In Vivo Clearance of Ternary Complexes of Vitronectin-Thrombin-Antithrombin Is Mediated by Hepatic Heparan Sulfate Proteoglycans*

Thrombin is inhibited by its cognate plasma inhibitor antithrombin, through the formation of covalent thrombin-antithrombin (TAT) complexes that are found as ternary complexes with vitronectin (VN-TAT). To determine whether the metabolism of VN-TAT ternary complexes is different from that previously reported for binary TAT complexes, plasma clearance studies were done in rabbits using human VN-TAT. 125I-VN-TAT was shown to be cleared rapidly from the circulation (t½α = 3.8 min) in a biphasic manner mainly by the liver. 125I-TAT had a similar initial clearance (t½α = 5.3 min) but had a significantly faster β-phase clearance (t½β = 42.8 min versus85.4 min for VN-TAT; p = 0.005). Protamine sulfate and heparin abolished the rapid initial α-phase of125I-VN-TAT clearance and reduced its liver-specific association and in vivo degradation. Heparin also reduced the α-phase clearance of 125I-TAT and was associated with the appearance of high molecular weight complexes, suggesting enhanced complex formation between VN and TAT. 125I-VN-TAT binding to HepG2 cells was reduced by competition with VN-TAT or heparin but to a much lesser extent in the presence of TAT. The binding of VN-TAT to HepG2 cells was not inhibited by competition with the low density lipoprotein receptor-related protein ligand, methylamine-α2-macroglobulin. 125I-VN-TAT binding was also inhibited by treating HepG2 cells with heparinase or by growing the cells in the presence of β-d-xyloside. Finally, both heparin and chloroquine, but not methylamine-α2-macroglobulin, reduced the internalization and degradation of VN-TAT by HepG2 cells. Taken together, these data indicate the importance of VN in TAT metabolism and demonstrate that VN-TAT binds to liver-associated heparan sulfate proteoglycans, which mediate its internalization and subsequent intracellular degradation.

I-VN-TAT binding was also inhibited by treating HepG2 cells with heparinase or by growing the cells in the presence of ␤-D-xyloside. Finally, both heparin and chloroquine, but not methylamine-␣ 2 -macroglobulin, reduced the internalization and degradation of VN-TAT by HepG2 cells. Taken together, these data indicate the importance of VN in TAT metabolism and demonstrate that VN-TAT binds to liver-associated heparan sulfate proteoglycans, which mediate its internalization and subsequent intracellular degradation.
Vitronectin (VN), a 78-kDa glycoprotein synthesized by the liver, is found in the circulation at a plasma concentration of 200 -400 g/ml (12). VN plays a key role in the attachment of cells to their surrounding matrix (13,14) and has been found to have various functions in proteolytic cascades. For example, VN has been shown to stabilize the major inhibitor of fibrinolysis PAI-1 (15), to inhibit the complement system (16,17), and to neutralize the heparin catalysis of AT inhibition of thrombin and factor Xa (18,19).
Although the metabolism of binary forms of SECs have been well studied, ternary complexes have not, despite the fact that VN has been shown to play an important role in the metabolism of various SECs. For example, TAT in serum or plasma has been found to exist in the form of a covalent ternary complex with VN (20,21). Other VN-SECs have also been identified, including VN-thrombin-HCII (22,23), VN-thrombin-PN-1 (24), VN-factor Xa-AT (25), VN-thrombin-PAI-1 (26), and VN-thrombin-protein C inhibitor (27) as well as ternary complexes between VN and thrombin with the mutant ␣ 1 -proteinase inhibitor Pittsburgh (27). The formation of ternary complexes with VN suggests that VN may be important physiologically in the metabolism of various SECs. In this study, we have examined the metabolism of 125 I-VN-TAT in vivo and have identified hepatocyte-associated heparan sulfate proteoglycans (HSPG) as major binding sites involved in the metabolism of VN-TAT both in vivo and in vitro.

EXPERIMENTAL PROCEDURES
Materials-Chloroquine, 5-bromo-4-chloro-3-indolyl phosphate, nitroblue tetrazolium, methylamine, bovine serum albumin (BSA), heparinase (heparin lyase I from Flavobacterium heparinum), p-nitrophenyl ␤-D-xylopyranoside (␤-D-xyloside), protamine sulfate (grade X, from * This work was supported, in part, by a grant from the Canadian Red Cross Society Research and Development Fund. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  salmon), and heparin (grade 1-A, from porcine intestines) were purchased from Sigma. IODO-GEN was purchased from Pierce. D-Phenylalanyl-L-prolyl-L-arginine-chloromethylketone (Ͼ99% active) was purchased from Calbiochem. The protein assay kit was purchased from Bio-Rad (Mississauga, ON).
Proteins-VN-TAT was purified by heparin-Sepharose chromatography from human serum as described by de Boer et al. (28). The protein concentration of the purified VN-TAT was determined with the Bio-Rad protein assay kit. The relative amounts of AT and HCII in the VN-TAT preparations were determined by enzyme-linked immunosorbent assay (23). Purified human AT concentrate was the generous gift of Bayer Canada Inc. (Toronto, Canada). Human ␣-thrombin (Ͼ2,000 units/mg and Ͼ99% active) was the generous gift of Dr. J. Fenton (New York State Division of Biologicals, Albany, NY). Rabbit anti-human VN antibodies were purchased from Celsus Laboratories Inc. (Cincinnati, OH). Goat anti-human HCII serum was purchased from Diagnostica Stago through Murex Diagnostics (Guelph, Canada). The use of sheep anti-human AT IgG has been described previously (29). Affinity-purified chicken anti-human ␣-thrombin IgG was raised as described (11) using purified human ␣-thrombin as antigen. Electrophoretically pure ␣ 2 -macroglobulin was purchased from Calbiochem and modified with methylamine as described previously (30).
Radioiodination of Proteins-For radioiodination of all proteins, IODO-GEN-coated glass vials were used as described previously (31). After radioiodination, VN-TAT had a specific activity of 1,400 -2,800 dpm/ng; human thrombin had a specific activity of 2,100 -2,700 dpm/ng; AT had a specific activity of 4,400 dpm/ng; and ␣ 2 M* had a specific activity of 12,000 dpm/ng.
TAT Formation-TAT was formed by incubating thrombin, or 125 Ithrombin, at a 1:3 molar ratio with AT, at 37°C for 30 min. To inhibit residual active thrombin, D-phenylalanyl-L-prolyl-L-arginine-chloromethylketone was added to the reaction mixture at a final concentration of 100 M and incubated at 37°C for 5 min. The TAT preparation was dialyzed against ␣-MEM buffered with 10 mM HEPES (pH 7.4) (HMEM) overnight at 4°C and then aliquoted and stored at Ϫ70°C. 125 I-TAT was aliquoted immediately after preparation and stored at Ϫ70°C.
Plasma Elimination of 125 I-VN-TAT, 125 I-TAT, and 125 I-AT-Young male New Zealand White rabbits, weighing between 2.2 and 2.7 kg were infused intravenously with 125 I-VN-TAT (12.6 -17.5 ϫ 10 6 dpm; 8 -16 g/animal), 125 I-TAT (3.4 -17.5 ϫ 10 6 dpm; 3-3.5 g/animal), or 125 I-AT (12.6 ϫ 10 6 dpm; ϳ2.9 g/animal) made up in 1 ml of saline. For competition experiments, heparin (10,000, 5,000, or 1,000 units/ml) or protamine sulfate (65 mg/ml) was dissolved in sterile saline, and 125 I-VN-TAT was added directly to each solution just prior to intravenous injection. All infusions were done using the marginal ear vein of the rabbit. Blood samples (1 ml) were taken at various time points, from the auricular artery of the opposite ear, into 200 l of anticoagulant (CP2D obtained from the Canadian Red Cross Society; it contained 0.016 M citric acid, 0.09 M sodium citrate, 0.016 M monobasic sodium phosphate, and 0.284 M dextrose). After thorough mixing, they were placed into microcentrifuge tubes and centrifuged for 10 min at 16,000 ϫ g. For each sample, 100 l was counted for radioactivity directly, while another 100-l aliquot was added to 100 l of 20% trichloroacetic acid, vortexed, and left on ice for 15 min. The trichloroacetic acid-treated samples were then centrifuged at 16,000 ϫ g for 10 min, and the supernatants were removed to determine radioactivity. The plasma from the blood sample taken at the 1-min time point was designated as representing 100% radioactivity.
Pharmacokinetic Analysis of the Clearance Data-The clearance data were analyzed using the technique of curve stripping as described previously (32). Briefly, the logarithm of the fraction of protein-associated radioactivity remaining in the circulation was plotted versus time. The terminal exponential phase of clearance was curve-fit by linear regression. To determine the initial clearance phase, the contribution of the terminal clearance phase was subtracted from it as follows. Initial clearance data were substituted into the equation of the line (for the terminal clearance curve) and then subtracted from their initial clearance values, and then the new data were plotted and analyzed by linear regression. This analysis revealed a biphasic clearance curve for VN-TAT and TAT, indicating a two-compartment clearance model. This clearance model is described by the equation, where C n are coefficients and a n are rate constants. The half-lives were determined from the a values. The data for the clearance of VN-TAT in the presence of heparin or protamine sulfate best fit a single-compartment clearance model. The clearance parameters C n , a n , and t1 ⁄2 were then compared using a two-tailed t test, and values are reported as the mean Ϯ S.E. Determination of Organ Uptake of 125 I-VN-TAT-For the determination of organ uptake, 125 I-VN-TAT (17.5 ϫ 10 6 dpm; 5.5 g) was injected into male New Zealand White rabbits (n ϭ 3). After 1 h, the rabbits were anesthetized, and the femoral veins were isolated and opened. One liter of saline was then perfused into each animal via a carotid artery cannula. The perfusion in all cases was judged to be successful by evidence of blanching of the liver and the lack of blood in the perfusate at the completion of the perfusion. The liver, lungs, kidneys, bladder, heart, and spleen were removed from each animal and weighed. The individual organs were then homogenized with a Polytron homogenizer (Brinkmann, Switzerland) in Tris-buffered saline, with ϳ25 ml of octanol added to prevent foaming, for ϳ2 min each, at ϳ75% of the highest setting. After measuring the volume of each homogenate, 2 ml was removed for the determination of the radioactivity present. In the experiments with 125 I-VN-TAT (ϳ12.6 ϫ 10 6 dpm) either alone or in competition with heparin (10,000 units/ml) (n ϭ 3 for each), the incubation period was reduced to only 30 min, and only the livers were removed and processed for determination of radioactive uptake.
Binding of 125 I-VN-TAT to HepG2 Cells-For these experiments, HepG2 cells were grown on 24-well plates to a final density of ϳ4.5 ϫ 10 5 cells/well. Binding experiments were performed in HMEM containing 0.5% BSA. The HepG2 cells were washed in binding buffer and then incubated with 125 I-VN-TAT, with or without competitors, at 4°C for 2 h. The cells were then washed three times with 10 mM HEPES (pH ϭ 7.4) containing 0.15 M NaCl, 1 mM CaCl 2 , 2 mM MgCl 2 , and 0.5% BSA (HBSB). The cells and associated radioactivity were solubilized by incubation with 2 M NaOH overnight before counting for radioactivity. For experiments with HepG2 cells using ␤-D-xyloside, the cells were trypsinized and replated on 24-well culture dishes, in the presence of a final concentration of 2.5 mM ␤-D-xyloside in the cell culture medium for 3 days, before conducting the binding experiment. This reagent was made up in dimethyl sufoxide, and an equal volume of dimethyl sulfoxide was added to control wells. In experiments using heparinase, the HepG2 cells were washed twice with HMEM. The heparinase was then added in HMEM and incubated with the cells for 2.5 h at 37°C, at a final concentration of 30 units/ml. VN-TAT Degradation Assay-HepG2 cells were grown in 24-well culture dishes to 80% confluency and washed in HBSB. 125 I-VN-TAT, at a final concentration of 10 nM, was then incubated with the HepG2 cells, in the presence or absence of competitors at 37°C, in HMEM containing 0.5% BSA. The competitors used were heparin (100 units/ml final concentration), chloroquine (100 M final concentration), or ␣ 2 M* (775 nM final concentration). 125 I-VN-TAT in reaction buffer was added to the wells of 24-well culture dishes that had been preincubated with ␣-MEM containing 10% fetal calf serum and penicillin-streptomycin, to control for non-cell-mediated protein degradation over the experimental time periods. At the appropriate time points (0, 1, 4, and 17 h) 100 l of the reaction medium was added to 100 l of 20% trichloroacetic acid on ice for 15 min and then centrifuged at 16,000 ϫ g for 10 min. To determine true cell-mediated degradation of the proteins, the total trichloroacetic acid-soluble counts from the control wells were subtracted from the trichloroacetic acid-soluble counts in the experimental wells.

RESULTS
Purification and Radiolabeling of Human VN-TAT-In order to investigate the metabolism of VN-TAT in vitro and in vivo, VN-TAT was purified by heparin-Sepharose chromatography. Under nonreducing conditions, the VN-TAT was seen as an ϳ160-kDa monomer but also as higher molecular weight multimers. When VN-TAT was electrophoresed under reducing conditions, it appeared to break down to form three new bands, one of 97 kDa that co-migrated with TAT and two additional bands of mobility intermediate to TAT and AT, which most likely represent the two forms of VN found in plasma. This observation is consistent with the concept that VN-TAT formation is through a disulfide interchange between VN and the thrombin in the TAT complex, as previously reported (21).
Composition of Purified VN-TAT-To demonstrate that the purified VN-TAT was composed of AT, VN, and thrombin, immunoblotting analyses were performed. Fig. 1 shows the immunological detection of VN (lane 1), AT (lane 2), and thrombin (lane 4). All three proteins were detected in the major band of ϳ160 kDa. Higher molecular weight multimers composed of VN, AT, and thrombin were also detected. Interestingly, a minor lower molecular weight band was also detected, which, from the Coomassie Blue-stained SDS-PAGE gels, corresponds to a molecular mass of ϳ60 kDa. This band most likely represents degradation of the ternary complex, since antibodies against VN (lane 1), AT (lane 2), and thrombin (lane 4) appeared to recognize it. Since VN can form ternary complexes with thrombin-HCII binary complexes, Western blot analysis was performed to determine whether these ternary complexes were present in the VN-TAT preparation. VN-thrombin-HCII complexes were detected and, like VN-TAT, appeared to be composed of multimers, with the majority of the complexes present in the monomeric form (Fig. 1, lane 3). The monomers appeared to have a slightly slower electrophoretic mobility than VN-TAT, which is most likely due to the fact than HCII is a larger protein than AT (Fig. 1, lane 3). Furthermore, there appeared to be more degradation products with VN-thrombin-HCII than seen with VN-TAT, which are of higher molecular mass than the VN-TAT degradation products. These lower molecular mass bands are most likely VN-thrombin-HCII degradation products, since they appear also to contain thrombin and VN. Enzyme-linked immunosorbent assay analysis of the purified VN-TAT preparation revealed that VN-HCII-thrombin complexes were a minor contaminant, since the preparation contained an approximately 10-fold molar excess of AT over HCII (data not shown).
Plasma Elimination of VN-TAT, TAT, and AT-125 I-VN-TAT plasma clearance experiments were done in rabbits to determine whether VN-TAT is metabolized in a fashion similar to TAT. The mean of trichloroacetic acid-corrected 125 I-VN-TAT plasma clearance curves is shown in Fig. 2. VN-TAT was removed rapidly from the circulation with a t1 ⁄2␣ ϭ 3.8 Ϯ 0.5 min and had a t1 ⁄2␤ ϭ 85.4 Ϯ 4.3 min. SDS-PAGE analysis showed that VN-TAT remained intact within the circulation over the entire period of the experiment (data not shown). Fig. 2 also shows the clearance curves of human TAT and human AT. AT clearance was slow, with its disappearance from the plasma being linear over the entire experimental time frame. As expected, the clearance of AT was much slower than that of VN-TAT and TAT, with only ϳ25% of the initially injected 125 I-AT cleared at the 1-h point. Both VN-TAT and TAT had a biphasic clearance with a rapid phase of initial clearance. TAT had a slightly higher t1 ⁄2␣ (5.3 Ϯ 0.32 min) than VN-TAT (3.8 Ϯ 0.5 min), although it is not significantly different (p ϭ 0.088). TAT had a slower second clearance phase with a half-life (t1 ⁄2␤ ) of 42.8 Ϯ 3.7 min.
Tissue Distribution of 125 I-VN-TAT following Clearance-To determine the organ localization of VN-TAT removal, 125 I-VN-TAT was injected into rabbits (n ϭ 3), and after 1 h the postmortem determination of radioactivity within different tissues was done. The organs examined were the liver, lungs, kidneys, spleen, bladder, and heart. Over 80% of the total organ-associated radioactivity was liver-associated (see Fig. 3A). The kidney contained the next highest fraction of radioactivity (ϳ10% of the total). Examining the amount of radioactivity/g of tissue showed that the liver had the highest specific activity (ϳ20,000 cpm/g) (Fig. 3B). Relatively high counts were found in the kidneys and spleen (ϳ14,000 cpm/g and ϳ15,000 cpm/g, respectively).
Heparin Sensitivity of VN-TAT Clearance-VN-TAT has been reported to interact with a heparin-like substance on the surface of human umbilical vein endothelial cells (HUVECs) (28). To determine whether the binding of VN-TAT to liver receptors was also heparin-sensitive, the plasma clearance of VN-TAT was observed under competition conditions using standard heparin. Fig. 4A shows the clearance curve of 125 I-VN-TAT co-injected with a bolus of 10,000 units of heparin, with the previous 125 I-VN-TAT clearance (without heparin) experiments as a reference. In the presence of heparin, the clearance data best approximated a single-compartment model, in which the initial rapid clearance phase with VN-TAT alone was abolished. The determined t1 ⁄2 from these experiments was 61.8 Ϯ 3.7 min, which was significantly different from the t1 ⁄2␤ for VN-TAT alone of 85.4 Ϯ 4.3 min (p ϭ 0.002). The t1 ⁄2 of initially injected 125 I-VN-TAT to clear from rabbits without heparin was ϳ10 min.
In further experiments, protamine sulfate was co-injected (final concentration, ϳ500 g/ml blood) with the 125 I-VN-TAT. Like heparin, protamine sulfate abolished the ␣-phase of VN-TAT clearance and reduced the clearance pattern to a monophasic one. From Fig. 4A, it can be seen that protamine, like heparin, decreased the clearance rate of VN-TAT in the plasma by approximately 5-fold compared with VN-TAT alone, with half of the total dose injected being cleared by 49.4 Ϯ 4.4 min.
To further determine the effects of heparin on 125 I-VN-TAT clearance, varying doses of heparin were used. Fig. 4B demonstrates the clearance of 125 I-VN-TAT in the presence of 10,000 (taken from Fig. 4A), 5,000, and 1,000 units of heparin or of saline control. In the presence of 10,000 units of heparin, 125 I-VN-TAT clearance is monophasic. However, with decreasing doses of competing heparin, the clearance becomes more biphasic, such that the more rapid ␣-phase of clearance reappears, similar to that seen with 125 I-VN-TAT alone.
To determine whether heparin treatment affected the liverspecific clearance of VN-TAT, experiments were done in the absence and presence of heparin, with the amount of radioactivity within the livers examined at 30 min (n ϭ 3) (data not shown). This time was chosen to maximize differences in VN-TAT removal. Comparing the amount of 125 I-VN-TAT in the livers of the two groups revealed that ϳ66% of cleared 125 I-VN-TAT was found in the rabbit livers (n ϭ 3). This clearance was reduced to only ϳ39% in the heparin-treated animals (n ϭ 2). Furthermore, there was 50% less radioactivity/g of liver tissue in the heparin-treated animals. Taken together, these data support the hypothesis that the liver is the major site for VN-TAT removal and that heparin-like materials are important for VN-TAT clearance.
In Vivo Internalization and Degradation of VN-TAT- Fig.  5A shows the appearance of trichloroacetic acid-soluble radioactivity in the plasmas of the rabbits injected with 125 I-VN-TAT. There was an initial decrease in the trichloroacetic acidsoluble plasma 125 I products, which most likely represents the clearance of free 125 I from the blood. At 20 min, the soluble plasma radioactivity began to increase, and it peaked at 60 min at a level 3 times greater than that seen initially in the plasma. This peak in appearance of 125 I degradation products was maintained for up to 120 min. The degradation of 125 I-VN-TAT co-injected with either heparin or protamine is shown in Fig.  5B. The initial levels of trichloroacetic acid-soluble counts were very similar for up to 20 min for all three experimental groups and probably represent the clearance of free 125 I. However, in the presence of protamine or heparin, the appearance of trichloroacetic acid-soluble degradation products barely exceeded (protamine) or did not exceed (heparin) the initial plasma levels of trichloroacetic acid-soluble products. The highest degradation levels were reduced Ͼ2.5-fold for VN-TAT competed with protamine or heparin compared with that seen with VN-TAT alone.
Effects of Heparin on 125 I-TAT Clearance-125 I-TAT clearance was determined in the absence or presence of 10,000 units of heparin (Fig. 6A). Similar to that seen with 125 I-VN-TAT clearance, the ␣-phase of 125 I-TAT clearance was greatly reduced in the presence of heparin. To determine if 125 I-TAT complexes remained intact during clearance, plasma samples were analyzed by nonreducing SDS-PAGE (Fig. 6B). The 125 I-TAT complexes injected with saline remained intact and were cleared rapidly. However, the 125 I-TAT complexes in the heparin competition experiments partitioned as normal TAT complexes and as a high molecular mass adduct that appeared immediately (1 min time point) and persisted throughout the experiment. Furthermore, the disappearance of the two labeled species was reduced in comparison with the rapid clearance seen with 125 I-TAT alone.
Radioligand Binding Studies with VN-TAT on Hepatoma Cells-To determine the affinity of VN-TAT for HepG2 cells, a competitive radioligand binding experiment was performed. A competition curve from the data was obtained giving an IC 50 value of ϳ1 M (Fig. 7A), resulting in an apparent K d of ϳ1 M. In further competitive radioligand binding experiments, the ability of VN-TAT, TAT, and ␣ 2 M* to compete for binding with 125 I-VN-TAT was compared. Fig. 7B shows a single concentration competition binding experiment with HepG2 cells, using 10 nM 125 I-VN-TAT in the absence or presence of a 240-fold molar excess of TAT or VN-TAT and a 76-fold excess of ␣ 2 M*. Excess VN-TAT effectively competed 125 I-VN-TAT binding, reducing its binding by 70%. However, TAT and ␣ 2 M* competed only modestly with VN-TAT binding, reducing its binding by only 25 and 20%, respectively. Competition with heparin resulted in an ϳ60% reduction in VN-TAT binding, compared with that seen for 125 I-VN-TAT alone.
VN-TAT Binds to Heparan Sulfate Proteoglycans on HepG2 Cells-To determine whether HSPGs might be involved in VN-TAT binding, radioligand binding experiments were performed on HepG2 cells in which either cellular HSPGs were removed or their synthesis was inhibited. Thus, the enzyme heparinase was incubated with HepG2 cells to see if this would reduce 125 I-VN-TAT binding. Heparinase is an enzyme that degrades heparin and heparan sulfate proteoglycans. Heparinase treatment of HepG2 cells reduced 125 I-VN-TAT binding by ϳ46% and is similar to the degree of inhibition observed by competition with a 30-fold molar excess of VN-TAT (Fig. 7B). To further demonstrate that cell-synthesized proteoglycans are important for VN-TAT binding, HepG2 cells were treated with ␤-D-xyloside, an inhibitor of glycosaminoglycan attachment to proteoglycans. Such treatment resulted in a 53% decrease in 125 I-VN-TAT binding (Fig. 7B). In all, the ability of competing heparin, heparinase treatment, and the reduction of glycosaminoglycan containing proteoglycans to decrease VN-TAT binding supports the hypothesis that heparinoid substances are important in VN-TAT binding to hepatic cells, both in vitro and in vivo.
Heparin Inhibits the Degradation of VN-TAT by HepG2 Cells-Different SECs have been found to be internalized and degraded after binding to hepatic receptors. To determine if VN-TAT is also internalized and degraded and what effect heparin might have on this process, competitive internalization experiments for 125 I-VN-TAT were performed. Fig. 8 shows the amount of 125 I-VN-TAT degraded over time, in the absence or presence of heparin, chloroquine, or ␣ 2 M*. At 17 h, VN-TAT was maximally degraded, with 149 fmol having been degraded (ϳ25 ng). To show that the trichloroacetic acid-soluble counts were truly from receptor-mediated degradation, chloroquine, an inhibitor of lysosomal degradation, was used. In the presence of 200 M chloroquine, there were negligible amounts of trichloroacetic acid-soluble products, such that the levels were very similar to those seen in control wells (autodegradation of VN-TAT seen in culture wells without cells). VN-TAT degradation was greatly reduced in the presence of heparin (100 units/ml) to only 33 fmol (ϳ5 ng) at 4 h. This level was maintained for over 17 h. To determine whether LRP might be involved in VN-TAT degradation, an excess of ␣ 2 M* (77.5ϫ; 775 nM) was used. The presence of ␣ 2 M* had no effect on VN-TAT internalization and degradation. As a positive control for the internalization experiment, 125 I-␣ 2 M* was used. The latter is known to be internalized by hepatocytes via LRP. The ␣ 2 M* was internalized to a lesser degree than that seen with VN-TAT, with only 43 fmol being degraded over 17 h (data not shown).

DISCUSSION
The impetus for these experiments was the finding that TAT is invariably found as a ternary complex with VN in human plasma (20,21). Since the metabolism of the binary TAT complex has been well studied (2, 3), we were interested in determining whether the addition of VN to TAT, to form a ternary complex, had any effect on TAT metabolism. The present studies indicate that VN-TAT, like TAT, is cleared in a rapid biphasic fashion mediated by hepatic receptors. This biphasic mode of clearance is reduced to a single phase in the presence of large doses of heparin or protamine, which suggests that the initial rapid clearance of VN-TAT is mediated through binding to liver HSPGs. Concurrent with the decreased clearance of VN-TAT in the presence of heparin or protamine is the reduction in both the liver uptake and the degradation of VN-TAT. It is noteworthy that 125 I-TFPI clearance in mice was shown to be inhibited similarly by preinjection with protamine sulfate, presumably by protamine blocking 125 I-TFPI interaction with HSPG (33). The physiological relevance of VN in TAT clearance is supported by the formation of a high molecular mass adduct, most likely VN-TAT, seen during the heparin competition clearance experiments with 125 I-TAT. These data indicate that TAT must first form ternary VN-TAT complexes before being cleared. These ternary complexes then bind to HSPGs before being passed onto other cellular clearance receptors. This hypothesis is supported by the fact that the high molecular mass 125 I-TAT adduct appeared in the plasma only in the presence of competitor heparin.
The suggestion that HSPGs are cellular receptors for VN-TAT in vivo is supported by the in vitro cell culture experiments with the human hepatoma cell line, HepG2. These experiments show that VN-TAT binding to HepG2 cells can be competed by the presence of excess heparin, reduced by the enzymatic removal of HSPGs by heparinase, and reduced when HSPG synthesis is inhibited by ␤-D-xyloside. These effects are consistent with the previously reported results obtained for VN-TAT binding to HUVEC proteoglycans (28). However, VN-TAT was not internalized and degraded by HUVECs; rather, it appeared to be translocated to the subendothelium intact (34). In contrast, VN-TAT was degraded, in a heparin-sensitive fash- ion, in vivo and by HepG2 cells in vitro. This data, in conjunction with the liver-specific accumulation of 125 I-VN-TAT, suggests that hepatocytes and not endothelial cells play a major role in VN-TAT metabolism.
From competitive radioligand binding experiments, the affinity of VN-TAT for hepatoma cells was found to be low (K d of ϳ1 M). This is substantially lower than that found for VN-TAT binding to HUVECs (K d ϭ 16 nM) (28). This low affinity could be due to the multimeric nature of the VN-TAT, which could result from different relative affinities of different moieties (i.e. monomeric versus dimeric versus trimeric, etc.). Relevantly, a low affinity thrombospondin 1-proteoglycan interaction has been described, with a K d of 289 nM (35). Both monomeric VN and multimeric VN (mVN; mass ϭ 420,000 Da) have been demonstrated to bind heparin with low affinity (K d values of 5 M and 200 nM, respectively) (36), similar to the low affinity binding of VN-TAT to HSPG observed in the present studies. mVN is believed to be in a similar conformation to VN in VN-TAT, since antibodies recognize newly exposed epitopes in these forms of VN, but not to those found in native monomeric VN (37,38). However, care must be taken in comparing mVN with VN-TAT, since mVN showed differences in clearance in rabbits (39), compared with the present findings, and in sensitivity to heparin binding to HUVECs (40) compared with that reported for VN-TAT previously (28).
Various SECs, including TAT, have been reported to bind to a number of candidate hepatic receptors including LRP (7), the SEC receptor (5,6), and CK18 (11). In addition to TAT, LRP binds various thrombin-serpin complexes, including thrombin-PAI-1 (9) and thrombin-PN-1 (41,42). However, recent evidence suggests that LRP is most likely the vehicle of internalization for SECs rather than the primary cellular binding site (4,11,(41)(42)(43). In the present studies, the LRP ligands TAT and ␣ 2 M* did not compete effectively for VN-TAT binding to HepG2 cells. However, ␣ 2 M* has been found not to compete with other known LRP binding ligands (44,45). The mild inhibition by TAT probably reflects the need for TAT to form ternary complexes with VN to compete for VN-TAT binding. This fact is supported by the in vivo 125 I-TAT clearance experiments in the presence of heparin. Furthermore, de Boer et al. (28) showed that TAT inhibition of 125 I VN-TAT binding to HUVECs was through the formation of VN-TAT. Together, these data suggest that VN contains important receptor-binding residues, and in this regard, VN has been proposed to promote the LRP-mediated clearance of thrombin-PAI-1 complexes (9). Additionally, VN-mediated clearance through LRP is supported by the fact that the heparin-binding domain of VN has been found to bind to a highly conserved sequence in complement proteins and perforin (14), which is not only found in LRP but is involved in LRP ligand binding (46,47).
In recent years, it has become apparent that cell surface proteoglycans have important biological roles in regulating the cellular binding of ligands and their internalization. Thus, proteoglycans have been reported to be involved in the cellular binding of TFPI (33), lipoprotein lipase (48), growth factors (49,50), hepatic lipase (51), thrombospondin-1 and 2 (52)(53)(54), and, more recently, thrombin-PN-1 (42). The presumed role for the proteoglycans in these interactions is that they serve as the initial binding sites for the various ligands, which can then be presented to various cognate cellular receptors. For example, proteoglycans probably bind thrombin-PN-1 (42), thrombospondin (52), TFPI (33,55), and lipoprotein lipase (56,57) and then present them to LRP for internalization.
For the first time, the in vivo metabolism of VN-TAT has been examined. These studies reveal that VN-TAT is cleared, at least in part, by HSPG. What other proteins may be involved in VN-TAT clearance remains to be elucidated. However, these studies shed light on the metabolism of TAT in ternary complex with VN. It will be interesting to see what other role(s) VN plays in the clearance of TAT and other SECs.