Chemistry: natural resins or derivatives; peptides or proteins; – Proteins – i.e. – more than 100 amino acid residues – Blood proteins or globulins – e.g. – proteoglycans – platelet...
Reexamination Certificate
1995-06-27
2002-11-05
Saunders, David (Department: 1644)
Chemistry: natural resins or derivatives; peptides or proteins;
Proteins, i.e., more than 100 amino acid residues
Blood proteins or globulins, e.g., proteoglycans, platelet...
C424S136100
Reexamination Certificate
active
06476198
ABSTRACT:
TECHNICAL FIELD
The present invention relates to multispecific and multivalent antigen-binding polypeptides and methods for producing them.
BACKGROUND
The ability to target and activate cytotoxic lymphocytes opens the way to utilize natural effector functions to treat cancer, auto-immune disorders and infectious diseases. See, Segal et al.,
Chem. Immunology,
Ed. Ishizaka et al., 47:179-213 (1989). Bispecific antibodies allow this linking of target cell to the effector cell. However, the bispecific antibodies generated so far have two major drawbacks.
Firstly, the specificities against human antigens are predominantly encoded by rodent antibodies and would result in a human anti-murine antibody (HAMA) response in repeated or prolonged use. Although the specificity problem has partially been overcome by forming chimeric (Knox, et al.,
Blood,
77:20, 1991); Shaw, et al.,
J. Biol. Response Med.,
7:204, (1988); and Oudin, et al.,
Proc. Natl. Acad. Sci. USA,
63:266, (1971)) and humanized (Hale, et al.,
Lancet,
2:1394, (1988)) antibodies, the problem of immunogenicity still persists. Advances in applying combinatorial antibody repertoire cloning for the generation of human monoclonal antibodies provides a source of human derived variable, framework and constant regions thus theoretically avoiding a HAMA response. This approach works well when an immune source for the desired specificity is available, but is of limited value in the search for antibodies which specifically bind human cell surface molecules.
Secondly, current methods for producing bispecific molecules rely on either chemical cross linking or heterohybridoma (quadroma) formation. The former approach results in a heteroconjugates resulting in preparations that vary from batch to batch in terms of composition and consequently potency. The latter approach involves fusing the two hybridoma cell lines producing the desired antibodies, giving rise to a quadroma cell encoding and expressing all the H and L chains. The desired H and L chain combination for the bispecific antibody is usually only 15% of the total antibody and is difficult to isolate from the closely related pool of antibodies as described by Milstein et al.,
Nature,
305:537 (1983).
Ideally human variable regions conferring the desired specificities should be used to construct a single molecule containing both the specificities as the major if not only product. Alternatively, variable regions conferring the desired specificities should be used to construct two polypeptide molecules containing immunologically distinct constant region domains wherein the majority of the two polypeptide molecules are linked by disulphide bonding. Such molecules would avoid the HAMA response while facilitating production of a homogeneous product for characterization and clinical evaluation.
The ability to PCR amplify, directionally clone and express antibody variable regions from cDNA has allowed hybridoma technology to be bypassed. Diverse high affinity antibodies have been generated to hapten, virus particles and protein antigens, thereby recapitulating functional molecules appearing during the natural immune response in animals and in humans. See, for example, Skerra et al.,
Science,
240:1038-1041 (1988); Better et al.,
Science,
240:1041-1043 (1988); Orlandi et al.,
Proc. Natl. Acad. Sci., USA,
86:3833-3837 (1989); Kang et al.,
Proc. Natl. Acad. Sci., USA,
88:4363-4366 (1991); and Barbas et al.,
Proc. Natl. Acad. Sci.. USA,
88:7978-7982 (1991). Marks et al.,
J. Mol. Biol.,
222:581-597 (1991) demonstrated that active single chain antibody Fv fragments with affinity constants in the range of 10
6
-10
7
M-l against a hapten or a small number of epitopes on a protein can be obtained directly from non-immune combinatorial immunoglobulin libraries. More recently, a combinatorial library approach was used to select monoclonal antibodies from non-immune mice and subsequently affinity mature the specificities, thereby establishing the principles of (i) accessing naive combinatorial antibody libraries for predefined specificities and (ii) increasing the affinity of the selected antibodies binding sites by random mutagenesis. See, Gram et al.,
Proc. Natl. Acad. Sci., USA,
89:3576-3580 (1992).
In addition, large libraries of antibody Fab fragments have been displayed on the surface of phage. See, for example, Kang et al.,
Proc. Natl. Acad. Sci., USA,
88:4363-4366 (1991); Hoogenboom et al.,
Nuc. Acids. Res.,
19:4133-4137 (1991); Burton et al.,
Proc. Natl. Acad. Sci., USA,
88:10134-10137 (1991); Griffiths et al.,
EMBO J.,
12:725-734 (1993); and Soderlind et al., Bio/Technology, 11:503-507 (1993). In essence, the antigen recognition unit has been linked to instructions for its production. An iterative process of mutation followed by selection has also been developed allowing for the rapid generation of specific antibodies from germ line sequences as describe by Gram et al., supra.
The B cell immune response to an antigen can be viewed to occur in two stages. The initial stage generates low affinity antibodies mostly of the IgM isotype from an existing pool of the B-cell repertoire available at the time of immunization. The second stage which is driven by antigen stimulation produces high affinity antibodies predominantly of the IgG isotype, starting with the VH and VL genes selected in the primary response. The predominant mechanism for affinity maturation is hypermutation of variable region genes (and possibly gene conversion) followed by selection of those cells which produce antibodies of the highest affinity. In its simplest form, the initial stage of the immune response can be recreated in vitro by generating a combinatorial library of PCR amplified IgM/G and light chains from the bone marrow of adults. This is a close approximation to the naive, unselected repertoire, since the majority of the B cells in the bone marrow expressing IgM/G chains have not been subjected to tolerance and antigen selection, and should therefore represent all the combinatorial diversity of immunoglobulin V-regions. See, Decker et al.,
J. Immunol.,
146:350-361 (1991). The phagemid pComb8 facilitates the display of multiple copies of the single chain antibody along the phage surface permitting the access to low affinity antibodies as described by Kang et al., supra and Gram et al., supra. Hence, specific VH and VL pairs could possibly be enriched from a diverse naive repertoire.
The second stage of the immune response in vivo involves affinity maturation of the selected specificities by mutation and selection. An efficient way to generate random mutations is by an error-prone replication mechanism, either by targeting the mutations to the antibody binding sites by error-prone PCR as described by Leung et al.,
J. Methods Cell Molecular Biol.,
1:1-15 (1989), or by passaging the phagemid carrying the genetic information for the antigen binding domain through an
E.coli
mutD strain, in which the spontaneous mutation frequency is 103 to 105 times higher than in a wild-type strain as described by Fowler et al.,
J. Bacteriol.,
167:130-137 (1986). Selected VH and VL pairs could be subjected to error prone PCR (also gene conversion by PCR is feasible) and the resulting products cloned into phagemid pComb3 which facilitates the display of a single copy of the mutant single chain Fv, such low level of display permits the isolation of the highest affinity molecules.
Variations of single chain bispecific molecules have been constructed in bacteria by linking two single chain heavy and light chain variable domains (sFv) with a synthetic linker. See, for example, Wels et al.,
Bio/Technology,
10:1128-1132 (1992); Stemmer et al.,
BioTechniques,
14:256-265 (1993); Goshorn et al.,
Cancer Res.,
53:2123-2127 (1993); and Bos et al.,
Biotherapy,
5:187-199 (1992). The molecules generated, however, are incorrectly folded and on denaturing and refolding result in very low yield. This may be intrinsically due to expressing both the specificities as a single protein giving rise to inter- and i
Fitting Thomas
Holmes Emily
Northrup Thomas E.
Saunders David
The Scripps Research Institute
LandOfFree
Multispecific and multivalent antigen-binding polypeptide... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Multispecific and multivalent antigen-binding polypeptide..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Multispecific and multivalent antigen-binding polypeptide... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2987757