The Low Density Lipoprotein Receptor-related Protein Mediates Fibronectin Catabolism and Inhibits Fibronectin Accumulation on Cell Surfaces*

Low density lipoprotein receptor-related protein (LRP) is a member of the low density lipoprotein receptor family, which functions as an endocytic receptor for diverse ligands. In this study, we demonstrate that murine embryonic fibroblasts (MEF-2 cells) and 13-5-1 Chinese hamster ovary cells, which are LRP-deficient, accumulate greatly increased levels of cell-surface fibronectin (Fn), compared with LRP-expressing MEF-1 and CHO-K1 cells. Increased Fn was also detected in conditioned medium from LRP-deficient MEF-2 cells; however, biosynthesis of Fn by MEF-1 and MEF-2 cells was not significantly different. When LRP-deficient cells were dissociated from monolayer culture, increased levels of Fn remained with the cells, as determined by cell-surface protein biotinylation, suggesting an intimate relationship with cell surface-binding sites. The LRP antagonist, receptor-associated protein (RAP), pro-moted Fn accumulation in association with MEF-1 cells, whereas expression of full-length LRP in MEF-2 cells substantially decreased Fn accumulation, confirming the role of LRP in this process. Purified LRP bound directly to immobilized Fn, and this interaction was inhibited by RAP. Furthermore, in a (cid:4) -counter. Specific 125 I-Fn degradation in MEF cul- tures was determined by subtracting degradation that occurred in the presence of unlabeled Fn and corrected for that which occurred in control wells that lacked cells. were in plates in and for 2 h. Depletion of intracellular methionine achieved by culturing in L -methionine-free DMEM for 4 h. The cells were then labeled with 35 (cid:3) Ci/ml L -[ 35 S]methionine for 4 h inserum- containing methionine-free DMEM. Cell extracts were prepared in RIPA buffer (0.1% SDS, 1% deoxycholate, 1% Nonidet P-40, 10 150 M NaCl, 0.4 m M sodium vanadate, 10 (cid:3) g/ml aprotinin, 10 (cid:3) g/ml E64, and 10 (cid:3) g/ml leupeptin). Labeled Fn was recovered by immunoprecipitation. Radioactivity in the precipitates was determined in a scintillation counter. The precipitates were then subjected to SDS-PAGE on 4–16% gradients. L -[ 35 S]Methionine incor-poration into Fn was determined using a Storm system (Molecular Dynamics and Amersham Biosciences).

Fibronectin (Fn) 1 is a multidomain glycoprotein found in the plasma as an ϳ450-kDa disulfide-linked dimer and in the extracellular matrix (ECM), in the form of larger multimers (1,2). Structural variants of Fn arise from a single gene by alternative mRNA splicing and post-translational modification (3,4). As one of the most ubiquitous and multifunctional ECM proteins, Fn plays a major role in such fundamental biological processes as cell adhesion, migration, growth, differentiation, cytoskeletal organization, hemostasis and thrombosis, and oncogenesis (5)(6)(7).
The structure of Fn includes well defined binding sites for multiple macromolecules, including cell-surface receptors, sulfated glycosaminoglycans, gelatin, and fibrin (8 -10). Many of the cellular receptors that bind Fn are members of the integrin family, including the major Fn receptor, ␣ 5 ␤ 1 , but also ␣ 4 ␤ 1 , ␣ 3 ␤ 1 , ␣ v ␤ 1 , and ␣ v ␤ 3 (11). Fn binding to integrins is particularly important because this interaction results in cell signaling involving multiple factors, including the Rho family of small GTP-binding proteins and phosphoinositide 3-OH kinase (12)(13)(14). Fn-␣ 5 ␤ 1 interactions are also critical in supporting the formation of extracellular Fn fibrils (15), which may then function to suppress signaling pathways that lead to apoptosis (16 -18). In addition to integrins, Fn fibril formation is controlled by macromolecules in the pericellular spaces and involves interaction of multiple Fn domains with other ECM components (19 -21).
Processes that regulate Fn catabolism are not well understood. One hypothesis is that Fn levels are controlled by local proteolysis (22). Serine proteinases, and in particular urokinase-type plasminogen activator (uPA) and plasmin, may play an important role (23,24); however, metalloproteinases may also be involved (25)(26)(27). In the liver, forms of Fn with terminal galactose residues may be cleared and catabolized by asialoglycoprotein receptors (28). Furthermore, certain integrins, including ␣ 5 ␤ 1 , are known to undergo endocytosis and recycling (29 -31). Thus, it is reasonable to assume that, under some circumstances, Fn may be internalized with ␣ 5 ␤ 1 .
The low density lipoprotein receptor-related protein (LRP) is a member of the LDL receptor family, which is constitutively transported through a pathway that includes rapid endocytosis in clathrin-coated pits and efficient recycling (32)(33)(34)(35). LRP binds many ligands, delivering these proteins to lysosomes, including activated ␣ 2 -macroglobulin, apolipoprotein E, plasminogen activators, proteinase-inhibitor complexes, and thrombospondin (36). Other receptors in this family, such as the VLDL receptor and megalin/LRP-2, may be partially redundant with regard to ligand binding specificity; however, members of the LDL receptor family have different patterns of expression, at the cellular level, and generate different phenotypes when the genes are eliminated in knock-out mice (37)(38)(39).
In the present study, we demonstrate that Fn accumulates to greatly increased levels in the medium and ECM surrounding fibroblasts and CHO cells that are LRP-deficient. LRP binds directly to immobilized Fn, and this interaction is inhibited by receptor-associated protein (RAP), a 39-kDa protein that functions as a general antagonist of specific LRP interactions (40 -42). Furthermore, we demonstrate that LRP mediates Fn catabolism by cells in culture. From these studies, we conclude that LRP may be a major regulator of Fn accumulation on cell surfaces.

EXPERIMENTAL PROCEDURES
Reagents and Proteins-Fn was purified from human plasma by the method of Ruoslahti et al. (43). LRP was purified from human placenta as described previously (44). GST-RAP was expressed in bacteria and purified as described previously (45) using a construct obtained from Dr. Joachim Herz (University of Texas Southwestern Medical Center, Dallas, TX). As a control, GST without fused RAP was also expressed and purified from bacteria transformed with the empty vector, pGEX-2T. Monoclonal antibody 8G1, which recognizes the 515-kDa LRP heavy chain, has been described previously (46). Fn-specific polyclonal antibody (F-3648) and globulin-free bovine serum albumin (BSA) were from Sigma. Peroxidase-conjugated donkey anti-rabbit IgG, sheep antimouse IgG, Na 125 I, and [ 35 S]methionine were from Amersham Biosciences. The expression construct, which encodes full-length human LRP in pcDNA3.1, has been described previously (47). The biotinylation reagent, sulfo-NHS-LC-biotin, was from Pierce. Other chemicals were from Sigma, unless otherwise indicated.
Biotinylation and Recovery of Extracellular Fn-Monolayer cultures of MEFs and CHO cells were washed three times with ice-cold 20 mM sodium phosphate, 150 mM NaCl, pH 7.4 (PBS), to remove contaminating fetal bovine serum and other soluble proteins and then treated with the membrane-impermeable biotinylation reagent, sulfo-NHS-LC-biotin (0.5 mg/ml), for 15 min at 22°C. Alternatively, cells were treated with sulfo-NHS-LC-biotin after dissociation from monolayer culture using enzyme-free cell dissociation buffer (Invitrogen). In these experiments, the cells were pelleted, washed with ice-cold PBS, and resuspended at 4 ϫ 10 6 cells/ml in Earle's balanced salt solution (EBSS) with 25 mM HEPES, pH 7.5, containing NHS-LC-biotin and proteinase inhibitors. Biotinylation reactions were terminated by the addition of 50 mM Tris-HCl, 150 mM NaCl, 100 mM glycine, pH 7.5, for 15 min at 22°C. After washing with PBS, the cells were counted and lysed in extraction buffer. Biotinylated membrane proteins were precipitated with streptavidin-Sepharose (Amersham Biosciences). The affinity precipitates were recovered by centrifugation, washed, boiled in SDS sample buffer, and subjected to SDS-PAGE and immunoblot analysis to detect Fn.
Fn Accumulation as a Function of Time-MEF-1 and MEF-2 cells (5 ϫ 10 5 ) were plated in 75-ml flasks in 10% fetal bovine serumcontaining medium and allowed to adhere for 2 h. The medium was then replaced with serum-free DMEM containing Nutridoma NS® sup-plement (Roche Molecular Biochemicals) and sodium bicarbonate. At various times, the cells and conditioned medium were recovered. The conditioned medium was concentrated 10-fold using Centricon units with 10-kDa exclusion limits (Millipore, Bedford, MA). Samples of conditioned medium and cells extracts were subjected to immunoblot analysis to detect Fn.
Binding of LRP to Immobilized Fn-Fn (10 g/ml) was immobilized in microtiter plates by incubation for 5 h at 22°C. The wells were blocked with 3% (w/v) BSA for 1 h. Purified LRP was incubated with immobilized Fn, or in control wells without Fn, for 12 h at 4°C. The incubation buffer was 20 mM Tris-HCl, 150 mM NaCl, pH 7.4, containing 1% (w/v) BSA, 5 mM CaCl 2 , and 0.05% Tween 20. In some experiments, GST-RAP was added together with the purified LRP. The wells were then washed, and bound LRP was detected with monoclonal antibody 8G1, followed by alkaline phosphatase-conjugated secondary antibody and the chromogenic substrate, 3,3Ј,5,5Ј-tetramethylbenzidine.
Cellular Degradation of Fn-Fn was radioiodinated to a specific activity of 0.5-1.0 Ci/g using IODO-BEADS (Pierce). MEF-1 and MEF-2 cells were transferred to 24-well dishes and cultured for 24 h. After washing the cells, 125 I-Fn (10 nM) in EBSS with 25 mM HEPES, pH 7.4, containing 5 mg/ml BSA was added. In some experiments, 125 I-Fn was added together with purified GST-RAP (200 nM) or a 50-fold molar excess of non-radiolabeled Fn. To detect degraded 125 I-Fn as a function of time, samples of the medium were assayed for trichloroacetic acid-soluble radioactivity. Each sample was incubated with 10% trichloroacetic acid at 4°C and subjected to centrifugation at 15,000 ϫ g. Trichloroacetic acid-soluble radioactivity in the supernatant was quantitated in a ␥-counter. Specific 125 I-Fn degradation in MEF cultures was determined by subtracting degradation that occurred in the presence of unlabeled Fn and corrected for that which occurred in control wells that lacked cells.
Biosynthetic Labeling of Fn Using [ 35 S]Methionine-MEF-1 and MEF-2 cells were seeded in 6-well plates in serum-containing medium and allowed to adhere for 2 h. Depletion of intracellular methionine was achieved by culturing in L-methionine-free DMEM for 4 h. The cells were then labeled with 35 Ci/ml L-[ 35 S]methionine for 4 h in serumcontaining methionine-free DMEM. Cell extracts were prepared in RIPA buffer (0.1% SDS, 1% deoxycholate, 1% Nonidet P-40, 10 mM sodium phosphate, 150 mM NaCl, 0.4 mM sodium vanadate, 10 g/ml aprotinin, 10 g/ml E64, and 10 g/ml leupeptin). Labeled Fn was recovered by immunoprecipitation. Radioactivity in the precipitates was determined in a scintillation counter. The precipitates were then subjected to SDS-PAGE on 4 -16% gradients. L-[ 35 S]Methionine incorporation into Fn was determined using a Storm system (Molecular Dynamics and Amersham Biosciences).

Fn Accumulation Is Significantly Increased in LRP-deficient
Cells-LRP-deficient MEF-2 cells and LRP-expressing MEF-1 cells were cultured until confluent in serum-containing medium and then dissociated non-enzymatically from monolayer culture. Extracts were prepared from an equal number of cells in suspension and assessed for Fn by immunoblot analysis (Fig.  1A). Fn levels were substantially increased in extracts from the LRP-deficient MEF-2 cells compared with LRP-expressing MEF-1 cells. An identical protocol was followed with LRPdeficient 13-5-1 CHO cells (49) and wild-type CHO-K1 cells. Once again, Fn levels were increased in the LRP-deficient cell line, compared with the control cells. Equivalent results were obtained when the amount of cell extract subjected to SDS-PAGE was standardized based on protein content instead of cell number (results not shown).
Fn forms fibrils at or near the cell surface, which may be variably recovered when cells are dissociated from monolayer culture (5)(6)(7). Furthermore, Fn fibrils may be incompletely solubilized by our standard deoxycholate-containing extraction buffer. To determine whether differential recovery of Fn from MEF-1 and MEF-2 cells affected our results, cell extracts were prepared without first dissociating the cells, using either our standard extraction buffer or 1% SDS (Fig. 1B). Under both sets of conditions, greatly increased levels of Fn were recovered from cultures of MEF-2 cells. When extracts were prepared in SDS, the amount of Fn recovered from the MEF-2 cells was increased by at least 20-fold, compared with that recovered from MEF-1 cells.
Fn Recovered from LRP-deficient Cells Is Produced by the Cells-The Fn detected by immunoblot analysis could have been derived from the cells in culture or the fetal bovine serum that was added to the medium. To distinguish between these possibilities, MEF-1 and MEF-2 cells were allowed to adhere for 2 h in serum-containing medium, which was then replaced by serum-free DMEM containing Nutridoma NS® supplement. Fn that was associated with the cells and in the medium was measured as a function of time. Under these conditions, the only source of Fn is cellular synthesis. As shown in Fig. 2A, cell-associated Fn was almost undetectable shortly after the cells adhered; however, Fn accumulated rapidly in the MEF-2 cell cultures. By 6 h, the level of Fn, which was recovered with MEF-2 cells, was significantly higher than that recovered with MEF-1 cells (p Ͻ 0.01). By 28 h, the difference in Fn recovery was further accentuated (p Ͻ 0.001).
As shown in Fig. 2B, Fn also accumulated to increased levels in conditioned medium from LRP-deficient MEF-2 cells, compared with conditioned medium from MEF-1 cells (p Ͻ 0.01). Thus, LRP deficiency is apparently associated with increased Fn accumulation in association with cells and in solution. To determine whether the differences in Fn accumulation were due to differential rates of Fn synthesis, L-[ 35 S]methionine metabolic labeling experiments were performed. In five separate experiments, we did not observe a significant difference in the rate of Fn synthesis between MEF-1 and MEF-2 cells (results not shown).
Fn Accumulates on the Cell Surface and in the Extracellular Spaces of LRP-deficient Cells-To determine the location of Fn, which was recovered from cultures of LRP-deficient and -ex-pressing cells, we labeled cell-surface proteins, in MEF-1 and MEF-2 cells, with the membrane-impermeable biotinylation reagent, sulfo-NHS-LC-biotin. Biotinylated proteins were selectively recovered from cell extracts by streptavidin affinity precipitation, and Fn was detected in the affinity precipitates by immunoblot analysis. When the cells were labeled with sulfo-NHS-LC-biotin while in monolayer culture, biotinylated Fn was detected almost exclusively with the LRP-deficient MEF-2 cells (Fig. 3). Equivalent results were obtained when the MEFs were dissociated from monolayer culture, using enzyme-free cell dissociation buffer, and then labeled with sulfo-NHS-LC-biotin in suspension. These results indicate that Fn accumulates more significantly in the extracellular spaces surrounding LRP-deficient MEF-2 cells. At least a significant fraction of the Fn is intimately associated with the MEF-2 cell membrane because the Fn partitions with MEF-2 cells that are dissociated from monolayer culture.
Cell-surface biotinylation experiments were also performed using another model system, by comparing LRP-deficient 13-5-1 CHO cells and wild-type CHO-K1 cells (Fig. 4). Once again, the membrane-impermeable biotinylation reagent selectively labeled Fn in the LRP-deficient cell line. Equivalent results were obtained regardless of whether the sulfo-NHS-LC-biotin reacted with cells in monolayer culture or with cells in suspension.
LRP Mediates the Degradation of Fn-LRP undergoes rapid and constitutive endocytosis and recycling, delivering many ligands to lysosomes for degradation (36). We hypothesized that LRP functions as a catabolic receptor for Fn, which might explain the increase in cell-associated Fn in LRP-deficient cells. To test this hypothesis, 125 I-Fn was incubated with MEF-1 and MEF-2 cells at 37°C. Fn degradation was detected

FIG. 1. LRP-deficient cells accumulate increased levels of Fn.
A, MEF-1 cells, MEF-2 cells, CHO-K1 cells, and CHO 13-5-1 cells were plated in serum-containing medium and maintained in culture until confluency was achieved. The cells were then dissociated, and extracts were prepared from 500,000 cells. The extracts were subjected to immunoblot analysis for Fn. Protein mobilities were compared with that of Fn, which was purified from plasma. B, confluent cultures of MEF-1 cells and MEF-2 cells were extracted while in monolayer, either using our standard deoxycholate-containing buffer or 1.0% SDS. The extracts were then subjected to immunoblot analysis to detect Fn. The position of the Fn monomer is indicated by an arrow.

FIG. 2. MEF-synthesized Fn accumulates to increased levels in the absence of LRP. A, MEF-1 cells (ࡗ) and MEF-2 cells (f)
were cultured in serum-free medium for the indicated times. The cells were then dissociated from monolayer culture. Cell extracts were prepared and subjected to immunoblot analysis. B, conditioned medium was recovered from each culture prior to dissociating the cells. The media samples were also subjected to immunoblot analysis. Each immunoblot was analyzed by densitometry. Relative band intensity is plotted (means Ϯ S.E., n ϭ 4). by measuring trichloroacetic acid-soluble radioactivity in the medium. Fig. 5 shows that LRP-expressing MEF-1 cells degraded 125 I-Fn at a significantly increased rate compared with MEF-2 cells (p Ͻ 0.01). 125 I-Fn degradation was specific because a 50-fold molar excess of unlabeled Fn blocked 125 I-Fn degradation by greater than 90%. To confirm that 125 I-Fn degradation by MEF-1 cells resulted from the activity of LRP, we added 200 nM GST-RAP to the cultures together with 125 I-Fn. GST-RAP binds directly to LRP and inhibits the binding of all other known LRP ligands (40 -42). As shown in Fig. 5, GST-RAP significantly inhibited specific 125 I-Fn degradation by MEF-1 cells (p Ͻ 0.01), decreasing the rate to that observed with LRP-deficient cells. By contrast, 125 I-Fn degradation by MEF-2 cells was not altered by GST-RAP, which was an anticipated result because MEF-2 cells do not express LRP. These results demonstrate that LRP mediates the catabolism of Fn and may alter cellular Fn accumulation by this mechanism.
Fn Binds Directly to LRP-Two possibilities were envisioned whereby LRP might promote Fn degradation. The first and most straightforward would involve direct binding of Fn to LRP so that the Fn is internalized by the cell and transferred to lysosomes. The second possibility is that LRP regulates the level of another plasma membrane protein, which in turn functions as a true catabolic receptor for Fn. To test whether Fn binds directly to LRP, we immobilized Fn in microtiter plates. Purified LRP demonstrated concentration-dependent binding to immobilized Fn and not to immobilized BSA in control wells (Fig. 6). GST-RAP blocked the interaction of purified LRP with immobilized Fn, providing evidence that the Fn-LRP interaction is specific. The inhibitory activity of GST-RAP was RAP concentration-dependent.
Blocking LRP Activity Promotes Fn Accumulation in MEF-1 Cells-From the results presented thus far, a model emerges in which LRP binds and internalizes Fn, promoting its degradation in lysosomes. LRP may also regulate levels of other plasma membrane proteins that function in Fn degradation or Fn fibril assembly. To confirm that the difference in Fn accumulation in MEF-1 and MEF-2 cultures was due to LRP, MEF-1 cells were treated with RAP, and MEF-2 cells were transfected to express LRP. If LRP is responsible for the observed differences in cell-associated Fn, then blocking the function of LRP with RAP should promote Fn accumulation. Conversely, stable expression of LRP in MEF-2 cells should inhibit Fn accumulation.
We confirmed that B41 cells, which are MEF-2 cells transfected with a cDNA construct encoding full-length human LRP, express LRP by immunoblot analysis (Fig. 7A). As shown in Fig. 7B, we also demonstrated that B41 cells internalize the LRP ligand, ␣ 2 -macroglobulin, which was purified and converted into its receptor-recognized conformation, as described previously (46). The rate of ␣ 2 -macroglobulin internalization was similar to that observed with heterozygous LRP-deficient All incubations were conducted in the presence or absence of a 50-fold molar excess of unlabeled Fn. At the indicated times, samples of medium were recovered, and trichloroacetic acid-soluble radioactivity in the medium was measured. Specific 125 I-Fn degradation was determined as the difference between trichloroacetic acid-soluble radioactivity that accumulated in the presence and absence of unlabeled Fn. Specific 125 I-Fn degradation (means Ϯ S.E.), obtained from six independent experiments run in triplicate, is shown.
MEFs (PEA-10 cells). These results demonstrate that LRP is functional in B41 cells.
When MEF-1 cells were cultured in the presence of 200 nM GST-RAP for 28 h, the level of cell-associated Fn was significantly increased, compared with cells that were cultured in the presence of 200 nM GST, as a control (Fig. 7C). An increase in Fn accumulation in conditioned medium was also observed (Fig. 7D). The amount of Fn detected in association with RAPtreated MEF-1 cells was comparable with that detected in LRP-deficient MEF-2 cells. These results support our model in which LRP functions to inhibit Fn accumulation on cell surfaces. B41 cells demonstrated substantially decreased Fn accumulation in the cell-associated fraction and in conditioned medium, compared with MEF-2 cells, further supporting the model.

DISCUSSION
The function of LRP and other proteins in the LDL receptor family, as endocytic receptors, is rather complicated at a number of levels. First, the breadth of ligands, which bind to LRP and are subsequently internalized, is extremely large (36). In some cases, LRP may bind complex ligands with potentially profound effects on cell physiology. For example, LRP may internalize growth factors including transforming growth factor-␤ and platelet-derived growth factor-BB that are bound to the primary LRP ligand, ␣ 2 -macroglobulin (50,51). Similarly, LRP mediates the endocytosis of matrix metalloproteinase 2 in association with another primary ligand, thrombospondin 2 (52).
There are a number of examples in which LRP functions as a catabolic receptor for a protein that binds first to another cell-surface site. Lipoprotein lipase and thrombin-protease nexin 1 are most frequently internalized by LRP after binding to cell surface heparan sulfates (53,54), and collagenase 3 probably binds to a distinct cell-surface receptor prior to interacting with LRP (55). Perhaps most interestingly, LRP may regulate plasma membrane levels of other receptors. The most commonly described mechanism involves internalization of multimeric complexes, in which one ligand bridges LRP to another plasma membrane receptor. The ligand and bridged receptor then undergo endocytosis with LRP. This mechanism is operational for uPA-plasminogen activator inhibitor-1 complex, which binds simultaneously to uPAR and LRP or the VLDL receptor, promoting uPAR internalization (45,56). As a result of this process, cell-surface levels of uPAR are typically decreased in LRP-expressing cells (35,57). Similarly, LRP may mediate the internalization of tissue factor via the bridging ligand, factor VIIa-tissue factor pathway inhibitor complex (58).
Fn has the capacity to interact with the cell surface by binding to multiple macromolecules, including integrins and glycosaminoglycans (8 -11); however, pathways that are responsible for Fn clearance from the extracellular spaces remain unclear. Because Fn is a major cell adhesion molecule that regulates cell signaling, pathways that alter cell-surface levels of Fn have the potential to impact on diverse aspects of cell physiology, including cell migration, proliferation, differentiation, and apoptosis. In this study, we identified LRP as a catabolic receptor for Fn. Direct binding of Fn to purified LRP with trypsin to dissociate cell-surface ␣ 2 -macroglobulin. Internalized ␣ 2 -macroglobulin was recovered in 1% SDS and 1.0 M NaOH and quantitated in a ␥-counter. C, MEF-1 cells were cultured in serum-free medium in the presence of GST-RAP (ϩRAP) or an equivalent concentration of GST (Con) for 28 h. MEF-2 cells (Con) and B41 cells (ϩLRP) were cultured under equivalent conditions. Cells were detached from monolayer culture and proteins extracted for immunoblot analysis. D, conditioned medium was isolated from the same cultures and also subjected to immunoblot analysis for Fn. was demonstrated, suggesting that Fn may bind directly to LRP in intact cells. However, it is also possible that Fn binds first to another cell-surface site and then to LRP in a second step. The characterization of Fn as an LRP ligand is intriguing because thrombospondins are the only other previously described ECM proteins that interact with LDL receptor family members (59,60).
In cells that are LRP-deficient, Fn accumulated to greatly increased levels. Our cell-surface biotinylation experiments studies suggest that the Fn is accumulating outside the cell, on the cell surface and in association with the ECM that forms surrounding LRP-deficient cells in culture. The ability of RAP to substantially increase Fn accumulation in MEF-1 cells supports the hypothesis that LRP-mediated Fn catabolism is directly linked to the differences in Fn accumulation observed in LRP-expressing and deficient cells. However, at the present time, other mechanisms cannot be excluded. For example, Aguirre-Ghiso et al. (61) demonstrated that uPAR may associate with ␣ 5 ␤ 1 , promoting the formation of cell-associated Fn fibrils. Because LRP decreases cell-surface levels of uPAR (35,45,57), LRP may indirectly regulate Fn accumulation at the cell surface by this mechanism.
LRP-deficient CHO 13-5-1 cells demonstrated increased Fn accumulation compared with LRP-expressing CHO-K1 cells, thus providing support for our hypothesis regarding the function of LRP in Fn catabolism in a second model system. The function of LRP as a regulator of Fn accumulation in both MEFs and CHO cells suggests that LRP may have an equivalent role in diverse cell types. In addition to LRP, CHO cells express the VLDL receptor, which is partially homologous to LRP in function (62). Ligands that bind to LRP and not to the VLDL receptor include ␣ 2 -macroglobulin and Pseudomonas exotoxin A (46,49). The increase in Fn accumulation in association with CHO 13-5-1 cells may suggest a selective role for LRP in Fn regulation; however, the activity of other LDL receptor homologues in this process should be taken into consideration and remains to be determined.
The relationship between cell-surface Fn and oncogenic transformation and cancer progression is intriguing but remains incompletely understood. Reduced cell-associated Fn levels may be linked to oncogenic transformation (64). However, cell-associated Fn fibrils may suppress the activity of p38 MAPK , leading to an increase in the activity of the MAP kinase, extracellular signal-regulated kinase, and increased cancer cell proliferation (61). Increased extracellular signalregulated kinase activity may also inhibit cancer cell apoptosis (65).
We have demonstrated an association between LRP deficiency and increased cell migration on vitronectin-and fibronectin-coated surfaces. MEF-2 cells migrate more rapidly than MEF-1 cells on vitronectin (35). HT 1080 cells, which express an LRP antisense RNA expression construct and are thus LRP-efficient, migrate more rapidly than LRP-expressing HT 1080 cells (57). Furthermore, MCF-7 breast cancer cells, HT 1080 cells, and normal human fibroblasts that are cultured in RAP for 3-5 days, so that sustained neutralization of LRP or the VLDL receptor is achieved, also migrate more rapidly on vitronectin and fibronectin (45,57). In all of these systems, LRP deficiency or neutralization of an LDL receptor homologue is associated with increased cell-surface uPAR levels and increased accumulation of endogenously produced uPA in conditioned medium. Furthermore, in RAP-treated MCF-7 cells and in antisense RNA-expressing HT 1080 cells, we have directly linked the increase in activity of the uPA/uPAR system to increased migration. However, in smooth muscle cells, LRP neutralization inhibits cell migration (66), suggesting that a uPAR-independent process may be operative. The effects of LRP on Fn accumulation represent a feasible mechanism whereby LRP may regulate cell migration. Similarly, LRP may regulate smooth muscle cell migration based on its ability to function as a receptor for apolipoprotein E (63). Understanding the full impact of Fn regulation by LRP, on cell phenotype, is an important goal for future studies.