Antibody engineering via genetic engineering of the mouse: XenoMouse strains are a vehicle for the facile generation of therapeutic human monoclonal antibodies

https://doi.org/10.1016/S0022-1759(99)00137-4Get rights and content

Abstract

The major impediment to the development of murine monoclonal antibodies (mAbs) for therapy in humans has been the difficulty in reducing their potential immunogenicity. XenoMouse™ mice obviate this problem while retaining the relative ease of generating mAbs from a mouse. XenoMouse strains include germline-configured, megabase-sized YACs carrying portions of the human IgH and Igκ loci, including the majority of the variable region repertoire, the genes for Cμ, Cδ and either Cγ1, Cγ2, or Cγ4, as well as the cis elements required for their function. The IgH and Igκ transgenes were bred onto a genetic background deficient in production of murine immunoglobulin. The large and complex human variable region repertoire encoded on the Ig transgenes in XenoMouse strains support the development of large peripheral B cell compartments and the generation of a diverse primary immune repertoire similar to that from adult humans. Immunization of XenoMouse mice with human antigens routinely results in a robust secondary immune response, which can ultimately be captured as a large panel of antigen-specific fully human IgGκ mAbs of sub-nanomolar affinities. Monoclonal antibodies from XenoMouse animals have been shown to have therapeutic potential both in vitro and in vivo, and appear to have the pharmacokinetics of normal human antibodies based on human clinical trials. The utility of XenoMouse strains for the generation of large panels of high-affinity, fully human mAbs can be made available to researchers in the academic and private sectors, and should accelerate the development and application of mAbs as therapeutics for human disease.

Introduction

A quarter century after the discovery of monoclonal antibodies (mAbs; Kohler and Milstein, 1975), their therapeutic utility is finally being realized. Monoclonal antibodies have now been approved for therapies in transplantation (Smith, 1996; Waldmann and O'Shea, 1998; Kahan et al., 1999; Nashan et al., 1999), cancer (Leget and Czuczman, 1998; Goldenberg, 1999), infectious disease (Storch, 1998), cardiovascular disease (Coller et al., 1995) and inflammation (Present et al., 1999). Many more mAbs are in late stage clinical trials for a broad range of disease indications. As a group, mAbs represent one of the largest classes of drugs in development.

In part, the potential therapeutic utility of mAbs stems from their specific and high affinity binding to targets coupled with their diversity of function. A mAb with a constant region with effector functions, e.g., human IgG1, can be used to direct complement dependent cytotoxicity or antibody-dependent cytotoxicity to a target cell (Scallon et al., 1995). Alternatively, a mAb with a constant region essentially lacking effector function, e.g., human IgG2 or IgG4, can be used to block signal transduction, either by binding to and neutralizing a ligand, or by blocking a receptor binding site (Yang et al., 1999a, Yang et al., 1999b). If conjugated to a radioisotope, chemical or biological toxin, mAbs can act as targeting agents, homing cytotoxic agents to a desired site such as a tumor while minimizing non-specific damage to other sites (Meredith et al., 1997; Reiter and Pastan, 1998). Additionally, mAbs can act as agonists, stimulating a receptor to cause a desired therapeutic effect (Melero et al., 1997; Schneider et al., 1997).

Many therapeutic applications for mAbs would require repeated administrations, especially for chronic diseases such as autoimmunity or cancer. Because mice are convenient for immunization and recognize most human antigens as foreign, mAbs against human targets with therapeutic potential have typically been of murine origin. However, murine mAbs have inherent disadvantages as human therapeutics. They require more frequent dosing to maintain a therapeutic level of mAb because of a shorter circulating half-life in humans than human antibodies (Weinstein et al., 1987; Zuckier et al., 1989). More critically, repeated administration of murine immunoglobulin (Ig) creates the likelihood that the human immune system will recognize the mouse protein as foreign, generating a human anti-mouse antibody (HAMA) response. At best, a HAMA response will result in a rapid clearance of the murine antibody upon repeated administration, rendering the therapeutic useless. Worse, a HAMA response can cause a severe allergic reaction (Jaffers et al., 1986; Abramowicz et al., 1992; Barjorin et al., 1992; Choy, 1998). As discussed below, this possibility of reduced efficacy and safety has lead to the development of a number of technologies for reducing the immunogenicity of murine mAbs.

An optimal technology for generation of mAbs lacking immunogenicity in humans should have several characteristics. Ideally, the technology should produce mAbs that are fully human in sequence, lacking any murine amino acid components. It should generally be able to generate mAbs against human antigens. These mAbs should routinely be of high affinity and high specificity. The technology should easily generate a mAb with the desired effector function. Critically, the technology should be easy to use so that many users can generate mAbs without the need for sophisticated molecular biology techniques.

Some of the early attempts at generating fully human mAbs used human hybridomas. However, numerous difficulties, including the instability of the human hybridomas, a preponderance of low affinity IgM mAbs, and the difficulty or ethical problems of finding humans immunized against the target antigen, resulted in this technology being used only rarely (Winter and Milstein, 1991). Others have chimerized mAbs by linking the murine variable region to human constant regions (Morrison and Oi, 1989). While requiring only relatively simple antibody engineering techniques, chimerized mAbs still may cause a human anti-chimeric antibody (HACA) response, an immune response in humans because of the fully murine framework and complementarity determining region (CDR) sequences of the antibody variable regions (reviewed in Adair, 1992; see also Elliott et al., 1994).

More sophisticated technologies such as humanization and phage display use tools of molecular biology and prokaryotic genetics to re-engineer antibodies to resemble their human counterparts. Humanization technology starts with a murine antibody with the desired characteristics and then changes, from a mouse-like to human-like usage, those amino acids thought to be non-critical to antigen recognition (Riechmann et al., 1988). Typically, this technology requires individual tailoring for each antibody, including extensive molecular modeling and manipulation of the DNA encoding the mAb. Even then, amino acid changes that would be predicted to have little effect on the antibody could unexpectedly abrogate antibody function. Finally, humanization leaves murine amino acids in the antibody, allowing the possibility of a HAMA response. In phage display technology, combinatorial libraries of human naive variable regions displayed on the surface of bacteriophage are used to develop antigen-specific antibodies (Burton and Barbas, 1993; Griffiths and Hoogenboom, 1993). However, this technology typically requires successive rounds of in vitro mutagenesis, V gene shuffling and/or panning to produce antibodies with sub-nanomolar affinity (Vaughan et al., 1998). The synthesis of a complete mAb from phage display technology requires that the DNA encoding the in vitro affinity-matured V regions be further manipulated by functionally linking them to DNA encoding constant regions in a suitable expression vector, expressing the construct in a tissue culture system, and finally confirming that the synthetic human mAb does indeed have the desired function.

An alternative strategy for producing human mAbs is to alter the mouse humoral immune system so that it will produce fully human antibodies, obviating the need for re-engineering of the actual antibodies themselves and leaving intact the powerful natural mechanisms for class switching and affinity maturation. This strategy would require considerable effort up front to genetically engineer the mouse, but the resulting technology would be user friendly, enabling the end-user to produce high-affinity human antibodies against human antigens using only standard hybridoma procedures. The XenoMouse technology represents the success of this vision.

XenoMouse strains are genetically engineered mice in which the murine IgH and Igκ loci have been functionally replaced by their human Ig counterparts on yeast artificial chromosome (YAC) transgenes. These human Ig transgenes carry the majority of the human variable repertoire and can undergo class switching from IgM to IgG isotypes. The large and complex human V repertoires on the YAC transgenes support development of a large B cell population and the formation of a broad and diverse primary immune repertoire. The human genes are compatible with mouse enzymes mediating class switching from IgM to IgG as well as somatic hypermutation and affinity maturation. The immune system of the XenoMouse strains recognizes administered human antigens as foreign, with a concomitant strong human humoral immune response. The use of XenoMouse mice in conjunction with well-established hybridoma procedures reproducibly results in IgG mAbs with sub-nanomolar affinities for human antigens and with suitability for repeated administration to humans.

Section snippets

Engineering of mice to produce fully human IgM and IgG

Recapitulation of a robust human humoral immune response in mice required two types of major genetic modifications of the mouse genome (Fig. 1). Mouse ES cells were used as the vehicle for altering the mouse genome because of their proven utility for introducing precise mutations, followed by efficient transmission into the germline. The first modification, ablation of the ability of the mouse to produce murine Ig, was achieved through inactivation of the endogenous murine IgH and Igκ loci

The large and complex V repertoire of XenoMouse mice is essential for efficient restoration of B cell development and a human-like humoral immune response

In XenoMouse mice, yH2 and yK2 were able to substitute for their murine counterparts, as demonstrated by restoration of the B cell compartments in all organs examined, the bone marrow, blood, spleen and lymph nodes (Mendez et al., 1997; Green and Jakobovits, 1998). For example, the spleens of XenoMouse have an abundant population of B220++ cells, about 80% of the population of B220++ cells in wild-type mice (Fig. 3A). In contrast, DI mice completely lack B cells in the splenic lymphocytes (

Generation and production of high affinity antigen-specific fully human mAbs

Supported by reconstituted B cell development and a broadly diverse antibody repertoire, XenoMouse strains allowed the derivation of high affinity, fully human mAbs against numerous human antigens. XenoMouse strains immunized with human antigens such as epidermal growth factor receptor (EGFr), interleukin-8 (IL-8), tumor necrosis factor α (TNF-α), interleukin-6 (IL-6), L-selectin, GROα, and CD147, amongst others, mounted a robust antigen-specific immune response (Mendez et al., 1997and data not

Conclusions

The XenoMouse strains described above are powerful tools for the generation of high affinity fully human mAbs suitable for use in humans as therapeutic agents. By using XenoMouse animals, extensive and successive rounds of antibody engineering by the user are unnecessary because the required engineering, genetic engineering of large portions of the native human Ig loci into the mouse germline, was performed at a much earlier stage. Thus, with the human Ig transgenes functionally replacing their

References (50)

  • D Abramowicz et al.

    Anaphylactic shock after retreatment with OKT3 monoclonal antibody

    N. Engl. J. Med.

    (1992)
  • Adair, J.R., 1992. Engineering antibodies for therapy. In: Immunological Reviews. Munksgarrd, Copenhagen, p....
  • D.F Barjorin et al.

    Treatment with high dose mouse monoclonal (anti-GD3) antibody R24 in patients with metastatic melanoma

    Melanoma Res.

    (1992)
  • H.P Brezinschek et al.

    Analysis of the heavy chain repertoire of human peripheral B cells using single-cell polymerase chain reaction

    J. Immunol.

    (1995)
  • Burrows, P.D., Schroeder, H.W., Jr., Cooper, M.D., 1995. B cell differentiation in humans. In: Honjo, T., Alt, F.W....
  • Burton, D.R., Barbas, C.F., 1993. Human antibodies from combinatorial libraries. In: Nottingham, C.M. (Ed.), Protein...
  • E.H Choy

    Clinical pharmacology and therapeutic potential of monoclonal antibody treatment in rheumatoid arthritis

    Drugs Aging

    (1998)
  • B.S Coller et al.

    New antiplatelet agents: platelet GPIIb/IIIa antagonists

    Thromb. Haemostasis

    (1995)
  • J.P.L Cox et al.

    A directory of human germ-line Vκ segments reveals a strong bias in their usage

    Eur. J. Immunol.

    (1994)
  • D.M Fishwild et al.

    High-avidity human Igκ monoclonal antibodies from a novel strain of minilocus transgenic mice

    Nat. Biotechnol.

    (1996)
  • L.L Green et al.

    Regulation of B cell development by variable gene complexity in mice reconstituted with human immunoglobulin yeast artificial chromosomes

    J. Exp. Med.

    (1998)
  • L.L Green et al.

    Antigen-specific human monoclonal antibodies from nice engineered with human Ig heavy and light chain YACs

    Nat. Genet.

    (1994)
  • Griffiths, A.D., Hoogenboom, H.R., 1993. Building an in vitro immune system: human antibodies from phage display...
  • G.J Jaffers et al.

    Monoclonal antibody therapy. Anti-idiotypic and non-anti-idiotypic antibodies to OKT3 arising despite intense immunosuppression

    Transplantation

    (1986)
  • A.J Jakobovits et al.

    Analysis of homozygous mutant chimeric mice: Deletion of the immunoglobulin heavy-chain joining region blocks B cell development and antibody production

    Proc. Natl. Acad. Sci. U.S.A.

    (1993)
  • Cited by (0)

    View full text