| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Tumor Biology |
Oral Cancer Research Center, Department of Biochemistry and Molecular Genetics, University of Alabama at Birmingham, Birmingham, Alabama, 35294-0005 [S. A., B. A. M., S. L. H., J. A. E.]; Department of Oral Biology, Indiana University, Indianapolis, Indiana, 46202 [L. J. W.]; and Research Genetics, Inc., Huntsville, Alabama 35801 [L. D. A-L.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Studies of MMP expression in oral SCCs have implicated mainly the secreted MMPs in invasion and metastasis. For example, individual studies have correlated the increased expression of MMPs -1, -2, -3, -9, and -13 in tissue sections of oral SCC with increased local invasion or incidence of lymph node metastases in the patients from whom they were derived (8, 9, 10, 11, 12) . In vitro experiments with established cell lines from oral SCCs have implicated the involvement of MMPs -1,-3, and -9 in the invasion of collagen gels and reconstituted basement membrane matrix (13, 14, 15, 16) . In several of these studies, epidermal growth factor and/or hepatocyte growth factor/scatter factor were required to initiate MMP-mediated invasion through three-dimensional matrices (17) . However, a role for MT1-MMP in oral cancer progression has received less attention. A limited number of studies have documented its expression in oral tumor specimens (12 , 18, 19, 20) , with one study detecting predominantly stromal expression of the mRNA (18) and another showing strong expression of the protein on tumor cells at the invasive edge (20) . Originally described as an activator for membrane-bound proMMP-2 (gelatinase A; 21 ), MT1-MMP has more recently been demonstrated to have matrix-degrading activity in its own right, including activity against interstitial collagen (22 , 23) .
In this study, we examined the interstitial collagen-degrading capacities of three established oral SCC cell lines derived from malignant lesions of the tongue, with the goal of identifying the relevant MMP(s). Collagen degradation was assessed with an assay developed to measure dissolution of reconstituted type I collagen fibrils by keratinocytes and fibroblasts (24, 25, 26, 27) , which is sensitive to low levels of cell-associated collagenase activity. Two of these cell lines, the SCC-25 and SCC-15 cells (28) , were capable of subjacent collagen degradation in the absence of exogenous growth factors, cytokines, or proteases such as trypsin or plasmin. Degradation was enhanced by the phorbol ester PMA and was strictly MMP-mediated as determined by its protease inhibitor profile. The third cell line, CAL 27 (29) , was incapable of collagen degradation under either basal or PMA-stimulated conditions. Limited MMP profiling by zymography and Western blotting was performed along with studies of MMP regulation by collagen culture and PMA. We present evidence that cell surface expression of active MT1-MMP is a requirement for pericellular collagen degradation in this system.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell Culture.
The SCC-25 (CRL-1628), SCC-15 (CRL-1623), CAL 27 (CRL- 2095), and HT-1080 (CCL-121) cell lines were purchased from American Type Culture Collection. The SCC-25, SCC-15, and CAL 27 cell lines were derived from SCCs of the tongue (28
, 29)
. Cell lines were cultured in DMEM supplemented with 10% (v/v) fetal bovine serum (HyClone, Logan, Utah), penicillin (10 units/ml)-streptomycin (10 µg/ml), and gentamicin (10 µg/ml; Sigma Chemical Co.) at 37°C in 10% CO2 in air. Each cell line was used within 15–20 passages after initiation of cultures from American Type Culture Collection.
Isolation of Rat-Tail Tendon Type I Collagen.
Type I collagen was isolated by standard procedures from tail tendons of Wistar rats 4–6 weeks of age as previously described in detail (33
, 34)
. Lyophilized collagen was stored at -80°C, and stock solutions were prepared as needed. Acid-soluble type I collagen forms fibrils at neutral pH which (at 37°C) are resistant to degradation by proteinases other than collagenases, similar to native collagen type I in vivo.
Preparation of Collagen-Coated Plates.
Six-well culture plates (35 mm diameter wells; Corning Glass Works, Corning, NY) were coated with a film of reconstituted type I collagen by a modification of techniques described previously (24, 25, 26, 27)
. Briefly, a stock solution of rat tail tendon type I collagen in 13 mM HCl was diluted and mixed with neutralizing phosphate buffer to a final concentration of 300 µg/ml. Aliquots of 1.5 ml/well (50 µg/cm2) were dispensed, and collagen fibrils were formed by heat gelation at 37°C for 2–4 h. Collagen gels were then air-dried down to a film and washed extensively with sterile distilled water to remove salt precipitates. Tissue culture plates (100 and 150 mm; Falcon) were coated similarly, maintaining a coating concentration of 50 µg/cm2.
Cell-Mediated Collagen Fibril Dissolution.
Cells were detached from subconfluent cultures in 0.25% trypsin, washed, and finally resuspended in serum-free DMEM supplemented with 0.1% BSA (DMEM/BSA). Cells were then seeded in collagen-coated, six-well plates as a droplet (60,000 cells in 50 µl) in the central part of each well. After 3 h in a humidified 37°C incubator, 2 ml of DMEM/BSA were added to each well, with or without PMA (160 nM; final DMSO concentration in media, 0.016%) plus other reagents/inhibitors as indicated. Culture was continued for up to 3 days, during which cells formed a compact monolayer and did not migrate outward from the colony. Collagen degradation by the cells was examined after removing cells with 0.25% trypsin/1 mM EDTA for 10 min before the addition of Triton X-100 to a final concentration of 0.3%. Wells were then washed and residual collagen stained with a solution of Coomassie Blue (0.2%) in 20% methanol/7.5% acetic acid. The fibrillar collagen film is resistant to trypsin at 37°C, therefore cellular collagen degradation was visualized as clear areas against a blue background. Samples done in duplicate were indistinguishable from one another visually. In selected experiments, the degree of clearing in equivalent areas under the cell button was quantitated with the Bio-Rad Gel Doc1000 using Molecular Analyst software (Bio-Rad, Richmond, CA).
Preparation of Cell Lysates.
Cells were plated at 70–80% confluence on substrate-coated or uncoated dishes in serum-containing DMEM (DMEM/+) and incubated for 16 h. Adherent cells were rinsed and then incubated in DMEM/BSA, with or without PMA for the duration indicated. Alternatively, in some experiments (Fig. 5A)
, cells were resuspended and plated directly into DMEM/BSA. Lysates were harvested by scraping the cells into a small volume of NP40 lysis buffer [50 mM Tris-Cl (pH 7.4),150 mM NaCl,1 mM CaCl2,1 mM MgCl2, 0.5% NP40, 2 µg/ml aprotinin, 10 µg/ml leupeptin, 1 mg/ml 4-(2-aminoethyl)-benzenesulfonyl fluoride, and 1 µM pepstatin A]. Extracts were incubated on ice for 45 min, then clarified by microfuge at 14,000 x g for 15 min at 4°C. Protein concentrations of lysates and membranes (below) were determined using the bicinchoninic acid Protein Assay Kit (Pierce, Rockford, IL).
|
Gelatin Zymography.
Samples were mixed with SDS sample buffer without heating or reduction and applied to 10% polyacrylamide gels copolymerized with 1 mg/ml gelatin. After electrophoresis, gels were washed for 1 h at room temperature in buffer containing 2.5% (v/v) Triton X-100 in 50 mM Tris-Cl (pH 7.5). Gels were then incubated at 37°C in 50 mM Tris-Cl (pH 7.5) with 5 mM CaCl2 and 1 µM ZnCl2 for 24 h. After staining with Coomassie Blue (0.2%), zones of gelatinolytic activity were detected as clear bands against a blue background.
SDS-PAGE and Western Blotting.
Samples (cell lysates, membranes, or conditioned medium) were mixed with reducing SDS sample buffer, heated, and electrophoresed on 10% polyacrylamide gels according to the method of Laemmli (35)
. Gels were then electroblotted onto polyvinylidene difluoride membrane for immunoblot analysis. After blocking for 1 h with 5% nonfat dry milk in PBS/0.1% Tween 20 (PBST; Bio-Rad), blots were probed with primary antibody diluted in 0.5% milk in PBST (for 1.5 h at 25°C) and then HRP-conjugated goat antimouse or antirabbit IgG (Pierce), diluted 1:50,000. Signals were detected by chemiluminescence using the ECL Western blotting detection reagents from Amersham Pharmacia. Where indicated, blots were stripped of antibodies after signal detection by incubation for 30 min at 50°C in 2% SDS in Tris-Cl (pH 6.7) with 2 mM ß-ME and then reblocked and reprobed with a different antibody.
Surface Biotinylation and Immunoprecipitation.
Adherent cells were rinsed with ice-cold PBS, then incubated with 0.5 mg/ml of water-soluble, cell-impermeable EZ-Link Sulfo-N-hydroxysuccinimide-long chain-biotin (Pierce, Rockford, IL) in PBS for 40 min at 4°C. After rinsing dishes with cold Tris-buffered saline [50 mM Tris-Cl (pH 7.5) and 150 mM NaCl], the reaction was quenched with 0.1 M glycine for 10 min. Cells were finally lysed with NP40 lysis buffer as described above.
For immunoprecipitation, a known quantity of lysate (300–1000 µg) was precleared with rabbit IgG immobilized onto Protein A-Sepharose CL-4B (Amersham Pharmacia). Precleared lysate was incubated for 2 h at 4°C with 5 µg anti-MT1-MMP (AB815) immobilized onto protein A-Sepharose beads. Immune complexes on beads were washed 5 times with cold 0.1% Triton X-100 in 50 mM Tris-Cl (pH 7.4), 300 mM NaCl. Complexes were eluted for 5 min at 95°C in reducing 2x SDS sample buffer, and eluates were electrophoresed and transblotted as described above. Biotinylated proteins were detected with ExtrAvidin Peroxidase Conjugate (Sigma Chemical Co.) and ECL. For negative controls, 5 µg of rabbit IgG was substituted for anti-MT1-MMP. Alternatively, samples were incubated with protein A-Sepharose beads in the absence of antibody; a similar pattern of nonspecific bands was seen in either case.
Immunofluorescence.
Cells were seeded on glass coverslips in DMEM/+ and grown to 60% confluency. The last 15 h of culture was in serum-free DMEM supplemented with 160 nM PMA. Cells were washed and fixed with phosphate-buffered 3% formaldehyde for 30 min at 25°C. After fixation, cells were washed with PBS and blocked with 10% normal goat serum (Zymed Laboratories; San Francisco, CA) in PBS for 30 min. MT1-MMP polyclonal antibody (AB8103) was diluted to 7 µg/ml in 10% normal goat serum in PBS and incubated with cells for 1 h. After washes, the cells were incubated with Texas-Red dye-conjugated goat antimouse secondary antibodies (Molecular Probes, Eugene, OR) for 1 h at 25°C. Slides were mounted in Prolong Antifade Medium (Molecular Probes) and stored at 4°C. Images were captured using an Olympus IX70 microscope using the Texas-Red filter and IP Lab software (Signal Analytics; San Jose, California).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(top) shows a typical degradation assay in which SCC-25 cells were cultured as a central colony on collagen films in the absence or presence of 160 nM PMA. Collagen degradation was assessed at 24, 48, and 72 h after seeding. After 24 h, both basal and PMA-stimulated collagen degradation were visible. Degradation increased 
2–3-fold between day 1 and day 3 under both conditions and did not extend beyond the boundaries of the colonies (see also bottom right panel). Below in Fig. 1A
, the stained residual collagen under the PMA-treated cell buttons is shown at a higher magnification. The left and middle panels are the central portions of cell colonies after removal of cells, whereas the right panel shows the edge of the colony at day 3. Consistent with previous studies (25
, 27)
, degradation occurred as focally discrete areas of lysis that merged over time. Enhancement of collagen degradation by PMA was already detectable within the first 24 h. In experiments (data not shown) in which PMA was removed after the first 24 h (media replaced with DMEM/BSA), there was still significant (2-fold) enhancement of collagen breakdown at day 3 relative to 3 days without PMA. This result indicated that the effects of the first 24 h of PMA were maintained for at least 2 more days.
|
2.5-fold by day 3 (data not shown). PMA enhanced basal degradation from 1.5-fold to 2.5- fold in these cells (data not shown).
In contrast to the SCC-25 and SCC-15 cells, the CAL 27 cells did not degrade the underlying collagen substrate in this assay (Fig. 1B)
, although they attached and spread on collagen as readily as the other two cell lines (data not shown). CAL 27 cells were assayed for collagen dissolution over 3 days, as in Fig. 1A
. There was no detectable degradation by the cells at any time point, even when similar experiments were allowed to proceed for 7 days in the presence of PMA (data not shown). Lack of degradation by these cells was verified by phase contrast microscopy (data not shown).
Fig. 1C
shows two independent assays of SCC-25 cells, illustrating the apparently complete inhibition of collagen degradation when natural and synthetic inhibitors of MMPs were included. In the top panel, collagen degradation after 3 days by PMA-treated SCC-25 cells (CTL, left well) was inhibited by the synthetic broad-spectrum MMP inhibitor BB-94 at 21 nM (right well). Similarly, the bottom panel shows PMA-treated SCC-25 cells in the absence (CTL, left well) or presence (right well) of 0.5 µg/ml TIMP-2. In contrast to the apparently complete inhibition by TIMP-2, TIMP-1 at 0.5 µg/ml (central well) resulted in only partial (50–60%) inhibition, illustrating the direct contribution of MT1-MMP to collagen degradation. Similar patterns of inhibition were also observed with SCC-15 cells (data not shown).
Serine and cysteine proteinases may contribute to the degradation of ECM components directly or indirectly by activation of latent MMPs (7
, 36)
. To determine whether these classes of proteinases participate in the extracellular degradation of collagen fragments, we tested a variety of inhibitors in assays with SCC-25 and SCC-15 cells. We found that the serine proteinase inhibitor aprotinin (200 µg/ml; Fig. 1D
), the plasmin inhibitor EACA (2–10 mM), and the cysteine proteinase inhibitor E-64 (10 µg/ml; data not shown) were each ineffective at blocking collagen dissolution. Likewise, leupeptin (10 µg/ml), an inhibitor of serine and cysteine proteinases with trypsin-like specificity, was also ineffective (data not shown). Taken together, these data show that subjacent collagen degradation by SCC-25 and SCC-15 cells occurred in the absence of growth factors, cytokines, and exogenous proteinases, was enhanced by PMA, and was strictly MMP-mediated.
Activation of Endogenous MMP-2 (Gelatinase A) by Culture on Collagen I Films.
Degradation was confined to the subjacent collagen, which implicated cell surface associated MMPs as opposed to MMPs in the bulk conditioned media. We began MMP profiling by identification of endogenous gelatinases in SCC-25 cells and examined their regulation by culture on type I collagen films. Fig. 2A
shows a zymographic analysis of cell membranes, lysates, and conditioned media of SCC-25 cells cultured on either tissue culture plastic (Lanes 1, 3, and 5) or collagen I films (Lanes 2, 4, and 6). Endogenous gelatinases comigrated with purified human proMMP-2 (72 kDa gelatinase A) and proMMP-9 (92 kDa gelatinase B; Chemicon zymography standards) in control Lanes (not shown). Culture on collagen increased total MMP-2 levels and up-regulated the activation of proMMP-2 to the fully active 62 kDa species relative to culture on plastic. (The molecular masses of 66, 64, and 62 kDa referred to below are those of the nonreduced pro-, intermediate, and active MMP-2 species, respectively, in zymograms.) The identity of the 66 kDa gelatinase in SCC-25 cells was verified as proMMP-2 using specific antibodies in Western blot analysis (data not shown).
|
confirmed the requirement for fibrillar type I collagen in inducing MMP-2 activation, as observed by others (39, 40, 41)
. SCC-25 cells were plated in parallel on tissue culture plastic (Fig. 2B
, Lane 1) and on dishes coated with poly-D-lysine (Lane 2), collagen I film (Lane 3), laminin-1 (Lane 4), fibronectin (Lane 5), and a thin coat of type I collagen (10 µg/ml; Lane 6). Lysates were collected after 48 h (no PMA) and compared by gelatin zymography. Whereas proMMP-2 was present in all culture conditions, the film of fibrillar type I collagen (Lane 3) was the only substrate to induce activation of cell associated proMMP-2. Increased levels of both MMP-2 and MMP-9 were apparent in this sample.
As observed in Fig. 2, A and B
, MMP-9 was detectable in conditioned medium, membranes, and lysates of SCC-25 cells. Although PMA did stimulate proMMP-9 production (Fig. 5A)
, active forms of MMP-9 were not readily observed under any conditions in SCC-25 cells.
Detection of MT1-MMP in Oral SCC Cells.
We next examined the cell lines for expression of MT1-MMP to examine its role in collagen degradation. Three major forms of MT1-MMP have been described in cells, including a proform of Mr 63,000 (sometimes seen as Mr 63,000/65,000 doublet), an active species at Mr 55,000–60,000, and a catalytically inactive truncated fragment of Mr 43,000–45,000, which includes the hinge domain (40
, 42, 43, 44, 45)
. To identify the forms present in the oral SCC cell lines, we analyzed cell membranes and lysates by Western immunoblot using a polyclonal antibody (AB815) to the hinge domain. Fig. 3A
shows typical Western blots of membranes from PMA-treated cells cultured on plastic (Lanes 1–3) and of lysates from collagen cultured cells (Lanes 4–6). In membranes from both HT-1080 fibrosarcoma cells (used as a positive control; Lane 1) and SCC-25 (Lane 2), a strong band of immunoreactivity was observed at Mr 57,000 (thick arrow), which corresponds in molecular weight to the active species (40
, 45)
. This band was present at greatly reduced levels in the CAL 27 membranes (Lane 3), although equal quantities of total membrane protein were loaded per lane. A minor band of Mr 63,000 (open arrow) was also detectable in Lanes 1–3, possibly the pro-form. Other membrane preparations (SCC-25 and CAL 27; data not shown) have contained variable levels of proMT1-MMP at Mr 63,000/65,000 as seen in Lanes 4–6, although the active Mr 57,000 form in SCC-25 was always predominant. Consistent with previous studies, PMA-treated HT-1080 cells contained a lower molecular weight fragment of MT1-MMP corresponding to the Mr 43,000 truncated form (Fig. 3A
, Lane 1, thin arrow; Refs. 42
, 43
, 45)
. A slightly smaller fragment was also observed in Lanes 1–3 (more prominent in CAL 27), perhaps the result of additional processing of the Mr 43,000 form. In separate Western blots of lysates from the three oral SCC cell lines (Fig. 3A
, Lanes 4–6), the proMT1-MMP doublet at Mr 63,000/65,000 was prominent. To confirm that the doublet was proMT1-MMP, Western blots of SCC-25 and CAL 27 lysates that had been probed with AB815 were stripped and reprobed with a prodomain-specific antibody (AB8101; Chemicon); AB8101 reacted with a band corresponding to the lower band of the doublet in both cell lines (data not shown). Therefore pro-, active, and processed species of MT1-MMP are detectable in the oral SCC cells. Membrane preparations of SCC-25 cells are enriched in the active species.
|
(left panel) shows the typical pattern of bands present in immunoprecipitates from surface biotinylated cells, detected with avidin-HRP and ECL. Surface proteins on adherent SCC-25 cells were biotinylated, and then membrane extracts were immunoprecipitated with AB815 as described (Fig. 3B
, Lane 2). Membrane extracts of biotinylated HT-1080 cells (Lane 1) were included for comparison. In both immunoprecipitates, avidin-HRP revealed a broad biotinylated band corresponding in molecular weight to the Mr 57,000–60,000 active form of MT1-MMP. The truncated Mr 43,000 fragment (arrowhead) also stained prominently in HT1080 cells. Separate aliquots of the corresponding nonimmunoprecipitated membrane extracts were Western blotted and probed with AB815 (Fig. 3B
, Lanes 4 and 5; HT-1080 and SCC-25, respectively) to compare the migration of the active species with the biotinylated bands (double-headed arrow). In comparison, the proform (open arrow) present in lysate from SCC-25 cells (Fig. 3B
, Lane 6) was clearly separated from the active form in these gels. We conclude that the major surface biotinylated, immunoprecipitated band from SCC-25 cells is a processed (and most likely active) form of MT1-MMP.
Effects of PMA on MMP-2 Activation and MT1-MMP Processing.
PMA amplifies the basal degradation of collagen, apparent after the first 24 h. To determine whether PMA also enhanced collagen-induced MMP-2 activation, SCC-25 cells were cultured on collagen films in the presence or absence of PMA (Fig. 4A)
. Lysates were collected at 6 h (Lanes 1 and 2) and 30 h (Lanes 3 and 4) and then analyzed by zymography. After 30 h of PMA treatment (Fig. 4A
, Lane 4), collagen-induced MMP-2 activation was noticeably enhanced relative to DMSO control (Lane 3). The level of proMMP-2 in the PMA-treated sample was correspondingly reduced. In contrast, enhancement by PMA was not yet apparent after only 6 h (Lane 2 versus Lane 1).
|
. Collagen-cultured SCC-25 cells were incubated in the presence (Lanes 2, 4, 6, and 8) or absence (Lanes 1, 3, 5, and 7) of PMA and sampled at several time points by surface biotinylation and then immunoprecipitation of lysates with AB815. In addition to the surface-labeled active MT1-MMP band that was present in all samples (thin arrow), a broad biotinylated band of Mr 43,000 (thick arrow) was immunoprecipitated only from cells treated with PMA for 24 h (Lane 8). The appearance of this band in SCC-25 cells corresponds approximately to the time (30 h; see Fig. 4A
, Lane 4) when enhanced MMP-2 activation was observed in PMA-treated cells. These results were also observed with the other collagen-degrading cell line SCC-15 (Fig. 5B
, Lane 3), although the intensity of the Mr 43,000 band was not as strong as in HT-1080 cells. Taken together, these data show that PMA treatment of SCC-25 cells accelerates collagen-induced MMP-2 activation, apparent by the first 24 h, and this corresponds in time to the appearance on cell surfaces of a Mr 43,000-processed form of MT1-MMP.
Comparison of CAL 27 and SCC-15 with SCC-25 Cells.
We next examined the SCC-15 and CAL 27 cells for collagen-induced MMP-2 activation and MT1-MMP expression. Each cell line was plated in parallel onto plastic or collagen-coated wells and cultured for 48 h in the presence or absence of PMA. Lysates from each sample were analyzed by gelatin zymography (Fig. 5A)
. SCC-25 cells (Lanes 1–4) showed the characteristic increase in proMMP-2 activation on collagen relative to plastic (Lane 3 versus Lane 1) with enhancement by PMA (Lane 4 versus Lane 3). Intermediate forms of MMP-2 (Mr 64,000) are also visible in these gels. Likewise, collagen culture of SCC-15 cells resulted in a pronounced increase in endogenous proMMP-2 activation (Fig. 5A
, Lane 11 versus Lane 9) and an apparent increase in levels of total MMP-2. PMA enhanced the activation of MMP-2 on both plastic and collagen in these cells (Lane 10 versus Lane 9 and Lane 12 versus Lane 11). Synthesis of MMP-9 was stimulated by PMA in both SCC-25 (Lanes 2 and 4) and SCC-15 cells (Lanes 10 and 12), as observed in other systems (15)
, and active MMP-9 was detectable in PMA-treated SCC-15 cells (Lanes 10 and 12). In marked contrast, lysates from CAL 27 cells (Fig. 5A
, Lanes 5–8) contained no detectable gelatinase activity under any conditions. In similar experiments, the CAL 27 cells did not express or activate gelatinases on other ECM substrates including laminin-1 and fibronectin, although proforms of MT1-MMP were detectable by Western blot in each sample (data not shown).
Cell surface expression of MT1-MMP was compared in the three cell lines by surface biotinylation and immunoprecipitation (Fig. 5B)
. Cells were cultured on collagen for 24 h in the presence of PMA and then surface biotinylated, lysed, and immunoprecipitated with AB815. Equal quantities of protein (1000 µg/sample) were immunoprecipitated from SCC-25 and CAL 27 lysates (Fig. 5B
, Lanes 1 and 2), whereas only 400 µg could be immunoprecipitated from SCC-15 lysate (Lane 3). Immunoprecipitates were electrophoresed, transblotted, and probed with avidin-HRP. In both SCC-25 (Lane 1) and SCC-15 (Lane 3) cells, active (Mr 57,000) MT1-MMP was prominently stained (thin arrow). The lower molecular weight Mr 43,000 band was also visible in the SCC-15 sample (Lane 3, thick arrow), although it was not detectable in the SCC-25 sample in this experiment. In contrast to the strong staining of active MT1-MMP in SCC-25 and SCC-15 cells, this band was barely visible in the CAL 27 cells (Lane 2) under the same conditions. Parallel cultures of CAL 27 cells, which continued for 48 h before biotinylation, did not contain increased levels of surface-labeled Mr 57,000 MT1-MMP (data not shown). These data are also consistent with the Western blot of membrane samples (Fig. 3A)
, which showed reduced expression of active MT1-MMP in CAL 27 cell membranes relative to SCC-25 membranes.
Reduced expression of MT1-MMP on the surface of CAL 27 cells was also demonstrated by immunofluorescence staining (Fig. 5C)
. SCC-25, CAL 27, and SCC-15 cells cultured on glass coverslips were stained with a polyclonal antibody against the hemopexin domain of MT1-MMP (AB8103) and then Texas-Red-labeled secondary antibody. High levels of membrane staining as well as stippled intracellular staining were evident in SCC-25 cells (Fig. 5C
, top). A similar intensity of immunofluorescence was observed in SCC-15 cells, with punctate membrane staining and stippled intracellular staining (bottom panel). In comparison, the CAL-27 cells showed much less total MT1-MMP staining (middle panel), with limited areas of weak punctate staining in membranes. The immunofluorescence data are in agreement with the surface biotinylation experiments (Fig. 5B)
, which show that the CAL-27 cells express lower levels of MT1-MMP on their surfaces relative to the collagen-degrading cells.
Inhibition of MT1-MMP Activation Inhibits Collagen Degradation.
To directly implicate MT1-MMP in collagen degradation, we assessed degradation in the presence of a cell-permeable synthetic furin inhibitor, decanoyl-Arg-Val-Lys-Arg-CMK (30
; Fig. 6A and B
). CMK at 10–100 µM has been reported to inhibit intracellular processing of proMT1-MMP to its active form, thereby inhibiting MT1-MMP-mediated functions (38
, 44
, 46)
. Fig. 6A
shows dose-dependent inhibition of collagen degradation at increasing concentrations of the furin inhibitor in PMA-treated SCC-25 cells, with 50–90% inhibition observed between 10 and 25 µM. Collagen degradation of SCC-15 cells was similarly inhibited by CMK, with >50% inhibition at 10 µM (data not shown). To confirm the effect of the furin inhibitor on MT1-MMP processing, lysates and conditioned media were collected and analyzed from collagen-cultured SCC-25 cells treated with 100 µM CMK for 48 h (Fig. 6B)
. At this concentration of furin inhibitor, collagen degradation was inhibited nearly 100% (left panel). Lysates of parallel cultures, which were surface biotinylated and immunoprecipitated with anti-MT1-MMP, showed reduced surface-labeled active form and increased surface-labeled proform in CMK-treated cells relative to controls (Lane 2 versus Lane 1). In agreement with this data, nonimmunoprecipitated lysates, which were Western blotted and probed for MT1-MMP, also showed reduced levels of active form and increased proform in CMK-treated cells (Lane 4 versus Lane 3). Correspondingly, zymography of conditioned media from the same cultures (Lanes 5 and 6) and lysates (Lanes 7 and 8) revealed much lower levels of active MMP-2 in CMK treated cells compared with controls (Lane 6 versus Lane 5, and Lane 8 versus Lane 7, respectively). Therefore, we conclude that CMK inhibits collagen degradation through the inhibition of proMT1-MMP processing.
|
shows a representative Western blot of lysates from SCC-25 and CAL 27 cells collected after 2 days of culture on either fibronectin (FN) or collagen films (Coll) in the absence of PMA, which was probed sequentially with the indicated antibodies. In these experiments, culture on fibronectin yielded results equivalent to those on tissue culture plastic. The top panel of Fig. 7
shows Mr 52,000 proMMP-1 in lysates of both cell lines under either condition. Likewise, stripping and reprobing of this blot with anti-MMP-13 Ab revealed expression of Mr 60,000 proMMP-13 in both cell lines (Fig. 7
, middle panel). Interestingly, in the CAL 27 cells, MMP-13 was apparently induced several-fold by collagen culture relative to culture on fibronectin (Lane 4 versus Lane 3). Lysates of SCC-15 cells also had comparable levels of proMMP-1 and proMMP-13 under all culture conditions (data not shown). In separate experiments (data not shown), Western blots of 20x concentrated serum-free conditioned media (48 h) detected secreted proMMP-1 and proMMP-13 from all three cell lines when probed with monoclonal Abs IM35L (anti-MMP-1) and ID3 (anti-MMP-13). As seen in lysates, collagen culture up-regulated the secretion of proMMP-13 from CAL 27 cells (data not shown). Taken together, these data reveal constitutive expression and secretion of proMMP-1 and proMMP-13 in the collagen-degrading SCC-25 and SCC-15 cells as well as the nondegrading CAL 27 cells, although active species were not detectable by Western blot even in the presence of PMA (data not shown).
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The SCC-25 and SCC-15 cells degraded subjacent collagen fibrils over a period of 3 days in the absence of exogenous growth factors, cytokines, or proteases, which have been used to induce/activate secreted MMPs in other systems (14
, 16
, 17
, 24, 25, 26
, 49) . Basal collagen degradation was enhanced 1.5–3-fold by the tumor- promoting phorbol ester, PMA. Both basal and PMA-stimulated collagen degradation were strictly MMP-mediated, as observed with keratinocytes and fibroblasts (25
, 26)
. The CAL 27 cells, in contrast, seemed incapable of collagen dissolution even in the presence of PMA. All three cell lines had in common the constitutive expression of collagenase-1 (MMP-1) and collagenase-3 (MMP-13), which are secreted MMPs with activity against triple-helical collagens (7
, 50)
. However, only the collagen-degrading SCC-25 and SCC-15 cells expressed detectable levels of active MT1-MMP on their surfaces, along with gelatinases A and B (MMP-2 and -9). Concomitantly, these cells were capable of MT1-MMP-mediated proMMP-2 activation, which was induced by collagen culture and further enhanced by PMA. When processing of the MT1-MMP zymogen was inhibited by the furin inhibitor peptide, collagen degradation was blocked. Taken together, we conclude that pericellular collagenolytic activity in oral SCC cells is mediated by membrane-associated MMPs, and that MT1-MMP is essential.
Tumor-promoting phorbol esters such as PMA induce the transcriptional activation of several secreted MMPs and MT1-MMP (7 , 15 , 32 , 51 , 52) . The up-regulation of endogenous MMPs by PMA allowed us to detect processing of cell-surface MT1-MMP to the Mr 43,000 species, providing evidence of its functional activity as seen in fibroblasts and ovarian carcinoma cells (40 , 42) .
Tumor cell invasion of ECM barriers is a highly complex process involving multiple interactions with host components (3 , 4) . In particular, dynamic adhesion of tumor cells to ECM, proteolytic modification of the matrix, and migration through the proteolyzed region must be coordinated both spatially and temporally for effective invasion to occur. Experimental analysis of such a complex process can be simplified somewhat by focusing on one aspect of invasion, i.e., proteolysis of matrix barriers or chemotactic migration of tumor cells. Our collagen degradation assay provides an assessment of the tumor cells’ proteolytic capacities in the absence of a migratory stimulus. Interstitial collagens are the most abundant proteins of the ECM, and the endogenous capacity of tumor cells to degrade this matrix may confer an advantage over nondegrading cells in situations where available growth factors and cytokines were limiting. Head and neck SCCs, which include oral SCC, are highly invasive cancers and a major cause of cancer morbidity and mortality (1 , 2 , 53) . Therefore, understanding the molecular mechanisms by which oral SCC cells proteolyze and invade interstitial collagen barriers is of vital importance. Preliminary data using Transwell invasion assays have shown that although all three cell lines can migrate at comparable rates over uncoated filters toward a chemotactic stimulus (10% fetal bovine serum in DMEM), their capacity to invade a film of type I collagen and translocate to the underside of the filter correlates with their capacity to proteolyze type I collagen in the two-dimensional degradation assay (data not shown). Furthermore, as with collagen degradation, invasion of collagen by SCC-25 and SCC-15 cells is inhibited >99% by TIMP-2 and only 50–60% by TIMP-1 (data not shown), supporting the hypothesis that active MT1-MMP is essential for this process.
The SCC-25 cells were shown in the current and in previous studies to express MMPs -1, -2, -3, and -9 under routine culture conditions (54, 55, 56) . In addition, we demonstrated constitutive expression of MT1-MMP and MMP-13 along with the same repertoire of MMPs in the SCC-15 cells. Upon discovering that MT1-MMP and MMP-2 were reduced or absent in the nondegrading CAL 27 cells, we examined their regulation in the degrading cells by collagen culture and PMA, conditions which mimicked the degradation assay. Culture on type I collagen films resulted in increased production and activation of cell-associated MMP-2 in both SCC-25 and SCC-15, which was an indirect indicator of increased MT1-MMP activity. PMA accelerated this process within 24 h and concomitantly up-regulated the processing of MT1-MMP to a Mr 43,000 fragment. Because MT1-MMP has been demonstrated to proteolyze type I collagen in cellular assay systems (48) , we hypothesized that MT1-MMP may be responsible for collagen degradation by the oral SCC cells either alone or in concert with MMP-2. In support of this hypothesis, inhibition of degradation by TIMP-2 seemed complete, whereas inhibition by TIMP-1 was only 50–60%. Because TIMP-1 is known to be a poor inhibitor of MT1-MMP (57) , residual degradation in the presence of TIMP-1 is directly mediated by MT1-MMP. Blocking the processing of the MT1-MMP zymogen to its active form with the furin inhibitor peptide resulted in near complete (>99%) inhibition of collagen degradation, providing additional evidence of the involvement of MT1-MMP. Taken together, we conclude that MT1-MMP has a direct role in the proteolysis of type I collagen, although the concerted action of MT1-MMP with other MMP(s)—possibly MMP-2—is required for complete clearance of collagen through to the plastic.
Culture of several different cell types on or within three-dimensional type I collagen gels has been shown to induce the MT1-MMP-mediated activation of proMMP-2 (17 , 38, 39, 40, 41 , 58) . In most of these cells, collagen culture up-regulates MT1-MMP mRNA and/or protein levels. These in vitro data, together with in vivo data from invasive tumors showing coexpression of MT1-MMP with MMP-2 and/or collagen I has led to the proposal that this mechanism contributes to tumor invasion and metastasis (18 , 58, 59, 60, 61) . Recent studies with tumor tissue from head and neck SCCs (including oral SCC) have shown overexpression of MT1-MMP mRNA in tumor tissue relative to corresponding normal tissue (20) , and MT1-MMP protein was detected in the tumor cells (12, 20) . High expression levels of MT1-MMP protein, along with its colocalization with MMP-2, was linked to the more invasive and metastatic cases (12) . Therefore, the collagen-induced proMMP-2 activation observed in vitro in the SCC-25 and SCC-15 cells may reflect their aggressiveness in vivo. The increased processing of MT1-MMP to a Mr 43,000 fragment when PMA was added to collagen cultured cells is similar to the effects of PMA on HT-1080 cells, and the presence of this form has been correlated with increased activation of MMP-2 (40 , 42) . Our studies have now demonstrated this correlation in the oral SCC cells as well. Because this fragment is catalytically inactive, it is believed to represent down-regulation of the MT1-MMP cycle of activity on the cell surface (42 , 43 , 62) .
MT1-MMP and stromelysin-3 are unique among the MMPs in having an RXKR recognition motif for furin and furin-like convertases at the COOH-terminal end of the propeptide domain (21 , 63) . Processing of proMT1-MMP by furin (64) is believed to be the major mechanism for conversion of the zymogen to its active form (65 , 66) . In previous studies, the furin inhibitor peptide CMK at 10–100 µM reduced intracellular processing of proMT1-MMP and inhibited both MMP-2 activation and in vitro invasion (38 , 44 , 46) . In our system, additional confirmation of the involvement of MT1-MMP in pericellular collagen degradation came from the clear dose-dependent inhibition of degradation in the presence of the furin inhibitor peptide. The corresponding increase in proMT1-MMP and reduced active species in peptide-treated cells confirmed the mechanism of action of the inhibitor.
We cannot explain the relative lack of active, surface-expressed MT1-MMP on CAL 27 cells with the available data. Using RT-PCR, furin mRNA was detected in all three cell lines (data not shown). We also amplified and sequenced RT-PCR products corresponding to MT1-MMP mRNA (first 732 nucleotides) from SCC-25 and CAL-27 cells and found that the sequences showed no mutations or other changes in the propeptide-coding region relative to wild-type human sequence (Ref. 67
; data not shown). Therefore, there is no evidence of a deficiency in the furin processing system in CAL 27 cells, which leaves open the question of why active MT1-MMP does not accumulate on their cell surfaces. A recent study demonstrated that optimal levels of TIMP-2 were required to stabilize the active form of MT1-MMP on cell membranes, and in the absence of sufficient TIMP-2, the active species was rapidly degraded by autocatalysis to the Mr 43,000 fragment (45)
. In agreement with this, Western blot detection of MT1-MMP in membranes showed a reduced level of the active Mr 57,000 form and increased levels of a lower molecular weight fragment in CAL 27 as compared with SCC-25 (Fig. 3A)
. Collagen mediated induction of MT1-MMP results from signaling pathways initiated by interaction of ß1 integrins with fibrillar collagen (37
, 40)
. All three cell lines express integrins 
2ß1 and 3ß1 (data not shown), so the failure to induce surface expression of active MT1-MMP cannot be explained by lack of collagen receptors.
Pericellular collagen degradation initiated by MT1-MMP in oral SCC cells is consistent with recent literature demonstrating its collagenolytic activity in transfected MDCK and COS-1 cells (48) . Pericellular degradation of laminin-5 and gelatin by endogenous MT1-MMP, alone or in concert with MMP-2, has also been demonstrated in carcinoma cells (68 , 69) . The focusing of collagenolytic and gelatinolytic activities to the membrane, either with integral membrane MMPs or membrane-associated MMPs, is increasingly recognized as the means by which cells invade ECM barriers (62) .
A key role for MT1-MMP, and possibly MMP-2, does not preclude the participation of the secreted collagenases MMP-1 and MMP-13 in collagen dissolution. MT1-MMP and MMP-2 have been demonstrated to activate proMMP-13 on cell surfaces (70) . Perhaps in SCC-25 and SCC-15 cells, this activating environment produces active MMP-13 and retains it on cell surfaces to an extent not possible in the CAL 27 cells. In vitro, active MMP-13 can activate proMMP-9 (71) , which would enhance further the pericellular gelatinolytic activity, if it occurred on cell surfaces. Therefore, we favor a model in which active MT1-MMP is at the "bottleneck" of an MMP activation cascade, activating proMMP-2 upon collagen culture and then possibly activating proMMP-13 in concert with active MMP-2. The combined collagenolytic and gelatinolytic activities of MT1-MMP, MMP-2, and MMP-13 could then dissolve the collagen fibril film. CAL 27 cells do not express sufficient active MT1-MMP to initiate such a cascade. They do express MMP-1 and MMP-13 constitutively, and MMP-13 is induced during collagen I culture, similar to human skin fibroblasts (72) . However, in the absence of a threshold level of active MT1-MMP, MMP-13 probably remains latent.
In conclusion, we postulate that tumor cells with an endogenous capacity for collagenolysis would be more independent of host factors at an early stage of progression and have an advantage over tumor cells without this capacity. Our data illustrate the potential importance of tumor cell-derived MT1-MMP in progression of oral SCC cells.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported by USPHS Grants P50 DE08228 (to S. M. M.), R01 DE10631 (to J. A. E.), 1P50 DE/CA 11910-01 (to J. A. E.), and AR44701 (to L. J. W.). Synthesis of oligonucleotide primers for RT-PCR was supported by National Cancer Institute Grant CA-13148 to the University of Alabama at Birmingham Comprehensive Cancer Center. Support for DNA sequence analysis computer programs used in this work was provided by the NIH Centers for AIDS Research Program Grant AI27767. 
2 To whom requests for reprints should be addressed, at University of Alabama at Birmingham, Department of Biochemistry and Molecular Genetics, 1530 Third Avenue South, Room 460 MCLM, Birmingham, AL 35294-0005. Phone: (205) 934-4734; Fax: (205) 934-0758; E-mail: jengler{at}bmg125.cmc.uab.edu
3 The abbreviations used are: SCC, squamous cell carcinoma; ECM, extracellular matrix; MMP, matrix metalloproteinase; MT1-MMP, membrane type I-MMP; PMA, phorbol 12-myristate 13-acetate; CMK, chloromethylketone; EACA, 
-amino-n-caproic acid; TIMP-1 and TIMP-2, tissue inhibitor of metalloproteinases-1 and -2; Ab, antibody; ß-ME, ß-mercapto-ethanol; HRP, horseradish peroxidase; ECL, enhanced chemiluminescence; RT-PCR, reverse-transcription PCR.
Received 1/23/01. Accepted 6/12/01.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
