% density and after 24 h, cells were washed twice with medium without FCS and then cultured with 0% FCS. After 48 h, cells reached 90–100% confluence and conditioned medium was removed and centrifuged at 800 g for 10 min and was finally stored at −80°C. For growth on Matrigel, 100,000 cells were directly plated into 24-well plates on a preset Matrigel layer and then covered with culture medium containing 10% FCS. In 3D Matrigel coculture outgrowth assays, 50,000 MCF-7 cells were resuspended at 4°C in Matrigel (0.2 ml, 10 mg/ml) and then added to a preset layer of Matrigel in 24-well plates. CD55−/−SV40, CCD45K, or HMF fibroblasts were resuspended in 0.2 ml Matrigel, added to the Matrigel layer containing embedded MCF-7 cells and were then covered with culture medium with 10% FCS (0.5 ml), as described previously for collagen I 3D coculture assays (Nakashiro et al., 2000). In other coculture outgrowth assays, 15,000 MCF-7 cells previously transfected with cath-D or luc siRNAs, or 200,000 cath-D transfected 3Y1-Ad12 cell lines were plated in 24-well plates. After 2 d, cells were covered with a layer of Matrigel (0.2 ml), then with a layer of Matrigel (0.2 ml) containing 50,000 CD55−/−SV40 fibroblasts, and finally were covered with culture medium with 10% FCS (0.5 ml). Cells were photographed with an Olympus inverted phase-contrast microscope using an Olympus camera with a 10× objective and negatives were scanned and processed using Adobe Photoshop software.

Wound healing, invasion, and motility assays

For wounding experiments, 500,000 cells were seeded on type I collagen-coated 6-well plates (5 μg/cm2) and cell layers were wounded by tip-scraping. Cells were washed with fresh medium to remove floating cells and refed with fresh medium supplemented with 10% FCS and wound healing was monitored with an inverted phase-contrast microscope (Olympus). The Boyden chamber chemo-invasion and migration assays were performed essentially as described previously (Glondu et al., 2002). For invasion assays, polycarbonate filters (12-μm pore) were coated with 35 μg of Matrigel. Cells were harvested in 4% FCS containing medium and added to the top chamber (600,000 cells/chamber) and 10% FCS containing DME medium was used in the bottom compartment. In some experiments, conditioned medium from transfected 3Y1-Ad12 cells was used in the bottom compartment either lacking pro-cath-D, or containing (30 nM) wild-type pro-cath-D supplemented with 10% FCS. Chambers were incubated for 24 h at 37°C, after which living cells that had traversed the Matrigel and spread on the bottom surface of the filter were quantified by the mitochondrial deshydrogenase enzymatic assay using 3(4,5-dimethyl-thiazol-2-yl)2,5-diphenol tetrazolium bromide and determination of OD540nm. Migration assays were performed as described for the chemo-invasion studies, with filters coated with 5 μg of rat tail collagen I and the cells were incubated in the top chamber in 0% FCS + 0.1% BSA.

Flow cytometry, incorporation of BrdUrd, electron microscopy, and in situ apoptosis

500,000 cells plated onto Matrigel (0.5 ml, 10 mg/ml) in 12-well plates were recovered after 3 d with 2.5 ml Matrisperse (Becton Dickinson), resuspended in 1 ml of a solution containing 0.1% tri-sodium citrate dehydrate, 0.1% Triton X-100, 25 μg/ml propidium iodide and 100 μg/ml RNase, before FACScan analysis (Becton Dickinson) with an argon ion laser turned at 488 nm, 20 mW. Propidium iodide fluorescence was measured at 585 nm. Data were collected and analyzed with Cellquest (Becton Dickinson) and ModFit (Verity Software) softwares, respectively. In other experiments, 60,000 cells plated in 6-well plates in the absence or presence of conditioned media supplemented with 2% FCS containing either no pro-cath-D, 30 nM wild-type, or 10 nM D231N pro-cath-D were analyzed by flow cytometry after 3 d. For staining for incorporated BrdUrd in nuclei, cells cultured on coverslips were labeled with BrdUrd (10 μM) followed by incubation with anti-BrdUrd mAb (Becton Dickinson) as described previously (Nakayasu and Rerezney, 1989). Secondary antibody was an FITC-conjugated anti–mouse IgG (Jackson ImmunoResearch Laboratories). Cells were then incubated with PBS containing DAPI (0.5 μg/ml) and coverslips were mounted on microscopy slides and counted under a fluorescent microscope (Axioplan MC100; Carl Zeiss MicroImaging, Inc.). For electron microscopy, ultra thin sections of cells embedded in Matrigel were classically observed using a JEOL 1200× transmission electron microscope. For in situ apoptosis, cells embedded in Matrigel were incubated for 15 min at 37°C with 10 μM cell permeant Hoechst 33342 (Molecular Probes) in complete DME medium. After three washes with complete DME medium, live cells in complete medium were observed with a DMRA2 microscope (Leica) equipped with a water immersion 63×/0.90 apochromatic objective and a Coolsnap FX cooled CCD camera (Princeton Instruments), both controlled by the MetaMorph imaging software (Universal Imaging Corp).

Immunoprecipitation and endocytosis of cath-D

For cath-D immunoprecipitation, fibroblasts were incubated in 70% methionine-free DME supplemented with 150 μCi/ml [35S]methionine (>1,000 mCi/mmol; Amersham Biosciences) and 10% FCS for 16 h. Labeled proteins (100 ×106 cpm for cell extract and 15 ×106 cpm for medium of total TCA-precipitable proteins) were immunoprecipitated with M1G8 anti–cath-D antibody and analyzed by 10% SDS-PAGE and fluorography (Capony et al., 1987). For cath-D endocytosis, MCF-7 cells were labeled with 200 μCi/ml [35S]methionine for 24 h in DME without methionine and FCS, and labeled conditioned culture medium was used directly for internalization studies. CD55−/− fibroblasts plated in 6-well plates were incubated for 18 h in 1 ml serum-free DME medium supplemented with the conditioned medium prepared from MCF-7 cells corresponding to 3 × 106 cpm of total TCA-precipitable proteins with an excess of nonradioactive methionine (10 mM). After incubation, the medium was discarded and cath-D endocytosed into cells was analyzed by immunoprecipitation (Laurent-Matha et al., 1998).

Western blot analysis of ERK1/2 activation

For ERK1/2 activation, cells were grown to 80% confluency in medium supplemented with 2% FCS for 3 d and ERK1/2 activation was analyzed with either an anti–phospho-p44/42 MAPK antibody (Cell Signaling Technology) or an anti-ERK2 antibody (sc-1647; Santa Cruz Biotechnology, Inc.).

Online supplemental material

Fig. S1 shows the paracrine action of pro-cath-D on fibroblasts in an aortic ring. Fig. S2 shows the characterization of human wild-type and D231N mutated cath-D in CD55−/− transfected fibroblasts.

Acknowledgments

We thank Dr. Christoph Peters for the kind gift of CD53+/+ and CD55−/− fibroblast cell lines, Dr. Ned Lamb for discussions and suggestions, Christophe Duperray for FACScan experiments (INSERM, Montpellier), Jean-Yves Cance for photographs, and Nadia Kerdjadj for secretarial assistance.

This work was supported by University of Montpellier I, Institut National de la Santé et de la Recherche Médicale, Association pour la Recherche sur le Cancer (grant 3344), and Ligue Nationale contre le Cancer for a fellowship to Mélanie Beaujouin.

References

References
Bar-Sagi, D., and A. Hall.
2000
. Ras and Rho GTPases: a family reunion.
Cell
103
:
227
–238.
Basset, P., J.P. Bellocq, C. Wolf, I. Stoll, P. Hutin, J.M. Limacher, O. Podhajcer, M.P. Chenard, M.C. Rio, and P. Chambon.
1990
. A novel metalloproteinase gene specifically expressed in stromal cells of breast carcinomas.
Nature.
348
:
699
–704.
Berchem, G.J., M. Glondu, M. Gleizes, J.P. Brouillet, F. Vignon, M. Garcia, and E. Liaudet-Coopman.
2002
. Cathepsin-D affects multiple steps of tumor progression: Proliferation, angiogenesis and apoptosis.
Oncogene.
21
:
5951
–5955.
Bidere, N., H.K. Lorenzo, S. Carmona, M. Laforge, F. Harper, C. Dumont, and A. Senik.
2003
. Cathepsin D triggers Bax activation, resulting in selective apoptosis-inducing factor (AIF) relocation in T lymphocytes entering the early commitment phase to apoptosis.
J. Biol. Chem.
278
:
31401
–31411.
Bisson, C., S. Blacher, M. Polette, J.F. Blanc, F. Kebers, J. Desreux, B. Tetu, J. Rosenbaum, J.M. Foidart, P. Birembaut, and A. Noel.
2003
. Restricted expression of membrane type 1-matrix metalloproteinase by myofibroblasts adjacent to human breast cancer cells.
Int. J. Cancer.
105
:
7
–13.
Briozzo, P., M. Morisset, F. Capony, C. Rougeot, and H. Rochefort.
1988
. In vitro degradation of extracellular matrix with Mr 52,000 cathepsin D secreted by breast cancer cells.
Cancer Res.
48
:
3688
–3692.
Capony, F., M. Morisset, A.J. Barrett, F. Capony, P. Broquet, F. Vignon, M. Chambon, P. Louisot, and H. Rochefort.
1987
. Phosphorylation, glycosylation, and proteolytic activity of the 52-kD estrogen-induced protein secreted by MCF7 cells.
J. Cell Biol.
104
:
253
–262.
Capony, F., T. Braulke, C. Rougeot, S. Roux, P. Montcourrier, and H. Rochefort.
1994
. Specific mannose-6-phosphate receptor-independent sorting of pro-cathepsin D in breast cancer cells.
Exp. Cell Res.
215
:
154
–163.
Clague, M.J.
1998
. Molecular aspects of the endocytic pathway.
Biochem. J.
336
:
271
–282.
Comoglio, P.M., L. Tamagnone, and C. Boccaccio.
1999
. Plasminogen-related growth factor and semaphorin receptors: a gene superfamily controlling invasive.
Exp. Cell Res.
253
:
88
–99.
Elenbaas, B., and R.A. Weinberg.
2001
. Heterotypic signaling between epithelial tumor cells and fibroblasts in carcinoma formation.
Exp. Cell Res.
264
:
169
–184.
Ferrandina, G., G. Scambia, F. Bardelli, B. Panici, S. Mancuso, and A. Messori.
1997
. Relationship between cathepsin-D content and disease-free survival in node-negative breast cancer patients: a meta-analysis.
Br. J. Cancer.
76
:
661
–666.
Foekens, J.A., M.P. Look, J. Bolt-de Vries, M.E. Meijer-van Gelder, W.L.J. van Putten, and J.G.M. Klijn.
1999
. Cathepsin-D in primary breast cancer: prognostic evaluation involving 2810 patients.
Br. J. Cancer.
79
:
300
–307.
Garcia, M., D. Derocq, P. Pujol, and H. Rochefort.
1990
. Overexpression of transfected cathepsin D in transformed cells increases their malignant phenotype and metastatic potency.
Oncogene.
5
:
1809
–1814.
Glondu, M., P. Coopman, V. Laurent-Matha, M. Garcia, H. Rochefort, and E. Liaudet-Coopman.
2001
. A mutated cathepsin-D devoid of its catalytic activity stimulates the growth of cancer cells.
Oncogene.
20
:
6920
–6929.
Glondu, M., E. Liaudet-Coopman, D. Derocq, N. Platet, H. Rochefort, and M. Garcia.
2002
. Down-regulation of cathepsin-D by antisense gene transfer inhibits tumor growth and metastatic potential of human breast cancer cells.
Oncogene.
21
:
5127
–5134.
Grinnell, F.
1994
. Fibroblasts, myofibroblasts, and wound contraction.
J. Cell Biol.
124
:
401
–404.
Heylen, N., L.M. Vincent, V. Devos, V. Dubois, C. Remacle, and A. Trouet.
2002
. Fibroblasts capture cathepsin D secreted by breast cancer cells: possible role in the regulation of the invasive process.
Int. J. Oncol.
20
:
761
–767.
Iwabu, A., K. Smith, F.D. Allen, D.A. Lauffenburger, and A. Wells.
2004
. Epidermal growth factor induces fibroblast contractility and motility via a PKCdelta -dependent pathway.
J. Biol. Chem.
279
:
14551
–14560.
Kataoka, H., H. Tanaka, K. Nagaike, S. Uchiyama, and H. Itoh.
2003
. Role of cancer cell-stroma interaction in invasive growth of cancer cells.
Hum. Cell.
16
:
1
–14.
Koblinski, J.E., J. Dosescu, M. Sameni, K. Moin, K. Clark, and B.F. Sloane.
2002
. Interaction of human breast fibroblasts with collagen I increased secretion of procathepsin B.
J. Biol. Chem.
277
:
32220
–32227.
Koike, M., M. Shibata, Y. Ohsawa, H. Nakanishi, T. Koga, S. Kametaka, S. Waguri, T. Momoi, E. Kominami, C. Peters, et al.
2003
. Involvement of two different cell death pathways in retinal atrophy of cathepsin D-deficient mice.
Mol. Cell. Neurosci.
22
:
146
–161.
Laurent-Matha, V., M.R. Farnoud, A. Lucas, C. Rougeot, M. Garcia, and H. Rochefort.
1998
. Endocytosis of pro-cathepsin D into breast cancer cells is mostly independent of mannose-6-phosphate receptors.
J. Cell Sci.
111
:
2539
–2549.
Liaudet, E., M. Garcia, and H. Rochefort.
1994
. Cathepsin D maturation and its stimulatory effect on metastasis are prevented by addition of KDEL retention signal.
Oncogene.
9
:
1145
–1154.
Liaudet, E., D. Derocq, H. Rochefort, and M. Garcia.
1995
. Transfected cathepsin D stimulates high density cancer cell growth by inactivating secreted growth inhibitors.
Cell Growth Differ.
6
:
1045
–1052.
Liotta, L.A., and E.C. Kohn.
2001
. The microenvironment of the tumor-host interface.
Nature.
411
:
375
–379.
Masson, R., O. Lefebvre, A. Noel, M.E. Fahime, M.P. Chenard, C. Wendling, F. Kebers, M. LeMeur, A. Dierich, J.M. Foidart, et al.
1998
. In vivo evidence that the stromelysin-3 metalloproteinase contributes in a paracrine manner to epithelial cell malignancy.
J. Cell Biol.
140
:
1535
–1541.
McKinnon, T., C. Chakraborty, L.M. Gleeson, P. Chidiac, and P.K. Lala.
2001
. Stimulation of human extravillous trophoblast migration by IGF-II is mediated by IGF type 2 receptor involving inhibitory G protein(s) and phosphorylation of MAPK.
J. Clin. Endocrinol. Metab.
86
:
3665
–3674.
Metcalf, P., and M. Fusek.
1993
. Two crystal structures for cathepsin D: the lysosomal targeting signal and active site.
EMBO J.
12
:
1293
–1302.
Nakanishi, H., J. Zhang, M. Koike, T. Nishioku, Y. Okamoto, E. Kominami, K. Von Figura, C. Peters, K. Yamamoto, P. Saftig, and Y. Uchiyama.
2001
. Involvement of nitric oxide released from microglia-macrophages in pathological changes of cathepsin D-deficient mice.
J. Neurosci.
21
:
7526
–7533.
Nakashiro, K., M. Okamoto, Y. Hayashi, and R. Oyasu.
2000
. Hepatocyte growth factor secreted by prostate-derived stromal cells stimulates growth of androgen-independent human prostatic carcinoma cells.
Am. J. Pathol.
157
:
795
–803.
Nakayasu, H., and R. Rerezney.
1989
. Mapping replicational sites in the eucaryotic cell nucleus.
J. Cell Biol.
108
:
1
–11.
Olumi, A.F., G.D. Grossfeld, S.W. Hayward, P.R. Caroll, T.D. Tslty, and G.R. Cunha.
1999
. Carcinoma-associated fibroblasts direct tumor progression of initiated human prostatic epithelium.
Cancer Res.
59
:
5002
–5011.
Pupa, S.M., S. Ménard, S. Forti, and E. Tagliabue.
2002
. New insights into the role of extracellular matrix during tumor onset and progression.
J. Cell. Physiol.
192
:
259
–267.
Richo, G., and G.E. Conner.
1991
. Proteolytic activation of human procathepsin D.
Adv. Exp. Med. Biol.
306
:
289
–296.
Rochefort, H.
1992
. Cathepsin D in breast cancer: a tissue marker associated with metastasis.
Eur. J. Cancer.
28A
:
1780
–1783.
Rochefort, H., and E. Liaudet-Coopman.
1999
. Cathepsin D in cancer metastasis: a protease and a ligand.
APMIS.
107
:
86
–95.
Rochefort, H., F. Capony, M. Garcia, V. Cavaillès, G. Freiss, M. Chambon, M. Morisset, and F. Vignon.
1987
. Estrogen-induced lysosomal proteases secreted by breast cancer cells: a role in carcinogenesis?
J. Cell. Biochem.
35
:
17
–29.
Saftig, P., M. Hetman, W. Schmahl, K. Weber, L. Heine, H. Mossmann, A. Koster, B. Hess, M. Evers, K. Von Figura, and C. Peters.
1995
. Mice deficient for the lysosomal proteinase cathepsin D exibit progressive atrophy of the intestinal mucosa and profound destruction of lymphoid cells.
EMBO J.
14
:
3599
–3608.
Schlessinger, J.
2000
. Cell signaling by receptor tyrosine kinases.
Cell
103
:
211
–225.
Shekhar, M.P., J. Werdell, S.J. Santner, R.J. Pauley, and L. Tait.
2001
. Breast stroma plays a dominant regulatory role in breast epithelial growth and differentiation: implications for tumor development and progression.
Cancer Res.
61
:
1320
–1326.
Sieuwerts, A.M., J.G. Klijn, S.C. Henzen-Logmans, and J.A. Foekens.
1999
. Cytokine-regulated urokinase-type-plasminogen-activator (uPA) production by human breast fibroblasts in vitro.
Breast Cancer Res. Treat.
55
:
9
–20.
Singer, C.F., N. Kronsteiner, E. Marton, M. Kubista, K.J. Cullen, K. Hirtenlehner, M. Seifert, and E. Kubista.
2002
. MMP-2 and MMP-9 expression in breast cancer-derived human fibroblasts is differentially regulated by stromal-epithelial interactions.
Breast Cancer Res. Treat.
72
:
69
–77.
Tamagnone, L., and P.M. Comoglio.
1997
. Control of invasive growth by hepatocyte growth factor (HGF) and related scatter factors.
Cytokine Growth Factor Rev.
8
:
129
–142.
Tang, Y., P. Kesavan, M.T. Nakada, and L. Yan.
2004
. Tumor-stroma interaction: positive feedback regulation of extracellular matrix metalloproteinase inducer (EMMPRIN) expression and matrix metalloproteinase-dependent generation of soluble EMMPRIN.
Mol. Cancer Res.
2
:
73
–80.
Tlsty, T.D., and P.W. Hein.
2001
. Stromal cells can contribute oncogenic signals.
Curr. Opin. Genet. Dev.
11
:
54
–59.
Vetvicka, V., J. Vektvickova, and M. Fusek.
1994
. Effect of human cathepsin D on proliferation of human cell lines.
Cancer Lett.
79
:
131
–135.
Vignon, F., F. Capony, M. Chambon, G. Freiss, M. Garcia, and H. Rochefort.
1986
. Autocrine growth stimulation of the MCF 7 breast cancer cells by the estrogen-regulated 52 K protein.
Endocrinology.
118
:
1537
–1545.
Vincent-Salomon, A., and J.P. Thiery.
2003
. Host microenvironment in breast cancer development: epithelial-mesenchymal transition in breast cancer development.
Breast Cancer Res.
5
:
101
–106.
Von Figura, K., and A. Hasilik.
1986
. Lysosomal enzymes and their receptors.
Annu. Rev. Biochem.
55
:
167
–193.
Westley, B.R., and F.E.D. May.
1999
. Prognostic value of cathepsin D in breast cancer.
Br. J. Cancer.
79
:
189
–190.

V. Laurent-Matha and S. Maruani-Hermann contributed equally to this work.

Sharon Maruani-Herrmann's present address is Dina Raveh's laboratory, Life Science Dept., Ben Gourion University of the Negev, Beer-Sheva 84105, Israel.

V. Laurent-Matha's present address is Cellular Biology Unit, IGH, CNRS UPR 1142, 34396 Montpellier, Cedex 05, France.

Abbreviations used in this paper: 3D, three-dimensional; cath-D, cathepsin D; HMF, human mammary fibroblast; luc siRNA, luciferase siRNA; Man-6-P, Mannose-6-phosphate.

Supplementary data