JPWO1999018212A1 - Naturally occurring humanized antibodies - Google Patents

Abstract

(57) [Abstract] A reshaped human anti-HM1.24 antibody comprising: (A) an L chain comprising (1) a human L chain C region, and (2) an L chain V region comprising human L chain FRs and the L chain CDRs of a mouse anti-HM1.24 monoclonal antibody; and (B) an H chain comprising (1) a human H chain C region, and (2) an H chain V region comprising human H chain FRs and the H chain CDRs of a mouse anti-HM1.24 monoclonal antibody. Because the majority of this reshaped human antibody is derived from a human antibody and the CDRs have low antigenicity, the reshaped human antibody of the present invention has low antigenicity to humans and is therefore expected to be used for medical therapy.

Description

[Detailed Description of the Invention] Natural Humanized Antibodies Field of the Invention The present invention relates to a method for producing natural humanized antibodies and natural humanized antibodies obtained by the production method. The present invention also relates to DNA encoding natural humanized antibodies, expression vectors containing the DNA, hosts containing the DNA, and methods for producing natural humanized antibodies from cells transfected with the DNA.

BACKGROUND ART Mouse monoclonal antibodies can be isolated relatively easily using standard hybridoma technology (Kohler, G. and Milstein, C. Nature (1975) 256, 495-497). While similar technology for human hybridomas has been anticipated, it has not yet become standard technology. Furthermore, antibodies against human antigens are required for clinical applications, and therefore, the generation of mouse monoclonal antibodies is essential for the development of antibody drugs.

Numerous mouse monoclonal antibodies against cancer cells and viruses have been isolated and their clinical applications have been investigated. However, mouse antibodies are foreign to humans, and they have been found to have significant drawbacks for clinical application, such as the induction of HAMA (Human Anti-Mouse Antibody) due to their strong antigenicity and poor ADCC induction (Schroff, R.W., Cancer Res. (1985) 45, 879-885; Shawler, D.L., et al., J. Immunol. (1985) 135, 1530-1535).

To address this issue, chimeric antibodies were first constructed (Neuberger, M.S. et al., Nature (1984) 312, 604-608; Boulianne, G.L. et al., Nature (1984) 312, 643-646). Chimeric antibodies combine the variable regions of mouse antibodies with the constant regions of human antibodies, replacing the constant regions responsible for particularly strong antigenicity with human ones. This is expected to enable physiological binding to human Fc receptors and induce Fc-mediated functions. In fact, clinical trials using chimeric antibodies have reported significantly reduced antigenicity (LoBuglio, A.F. et al., Proc. Natl. Acad. Sci. USA (1989) 86, 4220-4224). However, HAMA against mouse variable regions still occurs, and problematic cases have been identified (LoBuglio, A. F. et al., Proc. Natl. Acad. Sci. USA (1989) 86, 4220-4224).

Therefore, a more complex but more human-like method for creating humanized antibodies has been developed. This involves reconstructing the antigen-binding site of a mouse antibody onto a human antibody (Jones, P.T. et al., Nature (1986) 321, 522-525; Verhoeyen, M. et al., Science (1988) 239, 1534-1536; Riechmann, L. et al., Nature (1988) 332, 323-327). The antibody variable region, both the heavy chain and the light chain, consists of four framework regions (FRs) flanked by three complementarity-determining regions (CDRs).

The formation of the antigen-binding site is primarily carried out by the CDRs, and some amino acid residues on the FRs are also known to be directly or indirectly involved in this process. Because the basic structures of antibodies are similar, it was thought possible to graft the antigen-binding site of one antibody onto another. In fact, a research group led by G. Winter successfully created a humanized antibody with lysozyme-binding activity by grafting the CDRs of a mouse anti-lysozyme antibody onto a human antibody (CDR-grafting) (Jones, P.T. et al., Nature (1986) 321, 522-525).

However, humanization by CDR-grafting alone may not produce a humanized antibody with antigen-binding activity equivalent to that of the original mouse antibody. Therefore, as mentioned above, some FR amino acid residues are substituted. The FR amino acid residues to be substituted are those that contribute to the basic structure of the antibody molecule (canonical structure; Chothia, C. et al., Nature (1989) 342, 877-883; Chothia, C. and Lesk, A.M.J. Molec. Biol. (1987) 196, 901-917), are involved in maintaining the CDR structure, or directly interact with the antigen molecule.

In fact, most humanized antibodies contain amino acid substitutions in the FRs, resulting in artificial FR sequences not found in nature. Sometimes, numerous amino acid substitutions are made, casting doubt on the original purpose of CDR-grafting to minimize the antigenicity of mouse antibodies (Queen, C. et al., Proc. Natl. Acad. Sci. USA (1989) 86, 10029-10033; Co, M. S. et al., Proc. Natl. Acad. Sci. USA (1991) 88, 2869-2873).

The solution to this problem is to devise a more ingenious approach to the selection of human FRs. The number of FR amino acid residues to be substituted depends on the homology between the human antibody FRs selected for CDR-grafting and the original mouse antibody FRs. Therefore, human FRs are generally selected to be highly homologous to mouse FRs, with efforts made to minimize substitutions. However, even the FRs of humanized antibodies obtained in this way often contain amino acid sequences not previously found in nature, potentially resulting in antigenicity. A technology that can resolve this issue and construct safer humanized antibodies with a lower probability of inducing antigenicity is needed.

DISCLOSURE OF THE INVENTION The present invention improves upon existing methods for constructing humanized antibodies, providing a method for constructing humanized antibodies that fully retain the antigen-binding activity of the original mouse antibody and are composed entirely of naturally occurring human FRs—in other words, without any amino acid substitutions in the FRs.

Specifically, the present invention provides a method for producing a naturally occurring humanized antibody, which comprises conducting a homology search for the FRs of a primarily designed antibody and selecting naturally occurring human FRs that retain the artificial amino acid residues contained in the FRs of the primarily designed antibody and are homologous to the naturally occurring human FRs.

The primary designed antibody is a humanized antibody (also called a reshaped human antibody) produced by conventional CDR-grafting.

The present invention also provides a method for producing a natural humanized antibody, comprising: conducting a homology search for the FRs of a primarily designed antibody; selecting natural human FRs that retain the artificial amino acid residues contained in the FRs of the primarily designed antibody and are homologous; and then substituting one or more amino acid residues that differ between the FRs of the primarily designed antibody and the selected natural FRs.

Preferably, in these production methods, the primarily designed antibody comprises a CDR derived from a first animal species and an FR derived from a second animal species containing artificial amino acid residues. More preferably, in the primarily designed antibody, the first animal species is a non-human mammal, and the second animal species is a human. Examples of the first animal species, non-human mammals, include mice, rats, hamsters, rabbits, and monkeys.

The present invention also provides a method for producing a naturally occurring humanized antibody, comprising: conducting a homology search for the FRs of a primarily designed antibody; selecting naturally occurring human FRs that retain amino acid residues derived from non-human antibody FRs contained in the FRs of the primarily designed antibody and have high homology; and then substituting one or more amino acid residues that differ between the FRs of the primarily designed antibody and the selected naturally occurring FRs.

The present invention also provides a natural humanized antibody obtained by the above-mentioned production method.

The present invention also provides a naturally occurring humanized antibody comprising a CDR derived from a first animal species and a FR derived from a second animal species, wherein the FR has an amino acid sequence that differs from the FR used in CDR grafting by one or more amino acid residues, and the FR has been replaced with a FR derived from the second animal species containing the same amino acid residues at the same positions as the differing amino acid residues. Preferably, the first animal species is a non-human mammal, and the second animal species is human. Examples of the non-human mammal that is the first animal species include mouse, rat, hamster, rabbit, and monkey.

The present invention also provides DNA encoding the natural humanized antibody.

The present invention also provides an expression vector containing the DNA.

The present invention also provides a host containing the DNA.

The present invention further provides a method for producing a natural humanized antibody, comprising culturing cells into which an expression vector containing the above DNA has been introduced, and obtaining a desired natural humanized antibody from the cell culture.

The present invention further provides pharmaceutical compositions containing native humanized antibodies.

Brief Description of the Figures Figure 1 is a graph showing that, in FCM analysis using the human myeloma cell line KPMM2, the fluorescence intensity of the chimeric anti-HM1.24 antibody shifted compared to the control antibody, similar to the fluorescence intensity of the mouse anti-HM1.24 antibody.

Figure 2 shows that in Cell-ELISA using WISH cells, chimeric anti-HM1.24 antibody, like mouse anti-HM1.24 antibody, concentration-dependently inhibited the binding of biotinylated mouse anti-HM1.24 antibody to WISH cells.

Figure 3 is a graph showing that control human IgG1 or mouse anti-HM1.24 antibody had no cytotoxic activity against RPMI 8226 cells, whereas the cytotoxic activity of the chimeric anti-HM1.24 antibody against RPMI 8226 cells increased with increasing E/T ratio.

FIG. 4 is a schematic diagram showing a method for producing a reshaped human anti-HM1.24 antibody L chain by CDR grafting using PCR.

Figure 5 is a schematic diagram showing the method for assembling RVH1, RVH2, RVH3, and RVH4 oligonucleotides by PCR in the production of the reshaped human anti-HM1.24 antibody H chain.

Figure 6 is a schematic diagram showing the method for producing the human-mouse hybrid anti-HM1.24 antibody H chain V region by PCR.

Figure 7 is a schematic diagram showing a method for producing the mouse-human hybrid anti-HM1.24 antibody H chain V region by PCR.

Figure 8 is a graph showing that version a of the L chain of reshaped human anti-HM1.24 antibody has antigen-binding activity comparable to that of chimeric anti-HM1.24 antibody. Note that -1 and -2 indicate differences in lots.

Figure 9 is a graph showing the antigen-binding activity of reconstituted human anti-HM1.24 antibodies and chimeric HM1.24 antibodies combining L chain version a with H chain version a, b, f, or h.

FIG. 10 is a graph showing the binding activity of reshaped human anti-HM1.24 antibodies and chimeric HM1.24 antibodies that combine L chain version b with H chain versions a, b, f, or h.

FIG. 11 is a graph showing the binding inhibitory activity of reshaped human anti-HM1.24 antibodies and chimeric HM1.24 antibodies that combine L chain version a with H chain version a, b, f, or h.

FIG. 12 is a graph showing the binding inhibitory activity of a reshaped human anti-HM1.24 antibody and a chimeric HM1.24 antibody that combine L chain version b with H chain version a, b, f, or h.

FIG. 13 is a graph showing the antigen-binding activity of reshaped human anti-HM1.24 antibody H chain versions a, b, c, and d and chimeric anti-HM1.24 antibody.

Figure 14 is a graph showing the antigen-binding activity of reshaped human anti-HM1.24 antibody H chain versions a and e and chimeric anti-HM1.24 antibody. Note that -1 and -2 indicate different lots.

FIG. 15 is a graph showing the binding inhibitory activity of reshaped human anti-HM1.24 antibody H chain versions a, c, p, and r and chimeric anti-HM1.24 antibody.

FIG. 16 is a graph showing the antigen-binding activity of human-mouse hybrid anti-HM1.24 antibody, mouse-human hybrid anti-HM1.24 antibody, and chimeric anti-HM1.24 antibody.

FIG. 17 is a graph showing the antigen-binding activity of reshaped human anti-HM1.24 antibody H chain versions a, b, c, and f and chimeric anti-HM1.24 antibody.

Figure 18 is a graph showing the antigen-binding activity of reshaped human anti-HM1.24 antibody H chain versions a and g and chimeric anti-HM1.24 antibody.

Figure 19 is a graph showing the binding inhibitory activity of reshaped human anti-HM1.24 antibody H chain versions a and g and chimeric anti-HM1.24 antibody.

Figure 20 is a graph showing the antigen-binding activity of reshaped human anti-HM1.24 antibody H chain versions h and i and chimeric anti-HM1.24 antibody.

Figure 21 shows the antigen-binding activity of reshaped human anti-HM1.24 antibody H chain versions f, h, and j, and the chimeric anti-HM1.24 antibody.

Figure 22 is a graph showing the binding inhibitory activity of reshaped human anti-HM1.24 antibody H chain versions h and i and chimeric anti-HM1.24 antibody.

Figure 23 is a graph showing the binding inhibitory activity of reshaped human anti-HM1.24 antibody H chain versions f, h, and j and chimeric anti-HM1.24 antibody.

Figure 24 is a graph showing the antigen-binding activity of reshaped human anti-HM1.24 antibody H chain versions h, k, l, m, n, and o, and chimeric anti-HM1.24 antibody.

Figure 25 is a graph showing the antigen-binding activity of reshaped human anti-HM1.24 antibody H chain versions a, h, p, and q and chimeric anti-HM1.24 antibody.

Figure 26 is a graph showing the binding inhibitory activity of reshaped human anti-HM1.24 antibody H chain versions h, k, l, m, n, and o, and chimeric anti-HM1.24 antibody to WISH cells.

Figure 27 is a graph showing the binding inhibitory activity of reshaped human anti-HM1.24 antibody H chain versions a, h, p, and q and chimeric anti-HM1.24 antibody.

Figure 28 is a graph showing the antigen-binding activity of reshaped human anti-HM1.24 antibody H chain versions a, c, p, and r, and chimeric anti-HM1.24 antibody.

Figure 29 is a graph showing that the native humanized anti-HM1.24 antibody (secondary designed antibody) has antigen-binding activity comparable to that of the reshaped human anti-HM1.24 antibody (primary designed antibody).

Figure 30 is a graph showing that the native humanized anti-HM1.24 antibody (secondary designed antibody) has binding inhibitory activity comparable to that of the reshaped human anti-HM1.24 antibody (primary designed antibody).

Figure 31 is a graph showing that the purified reshaped human anti-HM1.24 antibody has antigen-binding activity comparable to that of the chimeric anti-HM1.24 antibody.

Figure 32 is a graph showing that the purified reshaped human anti-HM1.24 antibody has binding inhibitory activity comparable to that of the chimeric anti-HM1.24 antibody.

Figure 33 is a graph showing that the cytotoxic activity of the native humanized anti-HM1.24 antibody (secondary designed antibody) against KPMM2 cells increases with increasing E/T ratio.

Modes of Implementation 1. Natural FR Sequences To generate antibodies that respond to diverse antigens from the limited number of genes that make up antibody variable regions, the body possesses a mechanism for introducing random genetic mutations into antibody variable regions, known as somatic mutations. While this should theoretically result in an extremely diverse FR amino acid sequence, structural analysis of many human antibody FRs whose structures have been elucidated suggests that the positions and types of amino acid residues susceptible to mutation appear to be somewhat limited.

The FRs referred to in the present invention refer to those defined by Kabat, E. A. et al., Sequence of Proteins of Immunological Interest (1991). That is, in the H chain, FR1 is amino acid numbers 1 to 30, FR2 is amino acid numbers 36 to 49, FR3 is amino acid numbers 66 to 94, and FR4 is amino acid numbers 103 to 113. In the L chain, FR1 is amino acid numbers 1 to 23, FR2 is amino acid numbers 35 to 49, FR3 is amino acid numbers 57 to 88, and FR4 is amino acid numbers 98 to 107.

2. From Artificial Human FRs to Natural Human FRs Humanized antibodies (also called reshaped human antibodies) constructed using conventional CDR-grafting methods often contain FR amino acid sequences not found in nature. However, as mentioned above, diverse FR amino acid sequences have already been discovered through somatic mutation, suggesting the possibility of converting FRs containing artificial amino acid residues generated by humanization into naturally occurring human FRs.

The present invention creates humanized antibodies composed of naturally occurring human FRs rather than artificial FRs by further processing humanized antibodies constructed using conventional humanization techniques. By performing a homology search using known human antibody FRs and known databases such as SwissPlot (protein sequence database), GenBank (nucleic acid sequence database), PRF (protein sequence database), PIR (protein sequence database), and GenPept (translated protein sequence from GenBank), it is possible to identify human FRs with perfectly matching amino acid sequences or homologous human FRs.

In the former case, FR substitutions have been made from the perspective of the human FR used as the acceptor for CDR-grafting. However, what was thought to be an artificial FR exists in the natural FR and can be regarded as the acceptor, allowing for the generation of an FR without FR substitutions. In the latter case, by focusing on the amino acid sequence of a human FR that is highly homologous to the artificial FR, it is possible to make appropriate amino acid substitutions in the artificial FR that return it to a natural human antibody, thereby completely matching the natural human FR. This procedure refers to humanizing a CDR-grafted antibody.

In this case, a homology search of the amino acid sequences of human antibodies is performed. The selected human FRs belong to the same subgroup as the human FRs used in CDR-grafting, and highly homologous amino acid sequences can be found. Therefore, the naturally occurring human FRs obtained for each FR will more than fully satisfy the consensus sequence of the subgroup, even if they are derived from different antibodies.

3. Native Sequence Humanized Antibodies The native humanized antibodies obtained in this invention are composed of human antibody FRs known to exist in nature. While FR1 through FR4 may each originate from a different antibody, as mentioned above, a homology search between human antibodies allows the selection of only antibodies belonging to the same subgroup. The FR structures of antibodies within the same subgroup are highly similar, and humanized antibodies based on the consensus sequence of the subgroup have actually been created (Kettleborough, C.A. et al., Protein Eng. (1991) 4, 773-783; Satoh, K. et al., Molec. Immun. (1994) 31, 371-381).

As mentioned above, antibodies are thought to naturally have a wide variety of amino acid sequences due to somatic mutation, and only a portion of these structures have been characterized to date. If the FR sequence of a humanized antibody is not found in nature, it is unclear whether that FR exists in nature. Therefore, when considering antibodies as pharmaceuticals, the creation of CDR-grafted antibodies composed of naturally occurring human FRs provides antibodies with superior properties compared to conventional humanized antibodies, given the original goal of humanization: reduced antigenicity.

4. Novel Method for Constructing Humanized Antibodies The present invention overcomes the problem of humanized antibodies constructed using conventional humanization techniques, namely, the antigenicity caused by artificial FRs not found in nature. It is a technology for constructing humanized antibodies through CDR-grafting, composed of actual naturally occurring human FRs. An artificial FR amino acid sequence refers to an amino acid sequence of an FR that, as a whole, is not found in nature. Furthermore, an artificial amino acid residue contained in an FR refers to an amino acid residue contained in an FR that is not found in nature.

Examples of FR amino acid sequences not found in nature include, for example, FRs in humanized antibodies produced by existing humanization techniques, which have amino acid sequences that replace human amino acid residues with amino acid residues present in the FRs of antibodies derived from non-human mammals, which served as templates for humanization. Alternatively, examples of FRs in humanized antibodies produced by existing humanization techniques, which have amino acid sequences that replace human amino acid residues with amino acid sequences not found in antibodies derived from humans or non-human mammals, include, for example, FRs in humanized antibodies produced by existing humanization techniques, which have amino acid sequences that replace human amino acid residues with amino acid residues not found in antibodies derived from humans or non-human mammals.

The procedure for producing the natural humanized antibody of the present invention is shown below.

First, a human antibody FR for CDR-grafting is selected using conventional methods. Amino acid substitutions are then made in this FR to construct a humanized antibody with biological activity equivalent to or greater than that of a mouse antibody. While this would be considered the final product of a humanized antibody in conventional methods, in this invention it is merely an intermediate product for producing a natural humanized antibody with a native sequence. In this invention, this is referred to as a primary designed antibody.

Next, a homology search is performed for each FR of the primary antibody design. Perfectly matching FRs are already composed of natural FRs. Meanwhile, a list of natural human FRs that do not perfectly match but belong to the same subgroup as the homologous primary antibody FRs is generated. From the list, the optimal natural human FRs that retain the amino acid residues of FRs derived from non-human mammals, such as mice, that were important in constructing the primary antibody design and that are homologous to the primary antibody design are selected.

A homology search for FRs is performed using known databases. Examples of such databases include SwissPlot, GenBank, PRF, PIR, and GenPept. When a homology search is performed using such databases, "FRs homologous to the FRs of the primary antibody" listed in the homology search refer to FRs that share at least 80%, preferably at least 90%, more preferably at least 91%, more preferably at least 92%, more preferably at least 93%, more preferably at least 94%, more preferably at least 95%, more preferably at least 96%, more preferably at least 97%, more preferably at least 98%, and more preferably at least 99% amino acid sequence identity. Protein homology can be determined according to the algorithm described in Wilbur, W. J. and Lipman, D. J., Proc. Natl. Acad. Sci. USA (1983) 80, 726-730.

Non-human mammalian amino acid residues important in the construction of primary antibody designs refer to non-human FR-derived amino acid residues contained in the artificial FR. Such amino acid residues are frequently found in the canonical structure of antibody molecules, in the CDR structure, or in residues that directly interact with antigen molecules. While these residues vary from antibody to antibody, examples include amino acids at positions 71 and 94 of the heavy chain.

As described above, by substituting one or more amino acid residues that differ between the primary design antibody FR and the selected native FR with amino acid residues from native human FRs, the resulting humanized antibody (referred to as a "natural humanized antibody"; secondary design antibody) is composed entirely of native FRs. In this case, preferably, all human FRs belong to the same subgroup, and more preferably, they are derived from the same antibody. Furthermore, it is not necessary for all human FRs to be from the same subgroup, as long as they are reconstituted into an antibody and exhibit appropriate antigen-binding activity; they are not limited to human FRs belonging to the same subgroup. In the present invention, "multiple amino acid residues" refers to two or more amino acid residues in the amino acid sequence, preferably two to ten amino acid residues, more preferably two to five amino acid residues, even more preferably two to four amino acid residues, and even more preferably two to three amino acid residues.

The homology between the artificial FR and the native human FR is at least 80%, preferably at least 90%, more preferably at least 91%, more preferably at least 92%, more preferably at least 93%, more preferably at least 94%, more preferably at least 95%, more preferably at least 96% or more, more preferably at least 97% or more, more preferably at least 98% or more, and more preferably at least 99% or more.

Next, the secondary designed antibody is expressed in an appropriate expression system, such as animal cells, and its antigen-binding activity, etc. is evaluated.

Furthermore, the production method of the present invention can be performed without actually constructing a primary designed antibody. That is, a primary designed antibody can be designed as usual, and a secondary designed antibody can be designed and directly evaluated without evaluating its actual antigen-binding activity. However, in practice, identifying important FR residues may require experimentation, and it is therefore preferable to first experimentally construct a conventional primary designed antibody before preparing a secondary designed antibody.

Specifically, in one embodiment of the present invention, a native humanized antibody of the present invention was produced using mouse anti-HM1.24 antibody (Goto, T. et al., Blood (1994) 84, 1922-1930) as a template.

Genes encoding naturally occurring humanized antibodies designed as described above can be obtained by known methods. For example, several oligonucleotides with overlapping ends corresponding to the DNA encoding the amino acid sequence of the designed naturally occurring humanized antibody are synthesized. These oligonucleotides are used as primers for PCR. Next, a gene encoding the desired naturally occurring humanized antibody can be obtained by PCR using primers that specify both ends of the DNA encoding the amino acid sequence of the designed naturally occurring humanized antibody.

Genes encoding natural humanized antibodies constructed as described above can be expressed by known methods to obtain natural humanized antibodies. When mammalian cells are used, expression can be achieved using DNA or a vector containing a commonly used promoter/enhancer, the antibody gene to be expressed, and a poly(A) signal operably linked to the 3' downstream end of the gene. For example, the promoter/enhancer can be the human cytomegalovirus immediate early promoter/enhancer.

Other promoters/enhancers that can be used for expression of the antibodies of the present invention include viral promoters/enhancers derived from retroviruses, polyoma viruses, adenoviruses, simian virus 40 (SV40), and the like, as well as promoters/enhancers derived from mammalian cells, such as human elongation factor 1α (HEF1α).

For example, when using the SV40 promoter/enhancer, this can be easily achieved by the method of Mulligan et al. (Nature (1979) 277, 108), and when using the HEF1α promoter/enhancer, this can be easily achieved by the method of Mizushima et al. (Nucleic Acids Res. (1990) 18, 5322).

In E. coli, a commonly used promoter, a signal sequence for antibody secretion, and the antibody gene to be expressed can be operably linked to each other for expression. Examples of promoters include the lacZ promoter and the araB promoter. When using the lacZ promoter, the method of Ward et al. (Nature (1998) 341, 544-546; FASEB J. (1992) 6, 2422-2427) can be followed. When using the araB promoter, the method of Better et al. (Science (1988) 240, 1041-1043) can be followed.

When producing antibodies in the periplasm of E. coli, the pelB signal sequence (Lei, S.P. et al. J. Bacteriol. (1987) 169, 4379) can be used as a signal sequence for antibody secretion. After isolating the antibodies produced in the periplasm, the antibody structure is appropriately refolded before use (see, for example, International Patent Application Publication No. WO96-30394 and Japanese Patent Application Publication No. 7-93879).

Replication origins can be derived from SV40, polyomavirus, adenovirus, bovine papillomavirus (BPV), etc. Furthermore, to amplify the gene copy number in the host cell system, the expression vector can also contain selectable markers such as the aminoglycoside transferase (APH) gene, the thymidine kinase (TK) gene, the Escherichia coli xanthine guanine phosphoribosyltransferase (Ecogpt) gene, and the dihydrofolate reductase (dhfr) gene.

Any production system can be used to produce the native humanized antibodies of the present invention. Production systems for antibody production include in vitro and in vivo production systems. In vitro production systems include systems using eukaryotic cells and systems using prokaryotic cells.

Eukaryotic cells include animal, plant, and fungal cell production systems. Examples of animal cells include (1) mammalian cells such as CHO (J. Exp. Med. (1995) 108, 945), COS, myeloma, BHK (baby hamster kidney), HeLa, and Vero; (2) amphibian cells such as Xenopus oocytes (Valle et al., Nature (1981) 291, 358-340); and (3) insect cells such as sf9, sf21, and Tn5. CHO cells that are particularly suitable for use include dhfr-CHO (Proc. Natl. Acad. Sci. USA (1980) 77, 4216-4220), which are CHO cells lacking the DHFR gene, and CHO K-1 (Proc. Natl. Acad. Sci. USA (1968) 60, 1275).

Known plant cells include cells derived from Nicotiana tabacum, which can be cultured as calluses. Known fungal cells include yeasts, such as those from the genus Saccharomyces, e.g., Saccharomyces cerevisiae, and filamentous fungi, such as those from the genus Aspergillus, e.g., Aspergillus niger.

When prokaryotic cells are used, bacterial cells are used as production systems. Examples of bacterial cells include Escherichia coli (E. coli) and Bacillus subtilis. These cells are transformed with genes encoding the native humanized antibodies of the present invention, and the transformed cells are cultured in vitro to obtain native humanized antibodies. Culture is performed according to known methods. For example, DMEM, MEM, RPMI1640, or IMDM can be used as culture media. Serum supplements such as fetal calf serum (FCS) can be used in combination, or serum-free culture can be used. Alternatively, antibodies can be produced in vivo by transferring cells transfected with antibody genes into the peritoneal cavity of an animal.

On the other hand, in vivo production systems include those using animals and plants. Antibody genes are introduced into these animals or plants, and antibodies are produced within the animal or plant's body and then harvested.

When animals are used, there are production systems using mammals and insects.

Mammals that can be used include goats, pigs, sheep, mice, and cows (Vicki Glaser, SPECTRUM Biotechnology Applications, 1993). When mammals are used, transgenic animals can also be used.

For example, an antibody gene can be prepared as a fusion gene by inserting it into a gene encoding a protein specifically produced in milk, such as goat beta-casein. A DNA fragment containing the fusion gene with the antibody gene inserted is injected into a goat embryo, and the embryo is then introduced into a female goat. The antibody of the present invention can be obtained from the milk produced by the transgenic goat or its offspring, which are born to the recipient goat. Appropriate hormones may be administered to the transgenic goat to increase the amount of milk containing the antibody of the present invention produced by the transgenic goat. (Ebert, K.M. et al., Bio/Technology (1994) 12, 699-702)

Furthermore, as an insect, for example, silkworms can be used. When using silkworms, the silkworms are infected with a baculovirus into which the desired antibody gene has been inserted, and the desired antibody is obtained from the body fluids of the silkworms (Susumu, M. et al., Nature (1985) 315, 592-594).

Furthermore, plants, such as tobacco, can be used. When using tobacco, the antibody gene of interest is inserted into a plant expression vector, such as pMON530, and this vector is then introduced into bacteria such as Agrobacterium tumefaciens. Tobacco, such as Nicotiana tabacum, is infected with this bacterium, and the desired antibody is obtained from the tobacco leaves (Julian, K.-C. Ma et al., Eur. J. Immunol. (1994) 24, 131-138).

As mentioned above, in the present invention, the term "host" includes animals and plants that produce the desired native humanized antibody. When producing antibodies in an in vitro or in vivo production system, DNA encoding the antibody heavy and light chains may be separately incorporated into expression vectors and co-transformed into the host, or DNA encoding the heavy and light chains may be incorporated into a single expression vector and transformed into the host (see International Patent Application Publication No. WO 94-11523).

Methods for introducing an expression vector into a host include known methods, such as the calcium phosphate method (Virology (1973) 52, 456-467) and electroporation (EMBO J. (1982) 1, 841-845).

The native humanized antibodies of the present invention produced and expressed as described above can be isolated from the host, either intracellularly or extracellularly, and purified to homogeneity. The isolation and purification of the native humanized antibodies of the present invention can be performed using any isolation and purification method commonly used for protein purification, and is not limited in any way. For example, isolation and purification can be achieved by appropriately selecting and combining methods such as affinity chromatography columns, filters, ultrafiltration, salting out, and dialysis (Antibodies: A Laboratory Manual, Ed Harlow and David Lane, Cold Spring Harbor Laboratory, 1988).

Examples of columns used in affinity chromatography include protein A columns and protein G columns. Examples of supports used in protein A columns include Hyper D, POROS, and Sepharose F.F. (Pharmacia).

Examples of chromatography methods other than affinity chromatography include ion exchange chromatography, hydrophobic chromatography, gel filtration, reversed-phase chromatography, and adsorption chromatography (Strategies for Protein Purification and Characterization: A Laboratory Course Manual. Ed. Daniel R. Marshak et al., Cold Spring Harbor Laboratory Press, 1996).

These chromatographies can be performed using liquid-phase chromatographies such as HPLC and FPLC.

The concentration of the resulting native humanized antibody of the present invention can be measured by absorbance measurement or enzyme-linked immunosorbent assay (ELISA). Specifically, when measuring absorbance, the resulting native humanized antibody is appropriately diluted with PBS and then the absorbance at 280 nm is measured. For example, 1 mg/ml can be calculated as 1.35 OD.

Alternatively, ELISA can be used to measure the antibody concentration. Specifically, 100 ml of goat anti-human IgG antibody (TAGO) diluted to 1 mg/ml with 0.1 M bicarbonate buffer (pH 9.6) is added to a 96-well plate (Nunc) and incubated overnight at 4°C to immobilize the antibody. After blocking, 100 ml of an appropriately diluted native humanized antibody of the present invention or a sample containing the antibody, or 100 ml of human IgG (CAPPEL) at a known concentration as a concentration standard, is added and incubated at room temperature for 1 hour.

After washing, 100 ml of 5000-fold diluted alkaline phosphatase-labeled anti-human IgG (BIOSOURCE) was added and incubated at room temperature for 1 hour. After washing, substrate solution was added and incubated. The absorbance at 405 nm was measured using a Microplate Reader Model 3550 (Bio-Rad) to calculate the concentration of the antibody of the present invention. Alternatively, BIAcore (Pharmacia) can be used to measure antibody concentration.

The antigen-binding activity, binding-inhibitory activity, and neutralizing activity of the natural humanized antibodies of the present invention can be evaluated using commonly known methods. For example, ELISA, EIA (enzyme-linked immunosorbent assay), RIA (radioimmunoassay), or fluorescent antibody techniques can be used to measure the activity of the natural humanized antibodies of the present invention. BIAcore (Pharmacia) can be used to evaluate the activity of the above antibodies.

The native humanized antibodies of the present invention may be antibody fragments or modified antibodies. For example, antibody fragments include Fab, F(ab')2, Fv, or single-chain Fv (scFv). scFv has a structure in which the Fvs of the heavy and light chains are linked via an appropriate linker.

To obtain these antibody fragments, antibodies can be treated with enzymes such as papain or pepsin to generate antibody fragments, or genes encoding these antibody fragments can be constructed, introduced into expression vectors, and then expressed in appropriate host cells (see, for example, Co, M.S. et al., J. Immunol. (1994) 152, 2968-2976; Better, M. & Horwitz, A.H., Methods in Enzymology (1989) 178, 476-496, Academic Press, Inc.; Plueckthun, A. & Skerra, A., Methods in Enzymology (1989) 178, 476-496, Academic Press, Inc.; Lamoyi, E., Methods in Enzymology (1989) 121, 652-663; Rousseaux, J. et al., Methods in Enzymology (1989) 121, 652-663). al., Methods in Enzymology (1989) 121, 663- 669, Bird, R. E. et al., TIBTECH (1991) 9, 132-137).

An scFv can be obtained by linking the heavy chain V region and light chain V region of an antibody (see International Patent Application Publication No. WO 88-09344). In this scFv, the heavy chain V region and light chain V region are linked via a linker, preferably a peptide linker (Huston, J.S. et al., Proc. Natl. Acad. Sci. U.S.A. (1988) 85, 5879-5883). The heavy chain V region and light chain V region in the scFv can be derived from any of the antibodies described above. The peptide linker linking the V regions can be, for example, a single-chain peptide consisting of 12-19 amino acid residues (see U.S. Patent No. US 5,525,491).

DNA encoding the scFv can be obtained by using DNA encoding the antibody's H chain or H chain V region and DNA encoding the L chain or L chain V region as templates, amplifying the DNA portion encoding the desired amino acid sequence from those sequences by PCR using a primer pair that defines both ends of the DNA portion, and then further amplifying DNA encoding a peptide linker portion and a primer pair that defines both ends of the DNA portion so that it can be linked to the H chain and L chain, respectively.

Furthermore, once DNA encoding an scFv has been prepared, an expression vector containing the DNA and a host transformed with the expression vector can be obtained according to standard methods. Furthermore, scFv can be obtained using the host according to standard methods.

These antibody fragments can be produced in a host by obtaining and expressing the genes as described above. The term "antibody" as used in the claims of this application also encompasses these antibody fragments.

Modified antibodies may also be conjugated with various molecules, such as polyethylene glycol (PEG). The term "antibody" as used in the claims of this application encompasses these modified antibodies. Such modified antibodies can be obtained by chemically modifying the antibody. These methods have already been established in the art.

The native humanized antibodies of the present invention can be administered orally or parenterally, either systemically or locally. Parenteral administration can be selected, for example, by intravenous injection such as drip infusion, intramuscular injection, intraperitoneal injection, subcutaneous injection, or pulmonary administration. The administration method can be selected appropriately depending on the patient's age and symptoms.

The native humanized antibodies of the present invention are administered to a patient already suffering from a disease in an amount sufficient to cure or at least partially arrest the symptoms of the disease.

For example, an effective dose may be selected from the range of 0.01 mg to 100 mg per kg of body weight per administration. Alternatively, a dose of 1-1000 mg, preferably 5-50 mg, per patient may be selected. However, the native humanized antibodies of the present invention are not limited to these doses. The administration period may be appropriately selected depending on the patient's age and symptoms.

Depending on the route of administration, the native humanized antibodies of the present invention may contain pharmaceutically acceptable carriers and additives. Examples of such carriers and additives include water, pharmaceutically acceptable organic solvents, collagen, polyvinyl alcohol, polyvinylpyrrolidone, carboxyvinyl polymers, sodium carboxymethylcellulose, sodium polyacrylate, sodium alginate, water-soluble dextran, sodium carboxymethyl starch, pectin, methylcellulose, ethylcellulose, xanthan gum, gum arabic, casein, gelatin, agar, diglycerin, propylene glycol, polyethylene glycol, petrolatum, paraffin, stearyl alcohol, stearic acid, human serum albumin (HSA), mannitol, sorbitol, lactose, and surfactants acceptable as pharmaceutical additives. The additives used may be selected from the above list or in combination depending on the dosage form, but are not limited to these.

<Reference Examples> Before describing the present invention in detail through examples, the following reference examples will serve as a basis for the present invention.

Reference Example 1. Cloning of cDNA Encoding Mouse Anti-HM1.24 Antibody Variable Regions 1. Isolation of Messenger RNA (mRNA) mRNA was isolated from 2 x ¹⁰ hybridoma cells (FERM B P-5233) producing mouse anti-HM1.24 antibody using Fast Track mRNA Isolation Kit Version 3.2 (Invitrogen) according to the instructions provided with the kit.

2. PCR amplification of genes encoding antibody variable regions PCR was performed using a Thermal Cycler (Perkin Elmer Cetus).

2-1. Amplification and Fragmentation of the Gene Encoding the Mouse Light Chain Variable Region Single-stranded cDNA was synthesized from the isolated mRNA using the AMV Reverse Transcriptase First-Strand cDNA Synthesis Kit (Life Sciences) and used for PCR. The primers used for PCR were MKV (Mouse Kappa Variable) primers shown in SEQ ID NOS: 29–39, which hybridize with the mouse kappa light chain leader sequence (Jones, S.T. et al., Bio/Technology, 9, 88–89, (1991)).

A 100 μl PCR solution containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 0.1 mM dNTPs (dATP, dGTP, dCTP, dTTP), 1.5 mM _MgCl2 , 5 units of DNA polymerase AmpliTaq (Perkin Elmer Cetus), 0.25 mM MKV primers shown in SEQ ID NOs:29-39 and 3 mM MKC primer shown in SEQ ID NO:40, and 100 ng of single-stranded cDNA was covered with 50 μl of mineral oil and heated at an initial temperature of 94°C for 3 minutes, followed by 94°C for 1 minute, 55°C for 1 minute, and 72°C for 1 minute, in that order. This temperature cycle was repeated 30 times, and the reaction mixture was then further incubated at 72°C for 10 minutes. The amplified DNA fragment was purified with low-melting-point agarose (Sigma) and digested with XmaI (New England Biolabs) and SalI (Takara Shuzo) at 37°C.

2-2. Amplification and Fragmentation of cDNA Encoding the Mouse H Chain V Region The gene encoding the mouse H chain V region was amplified by the 5'-RACE method (Rapid Amplification of cDNA ends; Frohman, M.A. et al., Proc. Natl. Acad. Sci. USA, 85, 8998-9002, (1988); Edwards, J.B.D.M. et al., Nucleic Acids Res., 19, 5227-5232, (1991)). cDNA was synthesized using primer P1 (SEQ ID NO: 63), which specifically hybridizes to the mouse IgG2a constant region. The cDNA encoding the mouse H chain V region was then amplified using the 5'-AmpIiFINDER RACEKIT (CLONETEC) with primer MHC2a (SEQ ID NO: 64), which specifically hybridizes to the mouse IgG2a constant region, and the anchor primer (SEQ ID NO: 101) provided with the kit. The amplified DNA fragment was purified with low-melting-point agarose (Sigma) and digested with EcoRI (Takara Shuzo) and XmaI (New England Biolabs) at 37°C.

3. Ligation and transformation The DNA fragment containing the gene encoding the mouse kappa L chain V region prepared as described above was ligated to a pUC19 vector prepared by digestion with SalI and XmaI in a reaction mixture containing 50 mM Tris-HCl (pH 7.6), ₁₀ mM MgCl2, 10 mM dithiothreitol, 1 mM ATP, 50 mg/ml polyethylene glycol (8000), and 1 unit of T4 DNA ligase (GIBCO-BRL) at 16°C for 2.5 hours.

Similarly, a DNA fragment containing a gene encoding the mouse H chain V region was digested with EcoRI and XmaI to prepare a pUC19 vector, which was then reacted at 16°C for 3 hours and ligated.

Next, 10 μl of the ligation mixture was added to 50 μl of competent E. coli DH5α cells, and the cells were incubated on ice for 30 minutes, at 42°C for 1 minute, and again on ice for 1 minute. 400 μl of 2xYT medium (Molecular Cloning: A Laboratory Manual, Sambrook et al., Cold Spring Harbor Laboratory Press, (1989)) was then added, and the cells were incubated at 37°C for 1 hour. The E. coli cells were then plated on 2xYT agar medium (Molecular Cloning: A Laboratory Manual, Sambrook et al., Cold Spring Harbor Laboratory Press, (1989)) containing 50 μg/ml ampicillin and incubated overnight at 37°C to obtain E. coli transformants.

The transformants were cultured overnight at 37°C in 10 ml of 2xYT medium containing 50 μg/ml ampicillin, and plasmid DNA was prepared from the culture by the alkaline method (Molecular Cloning: A Laboratory Manual, Sambrook et al., Cold Spring Harbor Laboratory Press, (1989)).

The resulting plasmid containing the gene encoding the mouse kappa L chain V region derived from the hybridoma producing the anti-HM1.24 antibody was designated pUCHMVL9. The resulting plasmid containing the gene encoding the mouse H chain V region derived from the hybridoma producing the anti-HM1.24 antibody was designated pUCHMVHR16.

Reference Example 2. Determination of DNA sequence The nucleotide sequence of the cDNA coding region in the above-mentioned plasmid was determined using an automated DNA sequencer (manufactured by Applied Biosystems Inc.) and a Taq Dye Deoxy Terminator Cycle Sequencing Kit (manufactured by Applied Biosystems Inc.) according to the protocol specified by the manufacturer.

The nucleotide sequence of the gene encoding the L chain V region of the mouse anti-HM1.24 antibody contained in the plasmid pUCHMVL9 is shown in SEQ ID NO: 1. The nucleotide sequence of the gene encoding the H chain V region of the mouse anti-HM1.24 antibody contained in the plasmid pUCHMVHR16 is shown in SEQ ID NO: 3.

Reference Example 3. CDR Determination The overall structures of the V regions of light and heavy chains are similar to each other, with four framework regions connected by three hypervariable regions, i.e., complementarity-determining regions (CDRs). The amino acid sequences of the framework regions are relatively well conserved, while the amino acid sequences of the CDR regions are highly variable (Kabat, EA, et al., "Sequences of Proteins of Immunological Interest," U.S. Dept. Health and Human Services, 1983).

Based on these findings, the amino acid sequences of the variable regions of the anti-HM1.24 antibody were compared with the amino acid sequence database of antibodies prepared by Kabat et al. to examine the homology, thereby determining the CDR regions as shown in Table 1.

Table 1 Reference Example 4. Confirmation of Expression of Cloned cDNA (Preparation of Chimeric Anti-HM1.24 Antibody ) 1. Preparation of Expression Vector To prepare a vector expressing the chimeric anti-HM1.24 antibody, the cDNA clones pUCHMVL9 and pUCHMV HR16 encoding the L-chain and H-chain V regions of the mouse anti-HM1.24 antibody, respectively, were modified by PCR and then introduced into the HEF expression vector (see International Patent Application Publication No. WO 92-19759).

The reverse primer ONS-L722S (SEQ ID NO: 65) for the L chain V region and the reverse primer VHR16S (SEQ ID NO: 66) for the H chain V region were designed to hybridize to the DNA encoding the beginning of the leader sequence of each V region and contain a Kozak consensus sequence (Kozak, M. et al., J. Mol. Biol., 196, 947-950, (1987)) and a HindIII restriction enzyme recognition site. The forward primer VL9A (SEQ ID NO: 67) for the L chain V region and the forward primer VHR16A (SEQ ID NO: 68) for the H chain V region were designed to hybridize to the DNA sequence encoding the end of the J region and contain a splice donor sequence and a BamHI restriction enzyme recognition site.

A 100 μl PCR reaction mixture containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 0.1 mM dNTPs, _1.5 mM MgCl2, 100 pmoles of each primer, 100 ng of template DNA (pUCHMVL9 or pUCHMVHR16), and 5 units of Ampli Taq enzyme was covered with 50 μl of mineral oil and, after initial denaturation at 94°C, 30 cycles of 94°C for 1 minute, 55°C for 1 minute, and 72°C for 1 minute were performed, followed by a final incubation at 72°C for 10 minutes.

The PCR products were purified using a 1.5% low-melting-point agarose gel, digested with HindIII and BamHI, and cloned into HEF-VL-gκ (the L chain V region) and HEF-VH-gγ1 (the H chain V region). After DNA sequencing, plasmids containing DNA fragments with the correct DNA sequences were designated HEF-1.24L-gκ and HEF-1.24H-gγ1, respectively.

The regions encoding the variable regions of the HEF-1.24L-gκ and HEF-1.24H-gγ1 plasmids were digested with the restriction enzymes Hind III and BamHI to produce restriction fragments, which were then inserted into the Hind III and BamHI sites of the plasmid vector pUC19 and designated pUC19-1.24L-gκ and pUC19-1.24H-gγ1, respectively.

The E. coli strains containing the respective plasmids pUC19-1.24L-gκ and pUC19-1.24H-gγ1 are designated Escherichia coli DH5α (pUC19-1.24L-gκ) and Escherichia coli DH5α (pUC19-1.24H-gγ1), respectively, and were deposited under the Budapest Treaty on August 29, 1996, with the National Institute of Bioscience and Human-Technology, Agency of Industrial Science and Technology (1-1-3 Higashi, Tsukuba, Ibaraki Prefecture, Japan) as accessions FERM BP-5646 and FERM BP-5644, respectively.

2. Transfection into COS-7 Cells To observe the transient expression of the chimeric anti-HM1.24 antibody, the expression vector was tested in COS-7 (ATCC CRL-1651) cells. HEF-1.24L-gκ and HEF-1.24H-gγ1 were cotransfected into COS-7 cells by electroporation using a Gene Pulser apparatus (BioRad). Each DNA (10 μg) was added to a 0.8 ml aliquot of 1 × ¹⁰ cells/ml in PBS, and pulsed at 1500 V and 25 μF.

After a 10-minute recovery period at room temperature, the electroporated cells were added to 30 ml of DMEM medium (GIBCO) containing 10% γ-globulin-free fetal bovine serum. After 72 hours of incubation in _{a CO2} incubator BNA120D (TABAI), the culture supernatant was collected and centrifuged to remove cell debris, and then used in the following experiments.

The antigen-binding activity of the chimeric anti-HM1.24 antibody was analyzed by flow cytometry (FCM) using KPMM2 cells. 4.7 x ¹⁰ KPMM2 cells (JP Patent Publication No. 7-236475) were washed with PBS(-) and then incubated with 50 μl of the culture medium from COS-7 cells producing the chimeric anti-HM1.24 antibody and 50 μl of FACS buffer (PBS(-) containing 2% fetal bovine serum and 0.1% sodium azide), or with 5 μl of 500 μg/ml purified mouse anti-HM1.24 antibody and 95 μl of FACS buffer, on ice for 1 hour.

As controls, 50 μl of 2 μg/ml chimeric SK2 (International Patent Application Publication No. WO94-28159) and 50 μl of FACS buffer were added instead of the chimeric anti-HM1.24 antibody-producing COS cell culture medium, or 5 μl of 500 μg/ml purified mouse IgG2aκ (UP C10) (CAPPEL) and 95 μl of FACS buffer were added instead of the purified mouse anti-HM1.24 antibody, and the cells were incubated in the same manner. After washing with FACS buffer, 100 μl of 25 μg/ml FITC-labeled goat anti-human antibody (GAH) (CAPPEL) or 10 μg/ml FITC-labeled goat anti-mouse antibody (GAM) (Becton Dickinson) was added and incubated on ice for 30 minutes. After washing with FACS buffer, the cells were suspended in 1 ml of FACS buffer, and the fluorescence intensity of each cell was measured using a FACScan (Becton Dickinson).

As shown in Figure 1, the fluorescence intensity peak shifted to the right in cells treated with chimeric anti-HM1.24 antibody compared to the control, similar to the case of treatment with mouse anti-HM1.24 antibody, demonstrating that chimeric anti-HM1.24 antibody bound to KPMM2 cells. This confirmed that the cloned cDNA encoded the V region of mouse anti-HM1.24 antibody.

Reference Example 5. Establishment of a CHO cell line stably producing chimeric anti-HM1.24 antibody 1. Construction of a chimeric H chain expression vector The aforementioned plasmid HEF-1.24H-gγ1 was digested with restriction enzymes PvuI and BamHI, and an approximately 2.8 kb fragment containing DNA encoding the EFI promoter and the mouse anti-HM1.24 antibody H chain V region was purified using a 1.5% low-melting-point agarose gel. Next, the expression vector used in the human H chain expression vector DHFR-ΔE-RVh-PMlf (see International Patent Application Publication No. WO92/19759), which contains the DHFR gene and a gene encoding the human H chain constant region, was digested with PvuI and BamHI to prepare an approximately 6 kb fragment. This DNA fragment was inserted into this fragment to construct the chimeric anti-HM1.24 antibody H chain expression vector DHFR-ΔE-HEF-1.24H-gγ1.

2. Gene Transfer into CHO Cells To establish a stable production system for the chimeric anti-HM1.24 antibody, the expression vectors HEF-1.24L-gκ and DHFR-ΔE-HEF-1.24H-gγ1, which had been linearized by PvuI digestion, were simultaneously transfected into CHO cells DXB11 (provided by the Medical Research Council Collaboration Center) by electroporation under the same conditions as described above (transfection of COS-7 cells).

3. Gene Amplification with MTX Transfected CHO cells were cultured in nucleoside-free α-MEM medium (GIBCO-BRL) supplemented with 500 μg/ml G418 (GIBCO-BRL) and 10% fetal bovine serum. Only CHO cells transfected with both the L-chain and H-chain expression vectors survived and were selected. Next, 10 nM MTX (Sigma) was added to the culture medium, and clones that grew were selected for high production of chimeric anti-HM1.24 antibody. Clone #8-13, which exhibited a chimeric antibody production efficiency of approximately 20 μg/ml, was selected as the chimeric anti-HM1.24 antibody-producing cell line.

Reference Example 6. Production of chimeric anti-HM1.24 antibody Chimeric anti-HM1.24 antibody was produced by the following method: The chimeric anti-HM1.24 antibody-producing CHO cells were continuously cultured for 30 days in a high-density cell culture device, Veraxsystem 20 (CELLEX BIOSCIENCE Inc.), using Iscove's Modified Dulbecco's Medium (GIBCO-BRL) containing 5% γ-globulin-free newborn calf serum (GIBCO-BRL).

Culture medium was harvested on days 13, 20, 23, 26, and 30 after the initiation of cultivation. The medium was filtered using a SARTOBRAN pressure filtration unit (Sartorius). The chimeric anti-HM1.24 antibody was affinity-purified using the Afi-Prep System (NGK Insulators) and a Super Protein A column (100 ml bed volume, NGK Insulators) according to the manufacturer's instructions. PBS(-) was used as the adsorption/wash buffer, and 0.1 M sodium citrate buffer (pH 3) was used as the elution buffer. The pH of the eluted fraction was immediately adjusted to approximately 7.4 by adding 1 M Tris-HCl (pH 8.0). The antibody concentration was determined by measuring the absorbance at 280 nm, with 1 mg/ml being 1.35 OD.

Reference Example 7 Measurement of activity of chimeric anti-HM1.24 antibody The chimeric anti-HM1.24 antibody was evaluated for its binding inhibitory activity as described below.

1. Measurement of Binding Inhibitory Activity 1-1. Preparation of Biotin-Labeled Anti-HM1.24 Antibody Mouse anti-HM1.24 antibody was diluted to 4 mg/ml in 0.1 M bicarbonate buffer, and 4 μl of 50 mg/ml Biotin-N-hydroxy succinimide (EY LABS Inc.) was added. The reaction was allowed to proceed at room temperature for 3 hours. The reaction was then terminated by adding 1.5 ml of 0.2 M glycine solution and incubating at room temperature for 30 minutes. The biotinylated IgG fraction was then isolated using a PD-10 column (Pharmacia Biotech).

Measurement of binding inhibitory activity. The binding inhibitory activity of biotin-labeled mouse anti-HM1.24 antibody was measured by cell-ELISA using the human amniotic cell line WISH cells (ATCC CCL 25). Cell-ELISA plates were prepared as follows: 100 μl of a WISH cell suspension adjusted to 4 × ¹⁰ cells/ml in RPMI 1640 medium containing 10% fetal bovine serum was added to a 96-well plate. After overnight incubation, the cells were washed twice with PBS(-) and fixed with 0.1% glutaraldehyde (Nacalai Tesque).

After blocking, 50 μl of affinity-purified chimeric anti-HM1.24 antibody or mouse anti-HM1.24 antibody was serially diluted and added to each well. At the same time, 50 μl of 2 μg/ml biotin-labeled mouse anti-HM1.24 antibody was added. After incubation at room temperature for 2 hours and washing, peroxidase-labeled streptavidin (DAKO) was added. After incubation at room temperature for 1 hour and washing, the substrate solution was added, followed by incubation. The reaction was stopped with 50 μl of 6N sulfuric acid, and the absorbance at 490 nm was measured using a Microplate Reader Model 3550 (Bio-Rad).

As a result, as shown in Figure 2, the chimeric anti-HM1.24 antibody exhibited binding inhibitory activity against the biotin-labeled mouse anti-HM1.24 antibody equivalent to that of the mouse anti-HM1.24 antibody, demonstrating that the chimeric antibody has the same V region as the mouse anti-HM1.24 antibody.

Reference Example 8 Measurement of ADCC activity of chimeric anti-HM1.24 antibody ADCC (antibody-dependent cellular cytotoxicity) activity was measured according to the method described in Current protocols in immunology, Chapter 7, Immunologic studies in humans, Editor, John E., Coligan et al., John Wiley & Sons, Inc., 1993.

1. Preparation of effector cells. Mononuclear cells were isolated from peripheral blood and bone marrow of healthy donors and multiple myeloma patients by density centrifugation. An equal volume of PBS(-) was added to peripheral blood and bone marrow from healthy donors and multiple myeloma patients, layered on Ficoll (Pharmacia)-Conrey (Daiichi Pharmaceutical) (specific gravity 1.077), and centrifuged at 400 g for 30 minutes. The mononuclear cell layer was separated, washed twice with RPMI 1640 (Sigma) containing 10% fetal bovine serum (Witaker), and then cultured in the same medium to a cell density of 5 × ¹⁰ /ml.

2. Preparation of target cells Human myeloma cell line RPMI 8226 (ATCC CCL 155) was radiolabeled by incubating with 0.1 mCi of ^51Cr -sodium chromate in RPMI 1640 (Sigma) containing 10% fetal bovine serum (Witaker) for 60 minutes at 37°C. After radiolabeling, the cells were washed three times with Hanks balanced salt solution (HBSS) and adjusted to a concentration of 2 x ¹⁰⁵ /ml.

3. ADCC Assay 50 μl of radiolabeled target cells (2 × 10 ⁵ /ml) and 50 μl of affinity-purified chimeric anti-HM1.24 antibody, mouse anti-HM1.24 antibody, or control human IgG1 (Serotec) (1 μg/ml) were added to a 96-well U-bottom plate (Corning) and incubated at 4°C for 15 minutes.

Then, 100 μl of 5 × 10 ⁶ /ml effector cells were added and cultured in a CO 2 incubator for 4 hours at effector cell (E) to target cell (T) ratios (E:T) of 0:1, 5:1, 20:1, or 50:1.

100 μl of the supernatant was removed and the radioactivity released into the culture supernatant was measured using a gamma counter (ARC361, Aloka). 1% NP-40 (BRL) was used to measure maximum released radioactivity. The cytotoxicity (%) was calculated as (A-C)/(B-C) x 100. A represents the radioactivity (cpm) released in the presence of antibody, B represents the radioactivity (cpm) released by NP-40, and C represents the radioactivity (cpm) released in the culture medium alone without antibody.

As shown in Figure 3, when chimeric anti-HM1.24 antibody was added, cytotoxic activity increased with increasing E:T ratio compared to human control IgG1, demonstrating that this chimeric anti-HM1.24 antibody possesses ADCC activity. Furthermore, no cytotoxic activity was observed with the addition of mouse anti-HM1.24 antibody, indicating that the Fc portion of a human antibody is required for ADCC activity when effector cells are of human origin.

Reference Example 9. Construction of reshaped human anti-HM1.24 antibody 1. Design of reshaped human anti-HM1.24 antibody V region To construct a reshaped human antibody in which the CDRs of a mouse monoclonal antibody are grafted onto a human antibody, it is desirable that there be high homology between the FRs of the mouse monoclonal antibody and the FRs of the human antibody. Therefore, the V regions of the L and H chains of the mouse anti-HM1.24 antibody were compared with the V regions of all known antibodies whose structures have been elucidated using the Protein Data Bank.

The L chain V region of the mouse anti-HM1.24 antibody is most similar to the consensus sequence of human L chain V region subgroup IV (HSG I V), with 66.4% identity, while it shows 56.9%, 55.8%, and 61.5% identity with HSG I, HSG II, and HSG III, respectively.

Comparison of the L chain V region of the mouse anti-HM1.24 antibody with known human antibody L chain V regions revealed 67.0% homology to the human L chain V region REI, a member of subgroup I of human L chain V regions. Therefore, the REI FR was used as the starting material for constructing the reshaped human anti-HM1.24 antibody L chain V region.

Version a of the L chain V region of reshaped human anti-HM1.24 antibody was designed. In this version, the human FRs were identical to the REI-based FRs present in the reshaped human CAMPATH-1H antibody (see Riechmann, L. et al., Nature 322, 21-25, (1988); FRs contained in version a of the L chain V region of reshaped human PM-1, described in International Patent Application Publication No. WO92-19759), and the murine CDRs were identical to the CDRs in the L chain V region of the murine anti-HM1.24 antibody.

The H chain V region of the mouse anti-HM1.24 antibody is most similar to the HSG I consensus sequence of the human H chain V region, with 54.7% identity. Meanwhile, it shows 34.6% and 48.1% identity with HSG II and HSG III, respectively. Comparison of the H chain V region of the mouse anti-HM1.24 antibody with known human antibody H chain V regions reveals that FR1 through FR3 are highly similar to the H chain V region of the human antibody HG3, a member of subgroup I of human H chain V regions (Rechavi, G. et al., Proc. Nat. Acad. Sci. USA 80, 855-859), with 67.3% identity.

For this reason, the FR4 of human antibody HG3 was used as the starting material for constructing the H chain V region of reshaped human anti-HM1.24 antibody. However, since the amino acid sequence of FR4 of human antibody HG3 has not been described, we used the amino acid sequence of FR4 of human antibody JH6 (Ravetch, J.V. et al., Cell, 27, 583-591), which shows the highest homology with the FR4 of the H chain of mouse anti-HM1.24 antibody. FR4 of JH6 has the same amino acid sequence as the FR4 of the H chain of mouse anti-HM1.24 antibody, except for one amino acid.

In the first version a of the H chain V region of the reshaped human anti-HM1.24 antibody, FR1 to FR3 were identical to FR1 to FR3 of human HG3, except that the amino acids at positions 30 in human FR1 and 71 in human FR3 were identical to those in the mouse anti-HM1.24 antibody, and the CDRs were identical to those in the H chain V region of the mouse anti-HM1.24 antibody.

2. Construction of the Reshaped Human Anti-HM1.24 Antibody L Chain V Region The reshaped human anti-HM1.24 antibody L chain was constructed by CDR grafting using PCR. This method is shown schematically in Figure 4. Eight PCR primers were used to construct the reshaped human anti-HM1.24 antibody (version a) with FRs derived from the human antibody REI. External primers A (SEQ ID NO: 69) and H (SEQ ID NO: 70) were designed to hybridize with the DNA sequence of the HEF expression vector HEF-VL-gκ.

CDR-grafting primers L1S (SEQ ID NO: 71), L2S (SEQ ID NO: 72), and L3S (SEQ ID NO: 73) have sense DNA sequences, and CDR-grafting primers LIA (SEQ ID NO: 74), L2A (SEQ ID NO: 75), and L3A (SEQ ID NO: 76) have anti-sense DNA sequences and complementary DNA sequences (20-23 bp) to the 5'-terminal DNA sequences of primers L1S, L2S, and L3S, respectively.

Four reactions, A-L1A, L1S-L2A, L2S-L3A, and L3S-H, were performed in the first PCR step, and each PCR product was purified. The four PCR products from the first PCR were assembled by their own complementarity (see International Patent Application Publication No. WO 92-19759).

Next, external primers A and H were added to amplify the full-length DNA encoding the reshaped human anti-HM1.24 antibody L chain V region (second PCR). In this PCR, plasmid HEF-RVL-M21a (see International Patent Application Publication No. WO95-14041), encoding version a of the reshaped human ONS-M21 antibody L chain V region based on the FR from human antibody REI, was used as a template.

The first PCR step consisted of a PCR mixture containing 10 mM Tri-HCl (pH 8.3), 50 mM KCl, 0.1 mM dNTPs, _1.5 mM MgCl2, 100 ng of template DNA, 100 pmoles of each primer, and 5 μl of AmpliTaq. Each PCR tube was coated with 50 μl of mineral oil. After initial denaturation at 94°C, the reaction was cycled at 94°C for 1 minute, 55°C for 1 minute, and 72°C for 1 minute, followed by a 10-minute incubation at 72°C.

The PCR products A-L1A (215 bp), L1S-L2A (98 bp), L2S-L3A (140 bp), and L3S-H (151 bp) were purified using a 1.5% low-melting-point agarose gel and assembled in a second PCR. In the second PCR, 98 μl of PCR mixture containing 1 μg of each first PCR product and 5 μl of AmpliTaq was incubated for two cycles at 94°C for 2 minutes, 55°C for 2 minutes, and 72°C for 2 minutes, and then 100 pmoles of each outer primer (A and H) was added. The PCR tubes were covered with 50 μl of mineral oil, and 30 cycles of PCR were performed under the same conditions as above.

The 516-bp DNA fragment generated by the second PCR was purified from a 1.5% low-melting-point agarose gel and digested with BamHI and HindIII. The resulting DNA fragment was cloned into the HEF expression vector HEF-VL-gκ. After DNA sequencing, the plasmid containing the DNA fragment encoding the correct amino acid sequence of the reshaped human anti-HM1.24 antibody L chain V region was designated HEF-RVLa-AHM-gκ. The amino acid and nucleotide sequences of the L chain V region contained in this plasmid HEF-RVLa-AHM-gκ are shown in SEQ ID NO: 11.

Version b of the reshaped human anti-HM1.24 antibody L chain V region was generated by PCR-based mutagenesis. Mutagenic primers FTY-1 (SEQ ID NO: 77) and FTY-2 (SEQ ID NO: 78) were designed to mutate phenylalanine at position 71 to tyrosine.

After amplification using the above primers and the plasmid HEF-RVLa-AHM-gκ as a template, the final product was purified and digested with BamHI and HindIII. The resulting DNA fragment was cloned into the HEF expression vector HEF-VL-gκ to obtain the plasmid HEF-RVLb-AHM-gκ.

The amino acid sequence and nucleotide sequence of the L chain V region contained in this plasmid HEF-RVLb-AHM-gκ are shown in SEQ ID NO: 13.

3. Construction of the Reshaped Human Anti-HM1.24 Antibody H Chain V Region 3-1. Construction of Reshaped Human Anti-HM1.24 Antibody H Chain V Region Versions a-e DNA encoding the reshaped human anti-HM1.24 antibody H chain V region was designed as follows. The DNA sequence encoding FRs 1-3 of human antibody HG3 and FR4 of human antibody JH6 was linked to the DNA sequence encoding the CDRs of the mouse anti-HM1.24 antibody H chain V region to design a full-length DNA encoding the reshaped human anti-HM1.24 antibody H chain V region.

Next, a HindIII recognition site/Kozak consensus sequence and a BamHI recognition site/splice donor sequence were added to the 5' and 3' ends of this DNA sequence, respectively, to allow insertion into a HEF expression vector.

The DNA sequence thus designed was divided into four oligonucleotides, and then computer analysis was performed to determine secondary structures within the oligonucleotides that could interfere with their assembly. The four oligonucleotide sequences, RVH1 to RVH4, are shown in SEQ ID NOs: 79 to 82. These oligonucleotides are 119 to 144 bases in length and have overlapping regions of 25 to 26 bp. Among the oligonucleotides, RVH2 (SEQ ID NO: 80) and RVH4 (SEQ ID NO: 82) have sense DNA sequences, and the other oligonucleotides, RVH1 (SEQ ID NO: 79) and RVH3 (SEQ ID NO: 81), have antisense DNA sequences. The method for assembling these four oligonucleotides by PCR is shown in the diagram (see Figure 5).

A 98 μl PCR mixture containing 100 ng of each of the four oligonucleotides and 5 μl of AmpliTaq was subjected to an initial denaturation at 94°C for 2 minutes, followed by two cycles of incubation at 94°C for 2 minutes, 55°C for 2 minutes, and 72°C for 2 minutes. After adding 100 pmoles each of RHP1 (SEQ ID NO: 83) and RHP2 (SEQ ID NO: 84) as external primers, the PCR tube was covered with 50 μl of mineral oil and subjected to an initial denaturation at 94°C for 1 minute, followed by 38 cycles of 94°C for 1 minute, 55°C for 1 minute, and 72°C for 1 minute, followed by incubation at 72°C for 10 minutes.

The 438-bp DNA fragment was purified using a 1.5% low-melting-point agarose gel, digested with HindIII and BamHI, and then cloned into the HEF expression vector HEF-VH-gγ1. After DNA sequencing, the plasmid containing the DNA fragment encoding the correct amino acid sequence of the H chain V region was designated HEF-RVHa-AHM-gγ1. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid, HEF-RVHa-AHM-gγ1, are shown in SEQ ID NO: 11.

Versions b, c, d, and e of the reshaped human anti-HM1.24 antibody H chain V region were constructed as follows.

Version b was generated by PCR amplification using the mutagenic primers BS (SEQ ID NO: 85) and BA (SEQ ID NO: 86), designed to mutate arginine to lysine at position 66, and plasmid HEF-RVHa-AHM-gγ1 as template DNA. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid, HEF-RVHb-AHM-gγ1, are shown in SEQ ID NO: 17.

Version c was generated by PCR amplification using the mutagenic primers CS (SEQ ID NO: 87) and CA (SEQ ID NO: 88), designed to mutate threonine to lysine at position 73, and plasmid HEF-RVHa-AHM-gγ1 as template DNA. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid, HEF-RVHc-AHM-gγ1, are shown in SEQ ID NO: 19.

Version d was obtained using the mutagenic primers DS (SEQ ID NO: 89) and DA (SEQ ID NO: 90), which were designed to mutate arginine at position 66 to lysine and threonine at position 73 to lysine, and plasmid HEF-RVHa-AHM-gγ1 as template DNA to obtain plasmid HEF-RVHd-AHM-gγ1. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid HEF-RVHd-AHM-gγ1 are shown in SEQ ID NO: 21.

Version e was obtained by amplifying the HEF-RVHe-AHM-gγ1 plasmid as template DNA using mutagenic primers ES (SEQ ID NO: 91) and EA (SEQ ID NO: 92), which were designed to mutate valine at position 67 to alanine and methionine at position 69 to leucine. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid HEF-RVHe-AHM-gγ1 are shown in SEQ ID NO: 23.

3-2. Construction of Heavy Chain Hybrid V Regions Two types of heavy chain hybrid V regions were constructed. One was a mouse-human hybrid anti-HM1.24 antibody in which the amino acid sequences of FR1 and FR2 were derived from mouse anti-HM1.24 antibody and the amino acid sequences of FR3 and FR4 were derived from version a of the reconstructed human anti-HM1.24 antibody heavy chain V region. The other was a human-mouse hybrid anti-HM1.24 antibody in which the amino acid sequences of FR1 and FR2 were derived from version a of the reconstructed human anti-HM1.24 antibody heavy chain V region and the amino acid sequences of FR3 and FR4 were derived from mouse anti-HM1.24 antibody. All amino acid sequences in the CDR regions were derived from mouse anti-HM1.24 antibody.

Two H-chain hybrid V regions were constructed by PCR. This method is shown schematically in Figures 6 and 7. Four primers were used to construct the two H-chain hybrid V regions. External primers a (SEQ ID NO: 93) and h (SEQ ID NO: 94) were designed to hybridize with the DNA sequence of the HEF expression vector HEF-VH-gγl. The H-chain hybrid construction primer HYS (SEQ ID NO: 95) had a sense DNA sequence, and the H-chain hybrid primer HYA (SEQ ID NO: 96) had an antisense DNA sequence, and these primers were designed to be complementary to each other.

To construct an H-chain hybrid V region in which the amino acid sequences of FR1 and FR2 were derived from mouse anti-HM1.24 antibody and the amino acid sequences of FR3 and FR4 were derived from version a of the H-chain V region of reshaped human anti-HM1.24 antibody, PCR was performed in the first step using plasmid HEF-1.24H-gγ1 as a template with external primer a and H-chain hybrid primer HYA, and PCR was performed in the second step using plasmid HEF-RVHa-AHM-gγ1 as a template with H-chain hybrid primer HYS (SEQ ID NO: 95) and external primer h (SEQ ID NO: 94). The PCR products were then purified. The two PCR products from the first PCR were assembled by their own complementarity (see International Patent Application Publication No. WO 92-19759).

Next, external primers a (SEQ ID NO: 93) and h (SEQ ID NO: 94) were added to amplify the full-length DNA encoding the H-chain hybrid V region in which the amino acid sequences of FR1 and FR2 were derived from the mouse anti-HM1.24 antibody and the amino acid sequences of FR3 and FR4 were derived from version a of the H-chain V region of the reshaped human anti-HM1.24 antibody.

To construct an H-chain hybrid V region in which the amino acid sequences of FR1 and FR2 were derived from version a of the reshaped human anti-HM1.24 antibody H-chain V region and the amino acid sequences of FR3 and FR4 were derived from the mouse anti-HM1.24 antibody, PCR was performed in the first PCR step using plasmid HEF-RVH a-AHM-gγ1 as a template and external primer a and H-chain hybrid primer HYA, and PCR was performed using plasmid HEF-1.24H-gγ1 as a template and H-chain hybrid primer HYS and external primer h. The PCR products were then purified. The two PCR products from the first PCR were then assembled by their own complementarity (see International Patent Application Publication No. WO92-19759).

Next, external primers a and h were added to amplify in the second PCR step a full-length DNA encoding the H-chain hybrid V region in which the amino acid sequences of FR1 and FR2 were derived from version a of the H-chain V region of the reshaped human anti-HM1.24 antibody, and the amino acid sequences of FR3 and FR4 were derived from the mouse anti-HM1.24 antibody.

Purification of the first PCR product, assembly, second PCR, and cloning into the HEF expression vector HEF-VH-gγ1 were performed according to the method described in Example 9, "Construction of the reshaped human HM1.24 antibody L chain V region."

After DNA sequencing, a plasmid containing a DNA fragment encoding the correct amino acid sequence of an H-chain hybrid V region in which the amino acid sequences of FR1 and FR2 are derived from mouse anti-HM1.24 antibody and the amino acid sequences of FR3 and FR4 are derived from version a of the H-chain V region of reshaped human anti-HM1.24 antibody was designated HEF-MH-RVH-AHM-gγ1. The amino acid and nucleotide sequences of the H-chain V region contained in this plasmid, HEF-MH-RVH-AHM-gγ1, are shown in SEQ ID NO: 97. Furthermore, a plasmid containing a DNA fragment encoding the correct amino acid sequence of an H-chain hybrid V region in which the amino acid sequences of FR1 and FR2 are derived from version a of the reshaped human anti-HM1.24 antibody H-chain V region and the amino acid sequences of FR3 and FR4 are derived from mouse antibody HM1.24 antibody was designated HEF-HM-RVH-AHM-gγ1. The amino acid sequence and nucleotide sequence of the H chain V region contained in this plasmid HEF-HM-RVH-AHM-gγ1 are shown in SEQ ID NO: 99.

3-3. Construction of reshaped human anti-HM1.24 antibody H chain V region versions f–r The reshaped human anti-HM1.24 antibody H chain V region versions f, g, h, i, j, k, l, m, n, o, p, q, and r were constructed as follows.

Version f was obtained by PCR amplification using the mutagenic primers FS (SEQ ID NO: 102) and FA (SEQ ID NO: 103), which were designed to mutate threonine at position 75 to serine and valine at position 78 to alanine, and the plasmid HEF-RVHe-AHM-gγ1 as template DNA. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid HEF-RVHf-AHM-gγ1 are shown in SEQ ID NO: 25.

Version g was obtained by amplification using the mutagenic primers GS (SEQ ID NO: 104) and GA (SEQ ID NO: 105), which were designed to mutate alanine to arginine at position 40, and plasmid HEF-RVHa-AHM-gγ1 as template DNA. The amino acid sequence and nucleotide sequence of the H chain V region contained in this plasmid HEF-RVHg-AHM-gγ1 are shown in SEQ ID NO: 27.

Version h was generated by amplifying the HEF-RVHb-AHM-gγ1 plasmid as template DNA using mutagenic primers FS (SEQ ID NO: 102) and FA (SEQ ID NO: 103). The amino acid and nucleotide sequences of the H chain V region contained in this plasmid HEF-RVHh-AHM-gγ1 are shown in SEQ ID NO: 29.

Version i was generated by amplification using the HEF-RVHh-AHM-gγ1 plasmid as template DNA with mutagenic primers IS (SEQ ID NO: 106) and IA (SEQ ID NO: 107), which were designed to mutate arginine to alanine at position 83 and serine to phenylalanine at position 84. The HEF-RVHi-AHM-gγ1 plasmid was obtained by amplification using the HEF-RVHi-AHM-gγ1 plasmid as template DNA. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid are shown in SEQ ID NO: 31.

Version j was obtained by amplifying plasmid HEF-RVHf-AHM-gγ1 as template DNA using mutagenic primers JS (SEQ ID NO: 108) and JA (SEQ ID NO: 109), which were designed to mutate arginine to lysine at position 66. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid, HEF-RVHj-AHM-gγ1, are shown in SEQ ID NO: 33.

Version k was obtained by amplification using the HEF-RVHh-AHM-gγ1 plasmid as template DNA with mutagenic primers KS (SEQ ID NO: 110) and KA (SEQ ID NO: 111), which were designed to mutate glutamic acid at position 81 to glutamine. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid, HEF-RVHk-AHM-gγ1, are shown in SEQ ID NO: 35.

Version 1 was generated by amplification using the HEF-RVHh-AHM-gγ1 plasmid as template DNA with mutagenic primers LS (SEQ ID NO: 112) and LA (SEQ ID NO: 113), which were designed to mutate glutamic acid at position 81 to glutamine and serine at position 82B to isoleucine, to obtain the HEF-RVHl-AHM-gγ1 plasmid. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid, HEF-RVHl-AHM-gγ1, are shown in SEQ ID NO: 37.

Version m was generated by amplification using the HEF-RVHh-AHM-gγ1 plasmid as template DNA with mutagenic primers MS (SEQ ID NO: 114) and MA (SEQ ID NO: 115), which were designed to mutate glutamic acid at position 81 to glutamine, serine at position 82b to isoleucine, and threonine at position 87 to serine. The HEF-RVHm-AHM-gγ1 plasmid was obtained by amplification using the HEF-RVHh-AHM-gγ1 plasmid as template DNA. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid are shown in SEQ ID NO: 39.

Version n was obtained using the mutagenic primers NS (SEQ ID NO: 116) and NA (SEQ ID NO: 117), designed to mutate serine to isoleucine at position 82B, and plasmid HEF-RVHn-AHM-gγ1 as template DNA. The amino acid sequence and nucleotide sequence of the H chain V region contained in this plasmid HEF-RVHn-AHM-gγ1 are shown in SEQ ID NO: 41.

Version o was prepared by using mutagenic primers OS (SEQ ID NO: 118) and OA (SEQ ID NO: 119), designed to mutate threonine to serine at position 87, and plasmid HEF-RVHh-AHM-gγ1 as template DNA to obtain plasmid HEF-RVHo-AHM-gγ1. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid HEF-RVHo-AHM-gγ1 are shown in SEQ ID NO: 43.

Version p was amplified by PCR using mutagenic primers PS (SEQ ID NO: 120) and PA (SEQ ID NO: 121), designed to mutate valine to alanine at position 78, and plasmid HEF-RVHa-AHM-gγ1 as template DNA to obtain plasmid HEF-RVHp-AHM-gγ1. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid HEF-RVHp-AHM-gγ1 are shown in SEQ ID NO: 45.

Version q was obtained by PCR amplification using mutagenic primers QS (SEQ ID NO: 122) and QA (SEQ ID NO: 123), designed to mutate threonine to serine at position 75, and plasmid HEF-RVHa-AHM-gγ1 as template DNA. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid HEF-RVHq-AHM-gγ1 are shown in SEQ ID NO: 47.

Version r was amplified by PCR using the mutagenic primers CS (SEQ ID NO: 87) and CA (SEQ ID NO: 88) and the plasmid HEF-RVHp-AHM-gγ1 as template DNA to obtain the plasmid HEF-RVHr-AHM-gγ1. The amino acid and nucleotide sequences of the H chain V region contained in this plasmid HEF-RVHr-AHM-gγ1 are shown in SEQ ID NO: 49.

The regions encoding the variable regions of the HEF-RVLa-AHM-gκ and HEF-RVHr-AHM-gγ1 plasmids were isolated as restriction fragments using the restriction enzymes HindIII and BamHI, and these fragments were inserted into the HindIII and BamHI sites of the plasmid vector pUC19. The resulting plasmids were designated pUC19-RVLa-AHM-gκ and pUC19-RVHr-AHM-gγ1, respectively.

The E. coli strains containing the plasmids pUC19-RVLa-AHM-gκ and pUC19-RVHr-AHM-gγ1 are designated Escherichia coli DH5α (pUC19-RVLa-AHM-gκ) and Escherichia coli DH5α (pUC19-RVHr-AHM-gγ1), respectively. They were internationally deposited under the Budapest Treaty on August 29, 1996, with the National Institute of Bioscience and Human-Technology (1-1-3 Higashi, Tsukuba, Ibaraki Prefecture, Japan) as FERM BP-5645 and FERM BP-5643, respectively.

4. Construction of Reshaped Human Anti-HM1.24 Antibody, Chimeric Anti-HM1.24 Antibody, and Heavy Chain Hybrid Antibody To evaluate each chain of the reshaped human anti-HM1.24 antibody, we expressed a reshaped human anti-HM1.24 antibody and a chimeric anti-HM1.24 antibody as a positive control antibody. We then expressed a heavy chain hybrid antibody to determine which amino acid residues in the FR should be substituted when constructing each version of the reshaped human anti-HM1.24 antibody heavy chain V region from version b onward. We also expressed the reshaped human anti-HM1.24 antibody in combination with a chimeric heavy chain to evaluate version a of the light chain.

Expression of reshaped human anti-HM1.24 antibody. Ten micrograms of each of the expression vectors for the reshaped human anti-HM1.24 antibody heavy chain (HEF-RVHa-AHM-gγ1-HEF-RVHr-AHM-gγ1) and the expression vectors for the reshaped human anti-HM1.24 antibody light chain (HEF-RVLa-AHM-gκ or HEF-RVLb-AHM-gκ) were cotransfected into COS-7 cells by electroporation using a Gene Pulser apparatus (BioRad). Ten micrograms of each DNA were added to 0.8 ml aliquots containing 1 x ¹⁰ cells/ml in PBS, and the cells were pulsed at 1500 V and 25 μF.

After a 10-minute recovery period at room temperature, the electroporated cells were added to 30 ml of DMEM medium (GIBCO) containing 10% γ-globulin-free fetal bovine serum. After 72 hours of incubation at ₃₇ °C and 5% CO₂ in a _CO₂ incubator BNA120D (TABAI), the culture supernatant was collected and centrifuged at 1,000 rpm for 5 minutes in a centrifuge 15PR-22 (Hitachi) equipped with a centrifugal rotor 03 (Hitachi) at 1,000 rpm to remove cell debris. The supernatant was then subjected to ultrafiltration at 2,000 rpm using a microconcentrator (Centricon 100, Amicon) in a centrifuge J2-21 (Beckman) equipped with a centrifugal rotor JA-20·1 (Beckman) for cell ELISA.

4-2. Expression of Chimeric Anti-HM1.24 Antibody Using 10 μg each of the expression vector HEF-1.24H-gγ1 for the chimeric anti-HM1.24 antibody heavy chain and the expression vector HEF-1.24L-gκ for the chimeric anti-HM1.24 antibody light chain, chimeric anti-HM1.24 antibody for use in Cell-ELISA was prepared according to the method for expressing the reconstituted human anti-HM1.24 antibody described above.

4-3. Expression of anti-HM1.24 antibody composed of humanized L chain version a and chimeric H chain Using 10 μg each of the expression vector HEF-1.24H-gγ1 for the chimeric anti-HM1.24 antibody H chain and the expression vector HEF-RVLa-AHM-Gκ for the reshaped human anti-HM1.24 antibody L chain version a, anti-HM1.24 antibody composed of humanized L chain version a and chimeric H chain for use in Cell-ELISA was prepared according to the method for expressing reshaped human anti-HM1.24 antibody described above.

4-4. Expression of Heavy Chain Hybrid Antibodies H-chain hybrid antibodies for use in Cell-ELISA were prepared using 10 μg each of the expression vector for the heavy chain hybrid V region (HEF-MH-RVH-AHM-gγ1 or HEF-HM-RVH-AHM-gγ1) and the expression vector for the reshaped human anti-HM1.24 antibody light chain (HEF-RVLa-AHM-gκ) according to the method for expressing reshaped human anti-HM1.24 antibodies described above.

Measurement of antibody concentration The concentration of the obtained antibody was measured by ELISA. 100 μl of goat anti-human IgG antibody (BIO SOURCE) adjusted to a concentration of 1 μg/ml with coating buffer (0.1 _M NaHCO , 0.02% _NaN , pH 9.6) was added to each well of a 96-well ELISA plate (Maxisorp, NUNC), and the plate was incubated at room temperature for 1 hour for immobilization. After blocking with 100 μl of dilution buffer (50 mM Tris-HCl, 1 mM _MgCl2 , 0.15 M NaCl, 0.05% Tween ₂₀ , 0.02% NaN3, 1% bovine serum albumin (BSA), pH 8.1), the reconstituted human anti-HM1.24 antibody, chimeric anti-HM1.24 antibody, and H-chain hybrid antibody, which had been concentrated by ultrafiltration, were serially diluted and 100 μl each was added to each well.After incubation at room temperature for 1 hour and washing, 100 μl of alkaline phosphatase-labeled goat anti-human IgG antibody (manufactured by DAKO) was added.

After incubation at room temperature for 1 hour and washing, 100 μl of a 1 mg/ml substrate solution (Sigma 104, p-nitrophenyl phosphate, SIGMA) dissolved in substrate buffer (50 mM NaHCO ₃ , 10 mM MgCl ₂ (pH 9.8)) was added, and the absorbance at 405 nm was measured using a Microplate Reader Model 3550 (Bio-Rad). Human IgG1κ (The Binding Site) was used as a standard for concentration measurement.

5. Establishment of a CHO cell line stably producing reshaped human anti-HM1.24 antibody 5-1. Construction of a reshaped human anti-HM1.24 antibody H chain expression vector Plasmid HEF-RVHr-AHM-gγ1 was digested with restriction enzymes PvuI and BamHI, and an approximately 2.8 kbp fragment containing the EFI promoter and DNA encoding the reshaped human anti-HM1.24 antibody H chain V region was purified using a 1.5% low-melting-point agarose gel. Next, the expression vector used in the human H chain expression vector DHFR-ΔE-R Vh-PMlf (International Patent Application Publication No. WO 92-19759), which contains the DHFR gene and a gene encoding the human H chain constant region, was digested with PvuI and BamHI to prepare an approximately 6-kbp fragment. The above DNA fragment was then inserted into this fragment to construct the reconstructed human anti-HM1.24 antibody H chain expression vector DHFR-ΔE-HEF-RVHr-AHM-gγ1.

5-2. Gene Transfer into CHO Cells To establish a stable production system for the reconstituted human anti-HM1.24 antibody, the expression vectors DHFR-ΔE-HEF-RVHr-AHM-gγ1 and HEF-RVLa-AHM-gκ, which had been linearized by PvuI digestion, were simultaneously transfected into CHO cell DXB-11 by electroporation under the same conditions as described above (for transfection into COS-7 cells).

5-3. Gene Amplification with MTX The transfected CHO cells were cultured in nucleoside-free α-MEM medium (GIBCO-BRL) supplemented with 500 μg/ml G418 (GIBCO-BRL) and 10% fetal bovine serum. Only CHO cells transfected with both the L-chain and H-chain expression vectors were able to grow and were selected. Next, 10 nM MTX (Sigma) was added to the culture medium, and clones that produced the most reshaped human anti-HM1.24 antibody were selected. Clone #1, which produced approximately 3 μg/ml of reshaped human anti-HM1.24 antibody, was selected as a reshaped human anti-HM1.24 antibody-producing cell line.

Production of reshaped human anti-HM1.24 antibody Reshaped human anti-HM1.24 antibody was produced as follows: The reshaped human anti-HM1.24 antibody-producing CHO cells were cultured in a nucleoside-free α-MEM medium (GIBCO-BRL) supplemented with 500 μg/ml G418 (GIBCO-BRL) containing 10% γ-globulin-free fetal bovine serum (GIBCO-BRL) at 37°C and 5% CO for 10 days in _{a CO} incubator BNA120D (TABAI). The culture medium was collected on the 8th and 10th days after the start of cultivation and centrifuged at 2000 rpm for 10 minutes using a centrifuge RL-500SP (Tomy Seiko Co., Ltd.) equipped with a TS-9 rotor to remove cell debris from the culture medium. The culture medium was then sterilized by filtration using a bottle-top filter (Falcon) with a 0.45 μm membrane.

An equal volume of PBS(-) was added to the CHO cell culture medium producing the reconstituted human anti-HM1.24 antibody, and the reconstituted human anti-HM1.24 antibody was affinity purified using a high-speed antibody purification system ConSep LC100 (MILLIPORE) and a Hyper D Protein A column (NGK Insulators) according to the manufacturer's instructions, using PBS(-) as the adsorption buffer and 0.1 M sodium citrate buffer (pH 3) as the elution buffer. The eluted fraction was immediately adjusted to approximately pH 7.4 by adding 1 M Tris-HCl (pH 8.0), then concentrated using a centrifugal ultraconcentrator Centriprep10 (MILLIPORE) and buffer exchanged with PBS(-). The purified reconstituted human anti-HM1.24 antibody was obtained by filtration and sterilization using a 0.22 μm pore size membrane filter MILLEX-GV (MILLIPORE). The concentration of the purified antibody was determined by measuring the absorbance at 280 nm and calculating 1 mg/ml as 1.350 D.

Reference Example 11 Measurement of activity of reshaped human anti-HM1.24 antibody The reshaped human anti-HM1.24 antibody was evaluated for antigen binding activity and binding inhibitory activity as described below.

1. Measurement of Antigen Binding Activity and Binding Inhibition Activity 1-1. Measurement of Antigen Binding Activity Antigen binding activity was measured by Cell-ELISA using WISH cells. Cell-ELISA plates were prepared as described in Reference Example 7.1-2 above.

After blocking, serially diluted reconstituted human anti-HM1.24 antibody, obtained by concentrating COS-7 cell culture supernatant or purified from CHO cell culture supernatant, was added in 100 μl volumes to each well. After 2 hours of incubation at room temperature and washing, peroxidase-labeled rabbit anti-human IgG antibody (DAKO) was added. After 1 hour of incubation at room temperature and washing, substrate solution was added and the reaction was stopped with 50 μl of 6N sulfuric acid. The absorbance at 490 nm was measured using a Microplate Reader Model 3550 (Bio-Rad).

1-2. Measurement of Binding Inhibitory Activity The binding inhibitory activity of biotin-labeled mouse anti-HM1.24 antibody was measured by cell-ELISA using WISH cells. Cell-ELISA plates were prepared as described above. After blocking, serially diluted reconstituted human anti-HM1.24 antibody obtained by concentrating COS-7 cell culture supernatant or purified from CHO cell culture supernatant was added in 50 μl volumes to each well. At the same time, 50 μl of 2 μg/ml biotin-labeled mouse anti-HM1.24 antibody was added. The plate was incubated for 2 hours at room temperature and washed. After incubation for 1 hour at room temperature, peroxidase-labeled streptavidin (DAKO) was added. After incubation with substrate solution, the reaction was stopped with 50 μl of 6N sulfuric acid, and the absorbance at 490 nm was measured using a Microplate Reader Model 3550 (Bio-Rad).

2. Evaluation of Reshaped Human Anti-HM1.24 Antibody 2-1. Light Chain Reshaped human anti-HM1.24 antibody light chain version a was evaluated by measuring its antigen binding activity as described above. As shown in Figure 8, when expressed in combination with a chimeric heavy chain, light chain version a exhibited antigen binding activity comparable to that of the chimeric anti-HM1.24 antibody. However, considering further activity enhancement or compatibility with the heavy chain, a new light chain version b was constructed. Both light chain versions a and b were evaluated by measuring their antigen binding activity and binding inhibitory activity when combined with heavy chain versions a, b, f, or h. As shown in Figures 9, 10, 11, and 12, for all versions of heavy chain a, b, f, and h, light chain version a exhibited stronger activity than version b. Therefore, light chain version a of reshaped human anti-HM1.24 antibody was used in the following experiments.

2-2. H-Chain Versions A-E H-chain versions A-E of the reshaped human anti-HM1.24 antibody were evaluated in combination with L-chain version A by measuring the antigen binding activity and binding inhibitory activity described above. As shown in Figures 11, 13, 14, and 15, both activities were weaker in all versions compared to the chimeric anti-HM1.24 antibody, suggesting that further amino acid alterations are necessary.

2-3. Heavy Chain Hybrid Antibodies The heavy chain hybrid antibodies were evaluated by measuring their antigen-binding activity as described above. As shown in Figure 16, the human-mouse hybrid anti-HM1.24 antibody had antigen-binding activity equivalent to that of the chimeric anti-HM1.24 antibody, while the mouse-human hybrid anti-HM1.24 antibody had weaker activity than the chimeric anti-HM1.24 antibody. This indicates that, to generate a reshaped human HM1.24 antibody with antigen-binding activity equivalent to that of the mouse or chimeric anti-HM1.24 antibody, it is necessary to modify the amino acids in FR3 or FR4 of the heavy chain V region.

2-4. Heavy Chain Versions f-r The heavy chain version f of the reshaped human anti-HM1.24 antibody was evaluated by the antigen binding activity assay described above. As shown in Figure 17, the antigen binding activity was inferior to that of the chimeric anti-HM1.24 antibody, but was improved compared to versions a-c. This suggests that one of the four amino acids at positions 67, 69, 75, and 78 newly altered in this version is involved in the activity of the reshaped human antibody.

The heavy chain version g of the reshaped human anti-HM1.24 antibody was evaluated by measuring the antigen-binding activity and binding inhibitory activity described above. As shown in Figures 18 and 19, this version exhibited activity comparable to that of version a. This indicates that, as shown in the evaluation of the heavy chain human-mouse hybrid antibody described above, the amino acid at position 40, which was changed in this version, does not contribute to improving the activity of the reshaped human antibody.

The reshaped human anti-HM1.24 antibody H-chain versions h-j were evaluated by measuring the antigen-binding activity and binding inhibitory activity described above. As shown in Figures 20, 21, 22, and 23, both activities were weaker in all versions than in the chimeric anti-HM1.24 antibody and comparable to version f. This suggests that of the four amino acids newly altered in version f, amino acids 67 and 69 do not contribute to the improved activity of the reshaped human antibody.

The reshaped human anti-HM1.24 antibody H-chain versions k-p were evaluated by measuring the antigen-binding activity and binding inhibitory activity described above. As shown in Figures 24, 25, 26, and 27, both activities were weaker in all versions than in the chimeric anti-HM1.24 antibody and comparable to version h. This suggests that the newly modified amino acids from position 80 onward in these six versions do not contribute to improved activity of the reshaped human antibody.

The reshaped human anti-HM1.24 antibody H-chain version q was evaluated by measuring its antigen-binding activity and binding inhibitory activity. As shown in Figures 25 and 27, this version had weaker activity than versions h and p, and was only about the same level as version a, suggesting that the substitution of the amino acid at position 78 is essential for improving the activity of the reshaped human antibody.

The heavy chain version r of the reshaped human anti-HM1.24 antibody was evaluated using the above-described assay. As shown in Figures 15 and 28, the results demonstrated that version r had antigen-binding activity and binding inhibition activity comparable to those of the chimeric anti-HM1.24 antibody.

These results demonstrate that the minimum amino acid changes required in the H chain for a reshaped human anti-HM1.24 antibody to have antigen-binding activity comparable to that of mouse or chimeric anti-HM1.24 antibodies are at positions 30, 71, 78, and even 73.

The antigen binding activity and binding inhibitory activity of reshaped human anti-HM1.24 antibody H chain versions a-r are summarized in Table 2.

Table 2 The amino acid sequences of reshaped human anti-HM1.24 antibody L chain versions a and b are shown in Table 3, and the amino acid sequences of reshaped human anti-HM1.24 antibody H chain versions a-r are shown in Tables 4-6.

Table 3 Amino acid sequence of the L chain V region Table 4. Amino acid sequence of H chain V region (1) Table 5. Amino acid sequence of H chain V region (2) Table 6. Amino acid sequence of the H chain V region 3. Evaluation of Purified Reshaped Human Anti-HM1.24 Antibody The purified reshaped human anti-HM1.24 antibody was evaluated for its antigen-binding activity and binding inhibitory activity as described above. As a result, as shown in Figures 31 and 32, the reshaped human anti-HM1.24 antibody was shown to have antigen-binding activity and binding inhibitory activity comparable to those of the chimeric anti-HM1.24 antibody. This demonstrates that the reshaped human anti-HM1.24 antibody has the same antigen-binding ability as the mouse anti-HM1.24 antibody.

Reference Example 12 Preparation of Hybridoma Producing Mouse Anti-HM1.24 Monoclonal Antibody A hybridoma producing mouse anti-HM1.24 monoclonal antibody was prepared by the method described in Goto, T. et al., Blood (1994) 84, 1992-1930.

Epstein-Barr virus nuclear antigen (EBNA)-negative plasma cell line KPC-32 (1 x ¹⁰ cells) (Goto, T. et al., Jpn. J. Clin. Hematol. (11991) 32, 1400), derived from the bone marrow of a human multiple myeloma patient, was injected intraperitoneally into BALB/C mice (Charles River) twice, 6 weeks apart.

Three days before sacrifice, 1.5 x ¹⁰ KPC-32 cells were injected into the spleen of the mice to further increase their antibody titers (Goto, T. et al., Tokushima J. Exp. Med. (1990) 37, 89). After sacrifice, the spleens were removed and subjected to cell fusion with myeloma cells SP2/0 according to the method of Groth, de St. & Schreidegger (Cancer Research (1981) 41, 3465).

Antibodies in hybridoma culture supernatants were screened by ELISA using plates coated with KPC-32 cells (Posner, M.R. et al., J. Immunol. Methods (1982) 48, 23). ^{5 x 10} KPC-32 cells were suspended in 50 ml of PBS and dispensed into a 96-well plate (U-bottom, Corning, Iwaki). After blocking with PBS containing 1% bovine serum albumin (BSA), hybridoma culture supernatant was added and incubated for 2 hours at 4°C. The plates were then incubated with peroxidase-labeled goat anti-mouse IgG antibody (Zymed) for 1 hour at 4°C, washed once, and incubated with o-phenylenediamine substrate solution (Sumitomo Bakelite) for 30 minutes at room temperature.

The reaction was stopped with 2N sulfuric acid, and absorbance at 492 nm was measured using an ELISA reader (Bio-Rad). To remove hybridomas producing antibodies against human immunoglobulins, positive hybridoma culture supernatants were pre-absorbed with human serum and screened for reactivity against other cell subunits by ELISA. Positive hybridomas were selected and examined by flow cytometry for reactivity against various cell lines and human specimens. Finally, selected hybridoma clones were cloned twice and injected intraperitoneally into pristane-treated BALB/C mice to obtain ascites.

Monoclonal antibodies were purified from mouse ascites by ammonium sulfate precipitation and a Protein A affinity chromatography kit (Ampure PA, Amersham). The purified antibodies were conjugated to fluorescein isothiocyanate (FITC) using a Quick Tag FITC conjugation kit (Boehringer Mannheim).

As a result, monoclonal antibodies produced by 30 hybridoma clones reacted with KPC-32 and RPMI8226 cells. After cloning, the culture supernatants of these hybridomas were tested for reactivity with other cell lines and peripheral blood-derived mononuclear cells.

Of these, three clones produced monoclonal antibodies specifically reacting with plasma cells. Of these three clones, the hybridoma clone most useful for flow cytometry analysis and exhibiting complement-dependent cytotoxicity against RPMI8226 cells was selected and designated HM1.24. The subclass of the monoclonal antibody produced by this hybridoma was determined by ELISA using subclass-specific anti-mouse rabbit antibodies (Zymed). The anti-HM1.24 antibody possessed the IgG2aκ subclass. The hybridoma HM1.24, which produced the anti-HM1.24 antibody, was internationally deposited with the National Institute of Bioscience and Biotechnology, Agency of Industrial Science and Technology (1-1-3 Higashi, Tsukuba, Ibaraki Prefecture) under the Budapest Treaty on September 14, 1995, as FERM BP-5233.

Reference Example 13 Cloning of cDNA Encoding HM1.24 Antigen Polypeptide 1. Construction of cDNA Library 1) Preparation of Total RNA cDNA encoding the HM1.24 antigen, an antigen polypeptide specifically recognized by mouse monoclonal antibody HM1.24, was isolated as follows.

Total RNA was prepared from the human multiple myeloma cell line KPMM2 according to the method of Chirgwin et al. (Biology, 18, 5, 294 (1979)). Specifically, 2.2 x ¹⁰ KPMM2 cells were completely homogenized in 20 ml of 4 M guanidine thiocyanate (Nacalai Tesque).

The homogenate was layered on a 5.3 M cesium chloride solution in a centrifuge tube, and then the tube was centrifuged at 31,000 rpm in a Beckman SW40 rotor at 20°C for 24 hours to precipitate RNA. The RNA precipitate was washed with 70% ethanol and dissolved in 300 μl of 10 mM Tris-HCl (pH 7.4) containing 1 mM EDTA and 0.5% SDS. Pronase (Boehringer) was added to a concentration of 0.5 mg/ml, and the mixture was then incubated at 37°C for 30 minutes. The mixture was extracted with phenol and chloroform, and the RNA was precipitated with ethanol. The RNA precipitate was then dissolved in 200 μl of 10 mM Tris-HCl (pH 7.4) containing 1 mM EDTA.

2) Preparation of poly(A)+ RNA Approximately 500 μg of the total RNA prepared as described above was used to purify poly(A)+ RNA using the Fast Track 2.0mRNA Isolation Kit (Invitrogen) according to the instructions included with the kit.

3) Construction of a cDNA Library Double-stranded cDNA was synthesized using 10 μg of the poly(A) + RNA described above using the TimeSaver cDNA Synthesis Kit (Pharmacia) according to the kit's instructions. Furthermore, EcoRI adapters were ligated using the Directional Cloning Toolbox (Pharmacia) according to the kit's instructions. Kinase of the EcoRI adapters and digestion with the NotI restriction enzyme were performed according to the kit's instructions. Furthermore, adapter-tagged double-stranded cDNAs of approximately 500 bp or larger were separated and purified using a 1.5% low-melting-point agarose gel (Sigma), yielding approximately 40 μl of adapter-tagged double-stranded cDNA.

The thus prepared double-stranded cDNA with adapters was ligated with pCOS1 vector (Japanese Patent Application No. 8-255196) treated with restriction enzymes EcoRI and NotI and alkaline phosphatase (Takara Shuzo) using T4 DNA ligase (GIBCO-BRL) to construct a cDNA library. The constructed cDNA library was transformed into Escherichia coli cell line DH5α (GIBCO-BRL), and the total size was estimated to be approximately 2.5 x ¹⁰ independent clones.

2. Cloning by Direct Expression 1) Transfection into COS-7 Cells Approximately 5 x ¹⁰⁵ clones of the transduced E. coli described above were cultured in 2-YT medium containing 50 μg/ml ampicillin (Molecular Cloning: A Laboratory Manual, Sambrook et al., Cold Spring Harbor Laboratory Press (1989)) to amplify the cDNA, and plasmid DNA was recovered from the E. coli by the alkaline method (Molecular Cloning: A Laboratory Manual, Sambrook et al., Cold Spring Harbor Laboratory Press (1989)). The resulting plasmid DNA was transfected into COS-7 cells by electroporation using a Gene Pulser apparatus (Bio-Rad).

Specifically, 10 μg of purified plasmid DNA was added to 0.8 ml of COS-7 cells suspended in PBS at 1 x ¹⁰ cells/ml, and pulsed at 1500 V and 25 μFD. After a 10-minute recovery period at room temperature, the electroporated cells were cultured in DMEM medium (GIBCO-BRL) containing 10% fetal _bovine serum (GIBCO-BRL) at 37°C and 5% CO for 3 days.

2) Preparation of Panning Dishes. Panning dishes coated with mouse anti-HM1.24 antibody were prepared according to the method of B. Seed et al. (Proc. Natl. Acad. Sci. USA, 84 , 3365-3369 (1987)). Specifically, mouse anti-HM1.24 antibody was added to 50 mM Tris-HCl (pH 9.5) to a concentration of 10 μg/ml. 3 ml of the antibody solution was added to a 60 mm cell culture dish and incubated at room temperature for 2 hours. After washing three times with 0.15 M NaCl solution, the dish was blocked with PBS containing 5% fetal bovine serum, 1 mM EDTA, and 0.02% _NaN3 , and then used for the cloning procedure described below.

3) Cloning of cDNA Library. COS-7 cells transfected as described above were detached with 5 mM EDTA in PBS, washed once with 5% fetal bovine serum (FBS), and suspended in 5% fetal bovine serum and 0.02% _NaN3 in PBS at approximately 1 x ¹⁰⁶ cells/ml. The cells were added to the panning dish and incubated at room temperature for approximately 2 hours. After gentle washing three times with 5% fetal bovine serum and 0.02% _NaN3 in PBS, plasmid DNA was recovered from the cells attached to the panning dish using a solution containing 0.6% SDS and 10 mM EDTA.

The recovered plasmid DNA was again transformed into E. coli DH5α, and the plasmid DNA was amplified as described above and recovered by the alkaline method. The recovered plasmid DNA was transfected into COS-7 cells by electroporation, and plasmid DNA was recovered from the ligated cells as described above. This procedure was repeated once more, and the recovered plasmid DNA was digested with the restriction enzymes EcoRI and NotI, confirming the enrichment of an insert of approximately 0.9 kbp. Furthermore, E. coli transfected with a portion of the recovered plasmid DNA was inoculated onto a 2-YT agar plate containing 50 μg/ml ampicillin and grown overnight. Plasmid DNA was recovered from a single colony. Digestion with the restriction enzymes EcoRI and NotI yielded clone p3.19, which contained an insert of approximately 0.9 kbp.

This clone was subjected to a PRISM Dye Terminater Cycle Sequencing kit (Perkin Elmer) according to the instructions provided with the kit, and the nucleotide sequence was determined using an ABI 373A DNA Sequencer (Perkin Elmer). The resulting amino acid and nucleotide sequences are shown in SEQ ID NO: 128.

The cDNA encoding the polypeptide having the amino acid sequence shown in SEQ ID NO: 128 was inserted into the XbaI cleavage site of the pUC19 vector to prepare the plasmid pRS38-pUC19. E. coli containing this plasmid pRS38-pUC19 was internationally deposited under the Budapest Treaty on October 5, 1993, with the National Institute of Bioscience and Human-Technology (1-1-3 Higashi, Tsukuba, Ibaraki Prefecture) as Escherichia coli DH5α (pRS38-pUC19) under accession number FER M BP-4434 (see JP 7-196694).

Example Below, we present an example of a natural humanized antibody based on the humanized anti-HM1.24 antibody, as an example of a natural humanized antibody composed of the natural FR sequence of the present invention.

Example 1: Mouse monoclonal antibody HM1.24 antibody was humanized to form a reshaped human anti-HM1.24 antibody by CDR-grafting as described in the Reference Example. To construct the humanized H chain, FR1 through FR3 were selected from the human antibody HG3, and FR4 from human antibody JH6 was selected for FR4. Analysis of FR amino acid residues revealed that amino acid substitutions were necessary at four positions (FR1/30, FR3/71, 73, and 78) (Tables 7 and 8). This humanized antibody had antigen-binding activity equivalent to that of the original mouse antibody. This humanized antibody (a humanized antibody composed of RVLa and RVHr) was used as the primary design antibody.

(1) Construction of the H Chain A homology search for the FRs of the primary antibody design against naturally occurring human FRs was performed using the Swiss Plot, GenBank, PRF, PIR, and GenPept databases. First, 50 human FRs with perfectly identical amino acid sequences were found for FR1. In other words, FR1 of this primary antibody design already had a natural sequence. Since no amino acid substitutions were originally made in FR2 and FR4, there were 50 and 100 natural human antibody FRs, including HG3 and JH6, respectively.

On the other hand, no perfect matches were found for FR3. The most highly homologous FR3s were S46463, 1921296C, HUMIGHRF1, and U005831 (all symbols are database accession numbers), with 96.875% identity.

Therefore, in this primary design antibody, FR3 was an FR containing an artificial amino acid residue not found in nature. A comparison of the amino acid sequence with the human antibody S46463, which shows the greatest homology, is shown in Table 9.

The amino acid residue at position 70 in FR3 of the primary design antibody was methionine, while that in FR3 of human antibody S46463 was isoleucine. The other amino acid sequences were completely identical. Therefore, the amino acid residue at position 70 in the primary design antibody was replaced with isoleucine, converting it to a naturally occurring FR3. The resulting secondary design antibody is a CDR-grafted antibody composed of the natural human FRs of human antibody S46463. The secondary design antibody constructed in this way is composed entirely of naturally occurring FRs.

(2) Construction of the Natural Humanized Anti-HM1.24 Antibody H Chain V Region The natural humanized anti-HM1.24 antibody H Chain V region was constructed by PCR mutagenesis. Mutagenic primers SS (SEQ ID NO: 124) and SA (SEQ ID NO: 125) were designed to mutate methionine at position 69 to isoleucine.

Using the plasmid HEF-RVHr-AHM-gγ1 described in the Reference Example as a template, amplification was performed using the above primers. The final product was purified and digested with BamHI and HindIII. The resulting DNA fragment was cloned into the HEF expression vector HEF-VH-gγ1 to obtain the plasmid HEF-RVHs-AHM-gγ1. The amino acid sequence and nucleotide sequence of the H chain V region contained in this plasmid HEF-RVHs-AHM-gγ1 are shown in SEQ ID NO: 126.

The region encoding the variable region from the HEF-RVHs-AHM-gγ1 plasmid was digested with the restriction enzymes HindIII and BamHI to create a restriction fragment, which was then inserted into the BamHI and HindIII sites of the plasmid vector pUC19. The resulting plasmid was designated pUC19-RVHs-AHM-gγ1.

The E. coli strain containing the plasmid pUC19-RVHs-AHM-gγ1 was designated Escherichia coli DH5α (pUC19-RVHs-AHM-gγ1) and was deposited with the National Institute of Bioscience and Human-Technology, Agency of Industrial Science and Technology (1-1-3 Higashi, Tsukuba, Ibaraki Prefecture) on September 29, 1997, under the Budapest Treaty as FERM BP-6127.

2) Light Chain Analysis No amino acid substitutions were made in the FRs of the primary antibody light chain. However, the human antibody REI used had a modified FR (Riechmann, L. et al., Nature (1988) 332, 323-327) in which amino acid substitutions had already been made. Therefore, homology searches were also performed on these FRs. As a result, natural sequences corresponding to these modified FRs were found. Therefore, it was determined that amino acid substitutions in the light chain FRs were not necessary.

Example 2. Production of native humanized anti-HM1.24 antibody (1) Expression of native humanized anti-HM1.24 antibody COS cells were co-transfected by electroporation using a Gene Pulser device (BioRad) with 10 μg each of the expression vector for the native humanized anti-HM1.24 antibody H chain (HEF-RVHs-AHM-gγ1) and the expression vector for the reshaped human anti-HM1.24 antibody L chain (HEF-RVLa-AHM-gκ). Each DNA (10 μg) was added to a 0.8 ml aliquot of 1 x 10 cells/ml in PBS, and pulsed at 1,500 V and 25 μF.

After a 10-minute recovery period at room temperature, the electroporated cells were added to 30 ml of DMEM culture medium (GIBCO) containing 10% γ-globulin-free fetal bovine serum. After 72 hours of incubation at 37°C and 5% _CO2 in a _CO2 incubator BNA120D (TABAI), the culture supernatant was collected and centrifuged at 1,000 rpm for 5 minutes in a centrifuge 05PR-22 (Hitachi) equipped with a centrifugal rotor 03 (Hitachi) to remove cell debris. The supernatant was then ultrafiltrated at 2,000 rpm using a microconcentrator (Centricon 100, Amicon) in a centrifuge J2-21 (Beckman) equipped with a centrifugal rotor JA-20.1 (Beckman). The supernatant was then sterilized by filtration using a Millex GV 13mm filter (Millipore) and subjected to Cell-ELISA.

(2) Measurement of antibody concentration The concentration of the obtained antibody was measured by ELISA. 100 μl of goat anti-human IgG antibody (BIO SOURCE) adjusted to a concentration of 1 μg/ml with coating buffer (0.1 _M NaHCO , 0.02% _NaN , pH 9.6) was added to each well of a 96-well ELISA plate (Maxisorp, NUNC), and the plate was incubated at room temperature for 1 hour for immobilization. After blocking with 100 μl of dilution buffer (50 mM Tris-HCl, 1 mM _MgCl2 , 0.15 M NaCl, 0.05% Tween ₂₀ , 0.02% NaN3, 1% bovine serum albumin (BSA), pH 8.1), natural humanized anti-HM1.24 antibody concentrated by ultrafiltration was serially diluted and 100 μl was added to each well.After incubation at room temperature for 1 hour and washing, 100 μl of alkaline phosphatase-labeled goat anti-human IgG antibody (DAKO) was added.

After incubation at room temperature for 1 hour and washing, 100 μl of a 1 mg/ml substrate solution (Sigma 104, p-nitrophenyl phosphate, SIGMA) dissolved in substrate buffer (50 mM NaHCO ₃ , 10 mM MgCl ₂ (pH 9.8)) was added, and the absorbance at 405 nm was measured using a Microplate Reader Model 3550 (Bio-Rad). Human IgG1κ (The Binding Site) was used as a standard for concentration measurement.

(3) Establishment of a CHO cell line stably producing native humanized anti-HM1.24 antibody. A CHO cell line stably producing native humanized anti-HM1.24 antibody can be established according to the following method.

(3)-1. Construction of a Natural Humanized Anti-HM1.24 Antibody H Chain Expression Vector Plasmid HEF-RVHs-AHM-gγ1 was digested with restriction enzymes PvuI and BamHI, and an approximately 2.8 kbp fragment containing the EF1 promoter and DNA encoding the natural humanized anti-HM1.24 antibody H chain V region was purified using a 1.5% low-melting-point agarose gel. Next, the expression vector used in the human H chain expression vector DHFR-ΔE-RVh-PM1f (International Patent Application Publication No. WO 92-19759), which contains the DHFR gene and a gene encoding the human H chain constant region, was digested with PvuI and BamHI to prepare an approximately 6 kbp fragment. This DNA fragment was then inserted into the approximately 6 kbp fragment to construct the natural humanized anti-HM1.24 antibody H chain expression vector DHFR-ΔE-HEF-RVHs-AHM-gγ1.

(3)-2. Gene Transfer into CHO Cells To establish a stable production system for the native humanized anti-HM1.24 antibody, the expression vectors DHFR-ΔE-HEF-RVHs-AHM-gγ1 and HEF-RVLa-AHM-gκ, which had been linearized by digestion with PvuI, were simultaneously transfected into CHO cells DXB-11 by electroporation under the same conditions as described above (for transfection into COS-7 cells).

(3)-3. Gene Amplification with MTX Transfected CHO cells were cultured in nucleoside-free α-MEM medium (GIBCO-BRL) supplemented with 500 μg/ml G418 (GIBCO-BRL) and 10% fetal bovine serum. Only CHO cells transfected with both the light and heavy chain expression vectors were able to grow and were selected. Next, 10 nM MTX (Sigma) was added to the culture medium, and clones producing high amounts of natural humanized anti-HMI.24 antibody were selected.

(3)-4. Production of Native Humanized Anti-HM1.24 Antibody Native humanized anti-HM1.24 antibody is produced as follows: The native humanized anti-HM1.24 antibody-producing CHO cells are cultured in a nucleoside-free α-MEM medium (GIBCO-BRL) supplemented with 500 μg/ml G418 (GIBCO-BRL) containing 10% γ-globulin-free fetal bovine serum (GIBCO-BRL) at 37°C and 5% CO for 10 days in _{a CO} incubator BNA120D (TABAI). The culture medium was collected on the 8th and 10th day after the start of cultivation and centrifuged at 2000 rpm for 10 minutes using a centrifuge RL-500SP (Tomy Seiko Co., Ltd.) equipped with a TS-9 rotor to remove cell debris from the culture medium, followed by filtration sterilization using a bottle-top filter (Falcon) with a 0.45 μm membrane.

An equal volume of PBS(-) was added to the CHO cell culture medium producing the native humanized anti-HM1.24 antibody, and the native humanized anti-HM1.24 antibody was affinity purified using a high-speed antibody purification system ConSep LC100 (MILLIPORE) and a Hyper D Protein A column (NGK Insulators) according to the manufacturer's instructions, using PBS(-) as the adsorption buffer and 0.1 M sodium citrate buffer (pH 3) as the elution buffer. The eluted fraction was immediately adjusted to approximately pH 7.4 by adding 1 M Tris-HCl (pH 8.0), concentrated using a centrifugal ultraconcentrator Centriprep 10 (MILLIPORE), and buffer exchanged with PBS(-). The purified native humanized anti-HM1.24 antibody was then sterilized by filtration using a 0.22 μm pore size membrane filter MILLEX-GV (MILLIPORE). The concentration of the purified antibody is determined by measuring the absorbance at 280 nm and calculating 1 mg/ml as 1.350 D.

Example 3. Measurement of activity of natural humanized antibody Natural humanized anti-HM1.24 antibody was evaluated for antigen binding activity, binding inhibitory activity, and ADCC activity as described below.

(1) Measurement of Antigen Binding Activity and Binding Inhibition Activity (1)-1. Measurement of Antigen Binding Activity Antigen binding activity was measured by cell-ELISA using WISH cells. Cell-ELISA Plates were prepared as described in Reference Example 7.1-2 above.

After blocking, serial dilutions of native humanized anti-HM1.24 antibody, obtained by concentrating COS-7 cell culture supernatant, were added to each well in 100 μl volumes. After incubation at room temperature for 2 hours and washing, peroxidase-labeled rabbit anti-human IgG antibody (DAKO) was added. After incubation at room temperature for 1 hour and washing, substrate solution was added and the reaction was stopped with 50 μl of 6N sulfuric acid. The absorbance at 490 nm was measured using a Microplate Reader Model 3550 (Bio-Rad).

(1)-2. Measurement of Binding Inhibitory Activity The binding inhibitory activity of biotin-labeled mouse anti-HM1.24 antibody was measured by Cell-ELISA using WISH cells. Cell-ELISA plates were prepared as described above. After blocking, serial dilutions of native humanized anti-HM1.24 antibody obtained by concentrating COS-7 cell culture supernatant were added to each well in an amount of 50 μl. At the same time, 50 μl of 2 μg/ml biotin-labeled mouse anti-HM1.24 antibody was added. The plate was then incubated for 2 hours at room temperature and washed. After washing, peroxidase-labeled streptavidin (DAKO) was added. After incubation for 1 hour at room temperature, the plate was washed, and the substrate solution was added. The reaction was then stopped with 50 μl of 6N sulfuric acid. The absorbance at 490 nm was measured using a Microplate Reader Model 3550 (Bio-Rad).

(2) Antigen-binding activity and binding inhibitory activity The H chain of the native humanized anti-HM1.24 antibody was evaluated by measuring the antigen-binding activity and binding inhibitory activity described above in combination with the L chain version a. The results, as shown in Figures 29 and 30, demonstrated that the native humanized anti-HM1.24 antibody (secondary designed antibody) had antigen-binding activity and binding inhibitory activity comparable to that of the primary designed antibody (reshaped human anti-HM1.24 antibody: H chain version r).

(3) Measurement of ADCC Activity Measurement of ADCC (Antibody-dependent Cellular Cytotoxicity) activity was performed according to the method described in Reference Example 8.

1. Preparation of effector cells: Peripheral blood from a healthy donor was mixed with an equal volume of PBS(-), layered on Ficoll-Paque (Pharmacia Biotech), and centrifuged at 500 g for 30 minutes. The mononuclear cell layer was separated and washed twice with RPMI 1640 (GIBCO BRL) containing 10% fetal bovine serum (GIBCO BRL), and then resuspended in the same culture medium to a cell density of 5 × ¹⁰ /ml.

2. Preparation of target cells Human myeloma cell line KPMM2 (accession number P-14170, patent application number: Japanese Patent Application No. Hei 6-58082) was radiolabeled by incubating with 0.1 mCi of ^51Cr -sodium chromate in RPMI 1640 (GIBCO BRL) containing 10% fetal bovine serum (GIBCO BRL) at 37°C for 60 minutes. After radiolabeling, the cells were washed three times with the same culture medium and adjusted to a concentration of 2 x ¹⁰⁵ /ml.

3. Measurement of ADCC activity 50 μl of radiolabeled target cells (2 × ¹⁰⁵ /ml) and 50 μl of affinity-purified natural humanized anti-HM1.24 antibody (secondary designed antibody) solution (prepared at 4 μg/ml, 0.4 μg/ml, 0.04 μg/ml, and 0.004 μg/ml) were added to a 96-well U-bottom plate (Corning) and incubated for 15 minutes at 4° C. A solution containing no natural humanized anti-HM1.24 antibody (secondary designed antibody) was prepared in the same manner as above and used as a control.

Then, 100 μl of 5 × 10 ⁶ /ml effector cells was added and cultured for 4 hours in a CO 2 incubator, with effector cell (E) to target cell (T) ratios (E:T) of 0:1, 20:1, and 50:1. Each antibody was diluted 4-fold to achieve final concentrations of 1 μg/ml, 0.1 μg/ml, 0.01 μg/ml, 0.001 μg/ml, and no antibody.

100 μl of the supernatant was removed and the radioactivity released into the culture supernatant was measured using a gamma counter (ARC361, Aloka). 1% NP-40 (Nacalai Tesque) was used to measure maximum released radioactivity. The cytotoxicity (%) was calculated as (A-C)/(B-C) × 100. A represents the radioactivity (cpm) released in the presence of antibody, B represents the radioactivity (cpm) released by NP-40, and C represents the radioactivity (cpm) released in the culture medium alone without antibody.

4. Results As shown in Figure 33, when natural humanized anti-HM1.24 antibody (secondary designed antibody) was added, the specific chromium release rate increased in a concentration-dependent manner with increasing E:T ratio compared to when no antibody was added, demonstrating that natural humanized anti-HM1.24 antibody (secondary designed antibody) has ADCC activity.

The present invention relates to a method for producing naturally occurring humanized antibodies composed of naturally occurring human FRs and the naturally occurring humanized antibodies obtained by this method. This technology is a more sophisticated humanization technique that overcomes the problems inherent in CDR-grafting, devised by G. Winter (Jones, P.T. et al., Nature (1986) 321, 522-525). The construction of a primary designed antibody is considered an intermediate step toward the creation of humanized antibodies composed of naturally occurring human FRs. When developing antibodies as recombinant protein pharmaceuticals, naturally occurring humanized antibodies composed of naturally occurring human FRs offer superior antigenicity and safety.

EFFECTS OF THE INVENTION Natural humanized antibodies obtained by the production method of the present invention are expected to have low antigenicity in humans because they do not contain the non-naturally occurring artificial FR amino acid residues found in humanized antibodies obtained by conventional humanization techniques. Furthermore, natural humanized antibodies obtained by the production method of the present invention have been shown to have activity comparable to that of antibodies derived from non-human mammals that served as templates for humanization. Therefore, the natural humanized antibodies of the present invention are useful for therapeutic administration to humans.

Reference to deposited microorganisms and depository institution under Rule 13bis of the Patent Cooperation Treaty Depository institution Name: National Institute of Bioscience and Human-Technology, Agency of Industrial Science and Technology Address: 1-1-3 Higashi, Tsukuba, Ibaraki, Japan Microorganism (1) Identification: Escherichia coli DH5α (pRS38-pUC19) Deposit number: FERM BP-4434 Date of deposit: October 5, 1993 (2) Identification: Hybridoma HM1.24 Deposit number: FERM BP-5233 Date of deposit: September 14, 1995 (3) Identification: Escherichia coli DH5α (pUC19-RVHr-AHM-gγ１) Deposit number: FERM BP-5643 Date of deposit: August 29, 1996 (4) Identification: Escherichia coli DH5α (pUC19-1.24H-gγ1) Deposit number: FERM BP-5644 Deposit date: August 29, 1996 (5) Display: Escherichia coli DH5α (pUC19-RVLa-AHM-gκ) Deposit number: FERM BP-5645 Deposit date: August 29, 1996 (6) Table Indication: Escherichia coli DH5α(pUC19-RVHs-AMH- gγ1) Deposit number: FERM BP-6127 Deposit date: September 29, 1997

────────────────────────────────────────────────── Continued from front page (51) Int.Cl. ⁷ identification code FI C12N 1/15 1/19 1/21 C12P 21/08 C07K 16/00 A61K 39/395 (81) Designated countries EP (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, IT, LU, MC, NL, PT, SE), OA (BF, BJ, CF, CG, CI, CM, GA, GN, GW, ML, MR, NE, SN, TD, TG), AP(GH, GM, K E, LS, MW, SD, SZ, UG, ZW), EA (AM, AZ, BY, KG, KZ, MD, RU, TJ, TM), AL, AM, AT, AU, AZ, BA, BB, BG, BR, BY, CA, CH, CN, CU, CZ, DE, D K, EE, ES, FI, GB, GD, GE, GH, GM, HR, HU, ID, IL, IS, JP, KE, KG, KR, KZ, LC, LK, LR, LS, LT, LU, L V, MD, MG, MK, MN, MW, MX, NO, NZ , PL, PT, RO, RU, SD, SE, SG, SI, SK, SL, TJ, TM, TR, TT, UA, UG, US, UZ, VN, YU, ZW (Note) This publication is based on the gazette published internationally by the International Bureau of Patents (WIPO). The effect of the international publication of the Japanese patent application (Japanese utility model registration application) related to this publication arises pursuant to Article 184-10, Paragraph 1 of the Patent Act (Article 48-13, Paragraph 2 of the Utility Model Act), and is unrelated to this publication.

Claims (13)

[Claims] 1. A method for producing a naturally occurring humanized antibody, comprising: conducting a homology search for the FR of a primary design antibody and selecting a naturally occurring human FR that retains the artificial amino acid residues contained in the FR of the primary design antibody and is homologous to the FR of the primary design antibody. 2. A method for producing a natural humanized antibody according to claim 1, comprising: conducting a homology search for the FR of the primary antibody design; selecting a natural human FR that retains the artificial amino acid residue contained in the FR of the primary antibody design and is homologous to the natural human FR; and then substituting one or more amino acid residues that differ between the FR of the primary antibody design and the selected natural FR. 3. The method of claim 1 or 2, wherein the primarily designed antibody comprises a CDR derived from a first animal species and a FR derived from a second animal species that contains an artificial amino acid residue. 4. The method of claim 3, wherein the first animal species is a mouse or a rat and the second animal species is a human. 5. The method of any one of claims 1 to 4, wherein the artificial amino acid residue is an amino acid residue derived from a non-human antibody FR. 6. A natural humanized antibody obtained by the production method of any one of claims 1 to 5. 7. A naturally occurring humanized antibody comprising a CDR derived from a first animal species and a FR derived from a second animal species, wherein the FR has an amino acid sequence that differs from the FR used in CDR grafting by one or more amino acid residues, and the differing amino acid residues have been replaced with FRs derived from the second animal species containing the same amino acid residues at the same positions. 8. The native humanized antibody of claim 7, wherein the first animal species is mouse or rat, and the second animal species is human. 9. DNA encoding the native humanized antibody described in any one of claims 6 to 8. 10. An expression vector comprising the DNA of claim 9. 11. A host comprising the DNA of claim 10. 12. A method for producing a natural humanized antibody, comprising culturing cells into which an expression vector containing the DNA of claim 9 has been introduced, and obtaining the desired natural humanized antibody from the cell culture. 13. A pharmaceutical composition comprising a natural humanized antibody.