Cellular uptake of mammalian heparanase precursor involves low density lipoprotein receptor-related proteins, mannose 6-phosphate receptors, and heparan sulfate proteoglycans.

Mammalian heparanase, strongly implicated in the regulation of cell growth, migration, and differentiation, plays a crucial role in inflammation, angiogenesis, and metastasis. There is thus a clear need for understanding how heparanase activity is regulated. Cells can generate an active form of the enzyme from a larger inactive precursor protein by a process of secretion-recapture, internalization, and proteolytic processing in late endosomes/lysosomes. Cell surface heparan sulfate proteoglycans are the sole known components with a role in this trafficking of the heparanase precursor. Here, we provide evidence that heparan sulfate proteoglycans are not strictly required for this process. More importantly, by heparanase transfection, binding, and uptake experiments and by using a combination of specific inhibitors and receptor-defective cells, we have identified low density lipoprotein receptor-related proteins and mannose 6-phosphate receptors as key elements of the receptor system that mediates the capture of secreted heparanase precursor and its trafficking to the intracellular site of processing/activation.

. Up-regulation of heparanase-1 occurs in nephrosis, cirrhosis, diabetes, and a number of human cancers (3,4,7) and markedly promotes tumor angiogenesis and metastasis (9). Understanding how heparanase-1 activity is generated and controlled is thus of major medical interest.
Heparanase-1 is synthesized as an inactive ϳ65-kDa precursor that subsequently undergoes proteolytic cleavage, yielding ϳ8and ϳ50-kDa protein subunits that heterodimerize to form the active enzyme (10 -13). Heparanase-1 is thought to be mostly an intracellular enzyme, because cells rapidly bind and internalize secreted pro-heparanase-1 (14 -16), transferring the internalized precursor to late endosomes/lysosomes in which it is processed into the mature active form of the enzyme and stays localized (15). Recently, it has been suggested that the uptake or reuptake of secreted heparanase-1 precursor is mediated by cell membrane HSPGs, in particular the syndecans (16). The experiments reported here confirm a role for HSPGs in heparanase-1 uptake but indicate that these are not the sole elements involved and are not strictly required. Rather, we found that low density lipoprotein receptor-related protein (LRP) and possibly other receptor-associated protein (RAP)-sensitive receptor(s), as well as mannose 6-phosphate (Man-6-P) receptors are key to this process.
Antibodies-The synthetic peptide YGPDVGQPRRKTAKM, corresponding to a sequence in the ϳ50-kDa subunit of mature heparanase-1, was coupled to keyhole limpet hemocyanin for immunization. Rabbit anti-heparanase-1 antibodies were affinity-purified on the same peptide coupled to EAH-Sepharose. Mouse monoclonal anti-actin antibody was obtained from Sigma-Aldrich. Mouse monoclonal antibodies 3G10 (recognizing desaturated glucuronate created by enzymatic cleavage of heparan sulfate with bacterial heparitinase) (17) and 2E9 (recognizing an epitope in the cytoplasmic domain of syndecan-1) (18) were prepared as described previously.
Recombinant RAP-His 6 -RAP (plasmid provided by Dr. T. Willnow) was expressed in BL21(DE3)pLysS bacteria (Invitrogen) and purified using a nickel-nitrilotriacetic acid resin column (Qiagen) according to the manufacturer's instructions. His 6 -RAP was checked for purity by SDS-PAGE and stored at Ϫ80°C in Tris-buffered saline.
Cell Cultures and Transfection-All of the cells used were immortalized cell lines. Wild-type mouse embryonic fibroblasts (MEFs) were derived from wild-type mice. Wild-type HEK 293-T cells, wild-type CHO K1 cells, and MEFs that are genetically deficient in LRP (PEA13) (19) were obtained from ATCC. HS-deficient CHO cells (CHO 677) (20) were provided by Dr. J. D. Esko (University of California, San Diego). CHO cells deficient in LRP (CHO 13-5-1) (21) were provided by Dr. D. FitzGerald (National Institutes of Health, Bethesda, MD). MEF deficient in both the cation-independent (ϳ300-kDa) and cation-dependent (ϳ46-Da) Man-6-P receptor (22) were provided by Dr. K. von Figura (University of Göttingen, Germany). Cells were routinely grown in Dulbecco's modified Eagle's medium/nutrient mixture F-12 (Invitrogen) supplemented with 10% fetal calf serum (Hyclone). For transient expression studies, cells were grown in 6-well plates and transfected (1 g of plasmid DNA/well) using FuGENE 6 reagent (Roche Applied Science) for 24 h. In experiments in which the secreted protein or the effect of agents had to be analyzed, cells were transfected for 4 h, washed, and cultured for 20 h in serum-free medium with or without the addition of the agent monensin, tunicamycin A (both purchased from Sigma-Aldrich), and His 6 -RAP.
Western Blotting-After removal of the culture medium and several washes, cell extracts were prepared using ice-cold lysis buffer (1% Nonidet P-40 in Ca 2ϩ -and Mg 2ϩ -free phosphate-buffered saline (pH 7.4) supplemented with 1 mM phenylmethylsulfonyl fluoride, 1 g/ml pepstatin A, 10 g/ml leupeptin, and 10 mM EDTA). Cleared cell lysates (and equivalent volumes of the media) were fractionated by SDS-PAGE under reducing conditions (4 -15% gradient gels, Bio-Rad) and transferred to Hybond C-extra membranes (Amersham Biosciences). The membranes were probed with the appropriate antibody diluted in blocking buffer (0.1% Tween 20 and 5% nonfat dry skim milk). Bound antibody was detected with species-specific secondary antibody conjugated to peroxidase (goat anti-rabbit and goat anti-mouse, Bio-Rad) using the ECL detection system (PerkinElmer Life Sciences).
Heparanase-1 Binding and Uptake Studies-For binding and uptake studies, 0.75 ml of serum-free culture medium, conditioned by transiently heparanase-1-transfected HEK 293-T cells (heparanase-1 medium, containing ϳ5 nM of precursor), was added to subconfluent cell cultures (ϳ400,000 cells) at 4 and 37°C with or without the addition of heparin (Calbiochem), Man-6-P, glucose-6-phosphate, mannose (Sigma-Aldrich), or His 6 -RAP. After the indicated lengths of time, the medium was removed, cells were washed, and cell lysates were prepared as described above. Similar experiments with similar results were also conducted with purified heparanase-1 precursor isolated from heparanase-1 medium by heparin-affinity chromatography. To remove cell surface chondroitin sulfate or HS, cells were treated with chondroitinase ABC (0.1 unit/ml) or heparitinase (0.006 unit/ml), respectively (Seikagaku Corp.) for 4 h at 37°C, washed with serum-free medium, and further incubated with the enzyme in heparanase-1 medium. To distinguish between the membrane-bound and intracellular forms of heparanase-1, cells were treated with 250 g/ml trypsin (Sigma-Aldrich) at 4°C for 15 min. Soybean trypsin inhibitor (Sigma-Aldrich) was added, and after several washes total cell lysates were prepared as described above.
Endoglycosidase Digestions-Methanol/chloroform-precipitated cell extracts were suspended in 0.5% SDS and 0.1 M ␤-mercaptoethanol and boiled for 5 min. For PNGase F digestion (Roche Applied Science) the solution was adjusted to 3% Nonidet P-40, 50 mM Tris/HCl (pH 8.6), and 0.5 units of PNGase F. For Endo H digestion (Roche Applied Science) the solution was adjusted to 75 mM sodium acetate (pH 5.5), 0.05% phenylmethylsulfonyl fluoride, and 5 milliunits of Endo H. Both digestions were incubated overnight at 37°C followed by Western blotting, as described above.

RESULTS
HSPGs Are Not Strictly Required for Heparanase-1 Uptake-HEK 293-T cells that were transiently transfected with heparanase-1 accumulated both the inactive precursor (ϳ65 kDa) and the mature active form (ϳ50 kDa) of the enzyme in roughly similar amounts (Fig. 1A, lane 2). The culture medium of these cells contained large amounts of the pro-enzyme but no mature enzyme (Fig. 1A, lane 4). Yet, many other heparanase-1-transfected cells released hardly any pro-enzyme in their culture media (see below for examples). Monensin inhibited the accumulation of processed enzyme (Fig. 1B) and heparanase activity (Fig. 1C) in HEK 293-T cells, consistent with the late endosome/lysosome being the primary heparanase-1-processing organelle (15). The capacity of cells to internalize secreted heparanase-1 precursor and to convert the internalized protein into an active form of the enzyme was evident from adding spent culture medium, conditioned by heparanase-1-transfected HEK 293-T cells (Fig. 1A), to cultured mammalian cell lines, e.g. wild-type CHO K1 cells (Fig. 1, D and E) and wild-type MEFs (Fig. 1, F and G). Both cell types quickly bound the ϳ65-kDa heparanase-1 precursor and progressively converted the precursor into mature ϳ50-kDa protein ( Fig. 1, D and F). In MEFs, this resulted in markedly enhanced heparanase activity (Fig. 1G). Treating the cells with trypsin removed none of the mature form but, depending on the length of the incubation with precursor, removed most or part of the precursor form of heparanase-1 (Fig. 1E), suggesting that all processing occurs inside the cells. Below, we describe similar transfection and uptake experiments using various cell lines with known receptor deficiencies.
Consistent with a possible role for HSPG in heparanase-1 uptake or reuptake, HS-deficient CHO 677 cells (20) that were transiently transfected with heparanase-1 released large amounts of the ϳ65-kDa heparanase-1 precursor into their culture medium ( Fig. 2A, lane 4). Heparanase-1-transfected wild-type CHO K1 cells, in contrast, released none ( Fig. 2A, lane 3). Compared with wild-type CHO K1 cells, lysates of CHO 677 cells still contained considerable (almost similar) amounts of the ϳ50-kDa active enzyme ( Fig. 2A, lane 2 versus lane 1). Yet, in comparison with wild-type CHO K1 cells, HS-deficient CHO 677 cells showed a moderate delay in binding and uptake/processing of exogenous, added pro-heparanase-1 (Fig. 2B). Consistently, treatment of wild-type CHO K1 cells with bacterial heparitinase slightly decreased heparanase-1 binding and delayed uptake/processing (Fig. 2C, left panel). The heparitinase digestion was effective, as shown by the appearance of the 3G10 epitope (17), decorating discrete bands corresponding to the molecular masses of core proteins of different HSPGs (e.g. ϳ65-kDa syndecan-1, ϳ60-kDa glypican, ϳ35-kDa syndecan-4) (Fig. 2C, right panel). Because CHO 677 cells are deficient in HS but overexpress chondroitin sulfate (20), we then investigated whether chondroitin sulfate proteoglycans might be responsible for the remaining heparanase-1 binding and uptake, by pretreating CHO 677 cells with chondroitinase ABC. Although the treatment was effective (appearance of a discrete ϳ65-kDa band corresponding to the core protein of syndecan-1, a hybrid HS/chondroitin sulfate proteoglycan (18)) ( Fig. 2D, right panel), we could not detect any effect on heparanase-1 binding and internalization (Fig. 2D, left panel). It had been suggested that added heparin competes with cell surface HSPGs for heparanase-1 binding and thus inhibits uptake (16). Our data confirmed that the binding and uptake of heparanase-1 can be totally inhibited by the addition of increasing concentrations of heparin (1-50 g/ml). Yet, inhibition not only occurs in wild-type CHO K1 cells (Fig. 2E, left) but also in HSdeficient CHO 677 cells (Fig. 2E, right), implying that heparin also inhibits heparanase-1 binding to other receptor(s). From all of these data, we conclude that HSPGs are not strictly required for the capture or recapture and internalization of the heparanase-1 precursor and are not the key players in this process. Rather, they may have a role in regulating its efficiency and/or kinetics.
LRP Involvement in Heparanase-1 Precursor Uptake-Because the roles of LRP in the uptake of proteases, protease inhibitor complexes, lipoproteins, and lipases are well established (23,24), we then tested whether this receptor might be involved in the uptake or reuptake of secreted heparanase-1 precursor. MEFs deficient in LRP (PEA13 cells) (19) that were transiently transfected with heparanase-1 released large amounts of the ϳ65-kDa precursor in their culture medium (Fig. 3A,  left panel, lane 4). Transfected wild-type MEFs (which express LRP abundantly), in contrast, released none (Fig. 3A, left panel, lane 3). Comparing LRP-deficient CHO 13-5-1 cells (21) with wild-type CHO K1 cells yielded similar results (Fig. 3A, right panel). As these data were suggestive of a role for LRP, this role was further assessed by examining the effect of RAP, a receptor-associated protein that inhibits ligand binding to several members of the low density lipoprotein-receptor family. The addition of increasing concentrations of RAP to heparanase-1-transfected MEFs (Fig. 3B) or CHO K1 cells (data not shown) resulted in progressively increasing amounts of the ϳ65-kDa heparanase-1 precursor in the culture medium. The addition of RAP to transiently transfected PEA13 cells (Fig. 3C) or CHO 13-5-1 cells (data not shown) also increased the amount of heparanase-1 precursor that these cells released in the culture medium, suggesting that RAP-sensitive receptors other than LRP (e.g. sortilin, very low density lipoprotein receptor, megalin, etc.) might also be involved in heparanase-1 precur-sor uptake. Yet, RAP is a known ligand for heparin and may bind to HSPGs in some cell types (25), while not binding to HSPGs in cells that produce other, less or differently sulfated forms of HS (26). The addition of RAP to heparanase-1-transfected CHO 677 cells markedly enhanced the amount of heparanase-1 precursor that these cells released in the culture medium and reduced the amount of mature enzyme in the lysate of the cells (Fig. 3D), indicating that the effect of RAP is not due or solely due to the competition of binding to HSPGs. Consistent with a role for LRP, uptake experiments detected a moderate but clear difference between MEFs and PEA13 cells in the rate of heparanase-1 conversion (Fig. 3E, lane 1 versus lane 3, showing inverted ratios of precursor over mature protein after 120 min of incubation with heparanase-1 precursor at 37°C). Consistently, the processing of the heparanase-1 precursor by MEFs decreased when RAP was included during the uptake experiment, reducing it to the level of processing by PEA13 cells (Fig. 3E, lanes  2 and 3). At the same time, the precursor associated with MEFs remained more accessible to trypsin in the presence of RAP (Fig. 3F,  lanes 3 and 4) than in the absence of RAP (Fig. 3F, lanes 1 and 2), suggesting that the net effect of RAP is reduced uptake and, consequently, conversion. RAP also had a small effect on uptake/processing by PEA13 cells (Fig. 3E, lane 4), confirming that RAP-sensitive receptors other than LRP might also mediate heparanase-1 precursor uptake. The inclusion of RAP moderately reduced the levels of precursor and severely reduced the levels of mature enzyme associating with CHO 677 cells (Fig. 3G), confirming a role for LRP (and possibly other RAPsensitive receptors) in heparanase-1 binding and uptake and suggesting that this role might be less apparent in the presence of HSPGs. Consistently, increasing concentrations of RAP also progressively reduced the amount of heparanase-1 precursor that became associated with MEFs when these cells were incubated with heparanase-1 at 4°C (preventing internalization) for 60 min (time sufficient to reach maximal binding), implying that LRP and/or other RAP-sensitive receptors are involved in the binding of the heparanase-1 precursor (Fig. 3H).  A, B), and heparanase assay (C) of cell extract and medium samples derived from HEK 293-T cells, transiently transfected with empty () or heparanase-1 (H1) expression vector, and cultured without or with the addition of monensin. D-G, binding, internalization, processing, and activation of exogenous, added heparanase-1 precursor. Serum-free culture medium, conditioned by transiently heparanase-1-transfected HEK 293-T cells (as in A), was added to wild-type CHO K1 cells (D, E) and to wild-type MEFs (F, G) at 37°C for the indicated lengths of time. Nonaccessibility to trypsin was taken as a measure for heparanase-1 internalization (E). Total cell lysates were subjected to Western blotting with anti-heparanase-1 and antiactin (D-F) or tested for heparanase activity (G).
Open arrowheads point at the ϳ65-kDa heparanase-1 precursor; filled arrowheads point at the ϳ50-kDa mature active heparanase-1, and arrows point at actin. Depolymerization of [ 35 S]HS, taken as a measure for heparanase activity, was monitored by SDS-PAGE and detected by autoradiography (C, G).

Man-6-P Receptor Involvement in Heparanase-1 Precursor Uptake
-As most of the soluble lysosomal hydrolases are transported via Man-6-P receptors (27), we then investigated whether this receptor system might also play a role in the capture or recapture of heparanase-1. If so, then heparanase-1 glycosylation should affect the trafficking and activation of the precursor protein. Treatment of heparanase-1-transfected HEK 293-T cells with tunicamycin A (TM) (an inhibitor of N-glycosylation) or replacement of all six Asn residues in the protein that compose sites for N-glycosylation (28) by Gln residues resulted in heparanase-1 migrating as only one single band of ϳ50 kDa, corresponding to the predicted molecular mass of the nonglycosylated form of proheparanase-1 (Fig. 4, A and B, left panels), and no gain in heparanase activity (Fig. 4, A and B, right panels), suggesting that the maturation of the pro-enzyme is inhibited. These results are in clear contrast to an earlier claim that the N-glycosylation of heparanase-1 is not required for its activation and activity (28). Although we have no formal explanation for this discrepancy, we suggest that the metabolic stability of the (preformed) active enzyme (half-life of ϳ30 h) (16) may hamper detecting effects of TM on heparanase activity in stably transfected cells (28). Cell extracts of transiently heparanase-1-transfected HEK 293-T cells were then digested in vitro with either PNGase F (cleaving all Asn-linked oligosaccharides) or Endo H (removing only the high mannose and hybrid types of Asn-linked oligosaccharides). Western blots of these digestions (Fig. 4C) revealed that the ϳ65-kDa precursor and the ϳ50-kDa mature enzyme are equally sensitive to PNGase F and Endo H, both digestions leaving none of the proteins intact and leading to multiple bands (in the case of Endo H, including also major bands that correspond to the molecular masses of the nonglycosylated forms (ϳ50 kDa for the precursor and ϳ43 kDa for the mature enzyme)). A similar result was obtained for the secreted heparanase-1 precursor (data not shown). To ensure that these results were not due to the levels of expression or dependent on the cell type, we repeated these experiments with stably heparanase-1-transfected Madin-Darby canine kidney type II cells, obtaining similar results (Fig. 4D). Inclusion of protease inhibitors during the digestions had no effect on the banding pattern (shown for Endo H in Fig. 4E). Moreover, neither enzyme degraded a heparanase-1-mutant in which all six N-glycosylation Asn residues had been replaced by Gln residues (Fig. 4F). Both enzymes also failed to degrade the nongly-  cosylated heparanase-1 precursor accumulating in TM-treated cells (not shown), further ensuring that the degradation of the glycosylated precursor was not due to contaminating protease activities. These data imply that most of the Asn-linked oligosaccharides on the precursor and mature form of heparanase-1 are of the non-complex type. Possibly, these contain Man-6-P residues, preventing their conversion to complex carbohydrates.
The sole addition of Man-6-P (10 mM) had little or no effect on the uptake and conversion of heparanase-1 precursor by wild-type MEFs (Fig. 5A, lanes 1 and 3). However, in the presence of RAP (2 M) Man-6-P further inhibited the conversion of heparanase-1 precursor into the mature form (Fig. 5A, compare lanes 1 and 2 and lanes 2 and 4). A similar combination of Man-6-P and RAP completely prevented the accumulation of mature heparanase-1 in heparitinase-treated MEFs (Fig. 5A, lanes 5-8), and all of the precursor protein associating with these cells was accessible to trypsin (not shown). In contrast, in MEFs that lacked LRP (Fig. 5B) the sole addition of Man-6-P (10 mM) markedly reduced conversion of heparanase-1 precursor; the further addition of RAP had little additional inhibitory effect on this conversion. Similarly, in wild-type CHO K1 cells (Fig. 5C) and in LRP-deficient CHO 13-5-1 cells (Fig. 5D) the combination of Man-6-P and RAP was more effective than either Man-6-P or RAP alone in reducing the con- version of precursor, and in HS-deficient CHO 677 cells (Fig. 5E) it completely inhibited precursor uptake and conversion. Neither glucose-6-phosphate (10 mM) nor mannose (10 mM) had significant effects on uptake and conversion (Fig. 5, C-E). Finally, we also supplied proheparanase-1 to MEFs that lacked Man-6-P receptors (22) (Fig. 5F). All of the heparanase-1 precursors associating with these cells remained in precursor form, which was perhaps to be expected, because such MEFs are likely to be deficient in the lysosomal enzymes required for the processing of internalized heparanase-1 precursors. In the absence of RAP, most of this heparanase-1 resisted treatment of the cells with trypsin, confirming that substantial internalization can occur in the absence of Man-6-P receptors. Adding RAP markedly reduced the accumulation of trypsin-resistant (internalized) heparanase-1 precursor in these cells. Significantly, the internalization of heparanase-1 precursor was completely blocked by RAP when these MEFs were pretreated with heparitinase (Fig. 5F, lane 10). As expected, adding Man-6-P had no effect on heparanase-1 uptake by these cells. Together, these data clearly demonstrate that HSPGs, RAP-sensitive receptors, and Man-6-P recep-tors cooperate in the binding and internalization of heparanase-1 precursor.

DISCUSSION
In this study we show that receptors involved in the binding and internalization of secreted heparanase-1 precursor include HSPGs, LRP, possibly other RAP-sensitive receptor(s), and Man-6-P receptors. Mutant cell lines, undoubtedly deficient in at least one of these receptor systems, indicate that each of these receptors is individually dispensable. Thus, different types of cells are likely to use different combinations of these receptors for regulating their heparanase activities.
Persistent uptake and activation of latent heparanase-1 in the absence of HS, i.e. by HS-deficient CHO 677 cells and by heparitinase-treated cells, indicate that HSPGs are not the sole elements involved and are not strictly required for uptake. Moreover, adding RAP to MEFs that lack Man-6-P receptors or adding the combination of RAP and Man-6-P to wild-type MEFs or CHO cells severely reduces internalization of heparanase-1, suggesting that HSPGs, by themselves, are not effective internalization receptors. Yet, CHO 677 cells and heparitinase-treated cells are less effective at internalizing heparanase-1 precursor and are more sensitive to RAP and Man-6-P than are wild-type CHO cells or corresponding control cells, clearly indicating that HSPGs play a signif-icant role in the capture and internalization of heparanase-1. This suggests a scheme whereby HSPGs function in the context of other receptors (i.e. LRP and Man-6-P receptors), facilitating the binding of ligand to these receptors, a recurring theme shared with several other HS-dependent receptor systems. HSPGs can modulate receptor functions by concentrating ligands at the cell surface, presenting ligands to cell-surface receptors or participating in ternary ligand⅐receptor⅐HS complexes and stabilizing ligand-receptor interactions. These mechanisms could also apply to the internalization of heparanase-1, but further work is needed to distinguish between multiple additive and truly cooperative binding and internalization mechanisms.
In the absence of HS and functional Man-6-P receptors (e.g. CHO 667 cells that are exposed to Man-6-P, or Man-6-P receptor-deficient MEFs that are treated with heparitinase) binding and internalization of heparanase-1 precursor persists at a substantial rate. This residual uptake is totally abolished when RAP is also included, unambiguously identifying LRP and possibly other RAP-sensitive receptors as bona fide and major pro-heparanase-1 internalization receptors. The effects of Man-6-P on precursor internalization, evident in LRP-deficient cells or in the presence of RAP, and indirectly the extreme sensitivity of Man-6-P receptor-deficient MEFs to the combination of heparitinase and RAP indicate that Man-6-P receptors work along with LRP and/or other RAP-sensitive receptors in the internalization of the heparanase-1 precursor. Thus, the various mutants are equally (if not primarily) helpful in demonstrating the participation of the "alternative" receptor systems as they are in showing the contribution of the mutant receptor under question. Multiple different internalization receptors would seem to provide robustness to this uptake mechanism, possibly reflecting a strong need for cells to keep extracellular concentrations of heparanase-1 precursor at low levels.
It is not clear to what extent the secretion-recapture route represents a physiological path for cells to direct the heparanase-1 precursor from the endoplasmic reticulum (the site of biosynthesis) to lysosomes, in which it is processed into the mature active form of the enzyme. In most cells with easily detectable endogenous heparanase activity, e.g. nontransfected HEK 293-T cells, heparanase-1 is hardly if at all detectable at the protein level. Possibly, mechanisms for direct sorting from early secretory pathway compartments to lysosomes may be overwhelmed in instances of excessive heparanase-1 production as in transfected cells. In that case it remains to be established whether such direct sorting of heparanase-1 is mediated by the same receptors that were implicated here in the uptake or reuptake of the secreted heparanase-1 precursor, but LRP and in particular Man-6-P receptors are logical candidates.
Still, cells that are not synthesizing elevated levels of heparanase-1 on their own, e.g. quiescent vascular endothelial cells, may capture and activate heparanase-1 precursor secreted by cells that overexpress heparanase-1, e.g. tumor cells, with potential "paracrine" effects on HSdependent signaling pathways in these "recipient" cells. LRP levels and activity are known to be substantially decreased in many tumors (23). Based on frequent loss of heterozygosity and functional mutations in tumors from cancer patients, and from receptor expression studies in tumor cell lines, it is proposed that the ϳ300-kDa Man-6-P (Man-6-P⅐IGF-II) receptor represents an important tumor suppressor gene, possibly by mediating the endocytosis and degradation of IGF-II, in this way opposing its growth-promoting effects (27). Our findings could extend such a proposal whereby the loss of LRP and the Man-6-P⅐IGF-II receptor leads to the accumulation of (tumor) heparanase-1 in the extracellular space where it might be taken up and activated by (receptor-expressing) surrounding endothelial cells (promoting e.g. tumor angiogenesis). Moreover, it has been shown that the ϳ65-kDa heparanase-1 precursor can enhance Akt and in some cases mitogen-activated FIGURE 5. Man-6-P receptors are involved in the uptake of heparanase-1 precursor. Serum-free culture medium, conditioned by transiently heparanase-1-transfected HEK 293-T cells, was added at 37°C to wild-type MEFs (A), LRP Ϫ/Ϫ MEFs (PEA13) (B), wild-type CHO K1 cells (C), LRP Ϫ/Ϫ CHO 13-5-1 cells (D), HS-deficient CHO 677 cells (E), and Man-6-P receptor Ϫ/Ϫ MEFs (F) for the indicated lengths of time, either without any pretreatment or after treatment for 4 h at 37°C with bacterial heparitinase and either with or without the addition of the indicated concentrations of Man-6-P (M-6-P), RAP, glucose 6-phosphate (G-6-P), or mannose. Nonaccessibility to trypsin was taken as a measure for heparanase-1 internalization in Man-6-P receptor Ϫ/Ϫ MEFs. Total cell lysates were subjected to Western blotting using anti-heparanase-1 and anti-actin. Open arrowheads point at the ϳ65-kDa heparanase-1 precursor, filled arrowheads point at the ϳ50-kDa mature heparanase-1, and arrows point at actin. Note the nonspecific band at ϳ66 kDa occasionally appearing in the lysate of CHO 677 cells. protein kinase signaling by an unidentified receptor, stimulating endothelial cell migration and invasion (29). The signaling role of the inactive precursor may be enhanced by the loss of LRP and Man-6-P receptors in tumor cells. A corollary of identifying LRP (and/or other RAP-sensitive receptors) as a binding receptor for the heparanase-1 precursor could be that it also functions as the receptor that mediates the signaling functions of the inactive precursor in endothelial cells.
In summary, our studies identify HSPGs, LRP, possibly other RAPsensitive receptors, and Man-6-P receptors as responsible for the capture or recapture of the heparanase-1 precursor and its delivery to the intracellular compartments in which it is activated. This may have major implications for the development of novel strategies for inhibiting heparanase-1 activation, inflammatory processes, and cancer progression.