Neutralizing Human Fab Fragments against ... - Journal of Virology

JOURNAL OF VIROLOGY, Jan. 2002, p. 251–258 0022-538X/02/$04.00⫹0 DOI: 10.1128/JVI.76.1.251–258.2002 Copyright © 2002, American Society for Microbiology. All Rights Reserved.

Vol. 76, No. 1

Neutralizing Human Fab Fragments against Measles Virus Recovered by Phage Display Cristina de Carvalho Nicacio,1 R. Anthony Williamson,2 Paul W. H. I. Parren,2 Åke Lundkvist,1,3 Dennis R. Burton,2 and Ewa Bjo ¨rling1,4* Microbiology and Tumor Biology Center, Karolinska Institutet, S-171 77 Stockholm,1 Swedish Institute for Infectious Disease Control, S-171 82 Stockholm,3 and The South Hospital, Research Center, Karolinska Institutet, S-118 83 Stockholm,4 Sweden, and Departments of Immunology and Molecular Biology, The Scripps Research Institute, La Jolla, California 920372 Received 13 July 2001/Accepted 27 September 2001

Five human recombinant Fab fragments (Fabs) specific for measles virus (MV) proteins were isolated from three antibody phage display libraries generated from RNAs derived from bone marrow or splenic lymphocytes from three MV-immune individuals. All Fabs reacted in an enzyme-linked immunosorbent assay with MV antigens. In radioimmunoprecipitation assays two of the Fabs, MV12 and MT14, precipitated an ⬇80-kDa protein band corresponding to the hemagglutinin (H) protein from MV-infected Vero cell cultures, while two other Fabs, MT64 and GL29, precipitated an ⬇60-kDa protein corresponding the nucleocapsid (N) protein. In competition studies with MV fusion, H- and N protein-specific monoclonal antibodies (MAbs), the H-specific Fabs predominantly blocked the binding of H-specific MAbs, while the N-specific Fabs blocked MAbs to N. In addition, N-specific Fabs bound to denatured MV N protein in Western blotting. The specificity of the fifth Fab, MV4, could not be determined. By plaque reduction assays, three of the five Fabs, MV4, MV12, and MT14, exhibited neutralizing activity (80% cutoff) against MV (LEC-KI strain) at concentrations ranging between ⬇2 and 7 ␮g mlⴚ1. Neutralization capacity against MV strains Edmonston and Schwarz was also detected, albeit at somewhat higher Fab concentrations. In conclusion, three neutralizing Fabs were isolated, two of them reactive against the H glycoprotein of MV and another reactive against an undefined epitope. This is the first study in which MV-neutralizing human recombinant Fab antibodies have been isolated from phage display libraries.

the cellular immune response, which is thought to be directed predominantly against the N protein (12, 13), and the humoral immune response are important during an MV infection. As demonstrated by passive immunization against measles, antibodies alone are capable of protection against and contribute to the control of and recovery from MV infection (22). The importance of antibodies in the immunity against MV is also exemplified by protection of infants by maternal antibodies in the first months of life (16). Antibodies are induced to most viral proteins, but the major targets for the protective antibody responses are directed against the MV H and F proteins (5, 37). Although MV is generally considered to be an antigenically conserved virus, differences in the presence of specific epitopes defined by the binding of monoclonal antibodies (MAbs) have been described, showing that the H protein has the widest degree of variation between MV strains, while the F and N proteins are antigenically more conserved (34). This conclusion is supported by studies characterizing sequences of different MV strains (27, 28). H protein-specific MAbs have been shown to provide passive protection against encephalitis in rodents (14, 41), and vaccinia virus recombinants encoding H and F proteins have been shown to induce neutralizing antibodies in mice and protect them from lethal MV challenge (11, 39). Furthermore, in a cynomolgus monkey model similar results were obtained with recombinants expressing H and F proteins (36). Therefore, any new MV vaccine, be it a recombinant vector/protein, recombinant protein, or DNA vaccine,

Measles virus (MV) infection is rare in industrialized countries today, thanks to the safe and effective live attenuated vaccine. However, measles is still one of the most serious infectious diseases in children in developing countries, causing more than one million deaths annually (9). Measles is characterized by high fever, cough, coryza, conjunctivitis, and Koplik’s spots, followed by a maculopapular rash about 2 weeks after initial exposure (18). Immunization at an early age is necessary in countries with high levels of MV transmission and where MV infection is a serious and life-threatening disease. However, because of maternal antibodies and immaturity of the neonatal immune system, early immunization can result in low seroconversion rates, resulting in inadequate levels of immune protection (16). The World Health Organization has proposed a plan for the eradication of measles, but to achieve this goal, new alternative vaccination strategies and/or vaccines that are safe for young infants and not inhibited by maternal antibodies are needed (38). MV has a negative-sense RNA genome encoding six structural proteins. The hemagglutinin (H) and fusion (F) envelope glycoproteins and the nucleocapsid (N) protein surrounding the genome have been shown to be the most important proteins in terms of raising immunity against the virus (16). Both * Corresponding author. Mailing address: Microbiology and Tumor Biology Center, Box 280, Karolinska Institutet, S-171 77 Stockholm, Sweden. Phone: 46 8 728 63 22. Fax: 46 8 33 07 44. E-mail: ewa.bjorling @mtc.ki.se. 251

252

DE

CARVALHO NICACIO ET AL.

should induce neutralizing antibodies to the H and F proteins in addition to stimulating the cellular immune response. The preparation of combinatorial libraries from variable heavy- and light-chain antibody genes provides an efficient route for the isolation of human antibody Fab fragments (Fabs). Using antigen binding as a means of selection, Fab molecules of interest can be rescued from such libraries. The construction of antibody libraries on the surface of M13 phages has been described (2, 21), as well as their application for the generation of a large range of human MAbs against a variety of viruses (1, 4, 8, 10, 20, 32, 40). However, only three studies have generated recombinant human Fab molecules against MV (3, 6, 24), and none of them were able to neutralize the virus. Studies on the biological functions of generated Fabs have the potential to allow detailed analysis of human antibody responses, such as epitope specificities and neutralization mechanisms. In addition, identification of protective B-cell epitopes of specific viral proteins is vital for the development of rationally designed vaccines. This study describes the generation of three neutralizing and two nonneutralizing MV-specific human antibody Fab molecules by panning selection of combinatorial Fab libraries against viral proteins directly adsorbed or adsorbed with MVspecific MAbs onto microtiter surfaces. Additionally, studies of their biological properties and neutralization mechanism are described and discussed. MATERIALS AND METHODS Antibody phage display libraries. Three cDNA libraries generated from RNAs derived from three different donors (A, B, and C) were utilized in the present study. Two libraries were prepared from bone marrow RNAs recovered from two donors (A and B) possessing serum antibody titers of 1:400 against lysates of MV-infected Vero cells. In addition, serum immunoglobulin G (IgG) from these patients was demonstrated to specifically immunoprecipitate viral proteins from MV-infected-cell lysates. The RNA source of the third library originated from splenic lymphocytes from a splenectomized patient (donor C) possessing a serum antibody titer of 1:19,000 against MV measured in a standard diagnostic enzyme-linked immunosorbent assay (ELISA) performed at the Swedish Institute for Infectious Disease Control (SIIDC), Stockholm, Sweden. The same library has previously been used for isolation of Fab molecules specific for Puumala virus (10, 29). The IgG1 heavy-chain Fd regions and whole ␬ and ␭ light chains from the three libraries (A, B, and C) were expressed in the phage display vector pComb3H essentially as described earlier (8, 26, 29). Selection of antigen-binding clones by panning. Two different strategies were used for panning of the libraries. Libraries A and B were panned against MV antigen applied directly to microtiter plates, while library C was panned against MV antigen captured by MV-specific MAbs; both strategies were performed essentially as described before (10). Briefly, for libraries A and B, microtiter wells (Nunc, Roskilde, Denmark) were coated overnight at 4°C with 50 ␮l of detergent-treated MV lysate antigen prepared from approximately 107 MV-infected Vero cells which were lysed in 1 ml of phosphate-buffered saline (PBS) (pH 7.0) containing 1% NP-40 and 20 ␮g of phenylmethylsulfonyl fluoride (PMSF) ml⫺1 and diluted 35-fold in PBS. Library A was panned on two different occasions. On the second panning occasion, the MV lysate antigen was masked with a Fab (MT14) generated at the first panning occasion; 50 ␮g of purified Fab ml⫺1 was incubated for 1 h on the lysate-covered microtiter wells, followed by panning of phage resuspended in the presence of 50 ␮g of purified MT14 Fab ml⫺1. Thereafter the panning was performed as described below for library C. For library C, microtiter wells were coated overnight at 4°C with ascites containing MAb I-29, 7-AG11, 16-CD11, or 16-DE6, specific for the MV H protein, and MAb 19-FF10 or 19-GD6, specific for the MV F protein, at a dilution of 1:50 in 50 mM sodium carbonate buffer (pH 9.6). Wells were washed six times with PBS containing 0.05% Tween 20 (PBS-T) and blocked with PBS containing 3% bovine serum albumin (BSA) for 1 h at 37°C. Detergent-treated MV antigen used for panning of library C was prepared as follows. MV purified by sucrose gradient centrifugation (SIIDC) was diluted in detergent buffer (10 mM Tris-HCl [pH 7.4], 150

J. VIROL. mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 1% sodium dodecyl sulfate, 3 mM PMSF, 1% aprotinin) and freeze-thawed two times. The antigen preparation was diluted to a concentration of 20 ␮g ml⫺1 in PBS-T containing 0.5% BSA (dilution buffer) and incubated at 4°C overnight at 100 ␮l per well. To eliminate nonspecific binding, 100 ␮l of phage library resuspended in PBS containing 1% BSA was preadsorbed to wells coated with only MAbs. Phages were then incubated with antigen for 2 h at 37°C. Unbound phages were removed by vigorous washing with PBS-T after each round of panning. Bound phages, bearing antigen-binding surface Fabs, were eluted with HCl-glycine, pH 2.2. The eluted phages were amplified by infection of Escherichia coli XL1-Blue (Stratagene, La Jolla, Calif.) cultures followed by superinfection with helper phage VCS M13 (1012 PFU) (Stratagene). In all, four or five rounds of panning were performed. After the last round of panning, phagemid DNA was extracted from the overnight culture and the vector was modified by excision of the coat protein III-encoding gene fragment of the phage (8). Propagation and induction of Fab production were performed essentially as previously described (8), and the bacterial supernatants were selected by ELISA against sucrose gradient-purified MV (SIIDC) or MV lysate antigen for the presence of MV-specific Fabs. Fab production and purification. Clones positive for MV binding were prepared from isopropyl-␤-D-thiogalactopyranoside (IPTG)-induced 500-ml cultures and purified by affinity chromatography with goat anti-human IgG F(ab⬘)2 (Pierce, Rockford, Ill.) coupled to Affi-Gel 10 (Bio-Rad, Richmond, Calif.) as previously described (10, 30). The purity of Fab preparations was determined by reducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by Coomassie brilliant blue staining. Fab concentrations were determined by ELISA, including a standard (Fab fragment of human IgG; Nordic Immunology, Tilburg, The Netherlands), as described before (10). ELISA. Screening of crude Fab preparations, determination of Fab concentrations, confirmation of Fab antigenic specificities, and competition studies were performed by ELISA. For all assays, the following reagents were used. Antigens or antibodies were diluted in 50 mM sodium carbonate buffer (pH 9.6), applied to microtiter plates, and left overnight at 4°C. Coated plates were blocked with 3% BSA in PBS. All reagents were diluted in dilution buffer, and the plates were washed six times in PBS-T between each step. All incubations were carried out at 37°C for 1 h. The assays were developed with either o-phenylenediamine dihydrochloride (Sigma, St. Louis, Mo.) or p-nitrophenyl phosphate (Sigma) in diethanolamine buffer. Absorbances were measured at 490 or 405 nm, respectively. (i) ELISA for determination of MV specificity. Microtiter wells were coated with sucrose gradient-purified MV (SIIDC) at 2.5 ␮g ml⫺1 or with MV lysate antigen prepared as described above. After blocking, undiluted crude Fab supernatants or diluted purified Fabs were added. Bound Fabs were detected by horseradish peroxidase-labeled goat anti-human IgG F(ab⬘)2 antibody (Pierce) diluted 1:1,000, followed by o-phenylenediamine dihydrochloride. (ii) Cross-reactivity with unrelated proteins. The antigenic specificity of affinity-purified Fab fragments was evaluated by analyses of the reactivities to tetanus toxoid (SBL Vaccin, Stockholm, Sweden), ovalbumin (Worthington Biochemical Corp., Lakewood, N.J.), and BSA. The antigens (10 ␮g ml⫺1) were applied directly to microtiter plates, and the ELISA was performed as described above. (iii) Competitive binding assays. Twenty well-characterized MV H, F, and N protein-specific mouse MAbs were used in competitive binding assays to determine which of their epitopes were blocked by the Fabs generated in this study. Preparation and characterization of the MAbs have been described previously (23, 33, 34, 35). Seven H-specific (I-12, I-29, I-41, I-44, 7-AG11, 16-CD11, and 16-DE6), eight F-specific (16-AG5, 16-AE7, 16-DC9, 16-EE8, 19-FF4, 19-FF10, 19-GD6, and 19-HB4), and five N-specific (16-AC5, 16-CF7, R1B5, R2C2, and R2C3) MAbs were utilized. All MAbs were first titrated against sucrose gradientpurified MV (SIIDC) in ELISA, with the titer being defined as that dilution giving approximately half-maximal binding (optical density of ⬇1.3 at 405 nm). In each competition assay, Fabs were incubated in an MV-coated well at 20 ␮g ml⫺1, after which the MAb was added at a final concentration equal to its titer against MV. Bound MAbs were detected by alkaline phosphatase (ALP)-labeled goat anti-mouse IgG antibody (Sigma) diluted 1:1,000, followed by p-nitrophenyl phosphate (Sigma) in diethanolamine buffer. Cross-competition binding experiments including biotin-labeled and unlabeled Fabs were performed essentially as for MAbs. Biotinylation of Fabs was performed as previously described (19). Inhibiting Fabs (20 ␮g ml⫺1) were followed by labeled Fabs at a pretitrated dilution. Binding of biotin-labeled Fab was detected with Extravidin-ALP (Sigma) diluted 1:5,000 and developed as described above. Blocking was assessed and compared to that for controls (Puumala virus-reactive Fab or diluent only) applied to wells before addition of MAb or biotin-labeled Fab.

VOL. 76, 2002 Western blotting. The antigenic reactivities of Fab fragments against denatured antigen were evaluated by Western blotting. Detergent-treated, sucrose gradient-purified MV antigen (SIIDC) was mixed with Laemmli sample buffer, electrophoresed in a Tris-HCl–4 to 20% polyacrylamide gel (Bio-Rad), and transferred electrophoretically to a nitrocellulose membrane. Nitrocellulose strips blocked with 5% nonfat milk in PBS were incubated overnight at 4°C with Fabs (20 ␮g ml⫺1), MAbs, or sera diluted 1:200. After washing, an ALP–goat anti-human IgG F(ab⬘)2 (Pierce) or ALP–goat anti-mouse IgG antibody (Sigma) was added at a dilution of 1:1,000 in 5% nonfat milk in PBS. Strips were developed, after additional washings, with 5-bromo-4-chloro-3-indolylphosphate–nitroblue tetrazolium (Sigma). All MV-specific MAbs used in this study were tested and included as controls in addition to MV-antibody positive serum. Virus neutralization assay. The ability of Fabs to inhibit MV infectivity was assayed by the plaque reduction neutralization test (PRNT). Fabs were serially diluted in medium (Dulbecco’s minimal essential medium [DMEM] supplemented with 1% fetal calf serum [FCS], 2 mM L-glutamine, 50 IU of penicillin ml⫺1, and 50 ␮g of streptomycin ml⫺1) and mixed with an equal volume of MV (strain LEC-KI, Edmonston, or Schwarz [34]) at 1,000 PFU ml⫺1. The mixture was incubated under rotation for 1 h at 37°C and subsequently inoculated (200 ␮l) into 12-well tissue culture plates containing confluent Vero cell monolayers. After adsorption for 1 h at 37°C, wells were overlaid with a 42°C 1:1 (vol/vol) mixture of 1% low-melting-point agarose and DMEM supplemented with 6% FCS, 20 mM HEPES, 2 mM L-glutamine, and antibiotics as described above. Tissue culture plates were incubated at 37°C in a humidified 5% CO2 atmosphere for 7 days. For plaque detection, the agarose was gently removed and cells were fixed and stained with 0.15% crystal violet in 20% ethanol, followed by air drying. Plaques were enumerated, and the neutralization activity was expressed as the amount of Fab necessary to reduce the number of plaques by 80% compared to wells incubated with MV and medium only. As a negative control a Puumala virus-specific Fab was included, and neutralization assays were repeated at least twice with reproducible results. Radioimmunoprecipitation of 35S-labeled MV proteins with recombinant Fabs. For radioimmunoprecipitation assays (RIPA), a confluent monolayer of Vero cells (225 cm2) was infected with MV LEC-KI strain. The infected cells were incubated at 37°C with DMEM supplemented with 1% FCS, 2 mM Lglutamine, and antibiotics as described above until a 50 to 60% cytopathic effect was reached. At that time the cell culture medium was replaced with cysteineand methionine-free DMEM supplemented with 0.5% FCS, 2 mM L-glutamine, 60 ␮g of cysteine ml⫺1, and antibiotics as described above. Cells were incubated for 1 h at 37°C, after which they were radiolabeled with 1 mCi of [35S]methionine ([35S]Met) (Redivue; Amersham, Little Chalfont, United Kingdom). After 24 h, PMSF (2 mM) and aprotinin (1%; Sigma) were added to the cell flask, and the cells were harvested and pelleted. Cells were resuspended in 4 ml of RIPA buffer (10 mM Tris-HCl [pH 7.8], 150 mM NaCl, 0.6 M KCl, 5 mM EDTA, 2% Triton X-100, 1 mM PMSF, and 1% aprotinin), lysed by freeze-thawing five times, and centrifuged at 60,000 ⫻ g for 45 min in a Beckman SV60 rotor. A mixture of 150 ␮l of radiolabeled MV antigen, 20 ␮g of Fab in 100 ␮l of Tris-HCl buffer (10 mM Tris-HCl [pH 7.4], and 150 mM NaCl), and 250 ␮l of RIPA buffer was incubated at 4°C overnight. Precipitations of Fabs were performed by addition of 100 ␮l of protein G-coupled agarose beads (50% slurry) (Amersham Pharmacia Biotech, Uppsala, Sweden) with 10 ␮g of goat anti-human IgG F(ab⬘)2 (Pierce) ml⫺1 and incubation with rotation at 4°C for 4 h. For MAbs and control serum, only protein G-agarose beads were added. The beads were pelleted and washed six times with RIPA buffer and once with Tris-HCl buffer. Samples were resuspended in 30 ␮l of Laemmli sample buffer and boiled for 5 min, and after a brief centrifugation the samples were electrophoresed on a Tris-HCl–4 to 20% polyacrylamide gel (Bio-Rad). The gel was fixed in a solution of 20% methanol and 10% acetic acid for 30 min, washed twice with distilled deionized H2O, and then incubated in Amplify solution (Amersham) for 30 min. After drying, the gel was visualized by autoradiography on a phosphorimager. Nucleic acid sequencing. Cycle sequencing was carried out on phagemid DNA using the PRISM Ready BigDye Terminator Cycle sequencing kit with AmpliTaq DNA polymerase FS (Perkin-Elmer/Applied Biosystems Division, Foster City, Calif.). Amplification was performed using the GeneAmp PCR system 2400 (Perkin-Elmer, Norwalk, Conn.). The reaction mix (20 ␮l) contained 8 ␮l of PRISM premix, 3.2 pmol of sequencing primer, and 500 ng of DNA. The PCR products were analyzed using an ABI377 sequencer (Perkin-Elmer/Applied Biosystems Division). The following primers were used: Pelseq, ACC TAT TGC CTA CGG CAG CCG; Seqgz, GAA GTA GTC CTT GAC CAG; Seqlb, GAA GTC ACT TAT GAG ACA CAC; and Seqkb, ATA GAA GTT GTT CAG CAG GCA.

HUMAN ANTIBODIES NEUTRALIZING MEASLES VIRUS

253

TABLE 1. Amino acid sequences of heavy-chain complementaritydetermining region 3 (CDR3) and flanking framework regions FR3 and FR4 of generated MV-specific Fab fragments Fab

FR3

CDR3

FR4

MV4 MV12 MT14 MT64 GL29

YYCAR YFCAR YYCAT YFCAR YYCAGR

HHLLYPNVDYGMDV DISWDSSGYSDAFDI DRTFRTLKYMSV YRGLTMVGRVLKRGGLDI KVYQPHLRHYSYFMDV

WGQGTTVTVSS WGQGTMVTVSS WGEGTTVTVSS WGQGTDVTVSS WGKGTTVTVSS

RESULTS Isolation of MV-specific Fab fragments by panning. Three IgG1 ␭ and ␬ antibody libraries, A, B, and C, expressed on the surface of filamentous phage and derived from bone marrow or spleen lymphocyte RNAs from three MV-seropositive donors, were used for selection of MV-specific Fabs. Panning was performed on five different occasions, twice with library A, once with library B, and twice with library C. The second panning of library A was performed by first masking the panning antigen with the purified Fab generated in the first panning of the same library. The input phage library consisted of 8 ⫻109 PFU ml⫺1 for panning of library A, 5 ⫻109 PFU ml⫺1 for panning of library B, and 7.3 ⫻109 and 4.5 ⫻109 PFU ml⫺1, respectively, for the two panning occasions for library C. Libraries A and B were panned against viral proteins applied directly to microtiter wells, while library C was panned against viral proteins immobilized in microtiter wells with MV H and F protein-specific MAbs. Four or five rounds of panning resulted in 105 to 107 PFU of eluted phage ml⫺1 in all panning experiments. Phagemid DNA was prepared from the fourth or fifth round of panning, and the gene III fragment was removed by NheI and SpeI enzyme digestion followed by religation. The resulting phagemid was transformed into XL1-Blue E. coli cells to produce clones secreting soluble Fab fragments. To screen for antigen-specific Fab clones, 20 colonies each from the first and second pannings of library A, 30 colonies from the panning of library B, and 80 and 120 colonies from the first and second pannings of library C, respectively, were randomly picked, and freeze-thawed culture supernatants were screened for reactivity against MV in ELISA. Five of 20, 1 of 20, 2 of 30, 2 of 80, and 79 of 120 crude Fab bacterial lysates from the five different panning occasions reacted positively with MV antigen. Twenty-eight of the 79 ELISA-positive crude Fab suspensions from the second panning of library C were additionally screened by PRNT, where they showed 52 to 82% plaque reduction at a 1:2 dilution (data not shown). Five positive clones from the first and one positive clone from the second panning of library A, two positive clones from panning of library B, and two positive clones from the first panning of library C and eight of the clones with highest neutralizing capacity in PRNT from the second panning of library C were sequenced. Nucleic acid sequencing of Fab heavy- and light-chain variable regions. To evaluate whether the isolated Fab fragments represent individual clones, complementarity-determining regions 1 to 3 of the heavy and light chains were sequenced. This analysis revealed five predominant sequences with different heavy-chain sequences (Table 1) and light-chain sequences

254

DE


FIG. 1. Titration by ELISA of purified Fab fragments MV4 (f), MV12 (F), MT14 (}), MT64 (Œ), and GL29 (E) against sucrosepurified MV antigen.

(data not shown): MT14 and MT64 from the first and second pannings, respectively, of library A; GL29 from the panning of library B; and MV4 and MV12 from the first and second pannings, respectively, of library C. Five Fabs, MV4, MV12, MT14, MT64, and GL29, were treated as unique clones, which were amplified and affinity purified for further characterization. Fab specificity for MV. Fab clones MV4, MV12, MT14, MT64, and GL29 were tested for MV specificity in different assays. Affinity-purified Fabs from each of the five clones were tested for MV binding by titration of equivalent amounts of each Fab against sucrose gradient-purified MV. All clones bound to MV in a concentration-dependent manner but at various levels, suggesting different affinities for the antigen. Fabs MV12 and MT14 showed the highest and Fabs MV4 and MT64 showed the lowest reactivity (Fig. 1). Antigen specificity of the Fabs was verified by reactivity at the background level (optical density of ⬍0.1) in ELISA against tetanus toxoid, ovalbumin, and BSA (data not shown). Western blotting. A Western blot assay was applied to determine the nature of the epitopes recognized by the Fabs. Two of the clones, MT64 and GL29, visualized protein bands against denatured MV antigen. This protein, corresponding to MV N, was the same size (⬇60 kDa) as that seen with MV N protein-specific MAbs and MV antibody-positive serum, which were included as controls (data not shown). This indicated that these two Fabs were N protein specific and most likely recognized linear epitopes within this protein. RIPA with Fabs. Fabs were run in a RIPA to assess the MV protein specificity. Four of the isolated Fab clones were found to precipitate MV-specific proteins, identifying two different protein specificities (summarized in Table 2): those reactive with N protein and those reactive with H protein. The RIPA reactivities of Fabs MV12 and MT14 suggested that they were

J. VIROL.

specific to H in that they precipitated a band of ⬇80 kDa from MV-infected Vero cell lysates labeled with [35S]Met (Fig. 2). A band of the same size was also precipitated by the MV Hspecific MAb (I-41). Fabs MT64 and GL29 immunoprecipitated an ⬇60-kDa band corresponding to N (Fig. 2). The same band was also seen when an MV N-specific MAb (16-CF7) was used. Fab MV4 did not immunoprecipitate any clearly identifiable MV proteins with this method. Epitope mapping of Fabs. Cross-competition studies between Fabs and biotin-labeled Fabs were performed to investigate the presence of potentially shared epitopes among the different Fabs generated. All of the Fabs were blocked by the homologous Fab by 80% or more of maximal binding (Table 3). The two H-specific Fabs, MV12 and MT14, blocked each other’s binding to MV as efficiently as they blocked homologous binding, indicating that they had overlapping or shared epitopes. MV12 also blocked MV4 but to a much lower extent. In addition, Fab MV4 partially blocked GL29 binding. Fab competition studies against 20 MV-specific MAbs were performed to determine whether the Fabs generated in this study shared epitopes or parts of epitopes with these earliercharacterized MAbs (33, 34, 35). Four of the Fabs were able to inhibit MAb binding to MV to various degrees (Fig. 3). Fabs MV12 and MT14, which precipitated the H protein in RIPA, blocked anti-H MAb (I-41 and 16-DE6) binding (Fig. 3a and b). These MAbs were earlier mapped to the same antigenic group (group III) but to two different epitopes (33). In addition, these two Fabs inhibited binding of MAb I-44 (antigenic group IV) but to a much lower extent. Binding of MAbs I-12 and 7-AG11 (antigenic groups I and I⫹II) to MV was also inhibited by MV12 and MT14, respectively. Furthermore, MV12 and MT14 inhibited (⬍30%) the F-specific MAbs 19FF4 and 19-FF10, respectively, which are grouped in the same antigenic group within the anti-F MAb panel. Fabs MT64 and GL29 both blocked the binding of N-specific MAbs 16-AC5 and 16-CF7 (Fig. 3c and d). In addition, Fab MT64 blocked MAb R1B5, and Fab GL29 blocked MAbs R2C2 and R2C3. Fab MV4 did not inhibit binding of any MAb in the competition assays (data not shown). Virus neutralization by Fabs. Affinity-purified recombinant Fabs MV4, MV12, MT14, MT64, and GL29 were tested for their capacity to neutralize MV strains LEC-KI (laboratory strain), Edmonston (laboratory strain), and Schwarz (vaccine strain) by PRNT. Virus neutralization was set as the end point concentration of the Fab required to reduce the number of plaques by 80%. Three of five characterized Fabs had neutralizing activity against MV. Fab clones MV12 and MT14 neu-

TABLE 2. Antigen specificities of MV-specific Fabs Fab

MV4 MV12 MT14 MT64 GL29 a b

Competitive blocking NeutraliWestern Antigen RIPA of MV MAbsb zation blot specificity result a result H protein F protein N protein of MV

⫺ ⫺ ⫺ ⫹ ⫹

⫺ H H N N

0/7 4/8 4/7 0/7 0/7

0/8 1/8 1/8 0/8 0/8

0/5 0/5 0/5 3/5 4/5

⫹ ⫹ ⫹ ⫺ ⫺

Under reducing conditions. Number of MAbs blocked by Fab/total number of tested MAbs.

F or H? H H N N

VOL. 76, 2002


255

FIG. 2. Radioimmunoprecipitation of MV strain LEC-KI, illustrating the specificities of anti-MV recombinant Fabs. Lanes: 1, Fab MV4; 2, MV12; 3, MT14; 4, MT64; 5, GL29; 6, Puumala virus-specific Fab; 7, Tris-HCl buffer; 8, anti-H MAb I-41; 9, anti-N MAb 16-CF7; 10, anti-F MAb 16-AG5; 11, MV antibody-positive serum; 12, MV antibody-negative serum. The migration of molecular weight standards (in thousands) is indicated on the left.

tralized the MV LEC-KI strain at ⬇7 ␮g ml⫺1, while Fab MV4 neutralized the same strain at a concentration of below ⬇2 ␮g ml⫺1 (Fig. 4). When the neutralization cutoff was reduced to 50%, less than ⬇0.5 ␮g of Fabs MV4 and MT14 ml⫺1 was needed for neutralization, while ⬇1.5 ␮g of Fab MV12 ml⫺1 was needed (Fig. 4). Neutralization of the Edmonston strain was seen with Fab concentrations similar to those for the LEC-KI strain. Somewhat higher concentrations of Fabs MV4, MV12, and MT14 were needed to neutralize the Schwarz strain (Table 4). DISCUSSION The use of combinatorial libraries displayed on the surface of filamentous phage offers an efficient means of obtaining a diverse set of human MAbs from an immune donor (7). In this study we have used combinatorial phage antibody libraries generated from RNAs derived from bone marrow or splenic lymphocytes, which were selected against MV for the isolation of neutralizing H-specific and nonneutralizing N-specific recombinant human Fab molecules. Numerous murine MAbs specific for all of the structural proteins of MV have been isolated to date by using conventional hybridoma technology (15, 23, 31, 34). These MAbs are widely used as research tools and reagents for diagnosis. However, human MAbs are scarce, and only three earlier studies have generated human Fab molecules specific for MV proteins from recombinant antibody libraries. In the first of these studies, Bender et al. isolated three MV-specific Fabs, two of which were cross-reactive with a number of other antigens, from a library originating from bone marrow RNA from an individual with low antibody titers against MV (3). In the two remaining studies, an antibody library was prepared from the brain tissue of a patient with subacute sclerosing panencephalitis (SSPE) and selected either against MV-infected cells yielding one Nspecific and three phosphoprotein-specific Fab molecules (6) or against SSPE brain tissue sections, generating N-specific Fabs (24). Neutralizing antibodies are directed predominately against the surface glycoproteins of a virus, while intracellular proteins usually are more important for stimulation of the cellular immunity. However, isolating recombinant Fab molecules, from phage display libraries, directed against surface glycoproteins by selection against cellular extracts of virus-infected cells can be technically challenging. These selection procedures usually result in the isolation of antibodies specific for immunodominant proteins present in large amounts, such as the nucleocap-

sid protein. In this study two strategies were employed for selection of the libraries. Two libraries were selected against viral proteins directly applied to microtiter surfaces, while the third was selected against viral proteins captured on microtiter surfaces with MV MAbs specific for the H and F glycoproteins. The first method of selection resulted in the isolation of one H-specific MV neutralizing Fab (MT14) and two N-specific Fabs (MT64 and GL29). The second method, however, which was successfully used earlier to isolate neutralizing Fab molecules specific for Puumala virus glycoprotein 2 (10), resulted in the isolation of two MV-neutralizing Fabs; one (MV12) was H specific, while the specificity of the other (MV4) could not be determined. The second method seems to be the preferred method for the selection of antibodies reactive against proteins that otherwise are concealed by other, more abundant proteins. The specificities of the Fabs were determined by RIPA, where both H and N proteins were precipitated. However, in Western blotting under denaturing conditions only the N-specific Fabs generated bands, indicating that these Fabs most likely recognized linear sites, while the others are probably directed to conformational epitopes within the H protein. Fabs were tested in competition experiments against MV-specific MAbs to try to map their specificities. The MAb inhibition patterns of the two H-specific Fabs, MV12 and MT14, indicated that they were directed to the same (or to overlapping) epitopes as those of MAbs I-41 and 16-DE6, which they were able to block. These two MAbs have been mapped to the same antigenic group (III) but to two different epitopes in earlier competition studies (33, 34). In molecular characterization studies with I-41 and 16-DE6 MAb escape mutants (17), amino acid 552 in the H protein was of critical importance for the antigenic integrity of the I-41 epitope. In the same study, the epitope of MAb 16-DE6 was localized to amino acids 211, 388,

TABLE 3. Cross-competition between Fabs for MV binding Faba

MV4 MV12 MT14 MT64 GL29 a b c

% Inhibition of bindingb with blocking Fab: MV4

MV12

MT14

MT64

GL29

81 —c — — 34

34 100 96 — —

— 100 98 — —

— — — 85 —

— — — — 100

Biotin labeled. Boldface values indicate homologous inhibition. —, ⬍20% inhibition of binding.

256

DE


J. VIROL.

FIG. 3. Competitive binding data showing the ability of four recombinant Fabs to inhibit the binding of MAbs to MV. (a) Fab MV12; (b) Fab MT14; (c) Fab MT64; (d) Fab GL29. The 20 MV-specific MAbs are grouped according to their MV antigen specificity, i.e., anti-H, -F, and -N. The blocking is expressed as percent inhibition by the Fab compared to absorbance in wells without competing Fab.

and 532 to 533, indicating that this MAb has a conformational epitope. Together these data suggest that the epitopes of the Fabs generated in this study are located within these areas of the H protein, but further studies are needed to confirm this finding. It has not been possible to show by MAb competition studies to which epitopes the two N-specific Fabs are directed, as the MAbs used have not been previously mapped to specific epitopes within the N protein. Of the five isolated Fab clones, three conferred neutralizing activity against MV in PRNT. Two of three (MV12 and MT14) were shown by RIPA to be H specific, while the third Fab (MV4) was interpreted to be H or F specific, as these are the proteins responsible for inducing neutralizing antibodies. The fact that Fab MV4 could not precipitate any protein by RIPA using [35S]Met-labeled antigens may indicate that this Fab has low affinity for its target protein or that MV4 is specific for the F2 part of the F protein, which is not detectable with this method as it has a very low methionine content. The low MV reactivity of Fab MV4 as seen by ELISA titration could depend

on small amounts of F2 and/or F in the MV antigen preparation used. In addition, an F-specific MAb (19-FF10) was used for the selection of Fab MV4, adding support to this conclusion. The lowest Fab concentrations were needed for neutralization of the MV LEC-KI strain, which may have been expected since the Fabs were selected against this strain. Somewhat higher concentrations were needed for neutralizing the Edmonston and Schwarz strains. Differences in the amount of Fab needed for neutralization between strains could possibly reflect sequence differences between the H proteins of these three strains. While LEC-KI and Edmonston are laboratory strains, from an SSPE patient and a measles patient, respectively, the Schwarz strain is a vaccine strain. Although MV is a relatively conserved virus with low sequence diversity between strains, some differences occur, which could be of importance for neutralization. Neutralizing Fab fragments have been isolated for a number of viruses. Here we describe the first isolation of a number of human neutralizing Fab fragments against MV. A possible

VOL. 76, 2002


257

FIG. 4. Neutralization of MV by purified Fab fragments MV4 (f), MV12 (F), MT14 (}), MT64 (Œ), and GL29 (E) in PRNT using MV strain LEC-KI.

mechanism by which these Fab fragments neutralize is by coating of the virus surface leading to inhibition of attachment or interference with virus-cell fusion, which has been proposed to be a major mechanism of neutralization (25). It may be interesting to study this in more detail. The corresponding whole antibodies may inhibit MV in vivo by invoking additional, Fcdependent, mechanisms such as complement activation and antibody-dependent cellular cytotoxicity. Cloned antibodies from immune donors are helpful tools in the study of humoral responses. The analysis of antigen and epitope specificities of recombinant human antibodies during different infections can generate information about which proteins, and specifically which parts of proteins, are important for the generation of humoral immune responses. Particularly interesting are differences in antigen recognition in special forms and/or complications of a disease, such as atypical measles and SSPE. This approach was applied by Burgoon et al. when identifying the specificity of cloned antibodies from an SSPE brain (6). In the same line of thought, recombinant antibodies could be used to study differences in immunity induced by natural infection or vaccination. This information could be important when trying to circumvent the maternal antibodies present in the infant at the time for vaccination. In addition, Fab molecules could be reconstituted to whole immunoglobulins and be used for prophylaxis and/or immunotherapy in

individuals in whom an MV infection could lead to serious complications. Phage display antibody libraries could also be utilized in the search for antigens related to diseases with unknown etiology. In conclusion, this is the first report that describes the isolation of MV-specific neutralizing human Fab fragments. These antibodies, which were directed against the H (MV12 and MT14) and possibly the F (MV4) proteins of MV, could block the binding of MAbs to MV proteins (MV12 and MT14) and were able to neutralize heterologous MV strains in PRNT (MV4, MV12, and MT14). Resolving the epitope specificities of these antibodies within their target proteins could aid

TABLE 4. Fab neutralization of MV Fab

MV4 MV12 MT14 MT64 GL29 a

Concn (␮g ml⫺1) resulting in ⱖ80% plaque reduction of MV strain: LEC-KI

Edmonston

Schwarz

ⱕ0.625 ⱕ2.5 ⱕ0.625 —a —

ⱕ0.625 ⬎10 ⱕ0.625 — —

ⱖ2.5 ⬎10 ⱖ2.5 — —

—, no neutralization.

258

DE


in defining which antigenic determinants are crucial components in conferring immunity to MV infection in future vaccinees. ACKNOWLEDGMENTS This work was supported by grants from the Goljes Foundation, the Sven and Dagmar Sale´n Foundation, the National Institutes of Health (grant no. AI35165 to D.R.B.), the Swedish Medical Research Council (grant no. K2000-06P-13420-01A), and the Swedish Fund for Research without Animal Experiments (grant no. 8/99). REFERENCES 1. Barbas, C. F., III, J. E. Crowe, Jr., D. Cababa, T. M. Jones, S. L. Zebedee, B. R. Murphy, R. M. Chanock, and D. R. Burton. 1992. Human monoclonal Fab fragments derived from a combinatorial library bind to respiratory syncytial virus F glycoprotein and neutralize infectivity. Proc. Natl. Acad. Sci. USA 89:10164–10168. 2. Barbas, C. F., III, A. S. Kang, R. A. Lerner, and S. J. Benkovic. 1991. Assembly of combinatorial antibody libraries on phage surfaces: the gene III site. Proc. Natl. Acad. Sci. USA 88:7978–7982. 3. Bender, E., G. R. Pilkington, and D. R. Burton. 1994. Human monoclonal Fab fragments from a combinatorial library prepared from an individual with a low serum titer to a virus. Hum. Antib. Hybridom. 5:3–8. 4. Bjo ¨rling, E., E. von Garrelts, A. Mo ¨rner, M. Ehnlund, and M. A. A. Persson. 1999. Human neutralizing human immunodeficiency virus type 2-specific Fab molecules generated by phage display. J. Gen. Virol. 80:1987–1993. 5. Brinckmann, U. G., B. Bankamp, A. Reich, V. ter Meulen, and U. G. Liebert. 1991. Efficacy of individual measles virus structural proteins in the protection of rats from measles encephalitis. J. Gen. Virol. 72:2491–2500. 6. Burgoon, M. P., R. A. Williamson, G. P. Owens, O. Ghausi, R. B. Bastidas, D. R. Burton, and D. H. Gilden. 1999. Cloning the antibody response in humans with inflammatory CNS disease: isolation of measles virus-specific antibodies from phage display libraries of a subacute sclerosing panencephalitis brain. J. Neuroimmunol. 94:204–211. 7. Burton, D. R., and C. F. Barbas III. 1994. Human antibodies from combinatorial libraries. Adv. Immunol. 57:191–280. 8. Burton, D. R., C. F. Barbas III, M. A. A. Persson, S. Koenig, R. M. Chanock, and R. A. Lerner. 1991. A large array of human monoclonal antibodies to type 1 human immunodeficiency virus from combinatorial libraries of asymptomatic seropositive individuals. Proc. Natl. Acad. Sci. USA 88:10134–10137. 9. Clements, C. J., and F. T. Cutts. 1995. The epidemiology of measles: thirty years of vaccination. Curr. Top. Microbiol. Immunol. 191:13–33. 10. de Carvalho Nicacio, C., Å. Lundkvist, K. Brus Sjo ¨lander, A. Plyusnin, E.-M. Salonen, and E. Bjo ¨rling. 2000. A neutralizing recombinant human antibody Fab fragment against Puumala hantavirus. J. Med. Virol. 60:446–454. 11. Drillien, R., D. Spehner, A. Kirn, P. Giraudon, R. Buckland, F. Wild, and J. P. Lecocq. 1988. Protection of mice from fatal measles encephalitis by vaccination with vaccinia virus recombinants encoding either the hemagglutinin or the fusion protein. Proc. Natl. Acad. Sci. USA 85:1252–1256. 12. Fooks, A. R., D. Jeevarajah, A. Warnes, G. W. Wilkinson, and J. C. Clegg. 1996. Immunization of mice with plasmid DNA expressing the measles virus nucleoprotein gene. Viral Immunol. 9:65–71. 13. Fooks, A. R., E. Schadeck, U. G. Liebert, A. B. Dowsett, B. K. Rima, M. Steward, J. R. Stephenson, and G. W. Wilkinson. 1995. High-level expression of the measles virus nucleocapsid protein by using a replication-deficient adenovirus vector: induction of an MHC-1-restricted CTL response and protection in a murine model. Virology 210:456–465. 14. Giraudon, P., and T. F. Wild. 1985. Correlation between epitopes on hemagglutinin of measles virus and biological activities: passive protection by monoclonal antibodies is related to their hemagglutination inhibiting activity. Virology 144:46–58. 15. Giraudon, P., and T. F. Wild. 1981. Monoclonal antibodies against measles virus. J. Gen. Virol. 54:325–332. 16. Griffin, D. E. 1995. Immune responses during measles virus infection. Curr. Top. Microbiol. Immunol. 191:117–134. 17. Hu, A., H. Sheshberadaran, E. Norrby, and J. Ko¨vamees. 1993. Molecular characterization of epitopes on the measles virus hemagglutinin protein. Virology 192:351–354. 18. Katz, M. 1995. Clinical spectrum of measles. Curr. Top. Microbiol. Immunol. 191:1–12. 19. Lundkvist, Å., J. Ho ¨rling, L. Athlin, A. Rosen, and B. Niklasson. 1993. Neutralizing human monoclonal antibodies against Puumala virus, causative

J. VIROL.

20. 21. 22. 23. 24.

25. 26. 27. 28. 29.

30. 31. 32.

33. 34.

35.

36.

37. 38. 39. 40.

41.

agent of nephropathia epidemica: a novel method using antigen-coated magnetic beads for specific B cell isolation. J. Gen. Virol. 74:1303–1310. Maruyama, T., P. W. H. I. Parren, A. Sanchez, I. Rensink, L. L. Rodriguez, A. S. Khan, C. J. Peters, and D. R. Burton. 1999. Recombinant human monoclonal antibodies to Ebola virus. J. Infect. Dis. 179:S235–239. McCafferty, J., A. D. Griffiths, G. Winter, and D. J. Chiswell. 1990. Phage antibodies: filamentous phage displaying antibody variable domains. Nature 348:552–554. Norrby, E. 1995. The paradigms of measles vaccinology. Curr. Top. Microbiol. Immunol. 191:167–180. Norrby, E., S.-N. Chen, T. Togashi, H. Sheshberadaran, and K. P. Johnson. 1982. Five measles virus antigens demonstrated by use of mouse hybridoma antibodies in productively infected tissue culture cells. Arch. Virol. 71:1–11. Owens, G. P., R. A. Williamson, M. P. Burgoon, O. Ghausi, D. R. Burton, and D. H. Gilden. 2000. Cloning the antibody response in humans with chronic inflammatory disease: immunopanning of subacute sclerosing panencephalitis (SSPE) brain sections with antibody phage libraries prepared from SSPE brain enriches for antibody recognizing measles virus antigens in situ. J. Virol. 74:1533–1537. Parren, P. W. H. I., and D. R. Burton. 2001. The antiviral activity of antibodies in vitro and in vivo. Adv. Immunol. 77:195–262. Persson, M. A. A., R. H. Caothien, and D. R. Burton. 1991. Generation of diverse high-affinity human monoclonal antibodies by repertoire cloning. Proc. Natl. Acad. Sci. USA 88:2432–2436. Rota, J. S., K. B. Hummel, P. A. Rota, and W. J. Bellini. 1992. Genetic variability of the glycoprotein genes of current wild-type measles isolates. Virology 188:135–142. Rota, J. S., Z. D. Wang, P. A. Rota, and W. J. Bellini. 1994. Comparison of sequences of the H, F, and N coding genes of measles virus vaccine strains. Virus Res. 31:317–330. Salonen, E.-M., P. W. H. I. Parren, Y. F. Graus, Å. Lundkvist, P. Fisicaro, O. Vapalahti, H. Kallio-Kokko, A. Vaheri, and D. R. Burton. 1998. Human recombinant Puumala virus antibodies: cross-reaction with other hantaviruses and use in diagnostics. J. Gen. Virol. 79:659–665. ¨ hman, P. Putkonen, E. Norrby, and M. A. A. Samuelsson, A., F. Chiodi, P. O Persson. 1995. Chimeric macaque/human Fab molecules neutralize simian immunodeficiency virus. Virology 207:495–502. Sato, T. A., A. Fukuda, and A. Sugiura. 1985. Characterization of major structural proteins of measles virus with monoclonal antibodies. J. Gen. Virol. 66:1397–1409. Schofield, D. J., J. Glamann, S. U. Emerson, and R. H. Purcell. 2000. Identification by phage display and characterization of two neutralizing chimpanzee monoclonal antibodies to the hepatitis E virus capsid protein. J. Virol. 74:5548–5555. Sheshberadaran, H., and E. Norrby. 1986. Characterization of epitopes on the measles virus hemagglutinin. Virology 152:58–65. Sheshberadaran, H., S. N. Chen, and E. Norrby. 1983. Monoclonal antibodies against five structural components of measles virus. I. Characterization of antigenic determinants on nine strains of measles virus. Virology 128:341– 353. Sheshberadaran, H., E. Norrby, and K. W. Rammohan. 1985. Monoclonal antibodies against five structural components of measles virus. II. Characterization of five cell lines persistently infected with measles virus. Arch. Virol. 83:251–268. van Binnendijk, R. S., M. C. Poelen, G. van Amerongen, P. de Vries, and A. D. Osterhaus. 1997. Protective immunity in macaques vaccinated with live attenuated, recombinant, and subunit measles vaccines in the presence of passively acquired antibodies. J. Infect. Dis. 175:524–532. Varsanyi, T. M., B. Morein, A. Love, and E. Norrby. 1987. Protection against lethal measles virus infection in mice by immune-stimulating complexes containing the hemagglutinin or fusion protein. J. Virol. 61:3896–3901. Wild, T. F. 1999. Measles vaccines, new developments and immunization strategies. Vaccine 17:1726–1729. Wild, T. F., A. Bernard, D. Spehner, and R. Drillien. 1992. Construction of vaccinia virus recombinants expressing several measles virus proteins and analysis of their efficacy in vaccination of mice. J. Gen. Virol. 73:359–367. Williamson, R. A., R. Burioni, P. P. Sanna, L. J. Partridge, C. F. Barbas III, and D. R. Burton. 1993. Human monoclonal antibodies against a plethora of viral pathogens from single combinatorial libraries. Proc. Natl. Acad. Sci. USA 90:4141–4145. (Erratum, 91:1193, 1994.) Ziegler, D., P. Fournier, G. A. Berbers, H. Steuer, K. H. Wiesmuller, B. Fleckenstein, F. Schneider, G. Jung, C. C. King, and C. P. Muller. 1996. Protection against measles virus encephalitis by monoclonal antibodies binding to a cystine loop domain of the H protein mimicked by peptides which are not recognized by maternal antibodies. J. Gen. Virol. 77:2479–2489.