Mesangial Cells Possess an Asialoglycoprotein Receptor with Affinity for Human Immunoglobulin A

CARMEN GOMEZ-GUERRERO, NATALIA DUQUE, and JESUS EGIDO Fundaci#{243}n Jim#{233}nez DIaz, Universidad Aut#{243}norna, Madrid, Spain.

Abstract. Asialoglycoprotein receptor (ASGP-R), a hepatic tected by mRNA expression analysis (ratio 2: 1). Because the lectin involved in the clearance of galactose-terminal glyco- hepatic receptor presents avidity for the carbohydrates of IgAl, proteins, is also present in extrahepatic tissues, but its expres- a protein deposited in the glomerulus of patients with IgA sion in renal cells is not well established. This study examines nephropathy, the interaction of IgA 1 with the mesangial the presence of ASGP-R in cultured mesangial cells (MC), key ASGP-R was explored. As for the interaction with ASOR, cells involved in the removal of macromolecules deposited in catabolism of IgA1 by rat and human MC was Ca2tdependent the glomerulus. The binding of asialo-orosomucoid (ASOR) to and was reduced with galactose. In addition, the interaction of rat MC was saturable and galactose-specific. In addition, MC ASOR with rat MC was partially inhibited by incubation with internalized and degraded ASOR in a Ca2tdependent manner. IgAl and its desialylated form, but not by IgA2, as demon- Parallel studies were performed in a homologous system (hu- strated in binding experiments and in receptor purification. It is man MC), obtaining similar binding curve and competition concluded that MC possess ASGP-R specific for galactose

with unlabeled ASOR and carbohydrates. The purified receptor residues of several glycoproteins, including IgA1 . These data from rat MC consisted of two proteins (41 and 55 kD) with could be important for a better understanding of the pathogen- similar size to the hepatic receptor. Both subunits were de- esis oflgA nephropathy. (J Am Soc Nephrol 9: 568-576, 1998)

The asiaboglycoprotein receptor (ASGP-R) is a lectin be- equal numbers, state 2 mediates the majority of ligand process- longing to the C-type that binds and mediates endocytosis of ing and could be inactivated by diverse agents (1,9).

serum glycoproteins ( 1 ). The hepatic ASOP-R was first In contrast to the traditional concept that ASGP-R is exclu- recognized by its high affinity for the galactose-ending sively present in the liver, several works have shown that oligosacchanides of desialylated glycoproteins, but the car- mRNA for RHL-l and RHL-2/3 are widely expressed in sev- bohydrate specificity could also be extended to N-acetyl- eral rat tissues, and their expression is regulated during devel- galactosamine (2). There is a wide range of glycoproteins opment (10, 1 1). This suggests that other organs besides the that could bind to the ASGP-R in physiologic conditions, liver may participate in the clearance of serum glycoproteins, such as transcobalamin III and low-density lipoproteins (1). although the potential biological role of the extrahepatic Among these, asialo-orosomucoid (ASOR), a protein with ASGP-R has not been explored. Expression of extrahepatic tetra-antennary oligosaccharides, presents the highest affin- ASGP-R is also found in cultured intestinal epithelial cells ity (3,4). Physiologic ligands for ASGP-R are not limited to ( 12). However, the renal cells expressing ASGP-R remain soluble proteins. The hepatic ASGP-R mediates erythroag- unidentified. glutination, clearance of apoptotic cells, and binding and The glomerular mesangial region in the kidney is formed by uptake of hepatitis B virus (1,5,6). mesangiab cells (MC) and matrix. Morphologically, MC resem- The rat hepatic ASGP-R is an integral transmembrane gly- ble smooth muscle cells and they contract in response to coprotein composed of three subunits (RHL- 1, -2, and -3) several stimuli. Like macrophages, these cells can phagocytose encoded by two separate genes. RHL- 1 is the predominant several particles in association with the generation of prosta- subunit, with an apparent molecular mass of 41 to 43 kD. glandins and reactive oxygen metabolites (13). In this study, RHL-2 and RHL-3 (49 to 54 and 54 to 64 kD, respectively) we investigated the existence of ASGP-R in cultured rat and have the same core protein and differ only in the type and human MC, performing binding and uptake experiments of extent of posttranslational carbohydrate modification (7,8). radiolabeled ASOR, isolation of membrane receptors, and Two functionally different receptor populations, designated as mRNA expression analysis. state 1 and state 2, have been described. Although present in Elevated plasma levels of IgA 1 have been detected in pa- tients with a broad spectrum of diseases, including IgA ne- phropathy, and deposits of IgA 1 are characteristically found in Received June 23. 1997. Accepted September 26, 1997. the glomerular mesangium of these patients (14). In a previous Correspondence to Dr. Jesus Egido, Renal Research Laboratory, Fundaci#{243}n Jim#{233}nezDIaz, Avda Reyes CatOlicos 2, 28040 Madrid, Spain. work, we demonstrated that rat and human MC possess a specific receptor for the Fe region of IgA (FeaR) and that IgA 1 1046-6673/0904-0568$03.O0/O Journal of the American Society of Nephrology and its desialylated form bound to those cells in a dose- Copyright 0 1998 by the American Society of Nephrobogy dependent manner ( 15). In addition, the carbohydrates of IgAl Asialoglycoprotein Receptors in Mesangial Cells 569

seemed to play a role in the interaction with MC, although the and cytokeratin, excluding endothelial and epithebial contamination, exact role was not completely defined (15). Because serum respectively (13,15). IgA1 binds to the hepatic ASGP-R through the O-glycosid- ically linked oligosacchandes (galactose and N-acetyl-galac- Binding, Internalization, and Degradation of Proteins tosamine) present in the hinge region of the IgAl (16,17), in MC were trypsinized and subcultured for 3 to 6 d in 6-well plates this study we further investigated the possibility that IgA 1 (2.5 X l0 cells/well) and made quiescent by incubation for 48 h in could interact with the ASGP-R on cultured rat and human MC medium containing 0.5% FCS. For binding studies, MC were washed with ice-cold PBS, pH 7.3, containing 0.5% bovine serum albumin through those sugar residues. (PBS-BSA) and then incubated in 1 ml of Hepes buffer (2 mM Hepes, 150 mM NaCl, 5 mM KC1, 2 mM CaC12, 0.5% BSA, 0.1% glucose, Materials and Methods pH 7.4) with several concentrations of ‘251-ASOR. Incubation was Proteins earned out at 4#{176}Ctoavoid ligand internalization and degradation.

Monomeric human IgA1 (K light chain) was purified from serum of Nonspecific binding was determined in the presence of a 100-fold patients as described (15). Briefly, IgAl myeloma serum was precip- excess of unlabeled ASOR. Background was determined as the radio- itated with 50% (wtlvol) (NH4),SO4 and dialyzed against 0.015 M activity measured in the absence of cells. After 1 h of incubation, cells phosphate buffer, pH 7.5. The protein solution was chromatographied were rinsed with ice-cold PBS-BSA and then removed by addition of on a diethylaminoethyl cellulose column (Pharmacia, Uppsala, Swe- 0. lN NaOH, measuring the radioactivity contained in the cell fraction. den). Unbound protein, mostly belonging to the IgG class, was dis- For binding competition assays, a 30-mm preincubation with unba- carded, and monomeric IgA1 was eluted with 0.05 M phosphate beled proteins (ASOR, IgA I , deslgA 1 , and IgA2) or sugars (galac- buffer, pH 7.0. To eliminate IgO traces, the monomeric IgA 1 fraction tose, glucose, N-acetyb-galactosamine; Merck, Darmstadt, Germany) was passed through an anti-IgG-bound Sepharose column. The eluted was performed before addition of ‘251-ASOR. protein was shown to be free of the other contaminating serum For uptake studies, MC were incubated at 4#{176}Cin Hepes buffer proteins by immunoelectrophoresis and was shown to have the ap- containing 5 jg/ml radiobabeled proteins (ASOR or aggregates of propriate molecular size by sodium dodecyl sulfate-polyacrylamide IgAl, deslgAl, and IgA2). After 4 h of incubation, unbound proteins gel electrophoresis (SDS-PAGE) and gel filtration chromatography. were removed, and cells were incubated in Hepes buffer at 37#{176}Cinthe Purified human IgAl and human orosomucoid (Sigma Chemical Co., presence or absence of carbohydrates or ethylenediamine tetra-acetic St. Louis, MO) were desialylated by incubation for 6 h at 37#{176}Cin0.1 acid (EDTA) (Merck). At the appropriate time, the medium was M sodium acetate buffer, pH 5.0, containing neuraminidase (from removed, mixed with 20% tnichboroacetic acid, and centrifuged at Clostridium perfringens) attached to beaded agarose (Sigma) in a ratio 3000 X g, determining the amount of radioactivity in the supernatant. 0.03 U/mg of protein (17). The insoluble enzyme was removed by Degradation of ‘251-proteins was expressed as a percentage with centrifugation, and proteins were dialyzed against phosphate-buffered respect to total radioactivity bound to cells at 4#{176}C. saline (PBS). The extent of desialylation was monitored by the thio- barbituric acid assay (18), showing that 86 and 70% of sialic acid were Cell-Surface Biotinylation and Purification of ASGP-R removed from orosomucoid and IgAl, respectively. Human IgA2 was Biotinylation of MC was performed by the addition of 400 jsg/ml kindly provided by Dr. J. Mestecky (University of Alabama at Bin- sulfosuccinimidyl-6-(biotinamido) hexanoate-Biotin (Pierce Chemi- mingham). Proteins were radiolabeled with 1 mCi Na’251 (Amersham, cal Co., Rockford, IL) in 10 mM Na,B4O7, 150 mM NaCl, pH 8.8. Arlington Heights, IL) by the chloramine T method with a specific After 15 mm of incubation at room temperature, reactions were activity of 0. 1 mCi/mg. After radiolabeling, IgA1, desiabylated IgAl stopped by the addition of 2 mM NH4C1. Cells were washed, scraped (deslgAl), and IgA2 were heat-aggregated by incubation for 150 mm off the culture flasks, and spun down at 600 X g for 10 mm. The final at 63#{176}Cas described (19). These conditions do not seem to alter the biotinylated cell pellet was then resuspended in 1 ml of bysis buffer glycosylation of the IgAl molecule. ASOR was coupled to activated (20 mM Hepes, 150 mM NaCI, 50 mM NaF, I mM Na1VO4, 1 mM CNBr-Sepharose 4B (Pharmacia) at 5 mg proteinlml Sepharose in 0.2 ethylenegbycol-bis(3-aminoethyb ether)-N,N’-tetra-acetic acid, pH M NaHCO1, 0.5 M NaCl, pH 8.5. 7.5, 1% Nonidet P-40, 10 mM iodoacetamide, I mM phenylmethyb- sulfonyl fluoride, and 1 tg/mb antipain, chymostatin, beupeptin, and Cell Culture pepstatin A; Sigma) and incubated at 4#{176}Cfor 30 mm. Insoluble material was removed by centrifugation at I I ,600 X g for 10 mm. The MC were cultured from glomeruli according to a method described ASGP-R molecules were isolated from the cell extracts by affinity previously (15). Briefly, glomeruli were obtained from the renal chromatography on ASOR-Sepharose. Lysates were precleared with cortices of Sprague Dawley rats ( 150 to 200 g) or from human cadaver 200 p1 of BSA-Sepharose for 2 h at 4#{176}Cwith rotation, and then kidneys unsuitable for transplantation by sequential sieving and dif- incubated for 18 h at 4#{176}Cwith 200 pA of ASOR-Sepharose. The ferential centnifugation. The final preparation was treated with 500 pellets were washed, resuspended in Laemmbi’s sample buffer, sepa- U/ml collagenase (Type Ia; Sigma), and gbomeruli were resuspended rated by SDS-PAGE (10% acrylamide), and transferred to polyvinyli- in RPMI 1640 with 25 mM Hepes at pH 7.4, supplemented with 100 dene difluonide membranes. Detection of the biotinybated proteins U/mb penicillin, 100 p.g/ml streptomycin, 2 mM glutamine, and 10% bound to ASOR-Sepharose was made by incubation with streptavidin- fetal calf serum (FCS) (for rat MC) or 20% FCS (for human MC) horseradish peroxidase and enhanced chemibuminescence method (Life Technologies, Paisley, Scotland, United Kingdom), and cultured (Amersham). Films were quantified in a personal densitometer (Mo- at 37#{176}Cin 75-cm2 flasks (Costar, Cambridge, MA). Under such lecubar Dynamics, Sunnyvale, CA). conditions, MC began to spread after 7 to 14 d in culture, reaching

confluence by day 21 . MC were morphologically characterized by phase contrast microscopy, positive staining for the intermediate mRNA Expression filaments desmin and vimentin (which are considered specific for Total RNA was obtained by extraction with guanidine isothiocya- myogenic cells) and negative staining for factor VIII-related antigen nate, verifying the quality by ebectrophoresis in 1% agarose gel (20). 570 Journal of the American Society of Nephrology

One microgram of RNA was reverse-transcribed to single-stranded eDNA by incubation with 20 b of mixture (5 mM MgCl2, 10 mM A Tris-HCI, pH 8.8, 50 mM KC1, 0.1% Triton X-b00, 1 mM dNTP mixture, 50 ng of oligo(dT) primer, 20 U RNasin, and 15 U reverse -J transcriptase from avian myeboblastosis virus; Promega, Madison, E DO WI). The PCR was performed with specific primers (Genosys Bio- C technology, Cambridge, United Kingdom) for the two subunits of rat 0 C hepatic ASGP-R, RHL-l, and RHL-2 (21,22). The primer sequences 0 and the predicted length of eDNA amplification products were as follows: RHL-l (450 bp) sense: 5’-CGA AGC ITO AGC TGC CAG ATG GCC-3’, antisense: 5’-CCT TGC CCA ACT CTG TCT CAC A-3’; RHL-2 (400 bp) sense: 5’-TCC TGA GCA ACG GCA CAG AAT GC-3’, antisense: 5’-TFC ACA AGC CCA GCG OTt’ CAC 4 6 8 10 CT-3’. The mRNA expression for the glycerabdehyde 3-phosphate- Total (ug/mL) dehydrogenase bands was used as a control (23). Each PCR reaction containing 20 pmol of primers, 0.5 pCi [a32P]dCTP (3000 Cilmmol, Kd3.3x108 P.1 Amersham), and 3 U Taq DNA polymerase was incubated at 94#{176}Cfor n12x106 sites/cell 4 mm in a thermal cycler (Perkin Elmer, Norwalk, CT), and then a’ I several cycles were performed: 45 s at 94#{176}C;45 s at 60#{176}C;1 mm at a) 72#{176}C. Finally, extension was done at 72#{176}Cfor 7 mm. The PCR L&. 0 products were electrophoresed on 4% acrylamide gels and autoradio- C graphed. Films were quantified, and densitometric units of RHL- 1 and 0 ‘xl RHL-2 were calculated relative to the intensity of the glyceraldehyde 3-phosphate-dehydrogenase bands. In some cases, products were purified and used for hybridization in Northern blot analysis (24). Total cellular RNA (20 or 30 jtg per lane) was fractionated by a 1 % agarose gel containing formaldehyde and 10 15 transferred to nylon membranes (Genescreen, New England Nuclear, Bound (ng/mL) Boston, MA). The membranes were prehybridized for 6 h at 42#{176}Cin Figure 1. Binding of asialo-orosomucoid (ASOR) to cultured rat a 50% formamide solution, 1% SDS, 5X SSC, I X Denhardt’s sobu- mesangiab cells (MC). (A) Cells were incubated for 60 mm at 4#{176}Cin tion (0.02% Ficoll, 0.2% BSA, and 0.02% polyvinyl pyrrolidone), 0.1 Hepes buffer containing ‘251-ASOR (0.039 to 10 g/ml). Nonspecific mg/mI denatured DNA of salmon sperm, and 50 mM PBS, pH 6.5. binding and background were subtracted from each point. Mean ± SD Hybridization was performed at 42#{176}Cfor 20 h with 20% dextran of four experiments. (B) Scatchard analysis of the data from four sulfate in the presence of radiobabeled amplification products for different experiments. RHL- I and RHL-2. The membranes were washed, exposed to Kodak X-Omat films, and autoradiographed. As a control, to ensure that the same amount of RNA was loaded in each lane of the gel, the membranes were stripped and rehybridized with the ribosomal protein lation of binding sites with a Kd of 3.3 X lO_8 M and 1.2 X 285 probe (American Type Culture Collection, Rockvible, MD). Films 106 ASOR molecules bound/cell (Figure lB) were quantified, and densitometnic units of RHL- 1 and RHL-2 were We next examined the specificity of the binding of ASOR to calculated relative to the intensity of the 28S bands. rat MC. The competition curve for the 125I-ASOR binding is shown in Figure 2A. Inhibition of the binding increased with the concentration of unlabeled ASOR, reaching a 50% de- Statistical Analyses crease with a 57-fold excess of competitor. Because the ASOR Results are expressed as the mean ± SD. Significance was estab- molecule is composed of tetra-antennary oligosaccharides end- lished using GraphPAD Instat (GraphPAD Software). Dunn’s non- parametric test was used to compare the data. Differences were ing in galactose residues, we performed binding experiments in considered significant at P < 0.05. the presence of galactose. MC preincubation with different concentrations of this carbohydrate caused a gradual decrease in the binding of ‘25I-ASOR to rat MC (Figure 2B). No Results inhibition was observed in the presence of other carbohydrates, Binding of ‘251-ASOR to Rat and Human MC such as 0.5 M glucose and 0.5 M N-acetyl-galactosamine (6 ±

Quiescent MC were incubated with various concentrations 2 and 13 ± 2 percentages of inhibition, respectively, NS; n = of ‘25I-ASOR (0.039 to 10 pg/ml) for 60 mm. Binding studies 2 to 3). These results indicated that the ASOR interaction with were performed at 4#{176}Cto minimize internalization of the rat MC was specific for galactose. ligand, and nonspecific binding and background were sub- In some cases, we performed parallel binding studies using tracted from each point. As indicated in Figure 1A, the binding a homologous system (human ASOR and human MC), obtain- of ‘25I-ASOR to cultured rat MC was dose-dependent, and ing a binding curve similar to that seen with rat MC, and saturation was achieved at approximately S pg/ml. Scatchard reaching saturation at approximately 5 pg/ml ‘25I-ASOR (not analysis of the data from four different experiments revealed shown). Moreover, the binding of ‘251-ASOR to human MC that these receptors appear to represent a homogeneous popu- was blocked in the presence of a 700-fold excess of unlabeled Asiaboglycoprotein Receptors in Mesangial Cells 571

100 A

80 C 0 .4-, 0 C 6C 0 3 0 I- I!’ 0) 4C 0 ct: 0 2C U)

C ibo 1600 ioboo Unlabeled ASOR (ug/mL)

30 60 90 120 150 180 210 240 Time (mm)

Figure 3. Degradation of ‘251-ASOR by rat MC. Cells were incubated

C with 5 g/ml ‘251-ASOR. After 4 h at 4#{176}C.unbound proteins were 0 removed, and MC were incubated at 37#{176}Cincontrol conditions (S) or .0 -C in the presence of 5 mM ethylenediamine tetra-acetic acid (0) or 0. 1 C M galactose (A). The ability of MC to internalize and degrade the ASOR bound to the surface was determined as the increase in acid- soluble radioactivity accumulating in the medium with time. Data are expressed as percentage versus total binding at 4#{176}C.Mean ± SD of three experiments. 01 Calactose (M)

Figure 2. Specificity of the ‘251-ASOR binding to rat MC. MC were preincubated for 30 mm with several concentrations of unlabeled of 5 mM EDTA (Figure 3), suggesting that catabolism of ASOR (A) or galactose (B) before addition of ‘251-ASOR (7 tg/m1). ASOR is a process dependent on the presence of Ca2 in the Data are expressed as percentage versus total binding in the absence incubation medium. of unlabeled ASOR (A) or percentage of inhibition versus binding in the absence of galactose (B). Mean ± SD of four experiments. Affinity Purification ofASGP-R Molecules Rat MC were surface-biotinylated and solubibized with 1% Nonidet P-40 as described. The soluble fraction was initially ASOR (73 ± S percentage of inhibition, P < 0.01 ; n = 3) and precleared with BSA-Sepharose, and the supernatant was then by addition of 0.5 M galactose, but not with 0.5 M glucose adsorbed with ASOR-Sepharose. Analysis of the eluted mate- (76 ± 13 and S ± 4 percentage of inhibition, P < 0.01, and rial by SDS-PAGE and Western blot revealed two major bands NS, respectively; n = 2 to 3), thus confirming the specificity of responsible for the binding of ASOR. These two proteins had the interaction. apparent molecular weights of 41 and 55 kD (Figure 4). Quan- tification of the bands revealed a ratio of 1 .2: 1 (not shown). In Degradation of ‘251-ASOR by Cultured MC some experiments, lysates from MC were treated with several The ability of MC to internalize and degrade the ASOR doses of soluble galactose before affinity isolation on the bound to the surface was also investigated. The rate of degra- ASOR-Sepharose. This treatment caused a loss of intensity in dation was measured as the increase in acid-soluble radioac- the bands observed by SDS-PAGE and blotting, demonstrating tivity accumulating in the medium with time. As indicated in that the interaction was specific for galactose (Figure 4). Ga- Figure 3, degradation of ASOR was detectable after 15 mm of lactose incubation did not modify the relative amounts of the incubation at 37#{176}C,reaching a plateau at 120 mm. Degradation 41- and 55-kD proteins. was not a result of instability of ‘25I-ASOR because no change in the percentage of acid-soluble material was observed when mRNA Expression of the ASGP-R in Rat MC radiolabeled proteins were incubated in medium for 2 h at The expression of ASGP-R was studied by reverse tran- 37#{176}Cin the absence of cells (not shown). Degradation of scription-PCR with specific primers designed according to ‘251-ASOR was decreased by the addition of 0.1 M galactose, the published eDNA sequences of rat RHL-l and RHL-2 with maximal effect (approximately 80% inhibition) after 240 (21,22). As shown in Figure 5, mRNA expression for both mm of incubation (Figure 3). In contrast, neither glucose nor subunits of the ASGP-R was detected in growth-arrested rat N-acetyl-galactosamine blocked the degradation of ASOR by MC. The size of the PCR products corresponded with the rat MC (8 ± 3 and 14 ± 3, percentage of inhibition at I 20 mm, predicted size (450 bp for RHL-l; 400 bp for RHL-2). After

respectively, NS; n = 2). Degradation percentages were also quantification of the intensity bands and correction by the diminished when experiments were carried out in the presence glyceraldehyde 3-phosphate-dehydrogenase expression, we 572 Journal of the American Society of Nephrobogy

1234 97.4 0 C 0 - 68 U

IA

-46 -.0

I -31 a A -

Galactose (N)

Figure 4. Purification of asiaboglycoprotein receptor (ASGP-R). The ASGP-R molecules were purified from extracts of biotinylated rat MC by affinity chromatography on ASOR-Sepharose in the absence or presence of galactose. The left panel shows a representative experiment. Lane I, no addition: lanes 2, 3, and 4, addition of 0.01, 0. 1, and 0.5 M galactose, respectively. The position of molecular markers (in kibodaltons) is indicated. Bands corresponding to the cellular proteins bound to the ASOR-Sepharose are marked by arrows (41 and 55 kD). The right panel shows the densitometric analysis of both bands, expressed as percentage versus control in the absence of galactose. The data are representative of two experiments.

A transcripts was 2: 1 (not shown). We also analyzed the RHL-1 RHL-2 mRNA expression for RHL- 1 and RHL-2 in proliferating rat MC (culture medium with 10% FCS), but no significant cycles 35 40 35 40 differences in both subunits were detected compared with the expression in growth-arrested MC (not shown). #{149}45Obp #{149}-400bp

IgA I Is Able to Bind to the ASGP-R on Cultured MC Several works have revealed that hepatic ASGP-R binds B human IgAl , interacting with the 0-linked oligosacchanides in 1 2 12 the hinge region (I 6, 17). It has also been demonstrated that the carbohydrates present in the hinge region modulate the binding k... #{149} of IgA 1 to rat peritoneal macrophages and glomerular MC ( 15,25). Then, we explored the possibility that IgA 1 could

- RHL-2 mw..i#{243}-RHL-l Ip.O interact with the ASGP-R of rat and human MC through the 0-linked oligosaccharides present in the hinge region (galac- Figure 5. ASGP-R mRNA expression in rat MC. (A) The expression tose and N-acetyl-galactosamine). We studied the degradation of ASGP-R was studied by reverse transcription-PCR with specific rate of radiolabeled soluble IgA 1 aggregates (with similar primers for RHL-1 and RHL-2/3. After several amplification cycles, physicochemical properties to IgA 1 immune complexes) by rat the PCR products were electrophoresed and autoradiographed. The MC. In these experiments, we also used aggregates of IgAl length of cDNA amplification products is indicated. Blots are repre- desialybated with neuraminidase (deslgA 1), which exposes a sentative of three independent experiments. (B) Northern blot analysis of RHL- 1 and RHL-2 mRNA in rat MC. Total cellular RNA (30 jsg, greater number of sugars in the molecule. The kinetics of lane I : 20 g, lane 2) was electrophoresed in a 1 % agarose gel, catabolism of IgAl and deslgAl by rat MC was very similar, transferred to nylon membranes, and probed with the PCR amplifi- but percentages of degradation were bigger for deslgAl (27 ± cation products for RHL- I and RHL-2. Autoradiography results are 2 versus 21 ± S at 120 mm of incubation; ii = 3). As indicated representative of two experiments. Arrowheads indicate the migration in Table 1 , rat MC internalized and degraded IgAl and position of 285 and I 85 ribosomal RNA. deslgAl with dependence of the extracellular Ca2, because 5 mM EDTA treatment caused a reduction of the acid-soluble radioactivity. Degradation was also diminished by incubation observed that the RHL- 1 transcript was more abundant than of cells in the presence of several carbohydrates, observing the RHL-2 transcript, with a ratio of approximately 2: 1 (not maximal inhibition percentages when deslgAl degradation shown). Expression of both components for the ASGP-R was carried out in the presence of 0.5 M galactose (Table 1). was also analyzed by Northern blot, using the isolated PCR Human MC were also able to internalize and degrade soluble products for hybridization. As indicated in Figure 5, a band aggregates of IgA1 and deslgAl , and this process was partially for both RHL-l and RHL-2 mRNA was detected in RNA inhibited by the presence of 0.5 M galactose in the incubation from rat MC, comparable in size (approximately 1 .5 kb) to medium (46 ± 7 and 61 ± 9 percentage of inhibition, P < those described in hepatocytes (21,22). Densitometric anal- 0.01; n = 3). ysis of these bands, corrected by the 28S expression, re- To confirm that the 0-linked oligosacchanides present in the vealed that the relative amount of the RHL- 1 and RHL-2 hinge region of IgA 1 are essential for the interaction with Asiaboglycoprotein Receptors in Mesangial Cells 573

Table 1. Degradation of IgA 1, deslgA 1, and IgA2 Table 2. IgA1 and deslgAl, but not IgA2, inhibit the aggregates by rat MC0 binding of ‘251-ASOR to rat MC#{176}

Degr adation (% inhibition) Competitor ‘251-ASOR binding Addition (mg/mb) (% inhibition) ‘251-IgA1 ‘251-deslgAl ‘251-lgA2 ASOR (0.4) 68 ± 7h EDTA (5 mM) 51 ± 3” 54 ± 7” 38 ± 3C ASOR (1) 83 ± 2c 37 ± 5’ 48 ± 6b ND Galactose (0. 1 M) IgAl (0.4) 34 ± 8t) Galactose (0.5 M) 50 ± 3b 8’ 12 ± 7d IgAl (1) 39 ± 4C N-acetyl-galactosamine 35 ± 4” 33 ± 3b ND deslgA 1 (0.4) 43 ± 8h (0.1 M) deslgAl (1) 60 ± 5 N-acetyl-galactosamine 40 ± 1b 41 ± 5” 8 ± 6d IgAI (0.4) + deslgAl (0.4) 51 ± 7” (0.5 M) IgA2 (0.4) 1 1 ± 5d Glucose (0.5 M) 9 ± 3d 12 ± 4d 8 ± 3d a Binding experiments with ‘251-ASOR (7 g/ml) were

a MC were incubated with soluble aggregates of IgA 1 , deslgA 1, performed after a 30-mm preincubation with several concentrations or IgA2 (5 p.g/ml). After 4 h at 4#{176}C,unbound radiolabeled proteins of unlabeled ASOR, IgA 1, deslgA 1 , and IgA2. Inhibition were removed, and MC were incubated for 120 mm at 37#{176}Cin percentages were calculated in relation to total binding in the control conditions or in the presence of EDTA or several absence of competitor. Mean ± SD of two to four experiments. carbohydrates. Degradation was determined as acid-soluble ASOR, asiabo-orosomucoid. Other abbreviations as in Table 1. radioactivity in the incubation medium, expressed as percentage of “P < 0.001. C p < 0.05. inhibition versus control. Mean ± SD of two to four experiments. d Not significant with respect to control cells without competitor. deslgA 1, desialylated IgA 1; MC, mesangial cells; EDTA, ethybenediamine tetra-acetic acid; ND, not determined. bp < 0.001. cP < 0.05. Discussion ci Not significant with respect to control cells. The present study describes the first characterization of ASGP-R in cultured rat MC. The mesangial receptor displays characteristics nearby identical to those of the hepatic rat mesangial ASGP-R, catabolism experiments with soluble ag- ASGP-R: it is specific for asialoglycoproteins, interacts with gregates of IgA2 (which has a deletion in the hinge region and galactose terminal residues, and requires Ca2 for binding and lacks the 0-linked sugars) (26) were performed. Cultured rat internalization. In addition, MC exhibit a similar number of MC internalized and degraded IgA2 aggregates to a lesser ASGP-R molecules per cell (approximately 106) to those de- scnibed in freshly isolated hepatocytes and HepG2 cell line extent than IgA1 aggregates (15 ± 4 versus 21 ± 5; percent- (ranging from 0.5 X 106 to 1.2 X 106 receptors/cell) (3,27). ages of degradation at 120 mm of incubation; n 2 to 3), and However, affinity for the binding of ASOR to rat MC (Kd, this process was not affected by the presence of sugars (galac- approximately l0 M) was somewhat lower than that calcu- tose, N-acetyl-galactosamine, and glucose) in the incubation lated for hepatocytes (Kd, approximately l0 M) (3), but quite medium (Table 1). similar to that described in rat leukocytes (4). In other experiments, the binding of ‘251-ASOR to rat MC In this study, we described that the binding to extrahepatic was performed in the presence of unlabeled IgAl, deslgAl, ASGP-R is coupled to internalization and lysosomal degrada- and IgA2. As shown in Table 2, preincubation of rat MC tion of the ligand. Degradation of ASOR by rat MC occurred with IgAl and deslgAl caused a significant decrease in rapidly and was considerably reduced by treatment with ASOR binding. Inhibition of ASOR binding was dependent EDTA, suggesting that Ca2 may be necessary for the inter- on the concentration of unlabeled proteins, observing a action between ASGP-R and ASOR. A Ca2 requirement has maximal decrease in the presence of 1 mg/mI deslgA 1. also been found for several carbohydrate-binding proteins in Incubation with IgA2 did not cause any effect (Table 2). hepatocytes and peritoneal macrophages (3,17,28). Incubation with IgA 1 also resulted in a reduction of ASOR The stoichiometry and subunit composition of native rat binding to human MC (37 ± 12 percentage of inhibition, ASGP-R are still unknown. It has been suggested that rat P < 0.01; n = 3). hepatic ASGP-R is a heterohexamer composed of four RHL- I Competition of IgAl and deslgAl with ASOR for binding to subunits (41 to 43 kD) and two subunits of either RHL-2 (49 the ASGP-R was confirmed in the purification of the mem- to 54 kD) and/or RHL-3 (54 to 64 kD) (7,8,29). In rat MC, we brane receptor. Extracts from biotinylated rat MC were incu- have found by affinity chromatography on ASOR-Sepharose bated with IgA 1 or deslgA 1 before affinity purification of the two major bands responsible for the binding of ASOR, with ASGP-R with ASOR-Sepharose. As indicated in Figure 6, apparent molecular weights of 4 1 and 55 kD. These values are IgA1 or desigAl preincubation caused a reduction in the very close to those reported earlier for RHL- 1 and RHL-2/3 intensity of the two bands corresponding to the ASGP-R. A components of the rat ASGP-R in extrahepatic tissues (10). complete inhibition was obtained when lysates were preincu- Analysis of the relative amount of both components revealed bated with soluble ASOR. that membranes from rat MC contained more of the minor 574 Journal of the American Society of Nephrobogy

123

- 97.4

-68

- -46

- 31

Inhibitor Figure 6. Competition of IgA I and desiabybated IgAI (deslgAl) with ASOR for binding to the ASGP-R. Extracts from biotinybated rat MC were incubated with soluble ASOR, IgAl, and deslgAl before purification of the ASGP-R with ASOR-Sepharose. The left panel shows a representative experiment. Lane I . control cells; lane 2, pretreatment with 400 g/ml deslgA I ; lane 3, pretreatment with 400 p.g/ml ASOR. The right panel shows the densitometric analysis of the bands corresponding to the components of the ASGP-R, expressed as percentage versus control in the absence of inhibitor. Mean ± SD of three experiments.

component (RHL-2/3) relative to RHL-l, compared with the of two a 1 heavy chains, there are 10 closely adjacent 0-linked published hepatic content (10). This suggests that ASGP-R in sugars in the hinge region, which form a distinctive and Un- the kidney likely exists in the form of hetero-oligomeric com- usual feature of the IgAl molecule (26). plexes, but the stoichiometry could be different from that of the In a previous work, we described the interaction of IgA 1 hepatic receptor. The presence of both subunits of ASGP-R in with cultured rat and human MC. The binding was specific for cultured rat MC was also detected by mRNA expression anal- the Fe region of IgA, the isolated receptor consisted of a 60-kD ysis, with the RHL- I transcript more abundant than the RHL-2, protein, and its mRNA expression was confirmed by Northern as described in liver and other tissues ( 10). Similar mRNA blot using a probe for the myeloid FeaR (15). In leukocytes expressions of both subunits were detected both in proliferating and other cells, the ligand of FeaR lies at the very top of the and quiescent MC, indicating that ASGP-R expression is prob- CH2 domain of the IgAl molecule, adjacent to the hinge ably independent of the cell cycle. In contrast, the mitogenic region, and its sugars could be involved in the interaction stimulation of rat hepatocytes, by incubation with insulin and between IgA I and cells (3 1), but the existence of ASGP-R in epithelial growth factor, did modify the ASGP-R expression these cells has not been demonstrated. In our previous study, both at the mRNA and protein levels (30). we indirectly demonstrated that carbohydrates modulate the The receptor ligand specificity for terminal gabactosyl resi- interaction of IgAl with cultured MC. However, by Scatchard dues was also demonstrated. Binding, internalization, and deg- analysis of data binding, we did not detect a second type of radation of ASOR by rat MC were inhibited in the presence of binding site for IgAl on MC; therefore, there may exist two galactose, but not with other sugars, and isolation of membrane receptor types with similar affinities (15). To clarify this, in the receptors in an ASOR-Sepharose column was prevented by present work we investigated whether IgA 1 could interact with incubation with this carbohydrate. Thus, we conclude that the the ASGP-R on cultured MC through the 0-linked sugar res- renal ASGP-R resembles the classic hepatic receptor in many idues. The binding and degradation of IgAl by MC could be respects, including identical protein sizes and similar binding reduced by incubation with galactose or N-acetyl-galac- affinity and specificity. tosamine, and is probably dependent on the extracellular Ca2t Most of the data presented in this article were obtained using Both features have also been reported for the interaction of rat MC (heterobogous system). Therefore, we also performed ASOR and IgA 1 with the hepatic ASGP-R ( 16, 17). These parallel binding and uptake studies in a homologous system findings were corroborated by competition assays of the ASOR (human ASOR and human MC), obtaining similar binding binding in the presence of several forms of IgA. Binding of curves, as well as the same pattern of inhibition in the presence ASOR to cultured MC was reduced by addition of IgA1 , but of unlabeled ASOR, EDTA, or carbohydrates. not by IgA2. This result can be explained because the IgA2 has Several authors have reported indirect evidence (by ASOR a deletion of 13 amino acids in the hinge region (26). The lack binding inhibition experiments) that the ASGP-R expressed on of the five 0-linked sites of glycosylation in the IgA2 molecule hepatocytes is responsible for the binding of IgA 1 , interacting could explain the possible weak interaction with ASGP-R via with the 0-linked sugars present in the hinge region (16,17). the sugar residues in the two N-linked sites. In addition, inhi- The IgA 1 is a heavily glycosylated molecule, with two poten- bition of ASOR binding to MC in the presence of deslgAl was tial N-linked glycosylation sites in the CH3 domain (complex greater than that obtained with IgAl , indicating a major bind- and variable in structure) and five 0-glycosylation sites in the ing of deslgAl to the mesangial ASGP-R via the more abun- hinge region between the CH 1 and CH2 domains of the heavy dant galactose-ending oligosaccharides present in the desialy- chain (simple, composed by sialic acid, galactose, and N- lated molecule. Pretreatment with IgA1 and deslgAl also acetyb-gabactosamine). Because each IgAl monomer consists diminished the intensity of the bands corresponding to the Asiabogbycoprotein Receptors in Mesangial Cells 575

ASGP-R subunits obtained with the ASOR-Sepharose column itate their glomerular deposition with the subsequent triggering purification. Although all of these results are indirect evidence of immune reactions in MC. of the IgA1 interaction with the mesangial ASGP-R (because only inhibition studies were performed), they seem to confirm Acknowledgments that binding of IgA1 partially occurs via this receptor. This work was supported by grants from Comunidad Aut#{243}noma de Because catabolism of soluble IgA1 aggregates could be Madrid (CAM 08/076/96), Fondo de Investigaciones Sanitanias (94/ partially inhibited by IgA 1 Fe fragments ( 19) and carbohy- 370, 96/202 1), and Ministerio de Educaci#{243}n (PB-94/2 I 1 , PM-95/93). drates in vitro, we hypothesize that when IgA1 is presented as Dr. G#{243}mez-Guerrero is a postdoctoral fellow from CAM and Dr. a ligand for MC, it is recognized both by the FeaR (with Duque is a fellow from Fundaci#{243}nConchita R#{225}bago. affinity for the Fe region) and by the ASGP-R (with affinity for the hinge region carbohydrates), and both receptors could References contribute to the uptake and handling of IgAl complexes in the 1 . Stockert Ri: The asialoglycoprotein receptor: Relationships be- glomerular mesangium in viva. In this regard, the contribution tween structure, function, and expression. Phvsiol Rev 75: 591- of specific lectins to the phagocytosis mediated by Fe receptors 609, 1995 for IgG in mouse peritoneal macrophages has been reported 2. Ashwebb 0, Harford J: Carbohydrate-specific receptors of the (28,32). Immune aggregates of IgA, 1gM, and IgG may also be liver. Annu Rev Biochem 51: 531-554, 1982 3. Schwartz AL, Fnidovich SE, Knowles BB, Lodish HF: Charac- removed from the mesangium through cytokine-induced man- terization of the asiaboglycoprotein receptor in a continuous nose receptors (33). hepatoma line. J Biol Chem 256: 8878-8881, 1981 However, other functional events, such as intracellular Ca2 4. Bezouska K, Taborsky 0, Kubrycht J, Pospisil M, Kocourek J: mobilization, were abolished by IgA1 Fe fragments, but not by Carbohydrate-structure-dependent recognition of desialybated se- carbohydrates (34), suggesting different signal transduction rum glycoproteins in the liver and leucocytes. Biochem J 227: pathways for FeaR and ASGP-R. Recent works have indicated 345-354, 1985 that tyrosine phosphorylation is required for activation via 5. Dini L, Autuori F, Lentini A, Oliverio 5, Piacentini: The clear- FeaR in monocytes and MC (35,34). Although the RHL- 1 ance of apoptotic cells in the liver is mediated by the asiabogly- subunit of the hepatic ASGP-R is tyrosine-phosphorylated coprotein receptor. FEBS Lett 296: 174-b 78, 1992 before internalization (36), more information is needed for a 6. Treichel U, Meyer zum Btischenfelde KH, Stockert Ri, Porabla better understanding of the intracellular signals involved in the T, Gerken 0: The asiaboglycoprotein receptor mediates hepatic binding and uptake of natural hepatitis B virus particles derived activation through ASGP-R. from viraemic earners. J Gen Virol 75: 3021-3029, 1994 The data of the present study could be relevant for the 7. Dnickamer K: Membrane receptors that mediate glycoprotein pathogenesis of IgA-related diseases, such as IgA nephropathy. endocytosis: Structure and biosynthesis. Kidney hit 32: 5167- Recent works have suggested that the predominance of IgA1 in 5180, 1987 mesangial deposits in patients with IgA nephropathy may be 8. Habberg DF, Wager RE, Farrell DC, Hibdreth J, Quesenberry MS. due to an altered glycosylation pattern of IgAl molecules Loeb JA, Holland EC, Drickamer K: Major and minor forms of (3 1 ,37,38). Defective galactosylation reduces IgA 1 affinity for the rat liver asiaboglycoprotein receptor are independent galac- the hepatie ASGP-R, resulting in a failure of the normal tose-binding proteins: Primary structure and glucosylation het- catabolic route (37). We speculate that defective clearance of erogeneity of minor receptor forms. J Biol Che,iz 262: 9828- IgAl by the ASGP-R in liver and extrahepatic tissues (such as 9838, 1987 the kidney) in IgA nephropathy patients could contribute to the 9. Zeng FY, Weigel PH: Hydroxybamine treatment differentially inactivates purified rat hepatic asiaboglycoprotein receptors and persistence of circulating IgA 1 immune complexes and their distinguishes two receptor populations. J Biol Chem 270: deposition in the kidney. In these situations, the interaction of 21388-21395, 1995 immune complexes with mesangial FeaR could induce pro- 10. Mu JZ, Tang LH, Alpers DH: Asiaboglycoprotein receptor mR- duction of several cytokines (19,39) involved in mesangial NAs are expressed in most extrahepatic rat tissues during devel- proliferation and matrix protein synthesis, common features in opment. Am J Phvsiol 264: 0752-0762, 1993

immune complex-mediated glomerulonephritis. Recent data 1 1 . Pacifico F, Laviola L, Ulianich L, Porcelbini A, Ventra C, Con- have shown that cytokines inhibited the mRNA expression of sigbio E, Avvedimento VE: Differential expression of the asia- ASGP-R on hepatocytes and the binding and internalization of boglycoprotein receptor in discrete brain areas, in kidney and ligands (30,40). Thus, the presence of cytokines in the mesan- thyroid. Biochem Biopkvs Res Commun 210: 138 -144, 1995 gium milieu during glomerular injury could decrease the ca- 12. Mu JZ, Falbon Ri, Swanson PE, Carroll SB, Danaher M, Alpers tabolism of IgA1 complexes, favoring their persistence at this DH: Expression of an endogenous asialoglycoprotein receptor in a human intestinal epitheliab cell line, Caco-2. Biochem Biophys level. Acta 1222: 483-491, 1993 Taken together, the present data indicate that MC possess an 13. Men#{232}P. Simonson MS. Dunn Mi: Physiology of the mesangial ASGP-R with specificity for the galactose terminal residues of cell. Physiol Rev 69: 1347-1424, 1989 several glycoproteins, including IgAl , suggesting that these 14. Fogazzi GB, Sheerin NS: IgA-associated renal diseases. Curr cells may participate in the clearance of glycoproteins depos- Opin Nephrol Hypertens 5: 134-140, 1996 ited in the mesangial area. We propose that in situations with 15. G#{243}mez-Guerrero C, Gonzalez E, Egido J: Evidence for a specific IgA-defective galactosylation, such as in IgA nephropathy, the IgA receptor in rat and human mesangial cells. J Immunol 151: decreased systemic clearance of IgAl by ASGP-R could facil- 7172-7181, 1993 576 Journal of the American Society of Nephrobogy

16. Stockert Ri, Kressner MS. Collins JC, Sternlieb I, Morell AG: 29. Weigel PH: Endocytosis and function of the hepatic asiabogly- lgA interaction with the asiaboglycoprotein receptor. Proc Nail coprotein receptor. In Subcellular Biochemistry, edited by AcadSci USA 79: 6229-6231, 1982 Bergeron JJM, Harris JR. New York, Plenum, 1993, pp 125-16 I I 7. Tomana M, Kubhavy R, Mestecky J: Receptor-mediated binding 30. Conti-Devirgiliis L, Massimi M, Bruscabupi 0, Febici A, Dini L: and uptake of immunoglobulin A by human liver. Gastroenter- Regulation of asialoglycoprotein receptor expression in rat hepa- ologv 94: 762-770, 1988 tocytes cultured under proliferative conditions. Exp Cell Res 210: 18. Aminoff D: Methods for the quantitative estimation of N-acetyb- 123-129, 1994

neuraminic acid and their application to hydrolysates of siabo- 3 1 . Allen AC: Abnormal glycosylation of IgA: Is it related to the mucoids. Biochem J 81: 384-392, 1961 pathogenesis of IgA nephropathy? Nephrol Dial Transplant 10: 19. G#{243}mez-Guerrero C, Lopez-Armada MJ, Gonzalez E, Egido J: 1121-1124, 1995 Soluble IgA and IgG aggregates are catabolized by cultured rat 32. Nose M, Wigzelb H: Biological significance of carbohydrate mesangial cells and induce production of TNF-a and IL-6, and chains on monocbonal antibodies. Proc Nail Acad Sci USA 80: proliferation. J Immunol 153: 5247-5255, 1994 6632-6636, 1983 20. Chomczynski F, Sacchi N: Single-step of RNA isolation by acid 33. Lin ZH, Striker GE, Stetber-Stevenson M, Fukushima P. Pateb A, guanidinium thiocyanate-phenob-chboroform extraction. Anal Striker Li: TNFa and IL- 1a induce mannose receptors and Biochein 162: 156-159, 1987 apoptosis in gbomerular mesangial but not endothebiab cells. Am J

21 . Holland EC, Leung JO, Drickamer K: Rat liver asialoglycopro- Physiol 270: C1595-C1601, 1996 tein receptor lacks a cleavage NH2-terminab signal sequence. 34. G#{243}mez-Guerrero C, Duque N, Egido I: Stimulation of Fca Proc Nail Acad Sci USA 81: 7338-7342, 1984 receptors induces tyrosine phosphorylation of phosphobipase 22. McPhaul M, Berg P: Identification and characterization of eDNA C--yl , phosphatidybinositol phosphate hydrolysis, and Ca2 mo- clones encoding two homologous proteins that are part of the bilization in rat and human mesangiab cells. J Imniunol 156: asiaboglycoprotein receptor. Mol Cell Biol 7: 1841-1847, 1987 4369-4376, 1996 23. Fort P. Marty L, Piezchaczyk M, Sabrouty SE, Dani C, Jeanteur 35. Pfefferkorn LC, Yeaman OR: Association of IgA-Fc receptors P, Blanchard JM: Various rat adult tissues express only one (FeaR) with Fc#{128}R1y2subunits in U937 cells. J Immunol 153: major mRNA species from glyceraldehyde-3-phosphate-dehy- 3228-3233, 1994 drogenase multigenic family. Nucleic Acids Res 13: 1431-1442, 36. Haynes PA, Oka JA, Weigeb PH: The rat hepatie bectin 1 subunit 1985 of the rat asiaboglycoprotein receptor is a phosphoprotein and 24. Sambrook J, Fritsch EF, Maniatis T: Analysis of RNA. In: contains phosphotyrosine. J Biol Chem 269: 33 146-33 15 1, 1994 Molecular Cloning: A Laboratory Manual, 2nd Ed., edited by 37. Mestecky J, Tomana M, Crowley-Nowick PA, Moldoveanu Z, Nolan C, Cold Spring Harbor, NY, Cold Spring Harbor Labora- Jilian BA, Jackson 5: Defective galactosylation and clearance of tory Press, 1989, pp 7.37-7.52 IgA1 molecules as a possible etiopathogenic factor in IgA ne- 25. Gorter A, Hiemstra PS, Van der Voort EAM, Van Es LA, Daha phropathy. Contrib Nephrol 104: 172-182, 1993 MR: Binding of human IgAl to rat peritoneal macrophages. 38. Kokubo T, Hiki Y, Iwase H, Tanaka A, Nishikido J, Hotta K, Inununology 64: 207-212, 1988 Kobayashi Y: Increased reactivity of polycbonal anti-synthesized 26. Kerr MA: The structure and function of human IgA. Biochem J IgAl hinge peptide antibody against IgAl in IgA nephropathy 271: 285-296, 1990 [Abstract]. J Am Soc Nephrol 7: 1707, 1996 27. Steer C, Ashwelb G: Studies on a mammalian hepatic binding 39. Lopez-Armada Mi, G#{243}mez-Guerrero C, Egido J: Receptors for protein specific for asiaboglycoproteins: Evidence for receptor immune complexes activate gene expression and synthesis of recycling in isolated rat hepatocytes. J Biol Chem 255: 3008- matrix proteins in cultured rat and human mesangial cells: Role 3013, 1980 of TGF3. J Immunol 157: 2136-2142, 1996 28. Aramaki Y, Murai M, Tsuchiya 5: Contribution of N-acetyb-f3- 40. Treicheb U, Paietta E, Poralba T, Meyer zum BOschenfebde KH, D-galactosamine-specific bectin to Fe receptor-mediated phago- Stocker RJ: Effects of cytokines on synthesis and function of the cytosis by mouse peritoneal macrophages. Immunology 79: 403- hepatic asiaboglycoprotein receptor. J Cell Phvsiol 158: 527-534, 407, 1993 I 994