C1 Inhibitor-C1s Complexes Are Internalized and Degraded by the Low Density Lipoprotein Receptor-related Protein*

Like other serpin-enzyme complexes (SECs), proteinase-complexed C1 inhibitor (C1-INH) is rapidly cleared from the circulation and thought to be a neutrophil chemoattractant, suggesting that complex formation causes structural rearrangements exposing a domain which is recognized by specific cell surface receptors. However, the cellular receptor(s) responsible for the catabolism and potential mediation of chemotaxis by C1-INH-protease complexes remained obscure. To determine whether the SEC receptor mediates the binding and potential chemotaxis of C1-INH·C1̄s, we performed binding assays with HepG2 cells, neutrophils, and monocytes, and the results show that C1-INH·C1̄s neither bind to these cells nor cause a chemotactic response of neutrophils and monocytes. Furthermore, C1-INH·C1̄s, the COOH-terminal C1 inhibitor peptide, or the tetrameric C1-INH·C1̄s·C1̄r·C1-INH complex were found to be significantly less effective in competing with the SEC receptor ligand 125I-peptide 105Y for the binding to HepG2 cells than unlabeled 105Y, indicating that the SEC receptor does not sufficiently recognize C1-INH-protease complexes. The asialoglycoprotein receptor was also ruled out to be responsible for the removal of the heavily glycosylated C1-INH·C1̄s complex, since asialoorosomucoid did not compete for the clearance of C1-INH·125I-C1̄s and asialoglycoprotein receptor knockout mice showed no alterations in the C1-INH·125I-C1̄s clearance rate. We found that C1-INH·125I-C1̄s complexes were efficiently degraded by normal murine fibroblasts expressing the low density lipoprotein receptor-related protein (LRP) and cellular degradation was significantly reduced by chloroquine and the receptor-associated protein, which is a potent inhibitor of the binding of all known ligands to LRP. Moreover, receptor-associated protein inhibited thein vivo clearance of C1-INH·125I-C1̄s and murine fibroblasts genetically deficient for LRP did not degrade C1-INH·125I-C1̄s. Our results demonstrate that C1-INH·C1̄s complexes do not stimulate neutrophil or monocytic chemotaxis but are removed by LRP, further underscoring its role as a serpin-enzyme complex clearance receptor.

C1 inhibitor (C1-INH) 1 is the only known plasma inhibitor of C1 s and C1 r, the activated homologous serine proteases of the first component of complement (C1) (1). It is also one of the major inhibitors of plasma kallikrein and factor XIIa. Thus, C1-INH plays an essential role as a regulator of the activation of the classical complement pathway and the contact system, and genetic or acquired deficiencies of C1-INH result in recurrent episodes of angioedema (2). C1-INH is a member of the superfamily of serine proteinase inhibitors (serpins) and one of the most heavily glycosylated plasma proteins (3). Like other serpins, C1-INH inhibits target proteases by forming a stable covalent serpin-enzyme complex (SEC), resulting in the cleavage of the P1-P1Ј peptide bond of the reactive center loop of the inhibitor and in the generation of a COOH-terminal peptide of M r 4152 (4,5). The formation of a SEC is accompanied with a large structural rearrangement of the serpin, and it has been concluded that this reveals a receptor recognition site (6). Thus, a crucial function of serpins is to provide a tag for the removal of proteolytic enzymes from the circulation by cellular receptors, thereby preventing excessive and harmful proteolysis of blood and tissue proteins (7). In general, SECs are cleared much more rapidly than are the proteolytically inactivated or native serpins (6). Although the data of Mast et al. (6) have indicated that the target protease does not play a role for the clearance of SECs, recent findings of de Smet et al. (8) and of Malek et al. (9) have shown that the clearance rates of different C1-INH-protease complexes significantly differ. They have found that the turnover was most rapid for C1-INH⅐C1 s, followed by C1-INH-kallikrein and C1-INH-XIIa, suggesting a direct role of the protease in determining the clearance receptor affinity. However, the receptor(s) involved in the catabolism of C1-INH-protease complexes has not been identified. Different types of receptors have been proposed for the binding and uptake of SECs. HepG2 cells, monocytes, and neutrophils have been reported to express a not yet completely characterized SEC receptor, which binds, internalizes, and degrades several serpin-protease complexes, including those with ␣ 1 -antitrypsin (␣ 1 -AT), ␣ 1 -antichymotrypsin (␣ 1 -ACT), antithrombin III (ATIII), and, to a much lesser extent, C1-INH⅐C1 s (10). Furthermore, the SEC receptor has been shown to mediate the increase in synthesis of ␣ 1 -AT and elicits neutrophil chemotaxis after binding of ␣ 1 -AT-elastase (11). It has been proposed that the SEC receptor recognizes a specific pentapeptide sequence highly conserved among serpins, which is located carboxyl-terminal to the reactive center loop of the serpins (residues 370 -374 of ␣ 1 -AT) and is exposed after complex formation. This is based on studies using analogous peptides (e.g. peptide 105Y corresponding to residues 359 -374 of ␣ 1 -AT), which block the binding of SECs to HepG2 cells in a sequencespecific manner and by the finding that such peptides bind specifically and saturably to neutrophils, monocytes, and HepG2 cells, thereby stimulating the biosynthesis of ␣ 1 -AT and the chemotaxis of neutrophils (PMN) (11)(12)(13). A second SEC receptor has been proposed, termed SEC II receptor, which mediates the clearance of ␣ 2 -antiplasmin (␣ 2 -AP)-protease complexes. This has been concluded by the finding that ␣ 2 -APprotease complexes do not compete for the catabolic turnover of ␣ 1 -AT-, ␣ 1 -ACT-, and ATIII-proteinase complexes by the SEC receptor (6). Recent studies have demonstrated the role of the low density lipoprotein receptor-related protein (LRP) in the constitutive clearance of SECs (14 -16). LRP is a large multifunctional endocytic receptor, which is highly expressed in the liver. LRP mediates the cellular uptake of several unrelated ligands, including lipoproteins, lipoprotein lipase, proteinaseor methylamine-activated ␣ 2 -macroglobulin (␣ 2 -M*), proteases, SECs, bacterial toxins, and lactoferrin (reviewed in Ref. 17). Some of these ligands can compete with each other for the binding to LRP, whereas others bind to independent sites on the receptor (18). The expression of LRP is essential for embryogenesis, because embryos with a targeted disruption of the LRP gene die early in utero (19). The functional activity of LRP is controlled in vivo by a receptor-associated protein (RAP) of 39 kDa, which has recently been shown to act as a chaperone in the biosynthesis and maturation of the receptor (20). RAP blocks the binding of all known ligands to LRP and copurifies with the receptor (21)(22)(23).
In the present study, we have investigated the potential chemotactic activity and clearance of C1-INH⅐C1 s, and our results demonstrate that C1-INH⅐C1 s (i) does not stimulate neutrophil or monocytic chemotaxis, but (ii) is removed from the circulation by LRP.
Antibodies-Different rabbit polyclonal antibodies against hu C1-INH were obtained from Behring (Marburg, Germany). Monoclonal antibody (mAb) 13E1 ␣-hu C1-INH, mAb 303D4 ␣-hu C1 s, and mAb 98F12 ␣-hu C1 r have been described elsewhere (26). A polyclonal antibody against a peptide derived from the carboxyl-terminal end of trypsinized C1-INH was raised in a New Zealand White rabbit. The antibody was purified on protein G (Pharmacia Biotech Inc.) and specific antibody titers were determined by an enzyme-linked immunosorbent assay (ELISA).
Serpin-Enzyme Complex Preparation and Quantitation-C1-INH⅐C1 s complexes were formed as described previously (27). Complex formation was confirmed by SDS-PAGE and ELISA as described (26). Pure C1-INH⅐C1 s complexes were isolated by anion exchange chromatography on a Hydrobore AX column (Rainin, MA) fitted to an HPLC system (Gilson-Abimed, Langenfeld, Germany) applying a linear gradient of buffer A (10 mM sodium phosphate, pH 7.0) and buffer B (10 mM sodium phosphate, 0.3 M NaCl, pH 7.0). Tetrameric C1-INH⅐C1 s⅐C1 r⅐C1-INH complexes were prepared using a procedure described by Ziccardi and Cooper (28). Complexes of ␣ 1 -AT-trypsin or ␣ 1 -AT-elastase were prepared and purified as described (29). Serpinenzyme complexes were quantitated with the Micro BCA protein assay kit (Pierce). The amount of complexes in preparations containing residual unreacted serpin was determined by densitometry scanning of SDS-polyacrylamide gels (Herolab Densitometer, Herolab, Wiesloch, Germany).
Purification of the COOH-terminal Derived C1-INH Peptide-The cleavage of C1-INH by trypsin and the purification of the released COOHterminal C1-INH peptide were performed essentially as detailed previously (30). The peptide was subjected to NH 2 -terminal sequence analysis on a 473A protein sequencer (Applied Biosystems) and quantitated by amino acid hydrolysis (420A derivatizer analyzer system; Applied Biosystems). The COOH-terminal C1-INH peptide was stored at Ϫ20°C.
Removal of Sialic Acid from C1-INH and Orosomucoid-Desialylation of C1-INH was performed as described (33), and the preparation of ASOR has been described elsewhere (25). Release of sialic acid from C1-INH and orosomucoid was verified with sialic acid-specific lectin Sambucus nigra agglutinin (SNA) and Maackia amurensis agglutinin (MAA) using the DIG glycan differentiation kit (Boehringer Mannheim, Germany) according to the manufacturer's instructions.
Cells-A normal mouse fibroblast line (MEF-1, ATCC CRL-2214) and a mouse fibroblast cell line with a homozygous LRP deficiency (PEA-13, ATCC CRL-2216) were obtained from the American Type Culture Collection (ATCC, Rockville, MD) and maintained as described (34). HepG2 cells (ATCC HB-8065) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum and 2 mM Lglutamine. HepG2 cells used for flow cytometric analysis were removed from flasks using PBS with 1 mM EDTA. Cells were then washed in 0.1% BSA-PBS supplemented with 1 mM CaCl 2 . Primary rat hepatocytes were isolated and cultured as detailed previously (29). PMN and monocytes used for flow cytometric and radioreceptor binding experiments were prepared, cultured, and harvested as described previously (26).
Chemotaxis-Chemotactic activity was determined in modified Boyden chambers by the micropore-filter leading front assay essentially as described by Brenneis et al. (35) and in a 48-well chemotaxis microchamber (Neuroprobe, Rockville, MD) according to the methodology of Brandes et al. (36). Chemotaxis was evaluated using different dilutions of C1-INH, purified C1-INH⅐C1 s complexes, and the COOH-terminal derived C1-INH peptide. Medium alone served as a negative control, and FMLP (10 Ϫ8 M; Sigma) or zymosan-activated serum provided a positive control. In competitive experiments, PMN were incubated with either 10 nM ␣ 1 -AT-elastase complexes or 10 nM purified C1-INH⅐C1 s complexes for 15 min at 37°C, and then washed three times with medium. The cells were then assayed for a chemotactic response to 1 nM ␣ 1 -AT-elastase complexes. Controls were performed simultaneously with medium alone or with 10 nM FMLP.
Binding and Uptake Assays with 125 I-Labeled Ligands-Binding experiments with C1-INH⅐ 125 I-C1 s complexes were performed at 4°C as described previously (27). Cross-competition binding studies using peptide 125 I-105Y and unlabeled peptide 105Y, C1-INH⅐C1 s complexes, C1-INH⅐C1 s⅐C1 r⅐C1-INH complexes, and the COOH-terminal C1-INH peptide as competitors were carried out essentially as described by Joslin et al. (11). For binding studies with monocytes and PMN, triplicate samples of 1 ϫ 10 5 cells/tube in HBSS, 0.1% BSA were incubated for 2 h at 4°C with varying amounts of radiolabeled C1-INH⅐C1 s complexes (5-50 nM) in a final volume of 150 l. Unbound ligand was removed by layering 100 l of the samples on 300 l of 30% (w/v) sucrose in an Eppendorf microcentrifuge for subsequent centrifugation. After spinning for 1.5 min, the tips of the tubes containing the cell pellets were cut off for counting radioactivity in a ␥ counter (Canberra-Packard). Nonspecific binding was determined in the presence of a 50-fold molar excess of unlabeled complexes (reaction mixture). Uptake assays with primary rat hepatocytes and C1-INH⅐ 125 I-C1 s were performed as described previously (29).
Flow Cytometry-The incubation of PMN, monocytes, and HepG2 cells with purified C1-INH⅐C1 s and the flow cytometric analysis was conducted as described (26). The potential effect of cell activation on binding of C1-INH⅐C1 s complexes to PMN or monocytes was determined by preincubation of the cells (2 ϫ 10 6 /ml) with 100 nM FMLP for up to 30 min at 37°C or with 20 ng/ml phorbol 12-myristate 13-acetate for 15 min at 37°C. Cells were washed and chilled to 4°C, after which flow cytometric analysis was performed. Control cells were brought to 37°C for the same time in the absence of stimulus. Cell activation was controlled employing mAb ␣-CD35 (Dakopatts). In competition assays using fluorescein isothiocyanate (FITC)-conjugated ␣ 1 -AT-trypsin complexes (FITC-␣ 1 -AT-trypsin), PMN were incubated for 2 h at 4°C with 1 M FITC-␣ 1 -AT-trypsin alone or in the presence of increasing molar excesses (0.5-10-fold) of unlabeled ␣ 1 -AT-trypsin or C1-INH⅐C1 s complexes (reaction mixture). Cells were washed, and the cell-associated fluorescence was analyzed by flow cytometry.
Cellular Degradation of 125 I-Labeled Ligands-HepG2, MEF-1 and PEA-13 cells were seeded into 12-well plates and grown for 24 h in DMEM, 10% fetal calf serum, and 2 mM glutamine prior to the addition of 125 I-labeled ligands. The medium was then replaced by DMEM (without glutamine) containing 0.2% (w/v) BSA and the indicated iodinated ligands. Cellular degradation of 125 I-labeled proteins was measured as described previously (37) and is expressed as nanograms of 125 I-labeled trichloroacetic acid-soluble (non-iodide) material released into the culture medium per indicated number of cells.
Clearance and Tissue Distribution of C1-INH⅐ 125 I-C1 s-0.1-1 g of C1-INH⅐ 125 I-C1 s or asialo-C1-INH⅐ 125 I-C1 s in sterile PBS (total volume 100 l) were injected within 10 s in the lateral tail vein of 12-16-weekold BALB/c mice (weighing 20 -25 g) and in 15-week-old female ASGPR knockout mice in the absence or presence of competitors. At the indicated time points, blood samples were collected by retro-orbital bleeding of the anesthetized mice, and the amount of trichloroacetic acid-precipitable radioactivity in 20 l of plasma was determined as described (34). For tissue distribution studies, the abdomen of anesthetized mice were cut open and a cannula was inserted into the portal vein. Livers were perfused with 5 ml of PBS, 0.1% BSA to remove the blood-associated radioactivity. Subsequently, the major organs were removed, weighed, and assessed for radioactivity.

Physiological Association of the COOH-terminal C1-INH
Peptide with Complexed and Cleaved C1-INH-We have shown in a previous study that the 34-amino acid COOH-terminal C1-INH peptide remains tightly bound to the reactive centercleaved C1-INH, which was affinity-purified from the plasma of patients suffering from acquired angioedema (type II) (38). To test whether this COOH-terminal fragment remains also associated with the C1-INH⅐C1 s complex, the complex was purified under physiological conditions using aqueous buffers at neutral pH to remove any uncomplexed components. The purified complex was then analyzed by nonreducing Tricine-SDS-PAGE and by Western blots. Silver staining of the gels and detection of the blots using a polyclonal antibody against the COOHterminal peptide revealed that the 4.1-kDa COOH-terminal C1-INH peptide remained non-covalently associated with the C1-INH⅐C1 s complex (Fig. 1).
Chemotaxis Assays with C1-INH⅐C1 s Complexes and the COOH-terminal C1-INH Peptide-Chemotaxis assays were performed to determine whether the inhibitory complex of C1-INH and C1 s or the COOH-terminal C1-INH peptide is chemotactic for PMN and monocytes, as has previously been postulated (39). The COOH-terminal C1-INH peptide was isolated from trypsinized C1-INH by reversed-phase HPLC, where two major peaks could be detected (Fig. 2B). The peptide contained in peak 2 ( Fig. 2B) was subjected to NH 2 -terminal sequencing, which confirmed that it was derived from the cleavage between the P1 arginine and P1Ј threonine residues of the reactive center loop. The concentration of the 4.1-kDa C1-INH peptide was determined by amino acid hydrolysis. Dilutions of the 4.1-kDa C1-INH peptide, the purified C1-INH⅐C1 s complex, as well as native C1-INH were assayed for their ability to stimu- late migration of PMN in modified Boyden chambers or in a 48-well chemotaxis microchamber. As shown in Fig. 3A, there was no significant chemotactic migration of PMN toward the probes in modified Boyden chambers when compared with control medium (HBSS, 0.5% BSA). There was also no chemotactic activity detectable when the samples were assayed in 48-well chemotaxis microchambers using either PMN (data not shown) or monocytes (Fig. 3B). This was not due to a general impairment of the chemotactic activity of the cells, since diluted zymosan-activated serum (in modified Boyden chamber assays) or 10 Ϫ8 M FMLP (in 48-well microchambers) caused a significant chemotactic response of the cells (Fig. 3, A and B). Furthermore, preincubation of PMN with purified C1-INH⅐C1 s failed to suppress the chemotactic activity toward ␣ 1 -AT-elastase (Fig. 3C), which has been shown previously to be a potent chemoattractant (40). In contrast, preincubation of the cells with ␣ 1 -AT-elastase markedly reduced the number of migrated PMN toward 1 nM ␣ 1 -AT-elastase in the lower wells of the chamber (Fig. 3C). These results indicate that the COOH-terminal C1-INH peptide or the C1-INH⅐C1s complex do not cause chemotaxis of PMN and monocytes and thus are not pro-inflammatory.
Tissue Distribution of C1-INH⅐C1 s-Complexes of C1-INH⅐ 125 I-C1 s were formed immediately after the radioiodination of the protease as described under "Experimental Procedures" and were analyzed by SDS-PAGE and autoradiography. Fig. 4 (lane 3) shows that the majority of the 125 I-labeled protease was converted to an SDS-stable complex of the appropriate molecular mass (ϳ190 kDa), and quantitation of the dried SDS-polyacrylamide gel using a radioimager revealed that only 8% of 125 I-C1 s was uncomplexed. To identify the tissues responsible for C1-INH⅐C1 s clearance, the C1-INH⅐ 125 I-C1 s complex was injected into mice and the radioactivity in major organs was determined. Autopsies at 30 min revealed that about 70% of recovered radioactivity was localized to the liver (Table I), and clearance studies showed that about 32% of C1-INH⅐ 125 I-C1 s remained in the plasma at that time (Fig. 5, A  and C). These findings suggest that hepatic receptors are mainly involved in the clearance of the complex.
Probing the SEC Receptor for the Binding and Uptake of C1-INH⅐C1 s Complexes-To assess the role of the SEC receptor for the clearance of C1-INH⅐C1 s, binding studies were performed on HepG2 cells, PMN, and monocytes, which have been shown to express the SEC receptor (10,13). Using a concentration range of 5-50 nM C1-INH⅐ 125 I-C1 s and an 50-fold molar excess of unlabeled complex, no saturable and specific binding to HepG2 cells could be observed (data not shown). To test  A and B). A, the leading front is the distance from top of the filter to the level of the filter where at least five PMN/field could be found. Zymosan-activated serum (Act. NHS) provided a positive control, and HBSS plus 0.1% BSA (HBSS) a negative control. B, chemotaxis is defined as the number of monocytes per field that migrated through the filter pores and is expressed as the percentage of the maximal migration to FMLP (183.7 Ϯ 20.7/high power field). RPMI, 0.1% BSA (RPMI) was used as a negative control. C, chemotaxis of PMN to ␣ 1 -AT-elastase after exposure to excess C1-INH⅐C1 s, ␣ 1 -ATelastase, or medium. PMN were incubated for 15 min at 37°C with medium only, 10 nM C1-INH⅐C1 s, or 10 nM ␣ 1 -AT-elastase as described under "Experimental Procedures", and assayed for a chemotactic response to 1 nM ␣ 1 -AT-elastase in a 48-well chemotaxis microchamber. Chemotaxis is defined as the number of PMN per field that migrated through the filter pores and is expressed as the percentage of the maximal migration to FMLP (162 Ϯ 3.8/high power field). The data represent the mean of at least three separate experiments performed in triplicates, and error bars represent the standard error of the mean. To study the role of the SEC receptor for the binding and clearance of C1-INH⅐C1 s in more detail, we performed crosscompetition binding studies employing peptide 105Y, which has been shown previously to bind specifically and saturably to the SEC receptor (10 -12, 41). First, we assessed the binding of 125 I-105Y to separate monolayers of HepG2 cells for 2 h at 4°C and found specific and saturable binding of 125 I-105Y over the concentration range of 20 -100 nM, whereas nonspecific binding increased linearly (Fig. 6A). This result is consistent with previous reports demonstrating specific and saturable binding of 105Y to HepG2 cells with half-maximal saturation at 40 nM (10, 41). Next, we determined whether C1-INH⅐C1 s, C1-INH⅐C1 s⅐C1 r⅐C1-INH, and the COOH-terminal C1-INH peptide compete for the binding of 125 I-105Y to HepG2 cells. Before the experiment, the proper formation of the tetrameric C1-INH⅐C1 s⅐C1 r⅐C1-INH complex was confirmed by sucrose ultracentrifugation and ELISAs employing mAbs 13E1, 303D4, and 98F12 (data not shown). As displayed in Fig. 6B, neither C1-INH⅐C1 s nor C1-INH⅐C1 s⅐C1 r⅐C1-INH sufficiently competed with 125 I-105Y for binding to the SEC receptor, suggesting a low affinity of these ligands for the SEC receptor. Binding of 125 I-105Y to the SEC receptor was also blocked to a much lesser extent by the COOH-terminal C1-INH peptide than with unlabeled 105Y, although the competition was higher compared with the dimeric or tetrameric C1-INH-protease complex (Fig. 6B). Furthermore, C1-INH⅐C1 s complexes (reaction mixture) did not compete for the binding of FITC-conjugated ␣ 1 -AT-trypsin to PMN, in contrast to the same molar excesses of unlabeled ␣ 1 -AT-trypsin (data not shown), which is a high affinity ligand for the SEC receptor (41). Moreover, a 1000-fold molar excess of ␣ 1 -AT-trypsin failed to compete for the in vivo clearance of C1-INH⅐ 125 I-C1 s, in contrast to the same molar excess of unlabeled C1-INH⅐C1 s (Fig. 7C), further indicating that the SEC receptor is not operative in the uptake and clearance of C1-INH⅐C1 s. However, using a concentration range of 5-50 nM C1-INH⅐ 125 I-C1 s and a 50-fold molar excess of unlabeled complex, a specific uptake by primary rat hepatocytes could be observed (data not shown), suggesting that he- INH remained fully active, as judged by the ability to form a SDS-stable complex with 125 I-C1 s (Fig. 4, lane 4). We found that asialo-C1-INH⅐ 125 I-C1 s bound saturably to primary rat hepatocytes, and this binding was specifically competed by excess ASOR, N-acetylgalactosamine, or asialo-C1-INH⅐C1 s, indicating that the ASGPR binds complexes of asialo-C1-INH⅐ 125 I-C1 s (data not shown). To analyze the role of the AS-GPR for the clearance of C1-INH⅐C1 s under physiological conditions, the clearance of C1-INH⅐ 125 I-C1 s and asialo-C1-INH⅐ 125 I-C1 s from the circulation was determined in the absence or presence of excess ASOR. Fig. 7A shows that coinjection of ASOR into mice did not reduce the rate of C1-INH⅐ 125 I-C1 s clearance, in contrast to the clearance of asialo-C1-INH⅐ 125 I-C1 s, which was greatly reduced in the presence of ASOR (Fig. 7B). Furthermore, there was no difference between the clearance of C1-INH⅐ 125 I-C1 s in BALB/c and ASGPR knockout mice (Fig. 7A), suggesting that the hepatic ASGPR has no function in the turnover of C1-INH⅐C1 s.
Uptake and Degradation of C1-INH⅐C1 s by Murine Fibroblasts-Cellular degradation experiments in cultured murine embryonal fibroblasts that either express LRP abundantly (MEF-1) or are deficient for this receptor (PEA-13) were conducted to investigate the possibility whether LRP mediates the catabolism of C1-INH⅐C1 s. As shown in Fig. 8A, MEF-1 efficiently degraded C1-INH⅐C1 s, but not in the presence of chloroquine, indicating that lysosomal function was involved. Furthermore, the presence of RAP, a competitor for all known ligands of LRP (20), totally blocked the degradation of C1-INH⅐ 125 I-C1 s by MEF-1. This was not due to nonspecific interactions, since RAP did not affect the degradation of 125 I-ASOR by HepG2 cells, which have been demonstrated to express ASGPR (42) (data not shown). Control experiments were also performed employing the LRP ligand 125 I-␣ 2 -M*, which was efficiently degraded by MEF-1, but not in the presence of chloroquine or RAP (Fig. 8C). In contrast, 125 I-␣ 2 -M* and C1-INH⅐ 125 I-C1 s were not efficiently degraded by HepG2 cells, indicating that the cells used in these experiments did not effectively express LRP (data not shown). We next explored the ability of the LRP-deficient PEA-13 cell line to degrade C1-INH⅐ 125 I-C1s and, as a control, 125 I-␣ 2 -M*. As shown in Fig. 8B for C1-INH⅐ 125 I-C1s complexes and in Fig. 8D for 125 I-␣ 2 -M*, both ligands were not degraded by PEA-13, suggesting that LRP is the receptor essential for the uptake and degradation of C1-INH⅐C1s.
RAP Competes for the Plasma Clearance of C1-INH⅐ 125 I-C1 s in Vivo-Since RAP is known to compete for the in vivo plasma clearance of LRP ligands (16,43), we used RAP as competitor in turnover experiments employing C1-INH⅐ 125 I-C1 s and 125 I-␣ 2 -M* as a control. Fig. 9 (A and B) shows that the removal of C1-INH⅐ 125 I-C1 s and 125 I-␣ 2 -M* from the circulation was significantly delayed when an excess of RAP was co-injected with the ligand, further demonstrating the role of LRP for the clearance of C1-INH⅐C1 s complexes. DISCUSSION Previous studies by de Smet et al. (8) and by Malek et al. (9) have shown that proteinase-complexed C1-INH is rapidly removed from plasma in contrast to native or reactive center cleaved C1-INH. However, the receptor that mediates the rapid removal of C1-INH-enzyme complexes has not been identified.
Besides their function in the rapid removal of inactivated proteinases from plasma, some proteinase-complexed or cleaved serpins have been shown to initiate cellular signal transduction events (44,45). For example, Banda et al. (40) reported that the inhibitory complex of ␣ 1 -AT-elastase, as well as reactive center cleaved ␣ 1 -AT, is a potent neutrophil chemoattractant. Chemotaxis was stimulated by the 4.2-kDa COOH-terminal peptide of ␣ 1 -AT, which was found to remain I-C1 s complexes were injected in the absence of competitor in BALB/c (AS-GPR ϩ/ϩ) and ASGPR knockout mice (ASGPR Ϫ/Ϫ) or in the presence of a 3400-fold molar excess of ASOR in BALB/c mice. As described under "Experimental Procedures," blood samples were collected at the indicated times and trichloroacetic acid-insoluble radioactivity was determined. The initial time point, taken after 1 min of injection, was considered to represent 100% radioactivity in the circulation. B, 6 nM asialo-C1-INH⅐ 125 I-C1 s complexes were injected in BALB/c mice in the absence or presence of a 3400-fold molar excess of ASOR. Samples were processed as described above. C, 11 nM C1-INH⅐ 125 I-C1 s complexes were injected in BALB/c mice in the absence of competitor or in the presence of a 1000-fold molar excess of either C1-INH⅐C1 s or ␣ 1 -AT-trypsin. Samples were processed as described above. Each symbol represents the mean of at least duplicate experiments. tightly associated with the ␣ 1 -AT-elastase complex (40). However, we did not find a chemotactic response of neutrophils or monocytes toward native C1-INH, C1-INH⅐C1 s, and the COOH-terminal C1-INH peptide. C1-INH⅐C1 s complexes were also found to be ineffective to block the neutrophil chemotactic activity of ␣ 1 -AT-elastase in cross-desensitization experiments, suggesting that C1-INH⅐C1 s does not engage the SEC receptor.
The role of the SEC receptor for the potential mediation of chemotaxis by C1-INH⅐C1 s was further questioned by our findings that neither unlabeled nor iodinated C1-INH⅐C1 s bound to PMN, monocytes, and HepG2 cells, which are known to express the SEC receptor (10,13). C1-INH⅐C1 s complexes were also unable to compete the binding of FITC-conjugated ␣ 1 -AT-trypsin to PMN. On the other hand, it was found that unlabeled ␣ 1 -AT-trypsin was ineffective in delaying the clearance of C1-INH⅐ 125 I-C1 s in contrast to unlabeled C1-INH⅐C1 s. Most importantly, C1-INH⅐C1 s and the physiologically occurring tetrameric C1-INH⅐C1 s⅐C1 r⅐C1-INH complex did not efficiently compete with 125 I-peptide 105Y for the binding to the SEC receptor. These results confirm and extend the early observations of Perlmutter et al. (10) that the SEC receptor recognizes C1-INH⅐C1 s to a much lesser extent compared with ␣ 1 -ATelastase, ␣ 1 -ACT-cathepsin G, and antithrombin-thrombin. Furthermore, Patston et al. (27) found no specific binding of C1-INH⅐ 125 I-C1 s complexes to HepG2 cells in contrast to ␣ 1 -AT-125 I-trypsin complexes. Our results confirm this observation, because we were also unable to detect specific binding of C1-INH⅐ 125 I-C1 s to HepG2 cells.
The tight association of the COOH-terminal C1-INH peptide with the remaining complex would be one prerequisite for the binding of C1-INH⅐C1 s complexes to the SEC receptor, and we show here that the poor recognition of C1-INH⅐C1 s by the SEC receptor is not due to the absence of the COOH-terminal C1-INH peptide, because it was found to be tightly associated with the complex and with reactive center cleaved C1-INH (38). It is likely that the selectivity in the recognition of different SECs by the SEC receptor is due to differences within the conserved pentapeptide domain (12) or by the lack of accessibility of the pentapeptide domain of C1-INH for the receptor. The latter is supported by crystallographic data of cleaved serpins, showing that the COOH-terminal peptide is re-folded into a ␤-pleated sheet structure, which is at least partially buried within the modified serpin (7).
However, we determined that C1-INH⅐ 125 I-C1 s complexes are taken up by the liver, suggesting that hepatic receptors are involved that are different from the SEC receptor.
One likely receptor candidate was the well characterized hepatic asialogycoprotein receptor, which binds penultimate galactose residues of oligosaccharides, thus thought to mediate the turnover of glycoproteins after these have become desialylated (46). Several studies have shown that the functional activity of the heavily glycosylated C1-INH is independent of its carbohydrates (47,48). These findings prompted some authors to imply a possible role for the carbohydrate in clearance mechanisms after the inhibitor has bound C1 r and C1 s (49,50). The data presented here clearly exclude this possibility, al- though we cannot rule out this for the removal of native or cleaved C1-INH from plasma. However, there are data arguing strongly against a role of ASGPR in the normal turnover of plasma glycoproteins (25).
Another receptor candidate for the clearance of proteasecomplexed C1-INH was the low density lipoprotein receptorrelated protein, and this possibility has already been postulated by Orth et al. (14). The early finding that LRP mediates the uptake of ␣ 2 -macroglobulin-protease complexes suggested its role as an efficient protease inhibitor clearance receptor, and subsequent studies have expanded this as well for SECs like plasminogen activator inhibitor type 1-plasminogen activator, plasminogen activator inhibitor type 1-urokinase-type plasminogen activator, and ␣ 1 -AT-elastase (14,15,23,51). Recently, Kounnas et al. (16) have demonstrated that LRP is also responsible for the endocytosis and degradation of heparin cofactor II-thrombin, antithrombin-thrombin, and ␣ 1 -AT-trypsin. The cleaved or native serpins were not recognized by LRP, which is in line with results of previous in vivo clearance experiments (6). In the present study, we show that LRP is also the receptor candidate, which mediates the clearance and degradation of C1-INH⅐C1 s.
Although hepatocytes have been shown to express LRP abundantly (52), we found no efficient degradation of C1-INH⅐ 125 I-C1 s and 125 I-␣ 2 -M* by HepG2 cells, which is in line with our finding that C1-INH⅐ 125 I-C1 s complexes did not bind to HepG2 cells. Poller et al. (15) have also reported that 125 I-␣ 2 -M* was degraded much more efficiently by MEF-1 than by HepG2 cells. This suggests that HepG2 cells are heterogeneous and may not adequately represent a normal hepatocyte with regard to LRP expression. Differences between HepG2 cells and hepatocytes have also been reported by Patston et al. (27) with respect to C1-INH biosynthesis.
Our finding that ␣ 1 -AT-trypsin does not compete for the clearance of C1-INH⅐ 125 I-C1 s suggests that there are different SEC binding sites on LRP. This is supported by the finding that ␣ 2 -AP-plasmin does not compete the clearance of heparin cofactor II-thrombin, ATIII-thrombin, and ␣ 1 -AT-trypsin (6) and by the data of Pizzo (53) that proteinase-complexed C1-INH and ␣ 2 -AP clear via the same receptor system (SEC II receptor). However, it is not known yet which domain of the different serpin-enzyme complexes (SECs) is recognized by LRP. Obviously, LRP and the SEC receptor have an overlapping ligand binding specificity; however, Poller et al. (15) have provided evidence that LRP is unrelated to the SEC receptor. Furthermore, there are no reports yet showing that LRP mediates signal transductions events, in contrast to the SEC receptor, and both receptors differ largely in their molecular masses (80 kDa for the SEC receptor and 515 kDa for the ligand binding subunit of LRP, respectively) (54,55).
In summary, we found that C1-INH⅐C1 s has no signal transduction capabilities to stimulate a chemotactic response of PMN and monocytes, in contrast to ␣ 1 -AT-elastase or ␣ 1 -ACTcathepsin G. The SEC receptor and ASGPR are not involved in the clearance of C1-INH⅐C1 s, which is mediated by LRP, further underscoring its crucial role for the catabolism of serpinenzyme complexes.