Soluble low density lipoprotein receptor-related protein (LRP) circulates in human plasma.

Our studies have identified a soluble molecule in normal human plasma and serum with the characteristics of the alpha-chain of the low density lipoprotein receptor-related protein (LRP). LRP is a large multifunctional receptor mediating the clearance of diverse ligands, including selected lipoproteins, various protease inhibitor complexes, and thrombospondin. A soluble molecule (sLRP) has been isolated from plasma using an affinity matrix coupled with methylamine-activated alpha2-macroglobulin, the ligand uniquely recognized by LRP, and eluted with EDTA. This eluate contains a protein that co-migrates on SDS-polyacrylamide gel electrophoresis with authentic human placental LRP alpha-chain, is recognized by anti-LRP alpha-chain monoclonal antibodies, and binds the 39-kDa receptor-associated protein (RAP) and tissue plasminogen activator-inhibitor complexes. A similar RAP-binding molecule was detected in medium conditioned for 24 h by primary cultures of rat hepatocytes, suggesting that the liver may be the in vivo source of sLRP. In contrast, immunoprecipitation experiments failed to detect the production of sLRP by cultured HepG2 hepatoma and primary human fibroblast cells. Addition of a soluble form of LRP to cultured HepG2 cells resulted in a significant inhibition of capacity of these cells to degrade tPA, a process that has been demonstrated to be mediated by cell surface LRP. Preliminary data indicate that the concentration of sLRP is altered in the plasma of patients with liver disease. Increased levels of sLRP may antagonize the clearance of ligands by cell bound LRP perturbing diverse processes including lipid metabolism, cell migration and extracellular proteinase activity.

from the American Type Culture Collection (Rockville, MD). Antibodies were purified from culture supernatant using protein G-Sepharose (Pharmacia Biotech Inc., Uppsala, Sweden), according to the manufacturer's instructions. All cell culture reagents were purchased from ICN (Costa Mesa, CA) and culture-ware was from Costar (Cambridge, MA).
Color development was stopped after 20 min by addition of 50 l/well 3% oxalic acid, and the optical density at 405 nm (A 405 ) was determined.
Affinity Isolation of LRP-An affinity matrix was prepared by coupling 280 mg of methylamine-activated ␣ 2 M to 25 ml of CNBr-Sepharose (Pharmacia Biotech Inc.), according to the manufacturer's instructions. Fresh frozen human plasma (200 ml) containing a mixture of protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 2 mM benzamidine, 2 mM bacitracin, 1 M leupeptin) was thawed at 37°C. The plasma was adjusted to pH 7.4 by addition of 1 ⁄10 volume 0.1 M HEPES, pH 7.4, 10 IU/ml heparin, 5 mM Ca 2ϩ , clarified by centrifugation (4°C, 20 000g, 30 min), and filtered through a 0.4-m nitrocellulose filter (Millipore-Waters, Milford, MA). The plasma was mixed with blank Sepharose gel for 1 h at 4°C. The precleared plasma supernatant was then mixed with methylamine activated ␣ 2 M-Sepharose for 6 h at 4°C. The affinity matrix was washed on a scintered glass funnel with 250 ml of HBSC, packed into a column and eluted with 25 mM EDTA, 20 mM Mes, pH 6.0. Protease inhibitors (as above) were added to each fraction (1 ml). LRP was also isolated from detergent-solubilized human placental membranes, prepared as described previously (10), using the activated ␣ 2 M-Sepharose affinity matrix. The fractions were screened by specific LRP immunoassay and pooled positive fractions were stored at Ϫ20°C. The material eluted from the activated ␣ 2 M affinity matrix was fractionated by electrophoresis on 5-15% gradient SDS-polyacrylamide gels (SDS-PAGE) and analyzed by silver staining (11).
Western and Ligand Blots-Prior to Western and ligand blot analysis samples were electrophoresed on 6% SDS-PAGE minigels for 2 h at 150 V. The gels were electroblotted onto polyvinylidene difluoride membrane (NEN Life Science Products) and incubated, according to the manufacturer's instructions, with 5 g/ml mAb and 1/2000 dilution rabbit anti-mouse immunoglobulin-horseradish peroxidase conjugate (Dako). Bound antibody was visualized with chemiluminescence reagent (Renaissance, NEN Life Science Products) and exposed to film (Hyperfilm-MP, Amersham Corp.). For ligand blotting, blots were blocked in 5% milk, HBSC, 0.1% Tween 20 and incubated (4 h at room temperature) in either 4 nM 125 I-ligand (RAP or tPA), washed, and exposed to film (24 h, Ϫ80°C, intensifying screen). Blots were incubated in 125 I-tPA in the presence of 100 nM PAI-1, resulting in the complete conversion of 125 I-tPA to an SDS-stable tPA⅐PAI-1 complex (data not shown). The specificity of binding was indicated by the absence of 125 I-ligand binding in the presence of 1 M unlabeled RAP or tPA.
Cell Culture-All cell lines were cultured under standard conditions at 37°C, 5% CO 2 in a humidified incubator. Hepatocytes were isolated from 230 -270 g of anesthetized male Wistar rats by in situ two-stage collagenase perfusion of the liver, as described elsewhere (12). The liver was excised and dispersed in modified Waymouth medium containing sodium bicarbonate (24 mM), penicillin (100 units/ml), HEPES (18 mM), and insulin (25 milliunits/ml). Viability was determined by trypan blue exclusion, and only preparations with greater than 80% viable hepatocytes were used. Cells (3.0 ϫ 10 6 in a final volume of 3 ml of Williams' medium E, supplemented as above) were overlaid onto 60-mm culture plates (Medos Company, Sydney, Australia) coated with 400 l of matrigel prepared as described previously (13). The culture medium was replaced after a 3-h attachment period and at 24-h intervals thereafter. HepG2 cells (ATCC HB8065, provided by M. Gallicchio, Monash University, Melbourne, Australia) were maintained in Earle's modified Eagle's medium (EMEM) supplemented with 10% fetal bovine serum (FBS). Human foreskin fibroblasts (HFF) were kindly provided by Dr. Gabrielle Delbridge (Center for Thrombosis and Vascular Research, University of New South Wales) and maintained in Dulbecco's modified Eagle's medium supplemented with 10% FBS. For experiments, cell monolayers were trypsinized, with HepG2 cells plated at 2 ϫ 10 4 cells/ml and HFF plated at 5 ϫ 10 3 cells/ml. The cells were fed on day 3 and used for experiments on day 4 of culture when they were approximately 80% confluent.
Analysis of Primary Rat Hepatocyte Cultures-On day 7 of culture, cells and medium were harvested from rat hepatocyte cultures. Protease inhibitors were added to conditioned medium after it was centrifuged 4000 ϫ g, 10 min. The cells, on ice, were lysed in 2 ml of 0.25% Triton X-100, HBSC ϩ protease inhibitors (as above) and clarified by centrifugation 4000 ϫ g for 10 min at 4°C. The conditioned medium (8 ml) and the whole cell lysate (4 ml) were diluted 1/5 in 20 mM Mes, pH 6.0, and mixed with 1 ml of DEAE-Sephacel (Pharmacia Biotech Inc.) equilibrated in 20 mM Mes, pH 6.0. After washing with 20 mM Mes, pH 6.0, the gel was eluted with 0.5-ml aliquots of 20 mM Mes, pH 6.0, 0.5 M NaCl. The eluted fractions (25 l/lane) were electrophoresed on 6% SDS-PAGE gels, electroblotted onto polyvinylidene difluoride membrane, and incubated with 125 I-RAP, as described above.
Immunoprecipitation-HepG2 cells and HFF were plated in 60-mm Petri dishes in 10 ml of culture medium. When the cells were 80% confluent, the monolayers were washed twice with phosphate-buffered saline and incubated in 4 ml of cysteine/methionine-free EMEM, 10% FBS for 20 min at 37°C. The cultures were then pulsed by incubation, 1 h, in the presence of 66 Ci/ml [ 35 S]methionine/cysteine (Tran 35 Slabel, ICN Biomedical). Long term labeling was conducted overnight in the presence of 20% cysteine/methionine complete EMEM. The chase period (0, 30, 60, 180, or 240 min) was initiated by the addition of 4 ml of EMEM, 10% FBS containing 0.2 mM unlabeled cysteine/methionine. Medium was collected and centrifuged 1200 ϫ g for 5 min. Cell monolayers were washed twice with phosphate-buffered saline and lysed in the dish by the addition of 4 ml of 1% Triton, HBSC ϩ protease inhibitors (as above) and clarified by centrifugation 4000 ϫ g for 10 min at 4°C. Samples were mixed with 25 l of protein G-Sepharose (1:1 slurry, 30 min, 4°C) and then incubated (4°C, 2 h) in the presence or absence of 5 g of anti-LRP ␣-chain mAb. After incubation with 25 l of protein G-Sepharose (1:1 slurry, 30 min, 4°C), the Sepharose was washed five times in 20 mM HEPES, 0.5 M NaCl, 0.1% Triton X-100, pH 7.4, suspended in 50 l of SDS-PAGE sample buffer and electrophoresed on a 5-15% gradient SDS-PAGE gel. Gels were fixed in 10% acetic FIG. 1. Affinity isolation of sLRP from plasma. An affinity column of activated ␣ 2 M-Sepharose (280 mg of activated ␣ 2 M coupled to 25 ml of gel) was used to affinity-purify LRP from 200 ml of human plasma. A, human plasma contains immunoreactivity detected in an LRP immunoassay (q), which was removed by incubation with the activated ␣ 2 M-Sepharose (E). B, the immunoreactivity was recovered by eluting (1-ml fractions collected) the affinity matrix with 25 mM EDTA, 20 mM Mes, pH 6.0. Fractions eluting between 21 and 25 ml were pooled and analyzed. acid, 50% methanol, soaked in Amplify (Amersham Life Science, Inc., Little Chalfont, United Kingdom), and dried before exposing to film for 3 days.
Degradation of 125 I-tPA by HepG2 Cells-HepG2 cells plated in 24well culture dishes were washed three times in EMEM, 0.2% BSA and incubated in 2 ml of medium containing 3 nM 125 I-tPA in the presence or absence of the following ligands: 1) 1 M unlabeled tPA, 2) 2 g/ml (3.3 nM) purified placental LRP, and 3) 10 g/ml (16.7 nM) purified placental LRP. After incubation for appropriate times, binding medium was removed and mixed with an equal volume of 20% trichloroacetic acid containing 4% phosphotungstic acid, incubated on ice for 10 min, centrifuged 10,000 ϫ g for 5 min, and a 1-ml aliquot of supernatant was counted in a ␥-counter to determine counts/min of 125 I-tPA degraded from triplicate cultures at each time point.
Determination of sLRP Concentration in Human Plasma-The LRP immunoassay was used to determine the concentration of sLRP in human plasma. A standard curve was prepared by diluting affinitypurified placental LRP in the concentration range 0.125-2.5 g/ml. sLRP concentrations were calculated from the standard curve using a four parameter curve fit. A preliminary study was conducted to determine the effect of blood additives and storage conditions on the estimation of LRP concentration by the LRP immunoassay. To determine the range of LRP in normal human plasma, citrated plasma was obtained from 50 healthy blood donors giving informed consent. In addition, citrated plasma samples (n ϭ 45) were collected by the Clinical Chemistry Department, Prince of Wales Hospital, Randwick, New South Wales, Australia. These plasma samples were obtained from patients with clinical manifestation of liver disease, which was confirmed by routine liver function test assessing plasma and urinary bilirubin, plasma alkaline phosphatase, plasma transaminases, plasma ␥-glutamyltransferase. A third set of plasma samples from patients (n ϭ 49) with non-insulin-dependent diabetes mellitus (NIDDM) were kindly provided by Drs T. Mori and D. Dunstan, Department of Hematology, Royal Perth Hospital, Western Australia.

Human Plasma Contains a Soluble
Molecule with the Characteristics of LRP (sLRP)-An affinity matrix consisting of activated ␣ 2 M-Sepharose is able to deplete plasma of an LRPlike molecule detected using a specific LRP immunoassay (Fig.  1A). Consistent with Ca 2ϩ -dependent ligand binding, the immunoreactivity was recovered by washing the column with 25 mM EDTA, pH 6.0 (Fig. 1B). Analysis of the EDTA eluate of human plasma adsorbed to ␣ 2 M-Sepharose by SDS-PAGE (reducing) revealed the presence of a single chain molecule which co-migrates with the ␣-chain of LRP (M r 500,000) isolated from human placental membranes ( Fig. 2A). The high molecular weight protein appears to be structurally and functionally related to human LRP ␣-chain, as it is recognized by an anti-LRP mAb (8G1) on Western blot and binds 125 I-RAP in a ligand blot ( Fig. 2A). In contrast, a mAb specific for the COOH terminus of LRP ␤-chain (M r 85,000) failed to specifically recognize any species in the enriched plasma fraction either by specific immunoassay (data not shown) or Western blot (Fig. 2B). A second, distinct anti-LRP mAb detected sLRP in the enriched plasma fraction, providing further evidence that the affinity eluate of human plasma contains a molecule closely related to LRP (Fig. 2C). A molecule of identical electrophoretic mobility was also recognized by an affinity-purified rabbit anti-human LRP polyclonal antibody (R777, a generous gift of Dr. Dudley Strickland, American Red Cross; data not shown).
sLRP Is Secreted by Primary Cultures of Rat Hepatocytes-Ligand blot analysis of culture medium conditioned by primary cultures of rat hepatocytes revealed the presence of a 125 I-RAPbinding protein that co-migrated on SDS-PAGE with human placental LRP and a cell-associated molecule in cellular lysates of cell rat hepatocytes (Fig. 3A). Prior to analysis the conditioned medium was concentrated 16-fold by ion exchange chromatography on DEAE-Sephacel. Immunoprecipitation of metabolically labeled human HepG2 and HFF cells with an anti-LRP ␣-chain antibody (8G1) failed to detect the accumulation of a soluble form of LRP in the supernatant medium after a 4-h chase (Fig. 3B). Cell-associated LRP was readily detected at each chase time point, i.e. from the end of the 1-h pulse onwards. Long term labeling experiments (up to 24 h) also failed to detect soluble forms of LRP in medium conditioned by these cells (data not shown).
A Soluble form of LRP Is Able to Perturb the Degradation of tPA by HepG2 Cells-The uptake and degradation of tPA has been demonstrated to be mediated by LRP and can be inhibited by the LRP antagonist RAP (14). Ligand blot analysis demonstrates that both the purified placental LRP and the plasma-  1, arrow), was recognized by an anti-LRP ␣-chain mAb, 8G1, and bound 125 I-RAP on a ligand blot. B, LRP ␤-chain migrates on SDS-PAGE with M r 85,000. The placental LRP preparation contains two molecules specifically recognized by the anti-LRP ␤-chain antibody: the strong band at M r 85,000 represents the mature LRP ␤-chain (arrow), while the size of the weaker high molecular weight band (M r Ͼ500,000) is consistent with incompletely processed LRP prohormone, which has not been cleaved into the two LRP subunits. The plasma affinity eluate did not contain a molecule that specifically reacted with the anti-LRP ␤-chain antibody. The band present at M r 120,000 in lane 2 was a nonspecific contaminant in this preparation, which was recognized directly by the secondary (rabbit anti-mouse Ig-horseradish peroxidase) antibody. C, further confirmation that the plasma affinity eluate contains a molecule closely related to LRP is provided by the recognition of a band, which co-migrated with placental LRP ␣-chain (not shown), by both the 8G1 mAb (lane 1) and a second, distinct anti-LRP ␣-chain mAb (lane 2) on Western blot. derived sLRP bind 125 I-tPA⅐PAI-1 complexes (Fig. 4A). Fig. 4B demonstrates the inhibition of 125 I-tPA degradation by HepG2 cells when assays were conducted in the presence of 10 g/ml (16.7 nM) purified placental LRP.
Plasma sLRP Levels Are Increased in Patients with Liver Disease-sLRP levels were measured in healthy subjects to define the normal concentration range. The plasma sLRP concentration estimated using the LRP immunoassay did not vary significantly when blood from healthy donors was collected into a dry tube (serum, 6.9 g/ml), EDTA (6.5 g/ml), or citrate (6.8 g/ml), when samples were assayed at 10 -20-fold dilution in Ca 2ϩ -containing assay buffer. Addition of 10 IU/ml heparin to the assay buffer to prevent postdilution clotting also did not interfere with the estimation of sLRP concentration. The range of sLRP concentration detected in the plasma of healthy subjects (n ϭ 50; mean 6.1 Ϯ 1.2 g/ml; range 3.7-10.8 g/ml), patients with abnormal liver function (n ϭ 45; mean 7.1 Ϯ 3.44 g/ml, range ϭ 2.0 -22.4 g/ml), and patients with NIDDM (n ϭ 49; mean 6.5 Ϯ 0.9 g/ml, range ϭ 5.0 -9.0 g/ml) is shown in Fig. 5. The plasma concentration of sLRP was significantly altered (varied by Ͼ2 S.D. from the mean normal concentra-tion, i.e. outside the range 3.7-8.5 g/ml) in 24% plasma samples from patients with abnormal liver function. In the NIDDM disease control group, only two plasma samples (4%) contained a significantly altered sLRP concentration (both above 8.5 g/ ml). The altered plasma sLRP concentrations associated with

FIG. 3. sLRP is released by cultured primary rat hepatocytes.
A, rat hepatocytes were cultured for 7 days in serum-free medium on matrigel to maintain a differentiated phenotype characteristic of adult liver. Samples, prepared as described under " Materials and Methods," were electrophoresed on 6% SDS-PAGE gels for ligand blot analysis. abnormal liver function test did not correlate with disease etiology nor with the level of any plasma proteins assayed, including bilirubin, alkaline phosphatase, transaminases, ␥-glutamyltransferase or 5Ј-nucleotidase. DISCUSSION LRP belongs to the low density lipoprotein receptor family, the members of which share many structural and functional characteristics (3). The ␣-chain of the heterodimeric LRP contains multiple, Ca 2ϩ -dependent ligand binding domains and is noncovalently bound on the cell surface to the membranespanning ␤-chain (15). The characteristics of the soluble molecule circulating in human plasma suggest it is a molecule closely related to the ␣-chain of LRP: it displays Ca 2ϩ -dependent binding to two established LRP ligands RAP and activated ␣ 2 M, a ligand uniquely recognized by LRP; it is a single chain molecule which co-migrates on SDS-PAGE with authentic human placental LRP ␣-chain; and it is recognized by two distinct anti-LRP ␣-chain monoclonal antibodies. We were unable to detect the intracellular COOH terminus of the LRP ␤-chain in the affinity-isolated sLRP. However, it is possible that a truncated ␤-chain may be associated with the soluble ␣-chain. Precise determination of the exact structure of plasma sLRP and its relative affinity for the various LRP ligands will only be possible with a highly purified preparation of sLRP. These studies are currently under way. The presence of a circulating LRP-like molecule is not confined to human plasma, and a similar molecule has been detected in plasma and serum from a variety of mammals and the chicken. 2 LRP is most highly expressed in the liver, and analysis of culture medium conditioned by primary cultures of rat hepatocytes revealed the presence of an sLRP-like 125 I-RAP-binding protein. Our experiments could not detect the release of a soluble form of LRP from the human hepatoma cell line HepG2, nor from cultured normal human fibroblasts, suggesting that the release of sLRP is not a constitutive property of all cultured cells. The absence of sLRP in medium conditioned by the hepatoma cell line suggests that the production of sLRP may be associated with the more differentiated phenotype of the primary hepatocytes cultured on matrigel (16). Further studies are required to investigate the factors regulating the production of sLRP.
The detection of sLRP in hepatocyte-conditioned medium provides a model system for the further characterization of sLRP and the elucidation of the mechanism generating the soluble form. A wide variety of receptors and other plasma membrane proteins have been identified as having soluble counterparts in serum (17). There are examples in the literature of soluble receptors liberated by proteolytic cleavage of receptor exodomains (18,19) and those that derive from differential splicing of a common mRNA transcript or transcription of closely related, but distinct, genes (20). In the case of LRP, another potential mechanism for the generation of sLRP could be the disruption of the noncovalent bond that anchors the ␣-chain to the membrane-spanning ␤-chain (15). A soluble form of gp330, another member of the low density lipoprotein receptor family, has been detected in the supernatant of a yolk sac carcinoma cell line (21). Interestingly, in this cell line gp330 was present on the cell surface and in supernatant as a complex with RAP. No further characterization of the cellular and molecular events regulating gp330/RAP release has been reported.
Soluble receptors, generally, have reduced ligand affinity constants compared with their membrane-bound counterparts, and the circulating receptor concentration may be insufficient to effectively compete ligand binding to the cell-bound molecule (22). Our preliminary experiment indicates that the addition of soluble purified placental LRP at a physiologically relevant concentration (10 g/ml) was able to inhibit the degradation of tPA by cultured HepG2 cells, a process that has been demonstrated previously to be mediated by cell surface LRP (14). This indicates that the sLRP may act as a competitive inhibitor of ligand uptake by cell surface-bound LRP. The inhibition of tPA uptake by soluble LRP may have negligible consequences to the level of extracellular proteolysis, because the enzyme is rapidly inactivated before endocytosis by complex formation with plasminogen activator inhibitor-I (PAI-1), present in large excess in the extracellular matrix of HepG2 cells (23,24). However, LRP mediates the uptake of many biologically active molecules, including the potent angiogenesis inhibitor, thrombospondin (25), apoE-enriched lipoproteins (26), lipoprotein-lipase (27), and is able to bind directly active tPA (28) and urokinase PA (29). Increased levels of soluble LRP may extend the half-life of these active ligands and influence diverse biological processes, including lipid metabolism, cell growth, and migration and extracellular proteinase activity.
The identification of a soluble molecule with the characteristics of LRP introduces a new dimension into the biology of this unique molecule. Further studies are required to understand the biochemical mechanisms involved in the generation of sLRP and establish its physiological role. The concentration of plasma sLRP appears to alter in some patients with impaired liver function. A more extensive study is required to substantiate this preliminary data and elucidate the functional implications of the moderate changes. As LRP has been implicated to be overexpressed in a range of other pathological processes, including atherosclerosis (30) and Alzheimer's disease (31), the functional or prognostic implications of plasma sLRP concentration requires investigation.