Membrane-type-1 matrix metalloproteinase confers tumorigenicity on nonmalignant epithelial cells

Abstract

Overexpression of membrane-type-1 matrix metalloproteinase (MT1-MMP) in tumor cells has previously been shown to enhance tumor growth and metastasis. To establish if MT1-MMP is also able to confer tumorigenicity on nonmalignant epithelial cells, we transfected human MT1-MMP cDNA into Madin-Darby canine kidney (MDCK) cells expressing a tetracycline-repressible transactivator. Induction of MT1-MMP in the absence of doxycycline (Dox) was associated with activation of exogenous MMP-2 as well as with formation of large cysts and increased invasiveness in collagen matrices. Transfected cells were inoculated subcutaneously into two groups of nude mice, one of which received Dox to inhibit expression of MT1-MMP. Formation of tumor xenografts was observed in 11 of 17 mice maintained without Dox, but only in two of nine mice that received Dox (P<0.05). The xenografts were composed of tubular structures interspersed within a highly cellular stroma. The epithelial cells delimiting the lumen were polarized, as indicated by the basolateral distribution of Na,K-ATPase. Despite their differentiated appearance, the tumors lacked a well-defined boundary, and epithelial tubules invaded adjacent muscular layers. These results demonstrate that conditional expression of MT1-MMP in nonmalignant MDCK epithelial cells is by itself sufficient to drive formation of invasive tumors.

Introduction

The multistep process of carcinogenesis involves both genetic alterations (i.e. mutations that result in activation of oncogenes or inactivation of tumor suppressor genes) and epigenetic events (e.g. deregulated interaction of tumor cells with the stromal microenvironment) (Liotta and Kohn, 2001). A fundamental property that distinguishes malignant from benign tumor cells is the former's ability to cross structural barriers formed by basement membranes and interstitial matrix, to invade contiguous tissues, and ultimately to spread to distant organs (Mareel and Leroy, 2003). The invasive capacity of malignant cells relies on focal extracellular proteolysis and is largely mediated by matrix metalloproteinases (MMPs) (Sternlicht and Werb, 2001; Stetler-Stevenson and Yu, 2001), a family of zinc-dependent endopeptidases that, collectively, are able to degrade virtually all proteins of the extracellular matrix (ECM) and also cleave several surface-associated proteins (Egeblad and Werb, 2002; Lynch and Matrisian, 2002).

MMPs can be subdivided into two groups, secreted MMPs and membrane-type MMPs (MT-MMPs). MT-MMPs display the common structural domains of the MMP family, but are distinguished by the presence of an additional membrane-anchoring domain (Sato et al., 1994; Seiki, 2003; Zucker et al., 2003). Membrane-tethered MMPs, rather than soluble MMPs, are ideally positioned for regulating pericellular proteolysis. Among the six MT-MMPs described so far, MT1-MMP plays a critical role in ECM turnover owing to its ability to degrade various ECM components, either directly or via the activation of other MMPs such as MMP-2 or MMP-13 (Strongin et al., 1995; Weiss and Pei, 1996; Ohuchi et al., 1997; Seiki, 2003). In addition, through the proteolytic modification of cell surface-associated receptors, MT1-MMP can exert a long-lasting effect on cell behavior. Indeed, MT1-MMP is upregulated in several types of tumors and has been implicated in tumor progression and metastasis (Vihinen and Kahari, 2002; Yana and Seiki, 2002; Seiki, 2003). In particular, the overexpression of MT1-MMP in tumor cells has been shown to enhance tumor growth and metastatic spread (Tsunezuka et al., 1996; Ha et al., 2001; Deryugina et al., 2002b; Sounni et al., 2002; Hotary et al., 2003). It was not known, however, if overexpression of MT1-MMP in a well-differentiated, normally nontumorigenic cell line was sufficient, on its own, to elicit tumor formation. The current study was designed to specifically address this question. Using a xenograft model in immunodeficient nu/nu mice, we report that inoculation of Madin-Darby canine kidney (MDCK) cells that ectopically express MT1-MMP results in the formation of well-differentiated and locally invading tumors.

Results

Conditional expression of MT1-MMP alters the morphogenetic and invasive properties of MDCK cells

MDCK Tet-Off cells, which express the tTA tetracycline-repressible transactivator (Gossen and Bujard, 1992), were used to generate transfected cell lines in which the expression of MT1-MMP can be induced by withdrawing the tetracycline derivative doxycycline (Dox) from the culture medium. Induction of the MT1-MMP gene in the absence of Dox was assessed by Northern blot in several clones of transfected MDCK Tet-Off cells (not shown). Clone MT1-G, which exhibited robust induction of MT1-MMP in the absence of Dox (Figure 1a) was selected for further analysis and used throughout this study.

Figure 1
figure1

Conditionally expressed MT1-MMP has catalytic activity. (a) Northern blot analysis of MT1-MMP expression in MDCK Tet-Off cells and their transfected derivatives. Confluent monolayers were incubated with or without 2 μg/ml Dox for 48 h, after which RNA samples were prepared and hybridized with an MT1-MMP cRNA probe, as described in Materials and methods. A bovine P0 ribosomal phosphoprotein cRNA probe was used as an internal control for determining the amount of RNA loaded. The weaker P0 bands in MT1-G +dox and MT1-G -dox lanes resulted from an inadvertent nonuniform mRNA loading. MDCK-MT1-G cells express MT1-MMP in the absence of Dox. (b) Activation of exogenous MMP-2 by conditionally expressed MT1-MMP. Confluent monolayers of MT1-G cells were incubated with or without 2 μg/ml Dox for 24 h, at which time the monolayer was washed in serum-free medium and further incubated with 400 μl serum-free MRC-5 conditioned medium. After 3, 6, 9 and 24 h, the conditioned medium was collected and processed for gelatinolytic zymography as described in Materials and methods. Conditioned medium incubated with MT1-G cells grown in the presence of Dox shows bands of gelatin lysis corresponding to the higher molecular weight inactive form of MMP-2 (72 kDa) as well as a faint band corresponding to the lower molecular weight intermediate species of MMP-2 at 24 h (upper panel). In contrast, bands corresponding to both intermediate and fully active forms of MMP-2 are prominent in conditioned medium incubated for 24 h with MT1-G cells grown in the absence of Dox (lower panel)

To verify that conditionally expressed MT1-MMP possesses catalytic activity, we evaluated its ability to activate exogenous MMP-2. To this end, conditioned medium from MRC-5 cells, which contains latent MMP-2 (Reno et al., 2004), was incubated for 3–24 h with MT1-G cells maintained in the presence or absence of Dox and was subsequently subjected to zymographic analysis. Conditioned medium incubated for 3–9 h with MT1-G cells grown in the presence of Dox showed only a band corresponding to the inactive form of MMP-2. A faint band of gelatin lysis corresponding to the lower molecular weight intermediate species of MMP-2 appeared after 24 h of incubation (Figure 1b). In contrast, conditioned medium incubated with MT1-G cells grown in the absence of Dox showed already after 3 h a clear-cut band of lysis corresponding to the intermediate form of MMP-2. Of note, an additional band corresponding to the fully activated species of MMP-2 appeared after 24 h of incubation (Figure 1b). These findings provide evidence that transfected MT1-MMP possesses catalytic activity.

The biological effects of MT1-MMP expression were investigated by growing MDCK cell lines in collagen gels. When suspended in a collagen gel in the presence of Dox, MT1-G cells formed tiny spherical cysts (Figure 2a), similar to those observed with wild-type MDCK cells (Montesano et al., 1991b). In contrast, in the absence of Dox, MT1-G cells generated large, irregularly shaped cystic structures within the matrix (Figure 2b). These morphological changes were accompanied by an approximately four-fold increase in cell number in the gels when compared to MT1-G cells maintained in the presence of Dox (Table 1). Addition of the hydroxamate MMP inhibitor BB94 to collagen gel cultures abrogated both the formation of ectatic cysts (Figure 2c) and the increase in cell number (Table 1). In agreement with previous studies using MDCK cells constitutively transfected with MT1-MMP (Kadono et al., 1998b; Hotary et al., 2003), these findings show that conditional expression of MT1-MMP enhances the ability of MDCK cells to degrade, and to proliferate within, collagen matrices. They also indicate that these effects are dependent on the catalytic activity of MT1-MMP.

Figure 2
figure2

Expression of MT1-MMP modifies the in vitro properties of MDCK cells. (a) MT1-G cells grown in a collagen gel in the presence of Dox form small spherical cysts. (b) In the absence of Dox, the cells form large, irregularly shaped cystic structures. This effect is completely prevented by addition of 1 μ M BB94 (c), but not by addition of the related inactive isomer BB1268 (not shown). (d) MT1-G cells seeded atop of a collagen gel cast on an underlying gel layer populated by Swiss 3T3 fibroblasts (1 × 105 cells/ml) and grown for 10 days in the presence of Dox focally invade the collagen matrix, within which they form a few, short tube-like structures. (e) When co-cultured with 3T3 cells for 10 days in the absence of Dox, MT1-G cells extensively invade the underlying gel, within which they form long tubes enclosing a patent lumen. Formation of invading tubes by cells grown in the absence of Dox is suppressed by the addition of 2 μ M BB94 (f), but not by addition of the related inactive isomer BB1268 (data not shown). Three independent co-culture experiments performed with concentrations of 3T3 cells ranging from 1 × 105 to 4 × 105 per ml of collagen yielded similar results

Table 1 Expression of MT1-MMP stimulates MT1-G cell proliferation in three-dimensional collagen gels

To assess the impact of MT1-MMP expression on cell invasiveness, MT1-G cells were seeded onto the surface of a collagen gel cast atop of an underlying gel layer containing 3T3 fibroblasts. We have shown previously that in this experimental setting, fibroblast-derived hepatocyte growth factor (HGF) induces MDCK cells to penetrate the collagen gel, within which they form branching tubes (Montesano et al., 1991a,1991b). After 6–10 days of co-culture in the presence of Dox, only a few, narrow tube-like structures were seen to invade the underlying matrix (Figure 2d). In contrast, in co-cultures maintained in the absence of Dox during the same time period, MT1-G cells extensively invaded the collagen gel, within which they formed long tubes enclosing a wide lumen (Figure 2e). The formation of invading tubes was abrogated by addition of BB94 (Figure 2f), but not by addition of the related inactive isomer BB1268 (data not shown). In accord with previous studies in which MT1-MMP was induced in MDCK cells using a constitutive expression system (Hotary et al., 2000), these findings indicate that expression of MT1-MMP markedly enhances the ability of MDCK cells to invade a collagen matrix, and that this effect is dependent on the catalytic activity of the enzyme.

Conditional expression of MT1-MMP in MDCK cells promotes tumor formation in nude mice

MDCK cells are a spontaneously immortalized cell line derived from a normal dog kidney, which retain several differentiated features of renal tubular epithelium (Rabito et al., 1978; Rodriguez-Boulan and Nelson, 1989). As observed with most immortalized but nontransformed cell lines, MDCK cells do not form tumors when injected subcutaneously in adult nude mice (Stiles et al., 1976; U et al., 1985; Kadono et al., 1998a). To assess if the gain of invasive properties conferred on MDCK Tet-Off cells by MT1-MMP might be sufficient to sustain tumor formation, we evaluated the ability of MDCK cell lines to form tumor nodules after subcutaneous inoculation into nude mice.

As shown in Table 2, out of 17 mice that were injected with MT1-G cells but did not receive Dox, 11 developed tumor nodules after a latency period of 4–5 months. In contrast, over the same time period, tumor xenografts appeared in only two out of the nine mice that were injected with MT1-G cells and received Dox (P<0.05). Injection of mock-transfected MDCK Tet-Off cells (MDCK-Hygro cells) resulted in tumor xenograft formation after 5 months in one of 11 injected mice. The tumors formed in mice that did not receive Dox were scirrhous in nature and showed a characteristic retraction of the overlying skin (data not shown).

Table 2 Number of mice that developed a tumor after injection of MDCK cells

The tumor xenografts were composed of well-organized tubular structures, with a central lumen delimited by a single layer of epithelial cells (Figure 3a,b). Nuclear abnormalities typical of malignant tumors were not observed. The epithelial tubes were interspersed within an abundant, highly cellular stroma, which was clearly distinguishable from the surrounding, more fibrous subcutaneous connective tissue (Figure 3a, b). The stroma of a subset of tumors also contained focal collections of thin-walled venules (not shown), suggesting the local induction of angiogenesis. Contrary to what is observed in benign adenomas and many transplantable tumors, the xenografts were not delimited by a fibrous capsule. In addition, numerous epithelial tubes infiltrated the adjacent muscular layer (Figure 3c). The tubes within the muscular layer were often in continuity with those present within the primary, subcutaneous tumor mass (data not shown). This argues against proliferation of cells injected into the muscular layer and supports the occurrence of an active invasion process from the subcutaneous tumor.

Figure 3
figure3

Histology of the tumors formed by subcutaneously injected MDCK-MT1-G cells. (a) At low magnification, the major portion of the tumor mass appears to be localized between the subcutaneous adipose tissue and the underlying muscular layer, which is recognizable by its intensely eosinophilic fibers. The tumor consists of tubular structures enclosing a wide central cavity. (b) At higher magnification, the epithelial tubules appear to be dispersed within an abundant and highly cellular stroma. (c) Details of epithelial tubules that have invaded the muscular layer. Well-differentiated tubular structures are intermingled with the fibers of striated muscle

To assess if MDCK cells maintained differentiated properties in tumor xenografts, we applied antibodies to Na,K-ATPase, an enzyme that is selectively expressed on the basolateral membrane of most epithelial cell types (Dunbar and Caplan, 2001; Feraille and Doucet, 2001). Immunoreactivity for Na,K-ATPase was clearly restricted to the basolateral surface of tumoral MDCK cells (Figure 4), even in tubes that had invaded contiguous muscular tissue, indicating that MDCK-MT1 cells have retained a well-defined polarity. As the antibody we used does not crossreact with murine Na,K-ATPase, these results also confirm the exogenous origin of the tumor cells. Despite the well-differentiated appearance of the tumor, MDCK cell aggregates were detected in the lumen of intra- or peritumoral lymphatic vessels (Figure 5a) and venules (not shown). The epithelial nature and exogenous origin of the intralymphatic cell aggregates was confirmed by Na,K-ATPase staining (Figure 5b). In addition, the observation of epithelial aggregates apparently in the process of crossing the endothelial lining (not shown) suggested that penetration of tumor cells into the vascular compartment had occurred by active invasion rather than by passive uptake.

Figure 4
figure4

Tumor cells retain a well-defined polarity. Immunoperoxydase staining of Na,K-ATPase is restricted to the basolateral membrane of epithelial cells. Inset: × 600

Figure 5
figure5

(a) Longitudinal section of a collecting lymph vessel at the tumor periphery containing epithelial cell aggregates (arrows), two of which are closely associated with a valvular structure (V). (b) In a consecutive section, the plasma membrane of the epithelial cells shows strong Na,K-ATPase immunoreactivity

Sections of subcutaneous tissue adjacent to the site of cell inoculation in a mouse that received Dox and that did not develop a tumor revealed rare tubular structures scattered within a dense connective tissue (not shown). This indicates that under conditions that are not permissive for MT1-MMP expression, injected MDCK-MT1 cells can survive and form well-organized multicellular structures, yet are unable to generate a tumor, as defined by local overgrowth and invasion of adjacent tissues.

The expression of MT1-MMP in the tumor xenografts was assessed by RT–PCR using specific primers that recognize the exogenous (human) but not the canine or murine MT1-MMP gene (Figure 6a). MT1-MMP protein was also visualized in tumor cells by immunostaining with antibodies against MT1-MMP (Figure 6b). Although we did not detect expression of MMP-2 in MT1-G cells (data not shown), induction of MMP-2 may nonetheless have occurred in the in vivo environment. We therefore assessed the potential expression of MMP-2 in the tumors by RT–PCR using specific primers that recognize the canine but not the murine MMP-2 gene. In none of the seven tumors examined, expression of MMP-2 was detected (data not shown).

Figure 6
figure6

MT1-MMP expression in tumors. (a) PCR demonstrates expression of exogenous human MT1-MMP in tumors formed by MT1-G cells in four mice (no. 1, 2, 7 and 11) that did not receive Dox. Two additional tumors were analysed and found to express exogenous MT1-MMP (not shown). MT1-G cells cultivated in the absence of Dox were used as a positive control, while J3B1A, a cell line that expresses endogenous murine MT1-MMP, was used as a negative control. (b) Immunoperoxydase staining of MT1-MMP. Both the epithelial tubes located within the major portion of the tumor (upper panel) and those invading the muscular tissue (lower panel) show immunoreactivity for MT1-MMP

The finding that the tumor cells exhibit invasive properties and are able to enter lymphatic and blood vessels, led us to assess the potential occurrence of metastases. Visual inspection of lung, liver and lymph nodes did not reveal macroscopic metastases. Micrometastases were therefore searched for in lung, liver, brain, and lymph nodes by: (a) PCR amplification of the ampicillin resistance gene in the pTRE-MT1 plasmid, and (b) RT–PCR amplification of the human MT1-MMP gene. No signal was detected, suggesting either that intravasated cells did not reach the organs analysed, or that their number was below the sensitivity threshold of the PCR technique (1 : 106 cells for RT–PCR, data not shown).

Discussion

We have demonstrated that conditional expression of MT1-MMP in nontumorigenic MDCK epithelial cells is by itself sufficient to drive tumor formation in nude mice. Despite their ability to form well differentiated and correctly polarized tubular structures, tumor cells are endowed with a marked propensity to invade adjacent tissues.

Recent work has shown that overexpression of MT1-MMP in malignant, intrinsically tumorigenic cells enhances tumor growth and metastasis (Tsunezuka et al., 1996; Ha et al., 2001; Deryugina et al., 2002b; Sounni et al., 2002; Hotary et al., 2003). To our knowledge, the potential impact of MT1-MMP on the in vivo behavior of nontumorigenic epithelial cells has not been previously investigated. The finding that transgenic mice overexpressing MT1-MMP in the mammary gland develop adenocarcinomas (Ha et al., 2001), however, suggested the involvement of MT1-MMP in early stages of tumor evolution. Using a different approach, our present study complements those observations and strongly supports the hypothesis that MT1-MMP plays a role in tumor initiation. A similar property has previously been attributed to MMP-3 (stromelysin-1), whose expression in the mammary gland is sufficient to induce tumorigenesis (Sternlicht et al., 1999,2000).

In addition to cleaving ECM components, MT1-MMP activates other MMP, such as proMMP-2 (gelatinase A) and proMMP-13 (procollagenase 3) (Strongin et al., 1995; Weiss and Pei, 1996; Ohuchi et al., 1997). As we found that MT1-MMP-transduced MDCK cells do not produce MMP-2 and that expression of canine MMP-2 gene is also undetectable in the tumors, activation of endogenous MMP-2 is unlikely to mediate MT1-MMP-induced tumorigenesis. It cannot be excluded, however, that MDCK-derived MT1-MMP activates MMP-2 potentially present in mouse serum or produced by stromal cells in the tumor microenvironment.

Cancer cells can use different migration strategies to disseminate from the primary tumor (Friedl and Wolf, 2003). MT1-MMP-expressing MDCK cells appear to invade adjacent tissues, and particularly muscular layers, in the form of well-organized tubular structures (‘co-ordinated’ or ‘cohort’ migration) rather than as individual cells. During their evolution towards malignancy, epithelial tumors progressively lose histotypic organization, cell polarity and differentiated features (a pathological process that is referred to as anaplasia), while at the same time acquiring the ability to cross anatomical barriers and to disseminate to distant organs (invasiveness). An intriguing finding of our study is that MT1-MMP-induced tumors retain near-normal tissue-specific architecture, yet lack other features of benign adenomas, such as the presence of a well-defined capsule, and display a pronounced invasive capacity. The fact that MT1-MMP-expressing MDCK cells are able to transgress histological boundaries in the absence of major alterations in cell polarity and morphogenetic competence indicates that anaplasia and invasiveness are independently regulated processes that can be experimentally dissociated. An additional notable feature of MT1-MMP-induced tumors is their scirrhous appearance, suggestive of a desmoplastic stromal reaction. Although the mechanisms underlying this phenomenon have not been investigated, it is possible that they involve the release and/or activation of profibrotic cytokines by MT1-MMP.

Although tumor cells entered lymphatic and blood vessels, and were therefore potentially able to disseminate to distant sites, metastases were not observed in distant organs and regional lymph nodes. A likely explanation for this finding stems from the idea that metastasis is a complex process resulting from the accumulation of multiple genetic lesions. Our study shows that expression of MT1-MMP in nontumorigenic, well-differentiated epithelial cells is sufficient to activate cellular responses involved in tumor initiation, local tissue invasion and intravasation. In the absence of additional oncogenic lesions, however, MT1-MMP may not be able on its own to stimulate subsequent events that are required for the accomplishment of metastasis, such as survival in the circulation, adhesion to the endothelium and/or extravasation.

The mechanisms responsible for MT1-MMP-induced tumor formation are not known. MMPs were initially thought to facilitate invasion and metastasis by releasing the mechanical restraints imposed on tumor cells by the surrounding ECM. It is now well documented, however, that MMPs can cleave a host of nonmatrix substrates, thereby contributing to virtually all stages of cancer evolution (Egeblad and Werb, 2002; Stamenkovic, 2003). In light of these notions, several mutually nonexclusive mechanisms of MT1-MMP-induced tumorigenesis can be hypothesized. First, MT1-MMP may facilitate the invasion and colonization of the connective tissue stroma contiguous to the inoculation site, thereby increasing the chances of injected cells to survive and proliferate in the host microenvironment. MT1-MMP may exert such a proinvasive effect by degrading structural components of the ECM (Strongin et al., 1995; Weiss and Pei, 1996; Ohuchi et al., 1997), by processing cell surface receptors for ECM molecules such as αvβ3 integrin, CD44, and tissue transglutaminase (Deryugina et al., 2000,2002a; Belkin et al., 2001; Kajita et al., 2001), or by unmasking cryptic sites within ECM proteins that promote cell motility (Koshikawa et al., 2000; Schenk and Quaranta, 2003; Udayakumar et al., 2003). A second possibility is that the expression of MT1-MMP induces tumorigenesis by stimulating MDCK cell growth. This possibility is supported by the finding that MT1-MMP upregulates cyclin D3 kinase activity (Hotary et al., 2003) and increases the proliferation of MDCK cells grown either within (Hotary et al., 2003, and this study) or on the surface of a three-dimensional type I collagen matrix (our unpublished data). MT1-MMP may also promote proliferation in the in vivo environment by cleaving membrane-bound growth factor precursors, by releasing growth factors sequestered by the ECM, or by degrading growth factor-binding proteins (Stamenkovic, 2003). A third mechanism by which MT1-MMP may promote tumor formation is the induction of angiogenesis. Based on our observation of abundant venule-like vessels in some tumors, and the finding that MT1-MMP upregulates the production of vascular endothelial growth factor (VEGF) (Deryugina et al., 2002b; Sounni et al., 2002), we assessed the expression of VEGF in tumor cells and surrounding stroma by immunocytochemistry: weak, diffuse staining was observed. In addition, co-culture of MT1-MMP-expressing MDCK cells with microvascular endothelial cells in a collagen gel assay of angiogenesis (Montesano et al., 1993) did not induce the formation of capillary-like tubes (PS and RM, unpublished observations). Therefore, we have no evidence supporting a role for angiogenesis in MT1-MMP-induced tumorigenesis in our experimental setting. A fourth possibility is that, similar to what has been shown for MMP-3 (Thomasset et al., 1998), MT1-MMP recruits host cells and alters the stromal microenvironment, which may in turn promote neoplastic transformation (Liotta and Kohn, 2001; Maffini et al., 2004). Lastly, MT1-MMP may promote tumorigenesis by mechanisms independent of its proteolytic activity, for example, by activating intracellular signalling pathways (Gingras et al., 2001; Takino et al., 2004).

Unlike classical oncogenes, MMPs are not known to be upregulated by gene amplification or activating mutations (Egeblad and Werb, 2002). It is possible, however, that increased MT1-MMP expression resulting from transcriptional changes (Habelhah et al., 1999) enhances susceptibility to tumorigenesis.

Further work will be necessary to elucidate the molecular mechanism by which the expression of MT1-MMP promotes tumor formation by nonmalignant, well-differentiated epithelial cells. This knowledge could well lead to the development of more effective anticarcinogenic strategies.

Materials and methods

Cell culture and transfections

MDCK Tet-Off cells, which stably express the tTA tetracycline-repressible transactivator (Gossen and Bujard, 1992), were obtained from Clontech (Palo Alto, CA, USA). The cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 4 mM L-glutamine, 10% fetal calf serum (FCS) (Life Technologies) and 1 μg/ml puromycin (to maintain the selection pressure for tTA-expressing cells).

Human MT1-MMP cDNA from the pcDNA3-MT1-MMP plasmid (Deryugina et al., 1997) was inserted into the pTRE expression vector (Clontech) at the BamH1 site. Transfections of MDCK Tet-Off cells were performed using Effectene (Qiagen, Basel, Switzerland). The cells were either co-transfected with pTRE-MT1 and PNT Hygro (which carries the hygromycin-resistance gene) or transfected with PNT-Hygro alone. After selection in medium containing 300 μg/ml hygromycin plus 2 μg/ml doxycycline (Dox, Clontech) to turn off the expression of the transfected gene, surviving colonies were isolated and the resulting cell lines were grown in medium supplemented with 150 μg/ml hygromycin and 2 μg/ml Dox.

For experiments, parental MDCK Tet-Off cells as well as MDCK Tet-Off cells transfected with pTRE-MT1 (MDCK-MT1) or with PNT Hygro (MDCK-Hygro) were grown in DMEM with 4 mM L-glutamine and 5% tetracycline-free FCS (Clontech) with or without 2 μg/ml Dox.

RNA extraction and Northern blot

MDCK Tet-Off, MDCK-MT1 and MDCK-Hygro cells were seeded into 100-mm tissue culture dishes (106 cells per dish) in medium with and without 2 μg/ml Dox. Total RNA was extracted after 48 h with Trizol (Life Technologies) according to the manufacturer's instructions. Northern blot was performed as previously described (Pepper et al., 1990). RNAs were denatured with glyoxal, electrophoresed in a 1% agarose gel and transferred overnight onto nylon membranes (Hybond-N; Amersham, Buckinghamshire, UK). The filters were hybridized for 16 h at 65°C with 1.5 × 106 cpm/ml of 32P-labeled human MT1-MMP probe (gift of Dr M Seiki, Tokyo, Japan). As an internal control for determining the amount of RNA loaded, the filters were simultaneously hybridized with 2 × 105 cpm/ml of 32P-labeled P0 ribosomal phosphoprotein cRNA probe.

Gelatinolytic zymography

To assess whether transfected MT1-MMP has catalytic activity, MT1-G cells were tested for their ability to activate latent MMP-2 present in serum-free conditioned medium from MRC-5 fibroblasts (Reno et al., 2004). To prepare conditioned medium, MRC-5 fibroblasts were seeded into 100-mm dishes in minimal essential medium (MEM) with 10% FCS and grown to confluence, at which time the monolayer was washed three times with serum-free MEM and further incubated in 10 ml serum-free MEM. After 3 days, the conditioned medium was collected, centrifuged at low speed to remove floating cells and cell debris, and stored at −20°C. MT1-G cells were seeded in 22-mm wells at 3 × 105 cells/well and incubated with or without 2 μg/ml Dox for 24 h, at which time the monolayer was washed in serum-free MEM and further incubated with 400 μl serum-free MRC-5 conditioned medium. After 3, 6, 9 and 24 h, 30 μl of the conditioned medium was collected, supplemented with 0.5 mM PMSF and 15 mM HEPES, centrifuged for 5 min at 340 × g, and the resulting supernatants were stored at −20°C until use. The 30 μl of conditioned media were electrophoresed under nonreducing conditions in 7.5% SDS/polyacrylamide gels co-polymerized with 1 mg/ml gelatin. After soaking in 2.5% Triton X-100 for 20 min to remove SDS, the gels were incubated in reaction buffer (50 mM Tris-HCl pH 7.4 containing 150 mM NaCl, 10 mM CaCl2) at 37°C for 16 h and then stained with methanol : acetic acid : water (30 : 10: 60) containing 0.25% Coomassie Blue R250 for 4 h. The conditioned medium from MCF-7 cells, which is known to secrete MMP-2, was used as a positive control. Gelatinolytic activity was detected as a clear band on a background of uniform blue staining.

In vitro assays of cystogenesis and invasion

The cyst morphogenesis assay was carried out as described previously (Montesano et al., 1991b) by suspending MDCK cells at 1 × 104 or 2 × 104 cells/ml in a 8 : 1: 1 mixture of rat tail tendon collagen solution, 10 × concentrated MEM and sodium bicarbonate (11.76 mg/ml). After collagen gelation, the complete medium was added and changed every 2–3 days. To assess the ability of MDCK cells to invade a collagen matrix, the cells were seeded onto the surface of a collagen gel (0.5 ml) cast in 22-mm wells atop of an underlying gel layer (1 ml) containing 1 × 105 to 4 × 105 Swiss 3T3 fibroblasts per ml collagen (Montesano et al., 1991b) and grown for 7–11 days. In some experiments, the broad-spectrum hydroxamate MMP inhibitor BB94 or the related inactive isomer BB1268 (kindly provided by Dr P Brown, British Biotech Pharmaceuticals Ltd, Oxford, UK) was added to the cultures at the indicated final concentration.

Assay for tumorigenicity in nude mice

Swiss nu/nu mice (8 weeks old; 25 g body weight) were bred in the animal facilities of the Curie Institute (Paris, France) and maintained under specified pathogen-free conditions. Care and housing were in accordance with the institutional guidelines of the French Ethical Committee (Ministère de l'Agriculture et de la Forêt, Direction de la Santé et de la Protection Animale, Paris, France) and under the supervision of authorized investigators. MDCK-Hygro and MDCK-MT1-G cells were harvested by trypsinization of confluent cultures, and 0.1 ml aliquots of the suspensions containing 1.2 × 107 cells were injected subcutaneously into the flank region in each of three independent experiments. Mice injected with MT1-G cells were separated into two groups, one of which received 2 mg/ml Dox in the drinking water twice a week to inhibit expression of MT1-MMP. A cell population was considered tumorigenic when a tumor nodule appeared (after 4–5 months) and grew progressively at the site of injection. Tumor formation was statistically analysed using the χ2-test. P-values lower than 0.05 were considered significant.

Tumor nodules were excised after killing of the animal, fixed in Bouin's solution (for light microscopy) or in 4% paraformaldehyde in PBS (for immunocytochemistry), dehydrated, and embedded in paraffin. Sections (5-μm-thick) were stained with haematoxylin–eosin. Tumor fragments were also kept in ‘RNA later’ (Qiagen) for subsequent RNA extraction. To search for micrometastases, different organs (lung, liver, brain and lymph nodes) were either frozen in liquid nitrogen or kept in ‘RNA later’ for subsequent DNA and RNA extraction, respectively.

PCR and RT–PCR

DNA was extracted from the organs indicated above as follows: the tissues were incubated in lysis buffer (1 mM Tris-HCl, 1 mM EDTA, 250 mM NaCl, 0.2% SDS) and 100μg/ml proteinase K at 57°C for 16 h. Then 0.2 volumes of potassium acetate 5 M was added and the samples were centrifuged for 15 min at 14 000 r.p.m. The supernatant was precipitated with ethanol and DNA was resuspended in water. The ampicillin gene was amplified by PCR using the following conditions: forward primer 5′-IndexTermTTGCCGGGAAGCTAGAGTAA-3′, reverse 5′-IndexTermGATAACACTGCGGCCAACTT-3′; annealing temperature 55°C; 40 cycles. The size of the amplification product was 391 bp.

RNA was extracted from either tumors or organs using Trizol (Life Technologies) according to the manufacturer's instructions, and was submitted to retrotranscription. Amplification of sequences specific for the exogenous MT1-MMP gene was realized using primers targeting the human MT1-MMP cDNA (forward 5′-IndexTermTGATAAACCCAAAAACCCCA3′, reverse 5′-IndexTermCCTTCCTCTCGTAGGCAGTG-3′). These primers do not amplify mouse or canine MT1-MMP sequences. The annealing temperature was 57°C. The number of cycles required to detect a signal in the tumors or organs, respectively, was 35–50 (with re-addition of 1 U of Taq polymerase/reaction after 1 h of PCR running). The size of the amplification product was 209 bp. Canine MMP-2 was amplified using primers that discriminate between dog and mouse sequences (forward: 5′-IndexTermGAGACCGCCATGTCCACTAT-3′; reverse: 5′-IndexTermCAGAATGCTCCAGTCCCATT-3′). The annealing temperature was 65°C and the number of cycles was 50. The size of amplification product was 238 pb.

Immunohistochemistry

Deparaffined sections were incubated for 12–16 h at 4°C with monoclonal antibody to MT1-MMP (Ab-4; Oncogene, San Diego, CA, USA) or monoclonal antibody to Na,K-ATPase β1 subunit (clone 464.8, Upstate Biotechnology, Charlottesville, USA) at a concentration of 10 μg/ml. A biotinylated anti-mouse IgG was then applied for 1 h, followed by streptavidin-peroxydase (DAKO, Carpinteria, CA, USA) for a further 30 min. Immunoreactivity was visualized by applying diaminobenzidine tetrahydrochloride (DAB) (1/50 w/v) and 0.01% H2O2 for 10 min. Controls included the omission of the primary antibody.

References

  1. Belkin AM, Akimov SS, Zaritskaya LS, Ratnikov BI, Deryugina EI and Strongin AY . (2001). J. Biol. Chem., 276, 18415–18422.

  2. Deryugina EI, Bourdon MA, Jungwirth K, Smith JW and Strongin AY . (2000). Int. J. Cancer, 86, 15–23.

  3. Deryugina EI, Bourdon MA, Luo GX, Reisfeld RA and Strongin A . (1997). J. Cell Sci., 110, 2473–2482.

  4. Deryugina EI, Ratnikov BI, Postnova TI, Rozanov DV and Strongin AY . (2002a). J. Biol. Chem., 277, 9749–9756.

  5. Deryugina EI, Soroceanu L and Strongin A . (2002b). Cancer Res., 62, 580–588.

  6. Dunbar LA and Caplan MJ . (2001). J. Biol. Chem., 276, 29617–29620.

  7. Egeblad M and Werb Z . (2002). Nat. Rev. Cancer, 2, 161–174.

  8. Feraille E and Doucet A . (2001). Physiol. Rev., 81, 345–418.

  9. Friedl P and Wolf K . (2003). Nat. Rev. Cancer, 3, 362–374.

  10. Gingras D, Bousquet-Gagnon N, Langlois S, Lachambre MP, Annabi B and Beliveau R . (2001). FEBS Lett., 507, 231–236.

  11. Gossen M and Bujard H . (1992). Proc. Natl. Acad. Sci. USA, 89, 5547–5551.

  12. Ha HY, Moon HB, Nam MS, Lee JW, Ryoo ZY, Lee TH, Lee KK, So BJ, Sato H, Seiki M and Yu DY . (2001). Cancer Res., 61, 984–990.

  13. Habelhah H, Okada F, Kobayashi M, Nakai K, Choi S, Hamada J, Moriuchi T, Kaya M, Yoshida K, Fujinaga K and Hosokawa M . (1999). Oncogene, 18, 1771–1776.

  14. Hotary K, Allen E, Brooks P, Datta N, Long M and Weiss SJ . (2003). Cell, 114, 33–45.

  15. Hotary K, Allen E, Punturieri A, Yana I and Weiss SJ . (2000). J. Cell Biol., 149, 1309–1323.

  16. Kadono Y, Okada Y, Namiki M, Seiki M and Sato H . (1998a). Cancer Res., 58, 2240–2244.

  17. Kadono Y, Shibahara K, Namiki M, Watanabe Y, Seiki M and Sato H . (1998b). Biochem. Biophys. Res. Commun., 251, 681–687.

  18. Kajita M, Itoh Y, Chiba T, Mori H, Okada A, Kinoh H and Seiki M . (2001). J. Cell Biol., 153, 893–904.

  19. Koshikawa N, Giannelli G, Cirulli V, Miyazaki K and Quaranta V . (2000). J. Cell Biol., 148, 615–624.

  20. Liotta LA and Kohn EC . (2001). Nature, 411, 375–379.

  21. Lynch CC and Matrisian LM . (2002). Differentiation, 70, 561–573.

  22. Maffini MV, Soto AM, Calabro JM, Ucci AA and Sonnenschein C . (2004). J. Cell Sci., 117, 1495–1502.

  23. Mareel M and Leroy A . (2003). Physiol. Rev., 83, 337–376.

  24. Montesano R, Matsumoto K, Nakamura T and Orci L . (1991a). Cell, 67, 901–908.

  25. Montesano R, Pepper MS and Orci L . (1993). J. Cell Sci., 105, 1013–1024.

  26. Montesano R, Schaller G and Orci L . (1991b). Cell, 66, 697–711.

  27. Ohuchi E, Imai K, Fujii Y, Sato H, Seiki M and Okada Y . (1997). J. Biol. Chem., 272, 2446–2451.

  28. Pepper MS, Belin D, Montesano R, Orci L and Vassalli JD . (1990). J. Cell Biol., 111, 743–755.

  29. Rabito CA, Tchao R, Valentich J and Leighton J . (1978). J. Membr. Biol., 43, 351–365.

  30. Reno F, Lombardi F and Cannas M . (2004). Biomaterials, 25, 3439–3443.

  31. Rodriguez-Boulan E and Nelson WJ . (1989). Science, 245, 718–725.

  32. Sato H, Takino T, Okada Y, Cao J, Shinagawa A, Yamamoto E and Seiki M . (1994). Nature, 370, 61–65.

  33. Schenk S and Quaranta V . (2003). Trends Cell Biol., 13, 366–375.

  34. Seiki M . (2003). Cancer Lett., 194, 1–11.

  35. Sounni NE, Devy L, Hajitou A, Frankenne F, Munaut C, Gilles C, Deroanne C, Thompson EW, Foidart JM and Noel A . (2002). FASEB J., 16, 555–564.

  36. Stamenkovic I . (2003). J. Pathol., 200, 448–464.

  37. Sternlicht MD, Bissell MJ and Werb Z . (2000). Oncogene, 19, 1102–1113.

  38. Sternlicht MD, Lochter A, Sympson CJ, Huey B, Rougie JP, Gray JW, Pinkel D, Bissell MJ and Werb Z . (1999). Cell, 98, 137–146.

  39. Sternlicht MD and Werb Z . (2001). Annu. Rev. Cell Dev. Biol., 17, 463–516.

  40. Stetler-Stevenson WG and Yu AE . (2001). Semin. Cancer Biol., 11, 143–152.

  41. Stiles CD, Desmond W, Chuman LM, Sato G and Saier MH . (1976). Cancer Res., 36, 1353–1360.

  42. Strongin AY, Collier I, Bannikov G, Marmer BL, Grant GA and Goldberg GI . (1995). J. Biol. Chem., 270, 5331–5338.

  43. Takino T, Miyamori H, Watanabe Y, Yoshioka K, Seiki M and Sato H . (2004). Cancer Res., 64, 1044–1049.

  44. Thomasset N, Lochter A, Sympson CJ, Lund LR, Williams DR, Behrendtsen O, Werb Z and Bissell MJ . (1998). Am. J. Pathol., 153, 457–467.

  45. Tsunezuka Y, Kinoh H, Takino T, Watanabe Y, Okada Y, Shinagawa A, Sato H and Seiki M . (1996). Cancer Res., 56, 5678–5683.

  46. U HS, Boerner P, Rindler MJ, Chuman L and Saier Jr MH . (1985). J. Cell. Physiol., 122, 299–307.

  47. Udayakumar TS, Chen ML, Bair EL, Von Bredow DC, Cress AE, Nagle RB and Bowden GT . (2003). Cancer Res., 63, 2292–2299.

  48. Vihinen P and Kahari VM . (2002). Int. J. Cancer, 99, 157–166.

  49. Weiss SJ and Pei D . (1996). J. Biol. Chem., 271, 9135–9140.

  50. Yana I and Seiki M . (2002). Clin. Exp. Metastasis, 19, 209–215.

  51. Zucker S, Pei D, Cao J and Lopez-Otin C . (2003). Curr. Top. Dev. Biol., 54, 1–74.

Download references

Acknowledgements

We gratefully acknowledge the initial contribution of Dr J Soriano in setting up the MDCK Tet-Off expression system. We also thank Mrs P Couleru, J Rial-Robert and D Ben-Nasr for excellent technical assistance, Drs G Hosseini and A Wohlwend for helpful advice, N Dupont for secretarial work, Dr M Seiki for providing the human MT1-MMP probe and Dr L Orci for critical reading of the manuscript. This study was supported by Grants No. 31-61446.00 and 3100A0-101734 from the Swiss National Science Foundation (to RM) and NIH Grants CA83017 and CA77470 (to AYS).

Author information

Correspondence to Roberto Montesano.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Soulié, P., Carrozzino, F., Pepper, M. et al. Membrane-type-1 matrix metalloproteinase confers tumorigenicity on nonmalignant epithelial cells. Oncogene 24, 1689–1697 (2005). https://doi.org/10.1038/sj.onc.1208360

Download citation

Keywords

  • carcinogenesis
  • invasion
  • proteolysis
  • MDCK
  • MMP

Further reading