Identification of a Novel Function for 67-kDa Laminin Receptor: Increase in Laminin Degradation Rate and Release of Motility Fragments¹

The 67LR⁴ is overexpressed on the surface of a variety of tumor cells. The strong correlation between increased 67LR expression and the metastatic potential of tumor cells suggests that the receptor plays a role in development of the metastatic phenotype (1, 2). The physiological role of 67LR involves its activity as an accessory molecule for α₆β₄ integrin. The two receptors are coexpressed, coregulated, and physically associated on the cell surface, suggesting their mutual involvement in laminin binding (3). The interaction of 67LR with laminin induces conformational changes in the structure of the adhesion molecule, increasing the affinity of laminin for the cancer cell surface (4). Nevertheless, this activity cannot account for the involvement of 67LR in invasion because its adhesive function would be expected to stabilize the tumor cell at the primary site.

A key step in the metastatic process is the proteolytic degradation of basement membrane ECM components such as proteoglycans, collagen type IV, laminin-1, and laminin-5 through the action of specific proteases secreted by tumor and stromal cells. This proteolytic cleavage not only removes physical barriers to cell migration but also converts ECM components into substrates suitable for migration, presumably by exposure of motility-promoting cryptic sites (5). Although the precise mechanisms by which proteases alter ECM components remain unclear, it is well known that the invasive behavior of metastatic tumor cells correlates directly with the expression of many enzymes that have matrix hydrolytic activity, including cysteine proteinases [e.g., cathepsin B (6, 7)], aspartic proteinases [e.g., cathepsin D (6, 8)], serine proteinases [e.g., elastase (9)], and metalloproteases (5, 10). The conformation dependence of protease recognition domains on laminin-1 (11) raised the possibility that conformational modification of laminin by 67LR binding alters the proteolytic cleavage of this adhesion molecule such that basement membrane degradation is enhanced. In the present study, we analyzed the effect of laminin-1 modification by 67LR binding on the proteolytic cleavage of laminin-1 by specific proteases known to be involved in tumor progression. Our data point to a mechanism by which an adhesion receptor such as the 67LR plays a major role in tumor aggressiveness.

Human breast carcinoma cell line MDAMB231 and murine melanoma cell line B16F10 [from American Type Culture Collection (Manassas, VA) and Dr. I. J. Fidler (The University of Texas M. D. Anderson Cancer Center, Houston, TX), respectively] were maintained in RPMI 1640 (Sigma Chemical Co., St. Louis, MO) supplemented with 10% FCS, 1% l-glutamine, and antibiotics (100 μg/ml) at 37°C in a humidified 5% CO₂ atmosphere.

Peptide G (IPCNNKGAHSVGLMWWMLAR), corresponding to amino acids 161–180 of the 37-kDa precursor protein of 67LR; scrambled peptide X (PMLRWGCHIAMVNKLSWGNA); peptide YIGSR, corresponding to amino acids 929–933 of the laminin-1 β1 chain; and related peptide YIGSK were synthesized by Neosystem (Strasbourg, France). High-pressure liquid chromatography analysis indicated 95% purity. Peptides were dissolved in distilled water, and concentrations were evaluated spectroscopically. Murine laminin-1 purified from the mouse Engelbreth-Holm-Swarm tumor, fibronectin purified from human plasma, bovine spleen cathepsin B (all from Sigma Chemical Co.), bovine spleen cathepsin D, and human neutrophil cathepsin G (Calbiochem, La Jolla, CA) were used. Polyclonal antibodies directed against laminin-1 or fibronectin (Sigma Chemical Co.) were used.

Laminin-1 (10 μg for Coomassie Blue staining or 100 ng for Western blot analysis) was incubated with peptide G or peptide X at a 1:1 (w/w) ratio for 1 h at 37°C. An excess of peptide was used to obtain the maximal biological effect as described previously (4). Cathepsin B was activated by incubation in the presence of 10 mm DTT and 5 mm EDTA for 10 min at 37°C.

Laminin-1 in a final volume of 40 μl was digested for 5 h at 37°C. Cathepsin B was used at an enzyme:substrate ratio of 1:5 (w/w), and cathepsin D (2 units/μg) and cathepsin G (2 milliunits/μg) were used. For time course analyses, laminin-1 or fibronectin degradation was monitored for periods of up to 6 h. In the competition experiment, YIGSR was used at peptide:laminin-1 molar ratios of 5:1, 50:1, and 500:1. Reactions were stopped by adding E64 (10 μm) or SDS-PAGE sample buffer, depending on the assay system.

Samples obtained from proteolytic degradation were separated by 3–12% gradient or 7.5% polyacrylamide gels in the presence of SDS under reducing conditions. Gels were stained with Coomassie Blue or analyzed by Western blot of proteins separated by SDS-PAGE and transferred electrophoretically to nitrocellulose membranes (Hybond C Super; Amersham) using the enhanced chemiluminescence detection system (Amersham).

Laminin-1 samples incubated in the presence or absence of peptide G followed by treatment with cathepsin B were separated by chromatography (Akta system) on a Superose 12HR column (Amersham Pharmacia Biotech) equilibrated in 50 mm potassium phosphate buffer (pH 7.5), 10% glycerol, 2 mm EDTA, 5 mm 2-mercaptoethanol, and 250 mm NaCl. NH₂-terminal sequences were determined both on aqueous protein samples and on Pro-blot electrotransferred samples (12) using an automated protein sequencer (Applied Biosystems Model 492 Procise).

MDAMB231 cells (5 × 10³) were seeded in 96-well plates in the presence of laminin or cathepsin B-cleaved laminin (20 μg/ml) pretreated or not pretreated with peptide G. At 18 h after seeding, cells were fixed daily in ice-cold 10% trichloroacetic acid for 6 days and incubated with 0.4% sulforhodamine B in 1% acetic acid (100 μl/well) for 30 min. After three washes in 1% acetic acid, the dye was dissolved in 10 mm Tris (pH 10.5; 100 μl/well) and evaluated spectrophotometrically at 492 nm (13).

Equal amounts (20 μg/ml) of intact or cleaved laminin-1 were adsorbed on 96-well plates (Greiner Labortechnik, Frickenhausen, Germany) for 1 h at 37°C. MDAMB231 cells (1 × 10⁴ cells/well) in serum-free culture medium were added and allowed to adhere at 37°C. At different times, cells were photographed with a reverse-phase microscope. After 2 h, plates were filled with PBS, inverted, and shaken in a tank of PBS for 15 min. Adherent cells were fixed in ice-cold 10% trichloroacetic acid, labeled with sulforhodamine B, and evaluated as described above for the proliferation assay.

The ability of intact or cleaved laminin-1 to stimulate cell migration was assessed in a Boyden chamber (NeuroProbe, Gaithersburg, MD). For chemotaxis assay, the lower and upper compartments were separated by 8-μm-pore polycarbonate filters (Osmonics, Livermore, CA); 40 μl of serum-free culture medium and equal amounts (20 μg/ml) of intact or cleaved laminin-1 were added to the lower wells, whereas 40 μl of MDAMB231 cells resuspended in serum-free medium were transferred to the upper wells (4 × 10⁴ cells/well). For haptotaxis assays, filters were coated on the underside with intact or cleaved laminin-1 (20 μg/ml) for 16 h at 37°C. Lower chambers were filled with culture medium, and cells were seeded into the upper chambers as described for the chemotaxis analysis. Cells were allowed to migrate for 5 h at 37°C. Filters were fixed and stained, and cells migrating to the underside were counted in four microscopic fields. Laminin-1 binding to the membrane was verified by Coomassie Blue staining.

To investigate the effect of 67LR binding on the proteolytic cleavage of laminin-1, we used a 20-amino acid soluble peptide (peptide G) corresponding to the 67LR laminin binding site that was able to bind laminin and induce conformational changes mimicking the effect of the entire receptor (4, 14). Three different cathepsins were tested: (a) the cysteine protease cathepsin B; (b) the aspartic protease cathepsin D; and (c) the serine protease cathepsin G. All of these enzymes are overexpressed and secreted by the most aggressive tumors, and all are able to cleave laminin-1 (6, 8, 9, 15, 16).

Purified murine laminin-1 incubated with peptide G or scrambled peptide X or neither was digested with a suitable amount of proteolytic enzyme. Electrophoretic separation under reducing conditions of fragments generated by laminin cleavage revealed an altered pattern of degradation only when laminin was incubated with peptide G before treatment with cathepsin B (Fig. 1,A); in addition to cleavage products ranging in size from 100–200 kDa, a fragment of ∼60 kDa was detected (Lane 4), instead of the 70-kDa polypeptide observed after degradation of native (Lane 2) or peptide X-treated laminin (Lane 3). By contrast, incubation of laminin with peptide G did not affect the cleavage of the adhesion molecule by cathepsin D or cathepsin G (data not shown). Because the presence of many disulfide bridges linking the three laminin chains is likely to prevent the release in a soluble form of cleaved fragments (16), degradation products were analyzed under nonreducing conditions by gel filtration chromatography in fast protein liquid chromatography to characterize the fragments physiologically released after cathepsin B cleavage. Unlike the electrophoretic analysis under reducing conditions, which revealed fragments ranging in size from 60–200 kDa, the elution profiles from the gel filtration column revealed two major peaks in addition to a peak corresponding to eluted cathepsin B (Fig. 1, B and C). For native laminin digested with cathepsin B (Fig. 1,B), the first major peak (ln) represented a molecule with a slightly lower molecular mass than that of intact laminin-1, whereas the second peak was consistent with a single fragment of ∼70 kDa. NH₂-terminal microsequencing of this latter fragment demonstrated its origin from two cleavages on the laminin β1 chain (16), the first at Ala³¹, which generates a small NH₂ terminus fragment, and the second at about the level of domain V, leading to the release of the 70-kDa fragment (Fig. 1,D, left). When laminin-1 was incubated with peptide G immediately before cathepsin B cleavage, the elution profile from the gel filtration column again indicated laminin eluting as a single peak with a slightly lower molecular mass than that of the native protein, but a second peak corresponded to a 60-kDa fragment (Fig. 1,C). NH₂-terminal microsequencing of the latter protein indicated the presence of a single polypeptide derived from a single cleavage at Ala²⁷¹⁰ of the laminin α1 chain (Fig. 1,D, right). The altered pattern of degradation products obtained from peptide G-treated laminin provides evidence that the modification of laminin-1 structure by interaction with peptide G leads to the exposure of hidden cleavage sites, allowing the release of a new proteolytic fragment. Besides the different pattern of degradation, time course analysis by Western blot of laminin cleavage by cathepsin B revealed an increased degradation rate in peptide G-pretreated laminin as compared with degradation in the absence of peptide (Fig. 2,A). No fragments were seen, indicating the inability of anti-laminin polyclonal serum to detect cleavage products. This increase in degradation rate was not due to a nonspecific effect of peptide G on cathepsin B activity exerted through, for example, binding to the active site of the enzyme because no effect was observed in peptide G-treated and cathepsin B-digested fibronectin, another basement membrane component known to be a substrate for cathepsin B (data not shown). To further demonstrate the role of peptide G-induced changes on laminin cleavage rate, cathepsin B degradation of peptide G-treated laminin was evaluated in the presence of the pentapeptide YIGSR, which corresponds to the laminin sequence specifically bound by 67LR (17). Western blot analysis of laminin-1 incubated with peptide G before treatment with cathepsin B for 5 h confirmed the increase of degradation rate in peptide G-treated laminin compared with untreated laminin (Fig. 2,B). Moreover, the increase in laminin-1 degradation rates was abrogated by competition of peptide G-induced conformational modification of adhesion molecule with the pentapeptide YIGSR, whereas no effect was observed with the related but inactive peptide YIGSK (Ref. 17; Fig. 2 C), pointing to the specificity of the effects on laminin cleavage rate upon peptide G-induced changes. The need for 500× concentrations of YIGSR to detect abrogation of the peptide G-mediated degradation increase likely reflects our use of a 500-fold molar excess of peptide G to obtain the maximal biological effect on laminin (see “Materials and Methods”). Indeed, only at this concentration can YIGSR complex all of the free peptide G to abrogate its effect. The peptide G-induced increase in laminin degradation rates raises the possibility that 67LR overexpression combined with production and secretion of cathepsin B in a primary tumor drastically accelerates the loss of basement membrane integrity.

After basement membrane degradation, the metastatic potential of tumor cells rests mainly in the probability of receiving signals from the microenvironment. To determine whether fragments derived from laminin cleavage after 67LR binding modification affect tumor cell behavior, we evaluated tumor cell growth, adhesion, and migration in vitro in the presence of peptide G-modified cleaved laminin-1. MDAMB231 breast carcinoma cells seeded in wells containing native laminin or cathepsin B-digested laminin in the presence or absence of peptide G showed no difference in proliferation ratio under any of these culture conditions (data not shown), whereas adhesion analysis (Fig. 3,A) revealed the immediate development of filopodia and lamellipodia as well as stable binding with the adhesion molecule within 2 h when cells were seeded on laminin-1 or peptide G-modified laminin-1. When wells were coated with cathepsin B-cleaved laminin-1, cells remained round and formed aggregates, indicating the presence of homophylic cell-cell interactions. By contrast, no aggregates formed in cells plated on plastic alone. The results suggest that a laminin cleavage fragment(s) mediates some cell-cell interaction. Indeed, such rounded cells did not remain tethered to the plate surface after washing and showed an adhesion level sharply lower than that of control cells seeded on noncoated plates (Fig. 3,B). The morphological changes induced by peptide G-modified cleaved laminin were similar to those induced by native laminin, i.e., most cells spread immediately after seeding and acquired the typical star-shaped morphology within a few hours. The ability to spread did not correlate with the formation of a stable bond with the substrate because only uncleaved laminin, either native or peptide G- or peptide X-modified laminin, supported stable cell adhesion (Fig. 3,B). Thus, the slight cell spreading induced by cleaved laminin upon peptide G modification probably reflects the formation of weak bonds that are not involved in stable focal adhesion. In both chemotaxis and haptotaxis assays, the motility of MDAMB231 cells was increased ∼40-fold when native or peptide G- or peptide X-modified laminin was used as chemoattractant in a Boyden chamber, whereas no increase was observed using cathepsin B-digested laminin in the presence or absence of peptide X. On the other hand, peptide G-modified, cathepsin B-digested laminin-1 maintained chemotactic and haptotactic effects (Fig. 3 C). These data demonstrate that cathepsin B-cleaved laminin under any conditions is impaired in its ability to support cell adhesion, although the laminin cleaved after 67LR binding modification induces cell spreading. Thus, laminin still stimulates pseudopodia protrusion without generating stable focal adhesion. This behavior is typical of the invasive phenotype, in which coordinated and temporally limited attachment and detachment determine cell migration. Indeed, laminin digested by cathepsin B in the presence of peptide G promotes cell motility, whereas the cathepsin B proteolytic cleavage of unmodified laminin removes physical barriers to cell migration but leads to the complete loss of cell adhesion or migration capabilities.

To identify the fragments of peptide G-modified, cathepsin B-cleaved laminin that stimulate cell migration, we analyzed the chemotactic activity of the laminin cleavage products. Gel filtration chromatographic separation of peptide G-modified laminin cleaved by cathepsin B, followed by Boyden chamber assay of eluted fractions corresponding to the two major peaks shown in Fig. 2,C, revealed chemotactic activity in the fractions containing the 60-kDa fragment, whereas none of the fractions containing the remaining laminin molecule stimulated cell migration (Fig. 4). Thus, the chemotactic activity of laminin associated with the COOH-terminal part of the α1 chain. This region contains the sequence LQVQLSIR, one of the most important biologically active sites of laminin-1 (18), and was previously shown to promote tumor cell invasion in vitro and in vivo (19), probably by interacting with syndecan-1 (20). The COOH-terminal globular region of laminin contains other important sites and is involved in binding of α-dystroglycan, a high-affinity laminin-1 receptor (21, 22). The interaction between this molecule and the G domain of laminin is essential for the assembly of basement membrane in embryoid bodies and for the maintenance of its integrity in adult tissues (23). The release of a 60-kDa fragment upon cathepsin B cleavage might compete with uncleaved laminin for α-dystroglycan binding, destabilizing the entire framework of the basement membrane. However, interaction of this 60-kDa fragment with the tumor cell surface stimulates cell migration. Based on our results, we speculate that the proteolytic cleavage of unmodified laminin maintains an intact G domain but, through conformational changes in laminin, renders this mobility site unable to interact with cells.

Our data on the activity of 67LR in enhancing tumor cell motility shed light on a mechanism by which an adhesion receptor plays a major role in tumor aggressiveness and metastasis and raise the possibility of targeting this mechanism in therapeutic approaches to block the progression of tumors at a preinvasive stage. Treatment with the peptide YIGSR, which specifically blocks 67LR binding to laminin, might inhibit the allosteric modification of the laminin structure induced by the 67LR-laminin interaction and might influence the processes involving this modification, including proteolysis and release of chemotactic fragments. Such a therapeutic strategy is supported by the ability of the peptide YIGSR to prevent in vivo lung metastasis formation in mice (24) and by our present findings that peptide YIGSR inhibits peptide G-induced laminin modification, thereby reducing the degradation rate of peptide G-treated laminin.

We thank Piera Aiello for excellent technical assistance, Mario Azzini for photographic reproduction, and Laura Mameli for manuscript preparation.