High affinity human monoclonal antibodies directed against hepatitis B surface antigen

Journal of Virological Methods,

16 (1987) 281-292

281

Elsevier JVM 00602

High affinity human monoclonal antibodies directed against hepatitis B surface antigen Claude Desgranges’, Jacqueline Pairel, Christian Pichoud2, Sylvie Souche l, Dominique Frommel and Christian Trepo2 “Laboratoire d’Epid&niologie et Immunovirologie des Tumeurs, Faculd de Mtdecine Alexis Carrel, Lyon, France; 21Jnitd INSERM V 271 de Recherches sur les Hkpatites, Lyon, France; ‘Laboratoire de Parasitologic Exptrimentale, Facultk de Mddecine Piii&Salpt%rikre, France

(Accepted 12 March 1987)

Summary

Peripheral blood mononuclear cells from donors immunized with hepatitis B vaccine (Pasteur Hevac B) were transformed with Epstein-Barr virus. Two polyclonal cell lines, producing antibodies to hepatitis B surface antigen were established and cloned. Seven clones were isolated; they secreted between 10 and 20 l&ml of HBs specific IgGl K or A antibody with anti-HBs titer of 300400 IU/ml. These human antibodies expressed the anti ‘a’ specificities and had high affinity and avidity; their potential use as reagents for hepatitis B virus detection and for passive immunotherapy is under study. Hepatitis B; anti-HBs antibody; Epstein-Barr

virus; Lymphoblastoid

cell line

Introduction

Human hepatitis B virus (HBV) is a small hepatotropic DNA virus that produces acute and chronic liver diseases (Blumberg, 1977; Tiollais et al., 1981). The majority of persons infected with HBV develop an acute hepatitis from which they make a complete recovery. When HBV disappears from their serum, antibodies to hepatitis B surface antigen (HBsAg), an envelope antigen present on the three *Institution at which work was performed. Correspondence to: Dr. Claude Desgranges, Departement des Anticorps Monoclonaux, Laboratoire d’Epidemiologie et Immunovirologie des Tumeurs, Faculte de Medecine Alexis Carrel, 69372 Lyon Cedex 8. France.

0166-0934/87/$03.50 0 1987 Elsevier Science Publishers B.V. (Biomedical Division)

282

types of circulating HBV-related particles, appear in serum during convalescence; these antibodies clearly protect against future infection with HBV. A small percentage of HBV infected persons become chronic HBsAg carriers and fail to develop anti-HBs antibodies. Chronic HBV infection is often associated with chronic acute hepatitis and favors the development of hepatocellular carcinoma (Szmuness, 1978; Beasley et al., 1981a). Prompted by a number of reports on the efficiency of administration of HBV immune globulins to prevent acute infection, HBsAg carrier state and perinatal transmission (Kanai et al., 1985; Wong et al., 1984; Beasley et al., 1981b), or for immunization programs (Celis et al., 1986), we attempted to develop stable lymphoblastoid cell lines (LCLs) offering permanent sources of invariable monoclonal human anti-HBs antibodies (HBsAb). Herein, we report the successful establishment of seven such human monoclonal antibodies (HMAbs) against HBsAg and present the characteristics of the LCLs and of the antibodies secreted.

Materials Leucocyte

and Methods donors

Twenty ml of heparinized blood were obtained from two selected blood donors vaccinated with Hevac B, subtypes ad + ay (Institut Pasteur Production, Paris), who developed serum anti-HBs titer of 290 and 400 IUlml, respectively. Their antiHBs antibody titers were stable for about 1 yr and they were vaccinated for 3 yr. Epstein-Barr

virus transformation and cloning of B cells

Mononuclear cells were isolated by Ficoll-Hypaque density gradient centrifugation. One ml of virus supernatant from the Epstein-Barr virus (EBV) producing marmoset cell line B95-8 (Miller et al., 1972) was added to 5 x lo6 cells for 1 h at 37°C. The cells were then suspended at 5 x lO’/ml in RPM1 1640 supplemented with 15% fetal calf serum (FCS), penicillin (100 IU/ml), streptomycin (100 pg/ml), glutamine (2 mM) and cyclosporin A (0.1 ug/ml) and cultured at 37°C with 5% CO, in 384 well plates (Greiner, Bischwiller, France). Cell culture supernatants were screened 3 or 4 wk later for HBs-specific antibodies, using a solid-phase radioimmunoassay (RIA). Cloning of the positive wells was done in limiting dilution (1 or 10 cells/well) in plates with 96 round-bottom 0.2 ml wells, which contained a feeder layer of lo5 allogenic peripheral blood mononuclear cells, from an EBV antibody negative donor previously irradiated (2000 R). After 3 or 4 wk, the supernatants of wells with proliferating cells were tested for the presence of HBsAb. After cloning, the cells were grown in RPM1 with 5 or 10% FCS or in synthetic media (ADCM, Cegen, Lyon, France) in large volume.

283

Detection of hepatitis B surface antibodies

Supernatants of cell cultures were tested for the presence of HBsAb by RIA using AUSAB commercial kits (Abbott Laboratories, Chicago, U.S.A.). As for the titration of HBsAb in sera, the binding of labelled antigen in excess of 2.1 times of negative control indicated the presence of specific antibodies in culture supernatant. Antibody titers were determined in comparison with a reference preparation of human anti-HBs immunoglobulins (100 IU/ml ref. Centre National de Transfusion Sanguine, Paris). lmmunochemical

characterization

of the monoclonal antibodies

The heavy- and light-chain types of the secreted HMAbs were determined by immunoelectrophoresis. Rabbit antisera (specific for human lo,, (Y, y, A, and K chains) were obtained from Dakopatts (Copenhagen, Denmark). The quantitation of the amount of Igs secreted in the supernatant medium was achieved by a sandwich enzyme linked immunosorbent assay (ELISA) also with Dakopatts monaspecific antisera (Renversez et al., 1983). The IgG subclasses were characterized by double diffusion gel analysis with monoclonal antibody specific for the four subclasses (Organon Teknika, Paris, France). Electrophoresis of the monoclonal Igs was performed on a 4-30% of sodium dodecyl sulfate (SDS) polyacrylamide gel with standard protein markers and with a known human monoclonal IgGi. Isoelectric focusing (IEF): the monoclonal IgGs were focused on polyacrylamide gels with standard p1 markers (Pharmacia pH 3.5-9.5 Ampholine gradient). Cell surface or intracytoplasmic Igs of the cell lines producing the monoclonal antibodies were studied by direct immunofluorescence on fresh or methanol fixed cells, with fluorescein-conjugated monospecific antibodies of F(ab)‘, fragments to human l.~,OL,y, K and A chains (Biosys, Compiegne, France). Epstein-Barr nuclear antigen (EBNA) was detected by anti-complement immunofluorescence (Reedman et al., 1973), early antigen (EA) and viral capsid antigen (VCA) by indirect immunofluorescence (Henle and Henle, 1966). Purification and labelling of the monoclonal

antibodies

Immunoglobulins were isolated from culture supernatants of secreting clones by ammonium sulphate precipitation (33% w/v final), and sequential chromatographies on a DEAE-trisacryl (IBF, France) and on sepharose protein A (Pharmacia, France) columns. The purified monoclonal anti-HBsAb were radiolabelled with 1251by the chloramine T technique (Hunter et al., 1965). The specificity of the labelled antibodies was tested with polyclonal or monoclonal anti-HBs coated beads from the commercial kits (Ausria II and Ausria monoclonal Abbott and RIA monoclonal Centocor, Malvern, U.S.A.) incubated with a panel of HBsAg positive and negative sera.

284

Binding studies of the human monoclonal

antibodies

Anti-HBs binding activity of the seven HMAbs was measured after incubation with HBsAg coated beads (AUSAB, Abbott) for various times (10 min-24 h), in parallel with other radiolabelled mouse monoclonal antibodies (Ausria monoclonal Abbott and RIA Centocor) and with one human polyclonal antibody (Ausria II, Abbott) with similar specific activities (0.75 mCi/ml). The avidity after 10 min and 24 h was determined for the seven HMAbs. For evaluation of the constant association (K,) between HBsAg and the different HMAbs, 200 u,l of different dilutions of the seven radiolabelled antibodies were incubated with beads coated with a constant amount of HBsAg (AUSAB. Abbott) for 2 h at room temperature. After washings, the fixed radioactivity was determined comparatively to the free radioactivity. The apparent association constant (K,) was obtained by scatchard plot analysis (Pesce et al., 1978), after determination of specific activity of each monoclonal antibody; the antibody concentration ranging from 10 pg-1 mg per ml was achieved by optical density. Anti-HBs binding activity of 3 HMAbs was also assessed on the human hepatoma cell line: PLC/PRF/S (Alexander et al., 1976) which expresses HBsAg on the cell surface. The cells were trypsinized 48 h before seeding at lo5 cells in 24 well 12’1 labelled anti-HBs HMAbs or ‘251-labelled plate in 1 ml of complete medium. anti-delta antibodies (anti-delta, Abbott diagnostic kit) at 1.106 cpm were incubated 12 h at 37°C in triplicate with dilutions from l/4-1/1280 as described previously (Shouval et al., 1982). After three washings, 1 ml of 1% SDS was added to each well. After 4 h of incubation at 37”C, the cell lysates were counted for ‘*‘I bound antibodies. As control, binding of each radio-labelled antibody at their different dilutions was assessed on HBsAg coated beads (Ausria, Abbott). Immunodiffusion Double immunodiffusion analysis was performed in gel made of agarose 0.5% (Indubiose A37, Pharmindustrie, Clichy, France) and of 2% polyethylene glycol 6000 in 0.04 mol/l barbiturate buffer, pH 8.2. The HMAbs were tested against reference HBsAg positive sera of anti-ad and anti-ay subtypes, obtained from Centre National de Transfusion Sanguine (Paris). Polyacrylamide

gel electrophoresis

Polyacrylamide gel electrophoresis (PAGE) of ‘*‘I-HBs antigen (0.74 &i/ml, AUSAB, Abbott) precipitated by supernatants of monoclonal cultures positive or negative for anti-HBsAb, was carried out on SDS polyacrylamide gel (ll%), according to Laemmli (1970). lmmunohistologic

analyses

Six liver biopsies of HBV-associated hepatitis and 4 of extra-hepatic atresia with varying degrees of fibroplasia were studied. Sections were made from snap-frozen

285

tissues for analysis with fluorescent conjugates and from paraffin-embedded tissue, initially fixed in Bouin’s solution, for immuno-enzymatic reactions. In the latter case, sections were deparafhnized by sequential treatment with xylene and ethanol, endogenous peroxidases were inactivated by a 10 min incubation in”absolute methanol containing 0.1% of H,O,. The first antibody overlayed for 30 min at room temperature was the HMAbs or polyclonal antibodies (rabbit anti-HBsAg or goat anti-HBsAg). After washing in phosphate buffered saline, second labelled antibodies adsorbed on acetone-dried mouse and rat liver powder (25 mg/ml) were used (fluorescein-labelled goat anti-human Ig, Institut Pasteur Production, Paris, peroxidase-conjugated rabbit anti-human Ig, Hoechst Behring, Rueil Malmaison, France). Peroxidase activity was revealed by a 5 min incubation in 3,3’-diaminobenzidine saturated at pH 7.35 and filtered solution, supplemented with 0.01% of H,O*. Fluorescein-treated slides, mounted in PBS-glycerol, were examined with a Leitz Dialux 20 incident fluorescence microscope.

Results Generation of anti-HBs secreting cell lines

Seven clones secreting monoclonal antibodies against HBsAg with a titer above 250 IU/ml in the supernatant were grown and analyzed in details. One clone, Hl, was derived from the first donor and six clones, H2-H7, from the second (see Table 1). These clones were established after two cycles of cloning at one cell per well and are now stable for over 1 yr in continuous culture. The anti-HBs titers increased progressively with the different cloning cycles. Properties of cloned LCLs secreting HBs antibody

These positive clones are growing in RPM1 1640 + 5% FCS at 4 x lo5 cells/ml and have a double time of 24 h. In about 90% of the cells of the Hl line, membrane and cytoplasmic fluorescence were positive for IgGK, while the six clones deriving from the second donor stained for IgGh. All the LCLs express EBNA, but none EA or VCA or transforming EBV in the supernatants. For large scale production, these cells are cultured in the synthetic ADCM medium, either in large flasks or in fermenters. Properties of the HMAbs produced

by these LCLs

The IgG concentration of the different LCLs supernatants reaches 5-10 t.&ml after three days and 20-50 kg/ml after 1 wk of culture with anti-HBs titer ranging between 280-800 IU/ml (Table 1). Concentrated supernatants, analyzed by immunoelectrophoresis and double diffusion analysis had the same 1gGl~ and IgGlh specificities as those observed on their respective parental cells. The isoelectric focusing pattern of the Hl IgGlK reveals fine bindings limited to a narrow zone (p1

5.6 1.8 0.8 0.7 0.9 1.0

8.3

10 min

18.0 3.4 3.8 3.0 3.5 3.0

44.8

24 h

Binding activity on commercial beads” (%) after

8.5 ND ND 30.0 ND ND

3.2

Binding activity on PLCIPRFW (%) after 12 h

total ‘*‘I-Iabelled anti-HBs added ND = not done. For binding activities and apparent constant association determination,

all the assays were run in quadruplicate

a Culture supernatants were tested for anti-HBs activity by RIA using commercial kits (AUSAB, Abbott). b Values represented anti-HBs titer in culture supernatant after 3 days of culture with a cell concentration of lo6 cells/ml. c 1251-labelledanti-HBs bound on commercial beads x 100. total ‘Z’I-labelled anti-HBs added ‘, (‘251-labelled anti-HBs bound on PLCIPRFIS cells)i(‘2SI-labelled anti-HBs bound on commercial beads) x loo,

KgGlA

H2 H3 H4 H5 H6 H7

6/141

7148

2

620 780 280 730 320 350

IgGlk

Hl = 800

1134

4148

1

= = = = = =

Ig class

N stable clones with antiHBs activityb > 200 IUiml

N wells with anti-HBs activity”

Donors

Properties of human monoclonal anti-HBs antibodies.

TABLE 1

6.2 3.2 3.5 2.1 2.7 3.1

x x x x x x

lOi 10’” lo9 lo9 lo9 lo9

1.8 x 10”

Apparent K, limo1

287

Fig. 1. Agarose IEF. (1) and (4) = pI markers. (2) = 10 )~i supernatant of HI and (3) = 5 ~1 supernatant of Hl were applied on 0.5 mm thick gel with pH 3.5-9.5 Ampholine gradient (LKB, France) and run for 45 min at 15 V.

8.45-8.65) (Fig. 1). These homogeneous IEF pattern and restriction to a single heavy and light chain are consistent with the monoclonal nature of the LCLs. Anti-H&

specificity of the HMAbs

Hl cell line exhibited the highest anti-H& titer: 800 XUiml after three days of culture with the same final cellular concentration (Table 1). The values of rC,, representing the apparent constant association between the antigen and antibody molecules ranged between 2.1 X lop to 1.8 X 10” I/mol. The ability of the labelled HMAbs to interact with FIBsAg linked to a solid-phase support (commercial beads) is reported in Table 1, after incubation of 10 ruin and 24 h. II1 and H2 bind faster than the other HMAbs. Fig. 2 illustrates the time-

288

TIME

HOURS

Fig. 2. Binding of ‘251-labelled anti-HBs to HBsAg on commercial beads (Abbott, Ausria II), Hl, H2, H5 = human monoclonal antibodies, CM = commercial anti-HBs monoclonal reagent (0 = Centocor, + = Abbott), CP = commercial anti-HBs polyclonal reagent (Abbott).

course binding of three ‘251-labelled HMAbs, studied in comparison with that of commercially available monoclonal and polyclonal reagents. The binding capacity of Hl and H2 is similar to that of murine monoclonal but greater than that of the two other commercial preparations, while H5 has in this system a low avidity. The binding of the HMAbs to HBsAg present on the surface membrane of the PLC/PRF/S, hepatoma cell line is expressed in relation to the HMAbs uptake on HBsAg coated beads; in this instance, H5 shows highest binding value, 30% of the cpm added, versus 3.2 and 8.5% for Hl and H2, respectively (Table 1). No uptake of anti-delta antibodies by the PLC/PFR/S cells was observed. In double diffusion analysis, all HMAbs formed a precipitation line of complete identity with the eight different HBsAg subtypes by recognizing the common ‘a’ determinant. In immunoprecipitation of ‘251-HBsAg Hl and H2 monoclonals react with two HBsAg components, with molecular weights of 29 and 26 kDa, which correspond to the major Pl and P2 HBsAg proteins (Fig. 3). HBsAg localization in liver tissue with the HMAbs

The two biopsies of chronic active hepatitis, examined by indirect immunofluorescence, featured intra-cytoplasmic localization of HBsAg present in the form of either a homogeneous mass or of microgranular deposits, occupying often a portion only of the cytoplasm. In the four biopsy specimens of chronic persistent hepatitis examined with peroxidase-coupled reagents, HBsAg was visualized in the cytoplasm of hepatocytes, occasionally enlarged (Fig. 4). The number of HBsAg

289

Fig. 3. Immunoprecipitation of ‘251-labeled HBs antigens with supernatants of Hl and H2 anti-HBs positive clones (lanes 2 and 4) and of anti-HBs negative clones (lanes 1 and 3).

cells as well as the intensity of specific staining varied between lobules and individual biopsies. On the six biopsies, the localization of the HBs antigenic material by the monoclonal antibodies was superimposable to that observed with poIyclonal antibodies, aIthough the former delineated better the HBsAg cell membrane pattern. The sections of liver pathology unrelated to HBV did not show cells positively stained with these HMAbs. positive

Discussion We have obtained anti-HBsAg human monoclonal antibodies by immortalization of B cells from selected anti-HBs vaccinated blood donors. All established LCLs secrete IgG, with specific titers of 280-800 IU/ml; these cell lines are stable

290

Fig. 4. Section of liver biopsies with Hl monoclonal antibodies. Chronic persistent hepatitis: cytoplasmic localization of HBsAg in isolated hepatocytes (Hl HMAb -t anti-human IgG peroxidase-conjugated globulin). Magnification x 960. for not more than 1 yr. Two of these HMAbs display an avidity and affinity comparable to the murine monoclonal reagent described by Wands et al. (1981a). These HMAbs identify the two major HBsAg proteins and localized HBsAg on HBVrelated liver section of chronic active and persistent hepatitis. The HMAbs also bind to the,HBsAg secreting hepatoma cell line PLCIPRFIS; this binding, however, does not parallel that observed with HBsAg linked to commercial beads. Since our monoclonals recognize the native form of HBsAg by immunoblot analysis (unpublished results), it appears likely that there are co~formational differences between the viral antigens, either present on the outer membrane of the hepatoma cells or associated to artificial supports. Two previous reports described anti-HBsAg HMAbs after fusion of human B lymphocytes with a myeloma partner (Stricker et al., 1985; Maeda et al., 1986). We also developed such hybridomas, but in our hands their secretion was unstable

291

and decreased markedly after 6 mth; in contrast, some LCLs are producing monoclonal antibodies for 5 yr or more (Seigneurin et al., 1983). The anti-HBsAg cell lines are presently growing in large fermenters (120 1 Biolafitte, France) in a synthetic and defined medium and are yielding IgG at a concentration of 50 kg/ml with an anti-HBs titer of 2500 IU/ml (C. Blourde, personal communication). LCLs can thus provide HMAbs in large amounts to develop better tests for HBsAg detection, although some performance tests using murine monoclonal antibodies are available (Wands et al., 1981b; Goodall et al., 1982; Soulier et al., 1983). Beside numerous applications for diagnostic purpose, the main interest of these HMAbs should pertain to immunoprophylaxis. Successful strategies for active and/or passive prevention of HBV infection became possible with the availability of safe vaccines and high titer immunoglobulin preparation (Maupas et al., 1981; Hamilton, 1983). Administration of hyperimmune anti-HBsAg gammaglobulins successfully interrupted the transmission of HBV infection to infants born from HBsAg positive mothers (Kanai et al., 1985; Wong et al., 1984; Beasley et al., 1981b). LCLs producing high titer anti-HBsAg HMAbs may offer an unlimited source of well-defined antibodies, which could be made available at a lower cost than immunoglobulin extracted from plasma of hyperimmune donors. It has been suggested that antibody purified from culture supernatant of EBV-transformed cell lines carries an unacceptable risk of oncogenesis due to possible contamination with infectious EBV or viral DNA. However, EBV is a ubiquitous agent present in B lymphocytes and in saliva of 90% of the adult population and in most blood samples used for transfusion. No ill effect has been noted even in immunocompromised recipients. Nevertheless, this human monoclonal antibody will be purified by rigorous procedures with a scheme already described (Crawford et al., 1983). Finally, the finding that antibodies to an interspecies idiotype associated with antiHBsAg activity induced an anti-HBs response when injected into mice (Kennedy et al., 1984), urges us to try and produce human monoclonals with the ‘internal image’ of the anti HBs HMAbs with the same technology. These anti-idiotypes may become useful as vaccines against HBV, as they have the ability to mimic structurally very complex antigenic components by the nature of their three dimensional conformation (Kennedy et al., 1986). Acknowledgements

This work was supported in part by contracts with the Caisse Nationale de 1’Assurance Maladie des Travailleurs Salaries (CNAMTS), Minis&e de l’lndustrie et de la Recherche (MIR) and Pole Rhone-Alpes de Genie Biologique et Medical (GBM).

References Alexander, J.J., Bey, E.M. and Geddes, E.W. (1976) S. Afr. Med. J. 50, 2124. Beasley, R.P., Lin, C.C. and Hwang, L.Y. (1981a) Lancet ii, 1129.

292 Beasley, R.P., Hwang, L.Y., Lin, C.C., Stevens, C.E., Wang, K.Y., Sun, T.S., Hsieh, F.J. and Szmuness, W. (1981b) Lancet ii, 388. Blumberg, B.S. (1977) Science 197, 17. Crawford, D.H., Huehns, E.R. and Epstein, M.A. (1983) Lancet i, 1040. Celis, E., Chang, T.W., Stricker, L., Tiebout, R.F. and Zeijlemaker, W.P. (1986) Lancet i, 1219. Goodall, A.M., Meek, F.L., Waters, J.A., Miescher, G.C., Janossy, G. and Thomas, H. (1982) J. Immunol. Methods 52, 167. Hamilton, J.D. (1983) J. Am. Med. Assoc. 250, 2145. Henle, G. and Henle, H. (1966) J. Bacterial. 91, 1248. Hunter, W.M. and Greenwood, F.C. (1965) Nature 194, 495. Kanai, K., Takehiro, A., Noto, H., Nishida, M., Takahashi, K., Kawashima, Y., Igarashi, Y., Matsushita, K. and Shimizu, M. (1985) J. Infect. Dis. 151(2), 287. Kennedy, R.C., Melnick, J.L. and Dreesman, G.R. (1984) Science 223, 930. Kennedy, R.C., Henkil, R.D. and Dreesman, G.R. (1986) Stand. J. Immunol. 23, 481, Laemmli, U.K. (1970) Nature 227, 680. Maeda, T., Eda, Y., Nishiyama, K., Ishikawa, Y., Tashiro, A. and Watanabe, T. (1986) Hybridoma 5, 33. Maupas, P., Chiron, J.P. and Barin, F. (1981) Lancet i, 289. Miller, G., Shope, T., Lisco, H., Stitt, D. and Lipman, M. (1972) P rot. Natl. Acad. Sci. U.S.A. 69, 383. Pesce, A.J., Ford, D.J. and Gaizutis, M.A. (1978) Stand. J. Immunol. 8, 1. Reedman, B.M. and Klein, G. (1973) Int. J. Cancer 11, 499. Renversez, J.C., Roussel, B., Baccard, M. and Seigneurin, J.M. (1983) In: Immunoenzymatic Techniques (Avrameas, S., Druet, P., Masseyeff, R. and Feldmann, G., eds.), p. 313. Elsevier, Amsterdam. Seigneurin, J.M., Desgranges, C., Seigneurin, D., Paire, J., Renversez, J.C., Jacquemont, B. and Micouin, C. (1983) Science 221, 173. Shouval, D., Wands, J.R., Zurawski, V.R., Isselbacher, K.J. and Shafritz, D.A. (1982) Proc. Natl. Acad. Sci. U.S.A. 79, 650. Soulier, J.P., Courouce, A.M., Lee, H., Drouet, J., Muller, A. and Canavaggio, M. (1983) Biotest Bull. 4, 353. Stricker, E.A.M., Tiebout, R.F., Lelie, P.N. and Zeijlemaker, W.P. (1985) Stand. J. Immunol. 22, 337. Sugg, U., Schneider, W., Hoffmeister, H.E., Huth, C.H., Stephan, W., Lissner, R. and Haase, W. (1985) Lancet i, 405. Szmuness, W. (1978) Prog. Med. Virol. 24, 40. Tiollais, P., Charnay, P. and Vyas, G.N. (1981) Science 213, 406. Wands, J.R. and Zurawski, V.R. (1981a) Gastroenterology 80, 225. Wands, J.R., Carlson, R.I., Schoemaker, M., Isselbacher, J. and Zurawski, V.R. (1981b) Proc. Natl. Acad. Sci. U.S.A. 78, 1214. Wong, V.C.W., Reesink, H.W., Ip, H.M.H., Lelie, P.N., Reerink-Brongers, E.E., Yeung, C.Y. and Ma, H.K. (1984) Lancet i, 921.