Characterization of a Novel, Membrane-bound, 80-kDa Matrix-degrading Protease from Human Breast Cancer Cells
MONOCLONAL ANTIBODY PRODUCTION, ISOLATION, AND LOCALIZATION*

(Received for publication, August 29, 1996, and in revised form, January 28, 1997)

Chen-Yong Lin , Jehng-Kang Wang , Jeff Torri , Li Dou , Qingxiang Amy Sang Dagger and Robert B. Dickson §

From the Lombardi Cancer Center, Georgetown University Medical Center, Washington, D. C. 20007 and Dagger  Department of Chemistry, Florida State University, Tallahassee, Florida 32306

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

A major, apparently novel extracellular matrix-degrading protease was previously identified and partially isolated from hormone-dependent but not from hormone-independent human breast cancer cells (Shi, Y. E., Torri, J., Yieh, L., Wellstein, A., Lippman, M. E., and Dickson, R. B. (1993) Cancer Res. 53, 1409-1415). Although initially the 80-kDa protease was identified from breast cancer cell-conditioned medium, immunofluorescence staining of breast cancer cells with anti-80-kDa protease monoclonal antibody 21-9 showed that in addition to its detection in intracellular compartments, the protease was uniformly localized around periphery of the cells with more intensive staining on the pseudopodia and membrane ruffles. A surface biotinylation technique confirmed the plasma membrane localization of the protease. In addition, the 80-kDa protease could not be washed from the membrane fraction of homogenized breast cancer cells with high concentrations of salts or with EDTA.

The 80-kDa protease may noncovalently associate with other protein(s) to form complexes, the 95- and 110-kDa proteases. Both complexes showed gelatinolytic activity and bore the epitopes recognized by monoclonal antibody 21-9. Furthermore, both complexes could be converted to 80-kDa forms by boiling in SDS in the absence of reducing agents. Expression of this novel, integral membrane gelatinase could allow breast cancer cells an alternative to other previously described matrix-degrading enzymes for degradation of the extracellular matrix in close proximity to their surfaces.


INTRODUCTION

Fatality in breast cancer is the result of metastatic lesions. Degradation of the extracellular matrix (ECM),1 including the basement membrane and interstitial stroma, an important aspect of metastasis, is required for metastatic cancer cells to migrate through anatomical barriers and to invade tissues. Matrix-degrading proteases, including the plasminogen activator-plasmin system of serine proteases, 72-kDa gelatinase A (MMP-2) and 92-kDa gelatinase B (MMP-9) of the matrix metalloprotease (MMP) family, and cathepsins B (a cysteine protease) and D (an aspartic protease), have been implicated in tumor cell invasion (2). Morphological studies show that degradation of the ECM occurs in close proximity to the cell surface (3, 4), and that matrix-degrading proteases must be localized on the cell surface to perform their functions. This theory was further supported by the observation that the cancer cell-conditioned medium alone failed to degrade the ECM (5). Studies of two matrix-degrading proteases, the plasminogen activator-plasmin system and the activation of 72-kDa gelatinase A (MMP-2), clearly demonstrate that localized ECM degradation is achieved by protease binding to a membrane-bound protease receptor and/or activator followed by a proteolytic activation cascade. The urokinase-type plasminogen activator (uPA) is secreted as a soluble, inactive precursor by tumor cells and/or tumor stromal cells and recruited to tumor cell surfaces by binding to the membrane-bound uPA receptor. As a result of association with the membrane-bound uPA receptor, uPA activity, activated plasminogen, and degradation of the ECM are all localized close to the plasma membrane (6-8). MMPs such as MMP-2 are also secreted as soluble, inactive precursors. Although the specific mechanism of activation in vivo of gelatinase A is not fully understood (9-11), this process appears to occur on the cell surface and is believed to be the critical step for cancer cell invasion (12, 13). A subfamily of MMPs called membrane type metalloproteases (MT-MMPs) appears to be involved at an early step(s) in the proteolytic activation cascade (14, 15).

Previously we identified from human breast cancer cells a novel matrix-degrading protease, which is distinct from gelatinases A (72 kDa) and B (92 kDa) in the following characteristics: (a) lack of a gelatin binding site, (b) different electrophoretic mobility on gelatin zymograms, (c) a different inhibition profile, (d) broad metal ion dependence, and (E) insensitivity to activation by p-aminophenylmercuric acetate. In addition, the 80-kDa protease appears to be a major gelatinase, selectively produced by hormone-dependent human breast cancer cells (1). In the current study, an anti-protease monoclonal antibody was prepared and used for isolation of the 80-kDa enzyme by immunoaffinity chromatography; purification was to near homogeneity. Subcellular fractionation, immunofluorescence staining, and surface biotinylation were carried out to show that in addition to its detection in the secreted and intracellular compartments, the 80-kDa protease was localized around periphery of the cells, with more intensive staining on the pseudopodia and membrane ruffles. Our data indicate that the elevated expression of this integral membrane gelatinase at the pseudopod tip may allow breast cancer cells to disrupt the ECM locally and contribute to forward locomotion.


MATERIALS AND METHODS

Cell Lines and Culture Condition

The human hormone-dependent breast cancer cell line T47D was maintained in modified Iscove's minimal essential medium (Biofluids, Rockville, MD) supplemented with 5% fetal calf serum (Life Technologies, Inc). To isolate the 80-kDa protease, the monolayers of T47D cells were washed twice with phosphate-buffered saline (PBS) and were cultured in the absence of serum in Iscove's minimal essential medium supplemented with insulin/transferrin/selenium (Biofluids). The rat myeloma cell line YB2/0 was purchased from American Type Culture Collection and maintained in modified Iscove's minimal essential medium (Biofluids) supplemented with 20% fetal calf serum (Life Technologies). The hybridoma cell line 21-9 was maintained in Dulbecco's modified Eagle's medium (Biofluids) supplemented with 10% Nu-serum IV (Collaborative Biochemical Products, Bedford, MA). To produce mAb without contamination of bovine immunoglobulin, hybridoma cell line 21-9 was adapted into protein-free hybridoma medium II (Life Technologies).

Preparation of 80-kDa Protease as an Immunogen

Partially purified 80-kDa protease was prepared as described previously (1) with some modification. Briefly, T47D cell-conditioned serum-free media were collected, and cellular debris was removed by centrifugation. Ammonium sulfate powder was added to the supernatant with continuous mixing to 65% saturation and sat in a cold room overnight. The protein precipitates were obtained by centrifugation at 5,000 × g for 20 min and then dissolved in PBS. The 80-kDa protease was first separated from the major contaminating protein transferrin using hydroxylapetide chromatography (Bio-Rad). To do so, the 80-kDa protease was first dialyzed against 10 mM phosphate (pH 6.8) and then applied to a hydroxylapetide column equilibrated with 10 mM phosphate (pH 6.8). The column was washed with the same solution. The transferrin was eluted by 40 mM phosphate (pH 6.8), and then the 80-kDa protease was eluted by 150 mM phosphate (pH 6.8). The protease was further subjected to gel filtration through a Sephacryl S-200 HR column (HiPrep 16/60, Pharmacia Biotech Inc.) equilibrated with PBS. The column was eluted with the same buffer, and the positive fractions were collected. The 80-kDa protease obtained by gel filtration was then precipitated by ammonium sulfate as described above, dialyzed against 20 mM Tris-HCl (pH 8.0), and applied onto a DEAE-Sepharose FF column (Pharmacia), which was equilibrated with 20 mM Tris-HCl (pH 8.0). The column was washed with equilibration buffer. The 80-kDa protease was eluted with a linear gradient of 0-0.4 M NaCl in DEAE equilibration buffer. The enzymatic activity of 80-kDa protease was assessed by gelatin zymography.

Monoclonal Antibody Production

Eight-week-old female Harlan Sprague-Dawley rats were immunized at 2-week intervals with 20 µg of proteins including the 80-kDa protease obtained from DEAE fractions. The immune response against the 80-kDa protease was tested by precipitation of the 80-kDa protease using the antisera obtained from rats. The spleenocytes from an immunopositive rat were used to generated hybridoma according to a previously described method (16). Immunoblot analysis was applied in the primary screening (see below). The positive hybridomas were subjected to secondary screening by their ability to precipitate the gelatinase activity of the 80-kDa protease (see below).

Hybridoma Screening by Western Blot Analysis

To perform immunoblot screening, the 80-kDa protease was first resolved by a 7.5% SDS-polyacrylamide gel (minigel, 8 × 9 cm) and then transferred to nitrocellulose membranes. After blocking with 6% milk in PBS, each nitrocellulose membrane was cut into about 50 strips (1.6 mm in width). The cell-conditioned media (100 µl) from successful hybridomas were overlaid on nitrocellulose strips at room temperature for 2 h. The nitrocellulose strips were washed with 1% Triton X-100 in PBS. The immunoreactive polypeptides were visualized using a peroxidase-labeled anti-rat IgG antibody and the ECL detection system (Amersham Corp.).

Immunoblotting

Proteins were separated by 10% SDS-polyacrylamide gel electrophoresis, transferred overnight to nitrocellulose sheets (Schleicher & Schuell) by diffusion, and subsequently probed with mAbs. Blots were incubated with mAb 21-9 diluted in PBS containing 1% bovine serum albumin for 1 h at room temperature, and immunoreactive polypeptides were visualized using the ECL detection system.

Gelatin Zymography

Gelatin zymography was conducted as described previously with some modifications (1). Protease substrates used were gelatin (1 mg/ml). The electrophoresis was performed at a constant current of 15 mA. The gelatin gel was washed three times with PBS containing 2% Triton X-100 and incubated in PBS without any metal ions at 37 °C overnight.

Immobilization of Monoclonal Antibody on Sepharose 4B

Monoclonal antibody was isolated from protein-free cell-conditioned medium using DEAE chromatography. Immunoaffinity beads were made by coupling 5 mg of mAb/ml of CNBr-activated Sepharose 4B as specified in the manufacturer's instructions (Pharmacia). The beads were washed with 0.1 M glycine-HCl (pH 2.4) before use.

Immunoprecipitation

Twenty-five µl of packed mAb 21-9-Sepharose beads were incubated with 400 µl of the 80-kDa protease (Fig. 1, A, lane 1, and B, lane 1) for 2 h at 4 °C. Followed by centrifugation at low speed, the resulting supernatants were collected (Fig. 1, A, lane 2, and B, lane 2), and the beads were washed four times with 1% Triton X-100 in PBS. The bound proteins were eluted with 30 µl of 0.1 M glycine-HCl (pH 2.4) and immediately neutralized by 2 M Trizma base. Five continuous elutions were performed.


Fig. 1. Verification of mAb 21-9 directed against the 80-kDa protease. To verify that mAb 21-9 was directed against the 80-kDa protease, immobilized mAb 21-9 was used to precipitate the gelatinolytic activity of the 80-kDa protease assessed by a gelatin zymogram (A) and the immunoreactive polypeptides examined by immunoblotting using mAb 21-9 (B). The gelatinolytic activity and the corresponding polypeptides were both depleted by mAb 21-9 (comparing lanes 2 and 1). Moreover, the 80-kDa protease could be eluted by 0.1 M glycine buffer (pH 2.4) (lanes 3-7 represent fractions 1-5 in both panels). There was a possible degradation product with an apparent molecular mass of 40 kDa that was also recognized by mAb 21-9.
[View Larger Version of this Image (54K GIF file)]


Immunoaffinity Chromatography

The immunoaffinity matrix was equilibrated with PBS containing 1% Triton X-100. The 80-kDa protease was loaded onto a 1-ml column at a flow rate of 7 ml/h. The column was washed with 1% Triton X-100 in PBS. Bound protease was then eluted using 0.1 M glycine-HCl (pH 2.4). Fractions were immediately neutralized using 2 M Trizma base.

Subcellular Fractionation

T47D cells were washed with PBS three times and swelled in 20 mM Tris-HCl (pH 7.4). The cells were scraped and homogenized with a Dounce homogenizer. The homogenates were then centrifuged at 600 × g for 10 min. The pellets were referred as a nuclear fraction. The resultant supernatant was centrifuged at 20,000 × g for 20 min. The pellets were referred to as a membrane fraction and the supernatant as a cytosol fraction. Although additional membrane pellets could be obtained from the cytosol fraction by 100,000 × g centrifugation, and these membrane pellets also contained the 80-kDa protease, the yields of the membrane and the 80-kDa protease were too low to allow characterization.

Pulse-Chase

T47D cells were first starved from methionine for 1 h and pulse labeled with [35S]methionine for 30 min or 1 h as described (17, 18). The chases were intervals of 0, 1, 2, 3, and 4 h. Cell-free extracts were prepared in radioimmunoprecipitation assay buffer (150 mM NaCl, 1.0% Nonidet P-40, 0.5% deoxycholic acid, 0.1% SDS, and 50 mM Tris-HCl, pH 8.0) containing protease inhibitors (74 µM antipain, 0.3 µM aprotinin, 28 µM N-[N-(L-3-trans-carboxirane-2-carbonyl)-L-leucyl]agmatine, 1 µM leupeptin, 4 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride, 1.3 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride) and immunoprecipitated with mAb 21-9-Sepharose. The absorbed proteases were eluted either by boiling in SDS-sample buffer or 0.1 M glycine buffer (pH 2.4).

Immunofluorescent Staining

T47D cells were prefixed with 3.7% formaldehyde for 10 min. Cells with or without permeabilization were stained by an indirect immunofluorescent technique using mAb 21-9 at room temperature for 20 min, followed by fluorescein-conjugated secondary antibody (Sigma) for 20 min at room temperature. Control cells were stained in the absence of primary antibody.

Surface Biotinylation

Sulfosuccinimidobiotin (sulfo-NHS-biotin, Pierce) or sulfosuccinimidyl 6-(biotinamido) hexanoate (NHS-LC-biotin, Pierce) was used to label the cell surfaces as described (19). To label cell surfaces with biotinylation reagents, confluent T47D cells grown in 15-cm dishes were washed with PBS three times, and then sulfo-NHS-biotin or NHS-LC-biotin (2 mg in 10 ml of PBS/dish) was added. Cells were incubated at room temperature for 20 min. After incubation, cells were washed with 50 mM Tris-HCl (pH 7.4) and 150 mM NaCl three times. Biotinylated cells were then subjected to subcellular fractionation, and the membrane fractions were collected.

Amino Acid Composition Analysis

Immunoaffinity-purified 80-kDa protease was subjected to SDS-PAGE (7.5% acrylamide) and then transferred onto a polyvinylidene difluoride membrane (Bio-Rad). The 80-kDa protease was visualized by Coomassie Blue staining and excised for amino acid composition analysis (20, 21). To do so, bands were cut from the blot, and a section, equal to the sample band size, was cut from the edge of the blot to use as a control. The cut bands were placed in a 6 × 50-mm Pyrex tube. One hundred microliters of 6 N HCl was added to each tube, and then the samples were vapor hydrolyzed for 18 min using a CEM microwave digestion system. The bands were washed three times with 0.1 N HCl, 20% methanol. The rinses were combined, and the samples were dried on a Heto Speedvac. The samples were precolumn-derivitized with phenylisothiocyanate and run on reverse phase high performance liquid chromatography. The high performance liquid chromatography system comprised Waters M6000 pumps, a 680 gradient controller, and a 440 UV detector. The results were analyzed on a Waters Maxima software system. The separation was performed with a Waters free amino acid column (3.9 × 300 mm). The column temperature was 38 °C.


RESULTS

Monoclonal Antibody Production

To characterize the 80-kDa protease further, which we had previously identified in human breast cancer cells, we generated a mAb, which was directed against this protease. Since partially purified 80-kDa protease was used as an immunogen, we screened a hybridoma library by Western blot analysis to select mAbs that recognized polypeptides of approximately 80 kDa. The selected hybridoma lines were subcloned and further tested for their ability to precipitate the 80-kDa protease. The mAbs screened by both Western blot analysis and immunoprecipitation are convenient tools for both biochemical characterization and immunoaffinity purification.

An immunopositive rat whose serum contained antibodies against the 80-kDa protease was sacrificed, and its spleen cells were fused with rat YB2/0 myeloma cells (16). Successful hybridomas were screened by immunoblot analysis. One hybridoma line (mAb 21-9) was selected, subcloned, and further characterized. To verify that mAb 21-9 (IgG1) was directed against the 80-kDa protease, immunoprecipitation was performed. The gelatinase activity of the 80-kDa protease, as well as its corresponding polypeptide, was completely depleted by mAb 21-9 (Fig. 1, A, lane 2, and B, lane 2); the absorbed protease could be eluted from beads by 0.1 M glycine-HCl (pH 2.4) (Fig. 1, A and B, lanes 3-7). The 80-kDa protease remained enzymatically active after neutralizing the low pH eluate (Fig. 1A, lanes 3-7) with 2 M Trizma base. An apparently degraded form at 40 kDa was also precipitated by mAb 21-9, and its corresponding polypeptide was also recognized by mAb 21-9 by immunoblotting analysis (Fig. 1), suggesting that the active site of the 80-kDa protease and the epitope recognized by mAb 21-9 were both on the 40-kDa fragment.

One Step Purification of the 80-kDa Protease Using Immunoaffinity Chromatography

The results obtained from immunoprecipitation suggested that the 21-9 mAb was competent for immunoaffinity purification of the 80-kDa protease. The 80-kDa protease was concentrated from cell-conditioned medium by addition of ammonium sulfate to 65% saturation, and then the precipitated proteins were dialyzed against PBS. The immunoaffinity chromatography was carried out by applying concentrated medium onto a mAb 21-9-Sepharose column (1 ml). The protease was eluted by 0.1 M glycine-HCl (pH 2.4), and the eluate was subjected to three assays, including SDS-PAGE to examine the purity, gelatin zymography to assess its proteolytic activity, and immunoblot analysis to check its immunological reactivity (Fig. 2). The 80-kDa protease could easily be isolated to near homogeneity by immunoaffinity purification from the cell-conditioned medium (Fig. 2A). Approximately 2 µg of the 80-kDa protease was obtained from 1 liter of cell-conditioned medium, which contained approximately 40 mg of protein. It represented at least a 10,000-fold increase in the purity by this one step purification. The extent of recovery of protease was estimated at greater than 50% based on zymography. In addition, several minor polypeptides were co-purified, including 110- and 95-kDa polypeptides. Although the 80-kDa protease could be eluted from an immunoaffinity column by glycine-HCl (pH 2.4), its proteolytic activity could also be recovered by neutralization as determined by gelatin zymography (Fig. 2B). The reduced form of the 80-kDa protease appeared to be a single polypeptide chain with a slightly higher molecular mass than the nonreduced form but became inactive (data not shown). In addition to the 80-kDa protease, several weak gelatinolytic activities were seen on gelatin zymograms. The 110- and 95-kDa activities corresponded in size to the two polypeptides seen on SDS gels. The immunological relationships between the 80-kDa protease and these two higher molecular mass polypeptides were verified by Western blotting analysis using mAb 21-9 (Fig. 2C). Both 110- and 95-kDa species bore the epitope that was recognized by anti-80-kDa protease, suggesting that they were directly captured by mAb 21-9 during immunoaffinity column chromatography. To explore whether the 95- and 110-kDa proteases were the complexes of the 80-kDa protease, we performed nonboiling and boiling two-dimensional gels. Both 95- and 110-kDa proteases were resolved on a (first dimension) SDS-polyacrylamide gel at their expected sizes, but they were decreased to the 80-kDa size on another (second dimension) SDS gel after boiling of a sliced gel strip (Fig. 3) under nonreduced conditions. Furthermore, using a radioactive amino acid pulse-chase protocol, the 80-kDa protease appeared to be initially synthesized at approximately its final 80-kDa size but not from either of the larger size species (Fig. 4). Taking these data together, it appears likely that both 95- and 110-kDa species represent complexes of the 80-kDa protease with other unidentified protein(s).


Fig. 2. One step purification of the 80-kDa protease using an immunoaffinity column. The 80-kDa protease was isolated from the conditioned medium of T47D cells using a mAb 21-9 immunoaffinity column. The glycine-HCl-eluted fractions (fractions 1-3) were analyzed by SDS-PAGE (A), gelatin zymography (B), and Western blot analysis using mAb 21-9 (C). The SDS gels were visualized by silver staining. In addition to the 80-kDa protease, two minor 110- and 95-kDa polypeptides were co-purified (A), exhibited gelatinase activity (B), and were recognized by mAb 21-9 (C). Several minor low molecular mass polypeptides with or without gelatinolytic activity were also seen.
[View Larger Version of this Image (61K GIF file)]



Fig. 3. The 95- and 110-kDa proteases could be the complexes of the 80-kDa protease with unidentified protein(s). To detect both 95- and 110-kDa proteases clearly by immunoblot analysis using mAb 21-9, serum-free T47D cell-conditioned medium was concentrated to 100-fold and resolved by SDS-PAGE without boiling treatment (1st-D non boil). A gel strip containing these three proteases was sliced out and boiled for 5 min in 1 × SDS sample buffer in the absence of reducing agents and then electrophoresized on a second SDS-polyacrylamide gel (2nd-D boil). A duplicated 1st-D gel strip and a 2nd-D gel were transferred to nitrocellulose membranes, and immunoblot analysis was carried out using mAb 21-9.
[View Larger Version of this Image (61K GIF file)]



Fig. 4. Biosynthesis of the 80-kDa protease. A pulse-chase experiment was performed to study biosynthesis of the 80-kDa protease and whether both 95- and 110-kDa proteases were the precursors of the 80-kDa protease. After a 60-min pulse with [35S]methionine, cell-free extracts were collected at 0, 30, 60, 120, and 240 min (lanes 1-5) in radioimmunoprecipitation assay buffer containing protease inhibitors. The proteases were immunoprecipitated with mAb 21-9-Sepharose. The absorbed proteases were eluted by boiling in SDS sample buffer. The proteases were resolved by SDS-PAGE, and protein bands were viewed by fluorography.
[View Larger Version of this Image (64K GIF file)]


As noted in the Introduction, we previously observed the differences between the 80-kDa protease and previously described metalloproteases in biochemical characteristics such as gelatin binding site, electrophoretic mobility, and inhibitor profile. In addition, the amino acid composition of the 80-kDa protease was compared with the ones of gelatinase A (72 kDa), gelatinase B (92 kDa), and four members of MT-MMPs (14, 15, 22, 24). Results demonstrated that there was no similarity between the 80-kDa protease and these members of the MMP family in their amino acid composition (data not shown).

Plasma Membrane Localization of the 80-kDa Protease

Although the 80-kDa protease was initially identified and purified from the cell-conditioned medium, the protease was associated with the membrane fraction when the T47D cells were subjected to subcellular fractionation. For subcellular fractionation, the cells were swollen in hypotonic buffer, and the nuclei were separated by low speed centrifugation. The resultant supernatant was further fractionated into membrane and cytosol fractions by high speed centrifugation. The association of the 80-kDa protease with membrane fractions was sufficiently strong that it resisted salt washes with 1 M NaCl and 2 M KCl as well as 0.1 M EDTA. There was no 80-kDa protease detected from each salt wash (data not shown). However, it could be extracted by detergents from the washed membrane fractions (Fig. 5). Although less 80-kDa protease was detected from the EDTA-washed membrane fraction compared with the control sample, the decreased signal could result from the removal of divalent ions from the 80-kDa protease by a high concentration of EDTA, which affected the epitope recognized by mAb 21-9 to some extent. These results suggested that the association of the 80-kDa protease with the membrane was extremely strong and that the protease did not maintain this interaction through a loose association with other membrane-bound proteins or attach via a divalent cation. We propose that the 80-kDa protease could be an integral membrane protein. In addition, there was no prominent difference in the size of the 80-kDa protease comparing the membrane-bound and the secreted forms. The 95-, 110-, and 40-kDa putative degraded products were indistinctly seen in the membrane fraction (Fig. 5) The ratios of the 80-kDa form to the 95- and 110-kDa forms in the membrane fraction appeared to be similar to the ratios in the cell-conditioned medium.


Fig. 5. Extraction of membranes for the 80-kDa proteases. The membrane fraction of homogenated T47D cells was washed once by 20 mM Tris-HCl (lane 1), 1 M NaCl (lane 2), 2 M KCl (lane 3), and 0.1 M EDTA (lane 4), respectively, before extraction with Triton X-100. These four membrane extracts were subjected to Western blot analysis using 21-9 mAb. The 80-kDa protease was still associated with the membrane after salt washes.
[View Larger Version of this Image (35K GIF file)]


To examine more precisely the subcellular localization of the 80-kDa protease, T47D cells were grown on glass coverslips, fixed with or without permeabilization by detergent, and labeled with anti-80-kDa protease mAb 21-9. MAb 21-9 stained the surfaces of both nonpermeabilized and permeabilized cells most intensively at the membrane ruffles (Fig. 6, A, B, E, and I). The 80-kDa protease was also localized in intracellular compartments after cells were permeabilized with 0.5% Triton X-100 (Fig. 6I). The fact that the 80-kDa protease is specifically expressed on the cell surface was further verified with a surface biotinylation technique (Fig. 7B), in which the surface proteins of T47D cells were modified by sulfo-NHS-biotin or NHS-LC-biotin. These two protein modification reagents are water-soluble and not permeant to the plasma membrane; thus, biotinylation is restricted to the cell surface. After surface biotinylation, the membrane fractions were collected, and the Triton X-100 extraction was carried out using the membrane fractions. The total cellular 80-kDa protease pool (Fig. 7A, lane 1), including the biotinylated surface portion, was then specifically precipitated using mAb 21-9-Sepharose beads (Fig. 7A, lane 2), and the total biotinylated cell surface proteins (Fig. 7B, lane 1) and the surface biotinylated 80-kDa protease (Fig. 7B, lane 2) were detected by peroxidase-labeled streptavidin. These results strongly suggest that a portion of the 80-kDa protease pool is expressed on the cell surface.


Fig. 6. Immunofluorescent distribution of the 80-kDa protease in T47D cells. A, B, and E, mAb 21-9 fluorescent images of T47D cells at the plasma membrane and the leading edges of an individual cell (A), two cells (B), and a cell cluster (E). C, D, and F, corresponding phase contrast images. The cells were fixed without permeabilization by detergent. The 80-kDa protease was homogeneously expressed on the surface of T47D cells under different cellular conditions. More extensive localization was seen at membrane ruffles of spreading cells. G, control without primary antibody; H, corresponding phase contrast image. The cells were fixed without permeabilization. I, anti-80-kDa fluorescent image; J, corresponding phase contrast image. The cells were fixed and permeabilized with Triton X-100. The 80-kDa protease had a diffuse distribution; the most intense localization was at membrane ruffles. An increased level of immunoreactivity was seen in intracellular compartments. L, control without primary antibody; M, corresponding phase contrast. The cells were fixed and permeabilized with Triton X-100.
[View Larger Version of this Image (122K GIF file)]



Fig. 7. Surface biotinylation of the 80-kDa protease. Surface proteins of the T47D cells were biotinylated by sulfo-NHS-biotin, or NHS-LC-biotin. The membrane fractions were collected, and a Triton X-100 extract was prepared. The 80-kDa protease was specifically precipitated by mAb 21-9-Sepharose beads. The membrane extract (lane 1 in both panels), the immunoprecipitated 80-kDa protease (lane 2 in both panels), and the unbound fraction (lane 3 in both panels) were resolved in 10% SDS gels and transferred to nitrocellulose membranes. A nitrocellulose blot was probed by mAb 21-9 and then a horseradish peroxidase-labeled secondary antibody (A). A parallel blot was probed with an avidin-biotin complex kit to detect biotin (B).
[View Larger Version of this Image (56K GIF file)]



DISCUSSION

Isolation of the 80-kDa protease was achieved from cell-conditioned medium of T47D cells by combining concentration of the medium with ammonium sulfate and affinity chromatography using immobilized anti-80-kDa protease mAb 21-9. The purified protease consisted of three immunologically related forms, including the major activity, which was a doublet around 80 kDa, and two minor bands of activity at 95 and 110 kDa. In addition, a putative degradation product of 40 kDa was sometimes observed. These isolated proteolytic components corresponded to the polypeptides recognized by the mAb 21-9, as shown by Western blotting. MAb 21-9 did not bind to the chemically reduced protease, indicating that exposure or configuration of the epitope was dependent on an intramolecular disulfide bond(s). By analogy with the epitope recognized by mAb 21-9, the proteolytic activities of these enzymes were sensitive to reducing agents such as 2-mercaptoethanol and dithiothreitol. These proteases are, however, stable under acidic pH conditions down to 2.4; their activities after acidic pH treatment were recovered by neutralization. Several minor polypeptides were also co-purified from the immunoaffinity column. Of these polypeptides, some could potentially have gelatinolytic activity but were not recognized by mAb 21-9. These activities could be other proteases or the degraded fragments of the 80-kDa protease, which still bear the active site but have lost the epitope recognized by mAb 21-9. It is possible that the other polypeptides were co-purified from the immunoaffinity column, because they formed a complex with the 80-kDa protease through binding to a common binding protein(s) such as a protease inhibitor(s). The possible existence of complexes of the 80-kDa protease with other protein(s) was supported by the detection of the 95- and 110-kDa proteases. All three of these species exhibited gelatinolytic activity and were recognized by anti-80-kDa protease mAb 21-9. A precursor-mature form relationship among these three proteases could not be established using pulse-chase experiments. In addition, the 95- and 110-kDa proteases could be reduced to the 80-kDa size by boiling. These data were consistent with the higher molecular mass forms consisting of 80-kDa protease complexes with other protein(s).

A common feature of invasive cancer cells capable of penetrating anatomic barriers is their ability to degrade the ECM components, basement membranes, and interstitial stroma. ECMs are composed of collagens, glycoproteins (such as fibronectin, laminin, and entactin/nidogen), proteoglycans, and glycosaminoglycans. To degrade the ECM efficiently, many hydrolytic enzymes are believed to work in concert. Removal of one component of the ECM by a specific enzyme makes the other ECM components more available for degradation by other specific enzymes. It has been demonstrated that depletion of glycoproteins from the ECM by trypsin treatment is necessary for the maximal digestion of elastin and collagen (3). Another type of interaction between matrix-degrading enzymes is when one enzyme serves as an activator for others and may trigger a digestion cascade. For example, plasmin appears to activate several members of the MMP family, including interstitial collagenase and stromelysin 1 (25, 26). In the present work, we have isolated a novel matrix-degrading protease from the human breast cancer cell line T47D. The gelatinolytic activity of this novel protease could contribute to the degradation of the endothelial basement membrane by T47D cells observed by others (5). However, compared with other breast cancer cells, including MDA-MB-231, ZR-75-1, and MCF-7, the lack of proteoglycan-degrading activity for T-47D cells may result in less accessibility of the 80-kDa protease to its substrate(s) in vivo and thus a less invasive phenotype for this breast cancer cell type (5, 27).

Many studies have shown that degradation of the ECM occurs at the contact sites of invasive cells with substrate (3, 28). In addition to its presence in cell-conditioned medium, matrix-degrading activity has been described in cell cytosol fractions (29) and in the plasma membrane fractions of invasive cells (30). Secreted matrix-degrading proteases, such as the 72-kDa gelatinase A (MMP-2), have been reportedly associated with the cell surface through interaction with receptor-like molecule(s) (31, 32) and activated on the cell surface (33-35). More recently integrin-alpha vbeta 3 was suggested as a receptor for gelatinase A (36). Here we report that a novel, matrix-degrading protease, like other previously described matrix-degrading proteases, is associated with membrane fractions of cancer cells. The fact that the 80-kDa protease was extractable from the membrane fractions with detergents such as 1% Triton X-100 or radioimmunoprecipitation assay buffer but not with high concentrations of washes of salts (1 M NaCl or 2 M KCl) or 0.1 M EDTA is evidence that the 80-kDa protease is tightly associated with the membrane. Furthermore, the immunofluorescent staining using mAb 21-9 demonstrated that the 80-kDa protease is concentrated on the cell periphery at pseudopodia and membrane ruffles while cells are spreading. The membrane ruffle localization of the 80-kDa protease strongly substantiates its proposed function in invasiveness. This proposal has been based on morphological studies in which a specialized ruffled cellular edge that consists of a highly motile collection of membrane pseudopodia projects to dissolve and invade the ECM (37-40). Although the mechanisms by which the 80-kDa protease is associated with the plasma membrane or released into the medium were not explored in our study, release of membrane-bound protease into cell-conditioned medium, such as MT-MMP (23), has been previously described. Since MT-MMP was reported to exhibit gelatinolytic activity and plasma membrane localization (23), as does the 80-kDa protease, we compared the amino acid compositions of the 80-kDa protease with four MT-MMPs (14, 22-24), and no similarity was found (data not shown).

In summary, in this study, we isolated a novel matrix-degrading protease from human breast cancer cells and generated a mAb directed against this protease. This matrix-degrading protease was distributed on the surfaces of these cells, with more protease detected at the membrane ruffles. This observation is consistent with a role for this novel matrix-degrading protease in ECM degradation and breast cancer invasion.


FOOTNOTES

*   Supported by National Institutes of Health Specialized Program of Research Excellence Grant 1P50CA58185 in breast cancer.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§   To whom correspondence should be addressed: Lombardi Cancer Center, Georgetown University Medical Center, 3970 Reservoir Rd. N. W., Washington, D. C. 20007. Tel.: 202-687-4304; Fax: 202-687-7505.
1   The abbreviations used are: ECM, extracellular matrix; MMP, matrix metalloprotease; uPA, urokinase-type plasminogen activator; MT-MMP, membrane type metalloprotease; PBS, phosphate-buffered saline; mAb, monoclonal antibody; sulfo-NHS-biotin, sulfosuccinimidobiotin; NHS-LC-biotin, sulfosuccinimidyl 6-(biotinamido) hexanoate; PAGE, polyacrylamide gel electrophoresis.

REFERENCES

  1. Shi, Y. E., Torri, J., Yieh, L., Wellstein, A., Lippman, M. E., and Dickson, R. B. (1993) Cancer Res. 53, 1409-1415 [Abstract]
  2. Mignatti, P., and Rifkin, D. B. (1993) Physiol. Rev. 73, 161-195 [Free Full Text]
  3. Jones, P. A., and De Clerck, Y. A. (1980) Cancer Res. 40, 3222-3227 [Abstract]
  4. Kramer, R. H., Bensch, K. G., and Wong, J. (1986) Cancer Res. 46, 1980-1989 [Abstract]
  5. Yee, C., and Shiu, R. P. (1986) Cancer Res. 46, 1835-1839 [Abstract]
  6. Behrendt, N., Ronne, E., and Dano, K. (1995) Biol. Chem. Hoppe-Seyler 376, 269-279 [Medline] [Order article via Infotrieve]
  7. Duggan, C., Maguire, T., McDermott, E., O'Higgins, N., Fennelly, J. J., and Duffy, M. J. (1995) Int. J. Cancer 61, 597-600 [Medline] [Order article via Infotrieve]
  8. Conese, M., and Blasi, F. (1995) Bailliere's. Clin. Haematol. 8, 365-389 [Medline] [Order article via Infotrieve]
  9. Bergmann, U., Tuuttila, A., Stetler-Stevenson, W. G., and Tryggvason, K. (1995) Biochemistry 34, 2819-2825 [Medline] [Order article via Infotrieve]
  10. Atkinson, S. J., Crabbe, T., Cowell, S., Ward, R. V., Butler, M. J., Sato, H., Seiki, M., Reynolds, J. J., and Murphy, G. (1995) J. Biol. Chem. 270, 30479-30485 [Abstract/Free Full Text]
  11. Strongin, A. Y., Collier, I., Bannikov, G., Marmer, B. L., Grant, G. A., and Goldberg, G. I. (1995) J. Biol. Chem. 270, 5331-5338 [Abstract/Free Full Text]
  12. Kleiner, D. E., Jr., and Stetler-Stevenson, W. G. (1993) Curr. Opin. Cell Biol. 5, 891-897 [Medline] [Order article via Infotrieve]
  13. Stetler-Stevenson, W. G. (1994) Invasion & Metastasis 14, 259-268 [Medline] [Order article via Infotrieve]
  14. Sato, H., Takino, T., Okada, Y., Cao, J., Shinagawa, A., Yamamoto, E., and Seiki, M. (1994) Nature 370, 61-65 [CrossRef][Medline] [Order article via Infotrieve]
  15. Takino, T., Sato, H., Shinagawa, A., and Seiki, M. (1995) J. Biol. Chem. 270, 23013-23020 [Abstract/Free Full Text]
  16. Kilmartin, J. V., Wright, B., and Milstein, C. (1982) J. Cell Biol. 93, 576-582 [Abstract]
  17. Munemitsu, S. M., and Samuel, C. E. (1988) Virology 163, 643-646 [Medline] [Order article via Infotrieve]
  18. Thomis, D. C., and Samuel, C. E. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 10837-10841 [Abstract]
  19. Goodloe-Holland, C. M., and Luna, E. J. (1987) Methods Cell Biol. 28, 103-128 [Medline] [Order article via Infotrieve]
  20. Heinrikson, R. L., and Meredith, S. C. (1984) Anal. Biochem. 136, 65-74 [Medline] [Order article via Infotrieve]
  21. Tous, G. I., Fausnaugh, J. L., Akinyosoye, O., Lackland, H., Winter-Cash, P., Vitorica, F. J., and Stein, S. (1989) Anal. Biochem. 179, 50-55 [Medline] [Order article via Infotrieve]
  22. Will, H., and Hinzmann, B. (1995) Eur. J. Biochem. 231, 602-608 [Abstract]
  23. Imai, K., Ohuchi, E., Aoki, T., Nomura, H., Fujii, Y., Sato, H., Seiki, M., and Okada, Y. (1996) Cancer Res. 56, 2707-2710 [Abstract]
  24. Puente, X. S., Pendas, A. M., Llano, E., Velikova, G., and Lopez-Otin, C. (1996) Cancer Res. 56, 944-949 [Abstract]
  25. Goldberg, G. I., Marmer, B. L., Grant, G. A., Eisen, A. Z., Wilhelm, S., and He, C. S. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 8207-8211 [Abstract]
  26. Suzuki, K., Nagase, H., Ito, A., Enghild, J. J., and Salvesen, G. (1990) Biol. Chem. Hoppe-Seyler 371, (suppl.) 305-310
  27. Thompson, E. W., Paik, S., Brunner, N., Sommers, C. L., Zugmaier, G., Clarke, R., Shima, T. B., Torri, J., Donahue, S., Lippman, M. E., and Dickson, R. B. (1992) J. Cell. Physiol. 150, 534-544 [Medline] [Order article via Infotrieve]
  28. Chen, W. T., Olden, K., Bernard, B. A., and Chu, F. F. (1984) J. Cell Biol. 98, 1546-1555 [Abstract]
  29. Zucker, S., Turpeenniemi-Hujanen, T., Ramamurthy, N., Wieman, J., Lysik, R., Gorevic, P., Liotta, L. A., Simon, S. R., and Golub, L. M. (1987) Biochem. J. 245, 429-437 [Medline] [Order article via Infotrieve]
  30. Zucker, S., Wieman, J. M., Lysik, R. M., Wilkie, D. P., Ramamurthy, N., and Lane, B. (1987) Biochim. Biophys. Acta 924, 225-237 [Medline] [Order article via Infotrieve]
  31. Emonard, H. P., Remacle, A. G., Noel, A. C., Grimaud, J. A., Stetler-Stevenson, W. G., and Foidart, J. M. (1992) Cancer Res. 52, 5845-5848 [Abstract]
  32. Monsky, W. L., Kelly, T., Lin, C. Y., Yeh, Y., Stetler-Stevenson, W. G., Mueller, S. C., and Chen, W. T. (1993) Cancer Res. 53, 3159-3164 [Abstract]
  33. Ward, R. V., Atkinson, S. J., Slocombe, P. M., Docherty, A. J., Reynolds, J. J., and Murphy, G. (1991) Biochim. Biophys. Acta 1079, 242-246 [Medline] [Order article via Infotrieve]
  34. Strongin, A. Y., Marmer, B. L., Grant, G. A., and Goldberg, G. I. (1993) J. Biol. Chem. 268, 14033-14039 [Abstract/Free Full Text]
  35. Brown, P. D., Levy, A. T., Margulies, I. M., Liotta, L. A., and Stetler-Stevenson, W. G. (1990) Cancer Res. 50, 6184-6191 [Abstract]
  36. Brooks, P. C., Strömblad, S., Sanders, L. C., Von Schalscha, T. L., Aimes, R. T., Stetler-Stevenson, W. G., Quigley, J. P., and Cheresh, D. A. (1996) Cell 85, 683-693 [Medline] [Order article via Infotrieve]
  37. Ausprunk, D. H., and Folkman, J. (1977) Microvasc. Res. 14, 53-65 [Medline] [Order article via Infotrieve]
  38. Kramer, R. H., Gonzalez, R., and Nicolson, G. L. (1980) Int. J. Cancer 26, 639-645 [Medline] [Order article via Infotrieve]
  39. Kramer, R. H., and Vogel, K. G. (1984) J. Natl. Cancer Inst. 72, 889-899 [Medline] [Order article via Infotrieve]
  40. Mueller, S. C., and Chen, W. T. (1991) J. Cell Sci. 99, 213-225 [Abstract]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.