Glycobiology vol. 2 no. 3 pp. 217-224, 1992

Development and characterization of monoclonal antibodies against a mucin-type glycoprotein

Svein Haavik1-2, John F.Codington2-4 and Peter F.Davison3 'Department of Pharmacognosy, Institute of Pharmacy, PO Box 1068, Blindern, N-0316, Oslo 3, Norway, "Department of Cell and Molecular Biology, Boston Biomedical Research Institute, Boston, MA 02114, and Departments of 2Biological Chemistry and Molecular Pharmacology and 3 Ophthalmology, Harvard Medical School, Boston, MA 02115, USA "To whom correspondence should be addressed

Key words: epiglycanin/IgM/mucin-type glycoprotein/specificity of monoclonal antibodies/T-antigen

Introduction During the past years, many groups have demonstrated that carcinomas produce mucin-type glycoproteins that can be detected by monoclonal antibodies [see Hilkens (1988) for a review]. Some of the antibodies, e.g. B. 72.3 (Marsella et al, 1990), OC 125 (Dodd etai, 1985), CA 15.3 (Barak etai, 1990) and sialosyl Tn (Itzkowitz et al, 1990), have been used as cancer markers (Rasoul-Rockenschaub et al., 1989). Recently, the genes of some of the tumour-related mucins have been cloned and sequenced. This work indicated that the genes are © Oxford University Press

Results All anti-epiglycanin monoclonal antibodies investigated were found to be of the IgM class, based upon enzyme-linked immunosorbent assay (ELISA) of the positive hybridoma clone's supernatants and SDS-PAGE of the purified antibodies (results not shown). Their affinity constants (avidities) ranged from 7.5 x 108 (BF-11) to 6.6 x 1010 1/mol (CF-7) (Table I). These are among the highest avidities reported for IgM antibodies (Goding, 1986). The capacity for binding to epiglycanin of all the antibodies tested was reduced significantly when epiglycanin was subjected to either periodate oxidation, treatment with TPCKtrypsin or pronase (Table II). The extent of the reduction varied from 25 to 100% for the different antibodies. Incubation of epiglycanin with endo-a-N-acetyl-D-galactosaminidase (0-glycanase) reduced its ability to bind to eight of the antibodies. The activity against two of the antibodies (AD-7 and BH-11) was increased to 119 and 209%, respectively, by treatment of epiglycanin with this enzyme. In this experiment, Oglycanase treatment removed only - 5 0 % of the Gal/3(l-*3)GalNAc side chains on epiglycanin (data not shown). Treatment with sialidase increased the antigenic activity of epiglycanin when tested with the antibodies BF-11 and CEla. Fractionation of trypsin-treated epiglycanin by high-performance gel filtration, followed by assay of antigenic activity against monoclonal antibody CF-7, showed that fragments with 217

Downloaded from http://glycob.oxfordjournals.org/ at New York University on July 6, 2015

The preparation of > 30 different hybridomas, all secreting IgM class antibodies against epiglycanin, a glycoprotein at the surface of the mouse mammary carcinoma cell line TA3-Ha, is described. The specificities of 10 of the antibodies, with affinity constants in the range of lO8-!©10 1/mol were compared in an enzyme competitive binding assay. The affinity of epiglycanin was strongly reduced for all antibodies tested by incubation with periodate (10 mM, 4°C) and was reduced for most of the antibodies by endo-aiV-acetyl-D-galactosaminidase. This suggested that carbohydrate, and specifically the Gal/3(l-*3)GalNAc disaccharide, formed an integral part of the epitopes of most of the antibodies. The isolated disaccharide, however, exhibited 250 000 times less inhibitory activity in the competitive binding assay than epiglycanin. The binding capacity of epiglycanin was also reduced by incubation with trypsin or pronase, suggesting a high molecular weight dependency for binding. Incubation with sialidase increased its affinity for the antibodies. The binding of the antibodies to epiglycanin was strongly inhibited by peanut agglutinin, and to a lesser extent by lectins from Triticum vulgaris, Ricinus communis, Pisum sativum and Phaseolus vulgaris. None of the antibodies bound to any of eight different gangliosides immobilized on HPTLC plates. Mono- (Fab) and divalent [F(ab')2] fragments of the antibodies possessed very low affinity for epiglycanin. The results demonstrated that the specificities of the antibodies are related, but distinguishable, and they suggest that this epiglycanin-IgM model may be useful for studies on the general principles of the interaction between IgM antibodies and mucin-type glycoproteins.

polymorphic (Ligtenberg et al., 1990) and belong to a supergene family containing repetitive sequences (Gendler et al., 1988, 1990; Gum etai, 1990; Lan etai, 1990; Larocca etai, 1990). A correlation has been demonstrated between mucin production and metastatic potential in human colon cancer cells (Bresalier et al., 1991). Most of the mucin-specific antibodies raised against native glycoproteins bind to the carbohydrate moieties of the glycoproteins (Feizi 1984; Lan et al, 1985; Prat et al, 1989; Pancino et al, 1990), while others are reported to bind to the peptide cores of the molecules (Gendler etai, 1988; Layton etai, 1990). The detailed structures of most of the epitopes, however, are not known. Epiglycanin is the major sialomucin at the surface of the mouse mammary carcinoma ascites cell line TA3-Ha (Codington et al, 1972). It is an Mr 500 000 glycoprotein, - 8 0 % of which is carbohydrate, mostly as Gal/3(l-»3)GalNAc side chains linked to serine or threonine in the polypeptide chain (Van den Eijnden etai, 1979). A correlation has been demonstrated between the concentration of epiglycanin at the surface of the TA3-Ha cells and their allotransplantability (Codington etai, 1973, 1978). Poly- and monoclonal antibodies have been raised against epiglycanin. These antibodies have been demonstrated to cross-react with glycoproteins on human cancer cells (Codington et al, 1984). The purpose of the present paper was to study and compare the specificities of monoclonal IgM antibodies raised against epiglycanin.

S.Haavik, J.F.Codington and P.F.Davison

Table I. Affinity constants (1/mol, 20°C) of monoclonal anti-epiglycanin IgM antibodies (determined by ELISA) AD-7

ADD-12

BA-8

BF-6

BF-11

BH-11

CD-I

CEla

CElb

CF-7

8.7 x

3.2 x 109

6.5 x 109

1.1 X 10'°

7.5 x 108

2.0 x

4.0 x

2.6 x 10'°

5.6 x 10'

6.6 x 1010

Table D. Effect of modification of epiglycanin on its binding to anti-epiglycanin antibodies Remaining activity (% of native epiglycanin) Antibody

119 82 56 69 67 209 CD-I 51 CE-la 55 CE-lb 75 CF-7 62 Rab poly.* 41

Sialidaseb

Penodatec

Trypsind

Pronasee

135 278 186 217 92 204 223 97 154 132 210

8 1 12 18 9 0 1 0 0 20 8

38 1 32 74 27 5 2 27 30 12 45

36 1 36 1 31 0 9 0 9 56 19

Samples of epiglycanin were incubated with "endo-a-Af-acetyl-Dgalactosaminidase (O-glycanase), bsialidase (V.cholerae), dTPCK-trypsin, e pronase for 20 h at 37°C or c10 mM Na-meta periodate in 0.1 M NaOAc (pH 5.0) for 30 min at 4°C. Antigenic activities of modified samples were compared to that of native epiglycanin by an enzyme competitive binding assay. *Rabbit polyclonal anti-epiglycanin antibody (Codington et at., 1984).

mol. wts ranging from 50 000 to 400 000 were responsible for most of the activity remaining after the trypsin treatment (Figure 1). Fragments with mol. wts < ~50 000 expressed relatively little antigenic activity. In a competitive binding assay, comparison of the inhibitory activities of two mucins, before and after sialic acid removal (Table III), revealed that for all of the monoclonal antibodies tested, the binding of both the intestinal mucin from germ-free rats (Wold et al., 1975) and bovine submaxillary mucin was greater after desialylation. The antibodies that cross-reacted strongest with the mucins also exhibited the strongest crossreactivities with the desialylated mucins. Asialo intestinal mucin from germ-free rats exhibited the strongest cross-reactivity with all the antibodies tested. None of the Seven antibodies tested showed detectable binding to the gangliosides GM3, GD3, GM2, GM1, Fuc-GMl, GTlb, GDI a or GDlb, immobilized on HPTLC plates. The antibodies were tested at concentrations > 106-fold that required for epiglycanin detection. The binding of none of the antibodies was inhibited by the a-D-GalNAc-specific lectin from Griffonia simplicifolia or the CK-D-mannose/a-D-glucose-specific lectin from Lens culinaris (Table IV). The binding of all of the antibodies, except AD-7 and BA-8, was strongly inhibited by peanut agglutinin, a lectin which binds to the disaccharide Gal/3(l-*3)GalNAc (Liener et al., 1986). The antibodies AD-7 and BA-8 showed a pattern that differed from the rest of the antibodies by their relatively weak inhibition by peanut lectin and relatively strong inhibition by Ricinis communis agglutinin. 218

Discussion It has been stated that antibodies of the IgM type are of limited usefulness because of their relatively low affinity constants, multivalency and high molecular weights (Goding, 1986). All of the IgM antibodies studied in the present report, however, possessed relatively high affinity constants (Table I). They were sucessfully used in enzyme competitive binding assays capable of quantitating < 1 ng ( < 2 X 10~15 mol) of epiglycanin. Most of the antibodies were effectively purified by a combination of precipitation at low ionic strength and ABx™ HPLC. They were biotinylated without difficulty by the periodatehydrazide method (S.Haavik and J.F.Codington, in preparation). Purified native and biotinylated antibodies have been stored for > 1 year at 4°C without any notable loss of activity. All of the antibodies, including the rabbit antibody (Codington etai, 1984), followed the same general pattern of crossreactivity (Table III). This suggests that they bind to one or more structurally related epitope(s) on the molecule. The extent of cross-reactivity varied > 10-fold, however, demonstrating that the antibodies possess different, though related, properties. Marked differences among the monoclonal antibodies were also observed in their competition with lectins for sites on epiglycanin (Table IV). The observation that mild periodate oxidation of epiglycanin destroyed its antigenic activity against all the antibodies tested strongly suggests that the antibodies bind to carbohydrate-containing epitopes (Table II). Furthermore, the antigenic activity against eight of the monoclonal antibodies was reduced on treatment of epiglycanin with O-glycanase, an enzyme which removes only the disaccharide Gal|3(l-*3)GalNAc, O-glycosidically linked to serine or threonine residues on epiglycanin (Bhavanandan and Codington, 1983) (Table II). These results suggest that this disaccharide forms a part of the epitope recognized by some of the antibodies. The importance of the

Downloaded from http://glycob.oxfordjournals.org/ at New York University on July 6, 2015

AD-7 AD-12 BA-8 BF-6 BF-11 BH-11

O-Glycanase'

Each monosaccharidej methyl glycoside, /)-nitrophenyl glycoside, oligosaccharide, and dextran tested bound very weakly to the antibodies BF-11 and CF-7 (Table V). Although the disaccharide Gal/3(l-*3)GalNAc exhibited significant inhibitory activity in the competitive binding assay, the respective molar activities of the disaccharide for BF-11 and CF-7 were 3.3 X 108 and 6.7 X 107 times lower than that of epiglycanin. As shown in Table V, the two antibodies exhibited markedly different binding activities for a large number of monosaccharides and oligosaccharides. Some of these were as great as that of the disaccharide Galj8(l^*3)GalNAc. Assay of anti-epiglycanin activity in fractions from pepsincleaved antibodies (S.Haavik and J.F.Codington, in preparation) showed that F(ab')2"fragments of antibodies BH5b, CEla and AD-7 possessed very low anti-epiglycanin activity relative to that of the native IgMs (Figure 2). BF-11 was the only antibody tested that gave F(ab')2 fragments with anti-epiglycanin activity comparable to that of the native antibody.

Anti-epielycanin antibodies 300-

4

Epiglycanin

250-

3-

••

200-

t

Trypsin treated epiglycanin

4

150-

S

1

8

v

I

100-

50-

•J3

10

i

12

i

i

i

14

i

16

i

i

i

18

i

i

20

i

i

22

T

n n nn

i

24

i

26

T

T

i

28

i

30

r

i

32

i

i

i

V

i

34

i

36

i

i

38

40

i

i

i

42

i

44

r

i i

46

Fraction number (0.8 mL) Fig. 1. Chromatography of epiglycanin and trypsin-treated epiglycanin on coupled columns (2.5 x 250 mm each) of TSK G 2000-, 3000- and 4000 SW. Fractions were tested for anti-epiglycamn CF-7 binding activity by an enzyme competitive binding assay. Elution volumes of molecular weight markers are indicated. 1: IgM (900 000); 2: thyroglobulin (670 000); 3: ferritin (440 000); 4: catalase (232 000); 5: aldolase (158 000); 6: bovine serum albumin (67 000); 7. ovalbumin (43 000), 8: chymotrypsinogen A (25 000); 9: ribonuclease A (13 700); and 10: vitamin B l2 (1350).

Table IV. Cross-reactivities between monoclonal anti-epiglycanin antibodies and lectins in enzyme competitive binding assay

Table ID. Cross-reactivities between eipiglycanin and other mucin-typi glycoproteins. Inhibitory activity ( % by weight of epiglycanin)' Antibody

A

B

C

D

E

AD-7 AD-12 BA-8 BF-6 BF-11 BH-11 CD-I CE-la CE-lb CF-7 Rab poly.*

0.01 0.006 0.02 0.01 0.001 0.02 0.01 0.01 0.06