ANALYTICAL

BIOCHEMISTRY

20 1, l-8

(1992)

A Monoclonal Antibody (ST-l) Directed to the Native Heparin Chain’ Anita

H. Straus,*~2

Luiz R. Travassos,t

and Helio

K. Takahashi*

*Department of Biochemistry and TCell Biology Discipline, Escoh Paulista de Medicina, Caixa Postal 20372, Scio Pa&o, S6o Paul0 04023, Brazil Received

February

27,199l

A mouse monoclonal antibody, ST-l, was raised against heparin complexed to Salmonella minnesota. Characterization of this antibody showed that it recognizes an epitope in the intact molecule of heparin that is present regardless of its source or anticoagulant activity. ST-1 is the first monoclonal antibody specific for the intact unmodified molecule of heparin to be described. SH-labeled heparin in solution was immunoprecipitated by ST-l, and the formation of the ‘H-labeled immunocomplex was selectively inhibited by unlabeled heparin. No cross-reactivity of ST-l was observed with other glycosaminoglycans such as heparan sulfate, chondroitin sulfate, hyaluronic acid, dermatan sulfate, and keratan sulfate, or with polyanionic polymers such as dextran sulfate. Selective removal of the N-sulfate groups or N,O-desulfation of heparin strongly reduced the binding of ST-l. Inhibition of binding was also observed after carbodiimide reduction of the carboxyl groups of the uranic acid units of heparin. Competitive assays of ST- 1 binding to heparin immobilized on polyL-lysine-coated plates using oligosaccharides of different sizes that arose from HNO, cleavage of heparin showed that the minimum fragment required for reactivity of ST-l is a decasaccharide. Q 1992 Academic press, IIIC.

Monoclonal antibodies (MoAbs)3 have been successfully used in the last few years as powerful reagents for ’ This study was supported by FINEP (Financiadora de Estudos e Projetos) and CNPq (Conselho National de Desenvolvimento Cientifico e Tecnolbgico). * To whom correspondence should be addressed at Department of Biochemistry, Escola Paulista de Medicina, Rua Botucatu 862, C.P. 20.372, SBo Paulo, SP 04023, Brazil. 3 Abbreviations used: BSA, bovine serum albumin; PBS; phosphate-buffered saline (137 mM NaCl, 10 mM N%HPO,, pH 7.4); MoAb, monoclonal antibody; GAG, glycosaminoglycan; RIA, radioimmunoassay; dp, degree of polymerization; AT-III, antithrombin-III; If&,, inhibitor concentration that reduces by 50% the ST-l binding to heparin; PHI, methylated polymer of 1-ethyl-3-(3dimethylaminopropyl)carbodiimide; DEAE, diethylaminoethyl. ooo3-2697/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form

biochemical and biological investigations on glycoconjugates (l-4). Several hybridomas that secrete antibodies against carbohydrate epitopes of glycolipids or glycoproteins have been isolated (4-8). In contrast, only a few MoAbs that react with glycosaminoglycans (GAGS) have been described. Some of these MoAbs were produced against partially modified structures obtained by chemical or enzymatic cleavage of the carbohydrate chains (9,lO). MoAbs against intact chains of heparan sulfate (ll), keratan sulfate (12,13), and chondroitin 6sulfate (14) but not against heparin have been described. The apparent lack of immunogenicity of heparin, i.e., the virtual absence of anti-heparin antibodies, in heparinized patients, has been of great value for heparin use in medicine as an anticoagulant agent. However, a few reports have described the occurrence of low levels (~0.5 pg/ml) of anti-heparin antibodies in patients with Waldenstrom macroglobulinemia or other gammopathys (1516). These antibodies from human serum cross-reacted with other antigens and have not been sufficiently well characterized for an antigenic determinant in the heparin molecule to be defined. Recently, Pejler et al. (9) described a monoclonal antibody specific for chemically modified heparin that, however, did not react with the native polymer. The epitope recognized by this monoclonal antibody was a pentasulfated tetrasaccharide derived from nitrous acid degradation of heparin, and the resulting 2,5-anhydromannose-reducing end unit was shown to be essential for interaction with the antibody. Production of rabbit polyclonal antibodies against heparin has also been reported by Gitel et al. (17). These authors utilized heparin covalently linked to methyl-bovine serum albumin (BSA) and were able to produce polyclonal antibodies that recognized intact heparin adsorbed to a methylated polymer of 1-ethyl-3(3-dimethylaminopropyl)carbodiimide (PHI) but did not react with free heparin in solution. In the present report, we describe a monoclonal antibody, ST-1 (IgM), specific for intact heparin and sug1

Inc. reserved.

2

STRAUS,

TRAVASSOS,

gest the possible size and structure of the antigenic determinant recognized by this MoAb. The specificity of ST-1 was determined by: (a) reactivity with various standard glycosaminoglycans and other polysaccharides, (b) immunoprecipitation of tritium-labeled heparin by MoAb ST-l, (c) inhibition of antibody binding to heparin by different glycosaminoglycans, (d) inhibition of antibody binding by products of nitrous acid degradation of heparin, i.e., oligosaccharides ranging from dp 2 to dp 14, and (e) reactivity of modified heparins with the ST-l MoAb.

MATERIALS

AND

METHODS

GAGS, other polymers, and immunoglobulins. Hyaluranic acid from human umbilical cord; chondroitin 4-

sulfate from whale cartilage; chondroitin 6-sulfate from shark cartilage; heparan sulfate from bovine kidney; dermatan sulfate from hog skin; keratan sulfate from bovine cornea; heparin from hog intestine; dextran, M, 500,000; dextran sulfate, M, 500,000; dextran sulfate, M, 8000; and poly-L-lysine, M, 150,000-300,000, were purchased from Sigma Chemical Co. (St. Louis, MO). Rabbit anti-mouse IgM, IgG, IgGl, IgGBa, IgGBb, and IgG3 antibodies were purchased from Miles Scientific (Elkhart, IN). Purified, protein-free bovine lung heparin from Upjohn Co. was a kind gift of Dr. U. Lindahl (Department of Veterinary Medical Chemistry, Swedish University of Agricultural Sciences, Uppsala, Sweden). Heparins of high and low anticoagulant activity were obtained by affinity chromatography on a column of Sepharose-antithrombin-III as described (18). Heparins of high and low molecular weight and high and low degree of sulfation were obtained by ion-exchange in DEAE columns (19). Immunization procedure. About 4 mg of heparin was dissolved in 4 ml of distilled water and mixed with 16 mg of acid-treated heat-inactivated Salmonella minnesota. The mixture was stirred for 1 h at 57”C, lyophilized, and resuspended in 2 ml of 0.01 M phosphate-buffered saline (PBS), pH 7.2. Aliquots (100 ~1) of this suspension were used to immunize male BALB/c mice iv through the caudal vein once a week, for 5 weeks. After a rest period of 2 weeks, the animals’ immune response was boosted with 200 ~1 of the immunogenic complex. Three days later, the mice were sacrificed and their spleens removed. Derived B lymphocytes were then fused with a SP-2 myeloma cell line and the resulting hybrids grown in RPM1 1640 supplemented with 2 mM glutamine, 1 mM sodium pyruvate, and 15% fetal calf serum. Hybrids secreting immunoglobulins that reacted with heparin were detected by solid-phase radioimmunoassay (RIA). Only the clones showing positive reactivity with heparin and negative reactivity to other GAGS were recloned by limited dilution as described by Takahashi et al. (6).

AND TAKAHASHI

Binding assay. For solid-phase RIA different GAGS were adsorbed on 96-well plates coated with poly-L-lysine as described by Mehmet et al. (13), with slight modifications as follows: 96-well plates (Falcon Microtest III flexible assay plates) were precoated with 50 ~1 of aqueous 0.5% poly-L-lysine for 1 h at room temperature and then washed with PBS. Solutions (50 @well) of dextran (M, 500,000), dextran sulfate (M, 500,000 and S,OOO), or each of the GAGS at 40 pg/ml were added and serially diluted down to 1 rig/well. After 2 h, the wells were washed three times with PBS and blocked with 1% BSA in PBS (200 ~1) for 2 h. The plates were then incubated overnight with ST-1 (100 ~1) at 4°C. The amount of antibody bound to the GAGS was determined by reaction with anti-mouse IgM (50 ~1). After the plates were washed three times with PBS, approximately lo5 cpm in 50 11 of ‘251-labeled protein A in 1% BSA was added per well. After incubation for 1.5 h, the plates were washed five times with PBS and the radioactivity in each well was measured in a gamma counter. Irrelevant monoclonal antibodies (IgM and IgG3) were used in the solidphase RIA as negative controls in order to exclude the possibility of nonspecific interaction between heparin and immunoglobulins. Immunoprecipitation of 3H-labeledheparin. The MoAb ST-1/[3H]heparin complex was immunoprecipitated with affinity-purified goat anti-mouse IgM. About 1 mg of ST-l was incubated overnight at 4°C with 100 ~1 of [3H]heparin (1.000 cpm/pl/pg). Anti-mouse IgM (500 gg) was added; the resulting immunoprecipitate was separated by centrifugation, washed twice with cold PBS, and solubilized with Triton X-100; the radioactivity was measured in a scintillation counter. Controls were made using [3H]chondroitin 6-sulfate (860 cpm/pl/pg) or irrelevant mouse MoAbs: ST-2 (IgM) and CU(IgG3). Inhibition assays were carried out by incubating MoAb ST-l and [3H]heparin with 200 and 1,000 pg of unlabeled heparin or unlabeled chondroitin g-sulfate in PBS. The corresponding immunoprecitates obtained after addition of goat anti-mouse IgM were analyzed as above. Inhibition of antibody binding by different GAGS. Initially, 75 ~1 of a 100 pg/ml solution of heparin was serially diluted with PBS (2-l to 2054-l). All other GAGS, heparan sulfate, chondroitin 4-sulfate, chondroitin 6sulfate, dermatan sulfate, keratan sulfate, and hyaluranic acid, were utilized from stock solutions at 200 pglml. Each GAG solution was incubated with 75 ~1 of ST-l at room temperature. After 2 h, aliquots of 100 ~1 were taken and incubated overnight at 4°C in 96-well plates precoated with heparin (0.2 pg/well) as described under Binding assay. The inhibitory effect of the different GAGS on ST-1 binding to heparin was determined by solid-phase RIA as described above. Isolation of HNO, degradation products fdp Z-dp 14). Degradation of lung heparin was performed in pH

1.5 as described by Shively and Conrad (20) with slight

ANTI-HEPARIN

MONOCLONAL

ANTIBODY

3

in PBS. Each solution (75 ~1) was added to an equal volume of ST-l (culture supernatant) and incubated at room temperature. After 2 h, 100 ~1 of the monoclonal antibody/oligosaccharide mixture was added to wells

nM

4. 0.8 ; 0.8 u 0.4

2 g 0.2 5 80

120 ELUTION

160 VOLUME

200 (ml)

FIG. 1. Sephadex G-50 elution pattern of samples containing 2-14 monosaccharide units (dp 2-dp 14). Bovine lung heparin (10 mg) was cleaved with HNOz at pH 1.5 at 0°C for 10 min, as described under Materials and Methods. The products obtained were chromatographed in Sephadex G-50, equilibrated, and eluted with 1.0 M NH,HCO,. Fractions (2 ml) containing the different oligosaccharide elution peaks were collected and the pooled fractions for each peak were rechromatographed in the same column. Fractions were then analyzed for uranic acid, pooled as indicated by solid bars, and utilized in all subsequent experiments. 2, disaccharide fraction; 4, tetrasaccharide fraction; 6, hexasaccharide fraction; 8a, octasaccharide a fraction; 8b, octasaccharide b fraction; 10, decasaccharide fraction; 12, dodecasaccharide fraction; 14, tetradecasaccharide fraction.

modifications to obtain increased amounts of high-molecular-weight oligosaccharides. Oligosaccharides were reduced with 3H-labeled (135 mCi/mmol; Amersham Radiochemicals) or unlabeled NaBH, at 50°C for 30 min. The sample was then cooled to room temperature and the excess NaBH, destroyed in a fume hood by acidification to pH 3 with acetic acid. It was then neutralized with NaOH and evaporated to dryness in a stream of N, (21). The deamination products were separated by gel chromatography in a Sephadex G-50 column (1.5 X 110 cm) equilibrated with 1 M ammonium carbonate with a flow of 12 ml/h, previously calibrated with standards from 2 to 12 monosaccharide units (dp 2 to dp 12) kindly supplied by Dr. H. E. Conrad (Glycomed, San Francisco, CA). Fractions (2 ml) corresponding to the elution peaks of oligosaccharides of 2 to 14 monosaccharide units (dp 2 to dp 14) were pooled, lyophilized, and rechromatographed three times in Sephadex G-50 as above in order to obtain better resolution of each oligosaccharide peak (Fig. 1). The molecular weights of the oligosaccharides were estimated on 9% linear polyacrylamide gel electrophoresis with 0.06 M barbital buffer, pH 8.6, at 10 V/cm. The purity of each oligosaccharide fraction was confirmed after the gel was stained with toluidine blue (19). Only a single band was observed for each fraction. Inhibition of antibody binding by HNO, degradation products. Fractions of dp 2 to dp 14 were used to assess the correct size of the epitope recognized by ST-l. The different fractions were serially diluted from 2 mM to 1

precoated with heparin. The inhibitory effect of oligosaccharides of different size was determined by solidphase RIA as described in Binding assay. Chemical modifications of heparin. Selective N-desulfation and total N,O-desulfation of heparin were carried out as described by Kosakai and Yosizawa (22). Briefly, lung heparin was treated with dimethyl sulfoxide containing 5% water at 50°C for 1.5 h to obtain Ndesulfated heparin. N,O-desulfated heparin was prepared by treatment of heparin (pyridinium salt) with dimethyl sulfoxide containing 2% pyridine at 100°C for 9 h. N-acetylation following desulfation was carried out as described by Danishefsky and Steiner (23). The content of uranic acid in various preparations was determined as described by DiFerrante et al. (24). [3H]Heparin was obtained after reduction of 10 mg of heparin with 0.1 ml NaB3H, (135 mCi/mmol; Amersham Radiochemicals) in 0.1 M NaOH at 50°C for 30 min (35). The sample was then cooled to room temperature and the excess NaBH, destroyed in a fume hood by acidification to pH 3 with acetic acid. It was then neutralized with NaOH. 3H-labeled heparin was precipitated with 2 vol of ethanol and dialysed extensively against distilled water.

RESULTS

Specificity of the ST-l antibody. Among a few clones showing reactivity with heparin, a clone secreting an IgM antibody was established after repeated subcloning and termed ST-l. All subsequent studies were carried out using culture supernatants. Chondroitin 4-sulfate, chondroitin 6-sulfate, hyaluronic acid, dermatan sulfate, keratan sulfate, heparan sulfate, bovine lung heparin, pig mucosal heparin, and heparin of high and low anticoagulant activities obtained by affinity chromatography in Sepharose-AT-III were tested to determine their binding capacity to the ST-l antibody. The possibility that other less sulfated GAGS may not bind to poly-L-lysine-coated plates was discarded because another MoAb (ST-2) reacted specifically with chondroitin 6-sulfate under the same conditions. In Fig. 2 we show that only heparin, regardless of its source and anticoagulant activity, is recognized by ST-l. No significant difference in ST-l binding was observed with the different heparins tested. No reactivity was observed in solid-phase RIA when heparin was incubated either with an anti-chondroitin 6-sulfate MoAb or with anti-mouse IgG or anti-mouse IgM (data not shown), discarding the possibility of nonspecific binding of heparin to irrelevant antibodies.

STRAUS,

1

4

RECIPROCAL

16

64

266

OF ANTIGEN

TRAVASSOS,

2046 DILUTION

FIG. 2.

Specificity of binding of the monoclonal antibody ST-1 to beparin. Glycosaminoglycans and other polysaccharides (2 pglwell) were serially double-diluted to 20 rig/ml (1 rig/well). ST-l (100 ~1) was added and the mixture incubated overnight at 4Y!. The amount of antibody bound to wells coated with GAGS was determined by incubation with rabbit anti-mouse IgM (2 h) and ld cpm of ‘%I-labeled protein A in 1% BSA. (0) Bovine lung heparin; (0) hog intestine heparin; (A) high-anticoagulant-activity heparin; (A) low-anticoagulant-activity heparin. (m) Activities of heparan sulfate, chondroitin 4-sulfate, chondroitin 6-sulfate, dermatan sulfate, hyaluronic acid, keratan sulfate, dextran sulfate, M, 500,000, and dextran, M, 500,000.

The MoAb ST-1/[3H]heparin complex was immunoprecipitated with goat anti-mouse IgM. Under these conditions a small amount of labeled heparin (0.06 nmol) was precipitated by 1.3 nmol of ST-l. Addition of unlabeled heparin displaced virtually all labeled heparin from the immunoprecipitate. Conversely, addition of unlabeled chondroitin 6-sulfate did not reduce the amount of tritium-labeled heparin precipitated in the immunocomplex (Table 1). These data clearly confirm the specificity of MoAb ST-l for heparin. However, the MoAb has a low affinity for heparin when in solution. Competitive inhibition of ST-l binding to heparin. To test the inhibition of ST-l binding to heparin, six different GAGS and dextran sulfate, A4, 500,000 and 8000, and dextran, M, 500,000, were used for competitive binding in solid-phase RIA. All polysaccharides were used at an initial concentration of 200 pg/ml except heparin, which was used at 100 pg/ml. Figure 3 shows that only heparin can inhibit binding of ST-l to plates coated with heparin on poly+lysine. Heparin at 500 rig/well inhibits virtually 100% of binding of ST-l to heparin. All other glycosaminoglycans at concentra-

AND

TAKAHASHI

tions up to 10 pg/well were not inhibitory. Only a slight inhibition of ST-l could be observed (~5%) with dextran sulfate at 10 pg/well. Structural requirements for ST-l binding to heparin. In order to analyze the contribution of the carboxyl, O-sulfate, and N-sulfate groups of heparin for interaction with the antibody, different preparations of chemically modified heparins were tested as inhibitors of ST-l binding to unmodified heparin immobilized on poly-L-lysine-coated plates (Table 2). The N-desulfated or N,O-desulfated heparins were poor inhibitors (IC, 3.1 X 10m2 and 6.2 X 10-l mg/well, respectively). N-Acetylation or N- and N,O-desulfated heparins strongly reduced their inhibitory activity (IC, 1 and 10 mg/well, respectively). Surprisingly, an extensive loss of inhibitory activity was observed after reduction of the carboxy1 groups of uranic acid, indicating that carboxyl groups are also important for recognition of heparin by ST-l. Binding experiments using plates coated with Nand N,O-desulfated heparin showed a marked decrease of antigenic activity in comparison with unmodified heparin. Moreover, the ability of ST-l to bind to desulfated N-acetylated heparins was reduced lO,OOO-fold, confirming that the presence of acetyl groups strongly restricts ST-l recognition. Size of the epitope for ST-l reactivity. In order to verify the size of the epitope recognized by ST-l, a competitive binding assay was carried out using nitrous acid degradation products ranging from 2 to 14 monosaccharide units, which were isolated by gel chromatography in Sephadex G-50, as shown in Fig. 1. Each oligosaccharide solution was serially diluted and preincubated with

TABLE

1

Immunoprecipitation of (3H]Heparin with ST-l in Absence or Presence of Unlabeled Polysaccharides” Immunoprecipitation

system

ST-l+ [3H] heparin + goat anti-mouse IgM ST-l+ [sH] heparin + unlabeled heparin (1.0 mg) + goat anti-mouse IgM ST-l+ [‘HI heparin + unlabeled heparin (200 ag) + goat anti-mouse IgM ST-l+ [3H] heparin + unlabeled chondroitin 6-sulfate + goat anti-mouse IgM ST-l+[‘H]chondroitin 6 sulfate + goat anti-mouse IgM ST-2+[‘H]heparin + goat anti-mouse IgM CU-1+[3H]heparin + goat anti-mouse IgM

wm 1343 550 781 1237 503 465 390

a [‘H]Heparin or [3H]chondroitin B-sulfate (10’ cpm/lOO pg) was mixed with 1.0 mg of ST-l and the immunocomplex was precipitated by adding 500 pg of affinity-purified goat anti-mouse IgM. Inhibition of ST-l binding to labeled heparin was carried out by adding 1.0 or 0.2 mg of unlabeled heparin or unlabeled chondroitin 6-sulfate to the ST-l/[*H]heparin mixture.

ANTI-HEPARIN

MONOCLONAL

5

ANTIBODY

rin. The striking contrastbetweenthe inhibitory activities of intact heparin and decasaccharides derivedfrom

INHIBITOR

(Pg/Well)

FIG. 3. Inhibition of antibody binding to heparin by different GAGS and other polysaccharides. Different polysaccharides (200 pgl ml, except heparin, 100 wg/ml) were serially double-diluted with PBS and preincubated with the ST-l. The inhibition assay was carried out in 96-well plates coated with heparin. Aliquots (100 pl) of the mixture polysaccharide/ST-1 were added to the heparin-coated plates at 4°C. The amount of antibody bound to heparin-coated plates was determined by incubation with anti-mouse IgM (2 h) and ‘261-labeled protein A in 1% BSA. The effects of GAGS and other polysaccharides are expressed as percentages of inhibition of ST-1 binding to heparin in relation to ST-1 binding in the absence of a competitive ligand. (0) Bovine lung heparin; (0) hog intestine heparin; (A) dextran sulfate, M, 500,000. (m) Reactions with five other types of GAGS (heparan sulfate, chondroitin 4-sulfate; chondroitin B-sulfate, dermatan sulfate, hyaluronic acid, and keratan sulfate).

ST antibody. The inhibitory activity of each oligosaccharide was analyzed by solid-phase RIA in heparin-coated plates, as described under Materials and Methods. The data shown in Fig. 4 clearly demonstrate that at 100 nmol/well, disaccharides were inactive as inhibitors, and tetrasaccharides, hexasaccharides, octasaccharide A, and octasaccharide B inhibited binding in 10,22,18, and 30%, respectively. The decasaccharide fraction was capable of inhibiting 90% of the ST-l binding to heparin at 60 nmol/well. The same final inhibitory activity was achieved by dodecasaccharides and tetradecasaccharides at 25 nmol/well. The smallest fragment with the highest inhibitory activity is represented by the decasaccharide fraction, although relatively high concentrations (60 nmol/well) of decasaccharides are required for almost complete inhibition of the ST-l binding to heparin. Conversely, with the intact polymer, a similar inhibition was obtained at a concentration as low as 0.01 nmol/well of native hepa-

it suggests that the latter consist of a mixture of molecules with different structures; only a small fraction of these contain the correct sequence recognized as an epitope by ST-l. To test this possibility, an immunoaffinity chromatography was performed as follows. Affinity chromatography. About 5000 cpm of 3Hlabeled heparin was applied to a column of ST-1 covalently linked to Sepharose CL-4B (0.5 X 4.0 cm) equilibrated in 10 InM tris(hydroxymethyl)aminomethane-HCl buffer, pH 7.4. Under these conditions all the polysaccharide applied was quantitatively adsorbed on the column and was eluted with the same buffer plus 0.5 M NaCl, suggesting a weak affinity of the antibody or the fact that the heparin molecule contains only one or very few epitopes per molecule. In the next step 3H-labeled decasaccharides (lo4 cpm) were submitted to affinity gel chromatography under the same conditions as those for heparin. About 5% of the decasaccharide fraction was retained in the column, representing the highaffinity decasaccharides that contain the actual epitope recognized by ST-l. The fraction that did not bind to the ST-1 column continued to do so on rechromatography. It is clear that the decasaccharide fraction consists of a mixture of molecules and that only a small proportion of them have the correct sequence of monosaccharides recognized by ST-l. Under the same conditions, chondroitin 6-sulfate was quantitatively retained in a Sepharose CL-4B column of MoAb ST-2 (specific against chondroitin 6-sulfate), whereas heparin was not, therefore excluding the possibility of nonspecific binding of heparin to a nonspecific immunoglobulin.

TABLE Inhibitory

Activity and

2

of Chemically Modified Dextran Sulfate”

Inhibitor Heparin dp 14 oligosaccharides dp 12 oligosaccharides dp 10 oligosaccharides Carbodiimide-reduced heparin N-desulfated heparin N,O-desulfated heparin N-desulfated, N-acetylated heparin N,O-desulfated, N-acetylated heparin Dextran sulfate

Heparins

IGl bdml) 3.0 6.1 2.6 3.0 6.2 3.1 6.2

x 10-5 x lo-’ X 1O-3 x 10-3 x 10-r x lo+ X 10-l 1.25 10.0 4.0 x 1o-2

0 N- and N,O-desulfated heparin; N-acetylated, N- or N,O-desulfated heparin; carboxyl-reduced heparin; oligosaccharides derived from nitrous acid deamination of heparin (dp 10 to dp 14); and dextran sulfate, M, 500,000, were tested as inhibitors of ST-1 binding to unmodified heparin.

STRAUS,

INHIBITOR

TRAVASSOS,

(nmoles/well)

FIG. 4.

Inhibition of ST-1 antibody binding to heparin by heparinderived oligosaccharides of different sizes. The different oligosaccharides (2 mM) were serially double-diluted with PBS and preincubated for 2 h with the ST-1 antibody. Aliquots (100 cl) were then incubated in heparin-precoated plates at 4°C. The amount of antibody bound to heparin was determined by incubation with anti-mouse IgM (2.0 h) and ‘%I-labeled protein A in 1% BSA. The effects of the oligosaccharides are expressed as percentages of inhibition of ST-1 binding to the heparin-coated plate in relation to ST-1 binding in the absence of a competitive ligand. Oligosaccharides derived from nitrous acid deamination are as follows: (A) tetradecasaccharides, (A) dodecasaccharides, (O), decasaccharides, (V) octasaccharides a, (7) octasaccharides b, (0) hexasaccharides, (0) tetrasaccharides, (m), disaccharides.

Structural analysis of the ST-l-reactive epitope and the high-affinity decasaccharide is currently under study. DISCUSSION

In the present work we describe a monoclonal antibody, ST-l, specific for heparin. This is the first report supporting the assumption that native heparin can also act as a carbohydrate antigen able to elicit the production of monoclonal antibodies. The strict specificity of ST-1 for heparin and the lack of reactivity with other structurally related polymers like heparan sulfate probably depend on the low contents of N-acetylated disaccharides in these molecules and on the relatively high proportions of N-sulfated groups. In fact, we have shown that N- or N,O-desulfation of heparin reduced its affinity for the antibody. Moreover, N-acetylation of the previously desulfated heparin further reduced binding to ST-l. This suggests that acetyl groups may either induce a conformational change in the heparin molecule by increasing its hydro-

AND

TAKAHASHI

phobicity or promote a steric hindrance, inhibiting interaction with the antibody. The high content of N-acetylated disaccharide units in heparan sulfate probably explains its lack of reactivity with ST-l. A high number of sulfate groups is clearly not enough for ST-1 binding to heparin. In fact, heparin fractions of high molecular weight, high degree of sulfation, and high anticoagulant activity (M, 29K, 2.8 mol sulfate/m01 hexosamine, 298 IU/mg) and heparins of low molecular weight, low degree of sulfation, and low anticoagulant activity (M, 8K, 2.3 mol sulfate/m01 hexosamine, 40 IU/mg), both obtained by ion-exchange chromatography in a DEAEcellulose column (19), showed the same reactivity with ST-1 in solid-phase RIA (data not shown). Furthermore, dextran sulfate with 2.3 mol sulfate/mol glucose reacts poorly with this antibody (~5%). One could, however, argue that a correct conformation (tertiary structure) is also necessary for optimal binding of the antibody. It is not clear whether such conformation is determined only by a unique sequence of monosaccharides. The affinity of the MoAb varies when heparin is in solution or bound to poly+lysine-coated plates, as described below. Different conformations are expected under these conditions. The epitope recognized by ST-1 is tentatively defined as a decasaccharide on the basis of inhibition assays of antibody binding to heparin using different oligosaccharides prepared by partial deamination of heparin with nitrous acid. Decasaccharides were the smallest fragments able to completely inhibit binding of ST-1 to heparin. Octa-, hexa-, and tetrasaccharides were able to inhibit up to 30% of antibody binding. Tetradecasaccharides were better inhibitors at low concentrations as was the intact heparin molecule at a much lower concentration. These results suggest that conformation, in addition to the linear sequence of carbohydrate units, is important for ST-1 reactivity. The possibility that the 2,5-anhydromannose unit is a structural requirement for antibody recognition is discarded since intact heparin is highly reactive with ST-l. Recently, Gitel et al. (17) produced rabbit anti-heparin polyclonal antibodies able to react with heparin linked noncovalently to PHI. These antibodies, however, did not react with free heparin. The authors tentatively defined the size of the epitope in the deca- to tetradecasaccharide range. They also hypothesized that a favorable conformation of the heparin molecule was required for recognition by the anti-heparin antibodies. The MoAb ST-l, although selected by utilizing heparin linked noncovalently to plates precoated with poly-L-lysine, differs from the polyclonal antibodies described by Gitel et al. (17) in at least three aspects: it is able to interact with heparin or heparin-derived oligosaccharides in solution; it does not require binding of heparin or heparin-derived oligosaccharides to positively

ANTI-HEPARIN

MONOCLONAL

charged polymers like PHI or poly+lysine; and it does not bind to heparin bound to Sepharose. In fact, free heparin is able to inhibit the binding of ST-l to plates coated with poly+lysine-heparin. The MoAb ST-l also bound equally well to plates coated with AT-III heparin with no poly-L-lysine used for adsorbtion (data not shown). Comparing solid-phase RIA experiments in which heparin was noncovalently linked to a matrix of polylysine with immunoprecipitation experiments using free heparin in solution we observed that under the former condition heparin molecules where in a more favorable conformation to interact with ST-l. It is possible that the epitopes present on a heparin molecule when bound to poly-L-lysine permit a better fit of ST-l. Conversely, the MoAb itself might have been elicited by heparin molecules in a similar conformation when bound to the surface of S. minnesota used as carrier for immunization. Poly-L-lysine itself, when tested in solution, has an inhibitory effect on the binding of ST-l to the heparincoated plates, probably by making the acidic groups of the heparin epitopes unaccessible to ST-l (not shown). The possibility that the MoAb ST-l could react with contaminating proteins instead of heparin was eliminated by the fact that in all our assay we utilized highly purified heparin as well as heparin-derived oligosaccharides obtained by deamination at pH 1.5. Under these conditions it is highly improbable that MoAb ST-l is recognizing heparin-bound peptide or protein. Furthermore, an assay in which the plates were coated with heparin preincubated with AT-III did not show any alteration in the binding capacity of ST-l to heparin. This indicates that AT-III did not block or enhance binding of ST-l to heparin. In immunoprecitation experiments, we were able to immunoprecipitate 3H-labeled heparin in solution with ST-l. Such immunoprecipitation was specifically inhibited by unlabeled heparin but not by unlabeled chondroitin 6-sulfate, both in solution as described above; the reactivity of the MoAb with heparin in solution was much weaker than that with plate-bound heparin. The ST-l MoAb is of potential interest in biochemical studies involving heparin and heparin-like molecules. It can be of help for mapping heparin or heparinlike polymers on the cell surface, for analyzing their roles in cellular functions such as tumor angiogenesis (25,26), and for quantitative determination of heparin by immunoassays (l’i’), binding to growth factors (2729), proliferation of smooth muscle cells and endothelial cells (30,31), or binding to fibronectin (32). The advantage of a MoAb that recognizes native heparin is obvious, since it can be used in studies on heparin structure determination and expression without chemical modification. Microanalysis of heparin by enzymes (33,34), deamination cleavage (35,36), and discon-

7

ANTIBODY

tinuous gradient gel electrophoresis (37,38) can be performed with better accuracy using the ST-l antibody.

ACKNOWLEDGMENT The authors express corned, San Francisco, standards.

their gratitude to Dr. H. E. Conrad of GlyCA, for the heparin-derived oligosaccharide

REFERENCES 1. Houghton, A. N., Mintzer, D., Cordon-Cardo, C., Welt, S. W., Fliegel, B., Vadhan, S., Carswell, E., Melamed, M. R., Oettgen, H. F., and Old, L. J. (1985) Proc. N&l. Acad. Sci. USA 82,12421246. 2. Itzkowitz, S. H., Yuan, kahashi, H. K., Bigbee, 49, 197-204. 3. Bechtel, Thurin,

M., Montgomery, W. L., and Kim,

B., Wand, A. J., Wroblewiski, J. (1990) J. Biol. Chem. 265,

C. K., Kjeldsen, T., TaY. S. (1989) Cancer Res.

K., Koprowski, 2028-2037.

4. Kannagi, R., and Hakomori, S. (1986) in Handbook tal Immunology (Weir, D. M., Ed.), 4th ed., Chap. Scientific, Oxford.

H., and

of Experimen117, Blackwell

5. Pukel, C. S., Lloyd, K. O., Travassos, L. R., Dippold, W. G., Oettgen, H. F., and Old, L. J. (1982) J. Elep. Med. 156,1133-1147. 6. Takahashi, H. K., Metoki, Res. 48,4361-4367.

R., and

7. Dohi, (1990)

T., Ohta, S., Hanai, J. Biol. Chem. 265,

N., Yamaguchi, 7880-7885.

8. Feizi,

T., and Childs,

R. A. (1987)

Hakomori,

S. (1988)

Cancer

K., and Oshima,

Biochem.

J. 245,

M.

l-11.

U., Larm, O., Scholander, E., Sandgren, 9. Pejler, G., Lindahl, and Lundblad, A. (1988) J. Biol. Chem. 263,5197-5201.

E.,

10. Couchman, J. R., Caterson, B., Christener, J. E., and Baker, (1984) Nature 30’7,650-652. 11. Kure, S., and Yoshie, Y. (1986) J. Zmmunol. 13’7,3900-3908.

J. R.

12. Caterson, B., Christner, Chem. 258,8848-8854.

J. E., and

Baker,

J. R. (1983)

J. Biol.

13. Mehmet, H., Scudder, P., Tang, P. W., Hounsell, E. F., Caterson, B., and Feizi, T. (1986) Eur. J. Biochem. 157, 385-391. 14. Yamagata, M., Kimata, K., Oike, shida, K., Shimomura, Y., Yoneda, Biol. Chem. 262,4146-4152. 15. Levy, D. E., Horner, 153,883-896. 16. Beaumont, 17. Gitel, Med.

A. A., and Solomon,

J. L., and Lemort,

S. N., Medina, 109,672-678.

18. Hook, M., Bjork, Lett. 66, 90-93.

Y., Tani, K., Maeda, N., YoM., and Suzuki, S. (1987) J.

V. M.,

I., Hopwood,

N. (1974) and Wessler,

A. (1981) Pathol.

Biol.

S. (1987)

J., and Lindahl,

J. Exp. 22, J. Lab.

U. (1976)

Med. 67-76. Clin. FEBS

19. Takahashi, H. K., Nader, H. B., and Dietrich, C. P. (1985) Arzneim-Forsch.lDrug Res. 35,1620-1623. 20. Shively, J. E., and Conrad, H. E. (1976) Biochemistry 15, 39433950. 21. Knudson, W., Gundalach, M. W., Schmid, T. M., and Conrad, H. E. (1984) Biochemistry 23,368-375. 22. Kosakai, M., and Yosizawa, Z. (1979) J. Biochem. 86, 147-153.

8

STRAUS,

TRAVASSOS,

23. Danishefsky, I., and Steiner, H. (1965) Biochim. Biophys. Acta

31. Wright, T. C., Johnstone, T. V., Castelot, J. J., and Karnovsky, M. J. (1986) J. Cell. Physiol. 125,499-506. DiFerrante, N., Nichols, B. L., Donnely, P., Neri, G., Argovic, R., 32. Sekigushi, K., Hakomori, S., Funahashi, M., Matsumoto, I., and and Berglund, R. K. (1971) Proc. Nutl. Acctd. Sci. USA 68, 303Seno, N. (1983) J. Biol. Chem. 268, 14,359-14,365. 345. 33. Yoshida, K., Miyauchi, S., Kikuchi, H., Tawada, A., and Okuyasu, Folkman, J., Langer, R., Linhardt, R., Haudenschild, C., and K. T. (1989) And. Biochm. 177,327-332. Taylor, S. (1983) Science 221,719-725. 34. Yoshida, K., Miyazono, H., Tawada, A., Kikuchi, H., Morikawa, Chum, R., Szabo, S., and Folkman, J. (1985) Science 230,1375K., and Tokuyasu, K. (1989) in Proceedings, Xth International 1378. Symposium on Glycoconjugates, Jerusalem, Israel (Sharon, N., Lis, H., Duksin, D., and Kahane, I., Eds.), p. 330. Maciag, T., Mehlman, T., and Friesel, R. (1984) Science 225, 932-935. 35. Bienkowski, M. J., and Conrad, H. E. (1985) J. Bill. Ckm. 260, 356-365. Sullivan, R., and Klagsbrun, M. (1985) J. Biol. Chem. 260,23992403. 36. Guo, Y., and Conrad, H. E. (1989) And. Biochem. 176,96-104. Lobb, R. R., Sasse, J., Sullivan, R., Shing, Y., D’amores, P., Ja37. Rice, K. G., Rottink, M. K., and Linhardt, R. J. (1987) Biochem. J. cobs, J., and Klagsbrun, M. (1986) J. Bid. Chem. 261,1924-1928. 244,515-522. Wright, T. C., Castelot, J. J., Petitou, M., Lormeau, J.-C., Choay, 38. Turbull, J. E., and Gallagher, J. T. (1988) Biochm. J. 261,597C., and Karnovsky, M. J. (1989) J. Biol. Chem. 264, 1534-1542. 608.

101,37-45.

24.

25. 26. 27. 28. 29. 30.

AND TAKAHASHI