Low Density Lipoprotein (LDL) Receptor-related Protein 1B Impairs Urokinase Receptor Regeneration on the Cell Surface and Inhibits Cell Migration*

The low density lipoprotein (LDL) receptor-related protein 1B (LRP1B) is a newly identified member of the LDL receptor family and is closely related to LRP. It was discovered as a putative tumor suppressor and is frequently inactivated in lung cancer cells. In the present study, we used an LRP1B minireceptor (mLRP1B4), which mimics the function and trafficking of LRP1B, to explore the roles of LRP1B on the plasminogen activation system. We found that mLRP1B4 and urokinase plasminogen activator receptor (uPAR) form immunoprecipitable complexes on the cell surface in the presence of complexes of uPA and its inhibitor, plasminogen activator inhibitor type-1 (PAI-1). However, compared with cells expressing the analogous LRP minireceptor (mLRP4), cells expressing mLRP1B4 display a substantially slower rate of uPA·PAI-1 complex internalization. Expression of mLRP1B4, or an mLRP4 mutant deficient in endocytosis, leads to an accumulation of uPAR at the cell surface and increased cell-associated uPA and PAI-1 when compared with cells expressing mLRP4. In addition, we found that expression of mLRP1B or the mLRP4 endocytosis mutant impairs the regeneration of unoccupied uPAR on the cell surface and that this correlates with a diminished rate of cell migration. Taken together, these results demonstrate that LRP1B can function as a negative regulator of uPAR regeneration and cell migration.

The low density lipoprotein (LDL) receptor-related protein 1B (LRP1B) is a newly identified member of the LDL receptor family and is closely related to LRP. It was discovered as a putative tumor suppressor and is frequently inactivated in lung cancer cells. In the present study, we used an LRP1B minireceptor (mLRP1B4), which mimics the function and trafficking of LRP1B, to explore the roles of LRP1B on the plasminogen activation system. We found that mLRP1B4 and urokinase plasminogen activator receptor (uPAR) form immunoprecipitable complexes on the cell surface in the presence of complexes of uPA and its inhibitor, plasminogen activator inhibitor type-1 (PAI-1). However, compared with cells expressing the analogous LRP minireceptor (mLRP4), cells expressing mLRP1B4 display a substantially slower rate of uPA⅐PAI-1 complex internalization. Expression of mLRP1B4, or an mLRP4 mutant deficient in endocytosis, leads to an accumulation of uPAR at the cell surface and increased cell-associated uPA and PAI-1 when compared with cells expressing mLRP4. In addition, we found that expression of mLRP1B or the mLRP4 endocytosis mutant impairs the regeneration of unoccupied uPAR on the cell surface and that this correlates with a diminished rate of cell migration. Taken together, these results demonstrate that LRP1B can function as a negative regulator of uPAR regeneration and cell migration. The plasminogen activation system consists of a cascade of enzymes and plays a central role in many physiological processes requiring the degradation of basement membrane and components of the extracellular matrix. When the regulation of this system is disrupted, as occurs in the pathogenesis of cancer, malignant cells are able to invade surrounding tissue and metastasize to distant body regions (1)(2)(3). Urokinase plasminogen activator (uPA) 1 catalyzes the formation of plasmin from its inactive precursor, plasminogen. The activity of uPA is regulated by two proteins, the glycosylphosphatidylinositollinked uPA receptor (uPAR) and plasminogen activator inhibitor type 1 (PAI-1). The binding of uPA to uPAR at the cell surface greatly increases its catalytic rate. In contrast, uPA is inactivated by binding to PAI-1. When active uPA is bound to uPAR, it is not internalized but remains at the cell surface. However, when receptor-bound uPA is complexed to its inhibitor, PAI-1, the complex is rapidly internalized and degraded. The mechanism by which this occurs was unknown until it was determined that low density lipoprotein receptor-related protein (LRP) is responsible for this process (4). Following the internalization, uPAR and LRP recycle back to the cell surface, while uPA and PAI-1 are degraded in lysosomes (4 -7). The regeneration of unoccupied uPAR at the cell surface is thus critical for the maintenance of plasminogen activation and for regulation of cellular migration and invasion (8 -10).
LRP1B is a recently discovered member of the LDLR family (11,12). The LDLR family previously contained two large members, LRP (LRP1), a dimer of 515-and 85-kDa subunits, and its closely related homolog, megalin (LRP2), a single species of ϳ600 kDa. LRP1B is more closely related to LRP (59% identity at the amino acid level) than to megalin (11). Protein domain structure comparison of LRP and LRP1B reveals that the overall organization of the two proteins is almost identical, except for the fragments encoded by the two extra exons in LRP1B (11,12). Similar to LRP, LRP1B contains four putative ligandbinding domains (I, II, III, and IV from the N terminus) which consist of 2, 8, 10, and 12 cysteine-rich ligand-binding repeats, respectively. The one additional ligand-binding repeat in domain IV of LRP1B is encoded by one of its extra exons. Like LRP, LRP1B contains a furin endopeptidase processing site between its fourth ligand binding domain and transmembrane region. Once cleaved by furin in the Golgi, LRP1B exists as a noncovalently associated heterodimer, consisting of a large extracellular subunit and a smaller transmembrane subunit (12). Similar to that of LRP, the cytoplasmic tail of LRP1B contains five potential endocytosis motifs: two NPXY motifs, two dileucine motifs, and one YXXL motif. However, between the two NPXY motifs, LRP1B contains a unique insertion of a 33-amino acid sequence (encoded by the other extra exon), which is not present in LRP (11,12). We recently demonstrated that the YXXL motif, but not the two NPXY sequences of LRP, serves as the dominant signal for receptor-mediated endocytosis (13).
In a previous study, we demonstrated that LRP1B can bind and internalize single chain urokinase (scuPA) and PAI-1, two components of the plasminogen activation system (12). In the present study, we further explore the role of LRP1B in modulation of the plasminogen activation system. By using an LRP1B minireceptor, which mimics the function and trafficking of LRP1B, we show that the expression of LRP1B impairs the regeneration of unoccupied uPAR on the cell surface and diminishes the rate of cell migration.

EXPERIMENTAL PROCEDURES
Reagents and Antibodies-Recombinant human scuPA, uPA aminoterminal fragment (ATF; amino acids 1-135), and recombinant soluble urokinase receptor (suPAR; amino acids 1-281) were prepared and isolated as described (14 -16). Human two-chain uPA was obtained by limited plasmin digestion of recombinant human scuPA (17). Recombinant human PAI-1, uPA-specific monoclonal antibody 3471, PAI-1specific monoclonal antibody 379, and uPAR-specific monoclonal antibody 3936 were from American Diagnostica. Rabbit polyclonal anti-HA antibody was from Upstate. Monoclonal anti-HA antibody has been described before (13). Human vitronectin and goat anti-mouse IgFITC were from BD Biosciences. Transwell cell culture chambers were from Costar. Carrier-free Na 125 I was purchased from PerkinElmer Life Sciences. Proteins were iodinated using the IODO-GEN method as described previously for receptor-associated protein (RAP) (13).
Internalization of uPA⅐PAI-1 Complexes-125 I-uPA⅐PAI-1 complexes were prepared by the method described (18). Kinetic analysis of endocytosis was performed as described before (13).
Cross-linking Experiments-125 I-labeled, soluble recombinant human uPAR(1-281) (5 nM) was bound to stably transfected CHO cells incubated in serum-free F12/Ham's medium (containing 0.6% bovine serum albumin, 20 mM HEPES, pH 7.4) for 90 min at 4°C in the presence or absence of 5 nM uPA⅐PAI-1 complexes. Cells were then washed, and bound proteins were cross-linked using 1 mM of the crosslinker DTSSP (Pierce) for 10 min at 4°C. The reaction was quenched with 20 mM Tris, pH 7.5, containing 150 mM NaCl for 10 min at 4°C, and cell lysates were prepared in 10 mM CHAPS (Sigma). Immunoprecipitation was carried out with anti-HA antibody, as described before (19), and immunoprecipitated soluble uPAR was analyzed by SDS-PAGE and autoradiography.
Flow Cytometric Analysis of Cell Surface Proteins-Flow cytometric analysis of cell surface uPAR and cell surface-bound uPA and PAI-1 was performed essentially as described (9,13). Briefly, each cell suspension (50 l) was incubated with a saturating concentration of mouse monoclonal antibody at 4°C for 1 h. The cells were then washed and further incubated with saturating concentration of goat anti-mouse IgFITC at 4°C for 45 min. After washing, cells were analyzed by FACS Calibur. Background fluorescence intensity was assessed in the absence of primary antibody.
Analysis of Unoccupied uPAR Regeneration on the Cell Surface-LRP-null CHO cells stably transfected with mLRP4, mLRP4-Mu, or mLRP1B4 were incubated at 37°C for 30 min with 10 nM unlabeled uPA⅐PAI-1, in order to saturate cellular uPAR. Cells were then washed and further incubated in fresh medium at 37°C. At selected time points (0, 0.5, and 2.0 h), cells were chilled to 4°C and washed, and the presence of unoccupied cell surface uPAR was assessed by measuring the binding of 125 I-ATF. Specifically, assay buffer (serum-free medium containing 0.6% bovine serum albumin with 5 nM radioligand, 0.6 ml/well) was added to cell monolayers in the absence or presence of unlabeled 500 nM ATF, followed by incubation for 90 min at 4°C. Thereafter, overlying buffer containing unbound ligand was removed, and cell monolayers were washed and lysed in 0.5 N NaOH and counted.
Cell Migration Assay-Cell migration assays were carried out in 6.5-mm Transwell chambers as described previously, with minor modifications (9). The 8-m pore size Transwell membranes were coated with 2 g/ml vitronectin. Subconfluent CHO stable cells were harvested, washed with serum-free medium (Ham's F-12 medium containing 0.1% bovine serum albumin), resuspended in the same medium, and placed in the upper compartment of the Transwell chambers (5 ϫ 10 4 cells in 100 l). The lower compartment was filled with 400 l of serum-free medium. After incubation for 8 h at 37°C, cells on the lower surface of the filter were fixed and stained, and 8 random fields/filter were counted at ϫ200 magnification.

RESULTS
Internalization Rate of uPA⅐PAI-1 Complexes by mLRP1B4 -The very large size of LRP1B (ϳ600 kDa) limits molecular manipulations at the cDNA level and protein expression via transfection. In a previous study (for details, see Ref. 12), we generated an LRP1B minireceptor that includes the fourth cluster of ligand-binding repeats, transmembrane domain, and cytoplasmic tail (designated mLRP1B4). This LRP1B minireceptor is analogous to the LRP minireceptor mLRP4, which we developed previously to study the trafficking and function of the receptor. We have previously shown (13,19,20) that both the rate of endocytosis and the intracellular trafficking of mLRP4 are identical to those of the full-length receptor. mLRP1B4 includes both of the sequences that are not present within LRP (in domain IV and in the cytoplasmic tail) and therefore allows us to identify the functional differences between these two receptors. As assessed by Western blotting and flow cytometric analysis, mLRP1B stably transfected in LRPnull CHO cells is processed correctly by furin and expressed on the cell surface (12).
LRP1B displays a substantially slower rate of ligand internalization compared with LRP when measured with 125 I-receptor-associated protein (12). Because RAP is an endoplasmic reticulum chaperone (21) and not a physiological ligand for members of the LDLR family, we examined the internalization rates for 125 I-uPA⅐PAI-1 complexes in mLRP1B4-or mLRP4expressing CHO cells. As seen in Fig. 1, the rate and extent of 125 I-uPA⅐PAI-1 internalization is more rapid and more complete in cells expressing mLRP4 than in those expressing mLRP1B4. The mLRP4 endocytosis mutant (with the tyrosinebased and dileucine signals mutated, termed mLRP4-Mu; see Ref. 13), which also displays significantly slower rates of uPA⅐PAI-1 internalization when compared with mLRP4, was used as a control. At 1 and 2 min, mLRP1B4-expressing cells internalized 3 and 6%, respectively, of the total cell-associated 125 I-uPA⅐PAI-1, corresponding to an ϳ90% reduction in ligand endocytosis when compared with mLRP4.
mLRP1B4 Forms Complexes with uPAR in the Presence of uPA⅐PAI-1-Previous studies have shown that uPAR is randomly distributed along the plasma membrane in the absence of uPA⅐PAI-1, whereas the presence of uPA⅐PAI-1 complexes promotes formation of uPAR⅐LRP complexes and initiates re- distribution of occupied uPAR to clathrin-coated pits. uPAR⅐LRP complexes are then endocytosed via clathrin-coated vesicles and traffic together to early endosomes (10). To examine whether uPAR and mLRP1B4 form a molecular complex after binding of uPA⅐PAI-1, we analyzed the potential interaction between uPAR and mLRP1B4 on the plasma membrane by carrying out cross-linking experiments in the presence of recombinant soluble (lacking the glycosylphosphatidylinositolanchor) uPAR (amino acids 1-281). CHO cells stably transfected with mLRP1B4 were incubated with 125 I-uPAR(1-281) (5 nM) for 90 min at 4°C in the presence or absence of 5 nM uPA⅐PAI-1 complexes. Cells were then washed, and bound proteins were cross-linked with a thio-cleavable cross-linker DTSSP, followed by immunoprecipitation with anti-HA antibody. As seen in Fig. 2, in the absence of uPA⅐PAI-1 complexes, soluble uPAR is undetectable in the anti-HA immunoprecipitate. However, in the presence of uPA⅐PAI-1 complexes, the immunoprecipitate contains considerable amounts of 125 I-uPAR(1-281). Similar results were obtained with CHO cells expressing mLRP4 (data not shown). These results indicate that like LRP, LRP1B can form molecular complexes with uPAR at the cell surface in the presence of uPA⅐PAI-1 complexes.
Expression of mLRP1B4 or mLRP4-Mu Results in an Accumulation of Cell Surface uPAR-CHO cells express a relatively high level of cell surface uPAR (22). To examine the potential effects of LRP1B or LRP minireceptor expression on cell surface levels of uPAR, we analyzed cell surface uPAR in LRP-null CHO cells expressing various minireceptors via flow cytometric analyses. Fig. 3 shows that overexpression of mLRP1B4 or mLRP4-Mu results in a 3.5-fold increase in cell surface uPAR when compared with CHO cells stably transfected with pcDNA3 vector alone. In contrast, overexpression of wild-type mLRP4 results in no net change of cell surface uPAR expression. These results suggest that in these CHO cells mLRP1B4 may function in a dominant-negative fashion for uPAR endocytosis. As a result, LRP1B4-bound uPAR is endocytosed at a very slow rate, which in turn results in cell surface accumulation of uPAR.
Expression of mLRP1B4 or mLRP4-Mu Results in an Accumulation of Cell Surface-associated uPA and PAI-1-Accumulation of uPAR at the cell surface may reflect a decreased ability to remove proteolytically inactive uPA⅐PAI-1 complexes from the cell surface to regenerate unoccupied receptor. Thus, we next examined the levels of cell surface-associated uPA and PAI-1 in CHO cells. As seen in Fig. 3, stable expression of wild-type mLRP4 results in a significant decrease in cell surface-associated uPA and PAI-1, whereas stable expression of mLRP1B4 or mLRP4-Mu results in a significant increase in cell surface-associated uPA and PAI-1. These results indicate that stable expression of mLRP4 enhances unoccupied uPAR regeneration at the cell surface, whereas stable expression of mLRP1B4 or mLRP4-Mu has the opposite effect.
Expression of mLRP1B4 or mLRP4-Mu Impairs Regeneration of Unoccupied uPAR on the Cell Surface-To further analyze the role of LRP1B in uPAR trafficking, we assessed the effects of mLRP1B4 expression on the regeneration of unoccupied uPAR at the cell surface. To do this, we analyzed cells that had been saturated with unlabelled uPA⅐PAI-1 complexes. LRP-null CHO cells stably transfected with mLRP4, mLRP1B4, or mLRP4-Mu were incubated for 30 min at 37°C with 10 nM unlabeled uPA⅐PAI-1 complexes to saturate cellular uPAR. Cells were then washed and incubated at 37°C in fresh medium. At selected time points, cells were chilled to 4°C, and the media containing 125 I-ATF (a fragment of uPA that specifically binds to uPAR but not to LRP/LRP1B) was added. The amount of unoccupied cell surface uPAR was determined by measuring the amount of cell-bound 125 I-ATF. As shown in Fig.  4, without additional incubation at 37°C, CHO cells bearing mLRP4, mLRP1B4, and mLRP4-Mu bound equivalent small amounts of 125 I-ATF, as cell surface uPAR was largely occupied by uPA⅐PAI-1 complexes. After incubation at 37°C, progressively more 125 I-ATF was bound to CHO cells expressing mLRP4 (increases of 4.0, 7.5, and 10.0-fold at 0.5, 1, and 2 h, respectively). In contrast, CHO cells expressing mLRP1B4 or mLRP4-Mu showed significantly less change in 125 I-ATF binding. As shown in Fig. 3, steady-state uPAR levels at the cell surface were 3.5-fold greater in mLRP1B4-than mLRP4-expressing cells. However, when uPAR was initially saturated with uPA⅐PAI-1 complexes and allowed to regenerate to the cell surface for 2 h, the level of unoccupied uPAR on mLRP1Bexpressing cells was only one-third that of mLRP4-expressing cells. These results clearly indicate that LRP1B has a significantly reduced ability to regenerate unoccupied uPAR in CHO cells when compared with LRP.
Excess of mLRP1B4 or mLRP4-Mu Impairs Cell Migration-Having established that expression of mLRP1B4 or mLRP4-Mu impairs the regeneration of unoccupied uPAR on the cell surface, we next investigated the effects of mLRP1B4 and mLRP4-Mu on CHO cell migration. To study cellular migration, we utilized Transwell cell culture chambers in which the 8-mm pore membranes were precoated with vitronectin. Interestingly, stable expression of wild-type mLRP4 results in no change in CHO cell migration (Fig. 5), although these cells have lower levels of cell surface-bound uPA and PAI-1 (see Fig.  3). However, stable expression of mLRP1B4 or mLRP4-Mu results in an 80% decrease in CHO cell migration, compared with CHO cells transfected with empty pcDNA3 vector. These results demonstrate that expression of LRP1B results in a marked impairment in CHO cell migration. DISCUSSION The structural similarity between LRP and LRP1B suggests that those two giant receptors may bind similar sets of ligands and display overlapping functions. In support of this, we found that at the cell surface, LRP1B, like LRP, forms immunoprecipitable complexes with uPAR in the presence of uPA⅐PAI-1 complexes and that LRP1B can also mediate the delivery of uPA⅐PAI-1 to the lysosome for degradation (data not shown). On the other hand, the unique sequences in LRP1B suggest that it may also have functions distinct from those of LRP. In this study, we demonstrate that expression of mLRP4 enhances unoccupied uPAR regeneration on the cell surface, whereas expression of mLRP1B4, as well as an mLRP4 endocytosis-defective mutant, causes the accumulation of cell surface uPAR and cell-associated uPA and PAI-1, impairs unoccupied uPAR regeneration on the cell surface, and diminishes cell migration. FIG. 2. Binding of soluble uPAR to mLRP1B4 in the presence of uPA⅐PAI-1 complexes. LRP-null CHO cells stably transfected with mLRP1B4 were incubated with 5 nM 125 I-uPAR(1-281) for 90 min at 4°C in the presence or absence of 5 nM uPA⅐PAI-1 and followed by cross-linking with DTSSP and immunoprecipitation with anti-HA. Immunoprecipitated proteins were processed for autoradiography. mLRP1B4, like the mLRP4 endocytosis mutant, displays a substantially slower rate of uPA⅐PAI-1 internalization when compared with mLRP4. The cytoplasmic tails of LRP1B and LRP contain the same five potential endocytosis motifs, including two NPXY motifs, two dileucine motifs, and one YXXL motif. However, the tail of LRP1B has a unique insertion of 33 amino acid residues that is not present in the tail of LRP (11,12). This insert may be responsible for the slower internalization rate of LRP1B by binding cytosolic adapter proteins and masking the neighboring endocytosis signal(s). This slower endocytosis mediated by LRP1B may allow for more sustained signal transduction upon ligand binding. Consistent with this hypothesis, recent studies have revealed new roles for several members of the LDLR family as obligate components of signal transduction pathways (23). The 33-amino acid insert in the tail of LRP1B may itself, or together with a common motif in the LRP tail, also interact with some cytosolic proteins that participate in signal transduction. Future studies similar to those that have dissected the signals within the cytoplasmic tails of LRP and apolipoprotein E receptor-2 (24, 25) will be conducted to address this issue.
Continuous high level plasminogen activation at the cell surface depends upon the removal of uPA⅐PAI-1 complexes. LRP-mediated internalization of uPA⅐PAI-1⅐uPAR complexes results in the lysosomal degradation of uPA and PAI-1 and recycling of uPAR to the cell surface (4 -7). This process allows for the regeneration of unoccupied uPAR, immobilization of fresh uPA in the form of its zymogen, pro-uPA, and the eventual re-expression of active uPA at the cell surface (8,10). Failure to remove uPA⅐PAI-1complexes from the cell surface may diminish the cellular capacity for plasminogen activation and may impair cell migration/invasion mediated by uPAR (8 -10, 26). CHO cells bearing mLRP1B4 display approximately a 90% reduction in ligand endocytosis when compared with CHO cells expressing mLRP4. The physiological significance of this may be that LRP1B acts in a dominant-negative fashion to limit the effectiveness of LRP-or caveolae-mediated uPA⅐PAI-1⅐uPAR clearance. If cells co-express LRP and LRP1B, the net result of uPAR regeneration likely depends on the relative expression of the two receptors. For example, if the expression level of LRP1B in a given cell type exceeds that of LRP, LRP1B may function as a dominant-negative regulator for the regeneration of unoccupied uPAR at the cell surface and therefore inhibit uPAR-mediated cell migration. A model depicting this hypothesis is shown in Fig. 6. Consistent with this hypothesis, we have demonstrated that expression of mLRP1B4 in CHO cells impairs the regeneration of unoccupied uPAR on the cell surface and reduces cell migration. Although our previous studies with LRP minireceptors have clearly demonstrated that these receptors traffic and function similarly to their endogenous full-length receptor (13,20), it is important to realize that the use of minireceptors may abolish or diminish the functional differences between full-length LRP and LRP1B. Future studies with endogenous LRP1B should help us to fully understand the biological functions of this receptor.
In addition to regulating plasminogen activation at the cell surface, recent studies indicate that uPAR possesses other cellular activities, including an ability to generate intracellular signal transduction, to regulate integrin function, and to directly bind vitronectin (27)(28)(29). Thus, it is possible that LRP1B expression also mediates uPAR interaction with integrin subunits and uPAR/integrin-mediated signaling.
Two lines of evidence strongly support an important and apparently causal role for the plasminogen activation system in cancer metastasis: the results from experimental model systems with animal tumor metastasis and the finding that high levels of uPA, PAI-1, and uPAR in many tumor types predict poor patient prognoses (1). The discovery that LRP mediates the internalization of uPA⅐PAI-1⅐uPAR complexes has generated interest in the role of LRP in malignant transformed cells. Although the role of LRP in tumor progression remains unclear at present, several studies show a correlation between LRP expression and human cancer cell invasion and metastasis (9, 30 -36). LRP1B was originally discovered as a putative tumor suppressor gene and is frequently inactivated in non-small cell lung cancer cell lines (11) and in high grade of urothelial cancer (37). Therefore, our present results warrant further studies on the potential role of LRP1B as a tumor suppressor and in inhibition of tumor metastasis.
In summary, using an LRP1B minireceptor, we have demonstrated that expression of LRP1B impairs unoccupied uPAR regeneration on the cell surface and diminishes cell migration. These results suggest that LRP1B mimics an LRP endocytosis mutant and negatively regulates uPAR regeneration and function.