A Monoclonal Antibody That Specifically Recognizes a Glucuronic ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society of Biological Chemists, Inc.

Vol. 262, No. 9,Issue of March 25, pp. 4146-4152,1987 Printed in U.S.A.

A Monoclonal Antibody That Specifically Recognizes a Glucuronic Acid 2-Sulfate-containing Determinant in Intact Chondroitin Sulfate Chain* (Received for publication, December 9, 1986)

Masahito Yamagata, Koji Kimata, Yasuteru Oike, Katsuko Tani, Nobuaki Maeda,Keiichi YoshidaS, Yukio Shimomura, Masahiko Yoneda, andSakaru SuzukiP From the Department of Chemistry, Faculty of Science, Nagoya University, Nagoya464 and the $Tokyo ResearchInstitute, Seikngaku Kogyo Co. Ltd., Tokyo 189,Japan

Monoclonal antibodies produced against chick embryo limb bud proteoglycan (PG-M) were selected for their ability to recognize determinants on intact chondroitin sulfate chains. One of these monoclonal antibodies (IgM; designated MO-225) reacts with PG-M, chick embryo cartilage proteoglycans (PG-H, PG-Lb, and PG-Lt), and bovine nasal cartilage proteoglycan, but not with Swarm rat chondrosarcoma proteoglycan. The reactivity of PG-H to MO-225 is not affected by keratanase digestion but is completely abolished after chondroitinase digestion. Competitive binding analyses withvarious glycosaminoglycan samples indicate that the determinant recognized by MO-225 resides in a D-glucuronic acid 2-sulfate(~1+3)N-acetylgalactosamine 6-sulfate disaccharide unit (D-unit) common to antigenic chondroitin sulfates. A tetrasaccharide trisulfate containing D-unit at the reducing end is the smallest chondroitin sulfate fragment that can inhibit the binding of the antibody to PG-H. Decreasing the size of a D-unit-rich chondroitin sulfate by hyaluronidase digestion results in progressive reduction in its inhibitory activity. The results suggest that the epitope has a requirement for a long stretch of adisacchariderepeating structure for a better fit to the antibody.

In recent years a large number of apparently different chondroitin sulfate proteoglycans have been isolated. They are primarily located in extracellular matrices where they appear to help maintain tissue architecture and influence cell growth and differentiation (for reviews, see Refs. 1-3). Some, such as those intercalated in the cell membrane of melanoma cells (4) or those located in the secretory granules of mouse mast cells (5) and human natural killer cells (6), appear to have different, more specialized functions. In early studies two types of chondroitin sulfate were distinguished, chondroitin 4-sulfate composed of repeating GlcA’-GalNAc 4* This work was supported by grants-in-aid for Cancer Research, Scientific Research, and Special Project Research from the Ministry of Education, Science and Culture, Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. § To whom correspondence should be addressed. The abbreviations used are GlcA, glucuronic acid; PG-M, the major chondroitin sulfate proteoglycan of stage 22-23 chick embryo limb bud; PG-H, PG-Lb, and PG-Lt, three distinct proteoglycan components of 12-day-oldchick embryo cartilage; AlD1, large aggregating chondroitin sulfate proteoglycan monomer from bovine nasal cartilage; Agg-Dl, large aggregating chondroitin sulfate proteoglycan

SO4disaccharide units andchondroitin 6-sulfate composed of repeating GlcA4alNAc 6-S04disaccharide units (7). However, it has since been shown that considerable chemical variability can be superimposed upon the component sugar units. Thus,the galactosamine residues can be 4,6-bissulfated either at the nonreducing terminal (8) or in the internal portion (9) of chondroitin sulfate. The glucuronic acid residues, on the other hand, can be sulfated (10) at either position 2 or position 3 (11).It appears that most if not allchondroitin sulfate proteoglycans carry copolymers of these sulfated disaccharide units inwhich the percentages of the disaccharides vary with the source and age of the tissue (for a review, see Ref. 12). Recent immunological studies of proteoglycans have revealed that epitopes for antibody binding exist not only on the core proteins but also on intact keratan sulfate (13, 14) and chondroitin sulfate side chains (15). The latter antibody was prepared against ventral membranes of cultured chick fibroblast and has been shown to react primarily with the chondroitin sulfate associated with cell surface membranes. However, there has been no information on the precise antigenic structure recognized by this antibody. Recently, we have prepared a number of monoclonal antibodies against two distinct chondroitin sulfate proteoglycans, PG-H (16) and PG-M (17), isolated from 12-day-old chick embryo epiphysial cartilage and stage 22-23 chick embryo limb buds, respectively. In this paper, we report that one of these antibodies specifically recognizesa GlcA 2-S04-containing determinant contained in the chondroitin sulfate chains of proteoglycans. EXPERIMENTAL PROCEDURES

Materials Proteoglycans-The following proteoglycans were prepared by previously described methods: PG-Mand [35S]sulfate-labeledPG-M from stage 22-23 chick embryo limb buds (17); PG-H ( X ) , [35S] sulfate-labeled PG-H (X),PG-Lb (18),and PG-Lt (19) from 12-daymonomer from Swarm rat chondrosarcoma; GlcA 2-so4, D-glucuronic acid 2-sulfate; IdoA %so4, L-iduronic acid 2-sulfate; GlcNAc 4-so4, N-acetyl-D-galactosamine4-sulfate; GalNAc 6-SO4, N-acetyl-D-galactosamine 6-sulfate; GalNAc 4,6-bis-S04,N-acetyl-D-galactosamine 4,6-bissulfate; AGlcA,A‘-glucuronic acid; ADi-OS, 2-acetamido-2deoxy-3-O-(~-~-gluco-4-enepyranosyluronic acid)-D-galactose; ADi4S, ADi-GS, ADi-diSD, and ADi-diSE, derivatives ofADi-OS bearing one sulfate at position 4 of the hexosamine, one sulfate at position 6 of the hexosamine, two sulfates at position 2 of the hexuronate and position 6 of the hexosamine, and two sulfates at position 4 and position 6 of the hexosamine, respectively; ELISA, enzyme-linked immunosorbent assay; TPCK, N-tosyl-L-phenylalanyl chloromethyl ketone; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraaceticacid.

4146

Monoclonal Chondroitin Intact Sulfate Antibody to

4147

TABLEI Inhibition of binding of MO-225 to PG-H by chondroitin sulfates of different disaccharide unit compositions Disaccharide unit composition* ICd Source of chondroitin sulfate D-unit C-unit A-unitNonsulfated E-unit Pghl

%

Shark fin 6.7 21.4 21.2 43.4 0.004 0.8 Fraction I11 1.3 26.2 17.5 52.5 1.6 0.02 Fraction I1 0.8 23.7 13.9 58.6 0.08 2.0 Fraction I 0.9 8.4 14.8 73.4 2.0 0.43 Shark scapular cartilage 0.4 3.0 74.2 19.4 2.1 4.3 Whale cartilage 61.0 0‘ 21.2 9.8 0 30 Squid cartilage 0.5 2.0 29.4 68.1 NDd 170 Chick embryo cartilage (PG-H) 0.1 0.5 12.4 74.6 12.3 430 Human umbilical cord 0.2 0 3.3 92.5 1.3 >3300 Sturgeon notochord 0 0 0 0 100 >3300 Squid skin (chondroitin) 0 0 89.4 0.1 10.5 >3300 Swarm rat chondrosarcoma (Agg-Dl) . . Concentration (Fg/ml) giving 50% inhibition in ELISA on PG-H. Percent of total hexuronate content or total 35S activity (chick embryo cartilage preparation only) estimated from the yields of the unsaturated disaccharides produced by digestion with chondroitinases; nonsulfated unit, GlcA-+GalNAc; A-unit, GlcA-GalNAc &so4; C-unit, G l c A 4 a l N A c 6-S04; D-unit, GlcA 2-S04-GalNAc 6SO4;E-unit, GlcA-GalNAc 4,6-bis-SO4. e About 7% of the total hexuronate was recovered as glucose-carrying pentasaccharides and higher oligosaccharides with different sulfate contents (see the text for details). ND, not determined because 35Swould not label nonsulfated disaccharide units. ”

old chick embryo epiphysial cartilages; AlDl from bovine nasal cartilage (20); and Agg-Dl from Swarm rat chondrosarcoma (21). Glycosaminoglycans-Shark fin chondroitin sulfates (FractionI, 11, and 111) were prepared as follows. About 120 g of the fin of shark (Glyphisglaucus) was minced and digested a t 55 “C for 18 h with 60 mg of Prolisin in 60 mlof water. The digest was made 0.5 M in NaOH and kept at 40 “C for 1 h. After neutralization with HCl, insoluble materials were removedby filtration throughCelite. From the filtrate, chondroitin sulfate was precipitated with 1.5 volumes of 95% (v/v) ethanol containing 1.3% (w/v) sodium acetate, washed successively with ethanol and ether, and dried in a vacuum over P205; yield, 5 g. The chondroitin sulfate preparation was dissolved in 100 ml of 1 M NaCl and applied to a Diaion HPA-10 (anion exchange resin) column (2.6 X 55 cm) that had been equilibrated with 1 M NaCl. The column was washed with 600 ml of 1 M NaCl and then eluted stepwise with 400 ml of 1.5 M NaCl, 180 ml of 1.5 M NaCl, and 450 ml of 2.5 M NaCl. Effluents were dialyzed free of NaCl and evaporated at reduced pressure to a concentration near 10 mgof polysaccharide/ml. The chondroitin sulfate was precipitated with ethanol, washed, and dried as above:yield,1.9 g (the first 1.5 M NaCl fraction, designated Fraction I), 1.21 g (the second 1.5 M NaCl fraction, designated Fraction II), and 0.83 g (the 2.5 M NaCl fraction, designated Fraction 111). The composition of disaccharide repeat units of these preparations is given in Table I. Chick embryo cartilage chondroitin [35S]sulfateand Swarm rat chondrosarcoma chondroitin sulfate were prepared from [35S]sulfatelabeled PG-H (16) and unlabeled Agg-Dl (21), respectively, by 0elimination procedures followed by gel chromatography (22). Shark scapular cartilage chondroitin sulfate (Commercial name, “chondroitin sulfate C”) and whale cartilage chondroitin sulfate (Commercial name, “chondroitin sulfate A ) were the products of Seikagaku Kogyo Co., Tokyo (note that their structures are more complex than those suggested by the commercial names, see Table I). We are very grateful tothe following individuals for generous gifts of the substances indicated: human umbilical cord chondroitin sulfate and sturgeon notochord chondroitin sulfate from M. B. Mathews (Department of Pediatrics, University of Chicago); squid cartilage chondroitin sulfate (9,23) from T. Harada (Seikagaku Kogyo Co.);squid skin chondroitin (24) from H. Inoue (in this laboratory); bovine cornea keratan sulfate (25) from K. Nakazawa (Meijo University, Nagoya); and porcine lung heparan sulfate from S. Suzuki (Eisai Co., Toyosato, Japan). Other glycosaminoglycans and related compounds used for demonstrating the specificity of the monoclonal antibodies were as follows: calf thymus DNA and porcine intestinal mucosa heparin (type I) from Sigma; pig skin dermatan sulfate, cockscomb hyaluronic acid, and shark cartilage keratan sulfate from Seikagaku Kogyo Co. (Tokyo); and dextran sulfate from Nakarai Co. (Kyoto, Japan).

Oligosaccharides-A mixture of tetrasaccharides from whale cartilage chondroitin sulfate was prepared by digestion with testicular hyaluronidase followed by gelchromatography as described by Flodin et al. (26). ADi-OS, ADi-4S, ADi-6S,ADi-diSD,and ADi-diSE (9) were the products of Seikagaku Kogyo Co., Tokyo. The tetrasaccharides AGlcA((31-3)GalNAc 4-S0&31+4)GlcA 2S04(pl-3)GalNAc 6-SO4 and AGlcA(B1-3)GalNAc 4,6-bisSOa(p1+4)GlcA 2-S04(p14)GalNAc 6-SO4 were prepared from shark fin chondroitin sulfate Fraction 111 (see above) by taking advantage of the fact that chondroitinase AC I Flavobacterium heparinurn* acts much more slowly on the hexosaminidic linkages to GlcA % s o 4residues in chondroitin sulfate than on the hexosaminidic linkages to nonsulfated glucuronosyl residues (29). Briefly, a solution of 100 mgof shark fin chondroitin sulfate Fraction 111 in 2 ml of 0.025 M Tris buffer, pH 7.2, containing 0.4 mg of calcium acetate, 2.5 mg of bovine serum albumin, and 10 units of chondroitinase AC I F. heparinum (27) was incubated at 37 “C for 16 h. The reaction was stopped by placing the tube in a boiling water bath for 2 min. After centrifugation at 900 X g for 5 min, the clear supernatant was concentrated to a small volume at reduced pressure and then chromatographed on a Cellulofine GCL-90-m column (3 X 96 cm) equilibrated and eluted with 0.5 M NaC1. The eluates were screened for hexuronic acid-positive material. Three peak materials corresponding in Kd to hexasaccharide, tetrasaccharide, and disaccharide were obtained in the hexuronate ratio of 1:1617. The tetrasaccharide fractions were pooled and subjected to ion exchange chromatography on an AG 1-X4 column (2.2 X 30 cm) eluted with a linear NaCl concentration gradient (1-3.5 M). The column yielded two peak materials (hexuronate ratio = 3.2:l) eluting at about 2 M NaCl and 3 M NaCl, respectively. The identification of the 2 and 3 M NaCl fraction as the tetrasaccharide tri- and tetrasulfate, respectively, was based on ( a ) an absorption maximum a t 232 nm, ( b ) the molar ratios of total hexuronate, galactosamine, and sulfate to AGlcA, (c) the formation of equimolar amounts of ADi-4S and ADi-diSD (from the trisulfate) or ADi-diSE and ADi-diSD (from the tetrasulfate) on chondroitinase ABC digestion, and (d) the susceptibility of the GlcA 2-SOd-GalNAc 6-SO4disaccharide unit (in both tetrasaccharides) to reduction with NaBH4 (see Ref. 30 for the analytical methods based on the selective reduction with NaBH4). Other Materials-Chondroitinase ABC (Proteus vulgaris) (27), chondroitinase AC I F. heparinum (see Footnote 2 for the commercial name), chondroitinase AC I1 Arthrobacter aurescens (A. aurescens) The name given to a commercial preparation of Flavobacterium chondroitinase AC (27). A similar, but apparently different chondroitinase prepared from A. aurescens (28) has been designated as “chondroitinase AC I1 A. aurescens.”

4148

Chondroitin Intact Monoclonal Sulfate Antibody to

(28), karatanase (Pseudomonas sp.) (31), and Cellulofine GCL-90-m antibodies were then added and allowed to react with the antigens as were the products of Seikagaku Kogyo Co. (Tokyo). Pronase was the above. The concentration of inhibitor required for 50% inhibition product of Kaken Seiyaku Co. (Tokyo), and was kindly donated by ( I C d was calculated from plots of A d s o nm against inhibitor concentrathe company. TPCK trypsin and testicular hyaluronidase Type I were tion. purchased from Sigma; Prolisin was from Ueda Kagaku Co. (Osaka, Radioimmuoprecipitation" 100-pl solution of [36S]sulfate-laJapan); Diaion HPA-10 was from Mitsubishi Kasei Co. (Tokyo); AG beled PG-M or PG-H in phosphate-buffered saline/Tween 20 (see 1-X4 (200-400 mesh) was from Bio-Rad; fetal calf serum was from above) was mixedwith appropriate volumes of ascites fluids containCommonwealth Serum Laboratories (Melbourne, Australia); chemi- ing MO-225 or HM-110 and the mixtures were incubated for 12 h at cals for cell culture were from Nissui Seiyaku Co. (Tokyo); polyeth- 4 "C. Then, 10 pl of goat anti-mouse IgG antibody (y- and L-chain ylene glycol 6000 was from Nakarai Chemical Co. (Kyoto, Japan); specific) wasadded as thesecond antibody and incubation continued pristane was from Aldrich Chemical: goat anti-mouse IgG (7-and L- for an additional 12 h at 4 "C. The solution was mixedwith 50 pl(bed chain specific), peroxidase-conjugated goat anti-mouse IgM (p-chain volume) of Protein A-Sepharose in 100pl of phosphate-buffered specific), and peroxidase-conjugated goat anti-mouse IgG IgM (7-, saline/?'ween 20 and the mixture was allowed to stand at room p-, and L-chain specific) were from Tag0 Co.; polyvinyl ELISA plates temperature for 1 h. The resultant immunoprecipitates were pelleted (96-well)were fromSumitomo Bakelite (Osaka, Japan); mouse mono- by centrifugation at 1200 X g for 15 min a t 4 "C. The pellets were clonal isotyping kit was from Zymed Laboratories (South San Fran- washed three times with phosphate-buffered saline/Tween 20. Radiocisco, CA); and Sephadex G-100 and Protein A-Sepharose were from activity of the washed precipitates was measured in a liquid scintilPharmacia Japan (Tokyo). lation spectrometer. Treatment of Proteoglycans with Proteases-PG-H (2 mg) was Methods incubated for 10 h at 37 "C in 100 p1 of 0.1 M NH4HC03, pH 7.8, Immunization, Fusion, and Cloning-Three 6-week-old female containing 10 pg of TPCK trypsin or Pronase. Enzymes were inactiBALB/c mice were injected intraperitoneally with either PG-M or vated by heating in a boiling water bath for 5 min. The solutions PG-H solutions (50 pg of the antigen/mouse, emulsified in complete were lyophilized and the residues were dissolved in appropriate volFreund's adjuvant). Injection of immunogen in incomplete Freund's umes of phosphate-buffered saline and used for competitive inhibition adjuvant was repeated twice in %week intervals and theformation of tests in ELISA (see above). Treatment of Chondroitin Sulfates with Testicular Hyaluronidaseantibodies was monitored by ELISA (see below). Two days after the final injection, the spleens of two mice showing high antibody titers To a solution of 1 mg of shark scapular cartilage chondroitin sulfate were removed and the cells from spleen were fused with NS-1 mye- in 1ml of 0.01M sodium acetate/acetic acid, pH 5.0, containing 0.15 loma line (8-azaguanine-resistant) at a lymphocyte to myeloma ratio M NaC1, 50 pg of testicular hyaluronidase (415 units/mg) was added and the mixture was incubated at 37 'C for up to 24 h. At different of 101, using 20% (w/v) polyethylene glycol 4000 (32). The fusion time intervals, aliquots of the reaction mixture were withdrawn and products were suspended in hypoxanthine/aminopterin/thymidinethe enzyme was denatured by heating at 100 "C for 5 min. After containing medium (RPMI 1640, 10%fetal calf serum). The cell centrifugation a t 9000 x g for 5 min, the supernatantswere subjected suspension was aliquoted (1 X 10' cells/well) into 96-well plates to competitive inhibition tests by ELISA (see above). supplemented with feeder cells from a peritoneal lavage of a pristaneFor preparation of small-size chondroitin sulfate, shark cartilage treated BALB/c mouse and the plates were cultured at 37 "C in an chondroitin sulfate was digested with testicular hyaluronidase for 3 incubator in the presence of 5%COP. After 2 weeks of culture, h as above, and the supernatant from the digest was chromatographed hybridomas showing positive reaction in ELISA on PG-M or PG-H on a Sephadex G-100 column (calibrated with standard chondroitin were cloned by limiting dilution. From these, a clone that reacted sulfates of the known molecular weights) equilibrated and eluted with with PG-M but not with chondroitinase ABC-treated PG-M and a phosphate-buffered saline. Fractions corresponding in position to a clone that reacted with PG-H butnot with keratanase-treated PG-H M, = 10,000 chondroitin sulfate were pooled. were selected. Selected hybridoma clones were stored under liquid Other Methods-Hexuronic acid was measured by the method of nitrogen or grown up in larger tissue culture dishes for large scale Bitter and Muir (34) with glucuronolactone as astandard. The preparations of monoclonal antibodies. In some experiments, 1X 10s galactosamine content of oligosaccharides was determined, after hycloned hybridoma cells from confluent culture dishes were injected drolysis with 6 M HCI at 100 'C for 8 h, by the Elson-Morgan method intraperitoneally into male BALB/c mouse injected with 0.5 ml of as modified byStrominger et al. (35). The sulfate contentof oligosacpristane and theascites fluid was recovered 2 weeks later as a sample charides was determined, after hydrolysis as above, by the method of of monoclonal antibody. Isotype of monoclonal antibody was deter- Dodgson as modified by Picard et al. (36). Disaccharide products mined with the mouse monoclonal isotyping kit (Zymed Laborato- formed by cleavage of chondroitin sulfates with chondroitinases were ries). identified by chromatographic comparisons with previously characELISA for Characterization of the Monoclonal Antibodies-ELISA terized standards (9). Their amounts were measured spectrophotowas done as described by Rennard et al. (33) with a slight modifica- metrically a t 232 nm (9). Chemical desulfation of whale cartikage tion. Briefly, antigens (5 pg of protein/ml in carbonate buffer/azide, chondroitin sulfate was carried out by the method of Kantor and 200 pllwell) were coated on the plastic surface of the microtiter well Schubert (37). by passive adsorption overnight at 4 "C. The plates were then rinsed three times with phosphate-buffered saline, pH 7.4, containing 0.05% RESULTS (w/v) Tween 20 and appropriate culture supernatant (50 pl diluted with 100 p1 of phosphate-buffered saline, pH 7.4) or ascites fluids (0.1 Characteristics of Monoclonal Antibodies Produced against pl diluted with 100 pl of phosphate-buffered saline, pH 7.4)were PG-M andPG-H-The monoclonal antibody (MO-225) charadded and incubated for 2 h a t room temperature. The plates were then rinsed as above and horseradish peroxidase-conjugated goat acterized in this paper originates from a group of hybridoma anti-mouse IgM (or IgG + IgM) diluted 1:500 into phosphate-buffered clones prepared against PG-M, the chondroitin sulfate prosaline/Tween 20 was added as the second antibody. After 1 h of teoglycan isolated from stage 22-23 chick embryo limb buds incubation, o-Phenylenediamine (0.1 mg/ml in methanol diluted (17). Of a number of hybridoma culture supernatants with 1:100 into 0.03% (v/v) H,Oz) wasadded and color allowed to generate significant reactivity to the immunogen in ELISA, at least for 15 min, after which the reaction was stopped by adding 50 p1 of 8 two did not cross-react with chondroitinase ABC-digested M H,SO,. The brown color produced was measured spectrophotoPG-M, suggesting that thechondroitin sulfate chainsof PGmetrically at 480 nm. M are antigenic. From these, we cloned a hybridoma producFor ELISA to test the effects of chondroitinase or keratanase treatment on proteoglycan antigens, appropriate enzyme (5 units of ing an IgM ( p , K ) and injected the clone into mice. The chondroitinase ABC, 0.5 unit of chondroitinase AC I1 A. aurescens, resulting ascites fluid contained the same monoclonal antior 1 unit of keratanase/ml in enriched buffer containing protease body as that found in the culture supernatant, as judged by inhibitors (16);100 pllwell) was added to the antigen-coated plates specificity toward various proteoglycans and glycosaminoglyand incubated for 1 h at 37 "C. The plates were then rinsed and cans (see below). In the present study the antibody in culture subjected to ELISA as above. For competitive inhibition tests, testsamples were serially diluted supernatant was used unless otherwise indicated. When a purified PG-H preparation from 12-day-old chick into phosphate-buffered saline, pH 7.4, and 100-pl aliquots were added to the antigen-coated plates. Culture supernatants containing embryo epiphysial cartilage (16) was used as an immunogen,

+

Monoclonal Antibody to Intact Chondroitin Sulfate a number of positive hybridomas producing antibodies to the immunogen were obtained butall the hybridoma culture supernatants so far tested recognized either the core protein ora keratanase-sensitive part of PG-H(note that PG-H differs from PG-M in possessing keratan sulfate side chains (16, 17)). It appears that the keratansulfate moiety of PG-H is more immunogenic than thechondroitin sulfate moiety. In this study a monoclonal IgGl (yl, K ) (designated HM-110) prepared from one of these keratan sulfate-positive clones was used as a control. Reactivity of MO-225 to Native, Chondroitinase-treated, and Keratanase-treated Proteoglycam-Using ELISA, MO-225 was shown to react with the following native proteoglycans from different sources: chick embryo limb bud PG-M (17), chick embryo cartilage PG-H (16), chick embryo cartilage PG-Lb (18), chick embryo cartilage PG-Lt (19), and bovine nasal cartilage proteoglycan (AlD1 fraction, see Ref. 20) (Fig. 1).In contrast,Swarm rat chondrosarcoma proteoglycan monomer (Agg-Dl fraction, seeRef. 21) showed no significant reaction with this antibody. In control experiments, HM-110 was examined for its reactivity to the above six proteoglycans. The keratan sulfatecontaining proteoglycans, PG-H and AlD1, showed strong reaction but the other four (lacking keratan sulfate) showed no significant reaction, consistent with the notion that the antigenic determinant for HM-110 resides in the keratan sulfate chains. The difference between MO-225 and HM-110 in specificity was further illustrated by radioimmunoprecipitation analyses. Thus, when 5000 cpm each of [35S]sulfatelabeled PG-M and PG-H were subjected to immunoprecipitation with either MO-225 or HM-110, about 25% of the added PG-H was recovered in both of the MO-225 and HM110 immunoprecipitate, whereas the labeled PG-M was found only in the MO-225 immunoprecipitate. As Fig. 2 shows, enzymatic removal of the keratan sulfate side chains from PG-H did not alter its immunoreactivity to MO-225. However,the removal of its chondroitin sulfate side chains with chondroitinase ABC (or AC I1 A. aurescem) completely abolished the binding activity of the antigen to MO-225. In control experiments with HM-110 the reverse was true; i.e. keratanase could abolish the binding activity of PG-H to HM-110, whereas chondroitinase ABC (or AC I1 A. aurescens) could not. Together these results indicate that the determinant recognized by MO-225 resides in the chondroitin sulfate side chains. The failure of MO-225 to react with Agg-Dl could Reactivity to MO - 225

Antigen

I

HM-110

I

PG -M PG-H

I -

1

I

1

I

PG-Lb

-

1

PG Lt A1 D l

Agg- D 1 I

0

0.5 0 Absorbance at 480nm

0.5

FIG. 1. ELISA of antibodies MO-226 and HM-110 against proteoglycan preparations from different sources. The sources of proteoglycans are: PG-M, stage 22-23 chick embryo limb buds, PG-H, PG-Lb, and PG-Lt, 12-day-old chick embryo epiphysial cartilage; AlD1, bovine nasal cartilage; and Agg-Dl, Swarm rat chondrosarcoma. Antigens were passively absorbed to plastic microtiter plates at a Concentration of 5 pg/ml and reacted with antibody.

4149

FAC E,'F Reactivity to MO 2 2 5

1

Treatment

HM-110

ChOnd?&d=

Chond i i n a x

L'

Kcratanare

0

0.6 0 Absorbance at 480nm

0.;

FIG. 2. Changes in the immunoreactivity of PG-H to antibodies after treatment with chondroitinases and keratanase. PG-H (antigen) was passively absorbed to microtiter plates as described in the legend to Fig. l and then treated with 0.05 unit of chondroitinase AC I1 A. aurescens, 0.5 unit of chondroitinase ABC, or 0.1 unit of keratanase. Controls were treated with heat-denatured enzymes. The plates were rinsed and subjected to ELISA.

have arisen from an absence in the chondroitin sulfate of this particular proteoglycan of the determinant to be recognized by the antibody (see below for further evidence). Competitive Binding Analysiswith GlycosaminoglycansVarious glycosaminoglycans and related anionic polymers from different sources were tested as competitive inhibitors of the binding of MO-225 to PG-H. Of these, the following compounds were inactive at thehighest concentration tested, 3.3 mg/ml; heparin (porcine intestinal mucosa), heparan sulfate (porcine lung), dermatan sulfate (pig skin), keratan sulfate (shark cartilage), keratan sulfate (bovine cornea), hyaluronic acid (cockscomb), dextran sulfate, and DNA. In contrast, many if not most chondroitin sulfate preparations gave significant inhibition (Table I).It appears that theinhibitory activity of chondroitin sulfate, expressed as a concentration (microgram/ml) giving 50% inhibition (ICbo),reflected the content of GlcA2-S04+GalNAc 6-SO4disaccharide unit (designated D-unit), i.e. the higher the D-unitcontent the higher the inhibitory activity. Thus, the activity of shark fin chondroitin sulfate with the highest D-unit content (21.4% of the total hexuronate) was about 42,500-fold higher than that of chick embryo cartilage chondroitin sulfate (the side chains of PG-H) with a low D-unit content (2% of the total hexuronate). None of the D-unit-lacking polysaccharides, i.e. sturgeon notochord chondroitin sulfate, Swarm rat chondrosarcoma chondroitin sulfate, and squid skin chondroitin, gave any significant inhibition at thehighest concentration tested, 3.3 mg/ml. Since the chondroitin sulfate from Swarm rat chondrosarcoma represents the chondroitin sulfate moiety of AggDl, its failure to inhibit the binding of MO-225 may explain why the antibody did not bind to Agg-Dl (Fig. 1). It is also noteworthy that the inhibitory activity of whale cartilage chondroitin sulfate, which contains 3% D-units, was completely abolished after chemical desulfation (data notshown). In contrast, squid cartilage chondroitin sulfate, the preparation that did not release ADi-diSD after chondroitinase ABC (or AC I F. heparinum) digestion, gave significant inhibition (ICbo = 30 wg/ml). The structural basis for this apparently divergent finding is not known. Since, however, this chondroitin sulfate has been shown to differ from ordinary chondroitin sulfates in containing disaccharide repeat units with glucose branches linked by fi-~-(l-S) to the hexosamine residues (30), the observed inhibition activity may reflect the existence ofGlcA 2-so4 residues in some of those glucose carrying repeat units. The available data (30) suggest that many glucose carrying repeat units areadjacent to each other to form glucose branch-rich domains which are released after chondroitinase digestion as sulfated oligosaccharides larger than pentasaccharides.

4150

Monoclonal Antibody

to Intact Chondroitin Sulfate

Inhibition of Binding of MO-225 to PG-H with Oligosaccharides Derived from Chondroitin Sulfate-As Table 11 shows, the tetrasaccharide trisulfate and tetrasaccharide tetrasulfate obtained by enzymatic digestion of sharkfin chondroitin sulfate Fraction I11 gave significant inhibition with MO-225, whereas a tetrasaccharide fraction obtained by digestion of whale cartilage chondroitin sulfatewas inactive at thehighest concentration tested,10,000 pg/ml. Also inactive were a series of unsaturated disaccharides (ADi-OS, ADi-4S, ADi-6S, ADidiSE, and ADi-diSD)obtained by chondroitinase ABC digestion of shark cartilage chondroitin sulfate. In control experiments, neither of the shark fin tetrasaccharides inhibited the 6 12 18 24 I binding of HM-110 to PG-H at 10,000 pg/ml. Between the T i m e of digestion( h ) shark fin tetrasaccharides, the disaccharide sequence, GlcA2FIG. 3. Effect of glycolytic cleavage on ability of shark S04-GalNAc 6-S04,is common. Thus, the results are compatible with the notion that MO-225 specifically recognizes cartilage chondroitin sulfate to inhibit binding of MO-225to D-unit. The inability of ADi-diSDto inhibit the binding may PG-H. Shark scapular cartilage chondroitin sulfate (IC50= 0.43 pg/ ml, see Table I) was incubated with testicular hyaluronidase and at be due to the structural difference in hexuronosyl residue. the indicated times aliquots of the digests were tested for their ability Since glycosides or disaccharides consisting of D-glucuronic to inhibit the binding of MO-225 to PG-H in ELISA. acid 2-sulfate and an aglycone other than N-acetylgalactosamine 6-sulfate were not available for testing, it is not yet known whether the N-acetylgalactosamine 6-sulfate residue in D-unit is an obligate requirement for recognition by the antibody. Factors Affecting the Binding of PG-H to MO-225"Physicochemical studies (38, 39) have shown that divalent cations have significant effects on the conformation of sulfated glycosaminoglycans and proteoglycans. Using the ascites fluid containing MO-225, effects of 5 mM CaC12, 5 mM MgC12, 5 mM KCl, 1 mM EDTA, and 1 mM EGTA on the binding of the antibody to PG-H were tested. In no case, however, was the binding activity significantly influenced by these added reagents. When shark cartilage chondroitin sulfate (D-unit content = 8.4%)was digested with testicular hyaluronidase, its ability 0 0.5 5 CO 500 5ooo 5ooO WDX to inhibit the binding of MO-225 to PG-H was progressively Proteoglycan (ng/ m l ) decreased with digestion time (Fig. 3). The results indicate FIG.4. Inhibition of binding of antibody MO-225to a solid that not only the D-unit content but also the chain length is important in determining the binding activity of antigenic phase PG-H (in ELISA) by PG-H addedto the reaction medium before (0)or after digestion (10 h, 37 "C) with either TPCK chondroitin sulfate. trypsin (0)or Pronase (X). The binding of MO-225 to the PG-H substrate in ELISA was competitively inhibited by PG-H itself added to the medium (Fig.4). The inhibitory activity of PG-H was consid- bearing A*-glucuronicacid, were also reported in preimmune erably reduced after treatmentof the proteoglycan with either rabbits by Poole et al. (41). Some of these sera reacted with TPCK trypsin or Pronase, indicating that the epitope has a intact hyaluronic acid and chondroitin but never with intact requirement for protein core for a better fit of chondroitin chondroitin sulfate. It could be speculated then that chondroitin sulfate side chains of native proteoglycans are nonsulfate to theparatope of the antibody. antigenic. In two other studies, however, monoclonal antibodies directed to either intact chondroitin 4/6-sulfates (15) or DISCUSSION saturated chondroitin 6-sulfate oligosaccharides (42) have Christner et al. (40) reported that degradation of bovine been prepared by immunization with ventral membranes of nasal cartilage proteoglycan with chondroitinase ABC pro- cultured chick fibroblasts or testicular hyaluronidase-treated duced new antigenic determinants bearing nonreducing A4- chick embryo cartilage proteoglycans, respectively. This glucuronic acid end groups. Autoantibodies to degraded gly- would indicate that GlcA+GalNAc 4 - s o 4 and/or GlcAcosaminoglycans, including degraded chondroitin sulfate GalNAc 6-SO4,the common repeat units of many avian and mammalian chondroitin sulfates, can be the focus of an imTABLE 11 mune response in mice. Alternatively, the apparent antigenInhibition of binding of M0-225to PG-H by tetrasaccharides icity of the avian chondroitin sulfate preparations could be due to the fact that the preparations contained a small proTetrasaccharide ICE4 portion of some other structural units(such as D-unit)which are foreign to mice.3 In view of the results of the experiments AGlcA-*GalNac+GlcA-*GalNAc

'I

&

I

I

I

2-so4 6-SO4 AGlcA-&alNAc-GlcA+GalNAc 4-SO4

I

I

I

4,6-bis-S04 2 4 0 4 6304 Tetrasaccharide mixture from whale cartilage chondroitin sulfate

1,400 f 200

1,200 1 200 >10,000

~

Our analysis of a chondroitin [35S]sulfateprepared fromthe whole body of a newborn BALB/c mouse (which had been metabolically labeled with inorganic [35S]sulfate)indicated the composition (percent of the total 35S)of labeled disaccharide units to be 88.5 (A-unit), 7 (C-unit), and 3 (E-unit). N o trace of D-unit was detected by the analysis with chondroitinases.

Monoclonal Antibody to Intact Chondroitin Sulfate reported here, we feel a need of reliable information on the fine structure of the chondroitin sulfates used for specificity studies. The occurrence of sulfated glucuronosyl residues in shark chondroitin sulfate was demonstrated many years ago (9, 10) and the position of sulfate determined to be C-2 (11).Since, however, this residue is present at low levels in most of the chondroitin sulfate preparations so far isolated from avian and mammalian sources, the significance of such a residue in the structure andfunction of proteoglycan glycosaminoglycan chains has been discounted. Recently an unusual species of heparan sulfate has been found in the nuclei of rat hepatocytes which contains ahigh proportion of GlcA % s o 4residues (43). Furthermore,a sulfatase which specificallyhydrolyzes the sulfate group from GlcA % s o 4 but not from L-iduronic acid %sulfate has been observed in human skin fibroblasts and chick embryo chondrocytes (44), suggesting that GlcA % s o 4 residues in glycosaminoglycans might beof greater significance than previously b e l i e ~ e d . ~ We previously showedthat a dermatan sulfate preparation from pig skin contains IdoA2-S04+GalNAc 4-so4 units (56% of the total disaccharide units) (9). In the present study, this dermatan sulfate preparation was shown to give no inhibition to the binding of MO-225 to PG-H. Also shown to be inactive were porcine intestinal mucosa heparin and porcine lung heparan sulfate which contain IdoA % s o 4 residues linked to N - and/or 0-sulfated glucosamine units at considerably high levels. The results indicate that the IdoA2-S04containing disaccharide sequences involved in these glycosaminoglycans cannot be recognized by MO-225. The occurrence of GlcA3-S04+GalNAc 4-so4 disaccharide units has been suggested in king crab chondroitin sulfate (11). In view of its high sulfate to glucuronic acid ratio (-1.64), this polysaccharide must contain a high proportion of GlcA % s o 4 residues. It was of interest therefore to test the antigenicity of this polysaccharide to MO-225. Our competitive inhibition tests have shown that the inhibitory activity of a king crab chondroitin sulfate preparation5is far lower (IC5o= 5.0 pg/ml) than the activities of shark cartilage chondroitin sulfate preparations. Although the precise structure of king crab chondroitin sulfate is unknown, the results suggest that MO-225 maydistinguish between GlcA2-S04+GalNAc 6430, and GlcA3-S04+GalNAc 4-so4 residues. G l c A 4 a l N A c 4,6-bis-S04, the disulfated disaccharide unit frequently found in chondroitin sulfate chains of avian and mammalian proteoglycans (8) and designated “E-unit,” is unlikely to be recognized by MO-225, since squid cartilage chondroitin sulfate with a high proportion of E-units (61%of the total hexuronate) was far less inhibitory than shark fin chondroitin sulfates with only 0.8-6.7% E-units (Table I). There may be differences in average chain length among the chondroitin sulfate samples used for competitive inhibition tests, but this could not explain the observed difference in their inhibitory activity. Thus, a M , = 10,000 fragment prepared from shark scapular cartilage chondroitin sulfate was much higher (IC50 = 170 p(g/ml)‘j in inhibitory activity than sturgeon notochord chondroitin sulfate (IC5o= >3,300 pg/ml, see Table I) which has a comparable or slightly longer chain length (average M , = 12,000). Well characterized antibodies are now available which rec-

‘A small amount of the nuclear heparan sulfate was kindly given

by Dr. H. E. Conrad (University of Illinois) and was tested for its effect on the binding of MO-225 to PG-H. At a concentration of 10 pg/ml, the sample had no effect on the binding. Kindly donated by Dr. N. Seno (Ochanomizu University, Tokyo). M. Yamagata, unpublished observation.

4151

ognize epitopes present inchondroitin sulfate, keratan sulfate, and core protein structurescharacteristic of different subtypes of chondroitin sulfate proteoglycans. The use of monoclonal antibodies directed to the different parts of proteoglycans may offer great potential in the immunohistochemical characterization of the proteoglycan subtypes distributed in various tissues. Furthermore, it can be surmised that glycosaminoglycan chains of native proteoglycans from various sources may contain different antigenic structures to which specific monoclonal antibodies can be prepared. It is certainly desirable to have monoclonal antibodies which can detect differences in fine structures among glycosaminoglycan chains. Observations on the changes in their fine structure during biological changes in health and disease may help to understand thebiological roles of glycosaminoglycan chains. Acknowledgments-We wish to thank Dr. N. Seno (Ochanomizu University), Dr. I. Yamashina (Kyoto University), and Dr. T. Harada (Seikagaku Kogyo Co. Ltd.) for many helpful discussions and suggestions.

REFERENCES 1. Hascall, V. C., and Hascall, G.K. (1981) in Cell Biology of Extracellular Matrix (Hay, E. D., ed) pp. 39-63, Plenum Press, New York 2. Hassell, J. R., Kimura, J. H., and Hascall, V. C. (1986) Annu. Rev. Bwchem. 5 5 , 539-567 3. Poole, A. R. (1986) Biochem. J. 2 3 6 , l - 1 4 4. Bumol, T. F., Walker, L. E., and Reisfeld, R. A. (1984) J. Biol. Chem. 259,12733-12741 5. Razin, E., Stevens, R. L., Akiyama, F., Schmid, K., and Austen, K. F. (1982) J. Bwl. Chern. 2 5 7 , 7229-7236 6. MacDedmott, R. P., Schmidt, R. E., Caulfield, J. P., Hein, A., Bartley, G . T., Ritz, J., Schlossman, S. F., Austen, K. F., and Stevens, R. L. (1985) J. Exp. Med. 162, 1771-1787 7. Hoffman, P., Linker, A., and Meyer,K. (1958) Fed. Proc. 1 7 , 1078-1082 8. Otsu, K., Inoue, H., Tsuzuki, Y., Yonekura, H., Nakanishi, Y., and Suzuki, S. (1985) Biochem. J. 2 2 7 , 37-48 9. Suzuki, S., Saito, H., Yamagata, T., Anno, K., Seno, N., Kawai, Y., and Furuhashi, T. (1968) J. Biol. Chem. 243,1543-1550 10. Suzuki, S. (1960) J. Biol. Chem. 235, 3580-3588 11. Seno, N., and Murakami, K. (1982) Carbohydr. Res. 1 0 3 , 190194 12. Kimata, K., Okayama, M., Oohira, A., and Suzuki, S. (1973) Mol. Cell. Bwchem. 1 , 211-228 13. Caterson, B., Christner, J. E., and Baker, J. R. (1983) J. Bwl. Chem. 258,8848-8854 14. Mehmet, H., Scudder, P., Tang, P. W . , Fwnsell, E. F., Caterson, B., and Feizi, T. (1986) Eur. J. Biochern. 157, 385-391 15. Avnur, Z., and Geiger, B. (1984) Cell, 38,811-822 16. Oike, Y., Kimata, K., Shinomura, T., Nakazawa, K., and Suzuki, S. (1980) Biochem. J. 1 9 1 , 193-207 17. Kimata, K.,Oike, Y., Tani, K., Shinomura, T., Yamagata, M., Uritani, M., and Suzuki, S. (1986) J. Bwl. Chern. 2 6 1 , 1351713525 18. Shinomura, T., Kimata, K., Oike, Y., Noro, A., Hirose, N., Tanabe, K., and Suzuki, S. (1983) J.Bwl. Chem. 258,9314-9322 19. Noro, A., Kimata, K., Oike, Y., Shinomura, T., Maeda, N., Yano, S., Takahashi, N., and Suzuki, S. (1983) J. Bwl. Chem. 2 6 8 , 9323-9331 20. Heinegird, D. (1972) Biochim. Biophys. Acta 2 8 5 , 181-192 21. Kimata, K., Kimura, J. H., Thonar, E. J.-M. A., Barrach, H.-J., Rennard, S. I., and Hascall, V. C. (1982) J. Biol. Chem. 257, 3819-3826 22. Kimata, K., Okayama, M., Oohira, A., and Suzuki, S. (1974) J. Biol. Chem. 2 4 9 , 1646-1653 23. Kawai, Y., Seno, N., and Anno, K. (1966) J. Biochem. (Tokyo) 6 0 . 317-321 24. Anno, K., Kawai, Y., and Seno, N. (1964) Biochim. Biophys. Acta 83,348-349 25. Nakazawa, K., Suzuki, N., and Suzuki, S. (1975) J. Biol. Chem. 250,905-911 26. Flodin, P., Gregory, J. D., and RodBn, L. (1964) Anal. Biochem. 8,424-433 I

~~

~~~

4152

Monoclonal Antibody to Intact Chondroitin Sulfate

27. Yamagata, T., Saito, H., Habuchi, O., and Suzuki, S. (1968)J. Bwl. Chem. 243,1523-1535 28. Hiyama, K., and Okada, S. (1975)J.Biol. Chem. 250,1824-1828 29. Yoshida, K., Maeyama, K., Kikuchi, H., Watanabe, K., Sakurai, K., and Nagasawa, K. (1985)Seihguku 57,1189 30. Habuchi, O., Sugiura, K., Kawai, N., and Suzuki, S. (1977)J. Biol. Chem. 252,4570-4576 31. Nakazawa, K., and Suzuki, S. (1976)J. Biol. Chem. 250, 912917 32. Eshhar, Z., Blatt, C., Bergman, Y., and Haimovich, J. (1979)J. Zmmunol. 122,2430-2433 33. Rennard, S. I., Kimata, K., Dusemund, B., Barrach, H. J., Wilczek, J., Kimura, J. H., and Hascall, V.C. (1981)Arch. Bwchem. Biophys. 207,399-406 34. Bitter, T., and Muir, H. M.(1962)Anal. Biochem. 4,330-334 35. Strominger, J. L.,Park, J. T., and Thompson, R. E. (1959)J. Bwl. Chem. 234,3263-3268

36. Picard, J., Paul-Gardais, A., and Vedel, M. (1973)Biochim. Biophys. Acta 320,427-441 37. Kantor, T. G., and Schubert, M. (1957)J. Am. Chem. SOC.79, 152-153 38. Tanaka, K. (1978)J. Biochem. (Tokyo) 83,647-653 39. Boyd, J., and Williamson, F. B. (1980)J. Mol. Biol. 137, 175190 40. Christner, J. E., Caterson, B., and Baker, J. R. (1980)J. Biol. Chem. 255, 7102-7105 41. Poole, A. R., Reiner, A., Roughley, P. J., and Champion, B. (1985) J. Biol. Chem. 260,6020-6025 42. Jenkins, R. B., Hall, T., and Dorfman, A. (1981)J. Bwl. Chem. 256,8279-8282 43. Fedarko, N. S., and Conrad, H.E. (1986)J. Cell Bwl. 102,587599 44. Shaklee, P. N., Glaser, J. H., and Conrad, H. E. (1985)J. Biol. Chem. 260,9146-9149