Article

• The EMBO Journal (1997) 16, 2319 - 2332
• doi:10.1093/emboj/16.9.2319

Control of type IV collagenase activity by components of the urokinase–plasmin system: a regulatory mechanism with cell-bound reactants

Roberta Mazzieri¹, Laura Masiero², Lucia Zanetta¹, Sara Monea¹, Maurizio Onisto², Spiridione Garbisa² and Paolo Mignatti^1,3

1. Dipartimento di Genetica e Microbiologia, Università di Pavia, 27100 Pavia, Italy
2. Istituto di Istologia ed Embriologia, Università di Padova, 35121 Padova, Italy
3. Present address: Departments of Surgery and Cell Biology, New York University Medical Center, 550 First Avenue, New York, NY 10016, USA

Correspondence to:

Paolo Mignatti, E-mail: mignap01@mcrcr.med.nyu.edu

Received 6 May 1996; Revised 4 December 1996

• Keywords:

  • activation,
  • cell surface,
  • gelatinase,
  • plasmin,
  • urokinase

Introduction

The extracellular matrix (ECM) degradation that occurs during a variety of tissue remodeling processes, including tumor invasion, angiogenesis and metastasis formation, requires the concerted action of a number of extracellular proteinases. Among these enzymes the urokinase-type plasminogen activator (uPA) and a variety of matrix-degrading metalloproteinases (MMPs) play important roles (Mignatti et al., 1986, 1989; Kleiner and Stetler-Stevenson, 1993; Mignatti and Rifkin, 1993). Two MMPs, MMP-2 or gelatinase A, and MMP-9 or gelatinase B, degrade a variety of ECM proteins, including type IV collagen, fibronectin, laminin and interstitital (type I) collagen; in addition, they efficiently degrade denatured collagens (i.e. gelatins) of all types (Matrisian, 1990; Kleiner and Stetler-Stevenson, 1993; Aimes and Quigley, 1995). These proteinases, collectively referred to as type IV collagenases or gelatinases, have been implicated in a variety of invasive processes, including tumor invasion and metastasis both in experimental models and in human malignancies (Garbisa et al., 1987, 1992; Bernhard et al., 1990, 1994; Levy et al., 1991; Librach et al., 1991; Juarez et al., 1993; Mignatti and Rifkin, 1993; Montgomery et al., 1993; Watanabe et al., 1993; Crawford and Matrisian, 1995; Himelstein et al., 1995; Kohn and Liotta, 1995; Stetler-Stevenson, 1995).

Both uPA and MMPs are secreted in the form of inactive zymogens (prouPA, proMMPs) that are activated extracellularly by limited proteolysis. Trace amounts of plasmin can activate prouPA (Petersen et al., 1988), thus generating a self-maintained feed-back mechanism of prouPA and plasminogen activation. The relatively high concentration of plasminogen in virtually all tissues (Robbins and Summaria, 1970; Isseroff and Rifkin, 1983) can be rapidly converted to plasmin by uPA. Plasmin degrades a variety of ECM components and activates several MMPs, including MMP-1 (collagenase) and MMP–3 (stromelysin-1) but appears to have no effect on gelatinase activation in vitro (Werb et al., 1977; He et al., 1989; Wilhelm et al., 1989; Okada et al., 1990, 1992; Reith and Rucklidge, 1992; Desrivieres et al., 1993; Montgomery et al., 1993). The physiological mechanism(s) of proMMP-2 and proMMP-9 activation are presently unknown. proMMP activation involves the cleavage of the N-terminal 'pro' peptide with a subsequent decrease in M_r of 10 kDa (Matrisian, 1990; Docherty et al., 1992; Goldberg et al., 1992; Kleiner and Stetler-Stevenson, 1993). Several chemical compounds and certain proteinases, including trypsin, tissue kallikrein or cathepsin G, activate these enzymes in the test tube (Okada et al., 1992; Desrivieres et al., 1993; Watanabe et al., 1993); MMP-1 and MMP-3 can activate the progelatinases or 'superactivate' their active forms (He et al., 1989; Matrisian, 1990; Nagase et al., 1991; Goldberg et al., 1992; Kleiner and Stetler-Stevenson, 1993; Mignatti and Rifkin, 1993).

Both serine and metalloproteinase activities involved in tissue remodeling are localized and/or modulated on the cell surface in insoluble phase (Moscatelli and Rifkin, 1988; Saksela and Rifkin, 1988; Brown et al., 1990; Ward et al., 1991; Murphy et al., 1992b; Mignatti and Rifkin, 1993). prouPA binds to a high-affinity cell membrane binding site (uPAR) through a specific N-terminal sequence of its non-catalytic chain (Appella et al., 1987; Behrendt et al., 1990; Roldan et al., 1990). uPAR-bound prouPA is activated by trace amounts of plasmin or by other mechanisms, and remains active on the cell surface for several hours (Stoppelli et al., 1985; Vassalli et al., 1985). Binding of uPA to uPAR accelerates plasminogen activation on the cell surface and localizes the enzyme to focal contact sites (Hébert and Baker, 1988; Pöllanen et al., 1988; Ellis et al., 1991). In most cells, low-affinity, high-capacity binding sites for plasmin(ogen) are present in the chondroitin sulfate proteoglycans of the ECM and of the cell surface (Miles and Plow, 1985; Hajjar et al., 1986; Plow et al., 1986, 1995); some cells also bind plasmin(ogen) and the tissue-type PA (tPA) through a common high-affinity binding site (Hajjar, 1991; Kanalas and Makker, 1991; Felez et al., 1993; Hajjar et al., 1994). MMP-2 interaction with the cell membrane localizes the enzyme activity to specific regions of the cell surface (podosomes or invadopodia) involved in cell invasion (Emonard et al., 1992; Monsky et al., 1993). Recently, MMP-2 has been shown to bind to the _v3 integrin (Brooks et al., 1996). Cell membrane components are involved in both binding and activation of proMMP-2 (Brown et al., 1990; Ward et al., 1991; Murphy et al., 1992b; Strongin et al., 1993). The transmembrane proteins MT-MMPs (membrane-type MMPs) activate proMMP-2 but have no effect on proMMP-9 (Sato et al., 1994; Cao et al., 1995; Takino et al., 1995). MT1-MMP also appears to serve as a high-affinity binding site for MMP-2/TIMP-2 complex (Strongin et al., 1995). Expression of MT1-MMP stimulates in vitro invasion by several cell types, and is associated in vivo with the invasive tumor cells or with stromal components of a variety of human neoplasia (Sato et al., 1994; Nomura et al., 1995; Okada et al., 1995).

To characterize physiological mechanisms that control gelatinase activity in cell cultures we studied proMMP-9 and proMMP-2 activation by human fibrosarcoma HT1080 cells that express high levels of uPA and uPAR, and by clones of mouse L fibroblasts transfected with the human genes for different components of the uPA–plasmin system. These culture models offer several advantages relative to cell-free systems using pure reagents in soluble phase: (i) gelatinase activation can be studied in the presence of cell membrane and ECM components involved in the control of uPA and MMP activities; (ii) the level and cellular distribution of different components of the uPA–plasmin system can be modulated by co-cultivating the L cell transfectants in different relative proportions; and (iii) this co-culture model affords constant levels of the reactants to be maintained for many hours. The results show that physiological concentrations of plasmin can activate both progelatinases by a mechanism independent of metallo- or acid proteinase activities that requires interactions of the reactants with the cell surface. In contrast, in soluble phase plasmin degrades both MMP-9 and MMP-2. Thus, the uPA–plasmin system may represent a physiological mechanism for the control of gelatinase activity.

Plasmin activates proMMP-9 and proMMP-2 in cell cultures without the action of other MMPs

To characterize the role of plasmin in type IV collagen degradation, HT1080 cells that produce high levels of uPA and uPAR (Mazzieri et al., 1994) were tested for their ability to degrade ³H-type IV collagen films in the presence or absence of plasminogen. As shown in Figure 1A and B, low type IV collagenase activity was measured in the absence of plasminogen. Addition of 2 g/ml of pure plasminogen to HT1080 cell cultures considerably increased substrate degradation both by the cell-conditioned medium and by cells grown on a ³H-type IV collagen layer. This effect was abolished by 1,10-phenanthroline or EDTA (not shown), which inhibit metalloproteinases, and by aprotinin, a serine proteinase inhibitor, or ₂-antiplasmin, the physiological inhibitor of plasmin (Figure 2A). In contrast, pepstatin A, an inhibitor of acid proteinases, was ineffective. Since in the absence of cells plasmin had no collagenolytic activity (see legend to Figure 1), and none of these inhibitors was cytotoxic (Mignatti et al., 1986, 1989; data not shown), these results showed that type IV collagen degradation by HT1080 cells requires the concerted action of both metallo- and serine proteinases.

Figure 1.

Effect of plasminogen and proteinase inhibitors on type IV collagen degradation and gelatinase activation by HT1080 cells. Type IV collagen degradation. (A) Serum-free medium conditioned by HT1080 cells in the absence (-) or in the presence (+) of 2 g/ml of pure human plasminogen with or without 100 g/ml of aprotinin (Aprot), 10 g/ml of pepstatin A (Peps) or 10 g/ml of 1,10-phenanthroline (Phen) was assayed for type IV collagenase activity as described in Materials and methods. (B) HT1080 cells were grown in ³H-type IV collagen-coated wells with serum-free medium in the absence (-) or in the presence (+) of 2 g/ml of pure human plasminogen with or without addition of 100 g/ml of aprotinin (Aprot), 10 g/ml of pepstatin A (Peps) or 10 g/ml of 1,10-phenanthroline (Phen). After 16 h incubation, collagen degradation was measured as described in Materials and methods. 8–800 mU (Ploug units) of pure human uPA in 1 ml of serum-free medium with or without 2 g/ml of plasminogen, incubated as a control in parallel wells in the absence of cells, released 3–6% of the total radioactivity. Mean and variability of duplicate samples from a representative experiment are shown. This experiment was repeated three times with comparable results. Gelatin zymography of serum-free conditioned medium (C) or Triton X-100 extracts (D) of HT1080 cells grown for 16 h in the absence (-) or in the presence (+) of 2 g/ml of pure human plasminogen with or without addition of 100 g/ml of aprotinin (Aprot), 10 g/ml of pepstatin A (Peps) or 10 g/ml of 1,10-phenanthroline (Phen). The samples were concentrated by gelatin–Sepharose chromatography, and analyzed by gelatin zymography as described in Materials and methods. Arrowheads indicate the different M_r forms of MMP-9 and MMP-2. This experiment was repeated three times with comparable results. (E) Immunological characterization of the gelatinases associated with HT1080 cells. Western blotting of serum-free conditioned medium or Triton X-100 extracts of HT1080 cells with antibody to human MMP-9 (lane 1) or MMP-2 (lane 2). The experimental procedures are described in Materials and methods. This experiment was repeated three times with comparable results. (F) Biotinylation of the gelatinases associated with HT1080 cells. (Lanes a–c): to evidence intracellular forms of the gelatinases, Triton X-100 extracts of HT1080 cells were biotinylated and the gelatinases purified with gelatin–Sepharose. 20 l (a), 10 l (b) or 5 l (c) of the resin were loaded on a SDS/3–8% polyacrylamide gel, and biotinylated proteins were evidenced as described in Materials and methods. (Lane d): HT1080 cells were surface biotinylated and lysed, and the gelatinases purified with gelatin–Sepharose. 50 l of the resin was loaded on the gel. (Lane e): gelatin zymogram of HT1080 cell extract, to show the correspondence of the lysis bands with the bands evidenced by Western blotting and by biotinylation. Arrowheads show the different M_r forms of MMP-9 and MMP-2. proMMP-9 (92 kDa) is present both on the cell surface and intracellularly. In contrast, 84 kDa MMP-9 is exclusively found on the cell surface, and 79 kDa intracellularly. proMMP-2 not evidenced by gelatin zymography is biotinylated both inside the cell and on the cell surface. This experiment was repeated three times with comparable results.

View full figure (69 KB)

Figure 2.

Effect of plasminogen and ₂-antiplasmin on type IV collagen degradation and gelatinase activation by HT1080 cells. (A) Serum-free medium conditioned by HT1080 cells in the absence or in the presence of 2 g/ml of pure human plasminogen with or without 10 g/ml of ₂-antiplasmin was assayed for type IV collagenase activity as described in Materials and methods. Means and variability of duplicate samples from three experiments are shown. (B) Gelatin zymography of serum-free medium conditioned by HT1080 cells for 16 h in the absence (-) or in the presence of 2 g/ml of pure human plasminogen with or without addition of 10 g/ml of ₂-antiplasmin. APMA: medium conditioned in the absence of plasminogen and ₂–antiplasmin, incubated with p-aminophenylmercuric acetate 1 mM at 37°C for 1 h. The samples were concentrated by gelatin–Sepharose chromatography, and analyzed by gelatin zymography as described in Materials and methods. This experiment was repeated three times with comparable results.

View full figure (29 KB)

Gelatin zymography of serum-free cell-conditioned medium concentrated by ultrafiltration in Centricon tubes (not shown) or by gelatin–Sepharose as described in Materials and methods (Figure 1C) showed major bands with M_r >100 kDa, M_r 92 and 72 kDa, and a minor band of 84 kDa. In contrast, cell extracts (Figure 1D) showed major 92 and 84 kDa bands, and a minor band of 79 kDa, but no 72 kDa band. By Western blotting the 92, 84 and 79 kDa bands, and the 72 kDa band were recognized by specific antibodies to human MMP-9 and MMP-2, respectively (Figure 1E). The M_rs of the 84 and 79 kDa bands are consistent with those of active and of non-glycosylated MMP-9, respectively (Wilhelm et al., 1989; Juarez et al., 1993). The band with M_r >100 kDa, which may represent gelatinase–TIMP complexes or dimers of MMP-2 or MMP-9 (Goldberg et al., 1989, 1992), stained weakly only with anti-MMP-9 antibody. Biotinylation experiments showed that 92 and 84 kDa MMP-9, as well as low amounts of 72 kDa MMP-2 are located at the cell surface (Figure 1F); in contrast, 79 kDa MMP-9 is exclusively intracellular (Figure 1F). Thus, the gelatinases expressed by HT1080 cells are both secreted into the conditioned medium and associated with the cell surface. No bands with M_rs consistent with those of MMP-3 or MMP-7 (matrilysin) could be detected by gelatin or casein zymography of cell extracts or conditioned medium concentrated by ultrafiltration (data not shown), indicating that the type IV collagenase activity of HT1080 cells is associated with MMP-9 and MMP-2. As partial purification with gelatin–Sepharose afforded a better detection of the gelatinases by zymography, this method was used instead of ultrafiltration in all subsequent experiments.

Addition of plasminogen to HT1080 cell cultures had no effect on the gelatinases associated with the cell extracts (Figure 1D). In contrast, gelatin zymography of medium conditioned in the presence of plasminogen (Figure 1C) showed a considerable increase in the intensity of the 84 kDa band, and the generation of a 64–68 kDa band. These bands comigrated with those obtained by treatment of HT1080 cell-conditioned medium with trypsin (not shown) or with p-aminophenylmercuric acid (APMA), which activate proMMP-9 and proMMP-2 (Figure 2B) (Brown et al., 1990; Nagase et al., 1992; Strongin et al., 1993; Itoh et al., 1995).

To test whether plasmin-mediated activation of proMMP-9 and proMMP-2 requires the action of other proteinases, we characterized the effect of different proteinase inhibitors. In the presence of plasminogen, gelatinase activation was abolished by ₂-antiplasmin (Figure 2) or by aprotinin but not by pepstatin A (Figure 1C), showing that the generation of active MMP-9 and MMP–2 in cell-conditioned medium requires plasmin or plasmin-activated proteinase(s). Inhibition of metalloproteinase activity by 1,10-phenanthroline did not block gelatinase activation (Figure 1C): therefore, in the presence of plasmin MMP-9 and MMP-2 activation is not mediated by other MMPs, including MMP-1, MMP-3 or MT-MMPs (He et al., 1989; Goldberg et al., 1992; Sato et al., 1994), and does not result from autocatalysis (Shapiro et al., 1992; Desrivieres et al., 1993; Watanabe et al., 1993).

By the ³H-type IV collagen assay HT1080 cell extracts had low collagenase activity both in the presence and absence of plasminogen. In contrast, the activity of medium conditioned by HT1080 cells in the presence of plasminogen was 5-fold higher than that of medium conditioned in the absence of plasminogen (Figure 3A). Both extracts and conditioned medium of HT1080 cells grown in the presence of plasminogen had comparable levels of plasmin activity (Figure 3B), showing that HT1080 cells bind plasmin(ogen) on their surface. Thus, ³H-type IV collagen degradation by HT1080 cells was not mediated by cell-associated plasmin or MMP-9 but was catalyzed by the active forms of MMP-9 and MMP-2 generated by plasmin in the conditioned medium.

Figure 3.

Type IV collagenase (A) and plasmin (B) activities of extracts or conditioned medium of HT1080 cells grown in the presence (+) or absence (-) of plasminogen. Confluent HT1080 cells were grown for 16 h in serum-free medium in the absence or in the presence of 2 g/ml of pure plasminogen. At the end of the incubation, Triton X–100 cell extracts and conditioned media were assayed for type IV collagenase and for plasmin activities as described in Materials and methods. This experiment was repeated three times with comparable results.

View full figure (86 KB)

To characterize the relative contribution of the different components of the uPA–plasmin system to the control of gelatinase activity, we used clones of mouse L cells transfected with the human genes for uPA, uPAR, the type 1 plasminogen activator inhibitor (PAI-1) or with a cDNA encoding an N-terminal, non-catalytic peptide of uPA that contains the uPAR-binding sequence (ATF). Gelatin zymography of serum-free cell-conditioned media and cell extracts, concentrated by ultrafiltration in Centricon tubes (not shown) or by gelatin–Sepharose as described in Materials and methods, showed the same pattern of bands evidenced in HT1080 cells. Only L_ATF cell extracts showed a 72 kDa band (Figure 4). By Western blotting the 92, 84 and 79 kDa bands, and the 72 kDa band were recognized by specific antibodies to human MMP-9 and MMP-2, respectively (data not shown). No bands with M_rs consistent with those of MMP-3 or matrilysin could be detected by gelatin or casein zymography of L cell extracts or conditioned media concentrated by ultrafiltration. All the L cell clones showed no major differences in the expression of MMP-1, the type-1 and type-2 tissue inhibitors of metalloproteinases (TIMP-1 and TIMP-2), and uPA inhibitors other than PAI-1 (data not shown).

Figure 4.

Zymographic analysis of the gelatinases associated with serum-free conditioned media (A) or Triton X-100 extracts (B) of L cells transfected with the human genes for different components of the uPA plasmin system. Three mg of cell extract protein, or volumes of conditioned medium (3–4 ml) normalized to the protein concentration of the corresponding cell extracts were analyzed by gelatin zymography as described in Materials and methods. This experiment was repeated three times with comparable results. The MMP-2 band comigrates with that of the proMMP-2 secreted by HT1080 cells (not shown). The nature of the band present in L_ATF cell-conditioned medium with M_r higher than that of the MMP-2 secreted by the other clones has not been characterized. This band did not appear consistently in all the zymograms.

View full figure (53 KB)

By the ³H-type IV collagen assay all the L cell clones showed low collagenolytic activity in the absence of plasminogen. Addition of 2 g/ml of pure plasminogen increased the type IV collagenolytic activity of L_uPA cells by 7-fold but did not affect substrate degradation by the other cell clones (Figure 5A). L_uPA cells secrete a high level of human uPA comparable with that of HT1080 cells, whereas the other clones express no uPA. In the absence of cells, an amount of plasmin comparable with that generated by L_uPA cells did not degrade ³H-type IV collagen (Figure 5A).

Figure 5.

(A) Type IV collagen degradation by L cells transfected with the human genes for different components of the uPA plasmin system. Cells of the indicated clones were grown for 16 h in ³H-type IV collagen-coated microculture wells in serum-free medium with or without addition of 2 g/ml of pure human plasminogen (Plg). As a control, 400 mU (Ploug units) of pure human uPA in 1 ml of serum-free medium with or without 2 g/ml of plasminogen were incubated in parallel wells in the absence of cells. At the end of the incubation collagen degradation products were measured in the culture supernatant as described in Materials and methods. Mean and variability of duplicate samples from a representative experiment are shown. This experiment was repeated three times with comparable results. (B) Effect of plasminogen and proteinase inhibitors on gelatinase activation by L_uPA cells. Gelatin zymography of serum-free conditioned medium of L_uPA cells grown for 16 h in the absence (-) or in the presence (+) of 2 g/ml of pure human plasminogen with or without addition of 100 g/ml of aprotinin (Aprot), 10 g/ml of pepstatin A (Peps) or 10 g/ml of 1,10-phenanthroline (Phen). The samples were analyzed by gelatin zymography as described in Materials and methods. This experiment was repeated three times with comparable results.

View full figure (55 KB)

Addition of plasminogen had no effect on the gelatinases associated with L_uPA cell extracts (data not shown). In contrast, gelatin zymography of cell-conditioned medium showed that addition of plasminogen to serum-free L_uPA cell cultures resulted in the generation of active MMP-9 (84 kDa) and MMP-2 (64–68 kDa). This effect was abolished by aprotinin but not by pepstatin A. The metalloproteinase inhibitor 1,10-phenanthroline did not block gelatinase activation (Figure 5B); thus, as is the case for HT1080 cells, proMMP-9 and proMMP-2 activation by L cells requires plasmin but not other MMPs, and does not result from autocatalysis.

uPA binding to uPAR increases type IV collagenase activity

To test the effect of uPA binding to uPAR on the type IV collagenase activity of L cells, increasing amounts of pure human uPA or L_uPA cells were added to L_R cells grown on ³H-type IV collagen films in the absence or presence of 2 g/ml of pure plasminogen. L_R cells secrete no uPA but bind human uPA efficiently (Roldan et al., 1990). Addition of pure uPA or of L_uPA cells increased substrate degradation in a dose-dependent manner in the presence but not in the absence of plasminogen (Figure 6A and B). By gelatin zymography, medium conditioned by L_R cells alone in the presence of plasminogen showed MMP-9 and MMP-2 in proenzyme form; in contrast, in medium conditioned by co-cultures of L_R with 10% L_uPA cells in the presence of plasminogen both the gelatinases were converted into their active forms (Figure 6C). Addition of 1,10-phenanthroline, EDTA or aprotinin inhibited ³H–type IV collagen degradation by co-cultures of L_R + L_uPA cells. ₂-antiplasmin (not shown) had an equivalent effect. In contrast, pepstatin A was ineffective (Figure 7A and B).

Figure 6.

(A) Effect of uPA and plasminogen on the type IV collagenase activity of L_R cells. L_R cells () were grown for 16 h in ³H-type IV collagen-coated wells (310⁵ cells/well) in serum-free medium supplemented with 2 g/ml of plasminogen and the indicated concentrations of uPA. Medium containing plasminogen and uPA was assayed in the absence of cells as a control (). At the end of the incubation radioactive degradation products were measured as described in Materials and methods. Each point represents mean and variability of duplicate samples from a representative experiment. This experiment was repeated four times with comparable results. (B) Type IV collagen degradation by co-cultures of L_R with L_uPA cells. Increasing amounts of L_uPA cells alone () or mixed with 310⁵ L_R cells () were grown for 16 h in ³H-type IV collagen-coated wells in serum-free medium supplemented with 2 g/ml of plasminogen. The number of L_uPA cells is indicated on the abscissa as the percentage of the number of L_R cells. At the end of the incubation substrate degradation was measured as described in Materials and methods. Each point represents mean and variability of duplicate samples from a representative experiment. This experiment was repeated five times with comparable results. (C) Gelatinase activation by co-cultures of L_R with L_uPA cells. Gelatin zymography of medium conditioned by L_R cells or by co-cultures of L_R with 10% L_uPA cells in the presence of 2 g/ml of plasminogen. The samples were concentrated and analyzed as described in Materials and methods. Gelatin zymography of an amount of L_uPA cells equivalent to that used for the co-culture showed no visible bands. This experiment was repeated three times with comparable results. (D) uPA activity of co-cultures of L_R with L_uPA cells. Increasing amounts of L_uPA cells were co-cultivated with 310⁵ L_R cells for 16 h in microculture wells in serum-free medium. The number of L_uPA cells is indicated on the abscissa as the percentage of the number of L_R cells. At the end of the incubation soluble () and cell-bound () uPA activity was measured as described in Materials and methods. Each point represents mean and variability of duplicate samples from a representative experiment. This experiment was repeated four times with comparable results.

View full figure (66 KB)

Figure 7.

Effect of plasminogen and proteinase inhibitors on type IV collagen degradation by co-cultures of L_R with L_uPA cells. (A) Serum-free co-cultures of 310⁵ L_R cells with the indicated relative amounts of L_uPA cells were grown in ³H-type IV collagen-coated wells in the presence of 2 g/ml of pure human plasminogen with the addition of 100 g/ml of aprotinin (), 10 g/ml of pepstatin A (), 10 g/ml of 1,10-phenanthroline (), or with no addition (). After 16 h incubation, collagen degradation was measured as described in Materials and methods. (B) Serum-free medium conditioned by co–cultures of 310⁵ L_R cells with the indicated relative amounts of L_uPA cells in the presence of 2 g/ml of plasminogen with or without 100 g/ml of aprotinin (), 10 g/ml of pepstatin A (), EDTA 10 mM (), or equivalent volumes of PBS () was assayed for type IV collagenase activity as described in Materials and methods. Each point represents mean and variability of duplicate samples from a representative experiment. This experiment was repeated three times with comparable results. Effect of PAI-1 on the uPA (C) and type IV collagenase (D) activities of co-cultures of L_R with L_uPA cells. Increasing amounts of L_PAI-1 cells were mixed with 310⁵ L_R cells and 310⁴ L_uPA cells and grown for 16 h in microculture wells or in ³H-type IV collagen-coated wells with 1 ml of serum-free medium containing 2 g/ml of plasminogen. The number of L_PAI-1 cells is indicated on the abscissa as the percentage of the number of L_R cells. At the end of the incubation ³H-type IV collagen degradation was measured as described in Materials and methods. The co-cultures in the microculture wells were characterized for uPA and type IV collagenase activity in the conditioned medium, and for cell-bound uPA activity as described in Materials and methods. (C) soluble () and cell-bound () uPA activity; (D) type IV collagenase activity present in the conditioned medium () or associated with the cells grown on ³H-type IV collagen layers (). Mean and variability of duplicate samples from a representative experiment are shown. This experiment was repeated three times with comparable results.

View full figure (89 KB)

To confirm that the type IV collagenolytic activity of the co-cultures was mediated by the uPA secreted by L_uPA cells, increasing amounts of L_PAI-1 cells were added to co-cultures containing 10% L_uPA cells. The PAI-1 secreted by L_PAI-1 cells inhibited both cell-associated and soluble uPA activity in a dose-dependent manner (Figure 7C). Likewise, substrate degradation by co-cultures grown on ³H-type IV collagen films decreased as a function of the number of L_PAI-1 cells. In contrast, the type IV collagenase activity of the conditioned medium was not affected (Figure 7D).

Although L_uPA cells produce gelatinase levels higher than those of L_R cells (Figure 4), type IV collagen degradation by L_uPA cells was 6- to 7-fold lower than that of co-cultures with L_R cells (Figure 6B), suggesting that uPA binding to uPAR increased gelatinase activation. The collagenolytic activity of the co-cultures increased slowly in the presence of a proportion of L_uPA to L_R cells lower than 1%, and more rapidly in co-cultures containing 1–10% L_uPA cells (Figure 6B). The level of uPA activity also increased with the number of L_uPA cells. In co-cultures containing <1% L_uPA cells uPA was predominantly cell-associated. With higher proportions of L_uPA to L_R cells the amount of cell-bound uPA leveled off, and the level of uPA in the conditioned medium increased dramatically (Figure 6D). Saturation of the uPA-binding sites of L_R cells and subsequent release of excess uPA in soluble phase paralleled the rapid increase in ³H-type IV collagen degradation. These findings indicated that efficient type IV collagen degradation requires either uPAR-bound uPA or high levels of soluble uPA.

To discriminate between these two hypotheses, increasing amounts of L_uPA cells were co-cultivated on ³H-type IV collagen films with a constant number of either L_R or L_c cells. L_c cells do not bind human uPA (Roldan et al., 1990). In the presence of 2 g/ml of plasminogen co-cultures of L_uPA with L_R cells efficiently degraded the substrate. In contrast, co-cultures with L_c cells showed no collagenolytic activity (Figure 8A). Type IV collagen degradation by co-cultures of L_R cells with 10% L_uPA cells also increased as a function of plasminogen concentration, whereas co-cultures with L_c cells showed virtually no collagenolytic activity even with a high plasminogen concentration (Figure 8B). In a second set of experiments, increasing amounts of L_ATF cells were added to co-cultures of L_R cells with 10% L_uPA cells in the presence of 2 g/ml of plasminogen. L_ATF cells secrete an N-terminal, non-catalytic peptide (ATF) of human uPA that contains the uPAR-binding sequence (Appella et al., 1987; Pedersen et al., 1991). As ATF competes with uPA for binding to uPAR, addition of L_ATF cells resulted in a dose-dependent decrease in the level of cell-bound uPA and in increased soluble uPA in the culture supernatant (Figure 8C). Type IV collagen degradation by the cells or conditioned medium also decreased with increasing L_ATF cell number in a dose-dependent manner (Figure 8D). This effect paralleled the decrease in cell-bound uPA. Thus, uPA binding to uPAR on L_R cell membrane strongly accelerates gelatinase activation.

Figure 8.

Effect of uPA binding to uPAR on the type IV collagenase activity of L cells. (A) ³H-type IV collagen degradation by co-cultures of L_R cells () or L_c cells () with increasing amounts of L_uPA cells in serum-free medium containing 2 g/ml of plasminogen. Co-cultures of L_R cells with increasing amounts of L_c cells () corresponding to those of L_uPA cells, were assayed as a control. The number of L_uPA cells is indicated on the abscissa as the percentage of the number of L_R or L_c cells. Each point represents mean and variability of duplicate samples from a representative experiment. This experiment was repeated four times with comparable results. (B) ³H-type IV collagen degradation by co-cultures of L_R cells () or L_c cells () with 10% L_uPA cells in serum-free medium containing the indicated concentrations of plasminogen. Co-cultures of L_R cells with 10% L_c cells () were assayed as a control. Each point represents mean and variability of duplicate samples from a representative experiment. This experiment was repeated three times with comparable results. Effect of ATF on the uPA (C) and type IV collagenase (D) activities of co–cultures of L_R cells with L_uPA cells. Increasing amounts of L_ATF cells were mixed with 310⁵ L_R cells and 310⁴ L_uPA cells and grown for 16 h in microculture wells or in ³H-type IV collagen-coated wells with 1 ml of serum-free medium containing 2 g/ml of plasminogen. The number of L_ATF cells is indicated on the abscissa as the percentage of the number of L_R cells. At the end of the incubation ³H–type IV collagen degradation was measured as described in Materials and methods. The co-cultures in the microculture wells were characterized for uPA and type IV collagenase activity in the conditioned medium, and for cell-bound uPA activity as described in Materials and methods. (C) soluble () and cell-bound () uPA activity; (D) type IV collagenase activity present in the conditioned medium () or associated with the cells grown on ³H-type IV collagen layers (). Mean and variability of duplicate samples from a representative experiment are shown. This experiment was repeated three times with comparable results.

View full figure (87 KB)

Gelatinase activation by plasmin requires interactions with the cell surface

To understand whether gelatinase activation occurs in soluble phase or requires interaction with the cell surface, we studied activation of proMMP-9 and proMMP-2 in the absence of cells. ¹²⁵I-labeled recombinant MMP-2 was incubated at 37°C for 1 h in the presence or absence of 4 g/ml of plasminogen and increasing uPA concentration, and analyzed by SDS–PAGE and autoradiography. As shown in Figure 9A, addition of plasminogen and uPA resulted in dose-dependent MMP-2 degradation.

Figure 9.

Effect of plasmin on type IV collagenases in the absence of cells. (A) ¹²⁵I-labeled human recombinant MMP-2 (10 000 c.p.m.; 10⁷ c.p.m./g) was incubated for 1 h at 37°C in the absence or presence of the indicated concentrations of pure uPA or plasminogen alone, or with both uPA and plasminogen. At the end of the incubation the samples were electrophoresed in a SDS/3–8% polyacrylamide gel, and exposed to autoradiographic films. This experiment was repeated twice with similar results. (B) One ml of serum-free medium conditioned by L_R cells was incubated for 1 h at 37°C in the presence (+ Plg) or absence (- Plg) of 2 g/ml of plasminogen with or without addition of the indicated amounts (Ploug units) of pure uPA. The reaction was stopped by addition of 100 g of aprotinin, and the levels of MMP-2 were measured by the substrate–capture ELISA described in Materials and methods. Each point represents mean and standard deviation of two experiments with triplicate samples. (C) Gelatin zymography of L_R cell-conditioned medium incubated for 16 h at 37°C in the absence or presence of the indicated concentrations of pure uPA or plasminogen alone, or with both uPA and plasminogen. This experiment was repeated four times with similar results.

View full figure (28 KB)

Because recombinant MMP-2 may be more sensitive to proteolytic degradation than the native enzyme, we characterized the effect of plasmin on MMP-2 secreted by L cells. Serum-free medium conditioned by L_R cells was incubated for 1 h in the absence or presence of plasminogen and increasing uPA concentration, and MMP–2 levels were measured by a substrate-capture ELISA (Wacher et al., 1990). In the presence of plasminogen the level of MMP-2 detected by this assay decreased with increasing uPA concentration in a dose-dependent manner (Figure 9B). Thus, plasmin generated MMP-2 degradation products that either did not bind to the gelatin used as a substrate or were not recognized by the antibody.

To characterize further the effect of plasmin on MMP–2 and MMP-9 in the absence of cells, HT1080 cell-conditioned medium was incubated with pure uPA and plasminogen for 16 h—i.e. the time used to monitor gelatinase activation in cell cultures—and analyzed by gelatin zymography. As shown in Figure 9C, active MMP–2 and MMP–9 were not generated; in contrast, the intensity of the bands corresponding to the progelatinases decreased with increasing uPA concentration in a dose-dependent manner. Thus, in the absence of cells plasmin does not convert proMMP-9 and proMMP-2 into their active forms but rapidly generates degradation products that lack gelatin affinity and/or catalytic activity. These findings indicated that interactions with the cell surface are required for plasmin-mediated activation of proMMP-9 and proMMP-2.

To test this hypothesis, HT1080 cell-conditioned medium was incubated with L_R cells in the presence or absence of 4 g/ml of pure plasminogen. After 0–4 h of incubation, the conditioned medium was characterized by gelatin zymography. As shown in Figure 10A and B, incubation with L_R cells resulted in time-dependent generation of active MMP-2 and MMP-9 in the presence but not in the absence of plasminogen. To understand whether plasmin activates the gelatinases directly or through an intermediate serine proteinase(s), serum-free medium conditioned by HT1080 cells in the presence of 1,10-phenanthroline (10 g/ml) and pepstatin A (10 g/ml) was incubated with confluent L_R cells in the presence or in the absence of 4 g/ml of plasmin. After 2 h incubation, ₂-antiplasmin (10 g/ml) or ₂-antiplasmin plus aprotinin (100 g/ml) was added and the incubation was continued for 4 h. Gelatin zymography of the conditioned medium (Figure 10C) showed that after 2 h incubation with plasmin active gelatinases were generated; the subsequent addition of ₂-antiplasmin prevented the further generation of active forms. Surprisingly, inhibition of plasmin activity also resulted in the disappearance of the 84 and 64–68 kDa forms of MMP-9 and MMP-2 generated during the initial 2 h of incubation, indicating that these active forms of the gelatinases are rapidly processed to degradation products that lack either catalytic activity or gelatin affinity.

Figure 10.

Effect of plasmin on type IV collagenases in the presence of cells. Serum-free medium conditioned by HT1080 cells was incubated for the indicated time (hours) with confluent L_R cells in the absence (A) or in the presence (B) of 4 g/ml of pure plasminogen. At the end of the incubation the conditioned medium was analyzed by gelatin zymography as described in Materials and methods. This experiment was repeated three times with comparable results. C, control medium incubated with L_R cells for 4 h. (C) Effect of ₂-antiplasmin and proteinase inhibitors on gelatinase activation in the presence of cells. Serum-free medium conditioned by HT1080 cells in the presence of 10 g/ml of pepstatin A and 10 g/ml of 1,10-phenanthroline was incubated at 37°C with confluent L_R cells in the absence (-) or in the presence (+) of 4 g/ml of plasmin. After 2 h incubation, 10 g/ml of ₂–antiplasmin or 10 g/ml of ₂-antiplasmin plus 100 g/ml of aprotinin (+^*) was added, and incubation was continued for 4 h. At each time point the conditioned medium was analyzed by gelatin zymography as described in Materials and methods. This experiment was repeated three times with comparable results. (D) Effect of plasmin(ogen) binding to the cell surface. Gelatin zymography of serum-free HT1080 cell-conditioned medium incubated for 2 h at 37°C with confluent L_R cells in the presence (+) or absence (-) of 4 g/ml of plasminogen and 500 g/ml of -aminocaproic acid.

View full figure (69 KB)

To understand whether plasmin binding to the cell surface is required for gelatinase activation, serum-free HT1080 cell-conditioned medium was incubated for 2 h at 37°C with L_R cells in the presence or absence of 4 g/ml of plasminogen and 500 g/ml of -aminocaproic acid (EACA), which prevents plasmin(ogen) binding to the cell surface. Gelatin zymography of the cell-conditioned medium showed that addition of EACA inhibited gelatinase activation (Figure 10D). Thus, gelatinase activation requires the generation of plasmin activity on the cell surface.

The data reported show multiple roles of the uPA–plasmin system in the control of gelatinase activity: (i) in the presence of cells plasmin generates active forms of MMP-9 and MMP-2 without requiring the action of other metallo- or acid proteinases; (ii) this effect results in a 5- to 7-fold increase in type IV collagen degradation; (iii) binding of uPA and plasmin(ogen) on the cell surface, where the gelatinases are also located, accelerates gelatinase activation; (iv) in the absence of cells plasmin degrades soluble MMP-9 and MMP-2; and (v) progelatinase activation and degradation occur in a dose- and time-dependent manner in the presence of relatively low concentrations of plasminogen (1–4 g/ml) and uPA (0.3–100 mU/ml). Thus, the uPA–plasmin system may represent a physiological mechanism for the control of gelatinase activity.

By gelatin zymography, addition of pure plasminogen to HT1080 or L_uPA cells results in the generation of bands with M_r 84 and 64–68 kDa consistent with physiologically or chemically activated MMP-9 and MMP-2, respectively (Stetler-Stevenson et al., 1989; Brown et al., 1990; Okada et al., 1990, 1992; Keski-Oja et al., 1992; Nagase et al., 1992; Juarez et al., 1993; Montgomery et al., 1993; Strongin et al., 1993; Itoh et al., 1995). Under the same conditions in which the 84 and 64–68 kDa bands are generated, degradation of type IV collagen films also occurs. Several observations show that the collagenolytic activity we measured by the ³H-type IV collagen assay is associated with active MMP-9 or MMP-2: (i) by gelatin zymography, conditioned medium or extracts of HT1080 or L cells showed the presence of only MMP-9 and MMP-2; no bands with M_rs consistent with those of matrilysin or stromelysin, which also degrade type IV collagen, were detected; (ii) pure MMP-9 or MMP-2 efficiently solubilize type IV collagen films in a dose-dependent manner in the absence of other proteinases (Atkinson et al., 1992); and (iii) other MMPs, including MMP-3, are at least two orders of magnitude less efficient than MMP-9 or MMP–2 in degrading type IV collagen films by the same assay we used (Atkinson et al., 1992). Under our experimental conditions generation of plasmin resulted in partial activation of MMP-9 and MMP-2, as evidenced by gelatin zymography. However, this effect was associated with a 5- to 7-fold increase in type IV collagen degradation, indicating that the specific activity of the active forms of gelatinases generated by plasmin is considerably higher than that of the corresponding zymogens.

Plasmin generated on the cell surface can activate other MMPs, including MMP-1 and MMP-3 that in turn activate the gelatinases (He et al., 1989; Matrisian, 1990; Nagase et al., 1991; Goldberg et al., 1992; Kleiner and Stetler-Stevenson, 1993; Mignatti and Rifkin, 1993; Sato et al., 1994; Cao et al., 1995; Takino et al., 1995). The activation site of MT1-MMP appears to be compatible with a serine proteinase such as plasmin or uPA (Strongin et al., 1995). However, our finding that gelatinase activation by plasmin is not inhibited by metalloproteinase inhibitors rules out that in the presence of plasmin progelatinase activation is mediated by other MMPs or occurs by autocatalysis (Shapiro et al., 1992; Desrivieres et al., 1993; Watanabe et al., 1993).

Gelatinase activation was completely inhibited by ₂-antiplasmin (Figure 2). This finding is consistent with previous reports that ₂-antiplasmin or serum efficiently block or dramatically retard cell-mediated degradation of the ECM (Wong et al., 1992; Baricos et al., 1995) as well as tumor cell invasion in vitro (Mignatti et al., 1986). Thus, although ECM-bound plasmin is partly protected from ₂-antiplasmin inhibition (Knudsen et al., 1986), in in vitro experiments requiring relatively long incubation ₂-antiplasmin can greatly decrease the amount of active plasmin associated with the cell surface. Inhibition of uPA by PAI-1 blocked collagen degradation by cells grown on ³H-type IV collagen layers but appeared to have no effect on the collagenolytic activity present in their cell-conditioned medium (Figure 7D). This effect may result from the experimental conditions used to measure gelatinase activity in cell-conditioned medium. During incubation at 37°C for 16 h plasminogen might be activated through uPA-independent mechanisms. In addition, small amounts of gelatinase(s) may be activated during incubation with the cells; inhibition of plasmin formation by PAI-1 prevents plasmin-mediated gelatinase degradation in soluble phase.

Activation of proMMP-2 or proMMP-9 in vitro has been reported by other authors to be independent of plasmin (Brown et al., 1990; Okada et al., 1990, 1992; Reith and Rucklidge, 1992; Montgomery et al., 1993), or to require high concentrations of plasminogen (100–200 g/ml) or uPA (Okada et al., 1990, 1992; Keski-Oja et al., 1992; Murphy et al., 1992a; Desrivieres et al., 1993; Montgomery et al., 1993). We found that plasminogen or uPA concentrations 2–3 orders of magnitude lower than those used by other authors (Keski-Oja et al., 1992) can efficiently activate proMMP-2 and proMMP-9. Because the concentration of plasminogen in all tissues is relatively high (Robbins and Summaria, 1970; Isseroff and Rifkin, 1983), the concentrations we used (1–4 g/ml) are likely physiological or below physiological. Whereas different cells may use different mechanisms for gelatinase activation, the discrepancies between our results and those reported by other authors can be explained by the experimental conditions we used. Our cell culture models afforded the characterization of gelatinase activation in the presence of cell membrane and ECM components that contribute to modulating a variety of extracellular proteolytic activities. These conditions cannot be reproduced in the test tube using pure reagents in soluble phase. In addition, the use of co-cultures afforded constant levels of reactants to be maintained throughout the experiments. In contrast, addition of pure reagents to cell-free systems may result in relatively rapid inactivation. Several observations show that the results we obtained with our L cell co-cultures do not reflect nonspecific effects generated by our co-culture conditions: type IV collagen degradation occurred only in the presence of pure plasminogen, and was inhibited by the same proteinase inhibitors that blocked the gelatinase activity of single cell types; moreover, addition of a specific transfectant had parallel, dose-dependent effects on both PA and type IV collagenase activities.

Our data show that plasmin-mediated gelatinase activation occurs when the reactants—uPA, plasmin(ogen) and the gelatinases—can interact with the cell surface. Binding of uPA to uPAR accelerates plasminogen activation on the cell membrane (Ellis et al., 1991); high amounts of plasmin generated by uPAR-bound uPA can more rapidly activate the gelatinases, and contribute to further degrading type IV gelatin generated by MMP-9 or MMP-2 cleavage. Our ³H-type IV collagen film assays were done at physiological temperature (37°C). Under these conditions, the ¾- and ¼-length fragments of type IV collagen generated by MMP-9 or MMP-2 denature spontaneously into randomly coiled peptides that can be cleaved by a variety of enzymes, including plasmin and the same type IV collagenases (gelatinases). However, the single cleavage of the collagen triple helix is the rate-limiting step in collagen degradation (Welgus et al., 1981; Atkinson et al., 1992). In the absence of cells, soluble plasmin in control medium had no collagenolytic activity (Figures 1 and 5A). In addition, although HT1080 cell extract and conditioned medium possessed comparable levels of plasmin activity, the type IV collagenase activity of the cell-conditioned medium was considerably higher than that of the cell extract, which contained lower amounts of gelatinases (Figure 3). Therefore, the increase in collagen degradation in the presence of uPAR-bound uPA reflects efficient gelatinase activation, and is not the result of increased gelatin degradation by soluble or cell-associated plasmin. This effect may result from the juxtaposition of all the reactants on the cell surface. Our data show that both gelatinases are associated with the cell surface (Figure 1F), where uPA and plasmin(ogen) are also localized. Binding of uPA to uPAR accelerates gelatinase activation in a dose-dependent manner (Figure 8A); inhibition of uPA or plasmin binding to the cell surface blocks gelatinase activation (Figures 8D and 10D). Although L_uPA cells do not bind human uPA, they bind plasmin(ogen) on their cell surface or ECM as efficiently as other cell types (data not shown; Miles and Plow, 1985; Plow et al., 1986; Plow and Miles, 1990). The high level of uPA secreted by L_uPA cells may afford the generation of sufficient amounts of cell-bound plasmin to activate proMMP-9 and proMMP–2. Thus, gelatinase activation in cell cultures requires either high levels of soluble uPA and plasminogen or uPA binding to uPAR.

Our observation that progelatinase activation can occur in the presence of metalloproteinase inhibitors is consistent with previous findings (Lohi and Keski-Oja, 1995), and indicates that the proteolytic activity of MT-MMP(s) is not required for MMP-2 activation. However, MT-MMP(s) may serve as a cell membrane binding site for MMP-2 or MMP-2–TIMP-2 complex (Strongin et al., 1995). On the membrane of HT1080 cells MMP-2 and TIMP-2 form a trimolecular complex with an active form of MT1-MMP generated by proteolytic cleavage of its propeptide. The propeptide of MT1-MMP-1 appears to be an excellent substrate for serine proteinases, including plasmin or uPA (Strongin et al., 1995). In the light of these observations it can be speculated that plasmin plays a dual role in a proteolytic cascade which leads to proMMP-2 activation on the cell membrane. Plasmin may activate proMT1-MMP, that binds TIMP-2 and proMMP-2; bound proMMP–2 is then activated by plasmin.

In addition to activating the progelatinases, plasmin also provides a negative control for gelatinase activity by rapidly degrading both MMP-9 and MMP-2 in soluble phase. Thus, type IV collagenase activity in the presence of plasmin results from a balance between gelatinase activation and degradation; this balance being modulated by enzyme interactions with the cell surface. These conclusions prompt consideration of the mechanisms that control gelatinase activity in the extracellular microenvironment. Conformational changes resulting from gelatinase–cell surface interactions may partially protect these enzymes from plasmin degradation, and render the gelatinases available only to limited cleavage. In contrast, in soluble phase multiple cleavage sites may be available for degradation by plasmin. Autolysis or other plasmin-independent mechanisms can also contribute to degradation of active gelatinases. Consistent with previous findings (Nagase et al., 1992), our results indicate that the 84 and the 64–68 kDa forms of MMP-9 and MMP-2 generated in the presence of plasmin are rapidly degraded after addition of ₂-antiplasmin (Figure 10C).

Binding sites for MMP-2 have been described on the membrane of several cell types (Emonard et al., 1992; Monsky et al., 1993; Strongin et al., 1995). We found that both MMP-2 and MMP-9 are located on the surface of HT1080 cells (Figure 1E). MMP-9 was consistently associated with extracts of HT1080 or L cells, whereas MMP-2 could be detected by zymography or Western blotting only in conditioned medium. These results can be explained by several hypotheses, including an apparently higher intracellular concentration of MMP-9 resulting from a longer time for synthesis and/or post-translational modifications. Interestingly, gelatin zymography and Western blotting showed that in both HT1080 and L cells grown in the absence of plasminogen 84 kDa MMP-9 was more abundant in cell extracts than in conditioned medium where, on the contrary, 92 kDa MMP-9 was prevalent (Figure 1C and D; Figure 4). By pulse–chase experiments we found that 84 kDa MMP-9 remains associated with cell extracts, but not with the ECM, for several hours, whereas MMP-2 is rapidly secreted. Deglycosylation experiments showed that 84 kDa MMP–9 is not a glycosylation variant of proMMP-9. Inhibition of protein secretion with Brefeldin A prevents the generation of cell-associated 84 kDa MMP-9, indicating that this form of MMP-9 is generated extracellularly, and bound on the cell membrane (R.Mazzieri, S.Monea, L.Zanetta and P.Mignatti, manuscript in preparation). The generation of 84 kDa MMP-9 associated with HT1080 or L_uPA cell extracts appears to be independent of plasmin (Figures 1D and 4), whereas in cell-conditioned medium 84 kDa MMP-9 is upregulated by plasmin (Figures 1C, 5B and 6C). A further characterization of cell-associated 84 kDa gelatinase may be important to understand the mechanism of proMMP-9 activation.

Our results show that, like proMMP-2 (Brown et al., 1990; Ward et al., 1991; Murphy et al., 1992b; Strongin et al., 1993, 1995; Sato et al., 1994; Cao et al., 1995), proMMP-9 activation also requires interaction with the cell surface. A form of MMP-9 with M_r 84 kDa similar to that of the active form generated in the conditioned medium is associated with the cell surface. In contrast, no active forms of MMP-2 appear to be associated with the cells. Incubation of proMMP-2 with cells overexpressing MT1-MMP also results in generation of soluble forms of the active enzyme; no active MMP-2 is found associated with the cell surface (Sato et al., 1994; Cao et al., 1995). These findings indicate that, unlike MMP-9, MMP-2 interacts with the cell membrane only transiently, and is immediately released after activation. Thus, as proposed by several authors (Docherty et al., 1992; Kleiner and Stetler-Stevenson, 1993; Strongin et al., 1993; Lohi and Keski-Oja, 1995), MMP-9 and MMP-2 may have different mechanisms of activation. Our data show that plasmin represents a common pathway for the activation of both gelatinases.

ECM degradation during tissue remodeling requires a proteinase cascade that involves plasminogen activators and several MMPs (Mignatti et al., 1986, 1989; Mignatti and Rifkin, 1993; Stetler-Stevenson et al., 1993). This proteinase cascade can be modulated by specific interactions of both serine proteinases and MMPs with the cell surface. The comprehension of the molecular mechanisms that mediate gelatinase–cell surface interactions can provide useful tools for the development of novel strategies aimed at inhibiting type IV collagenase activity and controlling a variety of tissue remodeling processes, including tumor invasion and angiogenesis.

Materials

Human or bovine plasminogen and fibrinogen were purified as described (Unkeless et al., 1973). Human recombinant proMMP-2 (Fridman et al., 1993) and rabbit antisera to human MMP-2 and MMP-9 were a generous gift from Dr W.G.Stetler-Stevenson (NCI, Bethesda). Pure human uPA (Ukidan; M_r 54 kDa, 100 000 IU/mg) was purchased from Serono (Rome, Italy), biosynthetically [³H]proline-labeled type IV collagen from Amersham (Amersham, England), gelatin–Sepharose from Pharmacia Biotech AB (Uppsala, Sweden), gelatin from Merck (Darmstadt, Germany), aprotinin, -aminocaproic acid, 1,10-phenanthroline, pepstatin A, p-aminophenylmercuric acetate (APMA) and ₂-antiplasmin from Sigma (St Louis, MO). Protein concentration was measured by the BCA protein assay reagent (Pierce Chemical Co., Rockford, IL) using bovine serum albumin (BSA; Sigma) as a standard. Human fibrinogen and recombinant MMP-2 were labeled with ¹²⁵I (Amersham) using Iodogen (Pierce Chemical Co.) as described (Mignatti et al., 1991).

Cells and culture medium

The following clones of mouse L cells were used for the experiments described: L_uPA cells: mouse L cells transfected with human prouPA cDNA and the neomycin resistance gene (Cajot et al., 1989), kindly provided by Dr J.F.Cajot (ISREC, Lausanne, Switzerland); L_R cells: mouse LB6 cells transfected with human uPAR cDNA and the neomycin resistance gene (LB6 clone 19; Roldan et al., 1990), kindly provided by Drs F.Blasi and N.Pedersen (University of Copenhagen, Denmark); L_PAI-1 cells: mouse L cells transfected with human PAI-1 cDNA and the neomycin resistance gene (L_PAI cells, clone 6; Cajot et al., 1990), kindly provided by Drs J.F.Cajot and B.Sordat (ISREC, Lausanne, Switzerland); L_ATF cells: mouse LB6 cells transfected with a cDNA encoding a non-catalytic, N-terminal peptide of human uPA that contains the uPAR-binding sequence, and with the neomycin resistance gene (clone F; Pedersen et al., 1991), kindly provided by Drs F.Blasi and N.Pedersen (University of Copenhagen, Denmark); L_c cells: control L cells transfected with the neomycin resistance gene alone, kindly provided by the above laboratories.

The cells were grown in Dulbecco's minimum essential medium (DMEM; Gibco BRL, Gaithersburg, MD) supplemented with 10% fetal calf serum (FCS; Biochrom KG, Berlin, Germany), 100 units/ml of penicillin, 100 g/ml of streptomycin and 0.25 g/ml of amphotericin B (Gibco BRL). Every 4–5 weeks, the cells were grown for 5–7 days in the presence of 200 g/ml of geneticin (Sigma, St Louis, MO). The transfected cell clones were characterized for PA inhibitors by measuring inhibition of pure uPA activity by dilutions of conditioned media in [¹²⁵I]-fibrin assay (see below), and for MMP-1, MMP-2, MMP-3, MMP-9, TIMP-1 and TIMP-2 expression by semiquantitative RT–PCR as described (Onisto et al., 1993). The results obtained by RT–PCR for MMP-2 and MMP-9 were consistent with gelatin zymography results.

Human HT1080 fibrosarcoma cells were originally obtained from the American Type Culture Collection and grown in our laboratories for several years in DMEM supplemented with 10% FCS and L-glutamine 2 mM.

Preparation of cell extracts and conditioned media

L cells or HT1080 cells were seeded into 16-mm tissue culture wells at a density of 310⁵ cells/well, or into 10-cm culture dishes at a density of 2.610⁷ cells/dish. Where indicated, a constant number of L_R or L_c cells (310⁵ cells/well or 2.610⁷ cells/dish) were co-cultivated with increasing amounts of L_uPA, L_ATF or L_PAI-1 cells. The relative proportion of these cell types in the co-cultures is indicated as a percentage of L_R cell number. After 4–6 h incubation at 37 °C in DMEM supplemented with 10% FCS, the cells were washed twice with phosphate-buffered saline (PBS) to remove residual FCS, and incubated for 16 h with 1 ml/well or 4 ml/dish of serum-free DMEM with or without the indicated concentrations of plasminogen and the proteinase inhibitors to be tested. The culture supernatants were harvested, and cellular debris removed by centrifugation at 500 g for 10 min at 22 °C. The cells were washed twice with PBS, lysed for 10 min on ice with 100 l/well or 1 ml /dish of Triton X-100 0.5% (v/v) in Tris–HCl 0.1 M, pH 8.1 under constant shaking, and scraped with a rubber policeman. The cell lysates were centrifuged at 800 g for 10 min at 4 °C. Conditioned media and cell extracts were immediately processed for proteinase assays as described below.

Gelatin zymography

Cell extracts (1 ml) or conditioned media (4 ml) were incubated at 4 °C for 1 h in an end-over-end mixer with 50 l of gelatin–Sepharose equilibrated with Tris–HCl 50 mM, NaCl 150 mM, CaCl₂ 5 mM, 0.02% (v/v) Tween-20, EDTA 10 mM, pH 7.6 (Collier et al., 1988). After 4 washes with 1 ml of equilibration buffer containing NaCl 200 mM, the beads were resuspended in 30 l of 4 non-reducing Laëmmli buffer, and loaded on SDS/3–8% polyacrylamide gels containing 1 mg/ml of gelatin. After electrophoresis, the gels were washed twice with 200 ml of 2.5% (v/v) Triton X-100 for 2 h at 22°C to remove SDS, and 3 times with H₂O for 5 min to remove Triton X-100. The gels were incubated at 37°C for 24–48 h in Tris–HCl 50 mM, NaCl 0.2 M, CaCl₂ 20 mM, pH 7.4, stained overnight with Coomassie Brilliant Blue R-250 0.5% (w/v) in 45% (v/v) methanol, 10% (v/v) acetic acid and destained in the same solution without dye (Heussen and Dowdle, 1980). The M_rs of the lysis bands were determined by reference to high-M_r (14.3–200 kDa) standards (Rainbow Markers; Amersham, UK).

Western blotting

Gelatinases purified from cell extracts (1 ml) or conditioned medium (4 ml) by gelatin–Sepharose as described above were electrophoresed in a SDS/3–8% polyacrylamide gel and electroblotted to a nitrocellulose membrane (Hybond-C Extra, Amersham). The membrane was prehybridized at 42°C for 1 h in Tris base 20 mM, NaCl 137 mM, HCl 1.0 M, 0.1% Tween-20, pH 7.6 (TBS-T) containing 5% blocking agent (Amersham), hybridized at 22°C for 1 h in TBS-T containing 1% BSA and the indicated rabbit antisera diluted 1:200–1:300, and incubated at 22 °C for 45 min in TBS-T containing horseradish peroxidase (HPR)-labeled anti-rabbit IgG. Each step was followed by extensive washing in TBS-T (4 ml/cm²) at 22°C. After removing the TBS-T buffer, the membrane was incubated for 1 min at 22°C with 0.125 ml/cm² of ECL detection solution (Amersham) and exposed to films (Hyperfilm MP, Amersham) for 10 s to 5 min.

Cell surface biotinylation

To label cell surface protein, subconfluent HT1080 cells in 50 cm² tissue culture dishes were washed twice with cold PBS, and incubated at 4°C for 30 min with 3–5 ml of biotinylation buffer (bicarbonate buffer, pH 8.6; Amersham) containing 40 l/ml of biotin ester (Amersham). At the end of the incubation, the cells were washed twice with PBS and lysed with Triton X-100 as described above (preparation of cell extracts). To biotinylate total cell proteins, Triton X-100 cell extract protein was diluted to 1 mg/ml in biotinylation buffer containing 40 l of biotinylation reagent per mg protein, and incubated at 22°C for 1 h. Cell extracts were chromatographed on a G-25 column to remove unreacted biotin. The eluted protein was concentrated with 50 l of gelatin–Sepharose and run in an SDS–polyacrylamide gel as described above for gelatin zymography. After blotting to a nitrocellulose membrane, blocking and washing with TBS-T as described above for Western blotting, the membrane was incubated at 22°C for 1 h with horseradish peroxidase-labeled streptavidin (Amersham) 1:1500 in TBS-T. Following extensive washing with the same buffer, biotinylated proteins were evidenced by addition of ECL detection solution (Amersham) as described above. Films (Hyperfilm, Amersham) were exposed for 10 s to 1 min.

Assay for type IV collagenase activity

One ml of serum-free conditioned medium, or 310⁵ cells or 150–250 g of cell extract protein in 1 ml DMEM were added to ³H-type IV collagen-coated microculture wells (2 cm²; 3000–4000 d.p.m./well) (Garbisa et al., 1980). After 4–6 h incubation, when the cells had spread on the substrate, the medium was removed. The cells were washed twice with PBS to remove residual FCS, and incubated for 16 h with 1 ml of serum-free DMEM in the presence or absence of the indicated concentrations of plasminogen and the proteinase inhibitors to be tested. Serum-free medium with or without plasminogen and proteinase inhibitors was incubated in parallel wells as a control. At the end of the incubation, 800 l aliquots of the supernatants were collected, and radioactive degradation products measured in a Camberra Packard liquid scintillation counter (Garbisa et al., 1980). Samples and controls were assayed in duplicate. Unless otherwise indicated, type IV collagen degradation is reported in the figures as percentage of the total substrate after subtraction of background radioactivity released by control medium with or without plasminogen and the proteinase inhibitors (3–5% of the total radioactivity).

Assay for plasminogen activator (PA) or plasmin activity

1–10 g of cell extract protein or 10–50 l of cell-conditioned medium was tested for PA activity by the ¹²⁵I-fibrin assay (Gross et al., 1982) in the presence of 2 g/ml of pure plasminogen. To measure receptor-bound uPA activity, confluent cells in 16-mm culture wells were washed twice with PBS and incubated with 100 l/well of 50 mM glycine–HCl, 0.1 M NaCl, pH 3.1 at 22°C for 3 min with mild shaking (Stoppelli et al., 1986). After neutralization with 100 l of 0.5 M HEPES, 0.1 M NaCl, pH 7.5, 50 l aliquots of the acid wash were tested for uPA activity by the ¹²⁵I-fibrin assay (Gross et al., 1982). Equivalent volumes of a mixture of the glycine–HCl with the HEPES buffers were used as a control. To measure plasmin activity 150–250 g of cell extract protein or 150–250 l of cell-conditioned medium were tested by the ¹²⁵I-fibrin assay without adding plasminogen to the assay buffer. Samples and controls were assayed in duplicate.

Assay for PA inhibitor (PAI) activity

Increasing volumes of media conditioned by the different L cell clones, diluted in 0.5 ml of 0.1 M Tris–HCl, pH 8.1 containing 250 g/ml BSA, 4 g/ml plasminogen and 32 mU (International Units)/ml of pure human uPA, were tested for their ability to inhibit uPA activity in a ¹²⁵I-fibrin assay (Gross et al., 1982). Equivalent volumes of DMEM were used as controls. Samples and controls were assayed in duplicate.

ELISA for MMP-2

The level of MMP-2 in L_R cell-conditioned medium was quantitated by a modified substrate-capture immunoassay (Wacher et al., 1990). Ten l of conditioned medium diluted with 90 l of TSE (50 mM Tris–HCl, 200 mM NaCl, 10 mM EDTA, pH 7.6) were incubated overnight at 4°C in gelatin-coated, flat-bottom microtiter wells (Nunclon, Roskilde, Denmark). The captured MMP was detected by 3 h incubation at 4°C with a 1:500 dilution of rabbit antiserum (4221) to affinity-purified human MMP-2, followed by incubation for 2 h at 4°C with peroxidase-conjugated goat anti-rabbit IgG antibody (1:5000). After addition of o–phenylenediamine (Bio-Rad) for 10 min at 22°C, OD₄₀₅ was measured with a Titertek spectrophotometer. The specificity of the antibody has been assessed as described (Wacher et al., 1990). Human recombinant MMP-2 was used as a standard and irrelevant IgG as a negative control. The OD₄₀₅ readings of the samples were in the linear range of the assay. Samples and controls were assayed in triplicate.

We thank Drs F.Blasi, J.F.Cajot, N.Pedersen, and B.Sordat for kindly providing the transfected L cells; Drs D.B.Rifkin and D.Moscatelli for critically reading the manuscript and for their helpful discussions; Dr W.G.Stetler-Stevenson for kindly providing recombinant MMP-2 and antibodies to MMP-9 and MMP-2, and for his helpful comments; Prof. M.Spina for his precious collaboration; and F.Tredici and M.Bensi for their skillful technical assistance. This work was supported by grants from the Italian Association for Cancer Research (AIRC) and the Italian Ministry of University, and Scientific and Technological Research (MURST 60%) to P.M. and S.G., and from the Italian National Research Council (CNR Progetto Finalizzato Applicazioni Cliniche della Ricerca Oncologica, and Progetto Bilaterale) to S.G. R.M. was supported by a fellowship from AIRC.

Aimes RT and Quigley JP (1995) Matrix metalloproteinase-2 is an interstitial collagenase. J Biol Chem, 270, 5872–5876. | Article | PubMed | ISI | ChemPort |

Appella E, Robinson EA, Ullrich SJ, Stoppelli A, Corti MP, Cassani G and Blasi F (1987) The receptor-binding sequence of urokinase. A biological function for the growth factor module of proteases. J Biol Chem, 262, 4437–4440. | PubMed | ISI | ChemPort |

Atkinson SJ, Ward RV, Reynolds JJ and Murphy G (1992) Cell-mediated degradation of type IV collagen and gelatin films is dependent on the activation of matrix metalloproteinases. Biochem J, 288, 605–611. | PubMed | ISI | ChemPort |

Baricos WH, Cortez SL, el-Dahr SS and Schnaper HW (1995) ECM degradation by cultured human mesangial cells is mediated by a PA/plasmin/MMP-2 cascade. Kidney Int 47, 1039–47. | Article | PubMed | ISI | ChemPort |

Behrendt N et al. (1990) The human receptor for urokinase plasminogen activator. NH₂-terminal sequence and glycosylation variants. J Biol Chem, 265, 6453–6460. | PubMed | ISI | ChemPort |

Bernhard EJ, Muschel RJ and Hughes EN (1990) M_r 92,000 gelatinase correlates with the metastatic phenotype in transformed rat embryo cells. Cancer Res, 50, 3872–3877. | PubMed | ISI | ChemPort |

Bernhard EJ, Gruber SB and Muschel RJ (1994) Direct evidence linking expression of matrix metalloproteinase 9 (92-kD gelatinase/collagenase) to the metastatic phenotype in transformed rat embryo cells. Proc Natl Acad Sci USA, 91, 4293–4297. | PubMed | ChemPort |

Brooks PC, Stromblad S, Sanders LC, von Schalscha TL, Aimes RT, Stetler-Stevenson WG, Quigley JP and Cheresh DA (1996) Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrin alpha v beta 3. Cell, 85, 683–693. | Article | PubMed | ISI | ChemPort |

Brown PD, Levy AT, Margulies IMK, Liotta LA and Stetler-Stevenson WG (1990) Independent expression and cellular processing of M_r 72,000 type IV collagenase and interstitial collagenase in human tumorigenic cell lines. Cancer Res, 50, 6184–6191. | PubMed | ISI | ChemPort |

Cajot J-F, Schleuning W-D, Medcalf RL, Bamat J, Testuz J, Liebermann L and Sordat B (1989) Mouse L cells expressing human prourokinase-type plasminogen activator: effects on extracellular matrix degradation and invasion. J Cell Biol, 109, 915–925. | PubMed | ISI | ChemPort |

Cajot J-F, Bamat J, Bergonzelli GE, Kruithof EKO, Medcalf RL, Testuz J and Sordat B (1990) Plasminogen-activator inhibitor type 1 is a potent natural inhibitor of extracellular matrix degradation by fibrosarcoma and colon carcinoma cells. Proc Natl Acad Sci USA, 87, 6939–6943. | PubMed | ChemPort |

Cao J, Sato H, Takino T and Seiki M (1995) The C-terminal region of membrane type matrix metalloproteinase is a functional transmembrane domain required for pro-gelatinase A activation. J Biol Chem, 270, 801–805. | Article | PubMed | ISI | ChemPort |

Collier IE et al. (1988) H-ras oncogene-transformed human bronchial epithelial cells (TBE-1) secrete a single metalloproteinase capable of degrading basement membrane collagen. J Biol Chem, 263, 6579–6587. | PubMed | ISI | ChemPort |

Crawford HC and Matrisian LM (1995) Tumor and stromal expression of matrix metalloproteinases and their role in tumor progression. Invasion Metastasis, 14, 234–245. | ISI | ChemPort |

Desrivieres D, Lu H, Peyri N, Soria C, Legrand Y and Menashi S (1993) Activation of the 92 kDa type IV collagenase by tissue kallikrein. J Cell Physiol, 157, 587–593. | PubMed | ISI |

Docherty AJP, O'Connel J, Crabble T, Angal S and Murphy G (1992) The matrix metalloproteinases and their natural inhibitors: prospects for treating degenerative tissue diseases. Trends Biotechnol, 10, 200–207. | Article | PubMed | ISI | ChemPort |

Ellis V, Behrendt N and Danø K (1991) Plasminogen activation by receptor-bound urokinase. A kinetic study with both cell-associated and isolated receptor. J Biol Chem, 266, 12752–12758. | PubMed | ISI | ChemPort |

Emonard HP, Remacle AG, Noël AC, Grimaud JA, Stetler-Stevenson WG and Foidart JM (1992) Tumor cell surface associated binding site for M(r) 72,000 type IV collagenase. Cancer Res, 52, 5845–5848. | PubMed | ISI | ChemPort |

Felez J, Chanquia CJ, Fabregas P, Plow EF and Miles LA (1993) Competition between plasminogen and tissue plasminogen activator for cellular binding sites. Blood, 82, 2433–2441. | PubMed | ISI | ChemPort |

Fridman R et al. (1993) Expression of human recombinant 72 kDa gelatinase and tissue inhibitor of metalloproteinase-2 (TIMP-2): characterization of complex and free enzyme. Biochem J, 289, 411–416. | PubMed | ISI | ChemPort |

Garbisa S, Kniska K, Tryggvason K, Foltz CM and Liotta LA (1980) Quantitation of basement membrane collagen degradation by living tumor cells in vitro. Cancer Letters, 9, 359–366. | Article | PubMed | ISI | ChemPort |

Garbisa S, Pozzatti R, Mushel R, Saffiotti U, Ballin M, Goldfarb R, Khoury G and Liotta LA (1987) Secretion of type IV collagenolytic protease and metastatic phenotype: induction by transfection with c-Ha-ras but not c-Ha-ras + Ad2-E1a. Cancer Res, 47, 1523–1528. | PubMed | ISI | ChemPort |

Garbisa S, Scagliotti G, Masiero L, Di Francesco C, Caenazzo C, Onisto M, Micela M, Stetler-Stevenson WG and Liotta LA (1992) Correlation of serum metalloproteinase levels with lung cancer metastasis and response to therapy. Cancer Res, 52, 4548–4549. | PubMed | ISI | ChemPort |

Goldberg GI, Marmer BL, Grant GA, Eisen AZ, Wilhelm A and He C (1989) Human 72-kilodalton type IV collagenase forms a complex with a tissue inhibitor of metalloproteinases designated TIMP-2. Proc Natl Acad Sci USA, 86, 8207–8211. | Article | PubMed | ChemPort |

Goldberg GI, Strongin A, Collier IE, Genrich LT and Marmer BL (1992) Interaction of 92-kD type IV collagenase with the tissue inhibitor of metalloproteinases prevents dimerization, complex formation with interstitial collagenase, and activation of the proenzyme with stromelysin. J Biol Chem, 267, 4583–4591. | PubMed | ISI | ChemPort |

Gross JL, Moscatelli D, Jaffe EA and Rifkin DB (1982) Plasminogen activator and collagenase production by cultured capillary endothelial cells. J Cell Biol, 95, 974–981. | PubMed | ISI | ChemPort |

Hajjar KA (1991) The endothelial cell tissue plasminogen activator receptor. Specific interaction with plasminogen. J Biol Chem, 266, 21962–21970. | PubMed | ISI | ChemPort |

Hajjar KA, Harpel PC, Jaffe EA and Nachman RL (1986) Binding of plasminogen to cultured human endothelial cells. J Biol Chem, 261, 11656–11662. | PubMed | ChemPort |

Hajjar KA, Jacovina AT and Chacko J (1994) An endothelial cell receptor for plasminogen/tissue plasminogen activator. I. Identity with annexin II. J Biol Chem, 269, 21191–21197. | PubMed | ISI | ChemPort |

He C, Wilhelm SM, Pentland AP, Marmer BL, Grant GA, Eisen AZ and Goldberg GI (1989) Tissue cooperation in a proteolytic cascade activating human interstitial collagenase. Proc Natl Acad Sci USA, 86, 2632–2636. | Article | PubMed | ChemPort |

Hébert CA and Baker JB (1988) Linkage of extracellular plasminogen activator to the fibroblast cytoskeleton: colocalization of cell surface urokinase with vinculin. J Cell Biol, 106, 1241–1248. | Article | PubMed | ISI | ChemPort |

Heussen C and Dowdle EB (1980) Electrophoretic analysis of plasminogen activators in polyacrylamide gels containing sodium dodecyl sulfate and copolymerized substrates. Anal Biochem, 102, 196–202. | Article | PubMed | ISI | ChemPort |

Himelstein BP, Canete-Soler R, Bernhard EJ, Dilks DW and Muschel RL (1995) Metalloproteinases in tumor progression: the contribution of MMP-9. Invest Methods, 14, 246–258. | ChemPort |

Isseroff RR and Rifkin DB (1983) Plasminogen is present in the basal layer of epidermis. J Invest Dermatol, 80, 297–299. | Article | PubMed | ISI | ChemPort |

Itoh Y, Binner S and Nagase H (1995) Steps involved in activation of the complex of pro-matrix metalloproteinase 2 (progelatinase A) and tissue inhibitor of metalloproteinases (TIMP)-2 by 4-aminophenylmercuric acetate. Biochem J, 308, 645–651. | PubMed | ISI | ChemPort |

Juarez J, Clayman G, Nakajima M, Tanabe KK, Saya H, Nicolson GL and Boyd D (1993) Role and regulation of expression of 92-kD type IV collagenase (MMP-9) in 2 invasive squamous-cell-carcinoma cell lines of the oral cavity. Int J Cancer, 55, 10–18. | Article | PubMed | ISI | ChemPort |

Kanalas JJ and Makker SP (1991) Identification of the rat Heymann nephritis autoantigen (gp330) as a receptor site for plasminogen. J Biol Chem, 266, 10825–10829. | PubMed | ISI | ChemPort |

Keski-Oja J, Lohi J, Tuuttila A, Tryggvason K and Vartio T (1992) Proteolytic processing of the 72,000-Da type IV collagenase by urokinase plasminogen activator. Exp Cell Res, 202, 471–476. | Article | PubMed | ChemPort |

Kleiner DE and Stetler-Stevenson WG (1993) Structural biochemistry and activation of matrix metalloproteinases. Curr Opin Cell Biol, 5, 891–897. | Article | PubMed | ChemPort |

Knudsen BS, Silverstein RL, Leung LLK, Harpel PC and Nachman RL (1986) Binding of plasminogen to extracellular matrix. J Biol Chem, 261, 10765–10771. | PubMed | ChemPort |

Kohn EC and Liotta LA (1995) Molecular insights into cancer invasion: strategies for prevention and intervention. Cancer Res, 55, 1856–1862. | PubMed | ISI | ChemPort |

Levy AT, Cioce V, Sobel ME, Garbisa S, Grigioni W, Liotta LA and Stetler-Stevenson WG (1991) Increased expression of the 72 kDa type IV collagenase in human colonic adenocarcinoma. Cancer Res, 51, 439–444. | PubMed | ISI | ChemPort |

Librach CL et al. (1991) 92-kD type IV collagenase mediates invasion of human cytotrophoblasts. J Cell Biol, 113, 437–449. | Article | PubMed | ISI | ChemPort |

Lohi J and Keski-Oja J (1995) Calcium ionophores decrease pericellular gelatinolytic activity via inhibition of 92-kDa gelatinase expression and decrease of 72-kDA gelatinase activation. J Biol Chem, 270, 17602–17609. | Article | PubMed | ISI | ChemPort |

Matrisian LM (1990) Metalloproteinases and their inhibitors in matrix remodeling. Trends Genet, 6, 121–125. | Article | PubMed | ISI | ChemPort |

Mazzieri R, Høyer-Hansen G, Rønne E, Løber D, Vagnarelli P, Raimondi E, De Carli L, Danø K and Mignatti P (1994) Urokinase and urokinase receptor expression in somatic cell hybrids. Fibrinolysis, 8, 344–352. | ISI | ChemPort |

Mignatti P and Rifkin DB (1993) Biology and biochemistry of proteinases in tumor invasion. Physiol Rev, 73, 161–195. | PubMed | ISI | ChemPort |

Mignatti P, Robbins E and Rifkin DB (1986) Tumor invasion through the human amniotic membrane: requirement for a proteinase cascade. Cell, 47, 487–498. | Article | PubMed | ISI | ChemPort |

Mignatti P, Tsuboi R, Robbins E and Rifkin DB (1989) In vitro angiogenesis on the human amniotic membrane: requirement for basic fibroblast growth factor-induced proteinases. J Cell Biol, 108, 671–682. | Article | PubMed | ISI | ChemPort |

Mignatti P, Mazzieri R and Rifkin DB (1991) Expression of the urokinase receptor in vascular endothelial cells is stimulated by basic fibroblast growth factor. J Cell Biol, 113, 1193–1201. | Article | PubMed | ISI | ChemPort |

Miles LA and Plow EF (1985) Binding and activation of plasminogen on the platelet surface. J Biol Chem, 260, 4303–4311. | PubMed | ISI | ChemPort |

Monsky WL, Kelly T, Lin CY, Yeh Y, Stetler-Stevenson WG, Mueller SC and Chen WT (1993) Binding and localization of M(r) 72,000 matrix metalloproteinase at cell surface invadopodia. Cancer Res, 53, 3159–64. | PubMed | ISI | ChemPort |

Montgomery AM, De Clerck YA, Langley KE, Reisfeld BM and Mueller BM (1993) Melanoma-mediated dissolution of extracellular matrix: contribution of urokinase-dependent and metalloproteinase-dependent proteolytic pathways. Cancer Res, 53, 693–700. | PubMed | ISI | ChemPort |

Moscatelli D and Rifkin DB (1988) Membrane and matrix localization of proteinases: a common theme in tumor cell invasion and angiogenesis. Biochim Biophys Acta, 948, 67–85. | Article | PubMed | ISI | ChemPort |

Murphy G, Atkinson S, Ward R, Gavrilovic J and Reynolds JJ (1992a) The role of plasminogen activators in the regulation of connective tissue metalloproteinases. Ann NY Acad Sci, 667, 1–12. | Article | PubMed | ChemPort |

Murphy G, Willenbrock F, Ward RV, Cockett MI, Eaton D and Docherty AJ (1992b) The C-terminal domain of 72 kDa gelatinase A is not required for catalysis, but is essential for membrane activation and modulates interactions with tissue inhibitors of metalloproteinases. Biochem J, 283, 637–641. | PubMed | ISI | ChemPort |

Nagase H, Ogata Y, Suzuki K, Enghild JJ and Salvensen G (1991) Substrate specificities and activation mechanisms of matrix metalloproteinases. Biochem Soc Trans, 19, 715–718. | PubMed | ISI | ChemPort |

Nagase H, Suzuki K, Morodomi T, Enghild JJ and Salvesen G (1992) Activation mechanisms of the precursor of matrix metalloproteinases 1, 2 and 3. Matrix Suppl, 1, 237–244. | PubMed | ChemPort |

Nomura H, Sato H, Seiki M, Mai M and Okada Y (1995) Expression of membrane-type matrix metalloproteinase in human gastric carcinomas. Cancer Res, 55, 3263–3266. | PubMed | ISI | ChemPort |

Okada Y, Morodomi T, Enghild JJ, Suzuki K, Yasui A, Nakanishi I, Salvesen G and Nagase H (1990) Matrix metalloproteinase 2 from human rheumatoid synovial fibroblasts. Purification and activation of the precursor and enzymic properties. Eur J Biochem, 194, 721–730. | Article | PubMed | ISI | ChemPort |

Okada Y, Gonoji Y, Naka K, Tomita K, Nakanishi I, Iwata K, Yamashita K and Hayakawa T (1992) Matrix metalloproteinase 9 (92–kD gelatinase/type IV collagenase) from HT1080 human fibrosarcoma cells. Purification and activation of the precursor and enzymic properties. J Biol Chem, 267, 21712–21719. | PubMed | ISI | ChemPort |

Okada Y, Bellocq JP, Rouyer N, Chenard MP, Rio MC, Chambon P and Basset P (1995) Membrane-type matrix metalloproteinase (MT-MMP) gene is expressed in stromal cells of human colon, breast, head and neck carcinomas. Proc Natl Acad Sci USA, 92, 2730–2734. | PubMed | ChemPort |

Onisto M, Garbisa S, Caenazzo C, Freda M, Di Francesco C, Liotta LA and Stetler-Stevenson WG (1993) RT-PCR phenotyping of metalloproteinases and inhibitors involved in tumor matrix invasion. Diagn Mol Pathol, 2, 74–80. | PubMed | ISI | ChemPort |

Pedersen N, Masucci MT, Møller LB and Blasi F (1991) A non-catalytic human urokinase plasminogen activator derivative produced by mouse cells has full receptor-binding activity. Fibrinolysis, 5, 155–164. | Article | ISI | ChemPort |

Petersen LC, Lund LR, Nielsen LS, Danø K and Skriver L (1988) One chain urokinase-type plasminogen activator from human sarcoma cells is a proenzyme with little or no intrinsic activity. J Biol Chem, 263, 11189–11195. | PubMed | ISI | ChemPort |

Plow EF and Miles LA (1990) Plasminogen receptors in the mediation of pericellular proteolysis. Cell Differ Dev, 32, 293–298. | Article | PubMed | ISI | ChemPort |

Plow EF, Freaney DE, Plescia J and Miles LA (1986) The plasminogen system and cell surfaces: evidence for plasminogen and urokinase receptors on the same cell type. J Cell Biol, 103, 2411–2420. | Article | PubMed | ISI | ChemPort |

Plow EF, Herren T, Redlitz A, Miles LA and Hoover-Plow JL (1995) The cell biology of the plasminogen system. FASEB J, 9, 939–945. | PubMed | ISI | ChemPort |

Pöllanen J, Hedman K, Nielsen LS, Danø K and Vaheri A (1988) Ultrastructural localization of plasma membrane-associated urokinase-type plasminogen activator at focal contacts. J Cell Biol, 106, 87–95. | Article | PubMed | ISI

Reith A and Rucklidge GJ (1992) Invasion of brain tissue by primary glioma: evidence for the involvement of urokinase-type plasminogen activator as an activator of type IV collagenase. Biochem Biophys Res Commun, 186, 348–354. | PubMed | ISI | ChemPort |

Robbins KC and Summaria L (1970) Human plasminogen and plasmin. Methods Enzymol, 19, 184–199. | ChemPort |

Roldan AL, Cubellis MV, Masucci MT, Behrendt N, Lund LR, Danø K, Appella E and Blasi F (1990) Cloning and expression of the receptor for human urokinase plasminogen activator, a central molecule in cell surface, plasmin dependent proteolysis. EMBO J, 9, 467–474. | PubMed | ISI | ChemPort |

Saksela O and Rifkin DB (1988) Cell-associated plasminogen activation: regulation and physiological functions. Annu Rev Cell Biol, 4, 93–126. | Article | PubMed | ISI | ChemPort |

Sato H, Takino T, Okada Y, Cao J, Shinagawa A, Yamamoto E and Seiki M (1994) A matrix metalloproteinase expressed on the surface of invasive tumor cells. Nature, 370, 61–65. | Article | PubMed | ISI | ChemPort |

Shapiro SD, Griffin GL, Gilbert DJ, Jenkins NA, Copeland NG, Welgus HG, Senior RM and Ley TJ (1992) Molecular cloning, chromosomal localization, and bacterial expression of a Murine Macrophage Metalloelastase. J Biol Chem, 267, 4664–4671. | PubMed | ISI | ChemPort |

Stetler-Stevenson WG (1995) Progelatinase A activation during tumor cell invasion. Invest Methods, 14, 259–268. | ChemPort |

Stetler-Stevenson WG, Krutzsch HC, Wacher MP, Margulies IMK and Liotta LA (1989) The activation of human type IV collagenase proenzyme. Sequence identification of the major conversion product following organomercurial activation. J Biol Chem, 264, 1353–1356. | PubMed | ChemPort |

Stetler-Stevenson WG, Aznavoorian S and Liotta LA (1993) Tumor cell interactions with the extracellular matrix during invasion and metastasis. Annu Rev Cell Biol, 9, 541–573. | Article | PubMed | ChemPort |

Stoppelli MP, Corti A, Soffientini A, Cassani G, Blasi F and Assoian RK (1985) Differentiation-enhanced binding of the amino-terminal fragment of human urokinase plasminogen activator to a specific receptor on U937 monocytes. Proc Natl Acad Sci USA, 82, 4939–4943. | Article | PubMed | ChemPort |

Stoppelli MP, Tacchetti C, Cubellis MV, Corti A, Hearing VJ, Cassani G, Appella E and Blasi F (1986) Autocrine saturation of pro-urokinase receptors on human A431 cells. Cell, 45, 675–684. | Article | PubMed | ISI | ChemPort |

Strongin AY, Marmer BL, Grant GA and Goldberg GI (1993) Plasma membrane-dependent activation of the 72-kD type IV collagenase is prevented by complex formation with TIMP-2. J Biol Chem, 268, 14033–14039. | PubMed | ISI | ChemPort |

Strongin AY, Collier I, Bannikov G, Marmer BN, Grant GA and Goldberg GI (1995) Mechanism of cell surface activation of 72-kDa type IV collagenase. Isolation of the activated form of the membrane metalloproteinase. J Biol Chem, 270, 5331–5338. | Article | PubMed | ISI | ChemPort |

Takino T, Sato H, Shinagawa A and Seiki M (1995) Identification of the second membrane-type metalloproteinase (MT-MMP-2) gene from a human placenta cDNA library. J Biol Chem, 270, 23013–23020. | Article | PubMed | ISI | ChemPort |

Unkeless JC, Tobia A, Ossowski L, Quigley P, Rifkin DB and Reich E (1973) An enzymatic function associated with transformation of fibroblasts with oncogenic viruses. I. Chick embryo fibroblast cultures transformed by avian RNA tumor viruses. J Exp Med, 137, 85–111. | Article | PubMed | ISI | ChemPort |

Vassalli J-D, Baccino D and Belin D (1985) A cellular binding site for the M_r 55,000 form of the human plasminogen activator, urokinase. J Cell Biol, 100, 86–92. | Article | PubMed | ISI | ChemPort |

Wacher MP, Krutzsch HC, Liotta LA and Stetler-Stevenson WG (1990) Development of a novel substrate capture immunoassay for the detection of a neutral metalloproteinase capable of degrading basement membrane (type IV) collagen. J Immunol Methods, 126, 239–245. | Article | PubMed | ISI | ChemPort |

Ward RV, Atkinson SJ, Slocombe PM, Docherty AJP, Reynolds JJ and Murphy G (1991) Tissue inhibitor of metalloproteinases-2 inhibits the activation of 72 kDa progelatinase by fibroblast membranes. Biochim Biophys Acta, 1079, 242–246. | Article | PubMed | ISI | ChemPort |

Watanabe H, Nakanishi I, Yamashita K, Hayakawa T and Okada Y (1993) Matrix metalloproteinase-9 (92 kDa gelatinase/type IV collagenase) from U937 monoblastoid cells: correlation with cellular invasion. J Cell Sci, 104, 991–999. | PubMed | ISI | ChemPort |

Welgus HG, Jeffrey JJ and Eisen AZ (1981) The collagen substrate specificity of human skin fibroblast collagenase. J Biol Chem, 256, 9511–9515. | PubMed | ISI | ChemPort |

Werb Z, Mainardi C, Vater C and Harris ED (1977) Endogenous activation of latent collagenase by rheumatoid synovial cells. Evidence for a role of plasminogen activator. N Engl J Med, 296, 1017–1023. | PubMed | ISI | ChemPort |

Wilhelm SM, Collier IE, Marmer BL, Eisen AZ, Grant GA and Goldberg GI (1989) SV40-transformed human lung fibroblasts secrete a 92-kD type IV collagenase which is identical to that secreted by normal human macrophages. J Biol Chem, 264, 17213–17221. | PubMed | ISI | ChemPort |

Wong AP, Cortez SL and Baricos WH (1992) Role of plasmin and gelatinase in extracellular matrix degradation by cultured rat mesangial cells. Am J Physiol, 263, 1112–1118.

Main navigation

• Search EMBO Journal

  Search

• Advanced search

• The EMBO Journal

  • Home
  • Aims and scope
  • Current issue
  • Advance online publication
  • Web focuses
  • Archive:
  • Browse by issue
  • Browse by subject
  • Browse by category
  • Free online sample issue
  • Press releases

• Authors and Referees

  • Editorial process
  • Guide for authors
  • Submit an article
  • Guide for referees
  • Editors and Editorial Board
  • Contact editorial office

• Customer Service

  • Subscribe
  • Order sample copy
  • Purchase articles
  • Reprints and permissions
  • Contact NPG
  • Advertising