Extracellular Matrix Metalloproteinase 2 Levels Are Regulated by the Low Density Lipoprotein-related Scavenger Receptor and Thrombospondin 2*

We have recently shown that the adhesive defect observed in dermal fibroblasts derived from thrombospondin 2 (TSP2)-null mice results from an increase in matrix metalloproteinase 2 (MMP2) levels (Yang, Z., Kyriakides, T. R., and Bornstein, P. (2000) Mol. Biol. Cell 11, 3353–3364). Adhesion was restored by replacement of TSP2 and by inhibitors of MMP2 activity. In pursuing the observation that TSP2 and MMP2 interact, we now demonstrate that this interaction is required for optimal clearance of extracellular MMP2 by fibroblasts. Since TSP2 is known to be endocytosed by the scavenger receptor, low density lipoprotein receptor-related protein (LRP), we determined whether interference with LRP function affected fibroblast adhesion and/or extracellular MMP2 levels. Addition of heparin, which competes for the binding of TSP2 to LRP coreceptor proteoglycans, inhibited adhesion of control but not TSP2-null cells, and a blocking antibody to LRP as well as the LRP inhibitor, receptor-associated protein, also inhibited adhesion and increased MMP2 levels only in control fibroblasts. TSP2 did not inhibit active MMP2 directly and did not inhibit the activation of pro-MMP2. Finally, the internalization of 125I-MMP2 was reduced in TSP2-null compared with control fibroblasts. We propose that clearance of MMP2-TSP2 complexes by LRP is an important mechanism for the regulation of extracellular MMP2 levels in fibroblasts, and perhaps in other cells. Thus, some features of the phenotype of TSP2-null mice, such as abnormal collagen fibrillogenesis, accelerated wound healing, and increased angiogenesis, could result in part from increased MMP2 activity.

Thrombospondins (TSP) 1 1 and 2 are large extracellular macromolecules whose diverse functions reflect their ability to bind to multiple cell-surface receptors, cytokines, growth factors, and proteases, and to structural components of the matrix (1,2). TSP2-null mice display a complex phenotype that is characterized by changes in connective tissues, particularly in response to injury, an increase in vascular density and en-dosteal bone growth, and a bleeding defect (3)(4)(5). Dermal fibroblasts, isolated from adult animals, also show an adhesive defect in vitro that is most evident when cells are plated on a variety of pure protein substrates in the absence of serum. Adhesion was restored by prolonged (48 h) incubation of TSP2null cells with recombinant mouse TSP2 or by transfection with a TSP2 cDNA gene (6). The basis for this adhesive defect was recently investigated and was shown by zymography to result from an increase in matrix metalloproteinase 2 (MMP2) levels in both the conditioned media and cell layers of cultured cells (6). Although virtually all of the enzyme that was analyzed was in the zymogen or pro-MMP2 form, an increase in active MMP2 in TSP2-null cells was inferred from the observation that inhibitors of MMP2, including TIMP2 and a neutralizing anti-MMP2 antibody, corrected the adhesive defect (6).
Both TSP2 and its close relative, TSP1, are known to interact with the scavenger receptor, low density lipoprotein-related receptor protein (LRP), an interaction that results in the endocytosis and lysosomal degradation of the TSP (7)(8)(9)(10)(11). This interaction is mediated by the NH 2 -terminal heparin-binding domain of the TSP and is competed by heparin (7, 9 -11). TSP1 and TSP2 also interact with a number of serine proteases, including plasmin, cathepsin G, and neutrophil elastase, and function as competitive inhibitors of these enzymes (12). In view of the fact that LRP is capable of binding and endocytosing both ␣ 2 -macroglobulin proteinase and plasminogen activator inhibitor 1-tissue plasminogen activator and -urokinasetype plasminogen activator complexes (13), we postulated that TSP2-MMP2 complexes might also be cleared by the scavenger receptor.
In this study we find that although TSP2 binds both pro-MMP2 and MMP2 directly, the protein does not function as a direct binding inhibitor of the active protease nor does it prevent the activation of pro-MMP2. However, TSP2-null cells were defective in the uptake of extracellular MMP2. Furthermore, inhibitors of LRP function reduced adhesion of control skin fibroblasts and, correspondingly, increased MMP2 levels in these cells. We therefore propose that the interaction of MMP2 with TSP2, and possibly also with TSP1, and the subsequent uptake of the protein-enzyme complex by LRP serve as a means of regulating extracellular MMP2 levels. These results have implications not only for the phenotype of the TSP2-null mouse but also for the control of processes such as collagen fibrillogenesis, wound healing, and angiogenesis. purchased from Chemicon (Temecula, CA). Rabbit anti-MMP2 polyclonal antibody was purchased from Chemicon. For the direct-binding enzyme-linked immunosorbent plate assay, 96-well microtiter plates (Linbro®, Flow Laboratories, McLean, VA) were coated with TSP1, TSP2, gelatin (Fisher), and asialofetuin (Sigma) at 10 g/ml, 50 l/well in 0.1 M Tris-HCl, pH 7.2, at 4°C overnight. The plates were blocked with 1% bovine serum albumin in the same buffer containing 5 mM CaCl 2 for 1 h at room temperature. MMP2 (4 g/ml) in blocking solution was then added to wells for 2 h. Subsequently, rabbit anti-MMP2 antibody (1 g/ml) was added for another 2 h, followed by alkaline phosphatase-conjugated goat anti-rabbit IgG (Sigma). Color was developed with p-nitrophenyl phosphate substrate (1 mg/ml) in 10 mM diethanolamine, pH 9.5, containing 0.5 mM MgCl 2 . Between each incubation, the wells were washed with Tris-HCl buffer to remove unbound protein.
A 405 was measured in a microplate reader with SOFTmax®PRO software (Molecular Devices, Sunnyvale, CA). Each determination represents the average of four wells.
Gelatinolytic Assay-MMP2 activity was determined by a gelatinolytic assay with soluble gelatin as a substrate (15). The gelatin was radiolabeled with [ 3 H]acetic acid according to Cawston and Barrett (16). Two g of human MMP2 was incubated in 100 l of reaction buffer containing 50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 5 mM CaCl 2 , and 0.01% Brij-35 at 37°C for 2 h, with or without activation by 1 mM 4-aminophenylmercuric acetate (APMA). In experiments that tested the ability of TSP2 to inhibit MMP2, TSP2 was added in a chain molar ratio of 1 to 1 with respect to MMP2. Fifty g of radiolabeled gelatin was heat-denatured at 60°C for 15 min, cooled to 37°C, and added to the incubation mixture in a final volume of 200 l. The reaction was allowed to proceed for 24, 48, 72, and 96 h at 37°C in the presence of 0.03% toluene to prevent bacterial contamination. Undegraded gelatin was precipitated at 4°C with a mixture of 4% trichloroacetic acid and 0.8% tannic acid. The reaction mixture was centrifuged at 10,000 ϫ g for 15 min at 4°C, and aliquots of the resulting supernatants were counted for radioactivity in a liquid scintillation counter.
Cell Culture-Dermal fibroblasts were isolated by collagenase treatment of skin taken from the backs of adult (2-3-month-old) mice. Each cell preparation was derived from an individual mouse. Briefly, after removal of hair and subcutaneous tissues, skin was treated with 0.25% trypsin and antibiotics in a calcium-and magnesium-free solution.
After overnight incubation at 4°C, the epidermis was stripped off the dermis by holding the dermis with forceps and gently scraping off the epidermis with a scalpel. The isolated dermis was then digested in 0.25% bacterial collagenase (Sigma) in DMEM containing 0.02% CaCl 2 and 0.01% MgSO 4 at 37°C for 1-2 h until the cells were completely dissociated from the digested tissue. Cells derived from a 2-cm 2 segment of skin were collected by centrifugation and washed twice with DMEM, supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 units/ml penicillin G, and 100 g/ml streptomycin before plating on a 100-mm tissue culture dish in the same medium. Unattached cells and debris were removed, and the medium was replaced after 2 days of plating. Cells were passaged at confluence. After 2-3 passages, the cell population appeared, by light microscopy, to be composed entirely of fibroblasts.
Iodination and Uptake of MMP2 by Fibroblasts-MMP2 was labeled with 125 I using IODO-BEADS (Pierce) according the manufacturer's protocol. Briefly, 1 mCi of Na 125 I (Amersham Pharmacia Biotech) was incubated with IODO-BEADS for 5 min in 50 mM sodium phosphate buffer, pH 6.5, and then 20 g of MMP2 was added to a final volume of 200 l. After 10 min of incubation, the reaction supernatant was removed and added to another tube containing 0.5 ml of 0.1% bovine serum albumin as a carrier protein. The IODO-BEADS were washed twice with 150 l of reaction buffer, and the pooled iodinated protein was gel-filtered on a PD-10 column (Amersham Pharmacia Biotech) to remove free 125 I. Fibroblast monolayers in 12-well tissue culture plates were cultured in serum-free medium for 1 h, and then 30 nM of 125 I-MMP2 was added (0.5 ml/well). 125 I-MMP2-containing media were removed hourly over a 5-h period, and the cells were washed three times with DMEM, dissolved in 0.5 ml of 0.1 N NaOH, and neutralized with 30 l of 10% glacial acetic acid. In a separate experiment, fibroblasts in 6-well tissue culture plates were incubated in serum-free medium containing 30 nM of 125 I-MMP2 as described above. After a 5-h incubation, cells were washed three times with DMEM and then either dissolved in NaOH and neutralized or treated with 0.25% trypsin at 37°C for 10 min. The cells were then washed twice with DMEM, dissolved in NaOH, and neutralized with acetic acid. Aliquots of the resulting lysates were counted for radioactivity in a liquid scintillation counter to measure the internalization of 125 I-MMP2 by fibroblasts.
Cell Attachment Assays and Zymography-Analyses for attachment of fibroblasts on fibronectin-coated 96-well plates were performed as described previously (6). Fibronectin was coated at a concentration of 5 g/ml in phosphate-buffered saline. Dermal fibroblasts were incubated in cell culture medium containing heparin (5 M), anti-LRP IgG (50 g/ml), or RAP (1 M) for 48 h prior to analysis. The polyclonal rabbit anti-LRP functional blocking antibody (17) was affinity-purified (10), and recombinant human RAP, expressed in bacteria as a fusion protein with glutathione S-transferase, was prepared and purified as described previously (18). Normal rabbit IgG was purchased from Vector Laboratories (Burlingame, CA) and was used as a control. Zymography of conditioned media was performed as described by Yang et al. (6). Western Blotting-Serum-free conditioned media or lysates of mouse skin fibroblasts were subjected to SDS-PAGE in 7.5 or 4 -15% gradient gels. A buffer composed of 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, and 0.1% SDS was used to extract cellular proteins from fibroblasts. Separated proteins were transferred electrophoretically (19) from polyacrylamide gels to nitrocellulose membranes in a mini trans-blot cell (Bio-Rad) for 1 h at 100 V, followed by Western blot analysis with anti-LRP IgG or anti-TSP2 polyclonal antibody (14). The resulting antigen-antibody complexes were detected by incubation with alkaline phosphatase-conjugated antibody (Sigma) and alkaline phosphatase substrate (Bio-Rad) for determination of LRP expression. Incubation with horseradish peroxidase-linked antibody (New England Biolabs, Beverly, MA) and enhanced luminol reagent (PerkinElmer Life Sciences) and exposure to x-ray film were used to determine TSP2 levels.

RESULTS AND DISCUSSION
TSP2 Does Not Inhibit MMP2 Activity or the Activation of Pro-MMP2 by APMA-We recently reported that TSP2-null fibroblasts have a marked adhesive defect when plated on a number of protein substrates, and we showed that this defect resulted from a 2-fold increase in MMP2 protein in the conditioned media of the cells (6). The adhesive defect was corrected by inhibitors of MMP2, by transfection of a full-length TSP2 cDNA gene into the cells, or by long term preincubation with recombinant TSP2 but not by adsorption of TSP2 to tissue culture plastic. Furthermore, we and others (6,20) have found that TSP2 interacts directly with pro-MMP2. TSP2 could modulate MMP2 activity in a number of ways. TSP2 could function as a direct binding inhibitor of MMP2, as has been shown for the binding of both TSP2 and TSP1 to a number of serine proteases (12). TSP2 could also serve to sequester MMP2 in the extracellular matrix and reduce its bioavailability since TSP1, and probably TSP2 in view of its very similar sequence, bind to fibronectin and to a number of collagens and proteoglycans (2,21). This explanation seems unlikely in our cell culture system since not only MMP2 activity, but also protein, was increased in the absence of TSP2 (6). Finally, TSP2 could facilitate the clearance of TSP2 from the pericellular environment.
As shown in Fig. 1, both the zymogen, pro-MMP2, and APMA-activated MMP2 bound equally well to TSP1 and TSP2 in a direct binding solid phase assay. The extent of binding was only about 60% that to gelatin, but this difference is in keeping with a postulated role of TSP2 as a modulator of MMP2 activity rather than as a substrate for the enzyme. In preliminary experiments, the K d value for binding of pro-MMP2 to TSP2 was found to be in the micromolar range. 2 However, as revealed by a gelatinolytic assay with soluble [ 3 H]gelatin as a substrate, the binding of TSP2 to pro-MMP2 did not inhibit its activation by APMA (Fig. 2). Furthermore, TSP2 did not inhibit the activity of APMA-activated MMP2. Fig. 2 also shows that the small amount of activity associated with preparations of pro-MMP2 was not inhibited by TSP2.
Heparin Inhibits the Attachment of Control but Not TSP2null Fibroblasts-The accumulation of MMP2 protein in the conditioned media of TSP2-null cells argues strongly for a role for TSP2 in clearance of MMP2 from the pericellular environ-ment of fibroblasts. Both TSP1 and TSP2 are known to be endocytosed and catabolized by the LRP scavenger receptor (7)(8)(9)(10)(11). This process probably involves heparan sulfate proteoglycans as coreceptors since cellular uptake of the heparinbinding domain of TSP1 was inhibited by heparin, and degradation of this domain was reduced in Chinese hamster ovary cells that lacked heparan sulfate glycosaminoglycans (8). We hypothesized that TSP2-MMP2 complexes might also be bound and endocytosed by the LRP receptor. If so, the addition of heparin to wild-type cells in culture should lead to reduced adhesion, whereas a much smaller effect should be seen with TSP2-null cells. We therefore added heparin at 5 M to the culture media of dermal fibroblasts 48 h prior to determination of their attachment properties. As shown in Fig. 3, treatment of wild-type fibroblasts decreased their attachment to fibronectin by 57% and led to a level of attachment that equaled that of TSP2-null cells, whereas heparin reduced the attachment of TSP2-null cells by only 15%. These findings, together with the known ability of TSP2 to interact with LRP and MMP2, suggest that TSP2 could play a major role in regulating MMP2 levels and adhesion in mouse dermal fibroblasts.
Anti-LRP Antibodies and RAP Reduce Attachment and Increase MMP2 and TSP2 Levels in Control Mouse Dermal Fibroblasts-LRP was first identified as a homologue of the low density lipoprotein receptor (22) and is now recognized as a member of a family of at least six homologous mammalian receptors that include the low density lipoprotein receptor, LRP, LRP-DIT, gp330 (megalin), very low density lipoprotein receptor, and apoER-2. It was subsequently shown that LRP is identical to a receptor that had been described as responsible for the internalization of ␣ 2 -macroglobulin-proteinase complexes (23,24). LRP is synthesized as a single chain that is cleaved to form 515-and 85-kDa subunits (25). The distribution of LRP in human tissues and its subcellular location in human keratinocytes and fibroblasts has been described (26,27). Although LRP is known to be expressed in embryonic mouse fibroblasts (17) it has not, to our knowledge, been reported in adult mouse dermal fibroblasts. Cell lysates of dermal fibroblasts were therefore subjected to SDS-PAGE and Western blot analysis with an anti-LRP polyclonal antibody. As shown in Fig. 4, two specific components with molecular masses of 515 and 85 kDa, corresponding to ␣and ␤-subunits of LRP, respectively, were detected in dermal fibroblasts. It can be seen that the levels of LRP protein, as judged by Western blot analysis, are at least as high in TSP2-null fibroblasts as in wild-type cells. Thus a defect in internalization of TSP2-MMP2 complexes would not appear to result from a deficiency of LRP.
As a more stringent test of our hypothesis that the uptake of TSP2-MMP2 complexes by LRP serves as a means of regulating extracellular MMP2 levels, we treated dermal fibroblasts with a polyclonal rabbit anti-LRP functional blocking antibody or with the LRP inhibitor, RAP. RAP is a 39-kDa protein that is located in the endoplasmic reticulum but binds and antagonizes the function of LRP and other members of the low density lipoprotein receptor family (13). Since RAP and LRP are found in different cellular compartments, its mode of action is unclear. Willnow et al. (28) have proposed that RAP functions as a chaperone to prevent ligand-induced aggregation and degradation of LRP in the endoplasmic reticulum. In any event, extracellularly administered RAP has been found to be effective in inhibiting the internalization and degradation of both intact TSP1 and its heparin-binding domain (7,8,11). Treatment of fibroblasts with anti-LRP IgG at 50 g/ml or RAP at 1 M significantly decreased the attachment of wild-type cells to fibronectin to levels that are close to those of TSP2-null cells, but had essentially no effect on the attachment of TSP2-null cells (Fig. 5). These results are consistent with those of heparin treatment (Fig. 3). As predicted by our postulate that anti-LRP or RAP inhibited the binding and internalization of a TSP2-MMP2 complex, zymography of serum-free conditioned media from dermal fibroblasts demonstrated that treatment with these agents significantly increased MMP2 levels in wild-type FIG. 1. Both pro-MMP2 and activated MMP2 bind to TSP2. The interaction of MMP2 with substrate-bound proteins was determined in a direct binding solid phase assay. Pro-MMP2 and APMA-activated MMP2 bound equally well to TSP2 and TSP1, but the extent of binding was only about 60% that to gelatin. Binding to asialofetuin served as a negative control. This experiment is representative of three independent experiments.

FIG. 2. TSP2 neither inhibits the activation of pro-MMP2 by APMA nor the activity of active MMP2.
Pro-MMP2 activity, with or without activation by 1 mM APMA, was determined in a gelatinolytic assay using soluble [ 3 H]gelatin as a substrate. Preincubation of pro-MMP2 with TSP2 did not inhibit its activation by APMA: compare pro-MMP2/TSP2 ϩ APMA with MMP2/APMA. Also, the gelatinolytic activity of activated MMP2 was not affected by TSP2: compare MMP2/APMA ϩ TSP2 with MMP2/APMA. but not in TSP2-null cells (Fig. 6). Analysis of the zymogram in Fig. 6 indicated that 79 and 76% of the difference in MMP2 levels between wild-type and TSP2-null cells were restored by anti-LRP and RAP, respectively. Furthermore, SDS-PAGE and Western blot analysis of conditioned media revealed that the extracellular level of TSP2 in wild-type cells increased significantly after treatment with anti-LRP IgG or RAP (Fig. 7). Since tissue inhibitor of metalloproteinases 2 (TIMP2) levels, as determined by Western blot analysis, did not differ appreciably between wild-type and TSP2-null cells, 3 these experiments establish the TSP2/LRP system as a major regulator of extracellular MMP2 levels and activity.
There is precedence for the involvement of LRP in the cellular uptake of matrix metalloproteinases. Recently, Barmina et al. (29) reported that collagenase-3 (MMP13) bound to a recep-tor, tentatively identified as being identical to a member of the mannose-receptor type C lectin family and that the internalization of the enzyme-receptor complex was in turn mediated by LRP. To determine whether MMP2 could bind directly to LRP, solid phase assays were employed using purified reagents. These experiments revealed that MMP2 was not able to bind with high affinity to LRP-coated microtiter wells 4 and confirmed that MMP2 does not interact directly with LRP.
Internalization of MMP2 by Dermal Fibroblasts Is Reduced in TSP2-null Cells-As a final test of our hypothesis, we compared the uptake of exogenous MMP2 by wild-type and TSP2null dermal fibroblasts. Iodinated MMP2 was incubated with mouse dermal fibroblasts in serum-free medium for 5 h. The cellular uptake of 125 I-MMP2 was then measured by liquid scintillation counting of cell lysates after removal of media containing unbound 125 I-MMP2. As shown in Fig. 8A, the association of 125 I-MMP2 with TSP2-null fibroblasts was significantly lower than that for control cells at each time point of the experiment (p Ͻ 0.01). At the end of the 5-h incubation period, the association of 125 I-MMP2 with cells lacking TSP2 was only 67% that for wild-type cells. Equal amounts of dermal fibroblast lysate proteins, obtained directly from cell monolayers or from trypsinized cells, were subjected to SDS-PAGE in a 4 -15% gradient gel and Western blot analysis with polyclonal anti-LRP antibody. Two specific components with molecular masses of 515 and 85 kDa corresponding to the ␣and ␤-subunits of LRP, respectively, were detected in dermal fibroblasts. The expression of LRP in TSP2-null fibroblasts was at least equal to that in wild-type cells. Molecular masses were estimated from concurrently run molecular mass standards. The medium was then changed to serum-free medium containing the same reagents and culture resumed for another 20 -24 h. Equal amounts of conditioned medium protein were then applied to SDS-PAGE, 0.1% gelatin under nonreducing conditions. The gelatinolytic activity of MMP2 was significantly increased in media from wild-type fibroblasts treated with anti-LRP IgG or RAP and was almost equal to that of TSP2-null fibroblasts. Treatment of TSP2-null fibroblasts did not have a significant effect on the activity of MMP2. This experiment was performed twice with very similar results.
As a control to determine what fraction of cell-associated MMP2 was internalized, cells were incubated for 5 h with 125 I-MMP2 and then treated with trypsin prior to counting. As shown in Fig. 8B, 55% of the radioactivity associated with wild-type cells in monolayer culture, and 39% of that associated with TSP2-null cells, was retained after trypsinization. The 125 I-MMP2 released by trypsin presumably includes both surface-bound MMP2 and enzyme that was trapped nonspecifically in the monolayer. Nevertheless, the reduction in internalization of MMP2 by TSP2-null cells, compared with controls, is clearly preserved after trypsinization. In this experiment the uptake of MMP2 by TSP2-null cells was only 48% of controls, a greater difference than that seen at 5 h in Fig. 8A. In unpublished experiments 3 we have also shown that prior incubation of wild-type cells with 1 M RAP reduces uptake of MMP2 to the same extent as incubation with heparin, as would be predicted if LRP serves as the major mechanism for internalization of TSP2-MMP2 complexes. The basis for the non-TSP2mediated uptake of MMP2 is not known. It is not likely to result from binding to TSP1 and LRP-mediated endocytosis of TSP1-MMP2 complexes since the attachment of TSP2-null cells is not reduced much further by prior addition of heparin to the culture medium (Fig. 3). Bein and Simons (20) have reported recently that human TSP1 does not directly inhibit the degradation of type IV collagen by MMP2 in an in vitro assay. These authors also showed that TSP1 inhibited the activation of pro-MMP9 by MMP3 in vitro. However, their conclusion, based on cell culture experiments with bovine aortic endothelial cells, that TSP1 inhibits the activation of pro-MMP2 is open to the alternative explanation that TSP1 increased the clearance of both pro-MMP2 and active MMP2 by the very low density lipoprotein receptor in these endothelial cells. Furthermore, a role for TSPs in inhibi-tion of pro-MMP2 activation is not easily reconcilable with an increase in not only MMP2 activity but also MMP2 protein, as shown by Yang et al. (6) in TSP2-null cells.
Implications of Increased MMP2 Levels in Fibroblasts for the Phenotype of the TSP2-null Mouse-A number of features of the TSP2-null mouse, including abnormal collagen fibrillogenesis and increased angiogenesis in skin and in subcutaneous tissues in response to injury, remain unexplained. In recent experiments we have found that, despite the ready demonstration that TSP2-null dermal fibroblasts in culture accumulate increased levels of MMP2 (see Ref. 6 and this study), analyses of extracts of uninjured dermis by gelatin zymography failed to show a difference in MMP2 levels between normal and TSP2null tissue. 3 This apparent discrepancy can be explained by the fact that the TSP2 content of uninjured adult mouse skin is very low, at least by the criterion of immunohistochemistry (14). These findings are also in accord with increasing evidence that matricellular proteins such as TSP2 are expressed predominantly during development and growth and in response to injury (2). Indeed, cells in culture in serum-containing medium, which synthesize considerable amounts of TSP2, display many aspects of a reaction to injury, a phenomenon that has been termed "culture shock" (30). Recently our laboratory has shown that, in contrast to normal dermis, the MMP2 content of the provisional matrix or granulation tissue formed in a healing excisional skin wound is markedly increased. 5 Much of the newly formed TSP2 in a healing wound is associated with collagen fibers. It is therefore possible that TSP2 does play a role, either directly or indirectly, in normal collagen fibrillogenesis and that the abnormally sized and shaped der-5 T. R. Kyriakides and P. Bornstein, unpublished data. FIG. 7. TSP2 levels in conditioned media of dermal fibroblasts treated with anti-LRP IgG or RAP. Equal amounts of protein from conditioned media of dermal fibroblasts treated with anti-LRP IgG or RAP were subjected to SDS-PAGE and Western blot analysis with anti-TSP2 antibody. The levels of TSP2 in wild-type conditioned media were significantly increased after cells were treated with these agents. As expected, no TSP2 was detected in TSP2null cells.
FIG. 8. Uptake of 125 I-MMP2 by mouse skin fibroblasts. A, mouse skin fibroblasts were incubated with 30 nM 125 I-MMP2 in serum-free media for periods up to 5 h. At each time point, media containing unbound 125 I-MMP2 were removed; aliquots of cell lysates were measured by scintillation counting, and the results were normalized to cell protein. The association of 125 I-MMP2 with TSP2-null fibroblasts was significantly lower than that for wild-type cells. Each data point represents the average of four determinations. For most time points the error bars do not extend beyond the symbols on the graph. This experiment was performed twice with very similar results. B, after a 5-h incubation, cell monolayers were either lysed directly after washing or treated with 0.25% trypsin to remove noninternalized 125 I-MMP2, and the cells were lysed for scintillation counting. The internalization of 125 I-MMP2 was substantially reduced in TSP2-null cells, as compared with that in controls. mal and tendon collagen fibrils seen by electron microscopy in the TSP2-null mouse (3) reflect the absence of TSP2 and consequent increase in MMP2 during this critical period of morphogenesis. More recent, albeit circumstantial, evidence for this view comes from electron microscopic examination of 4and 8-day postnatal mouse hindlimb flexor tendons. It was found that the tendon fibroblast processes that delimit developing collagen fibril bundles in the growing tendon were less regular and not as closely apposed to the collagen fibrils in TSP2-null as compared with normal tissues (31). Although by no means diagnostic, such changes are compatible with alterations in MMP activity.
There is a substantial literature that documents a role for MMP2 in angiogenesis (32) and in tumor growth and metastasis (33). A direct demonstration is seen in MMP2 knockout mice that manifest reduced angiogenesis and tumor progression (34). It is therefore possible that the increased dermal vascular density observed in TSP2-null mice results, in part, from increased MMP2 levels. Since LRP has not been found in umbilical vein endothelial cells (7) and may not exist in other endothelial cells, TSP2-MMP2 complexes may be bound and internalized by another member of the LRP receptor family, notably the very low density lipoprotein receptor that has been documented to bind and internalize TSP1 (11). In tissues such as skin, endothelial cells may also be influenced by the paracrine effects of fibroblasts and pericytes. 6 Conclusions and Directions for Future Research-The LRP scavenger receptor has been implicated in the binding and internalization of a number of protein inhibitor-enzyme complexes of the fibrinolytic pathway (13,35), in the cellular uptake of chylomicron remnants (36), and Pseudomonas toxin A (37), in the regulation of cell surface levels of the urokinase receptor and tissue factor (38,39), and in the processing of the ␤-amyloid precursor protein (40). Our study and that of Barmina et al. (29) now implicate LRP, directly or indirectly, in the clearance of MMP2 and MMP13 from the pericellular environment. It seems likely that extracellular levels of other members of the MMP family will also be regulated in a similar fashion, either by LRP or by other members of the LRP family. Internalization and lysosomal degradation of protein-MMP complexes represent a definitive mechanism for disposal of excessive extracellular proteolytic activity, a function that cannot be provided by the noncovalent, albeit high affinity, interaction between MMPs and TIMPs.
A major issue that requires additional study is the means by which increased MMP2 activity reduces the adhesion of fibroblasts, and possibly other cells, in vitro and perhaps in vivo. A common explanation that is given is that MMP2 is capable of degrading many of the extracellular proteins to which cells attach, including fibronectin and several collagens (41). However, this explanation may not provide the whole answer since reduced attachment of TSP2-null cells can be documented in assays that are performed for 1 h in the absence of serum, conditions that are not likely to support the deposition of a substantial matrix. Another possibility is that increased MMP2 activity may alter the cellular surface, perhaps by proteolysis of adhesion receptors, as suggested by Ray and Stetler-Stevenson (42). Finally, it is not excluded that MMP2 could interact with cell-surface receptors and generate changes in intracellular signaling that in turn compromise focal adhesions. These possibilities are currently under study in our laboratories.