- Email: [email protected]
Humanization of a chicken anti-IL-12 monoclonal antibody
Journal of Immunological Methods 295 (2004) 9 – 19 www.elsevier.com/locate/jim
Research paper
Humanization of a chicken anti-IL-12 monoclonal antibody Naoya Tsurushita*, Minha Park, Kanokwan Pakabunto, Kelly Ong, Anamarija Avdalovic, Helen Fu, Audrey Jia, Max Va´squez, Shankar Kumar Protein Design Labs, Inc., 34801 Campus Drive, Fremont, CA 94555, USA Received 9 July 2004; accepted 12 August 2004 Available online 28 September 2004
Abstract Chicken anti-IL-12 monoclonal antibodies were isolated by phage display using spleen cells from a chicken immunized with human and mouse IL-12 as a source for library construction. One of the chicken monoclonal antibodies, DD2, exhibited binding to both human and mouse IL-12 in the single-chain Fv form and also after conversion to chicken–human chimeric IgG1/E antibody. The chicken DD2 variable regions were humanized by transferring their CDRs and several framework amino acids onto human acceptor variable regions. In the VE, six chicken framework amino acids were identified to be important for the conformation of the CDR structure by computer modeling and therefore were retained in the humanized form; likewise, five chicken amino acids in the VH framework regions were retained in the humanized VH. The affinities of humanized DD2 IgG1/ E to human and mouse IL-12 measured by competitive binding were nearly identical to those of chicken–human chimeric DD2 IgG1/E. This work demonstrates that humanization of chicken monoclonal antibodies assisted by computer modeling is possible, leading to a new way to generate therapeutic humanized antibodies against antigens to which the rodent immune system may fail to efficiently raise high affinity antibodies. D 2004 Elsevier B.V. All rights reserved. Keywords: Antibody engineering; Phage display; Cross-reactive antibody; Tolerance
1. Introduction Monoclonal antibodies are one of the most important experimental tools in bioscience. Since the first report of hybridoma technology (Kohler and Abbreviations: CDR; complementarity determining region; scFv; single chain Fv; TMB; 3,3V,5,5V-tetramethylbenzidine. * Corresponding author. Tel.: +1 510 574 1684; fax: +1 510 574 1500. E-mail address: [email protected] (N. Tsurushita). 0022-1759/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2004.08.018
Milstein, 1975), a number of rodent, and particularly murine, monoclonal antibodies have been generated against a wide range of molecules. The utility of murine monoclonal antibodies has also been extended to the medical field; however, the strong immunogenicity of murine antibodies in humans hampered their application for therapeutic uses (Fagnani, 1994; Khazaeli et al., 1994; Kuus-Reichel et al., 1994). This problem was essentially solved by the invention of a computer-guided humanization technology that systematically enabled reduction, or even elimination,
10
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
of immunogenicity of murine monoclonal antibodies in humans while preserving biological activity (Tsurushita and Vasquez, 2004). Since the approval of daclizumab (ZenapaxR, humanized anti-CD25 IgG1 monoclonal antibody) for treatment of renal allograft rejection in 1997, a total of eight humanized monoclonal antibodies have been approved for therapeutic uses in the United States (Reichert and Pavlou, 2004). While murine monoclonal antibodies have proven their usefulness, it is not always possible to raise high affinity murine antibodies against certain antigens or epitopes that are highly conserved in mammals, as such antigens have little or no immunogenicity in mice due to tolerance (Rajewsky, 1996). To overcome this problem, several new approaches have been developed. Knockout mice lacking the gene encoding a target antigen have been successfully used for raising antibodies when the antigen is not strongly immunogenic in mice (Korth et al., 1997; Walter et al., 1997; Lunn et al., 2000); however, generation of knockout mice is not necessarily a simple task, and some genes are essential for survival of mice or for development of the immune system. The use of phage and yeast display libraries expressing antibody fragments on the surface represents an alternative method for bypassing immunization of mice (Winter et al., 1994; Hoogenboom et al., 1998; Feldhaus et al., 2003). Another approach is to use non-rodent animals for immunization. Hybridoma technology that uses rabbit or chicken B cells for fusion has been described (Spieker-Polet et al., 1995; Nishinaka et al., 1996). However, these methods are not as robust in efficiently generating monoclonal antibodies as the well-established hybridoma method with rodents. Although the isolation of non-rodent antibodies by phage display has been reported with rabbit (Ridder et al., 1995; Spieker-Polet et al., 1995; Rader et al., 2000), sheep (Charlton et al., 2000; Li et al., 2000), cow (O’Brien et al., 1999), llama (Tanha et al., 2002) and chicken (Carnemolla et al., 1996; Yamanaka et al., 1996; Andris-Widhopf et al., 2000), their humanization would be required for use as human therapeutics. The use of chickens for generating monoclonal antibodies is an attractive alternative to standard mouse hybridoma technology because the chicken is evolutionally distant from mammals and therefore antigens highly conserved in mammals may trigger strong immune responses in chickens (Gassmann et
al., 1990; Camenisch et al., 1999). With proper immunization protocols and selection procedures, it may even be possible to isolate chicken monoclonal antibodies that bind to the same antigen from multiple mammals. Chicken antibodies that cross-react with antigens of human and other mammals may be tested both in preclinical studies using relevant animal models and in human clinical studies to assess their therapeutic potential. In order to use chicken monoclonal antibodies as human therapeutics, however, it is essential to minimize their immunogenicity in humans. This paper describes the isolation of chicken monoclonal antibodies recognizing both human and mouse IL-12 by phage display. One of the chicken anti-human and mouse IL-12 monoclonal antibodies was successfully humanized without losing its affinity to either of the antigens.
2. Materials and methods 2.1. Cell culture Mouse myeloma cell line Sp2/0-Ag14 (referred to as Sp2/0 in this text) was obtained from American Type Culture Collection (Manassas, VA) and maintained in DME medium containing 10% FBS (HyClone, Logan, UT) at 37 8C in a 7.5% CO2 incubator. 2.2. Phage display library The M13 phage display vector pNT3206 (Fig. 1) was constructed by replacing the human Cn gene in pScUAGDcp3 (Akamatsu et al., 1993) with the human CE1 gene. In pNT3206, single-chain Fv (scFv) fragments are produced as a fusion to human CE1. Construction of phage display libraries was carried out essentially as described previously (Akamatsu et al., 1993) using pNT3206 as a vector and mRNA isolated from chicken spleen cells as a source for cDNA synthesis. Chicken VH genes were amplified by PCR using a 5V primer CCAGCACCCATGGCCGCCGTGACGTTGGACGAGTCCG (NT561) and a 3V primer CGTCAAGCTAGCGGAGGAGACGATGACTTCGGTCCC (NT563). The NcoI site in NT561 and the NheI site in NT563 are underlined.
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
Fig. 1. Structure of the phage display vector pNT3206. (A) Schematic diagram of pNT3206. Symbols used: Amp, h-lactamase gene for ampicillin resistance; pUC ori, replication origin of pUC19; lacP, E. coli lac promoter; pelB, pelB leader peptide; linker, synthetic sequence encoding a short polypeptide linker to connect VH and VE; CE, constant region of human E light chain gene; TAG, amber termination codon; Dcp3, carboxyl terminal domain of M13 gene III minor coat protein. Arrows indicate direction of transcription. Relevant restriction enzyme sites are shown. The diagram is not drawn to scale. (B) Amino acid sequence surrounding the cloning sites for VH and VE. Amino acid sequence is shown in single letter code. A vertical arrow indicates the cleavage site of the leader peptide.
Chicken VE genes were amplified using a 5V primer CACGCAGAGCTCGCGCTGACTCAGCCG(TG)CCTC(GA)GT (NT562) and a 3V primer A G C C A C A G AT C T TA G G A C G G T C A G G G TTGTCCCG (NT564). The SstI site in NT562 and the BglII site in NT564 are underlined. PCR amplified fragments were digested with NcoI and NheI (for VH) or SstI and BglII (for VE), gel-purified, and ligated with correspondingly digested pNT3206 to construct VH and VE libraries. Plasmid DNA of the two libraries were then digested with EcoRI and NheI. The VH- and VE-containing fragments were gelpurified, ligated together, and electroporated into E. coli DH5a/FVIQ. The library size was 3.5107. Phage libraries were rescued by superinfection of VCSM13 helper phage, concentrated by precipitation with polyethylene glycol, and suspended in HEPESbuffered saline (pH 7.5). 2.3. ELISA MaxiSorp plates (Nalge Nunc, Rochester, NY) were coated with 0.1 Ag/ml of human IL-12 (R&D, Minneapolis, MN), mouse IL-12 (R&D), human globin (Sigma, St. Louis, MO), hen egg lysozyme
11
(Sigma) or concanavalin A (Sigma) in 0.2 M sodium carbonate–bicarbonate buffer (pH 9.4), and then blocked with SuperBlock blocking buffer (Pierce, Rockford, IL). After incubating with appropriately diluted test antibodies in ELISA buffer (PBS containing 1% BSA and 0.1% Tween 20), plates were washed with Washing Buffer (PBS containing 0.1% Tween 20) and bound antibodies were detected with HRPconjugated goat anti-human g chain antibodies (Jackson ImmunoResearch, West Grove, PA). Color development was performed with TMB substrate (Kirkegaard and Perry Laboratories, Gaithersburg, MD) as described by the supplier. Absorbance was read at 450 nm using a VERSAmax microplate reader (Molecular Devices, Menlo Park, CA). 2.4. Conversion of scFv to whole antibody The VE and VH genes of chicken anti-IL-12 scFv antibodies were converted to exons by the recombinant PCR method (Higuchi, 1990) as outlined in Fig. 2. At the first-step PCR, the signal peptide-coding region of the Vn (or VH) of the mouse anti-human CD33 monoclonal antibody M195 (Co et al., 1992) was amplified by PCR in such a way that the 5V end contained an MluI site and the 3V end was attached to a sequence homologous to the 5V end of the DD2 VE (or VH) coding region (left-side fragment). The VE (or VH) of the chicken scFv antibody DD2 was amplified by PCR in such a way that the 5V end was attached to a sequence homologous to the 3V end of the signal peptide-coding region of M195 Vn (or VH) while the 3V end carried a splicing donor signal and an XbaI site (right-side fragment). The left- and rightside fragments for each of DD2 VE and VH were combined and amplified by PCR to make a mini-exon flanked by MluI and XbaI sites. 2.5. Humanization Humanization of chicken antibody V regions was performed with the assistance of a molecular model generated by the algorithms of Levy et al. (1989) and Zilber et al. (1990). The human V region framework used as an acceptor for the CDRs of the chicken antiIL-12 monoclonal antibody DD2 was chosen based on sequence homology. Amino acid residues in the humanized V regions predicted from the three-dimen-
12
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
Fig. 2. Conversion of chicken V region cDNAs to exons. The coding region of the VE (or VH) of chicken anti-IL-12 DD2 monoclonal antibody was connected to an MluI site and the Vn (or VH) signal peptide-coding region of the mouse anti-CD33 monoclonal antibody M195 (Co et al., 1992) at the 5V end and to a splicing donor site and an XbaI site at the 3V end, as described in Materials and methods. The amplified VE and VH exons were digested with MluI and XbaI, and cloned into appropriate expression vectors.
sional model to be important for proper formation of the CDR structure were substituted with the corresponding residues of chicken DD2. 2.6. Expression of antibodies Isolation of Sp2/0 stable transfectants producing monoclonal antibodies was carried out by electroporation and selection in mycophenolic acid medium as described previously (Co et al., 1992). Expression of chimeric and humanized DD2 antibodies was measured by sandwich ELISA using goat anti-human g chain polyclonal antibodies (Biosource, Camarillo, CA) for coating and HRP-conjugated goat anti-human E polyclonal antibodies (Southern Biotechnology, Birmingham, AL) for detection of bound antibodies. 2.7. Purification of anti-IL-12 antibodies Sp2/0 stable transfectants expressing high levels of antibodies were adapted to growth in serum-free medium using Hybridoma SFM (Invitrogen, Carlsbad, CA) and grown to exhaustion in roller bottles. AntiIL-12 monoclonal antibodies were purified from
culture supernatants with a protein-A Sepharose column. The column was washed with PBS before the antibody was eluted with 0.1 M glycine–HCl (pH 2.8), 0.1 M NaCl. After neutralization with 1 M Tris– HCl (pH 8), the eluted protein was dialyzed against PBS and stored at 4 8C. Antibody concentration was determined by measuring absorbance at 280 nm (1 mg/ml=1.4 A280). SDS-PAGE in Tris-glycine buffer was performed according to standard procedures. 2.8. Affinity measurement by competitive binding MaxiSorp plates were coated with 0.1 Ag/ml human or mouse IL-12. A mixture of biotinylated chimeric DD2 (0.5 Ag/ml final concentration) and competitor antibody (chimeric or humanized DD2; starting at 0.5 mg/ml final concentration and serial 3fold dilutions) in 100 Al ELISA buffer was added to an IL-12-coated well in triplicate. ELISA plates were incubated at room temperature for 2 h. After washing the wells with Washing Buffer, 0.5 Ag/ml HRPconjugated streptavidin (Pierce) was added to each well. Color development was performed with TMB substrate.
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
13
3. Results 3.1. Isolation of chicken anti-IL-12 monoclonal antibodies A female white leghorn chicken was immunized at BAbCO (Richmond, CA), initially with 100 Ag of recombinant human IL-12 in complete Freund’s adjuvant, boosted at day 21 with 25 Ag of mouse IL-12 in incomplete Freund’s adjuvant, and further boosted at day 35 with 25 Ag of human IL-12 in incomplete Freund’s adjuvant. The spleen was harvested at day 40. Chicken VE and VH genes were amplified by PCR using the cDNA derived from the spleen cells and cloned into the phage display vector pNT3206 (Fig. 1) to make a scFv library as described in Materials and methods. Isolation of phage antibodies that bind to human IL12 was carried out by panning as described (Akamatsu et al., 1993). After three cycles of binding to and elution from human IL-12-coated microtiter plates, infection of E. coli TG1DrecA, and rescue by superinfection of VCSM13 helper phage, several phage clones were found to produce scFv antibodies that bind specifically to both human and mouse IL-12 (data not shown). One of the clones, DD2, was chosen for further analysis. To express DD2 in the form of whole antibody, each of the VE and VH genes in the phage display vector was converted to an exon structure containing a signal peptide and a splicing donor site as outlined in Fig. 2. After digestion with MluI and XbaI, the VE and VH genes were cloned into the corresponding sites of the mammalian expression vectors pVE2 (Hinton et al., 2004) and pVg1 (Co et al., 1992), respectively (Fig. 3). The resultant plasmids were named pVE2-DD2 and pVg1-DD2. The amino acid sequences of mature DD2 VE and VH regions are shown in Fig. 4. To obtain cell lines stably producing chimeric DD2 IgG1/E monoclonal antibodies, mouse myeloma cell line Sp2/0 was cotransfected with pVE2-DD2 and pVg1-DD2 after linearization with FspI as described in Materials and methods. One of the high-producing cell lines for chimeric DD2 was adapted to growth in serum-free medium (Hybridoma SFM; Invitrogen) and expanded in roller bottles. Chimeric DD2 antibody was purified from spent
Fig. 3. Structure of antibody expression vectors. The mammalian expression vectors pVE2 and pVg1 for production of human E2 light and E1 heavy chains, respectively, were described previously (Co et al., 1992; Hinton et al., 2004). After digestion with MluI and XbaI, the VE and VH exon fragments were cloned between the corresponding sites of pVE2 and pVg1, respectively. Symbols used: CMV-P, human cytomegalovirus immediate early promoter; (A), polyadenylation signal; SV40-P, SV40 early promoter; ori, replication origin of pBR322; Amp, h-lactamase gene; gpt, E. coli gpt gene; dhfr, mouse dhfr gene; CE, constant region of human E2 gene; CH1, H, CH2 and CH3, constant regions of human E1 gene. Arrows show direction of transcription. Locations of relevant enzyme sites are indicated. The diagram is not drawn to scale.
culture supernatant with a protein-A Sepharose column. SDS-PAGE analysis indicated that the purity of chimeric DD2 was more than 95% (data not shown). Chimeric DD2 showed specific binding to both human and mouse IL-12 (Fig. 5). 3.2. Design of humanized DD2 variable regions For humanization of chicken DD2 variable regions, a molecular model of the DD2 variable regions (Fig. 6) was first constructed as described in Materials and methods. Next, based on a homology search against human germline V and J segment sequences, the VE segment DPL16 (Williams et al., 1996) and the J segment JE2 (Udey and Blomberg, 1987) were selected to provide the frameworks for the DD2 VE region. For the DD2 VH region, the germline VH segment DP-54 (Tomlinson et al., 1992) and the J segment JH1 (Ravetch et al., 1981) were used. The identity of the framework amino acids between the chicken DD2 VE region and the acceptor human DPL16 and JE2 segments was 68%, while the identity
14
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
Fig. 4. Alignment of the V region amino acid sequences. Amino acid sequences of the VE (top) and VH (bottom) regions of chicken DD2 (DD2), humanized DD2 (HuDD2), and the human acceptor germline V and J segments are shown in single letter code. The CDR sequences based on the definition of Kabat et al. (1991) are underlined in the chicken DD2 VE and VH sequences. The CDR sequences in the acceptor human V segments are omitted in the figure. Asterisks indicate gaps in the alignment. Note that an amino acid at position 10 is missing in both human and chicken VE sequences. The single underlined amino acids in the humanized VE and VH sequences were predicted to contact the CDR and therefore were substituted with the corresponding chicken residues. Numbers written vertically show amino acid positions according to Kabat et al. (1991). The location of an extra amino acid in framework 2 of chicken VE is designated 39A.
between the chicken DD2 VH and the human DP-54 and JH1 segments was 71%. Chicken VE regions contain two amino acid deletions and one amino acid insertion compared to human VE sequences (Fig. 4). The N-terminal amino acid of mature chicken VE corresponds to the third amino acid from the N-terminus of mature human VE. Framework 2 of chicken VE contains an extra amino acid at position 39A. Although the first two amino acids at the N-terminus of mature human VE exist in close proximity to CDR amino acids, detailed
examination of the molecular model suggested that the transfer of two serine residues from the DPL16 VE segment to the N-terminus of DD2 VE would not drastically change the CDR structure. A serine residue at position 39A in the chicken DD2 VE is located in the loop opposite and away from the antigen-binding site (Fig. 6). Examination of the model suggested that removal of serine at position 39A during humanization was unlikely to introduce a significant change in CDR or framework structure. Therefore, in the humanized DD2 VE sequence, the two N-terminal
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
15
fragments were digested with MluI and XbaI, and subcloned into mammalian expression vectors, pVE2 and pVg1 (Fig. 3), respectively. The resultant plasmids were designated pVE2-HuDD2 and pVg1-HuDD2. Sp2/0 stable transfectants producing humanized DD2 were obtained by cotransfection of VE2-HuDD2 and pVg1-HuDD2 by electroporation after linearization of the vectors with FspI as described in Materials and methods. One of the high-producing Sp2/0 transfectants was adapted to growth in Hybridoma SFM and expanded in roller bottles. Humanized DD2 antibody was purified with a protein-A Sepharose column. The purity measured by SDS-PAGE was more than 95% (data not shown). 3.4. Binding properties of humanized DD2
Fig. 5. Binding of humanized and chimeric DD2 antibodies to various proteins. Binding of humanized and chimeric DD2 to human IL-12, mouse IL-12, human globin, chicken lysozyme, and concanavalin A was analyzed by ELISA as described in Materials and methods.
amino acids of the acceptor human DPL16 segment were added and the serine residue at position 39A in the chicken DD2 VE was deleted. At framework positions for which examination of the computer model suggested significant contact with the CDRs, the amino acids from the chicken V regions were substituted for the original human framework amino acids. This was performed at positions 36, 46, 57, 60, 66 and 69 of the light chain (Fig. 4). Similarly, for the heavy chain, replacements were made at positions 28, 49, 67, 78 and 93 (Fig. 4). The alignment of the amino acid sequences of the chicken DD2, humanized DD2, and human acceptor germline V and J segments are shown for both light and heavy chain variable regions in Fig. 4.
Binding of chimeric and humanized DD2 to human or mouse IL-12 was examined by ELISA. Fig. 5 shows that humanized DD2 at 1 Ag/ml bound to human IL-12 as well as chimeric DD2 did. In addition, humanized DD2 bound to mouse IL-12 at the same level as chimeric DD2 did. Similar results were obtained with different antibody concentrations in binding to either human or mouse IL-12 (data not
3.3. Expression of humanized DD2 The humanized DD2 VEand VH genes, designed as a mini-exon including a signal peptide, a splice donor signal, and appropriate restriction enzyme sites, were constructed by extension of eight overlapping synthetic oligonucleotides and PCR amplification as described (He et al., 1998). The resulting V gene
Fig. 6. Three-dimensional model of the chicken DD2 variable region. The backbone model of the structure of the variable region of the chicken anti-IL12 monoclonal antibody DD2 is shown. Color scheme: red, CDR residues; magenta, a serine residue at position 39A in the VE region; green, the N-terminal residue of the VE which corresponds to the third residue from the N-terminus of the humanized DD2 VE.
16
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
shown). As shown in Fig. 5, both humanized and chimeric DD2 antibodies exhibited little binding to three control antigens (human globin, hen egg lysozyme, and concanavalin A), suggesting that the
binding specificity of chimeric DD2 was retained in the humanized form. The affinities of chimeric and humanized DD2 to human and mouse IL-12 were compared by competition ELISA as described in Materials and methods. Representative results are shown in Fig. 7. Both chimeric and humanized DD2 competed with biotinylated chimeric DD2 for binding to human and mouse IL-12 in a concentration-dependent manner. In each of the five independent experiments, chimeric and humanized DD2 showed very similar competition patterns in binding to human IL-12. The average IC50 values for chimeric and humanized DD2, obtained using the computer software Prism (GraphPad Software, San Diego, CA), were 2.4 and 2.5 Ag/ml, respectively. The average IC50 value of three independent experiments testing the binding to mouse IL12 was 2.4 Ag/ml for both chimeric and humanized DD2. These results clearly indicate that the affinity to human and mouse IL-12 was fully retained during the process of humanization of the chicken monoclonal antibody DD2.
4. Discussion
Fig. 7. Comparison of the affinity to human IL-12 (A) and mouse IL-12 (B) between humanized and chimeric DD2 antibodies by competition ELISA. Binding of biotinylated chimeric DD2 to human or mouse IL-12 was analyzed in the presence of increasing amounts of competitor as described in Materials and methods. As a control, human IgG1/E monoclonal antibody OST577 (Ehrlich et al., 1992), that recognizes hepatitis B virus surface antigen, was used. A representative result of five (A) or three (B) independent experiments is shown.
The advantage of using chickens for generating monoclonal antibodies is that the chicken immune system can react strongly against antigens conserved in mammals, which in turn may be poor immunogens in most, if not all, mammalian species. Chickens can also produce antibodies that bind to a target antigen of human origin as well as other species (e.g., nonhuman primates, mouse, rat, and rabbit), which may be used for disease models. Such antibodies may be useful for both preclinical studies with model animals and clinical studies with human patients. Moreover, chickens may generate monoclonal antibodies that block multiple, functionally related proteins present in humans. In this paper, we reported the isolation of chicken monoclonal scFv antibodies that specifically recognize both human and mouse IL-12. We also reported the successful humanization of a chicken monoclonal antibody. Chicken immunoglobulin genes consist of three classes of heavy chain genes (A, g and a) and only one class of the light chain gene (E). The molecular mechanism for diversification of the antibody reper-
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
toire in chickens is distinct from that in mice and humans. In chickens, functional V genes are generated by V(D)J rearrangement using a single germline V segment (VE1 for VE genes and VH1 for VH genes), and gene conversion, instead of somatic mutation, plays a central role for affinity maturation (Reynaud et al., 1989). The framework amino acid sequences of the chicken genomic VE1 and VH1 segments have highest overall homologies to the human genomic VE segments of subclass III and VH segments of subclass III (Kabat et al., 1991), respectively. For humanization of chicken anti-IL-12 monoclonal antibody DD2, the human genomic V segments DPL16 of VE subgroup III (Williams et al., 1996) and DP-54 of the VH subgroup III (Tomlinson et al., 1992) were used. Since the heterogeneity of framework amino acid sequences of affinity-matured V genes is not as extensive in chickens as that observed in mice or humans, humanization of chicken monoclonal antibodies may be achieved with a small set of human variable region sequences as acceptors. This was indeed indicated when V region framework sequences of eight chicken monoclonal antibodies isolated in this work were aligned with human germline V segment sequences (data not shown). The human DPL16 VE segment was one of the most similar sequences to all of the eight chicken VE sequences. For the VH sequences, a small set of human germline VH segments that belong to subgroup III, including DP-47, DP-49 and DP-54 (Tomlinson et al., 1992), was identified to provide the most similar acceptor sequences. In addition, comparison of the chicken V region sequences with their most similar human counterparts revealed that certain framework amino acids that could potentially interact with the CDRs were almost always different between chickens and humans. These amino acids are located at positions 46, 66 and 69 in VE and 67, 78 and 93 in VH (cf. Fig. 4). Therefore, successfully humanized chicken monoclonal antibodies will most likely require chicken-derived residues at these six locations. Three-dimensional modeling and humanization of the VE domain of the chicken anti-IL-12 monoclonal antibody DD2 were particularly challenging because chicken VE regions, when compared to V regions of known structure and to human VE regions, contain two amino acid deletions at the N-terminus of the
17
mature protein and one amino acid insertion at position 39A (Fig. 4). The human genomic VE segments of subgroup III (Kabat et al., 1991), which have a high homology to chicken VE sequences, start with either Ser–Tyr or Ser–Ser at the N-terminus of the mature protein (Williams et al., 1996). While examination of the molecular model of the DD2 variable region suggested that the addition of two serine residues at the N-terminus of the humanized form would not drastically alter the CDR conformation, the addition of a tyrosine residue at the second position was predicted to potentially cause a structural change in the CDRs. The DPL16 VE segment, which has high similarity to the DD2 VE sequence and starts with Ser–Ser at the N-terminus, was therefore chosen as an acceptor sequence for humanization. In addition, examination of the molecular model suggested that removal of a serine residue at position 39A would not cause a drastic change in either the CDR or framework structure. Although a subtle change of the CDR structure in the humanized form could have resulted in reduction of the affinity, our data showed that no loss in the affinity to human and mouse IL-12 was detected for humanized DD2, suggesting that humanization technology based on computer-guided molecular modeling is reliable and can be successfully applied beyond antibodies of mammalian origin. The ultimate goal of antibody humanization is to eliminate potential immunogenicity of non-human antibodies, such as chicken antibodies, in human patients. In this work, a total of 11 chicken framework amino acids (six from VE and five from VH) were retained in humanized DD2. When another chicken anti-IL-12 monoclonal antibody was successfully humanized, a total of eight chicken framework residues were retained in the humanized form to preserve its antigen-binding affinity (data not shown). These numbers are not unusually high when compared with the humanization of murine monoclonal antibodies. In the case of humanizing the mouse antiTac monoclonal antibody to generate daclizumab, for example, a total of nine mouse-specific framework amino acids were retained to preserve CDR conformation (Queen et al., 1989). In clinical trials with daclizumab involving 535 patients, no severe adverse events due to immunogenicity were reported (Bumgardner et al., 2001). Therefore, even though humanized chicken antibodies may tend to retain
18
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
approximately 10 chicken framework amino acids, it seems unlikely that they will be highly immunogenic in humans. In this paper, we reported the first successful humanization of a non-mammalian monoclonal antibody with the aid of computer-guided molecular modeling. Since the structure of the variable regions of antibodies is relatively well conserved across species, our data implies that humanization of monoclonal antibodies of any species is achievable by the same general approach using molecular modeling. Chicken antibodies can recognize antigens and epitopes to which the rodent and other mammalian immune systems cannot respond effectively due to tolerance. In addition, since certain chicken antibodies can simultaneously recognize an antigen of human and murine (or other mammalian) origin, a single humanized chicken monoclonal antibody may be useful for both preclinical studies with disease animal models and clinical trials in patients. With advances in methodologies for more efficiently isolating desired chicken monoclonal antibodies, chickens should become an important source of potent humanized antibodies for treatment of human diseases. Acknowledgements The authors thank Catherine Huey, Brett Jorgensen, Linh Le, David Maciejewski, Thang Pham, and Raymond Ogawa for their technical assistance, J. Tso for valuable discussion throughout the work, Cary Queen for his support of the project, and Yoshiko Akamatsu, Paul Hinton, and David Powers for their comments on the manuscript. References Akamatsu, Y., Cole, M.S., Tso, J.Y., Tsurushita, N., 1993. Construction of a human Ig combinatorial library from genomic V segments and synthetic CDR3 fragments. J. Immunol. 151, 4651. Andris-Widhopf, J., Rader, C., Steinberger, P., Fuller, R., Barbas III, C.F., 2000. Methods for the generation of chicken monoclonal antibody fragments by phage display. J. Immunol. Methods 242, 159. Bumgardner, G.L., Hardie, I., Johnson, R.W., Lin, A., Nashan, B., Pescovitz, M.D., Ramos, E., Vincenti, F., 2001. Results of 3year phase III clinical trials with daclizumab prophylaxis for
prevention of acute rejection after renal transplantation. Transplantation 72, 839. Camenisch, G., Tini, M., Chilov, D., Kvietikova, I., Srinivas, V., Caro, J., Spielmann, P., Wenger, R.H., Gassmann, M., 1999. General applicability of chicken egg yolk antibodies: the performance of IgY immunoglobulins raised against the hypoxia-inducible factor 1alpha. FASEB J. 13, 81. Carnemolla, B., Neri, D., Castellani, P., Leprini, A., Neri, G., Pini, A., Winter, G., Zardi, L., 1996. Phage antibodies with panspecies recognition of the oncofoetal angiogenesis marker fibronectin ED-B domain. Int. J. Cancer 68, 397. Charlton, K.A., Moyle, S., Porter, A.J., Harris, W.J., 2000. Analysis of the diversity of a sheep antibody repertoire as revealed from a bacteriophage display library. J. Immunol. 164, 6221. Co, M.S., Avdalovic, N.M., Caron, P.C., Avdalovic, M.V., Scheinberg, D.A., Queen, C., 1992. Chimeric and humanized antibodies with specificity for the CD33 antigen. J. Immunol. 148, 1149. Ehrlich, P.H., Moustafa, Z.A., Justice, J.C., Harfeldt, K.E., Kelley, R.L., Ostberg, L., 1992. Characterization of human monoclonal antibodies directed against hepatitis B surface antigen. Hum. Antib. Hybrid. 3, 2. Fagnani, R., 1994. The immunogenicity of foreign monoclonal antibodies in human disease applications: problems and current approaches. Immunol. Ser. 61, 3. Feldhaus, M.J., Siegel, R.W., Opresko, L.K., Coleman, J.R., Feldhaus, J.M., Yeung, Y.A., Cochran, J.R., Heinzelman, P., Colby, D., Swers, J., Graff, C., Wiley, H.S., Wittrup, K.D., 2003. Flow-cytometric isolation of human antibodies from a nonimmune Saccharomyces cerevisiae surface display library. Nat. Biotechnol. 21, 163. Gassmann, M., Thommes, P., Weiser, T., Hubscher, U., 1990. Efficient production of chicken egg yolk antibodies against a conserved mammalian protein. FASEB J. 4, 2528. He, X.Y., Xu, Z., Melrose, J., Mullowney, A., Vasquez, M., Queen, C., Vexler, V., Klingbeil, C., Co, M.S., Berg, E.L., 1998. Humanization and pharmacokinetics of a monoclonal antibody with specificity for both E- and P-selectin. J. Immunol. 160, 1029. Higuchi, R., 1990. Recombinant PCR. In: Innis, M.A., Gelfand, D.H., Sninsky, J.J., White, T.J. (Eds.), PCR Protocols. Academic Press, San Diego, CA, p. 177. Hinton, P.R., Johlfs, M.G., Xiong, J.M., Hanestad, K., Ong, K.C., Bullock, C., Keller, S., Tang, M.T., Tso, J.Y., Vasquez, M., Tsurushita, N., 2004. Engineered human IgG antibodies with longer serum half-lives in primates. J. Biol. Chem. 279, 6213. Hoogenboom, H.R., de Bruine, A.P., Hufton, S.E., Hoet, R.M., Arends, J.W., Roovers, R.C., 1998. Antibody phage display technology and its applications. Immunotechnology 4, 1. Kabat, E.A., Wu, T.T., Perry, H.M., Gottesman, K.S., Foeller, C., 1991. Sequences of Proteins of Immunological Interest, vol. 1. U.S. Department of Health and Human Services, Bethesda, MD. Khazaeli, M.B., Conry, R.M., LoBuglio, A.F., 1994. Human immune response to monoclonal antibodies. J. Immunother. 15, 42. Kohler, G., Milstein, C., 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256, 495.
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19 Korth, C., Stierli, B., Streit, P., Moser, M., Schaller, O., Fischer, R., Schulz-Schaeffer, W., Kretzschmar, H., Raeber, A., Braun, U., Ehrensperger, F., Hornemann, S., Glockshuber, R., Riek, R., Billeter, M., Wuthrich, K., Oesch, B., 1997. Prion (PrPSc)specific epitope defined by a monoclonal antibody. Nature 390, 74. Kuus-Reichel, K., Grauer, L.S., Karavodin, L.M., Knott, C., Krusemeier, M., Kay, N.E., 1994. Will immunogenicity limit the use, efficacy, and future development of therapeutic monoclonal antibodies? Clin. Diagn. Lab. Immunol. 1, 365. Levy, R., Assulin, O., Scherf, T., Levitt, M., Anglister, J., 1989. Probing antibody diversity by 2D NMR: comparison of amino acid sequences, predicted structures, and observed antibody– antigen interactions in complexes of two antipeptide antibodies. Biochemistry 28, 7168. Li, Y., Kilpatrick, J., Whitelam, G.C., 2000. Sheep monoclonal antibody fragments generated using a phage display system. J. Immunol. Methods 236, 133. Lunn, M.P., Johnson, L.A., Fromholt, S.E., Itonori, S., Huang, J., Vyas, A.A., Hildreth, J.E., Griffin, J.W., Schnaar, R.L., Sheikh, K.A., 2000. High-affinity anti-ganglioside IgG antibodies raised in complex ganglioside knockout mice: reexamination of GD1a immunolocalization. J. Neurochem. 75, 404. Nishinaka, S., Akiba, H., Nakamura, M., Suzuki, K., Suzuki, T., Tsubokura, K., Horiuchi, H., Furusawa, S., Matsuda, H., 1996. Two chicken B cell lines resistant to ouabain for the production of chicken monoclonal antibodies. J. Vet. Med. Sci. 58, 1053. O’Brien, P.M., Aitken, R., O’Neil, B.W., Campo, M.S., 1999. Generation of native bovine mAbs by phage display. Proc. Natl. Acad. Sci. U. S. A. 96, 640. Queen, C., Schneider, W.P., Selick, H.E., Payne, P.W., Landolfi, N.F., Duncan, J.F., Avdalovic, N.M., Levitt, M., Junghans, R.P., Waldmann, T.A., 1989. A humanized antibody that binds to the interleukin 2 receptor. Proc. Natl. Acad. Sci. U. S. A. 86, 10029. Rader, C., Ritter, G., Nathan, S., Elia, M., Gout, I., Jungbluth, A.A., Cohen, L.S., Welt, S., Old, L.J., Barbas III, C.F., 2000. The rabbit antibody repertoire as a novel source for the generation of therapeutic human antibodies. J. Biol. Chem. 275, 13668. Rajewsky, K., 1996. Clonal selection and learning in the antibody system. Nature 381, 751. Ravetch, J.V., Siebenlist, U., Korsmeyer, S., Waldmann, T., Leder, P., 1981. Structure of the human immunoglobulin mu locus: characterization of embryonic and rearranged J and D genes. Cell 27, 583. Reichert, J., Pavlou, A., 2004. Monoclonal antibodies market. Nat. Rev., Drug Discov. 3, 383.
19
Reynaud, C.-A., Dahan, A., Anquez, V., Weill, J.-C., 1989. Development of the chicken antibody repertoire. In: Honjo, T., Alt, F.W., Rabbitts, T.H. (Eds.), Immunoglobulin Genes. Academic Press, San Diego, CA, p. 151. Ridder, R., Schmitz, R., Legay, F., Gram, H., 1995. Generation of rabbit monoclonal antibody fragments from a combinatorial phage display library and their production in the yeast Pichia pastoris. Biotechnology (NY) 13, 255. Spieker-Polet, H., Sethupathi, P., Yam, P.C., Knight, K.L., 1995. Rabbit monoclonal antibodies: generating a fusion partner to produce rabbit–rabbit hybridomas. Proc. Natl. Acad. Sci. U. S. A. 92, 9348. Tanha, J., Dubuc, G., Hirama, T., Narang, S.A., MacKenzie, C.R., 2002. Selection by phage display of llama conventional V(H) fragments with heavy chain antibody V(H)H properties. J. Immunol. Methods 263, 97. Tomlinson, I.M., Walter, G., Marks, J.D., Llewelyn, M.B., Winter, G., 1992. The repertoire of human germline VH sequences reveals about fifty groups of VH segments with different hypervariable loops. J. Mol. Biol. 227, 776. Tsurushita, N., Vasquez, M., 2004. Humanization of monoclonal antibodies. In: Honjo, T., Alt, F.W., Neuberger, M.S. (Eds.), Molecular Biology of B Cells. Elsevier Academic Press, San Diego, CA, p. 533. Udey, J.A., Blomberg, B., 1987. Human lambda light chain locus: organization and DNA sequences of three genomic J regions. Immunogenetics 25, 63. Walter, U.M., Ayer, L.M., Wolitzky, B.A., Wagner, D.D., Hynes, R.O., Manning, A.M., Issekutz, A.C., 1997. Characterization of a novel adhesion function blocking monoclonal antibody to rat/ mouse P-selectin generated in the P-selectin-deficient mouse. Hybridoma 16, 249. Williams, S.C., Frippiat, J.P., Tomlinson, I.M., Ignatovich, O., Lefranc, M.P., Winter, G., 1996. Sequence and evolution of the human germline V lambda repertoire. J. Mol. Biol. 264, 220. Winter, G., Griffiths, A.D., Hawkins , R.E., Hoogenboom, H.R., 1994. Making antibodies by phage display technology. Annu. Rev. Immunol. 12, 433. Yamanaka, H.I., Inoue, T., Ikeda-Tanaka, O., 1996. Chicken monoclonal antibody isolated by a phage display system. J. Immunol. 157, 1156. Zilber, B., Scherf, T., Levitt, M., Anglister, J., 1990. NMR-derived model for a peptide–antibody complex. Biochemistry 29, 10032.
Research paper
Humanization of a chicken anti-IL-12 monoclonal antibody Naoya Tsurushita*, Minha Park, Kanokwan Pakabunto, Kelly Ong, Anamarija Avdalovic, Helen Fu, Audrey Jia, Max Va´squez, Shankar Kumar Protein Design Labs, Inc., 34801 Campus Drive, Fremont, CA 94555, USA Received 9 July 2004; accepted 12 August 2004 Available online 28 September 2004
Abstract Chicken anti-IL-12 monoclonal antibodies were isolated by phage display using spleen cells from a chicken immunized with human and mouse IL-12 as a source for library construction. One of the chicken monoclonal antibodies, DD2, exhibited binding to both human and mouse IL-12 in the single-chain Fv form and also after conversion to chicken–human chimeric IgG1/E antibody. The chicken DD2 variable regions were humanized by transferring their CDRs and several framework amino acids onto human acceptor variable regions. In the VE, six chicken framework amino acids were identified to be important for the conformation of the CDR structure by computer modeling and therefore were retained in the humanized form; likewise, five chicken amino acids in the VH framework regions were retained in the humanized VH. The affinities of humanized DD2 IgG1/ E to human and mouse IL-12 measured by competitive binding were nearly identical to those of chicken–human chimeric DD2 IgG1/E. This work demonstrates that humanization of chicken monoclonal antibodies assisted by computer modeling is possible, leading to a new way to generate therapeutic humanized antibodies against antigens to which the rodent immune system may fail to efficiently raise high affinity antibodies. D 2004 Elsevier B.V. All rights reserved. Keywords: Antibody engineering; Phage display; Cross-reactive antibody; Tolerance
1. Introduction Monoclonal antibodies are one of the most important experimental tools in bioscience. Since the first report of hybridoma technology (Kohler and Abbreviations: CDR; complementarity determining region; scFv; single chain Fv; TMB; 3,3V,5,5V-tetramethylbenzidine. * Corresponding author. Tel.: +1 510 574 1684; fax: +1 510 574 1500. E-mail address: [email protected] (N. Tsurushita). 0022-1759/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2004.08.018
Milstein, 1975), a number of rodent, and particularly murine, monoclonal antibodies have been generated against a wide range of molecules. The utility of murine monoclonal antibodies has also been extended to the medical field; however, the strong immunogenicity of murine antibodies in humans hampered their application for therapeutic uses (Fagnani, 1994; Khazaeli et al., 1994; Kuus-Reichel et al., 1994). This problem was essentially solved by the invention of a computer-guided humanization technology that systematically enabled reduction, or even elimination,
10
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
of immunogenicity of murine monoclonal antibodies in humans while preserving biological activity (Tsurushita and Vasquez, 2004). Since the approval of daclizumab (ZenapaxR, humanized anti-CD25 IgG1 monoclonal antibody) for treatment of renal allograft rejection in 1997, a total of eight humanized monoclonal antibodies have been approved for therapeutic uses in the United States (Reichert and Pavlou, 2004). While murine monoclonal antibodies have proven their usefulness, it is not always possible to raise high affinity murine antibodies against certain antigens or epitopes that are highly conserved in mammals, as such antigens have little or no immunogenicity in mice due to tolerance (Rajewsky, 1996). To overcome this problem, several new approaches have been developed. Knockout mice lacking the gene encoding a target antigen have been successfully used for raising antibodies when the antigen is not strongly immunogenic in mice (Korth et al., 1997; Walter et al., 1997; Lunn et al., 2000); however, generation of knockout mice is not necessarily a simple task, and some genes are essential for survival of mice or for development of the immune system. The use of phage and yeast display libraries expressing antibody fragments on the surface represents an alternative method for bypassing immunization of mice (Winter et al., 1994; Hoogenboom et al., 1998; Feldhaus et al., 2003). Another approach is to use non-rodent animals for immunization. Hybridoma technology that uses rabbit or chicken B cells for fusion has been described (Spieker-Polet et al., 1995; Nishinaka et al., 1996). However, these methods are not as robust in efficiently generating monoclonal antibodies as the well-established hybridoma method with rodents. Although the isolation of non-rodent antibodies by phage display has been reported with rabbit (Ridder et al., 1995; Spieker-Polet et al., 1995; Rader et al., 2000), sheep (Charlton et al., 2000; Li et al., 2000), cow (O’Brien et al., 1999), llama (Tanha et al., 2002) and chicken (Carnemolla et al., 1996; Yamanaka et al., 1996; Andris-Widhopf et al., 2000), their humanization would be required for use as human therapeutics. The use of chickens for generating monoclonal antibodies is an attractive alternative to standard mouse hybridoma technology because the chicken is evolutionally distant from mammals and therefore antigens highly conserved in mammals may trigger strong immune responses in chickens (Gassmann et
al., 1990; Camenisch et al., 1999). With proper immunization protocols and selection procedures, it may even be possible to isolate chicken monoclonal antibodies that bind to the same antigen from multiple mammals. Chicken antibodies that cross-react with antigens of human and other mammals may be tested both in preclinical studies using relevant animal models and in human clinical studies to assess their therapeutic potential. In order to use chicken monoclonal antibodies as human therapeutics, however, it is essential to minimize their immunogenicity in humans. This paper describes the isolation of chicken monoclonal antibodies recognizing both human and mouse IL-12 by phage display. One of the chicken anti-human and mouse IL-12 monoclonal antibodies was successfully humanized without losing its affinity to either of the antigens.
2. Materials and methods 2.1. Cell culture Mouse myeloma cell line Sp2/0-Ag14 (referred to as Sp2/0 in this text) was obtained from American Type Culture Collection (Manassas, VA) and maintained in DME medium containing 10% FBS (HyClone, Logan, UT) at 37 8C in a 7.5% CO2 incubator. 2.2. Phage display library The M13 phage display vector pNT3206 (Fig. 1) was constructed by replacing the human Cn gene in pScUAGDcp3 (Akamatsu et al., 1993) with the human CE1 gene. In pNT3206, single-chain Fv (scFv) fragments are produced as a fusion to human CE1. Construction of phage display libraries was carried out essentially as described previously (Akamatsu et al., 1993) using pNT3206 as a vector and mRNA isolated from chicken spleen cells as a source for cDNA synthesis. Chicken VH genes were amplified by PCR using a 5V primer CCAGCACCCATGGCCGCCGTGACGTTGGACGAGTCCG (NT561) and a 3V primer CGTCAAGCTAGCGGAGGAGACGATGACTTCGGTCCC (NT563). The NcoI site in NT561 and the NheI site in NT563 are underlined.
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
Fig. 1. Structure of the phage display vector pNT3206. (A) Schematic diagram of pNT3206. Symbols used: Amp, h-lactamase gene for ampicillin resistance; pUC ori, replication origin of pUC19; lacP, E. coli lac promoter; pelB, pelB leader peptide; linker, synthetic sequence encoding a short polypeptide linker to connect VH and VE; CE, constant region of human E light chain gene; TAG, amber termination codon; Dcp3, carboxyl terminal domain of M13 gene III minor coat protein. Arrows indicate direction of transcription. Relevant restriction enzyme sites are shown. The diagram is not drawn to scale. (B) Amino acid sequence surrounding the cloning sites for VH and VE. Amino acid sequence is shown in single letter code. A vertical arrow indicates the cleavage site of the leader peptide.
Chicken VE genes were amplified using a 5V primer CACGCAGAGCTCGCGCTGACTCAGCCG(TG)CCTC(GA)GT (NT562) and a 3V primer A G C C A C A G AT C T TA G G A C G G T C A G G G TTGTCCCG (NT564). The SstI site in NT562 and the BglII site in NT564 are underlined. PCR amplified fragments were digested with NcoI and NheI (for VH) or SstI and BglII (for VE), gel-purified, and ligated with correspondingly digested pNT3206 to construct VH and VE libraries. Plasmid DNA of the two libraries were then digested with EcoRI and NheI. The VH- and VE-containing fragments were gelpurified, ligated together, and electroporated into E. coli DH5a/FVIQ. The library size was 3.5107. Phage libraries were rescued by superinfection of VCSM13 helper phage, concentrated by precipitation with polyethylene glycol, and suspended in HEPESbuffered saline (pH 7.5). 2.3. ELISA MaxiSorp plates (Nalge Nunc, Rochester, NY) were coated with 0.1 Ag/ml of human IL-12 (R&D, Minneapolis, MN), mouse IL-12 (R&D), human globin (Sigma, St. Louis, MO), hen egg lysozyme
11
(Sigma) or concanavalin A (Sigma) in 0.2 M sodium carbonate–bicarbonate buffer (pH 9.4), and then blocked with SuperBlock blocking buffer (Pierce, Rockford, IL). After incubating with appropriately diluted test antibodies in ELISA buffer (PBS containing 1% BSA and 0.1% Tween 20), plates were washed with Washing Buffer (PBS containing 0.1% Tween 20) and bound antibodies were detected with HRPconjugated goat anti-human g chain antibodies (Jackson ImmunoResearch, West Grove, PA). Color development was performed with TMB substrate (Kirkegaard and Perry Laboratories, Gaithersburg, MD) as described by the supplier. Absorbance was read at 450 nm using a VERSAmax microplate reader (Molecular Devices, Menlo Park, CA). 2.4. Conversion of scFv to whole antibody The VE and VH genes of chicken anti-IL-12 scFv antibodies were converted to exons by the recombinant PCR method (Higuchi, 1990) as outlined in Fig. 2. At the first-step PCR, the signal peptide-coding region of the Vn (or VH) of the mouse anti-human CD33 monoclonal antibody M195 (Co et al., 1992) was amplified by PCR in such a way that the 5V end contained an MluI site and the 3V end was attached to a sequence homologous to the 5V end of the DD2 VE (or VH) coding region (left-side fragment). The VE (or VH) of the chicken scFv antibody DD2 was amplified by PCR in such a way that the 5V end was attached to a sequence homologous to the 3V end of the signal peptide-coding region of M195 Vn (or VH) while the 3V end carried a splicing donor signal and an XbaI site (right-side fragment). The left- and rightside fragments for each of DD2 VE and VH were combined and amplified by PCR to make a mini-exon flanked by MluI and XbaI sites. 2.5. Humanization Humanization of chicken antibody V regions was performed with the assistance of a molecular model generated by the algorithms of Levy et al. (1989) and Zilber et al. (1990). The human V region framework used as an acceptor for the CDRs of the chicken antiIL-12 monoclonal antibody DD2 was chosen based on sequence homology. Amino acid residues in the humanized V regions predicted from the three-dimen-
12
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
Fig. 2. Conversion of chicken V region cDNAs to exons. The coding region of the VE (or VH) of chicken anti-IL-12 DD2 monoclonal antibody was connected to an MluI site and the Vn (or VH) signal peptide-coding region of the mouse anti-CD33 monoclonal antibody M195 (Co et al., 1992) at the 5V end and to a splicing donor site and an XbaI site at the 3V end, as described in Materials and methods. The amplified VE and VH exons were digested with MluI and XbaI, and cloned into appropriate expression vectors.
sional model to be important for proper formation of the CDR structure were substituted with the corresponding residues of chicken DD2. 2.6. Expression of antibodies Isolation of Sp2/0 stable transfectants producing monoclonal antibodies was carried out by electroporation and selection in mycophenolic acid medium as described previously (Co et al., 1992). Expression of chimeric and humanized DD2 antibodies was measured by sandwich ELISA using goat anti-human g chain polyclonal antibodies (Biosource, Camarillo, CA) for coating and HRP-conjugated goat anti-human E polyclonal antibodies (Southern Biotechnology, Birmingham, AL) for detection of bound antibodies. 2.7. Purification of anti-IL-12 antibodies Sp2/0 stable transfectants expressing high levels of antibodies were adapted to growth in serum-free medium using Hybridoma SFM (Invitrogen, Carlsbad, CA) and grown to exhaustion in roller bottles. AntiIL-12 monoclonal antibodies were purified from
culture supernatants with a protein-A Sepharose column. The column was washed with PBS before the antibody was eluted with 0.1 M glycine–HCl (pH 2.8), 0.1 M NaCl. After neutralization with 1 M Tris– HCl (pH 8), the eluted protein was dialyzed against PBS and stored at 4 8C. Antibody concentration was determined by measuring absorbance at 280 nm (1 mg/ml=1.4 A280). SDS-PAGE in Tris-glycine buffer was performed according to standard procedures. 2.8. Affinity measurement by competitive binding MaxiSorp plates were coated with 0.1 Ag/ml human or mouse IL-12. A mixture of biotinylated chimeric DD2 (0.5 Ag/ml final concentration) and competitor antibody (chimeric or humanized DD2; starting at 0.5 mg/ml final concentration and serial 3fold dilutions) in 100 Al ELISA buffer was added to an IL-12-coated well in triplicate. ELISA plates were incubated at room temperature for 2 h. After washing the wells with Washing Buffer, 0.5 Ag/ml HRPconjugated streptavidin (Pierce) was added to each well. Color development was performed with TMB substrate.
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
13
3. Results 3.1. Isolation of chicken anti-IL-12 monoclonal antibodies A female white leghorn chicken was immunized at BAbCO (Richmond, CA), initially with 100 Ag of recombinant human IL-12 in complete Freund’s adjuvant, boosted at day 21 with 25 Ag of mouse IL-12 in incomplete Freund’s adjuvant, and further boosted at day 35 with 25 Ag of human IL-12 in incomplete Freund’s adjuvant. The spleen was harvested at day 40. Chicken VE and VH genes were amplified by PCR using the cDNA derived from the spleen cells and cloned into the phage display vector pNT3206 (Fig. 1) to make a scFv library as described in Materials and methods. Isolation of phage antibodies that bind to human IL12 was carried out by panning as described (Akamatsu et al., 1993). After three cycles of binding to and elution from human IL-12-coated microtiter plates, infection of E. coli TG1DrecA, and rescue by superinfection of VCSM13 helper phage, several phage clones were found to produce scFv antibodies that bind specifically to both human and mouse IL-12 (data not shown). One of the clones, DD2, was chosen for further analysis. To express DD2 in the form of whole antibody, each of the VE and VH genes in the phage display vector was converted to an exon structure containing a signal peptide and a splicing donor site as outlined in Fig. 2. After digestion with MluI and XbaI, the VE and VH genes were cloned into the corresponding sites of the mammalian expression vectors pVE2 (Hinton et al., 2004) and pVg1 (Co et al., 1992), respectively (Fig. 3). The resultant plasmids were named pVE2-DD2 and pVg1-DD2. The amino acid sequences of mature DD2 VE and VH regions are shown in Fig. 4. To obtain cell lines stably producing chimeric DD2 IgG1/E monoclonal antibodies, mouse myeloma cell line Sp2/0 was cotransfected with pVE2-DD2 and pVg1-DD2 after linearization with FspI as described in Materials and methods. One of the high-producing cell lines for chimeric DD2 was adapted to growth in serum-free medium (Hybridoma SFM; Invitrogen) and expanded in roller bottles. Chimeric DD2 antibody was purified from spent
Fig. 3. Structure of antibody expression vectors. The mammalian expression vectors pVE2 and pVg1 for production of human E2 light and E1 heavy chains, respectively, were described previously (Co et al., 1992; Hinton et al., 2004). After digestion with MluI and XbaI, the VE and VH exon fragments were cloned between the corresponding sites of pVE2 and pVg1, respectively. Symbols used: CMV-P, human cytomegalovirus immediate early promoter; (A), polyadenylation signal; SV40-P, SV40 early promoter; ori, replication origin of pBR322; Amp, h-lactamase gene; gpt, E. coli gpt gene; dhfr, mouse dhfr gene; CE, constant region of human E2 gene; CH1, H, CH2 and CH3, constant regions of human E1 gene. Arrows show direction of transcription. Locations of relevant enzyme sites are indicated. The diagram is not drawn to scale.
culture supernatant with a protein-A Sepharose column. SDS-PAGE analysis indicated that the purity of chimeric DD2 was more than 95% (data not shown). Chimeric DD2 showed specific binding to both human and mouse IL-12 (Fig. 5). 3.2. Design of humanized DD2 variable regions For humanization of chicken DD2 variable regions, a molecular model of the DD2 variable regions (Fig. 6) was first constructed as described in Materials and methods. Next, based on a homology search against human germline V and J segment sequences, the VE segment DPL16 (Williams et al., 1996) and the J segment JE2 (Udey and Blomberg, 1987) were selected to provide the frameworks for the DD2 VE region. For the DD2 VH region, the germline VH segment DP-54 (Tomlinson et al., 1992) and the J segment JH1 (Ravetch et al., 1981) were used. The identity of the framework amino acids between the chicken DD2 VE region and the acceptor human DPL16 and JE2 segments was 68%, while the identity
14
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
Fig. 4. Alignment of the V region amino acid sequences. Amino acid sequences of the VE (top) and VH (bottom) regions of chicken DD2 (DD2), humanized DD2 (HuDD2), and the human acceptor germline V and J segments are shown in single letter code. The CDR sequences based on the definition of Kabat et al. (1991) are underlined in the chicken DD2 VE and VH sequences. The CDR sequences in the acceptor human V segments are omitted in the figure. Asterisks indicate gaps in the alignment. Note that an amino acid at position 10 is missing in both human and chicken VE sequences. The single underlined amino acids in the humanized VE and VH sequences were predicted to contact the CDR and therefore were substituted with the corresponding chicken residues. Numbers written vertically show amino acid positions according to Kabat et al. (1991). The location of an extra amino acid in framework 2 of chicken VE is designated 39A.
between the chicken DD2 VH and the human DP-54 and JH1 segments was 71%. Chicken VE regions contain two amino acid deletions and one amino acid insertion compared to human VE sequences (Fig. 4). The N-terminal amino acid of mature chicken VE corresponds to the third amino acid from the N-terminus of mature human VE. Framework 2 of chicken VE contains an extra amino acid at position 39A. Although the first two amino acids at the N-terminus of mature human VE exist in close proximity to CDR amino acids, detailed
examination of the molecular model suggested that the transfer of two serine residues from the DPL16 VE segment to the N-terminus of DD2 VE would not drastically change the CDR structure. A serine residue at position 39A in the chicken DD2 VE is located in the loop opposite and away from the antigen-binding site (Fig. 6). Examination of the model suggested that removal of serine at position 39A during humanization was unlikely to introduce a significant change in CDR or framework structure. Therefore, in the humanized DD2 VE sequence, the two N-terminal
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
15
fragments were digested with MluI and XbaI, and subcloned into mammalian expression vectors, pVE2 and pVg1 (Fig. 3), respectively. The resultant plasmids were designated pVE2-HuDD2 and pVg1-HuDD2. Sp2/0 stable transfectants producing humanized DD2 were obtained by cotransfection of VE2-HuDD2 and pVg1-HuDD2 by electroporation after linearization of the vectors with FspI as described in Materials and methods. One of the high-producing Sp2/0 transfectants was adapted to growth in Hybridoma SFM and expanded in roller bottles. Humanized DD2 antibody was purified with a protein-A Sepharose column. The purity measured by SDS-PAGE was more than 95% (data not shown). 3.4. Binding properties of humanized DD2
Fig. 5. Binding of humanized and chimeric DD2 antibodies to various proteins. Binding of humanized and chimeric DD2 to human IL-12, mouse IL-12, human globin, chicken lysozyme, and concanavalin A was analyzed by ELISA as described in Materials and methods.
amino acids of the acceptor human DPL16 segment were added and the serine residue at position 39A in the chicken DD2 VE was deleted. At framework positions for which examination of the computer model suggested significant contact with the CDRs, the amino acids from the chicken V regions were substituted for the original human framework amino acids. This was performed at positions 36, 46, 57, 60, 66 and 69 of the light chain (Fig. 4). Similarly, for the heavy chain, replacements were made at positions 28, 49, 67, 78 and 93 (Fig. 4). The alignment of the amino acid sequences of the chicken DD2, humanized DD2, and human acceptor germline V and J segments are shown for both light and heavy chain variable regions in Fig. 4.
Binding of chimeric and humanized DD2 to human or mouse IL-12 was examined by ELISA. Fig. 5 shows that humanized DD2 at 1 Ag/ml bound to human IL-12 as well as chimeric DD2 did. In addition, humanized DD2 bound to mouse IL-12 at the same level as chimeric DD2 did. Similar results were obtained with different antibody concentrations in binding to either human or mouse IL-12 (data not
3.3. Expression of humanized DD2 The humanized DD2 VEand VH genes, designed as a mini-exon including a signal peptide, a splice donor signal, and appropriate restriction enzyme sites, were constructed by extension of eight overlapping synthetic oligonucleotides and PCR amplification as described (He et al., 1998). The resulting V gene
Fig. 6. Three-dimensional model of the chicken DD2 variable region. The backbone model of the structure of the variable region of the chicken anti-IL12 monoclonal antibody DD2 is shown. Color scheme: red, CDR residues; magenta, a serine residue at position 39A in the VE region; green, the N-terminal residue of the VE which corresponds to the third residue from the N-terminus of the humanized DD2 VE.
16
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
shown). As shown in Fig. 5, both humanized and chimeric DD2 antibodies exhibited little binding to three control antigens (human globin, hen egg lysozyme, and concanavalin A), suggesting that the
binding specificity of chimeric DD2 was retained in the humanized form. The affinities of chimeric and humanized DD2 to human and mouse IL-12 were compared by competition ELISA as described in Materials and methods. Representative results are shown in Fig. 7. Both chimeric and humanized DD2 competed with biotinylated chimeric DD2 for binding to human and mouse IL-12 in a concentration-dependent manner. In each of the five independent experiments, chimeric and humanized DD2 showed very similar competition patterns in binding to human IL-12. The average IC50 values for chimeric and humanized DD2, obtained using the computer software Prism (GraphPad Software, San Diego, CA), were 2.4 and 2.5 Ag/ml, respectively. The average IC50 value of three independent experiments testing the binding to mouse IL12 was 2.4 Ag/ml for both chimeric and humanized DD2. These results clearly indicate that the affinity to human and mouse IL-12 was fully retained during the process of humanization of the chicken monoclonal antibody DD2.
4. Discussion
Fig. 7. Comparison of the affinity to human IL-12 (A) and mouse IL-12 (B) between humanized and chimeric DD2 antibodies by competition ELISA. Binding of biotinylated chimeric DD2 to human or mouse IL-12 was analyzed in the presence of increasing amounts of competitor as described in Materials and methods. As a control, human IgG1/E monoclonal antibody OST577 (Ehrlich et al., 1992), that recognizes hepatitis B virus surface antigen, was used. A representative result of five (A) or three (B) independent experiments is shown.
The advantage of using chickens for generating monoclonal antibodies is that the chicken immune system can react strongly against antigens conserved in mammals, which in turn may be poor immunogens in most, if not all, mammalian species. Chickens can also produce antibodies that bind to a target antigen of human origin as well as other species (e.g., nonhuman primates, mouse, rat, and rabbit), which may be used for disease models. Such antibodies may be useful for both preclinical studies with model animals and clinical studies with human patients. Moreover, chickens may generate monoclonal antibodies that block multiple, functionally related proteins present in humans. In this paper, we reported the isolation of chicken monoclonal scFv antibodies that specifically recognize both human and mouse IL-12. We also reported the successful humanization of a chicken monoclonal antibody. Chicken immunoglobulin genes consist of three classes of heavy chain genes (A, g and a) and only one class of the light chain gene (E). The molecular mechanism for diversification of the antibody reper-
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
toire in chickens is distinct from that in mice and humans. In chickens, functional V genes are generated by V(D)J rearrangement using a single germline V segment (VE1 for VE genes and VH1 for VH genes), and gene conversion, instead of somatic mutation, plays a central role for affinity maturation (Reynaud et al., 1989). The framework amino acid sequences of the chicken genomic VE1 and VH1 segments have highest overall homologies to the human genomic VE segments of subclass III and VH segments of subclass III (Kabat et al., 1991), respectively. For humanization of chicken anti-IL-12 monoclonal antibody DD2, the human genomic V segments DPL16 of VE subgroup III (Williams et al., 1996) and DP-54 of the VH subgroup III (Tomlinson et al., 1992) were used. Since the heterogeneity of framework amino acid sequences of affinity-matured V genes is not as extensive in chickens as that observed in mice or humans, humanization of chicken monoclonal antibodies may be achieved with a small set of human variable region sequences as acceptors. This was indeed indicated when V region framework sequences of eight chicken monoclonal antibodies isolated in this work were aligned with human germline V segment sequences (data not shown). The human DPL16 VE segment was one of the most similar sequences to all of the eight chicken VE sequences. For the VH sequences, a small set of human germline VH segments that belong to subgroup III, including DP-47, DP-49 and DP-54 (Tomlinson et al., 1992), was identified to provide the most similar acceptor sequences. In addition, comparison of the chicken V region sequences with their most similar human counterparts revealed that certain framework amino acids that could potentially interact with the CDRs were almost always different between chickens and humans. These amino acids are located at positions 46, 66 and 69 in VE and 67, 78 and 93 in VH (cf. Fig. 4). Therefore, successfully humanized chicken monoclonal antibodies will most likely require chicken-derived residues at these six locations. Three-dimensional modeling and humanization of the VE domain of the chicken anti-IL-12 monoclonal antibody DD2 were particularly challenging because chicken VE regions, when compared to V regions of known structure and to human VE regions, contain two amino acid deletions at the N-terminus of the
17
mature protein and one amino acid insertion at position 39A (Fig. 4). The human genomic VE segments of subgroup III (Kabat et al., 1991), which have a high homology to chicken VE sequences, start with either Ser–Tyr or Ser–Ser at the N-terminus of the mature protein (Williams et al., 1996). While examination of the molecular model of the DD2 variable region suggested that the addition of two serine residues at the N-terminus of the humanized form would not drastically alter the CDR conformation, the addition of a tyrosine residue at the second position was predicted to potentially cause a structural change in the CDRs. The DPL16 VE segment, which has high similarity to the DD2 VE sequence and starts with Ser–Ser at the N-terminus, was therefore chosen as an acceptor sequence for humanization. In addition, examination of the molecular model suggested that removal of a serine residue at position 39A would not cause a drastic change in either the CDR or framework structure. Although a subtle change of the CDR structure in the humanized form could have resulted in reduction of the affinity, our data showed that no loss in the affinity to human and mouse IL-12 was detected for humanized DD2, suggesting that humanization technology based on computer-guided molecular modeling is reliable and can be successfully applied beyond antibodies of mammalian origin. The ultimate goal of antibody humanization is to eliminate potential immunogenicity of non-human antibodies, such as chicken antibodies, in human patients. In this work, a total of 11 chicken framework amino acids (six from VE and five from VH) were retained in humanized DD2. When another chicken anti-IL-12 monoclonal antibody was successfully humanized, a total of eight chicken framework residues were retained in the humanized form to preserve its antigen-binding affinity (data not shown). These numbers are not unusually high when compared with the humanization of murine monoclonal antibodies. In the case of humanizing the mouse antiTac monoclonal antibody to generate daclizumab, for example, a total of nine mouse-specific framework amino acids were retained to preserve CDR conformation (Queen et al., 1989). In clinical trials with daclizumab involving 535 patients, no severe adverse events due to immunogenicity were reported (Bumgardner et al., 2001). Therefore, even though humanized chicken antibodies may tend to retain
18
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19
approximately 10 chicken framework amino acids, it seems unlikely that they will be highly immunogenic in humans. In this paper, we reported the first successful humanization of a non-mammalian monoclonal antibody with the aid of computer-guided molecular modeling. Since the structure of the variable regions of antibodies is relatively well conserved across species, our data implies that humanization of monoclonal antibodies of any species is achievable by the same general approach using molecular modeling. Chicken antibodies can recognize antigens and epitopes to which the rodent and other mammalian immune systems cannot respond effectively due to tolerance. In addition, since certain chicken antibodies can simultaneously recognize an antigen of human and murine (or other mammalian) origin, a single humanized chicken monoclonal antibody may be useful for both preclinical studies with disease animal models and clinical trials in patients. With advances in methodologies for more efficiently isolating desired chicken monoclonal antibodies, chickens should become an important source of potent humanized antibodies for treatment of human diseases. Acknowledgements The authors thank Catherine Huey, Brett Jorgensen, Linh Le, David Maciejewski, Thang Pham, and Raymond Ogawa for their technical assistance, J. Tso for valuable discussion throughout the work, Cary Queen for his support of the project, and Yoshiko Akamatsu, Paul Hinton, and David Powers for their comments on the manuscript. References Akamatsu, Y., Cole, M.S., Tso, J.Y., Tsurushita, N., 1993. Construction of a human Ig combinatorial library from genomic V segments and synthetic CDR3 fragments. J. Immunol. 151, 4651. Andris-Widhopf, J., Rader, C., Steinberger, P., Fuller, R., Barbas III, C.F., 2000. Methods for the generation of chicken monoclonal antibody fragments by phage display. J. Immunol. Methods 242, 159. Bumgardner, G.L., Hardie, I., Johnson, R.W., Lin, A., Nashan, B., Pescovitz, M.D., Ramos, E., Vincenti, F., 2001. Results of 3year phase III clinical trials with daclizumab prophylaxis for
prevention of acute rejection after renal transplantation. Transplantation 72, 839. Camenisch, G., Tini, M., Chilov, D., Kvietikova, I., Srinivas, V., Caro, J., Spielmann, P., Wenger, R.H., Gassmann, M., 1999. General applicability of chicken egg yolk antibodies: the performance of IgY immunoglobulins raised against the hypoxia-inducible factor 1alpha. FASEB J. 13, 81. Carnemolla, B., Neri, D., Castellani, P., Leprini, A., Neri, G., Pini, A., Winter, G., Zardi, L., 1996. Phage antibodies with panspecies recognition of the oncofoetal angiogenesis marker fibronectin ED-B domain. Int. J. Cancer 68, 397. Charlton, K.A., Moyle, S., Porter, A.J., Harris, W.J., 2000. Analysis of the diversity of a sheep antibody repertoire as revealed from a bacteriophage display library. J. Immunol. 164, 6221. Co, M.S., Avdalovic, N.M., Caron, P.C., Avdalovic, M.V., Scheinberg, D.A., Queen, C., 1992. Chimeric and humanized antibodies with specificity for the CD33 antigen. J. Immunol. 148, 1149. Ehrlich, P.H., Moustafa, Z.A., Justice, J.C., Harfeldt, K.E., Kelley, R.L., Ostberg, L., 1992. Characterization of human monoclonal antibodies directed against hepatitis B surface antigen. Hum. Antib. Hybrid. 3, 2. Fagnani, R., 1994. The immunogenicity of foreign monoclonal antibodies in human disease applications: problems and current approaches. Immunol. Ser. 61, 3. Feldhaus, M.J., Siegel, R.W., Opresko, L.K., Coleman, J.R., Feldhaus, J.M., Yeung, Y.A., Cochran, J.R., Heinzelman, P., Colby, D., Swers, J., Graff, C., Wiley, H.S., Wittrup, K.D., 2003. Flow-cytometric isolation of human antibodies from a nonimmune Saccharomyces cerevisiae surface display library. Nat. Biotechnol. 21, 163. Gassmann, M., Thommes, P., Weiser, T., Hubscher, U., 1990. Efficient production of chicken egg yolk antibodies against a conserved mammalian protein. FASEB J. 4, 2528. He, X.Y., Xu, Z., Melrose, J., Mullowney, A., Vasquez, M., Queen, C., Vexler, V., Klingbeil, C., Co, M.S., Berg, E.L., 1998. Humanization and pharmacokinetics of a monoclonal antibody with specificity for both E- and P-selectin. J. Immunol. 160, 1029. Higuchi, R., 1990. Recombinant PCR. In: Innis, M.A., Gelfand, D.H., Sninsky, J.J., White, T.J. (Eds.), PCR Protocols. Academic Press, San Diego, CA, p. 177. Hinton, P.R., Johlfs, M.G., Xiong, J.M., Hanestad, K., Ong, K.C., Bullock, C., Keller, S., Tang, M.T., Tso, J.Y., Vasquez, M., Tsurushita, N., 2004. Engineered human IgG antibodies with longer serum half-lives in primates. J. Biol. Chem. 279, 6213. Hoogenboom, H.R., de Bruine, A.P., Hufton, S.E., Hoet, R.M., Arends, J.W., Roovers, R.C., 1998. Antibody phage display technology and its applications. Immunotechnology 4, 1. Kabat, E.A., Wu, T.T., Perry, H.M., Gottesman, K.S., Foeller, C., 1991. Sequences of Proteins of Immunological Interest, vol. 1. U.S. Department of Health and Human Services, Bethesda, MD. Khazaeli, M.B., Conry, R.M., LoBuglio, A.F., 1994. Human immune response to monoclonal antibodies. J. Immunother. 15, 42. Kohler, G., Milstein, C., 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256, 495.
N. Tsurushita et al. / Journal of Immunological Methods 295 (2004) 9–19 Korth, C., Stierli, B., Streit, P., Moser, M., Schaller, O., Fischer, R., Schulz-Schaeffer, W., Kretzschmar, H., Raeber, A., Braun, U., Ehrensperger, F., Hornemann, S., Glockshuber, R., Riek, R., Billeter, M., Wuthrich, K., Oesch, B., 1997. Prion (PrPSc)specific epitope defined by a monoclonal antibody. Nature 390, 74. Kuus-Reichel, K., Grauer, L.S., Karavodin, L.M., Knott, C., Krusemeier, M., Kay, N.E., 1994. Will immunogenicity limit the use, efficacy, and future development of therapeutic monoclonal antibodies? Clin. Diagn. Lab. Immunol. 1, 365. Levy, R., Assulin, O., Scherf, T., Levitt, M., Anglister, J., 1989. Probing antibody diversity by 2D NMR: comparison of amino acid sequences, predicted structures, and observed antibody– antigen interactions in complexes of two antipeptide antibodies. Biochemistry 28, 7168. Li, Y., Kilpatrick, J., Whitelam, G.C., 2000. Sheep monoclonal antibody fragments generated using a phage display system. J. Immunol. Methods 236, 133. Lunn, M.P., Johnson, L.A., Fromholt, S.E., Itonori, S., Huang, J., Vyas, A.A., Hildreth, J.E., Griffin, J.W., Schnaar, R.L., Sheikh, K.A., 2000. High-affinity anti-ganglioside IgG antibodies raised in complex ganglioside knockout mice: reexamination of GD1a immunolocalization. J. Neurochem. 75, 404. Nishinaka, S., Akiba, H., Nakamura, M., Suzuki, K., Suzuki, T., Tsubokura, K., Horiuchi, H., Furusawa, S., Matsuda, H., 1996. Two chicken B cell lines resistant to ouabain for the production of chicken monoclonal antibodies. J. Vet. Med. Sci. 58, 1053. O’Brien, P.M., Aitken, R., O’Neil, B.W., Campo, M.S., 1999. Generation of native bovine mAbs by phage display. Proc. Natl. Acad. Sci. U. S. A. 96, 640. Queen, C., Schneider, W.P., Selick, H.E., Payne, P.W., Landolfi, N.F., Duncan, J.F., Avdalovic, N.M., Levitt, M., Junghans, R.P., Waldmann, T.A., 1989. A humanized antibody that binds to the interleukin 2 receptor. Proc. Natl. Acad. Sci. U. S. A. 86, 10029. Rader, C., Ritter, G., Nathan, S., Elia, M., Gout, I., Jungbluth, A.A., Cohen, L.S., Welt, S., Old, L.J., Barbas III, C.F., 2000. The rabbit antibody repertoire as a novel source for the generation of therapeutic human antibodies. J. Biol. Chem. 275, 13668. Rajewsky, K., 1996. Clonal selection and learning in the antibody system. Nature 381, 751. Ravetch, J.V., Siebenlist, U., Korsmeyer, S., Waldmann, T., Leder, P., 1981. Structure of the human immunoglobulin mu locus: characterization of embryonic and rearranged J and D genes. Cell 27, 583. Reichert, J., Pavlou, A., 2004. Monoclonal antibodies market. Nat. Rev., Drug Discov. 3, 383.
19
Reynaud, C.-A., Dahan, A., Anquez, V., Weill, J.-C., 1989. Development of the chicken antibody repertoire. In: Honjo, T., Alt, F.W., Rabbitts, T.H. (Eds.), Immunoglobulin Genes. Academic Press, San Diego, CA, p. 151. Ridder, R., Schmitz, R., Legay, F., Gram, H., 1995. Generation of rabbit monoclonal antibody fragments from a combinatorial phage display library and their production in the yeast Pichia pastoris. Biotechnology (NY) 13, 255. Spieker-Polet, H., Sethupathi, P., Yam, P.C., Knight, K.L., 1995. Rabbit monoclonal antibodies: generating a fusion partner to produce rabbit–rabbit hybridomas. Proc. Natl. Acad. Sci. U. S. A. 92, 9348. Tanha, J., Dubuc, G., Hirama, T., Narang, S.A., MacKenzie, C.R., 2002. Selection by phage display of llama conventional V(H) fragments with heavy chain antibody V(H)H properties. J. Immunol. Methods 263, 97. Tomlinson, I.M., Walter, G., Marks, J.D., Llewelyn, M.B., Winter, G., 1992. The repertoire of human germline VH sequences reveals about fifty groups of VH segments with different hypervariable loops. J. Mol. Biol. 227, 776. Tsurushita, N., Vasquez, M., 2004. Humanization of monoclonal antibodies. In: Honjo, T., Alt, F.W., Neuberger, M.S. (Eds.), Molecular Biology of B Cells. Elsevier Academic Press, San Diego, CA, p. 533. Udey, J.A., Blomberg, B., 1987. Human lambda light chain locus: organization and DNA sequences of three genomic J regions. Immunogenetics 25, 63. Walter, U.M., Ayer, L.M., Wolitzky, B.A., Wagner, D.D., Hynes, R.O., Manning, A.M., Issekutz, A.C., 1997. Characterization of a novel adhesion function blocking monoclonal antibody to rat/ mouse P-selectin generated in the P-selectin-deficient mouse. Hybridoma 16, 249. Williams, S.C., Frippiat, J.P., Tomlinson, I.M., Ignatovich, O., Lefranc, M.P., Winter, G., 1996. Sequence and evolution of the human germline V lambda repertoire. J. Mol. Biol. 264, 220. Winter, G., Griffiths, A.D., Hawkins , R.E., Hoogenboom, H.R., 1994. Making antibodies by phage display technology. Annu. Rev. Immunol. 12, 433. Yamanaka, H.I., Inoue, T., Ikeda-Tanaka, O., 1996. Chicken monoclonal antibody isolated by a phage display system. J. Immunol. 157, 1156. Zilber, B., Scherf, T., Levitt, M., Anglister, J., 1990. NMR-derived model for a peptide–antibody complex. Biochemistry 29, 10032.