Heparanase Uptake Is Mediated by Cell Membrane Heparan Sulfate Proteoglycans*

Heparanase is a mammalian endoglycosidase that degrades heparan sulfate (HS) at specific intrachain sites, an activity that is strongly implicated in cell dissemination associated with metastasis and inflammation. In addition to its structural role in extracellular matrix assembly and integrity, HS sequesters a multitude of polypeptides that reside in the extracellular matrix as a reservoir. A variety of growth factors, cytokines, chemokines, and enzymes can be released by heparanase activity and profoundly affect cell and tissue function. Thus, heparanase bioavailability, accessibility, and activity should be kept tightly regulated. We provide evidence that HS is not only a substrate for, but also a regulator of, heparanase. Addition of heparin or xylosides to cell cultures resulted in a pronounced accumulation of, heparanase in the culture medium, whereas sodium chlorate had no such effect. Moreover, cellular uptake of heparanase was markedly reduced in HS-deficient CHO-745 mutant cells, heparan sulfate proteoglycan-deficient HT-29 colon cancer cells, and heparinase-treated cells. We also studied the heparanase biosynthetic route and found that the half-life of the active enzyme is ∼30 h. This and previous localization studies suggest that heparanase resides in the endosomal/lysosomal compartment for a relatively long period of time and is likely to play a role in the normal turnover of HS. Co-localization studies and cell fractionation following heparanase addition have identified syndecan family members as candidate molecules responsible for heparanase uptake, providing an efficient mechanism that limits extracellular accumulation and function of heparanase.

Heparanase is a mammalian endo-␤-D-glucuronidase that cleaves heparan sulfate (HS) 1 side chains at a limited number of sites (1)(2)(3). Such enzymatic activity is thought to participate in degradation and remodeling of the extracellular matrix and to facilitate cell invasion associated with cancer metastasis and inflammation (1, 4 -6). Under normal conditions, heparanase activity is restricted to the placenta and skin tissues and to blood-borne cells such as platelets, neutrophils, monocytes, mast cells, and T lymphocytes (1,4,5,(7)(8)(9)(10)(11)(12). In these cells, heparanase is thought to be stored in specific granules, and its release by degranulation has been implicated in diapedesis and extravasation of a number of immune cells (4,5,(11)(12)(13). Heparanase up-regulation has been documented in a variety of human tumors correlating, in some cases, with increased vascular density and poor postoperative survival (14 -17). Heparanase overexpression has also been noted in several other pathologies such as cirrhosis (18), nephrosis (19), and diabetes (20). In addition to its intimate involvement in the egress of cells from the blood stream, heparanase activity may release a multitude of HS-bound, extracellular matrix-resident growth factors, cytokines, chemokines, and enzymes that might profoundly affect cell and tissue function (1,21). Thus, heparanase activity and bioavailability should be kept tightly regulated. Mechanisms that dictate heparanase regulation are only poorly understood, but are expected to operate at several distinct levels. Induced heparanase expression under pathological conditions suggests a transcriptional regulation. Heparanase gene expression has been shown to involve promoter methylation (22), eukaryotic initiation factor 4E (23), and the Ets (24) and Egr1 (25) transcription factors. Recently, estrogen has been shown to induce heparanase promoter activation in estrogen receptor-positive breast cancer cells (26). Regulation at the post-translational level, viz. heparanase processing, cellular localization, and secretion, has also been implicated as a major regulatory mechanism (27)(28)(29)(30).
We have previously shown that exogenously added heparanase rapidly interacts with primary human fibroblasts (31) as well as with tumor-derived cells (32), followed by internalization and processing into a highly active enzyme, collectively defined as heparanase uptake. The role of heparan sulfate proteoglycans (HSPGs) in heparanase uptake has not been studied in detail. Here, we provide evidence that the addition of heparin to cell cultures results in a pronounced accumulation of heparanase in the culture medium, whereas treatment with sodium chlorate had no such effect. Moreover, cellular uptake of heparanase was markedly reduced in HSdeficient CHO-745 cells, HSPG-deficient HT-29 cells, and heparinase-treated cells. We also studied the heparanase biosynthetic route and estimated the half-life of the active enzyme to be ϳ30 h. Co-localization and cell fractionation studies following heparanase addition have identified syndecans, rather than glypicans, as candidate HSPGs responsible for heparanase uptake.
Cell Culture and Transfection-Human U87-MG glioma, MDA-MB-435, and MDA-MB-231 breast carcinoma cells; HT-29 colon carcinoma cells; and CHO-K1 cells were purchased from American Type Culture Collection. Cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and antibiotics. Mutant Chinese hamster ovary (CHO) cells (pgsA-745) deficient in xylosyltransferase and unable to initiate glycosaminoglycan synthesis were kindly provided by Dr. J. D. Esko (University of California, San Diego, CA) and grown in RPMI 1640 medium supplemented with 10% fetal calf serum and antibiotics. CHO cells deficient in glycosylphosphatidylinositol (GPI)-anchored proteins and thus lacking cell-surface glypicans were kindly provided by Dr. F. Gisou van der Goot (Department of Biochemistry, University of Geneva, Switzerland) (35). Human umbilical vein endothelial cells were grown in gelatin-coated flasks essentially as described (36). HEK-293 cells stably transfected with the human heparanase cDNA cloned into the pSecTag vector were kindly provided by Dr. H.-Q. Miao (ImClone Systems Ltd., New York, NY), and rat C6 glioma cells were kindly provided by Dr. Eli Keshet (Hebrew University Hadassah Medical School). U87 and C6 glioma cells stably transfected with the chimeric hpa gene construct were described previously (28,33). For stable transfection, subconfluent MDA-435, MDA-231, CHO-K1, and CHO-745 cells were transfected with the pSecTag2 vector (Invitrogen) containing the full-length heparanase cDNA using the FuGENE 6 reagent (Roche Applied Science) according to the manufacturer's instructions. The pSecTag2 vector is designed for efficient protein secretion driven by the IgG signal peptide and contains Myc and His tags at the protein C terminus. Transfection proceeded for 48 h, followed by selection with Zeocin for 2 weeks. Stable transfectant pools were further expanded and analyzed.
Immunoblotting, Metabolic Labeling, and Immunoprecipitation-Cells were incubated with the indicated concentrations of heparin or chlorate for 24 h under serum-free conditions. The medium was collected, and following several washes, cell extracts were prepared using lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.5% Triton X-100 supplemented with a mixture of protease inhibitors (Roche Applied Science). Protein concentration was determined using the Bradford reagent (Bio-Rad), and 30 g of protein and the equivalent volume of medium were resolved by SDS-PAGE under reducing conditions. After electrophoresis, proteins were transferred to polyvinylidene difluoride membrane (Bio-Rad) and probed with the appropriate antibody, followed by horseradish peroxidase-conjugated secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) and an enhanced chemiluminescent substrate (Pierce) (30).
Metabolic labeling was performed essentially as described (32). Briefly, confluent cell cultures were methionine-starved for 30 min prior to the addition of 150 Ci/ml [ 35 S]methionine (Amersham Biosciences). Cells were pulsed for 20 min and chased for the indicated time periods in 1 ml of complete growth medium containing excess unlabeled methionine and the indicated concentrations of heparin or chlorate. For immunoprecipitation, equal volumes (0.1 ml) or equal trichloroacetic acid-precipitable cpm of cell lysates and medium samples were brought to a volume of 1 ml with 50 mM Tris-HCl (pH 7.4), 5 mM EDTA, 150 mM NaCl, and 0.5% Nonidet P-40 (buffer A) and incubated with antiheparanase monoclonal antibody 130 or anti-Myc monoclonal antibody for 2 h at 4°C. Protein A/G-Sepharose beads (Santa Cruz Biotechnology) were then added for an additional 30 min. Beads were collected by centrifugation and washed three times with buffer A supplemented with 300 mM NaCl and 5% sucrose and once with buffer A. Sample buffer was then added, and after boiling at 100°C for 5 min, samples were subjected to SDS-PAGE as described above. Gels were fixed for 30 min in 25% isopropyl alcohol and 10% acetic acid and fluorographed for 30 min in Amplify (Amersham Biosciences) before drying and autoradiography.
Cell Fractionation-Isolation of plasma membrane and endosomal/ lysosomal cell fractions was carried out essentially as described (37). Briefly, human umbilical vein endothelial cells (3 ϫ 10 8 ) were harvested with trypsin/EDTA, washed twice with phosphate-buffered saline (PBS), suspended in 2 ml of extraction buffer (10 mM Tris acetic acid (pH 7.0) supplemented with 250 mM sucrose), and homogenized in a 5-ml Potter-Elvehjem homogenizer. The cell homogenate was then centrifuged at 2000 ϫ g for 2 min to pellet cell debris. The supernatant was collected and centrifuged at 4000 ϫ g for 2 min to pellet a fraction enriched with plasma membranes. The supernatant was centrifuged at 100,000 ϫ g for 10 min to sediment endosomes and lysosomes. Pellets were resuspended in lysis buffer containing 1% Triton X-100, and 0.5% deoxycholate-containing lysis buffer and equal amounts of protein were analyzed by immunoblotting as described above.
Heparanase Purification and Uptake Studies-The latent Myctagged 65-kDa heparanase precursor was purified from the culture medium of heparanase-transfected HEK-293 cells essentially as described (33). For uptake studies, the 65-kDa heparanase precursor was added to confluent cell cultures at a concentration of 1 g/ml under serum-free conditions. At the indicated time points, the medium was aspirated; cells were washed twice with ice-cold PBS; and total cell lysates were prepared as described above. Heparanase uptake and processing were analyzed by immunoblotting with anti-heparanase (1453) or anti-Myc tag antibody.
Immunocytochemistry-Human U87 glioma cells were left untreated or were incubated with exogenously added 65-kDa heparanase (10 g/ml) for 15 min and subjected to indirect immunofluorescence staining essentially as described (32,33). Briefly, cells were fixed with cold methanol for 10 min. Cells were then washed with PBS and incubated in PBS containing 10% normal goat serum for 1 h at room temperature, followed by a 2-h incubation with the indicated primary antibodies. Cells were washed extensively with PBS, incubated with the relevant (Cy2/Cy3-conjugated) secondary antibody (Jackson ImmunoResearch Laboratories, Inc.) for 1 h, washed, and mounted (Vectashield, Vector Laboratories, Burlingame, CA).

Heparin, but Not Sodium Chlorate, Enhances Heparanase
Accumulation in Cell Culture Medium-To explore the role of cell membrane HS in heparanase uptake, heparanase-transfected cell lines were incubated with increasing concentrations of heparin, and accumulation of the enzyme in the culture medium was evaluated by immunoblotting ( Fig. 1 In contrast with 293 cells, heparanase was not detected in the culture medium of untreated MDA-435, C6, and U87 cells, and heparin at concentrations of 5 g/ml or higher was required for heparanase accumulation in the culture medium. Heparin had no effect on the levels of the 65-kDa heparanase in the cell lysates ( Fig. 1, A-D, upper panels, Lysate). These results suggest that heparanase, secreted by each of the cell lines, rapidly interacts with membranous HSPGs, which is presumably followed by uptake and internalization (see below) and is competed by the addition of heparin.
Interaction of a multitude of molecules with HS is determined by the sequence and sulfation level of the sugar moieties (38,39). Sulfation can be partially inhibited by sodium chlorate (40 -42). Interestingly, pretreatment with sodium chlorate at concentrations up to 50 mM had no effect on heparanase accumulation in the culture medium of all cell lines examined ( Fig.  1, A-D, lower panels). Heparanase could not be detected in the culture medium of the other cell lines, unlike 293 cells, regardless of chlorate pretreatment. The effect of heparin and chlorate on extracellular heparanase accumulation was next eval-uated by metabolic labeling and immunoprecipitation analysis. C6 (Fig. 1E, upper panel) and 293 (lower panel) cells were pulsed for 20 min with [ 35 S]methionine, followed by a 2-h incubation without or with heparin (1-5 g/ml), chlorate (50 mM), or both. Medium samples were then collected and subjected to immunoprecipitation with anti-heparanase (Fig. 1E, upper panel) or anti-Myc (lower panel) antibody. Heparanase was readily detected in the medium of untreated 293 cells, and the addition of heparin resulted in further accumulation of the de novo synthesized heparanase. Chlorate treatment had no effect on heparanase accumulation, and the combination of heparin and chlorate was as effective as heparin alone (Fig. 1E, lower panel). Similar qualitative results were obtained with C6 glioma cells (Fig. 1E, upper panel), in agreement with the immunoblotting results (Fig. 1C). Chlorate at a concentration of 50 mM has been shown to inhibit overall O-sulfation by 70%, whereas N-sulfation remains unchanged in Madin-Darby canine kidney cells (40). This suggests that O-sulfation is not involved in HS-heparanase interaction, yet N-sulfation may still be necessary. To test this possibility, heparanase-transfected 293 cells were incubated with chemically modified heparin that was totally N-desulfated, followed by N-acetylation. Modified heparin lacking 2-O-sulfation was also examined for its ability to bind heparanase. Indeed, N-desulfated heparin lost its ability to bind heparanase as revealed by the low levels of heparanase accumulated in the culture medium (Fig. 1F, N-Ac). In contrast, 2-O-desulfated heparin was as efficient as unmodified heparin, further suggesting that N-sulfation is necessary for the interaction of HS with heparanase. Moreover, the addition of 4-methylumbelliferyl 7-␤-D-xyloside, a xyloside that substitutes for the linker moiety to the proteoglycan core protein and thus functions as a soluble primer for glycosaminoglycan biosynthesis, resulted in accumulation of heparanase in the culture medium of 293 cells (Fig. 1G), mimicking the effect of heparin.
Reduced Heparanase Uptake by HS-deficient Cells-Next, we examined the uptake of heparanase by HS-deficient cells.
To this end, the Myc-tagged 65-kDa heparanase precursor was exogenously added to the culture medium, and uptake was analyzed by immunoblotting. Heparanase binding to wild-type CHO-K1 cells was rapid and appeared maximal by 15 min, followed by a gradual decrease ( Fig. 2A, left panels, K1). Interestingly, the anti-Myc tag antibody detected only a single protein band that corresponded in its molecular mass to the added 65-kDa heparanase precursor ( Fig. 2A, upper left panel, K1). Heparanase processing into the 50-kDa active enzyme was clearly detected by the anti-heparanase 1453 antibody ( Fig. 2A,  middle left panel, K1), suggesting that the heparanase C terminus was subjected to processing that removed the Myc tag sequence. Heparanase processing at the C terminus has not been reported so far. A marked decrease in heparanase binding and uptake was observed upon its addition to HS-deficient CHO-745 cells (Fig. 2A, upper and middle left panels, 745). The requirement of HS for efficient heparanase uptake was confirmed using the human HT-29 colon carcinoma cell line (also known as WiDr), reported to express perlecan, a secreted HSPG found mainly in the extracellular matrix, but no other HSPG (42). Heparanase uptake by the HT-29 cells was markedly attenuated (Fig. 2A, upper and middle right panels, HT29) and was similar in magnitude to that by the CHO-745 cells, further confirming the involvement of cell-surface HS in heparanase uptake. Similarly, pretreatment of CHO-K1 cells with bacterial heparinase I resulted in reduced heparanase binding and uptake (Fig. 2B). In contrast with the rapid heparanase uptake observed with CHO-K1 cells at 37°C (Fig.  2, A, K1; and B, upper left panel, Control), binding at 4°C was time-dependent and appeared maximal by 60 min (Fig. 2B, middle left panel, Control). Under these conditions, heparanase binding (4°C) and uptake (37°C) were barely detected in the heparinase-treated cells (Fig. 2B, upper and middle left panels, Heparinase I). In contrast, pretreatment with chondroitinase ABC had no effect on heparanase binding and uptake (Fig. 2B, right panels, ABC), supporting the specificity of the HS-heparanase interaction. The pronounced decrease in heparanase uptake by HS-deficient cells ( Fig. 2A, left panels, 745; and right panels, HT29) rationalizes that heparanase expression by these cells would result in reduced uptake and increased accumulation of the enzyme in the culture medium. Indeed, large amounts of the 65-kDa heparanase precursor were found in the culture medium of heparanase-transfected CHO-745 cells (Fig.  2C, upper panel, 745). In contrast, heparanase was not detected in the culture medium of wild-type CHO-K1 cells (Fig. 2C,  upper panel, K1), unless heparin was supplemented (lower panel). Accumulation of heparanase in the culture medium of transfected CHO-745 cells correlated with reduced amounts of the 50-kDa active enzyme in the cell lysates (Fig. 2C, middle  panel). These results imply that the 50-kDa heparanase present in cell lysates originated from the internalized 65-kDa heparanase precursor.
Heparanase Biosynthetic Route and Half-life-To further study the ratio between extracellular (65 kDa, latent) and intracellular (50 kDa, active) heparanase forms, we utilized metabolic labeling and immunoprecipitation analysis to follow heparanase biosynthesis. Heparanase-transfected 293 cells, which exhibited the highest enzyme expression and secretion levels (Fig. 1A), were pulsed for 20 min with [ 35 S]methionine and chased for the indicated time periods with complete growth medium supplemented with unlabeled methionine. At each time point, the medium and lysate samples were subjected to immunoprecipitation with anti-heparanase monoclonal antibody 130 (lysate) or anti-Myc monoclonal antibody (medium) and autoradiography. As shown in Fig. 3A (lower panel, To), the newly synthesized heparanase first appeared as a single protein band that corresponded in its molecular mass to the 65-kDa heparanase precursor. The amount of the 65-kDa protein band found in the cell lysates rapidly decreased thereafter and practically disappeared after 2 h of chase, when a 65-kDa protein band was first detected and started to accumulate in the culture medium (Fig. 3A, upper panel). Heparanase accumulation in the culture medium peaked at 4 h of chase, followed by a gradual decrease. Interestingly, the 50-kDa active heparanase appeared first in lysate samples after 4 h of chase (Fig. 3A, lower panel), correlating with maximal heparanase accumulation in the culture medium. This result supports the notion that intracellular active heparanase originates from uptake and internalization of the extracellular heparanase precursor. Densitometry analysis of the kinetics of heparanase biosynthesis yielded an estimated half-life for the 50-kDa active heparanase of ϳ30 h (Fig. 3B). Next, we investigated the half-life of the 50-kDa enzyme following uptake. To this end, CHO cells were incubated with the 65-kDa heparanase precursor for 4 h, washed, and chased for an additional 24 or 48 h with complete growth medium. The amount of the 50-kDa heparanase form was determined by immunoblotting and densitometry (Fig. 3, C and D). Prominent levels of the 50-kDa heparanase were observed following 4 h of uptake (Fig. 3C, 4), in agreement with the previous CHO studies presented in Fig. 2A  (middle left panel, K1). Densitometry analysis revealed 30 and 80% decreases in the levels of the 50-kDa heparanase protein following 24 and 48 h of chase, respectively (Fig. 3D), and the half-life of the exogenously added heparanase was calculated to be ϳ30 h. Thus, endogenously expressed heparanase (Fig. 3, A  and B) and exogenously added heparanase (Fig. 3, C and D) exhibited similar half-lives, suggesting targeting to the same cellular compartment, most likely late endosomes and lysosomes (32,43).
If indeed the intracellular active heparanase originates from uptake and internalization of the 65-kDa heparanase precursor, prevention of uptake and accumulation of the 65-kDa enzyme in the culture medium should result in decreased intracellular levels of the 50-kDa heparanase. Addition of heparin to the culture medium of heparanase-transfected MDA-435 cells resulted in a typical accumulation of the 65-kDa heparanase precursor in the cell culture medium (Figs. 1B and Fig. 3E, upper panel), correlating with a proportional decrease in the intracellular levels of the 50-kDa active heparanase (Fig. 3E,  middle panel). Densitometry analysis revealed a 60 -70% decrease in the intracellular pool of the 50-kDa heparanase in response to heparin treatment (Fig. 3F). This decrease was further confirmed by metabolic labeling analysis (data not shown).
Syndecans as Mediators of Heparanase Uptake-We searched for HSPGs that interact with heparanase and mediate its uptake. Syndecans are transmembrane HSPGs implicated in the uptake of several classes of ligands, including lipoproteins and pathogens (44 -47). As expected, syndecan-1 and syndecan-4 staining was largely restricted to the plasma membrane of human U87 glioma cells (Fig. 4A, first panels,  red). Interestingly, however, shortly after heparanase addition, syndecan-1 and syndecan-4 were found to be localized mainly in endocytic vesicles (Fig. 4A, second panels), colocalizing, to a large extent, with heparanase (fourth panels). This marked change in syndecan localization suggests that syndecan family members and heparanase are subjected to endocytosis and internalization as a complex, resulting in co-localization. Moreover, such rapid internalization would transiently reduce syndecan membrane localization and thus may modify cell behavior. Redistribution of syndecan upon heparanase addition was further examined biochemically by cell fractionation. Since the expression levels of syndecans and glypicans are often altered in tumor-derived cells, we chose primary endothelial cells (human umbilical vein endothelial cells) for the fractionation studies. Heparanase was noted to be highly abundant in the endosomal/lysosomal (E/L) fraction 40 min after its addition (Fig. 4B, upper panel), in agreement with the immunofluorescence staining (Fig. 4A)  (31, 32, 43). Moreover, a shift of syndecan-1 from the plasma membrane (M) into the endosomal/lysosomal compartment FIG. 4. Altered syndecan localization in response to heparanase addition. A, immunostaining. U87 glioma cells were left untreated (ϪHepa) or were incubated with the latent heparanase (ϩHepa; 10 g/ml) for 15 min. Cells were then fixed and stained with anti-syndecan-1 (left, first and second panels) or anti-syndecan-4 (right, first and second panels) monoclonal antibody or with anti-Myc polyclonal antibody (Hepa; third panels). Merged images are shown in the fourth panels. Note internalization of syndecan (Syn) into endocytic vesicles upon heparanase addition. B, cell fractionation. Human umbilical vein endothelial cells were left untreated (control (Con)) or were incubated with heparanase (Hepa; 1 g/ml) for 40 min. Cells were then washed and subjected to cell fractionation as described under "Materials and Methods." Membrane (M) and endosomal/lysosomal (E/L) fractions, representing equal amounts of proteins (30 g), were analyzed with anti-Myc (upper panel), anti-syndecan-1 (Syn 1; middle panel), or anti-glypican-3 (Gly 3; third panel) antibody by immunoblotting. Note the accumulation of syndecan-1, but not glypican-3, in the endosomal/lysosomal fraction upon heparanase addition. C, heparanase uptake by GPI-deficient cells. CHO cells deficient in GPI-anchored proteins (⌬GPI) or control (Con) cells were left untreated (0) or were incubated with the latent 65-kDa heparanase precursor (1 g/ml). At the indicated time points, cells were washed, and heparanase uptake was evaluated by subjecting total cell lysates to immunoblotting with anti-heparanase antibody 1453. upon heparanase addition was observed (Fig. 4B, middle  panel), supporting the immunostaining results (Fig. 4A). Interestingly, no significant change in glypican-3 distribution was observed upon heparanase addition (Fig. 4B, lower  panel), suggesting that syndecans, rather than glypicans, mediate heparanase uptake. The role of glypicans in heparanase regulation was further evaluated by applying GPI-deficient cells that lack glypicans on their cell surface (35). Heparanase uptake by these cells appeared unchanged compared with the control cells (Fig. 4C), supporting the notion that glypicans, unlike syndecans, are not critically important for heparanase uptake. DISCUSSION Virtually all cells express at least one HS-bound core protein (HSPG) on their surface. From mice to worms, embryos that lack HS die during gastrulation (48), positioning HSPGs as critical regulators of cell-cell signaling during embryogenesis. Such critical function is not restricted to developmental processes. HSPGs are thought to play key roles in numerous biological settings, including cytoskeleton organization, cell migration, wound healing, inflammation, cancer metastasis, and angiogenesis (48 -51). HSPGs exert their multiple functions via several distinct mechanisms, combining biochemical, structural, and regulatory aspects. Our results indicate that HS is not only the substrate for, but also a regulator of heparanase. HS is well known for its ability to assemble ligands and receptors into ternary signaling complexes, best exemplified by the fibroblast growth factor-fibroblast growth factor receptor-heparin complex (52). The multitude of polypeptides sequestered and regulated by HS (53) and the ability of heparanase to convert these into bioavailable molecules (1,21) require that these activities will be kept tightly regulated. Several lines of evidence indicate that HS mediates cellular uptake of heparanase and thus regulates its extracellular retention. Addition of heparin to cell cultures resulted in extracellular accumulation of heparanase in essentially all cell lines examined and was most pronounced in cells that exhibit low heparanase secretion levels (i.e. MDA-435, C6, U87, and CHO). This consistent behavior suggests that heparanase is being secreted but does not normally accumulate extracellularly unless the cell membrane HS is removed (i.e. by heparinase treatment) (Fig. 2B) or competed with heparin or soluble xyloside primers (Fig. 1). In contrast, chlorate treatment had no such effect on extracellular heparanase accumulation (Fig. 1), suggesting that N-rather than O-sulfate groups mediate the interaction of HS with heparanase, as also indicated by the inability of N-acetylated heparin to promote extracellular accumulation of heparanase (Fig. 1F). Such a sulfation pattern is different from that shown to mediate the uptake of atherogenic lipoproteins, which is significantly inhibited by chlorate treatment (54 -56), arguing for sulfation pattern specificity of HS-mediated cellular uptake of ligands. In line with the effect of heparin, heparanase uptake was markedly inhibited in HS-deficient CHO-745 cells and HSPG-deficient HT-29 cells (Fig. 2A). In contrast with the CHO-745 cells, which bear defective xylosyltransferase responsible for assembly of the proteoglycan saccharide linkage region, resulting in cell-surface HS deficiency, HT-29 cells synthesize perlecan, a proteoglycan that is primarily secreted and assembled in the extracellular matrix. However, perlecan can also be found on the cell surface, associated with integrin molecules, and thereby may mediate uptake of specific ligands (42). Heparanase uptake by CHO-745 and HT-29 cells was marginal and appeared similar in terms of kinetics and magnitude ( Fig. 2A), indicating that perlecan does not play a significant role in this process. This observation implies that the core protein together with its HS side chains determine the specificity of HSPG interaction with various protein ligands. Analysis of the localization of syndecans and glypicans following heparanase addition further supports this notion. Heparanase was noted to accumulate in perinuclear vesicles already 15 min following its addition to U87 cells (Fig. 4), in agreement with previous reports (31,32). Interestingly, both syndecan-1 and syndecan-4 were shown to accumulate in endocytic vesicles shortly after heparanase addition, co-localizing with heparanase (Fig. 4A), a redistribution that was further confirmed biochemically (Fig. 4B). This observation suggests that heparanase and syndecans are internalized as a complex and thus co-localize and that both syndecan-1 and syndecan-4 participate in heparanase uptake. Such a rapid and efficient process would ultimately reduce the amounts of syndecans on the cell surface, altering the ability of the cells to respond to extracellular cues (57). In contrast, glypican-3 was not subjected to such redistribution upon heparanase addition (Fig. 4B), and heparanase uptake by GPI-deficient cells that lack glypicans on their surface was as efficient as that by control cells (Fig.  4C). These results suggest that syndecans rather than glypicans are the major mediators of heparanase uptake and regulation. It is important to note that although we have utilized the heparanase precursor in our uptake studies, the active enzyme exhibits an even higher affinity for HS (31) and is therefore likely to be similarly regulated by HS.
Activation of the latent 65-kDa heparanase precursor involves proteolytic cleavage at two potential sites located at the N terminus of the molecule (Glu 109 -Ser 110 and Gln 157 -Lys 158) , resulting in the formation of two subunits that heterodimerize and form the active heparanase enzyme (27,29,30). Interestingly, when the uptake process was studied with the anti-Myc tag antibody, we consistently observed a single protein band that corresponded in its molecular mass to the added 65-kDa heparanase precursor. The 50-kDa protein was not detected by the anti-Myc antibody even at later time points, when heparanase processing was evident, suggesting that the Myc tag was removed. Such processing at the protein C terminus has not been recognized so far. Moreover, processing at the C terminus preceded the N-terminal processing (36), raising the possibility that cleavage at the C terminus is a prerequisite for further processing events. Studies exploring this possibility are currently under way.
The half-life of heparanase has not been elucidated so far. By employing metabolic labeling and immunoprecipitation analyses, we followed heparanase biosynthesis, secretion, and activation. A single 65-kDa protein band appeared following 20 min of pulse and practically disappeared after 2 h of chase. This rapid decrease correlated with the appearance of a Myc-tagged 65-kDa protein in the culture medium (Fig. 3A). The secreted protein continued to accumulate by 4 h of chase, followed by a gradual decrease. Several lines of evidence suggest that the intracellular active heparanase originates from uptake, internalization, and processing of the extracellular 65-kDa heparanase precursor. The appearance of a 50-kDa active heparanase in the cell lysates correlated with accumulation of de novo synthesized 65-kDa heparanase in the culture medium (Fig. 3). Direct intracellular processing of the 65-kDa precursor into its active form would have been expected sooner, parallel to the decrease in the 65-kDa band. Furthermore, overexpression of heparanase by HS-deficient CHO-745 cells resulted in high levels of heparanase in the culture medium, correlating with reduced amounts of the intracellular 50-kDa enzyme. Similarly, heparanase accumulation in the culture medium upon addition of heparin was accompanied by a comparable decrease in the amount of the intracellular processed enzyme. Collectively, these findings suggest that the intracellular active heparanase originates, at least in part, from uptake of the extracellular heparanase precursor, possibly reflecting the in vivo trafficking route.
The half-life of the newly formed processed heparanase exceeded 24 h and was estimated to be ϳ30 h. A similar value was calculated for the half-life of the 50-kDa subunit generated upon processing of exogenously added latent heparanase. Thus, endogenously secreted or exogenously added latent 65-kDa heparanase appears to follow similar routes, resulting in generation of the active enzyme, which, in turn, exhibits a relatively long half-life. The attended stability of the processed active heparanase is of special significance given its lysosomal localization (32,43) and stands in contrast with the relatively short half-life (2-6 h) of HSPGs with a transmembrane domain or the even shorter t1 ⁄2 (ϳ25 min) for GPI-anchored HSPGs (58). These studies suggest that heparanase may normally function in the turnover of endosomal/lysosomal HSPGs, whereas heparanase secretion (i.e. by atherogenic agents or inflammatory cytokines) (59) is responsible for its pathological aspects. Fig. 5 summarizes our proposed model for heparanase biosynthesis and trafficking. Altogether, our results indicate that HSPGs efficiently limit the bioavailability of extracellular heparanase and hence protect cells and tissues from undesirable effects of the enzyme.