Humanized antibodies

Abstract

Chimeric human antibody expression vectors are constructed by inserting the antibody heavy chain variable region-encoding cDNA and antibody light chain variable region-encoding cDNA isolated from hybridomas producing a mouse or rat monoclonal antibody reacting with the ganglioside GM2 respectively into an expression vector for use in animal cells which contains the human antibody heavy chain constant region- or human antibody light chain constant region-encoding cDNA. The expression vectors are introduced into animal cells and the transformant thus obtained is cultured for the production of a chimeric human antibody reacting with the ganglioside GM2.

Description

FIELD OF THE INVENTION

[0002] The present invention relates to humanized antibodies reacting with the ganglioside GM2. The humanized antibodies do not cause production of anti-mouse immunoglobulins in the patient's body as compared with mouse monoclonal antibodies, hence the incidence of adverse effects possibly caused by them is much lower, their blood half-lives are longer and, further, their anti-tumor effector effect is greater. Therefore, the humanized antibodies are expected to produce improved therapeutic effects as compared with mouse monoclonal antibodies.

BACKGROUND OF THE INVENTION

[0003] When administered to humans, mouse antibodies are generally recognized as foreign matters, inducing production of anti-mouse immunoglobulin antibodies in the human body. It is known that the former antibodies react with the latter antibodies to produce adverse effects [J. Clin. Oncol., 2, 881 (1984); Blood, 65, 1349 (1985); J. Natl. Cancer Inst., 80, 932 (1988); Proc. Natl. Acad. Sc. U.S.A., 82, 1242 (1985)] and that the mouse antibodies undergo rapid clearance [J. Nucl. Med., 26, 1011 (1985); Blood, 65, 1349 (1985); J. Natl. Cancer Inst., 80, 937 (1988)], thus showing only a reduced efficacy [J. Immunol., 135, 1530 (1985); Cancer Res., 46, 6489 (1986)]. Attempts have been made to solve these problems by deriving, from mouse monoclonal antibodies, chimeric human antibodies or CDR (complementarity determining region)-transplanted antibodies (reshaped antibodies) using gene engineering technique. In a human chimeric antibody, the variable regions thereof are of mouse origin and the constant regions thereof are of human origin [Proc. Natl. Acad. Sci. U.S.A., 81, 6851 (1984)] and it is reported that when administered to humans, said antibody causes litte human anti-mouse immunoglobulin antibody production, its blood half-life being 6-fold longer [Proc. Natl. Acad. Sci. U.S.A., 86, 4220 (1989)]. The CDR-grafted antibodies are antibodies resulting from replacement of the CDRs in a human antibody alone with the CDRs from an animal other than the human [Nature, 32, 522 (1986)] and, in an experiment with monkeys, such antibodies showed reduced immunogenicity and 4-to 5-fold higher serum half-lives as compared with mouse antibodies [J. Immunol., 147, 1352 (1991)].

[0004] As regards the cytocidal activity of antibodies, it is reported that the Fc region of a human antibody is more potent in activating human complement and human effector cells than the Fc region of a mouse antibody. Thus, for instance, a chimeric antibody derived from a mouse monoclonal antibody to the ganglioside GD2 and containing a human antibody FC region enhances the human effector cell-mediated antitumor effect [J. Immunol., 144, 1382 (1990)]. Similar results are reported for CDR-grafted antibodies [Nature, 332, 323 (1988)]. Such results indicate that, for clinical use, humanized monoclonal antibodies are preferred to mouse monoclonal antibodies.

[0005] The antibody classes include IgA, IgM, IgG, IgD and IgE and, in mice, the class IgG includes Four subclasses, namely IgG1, IgG2a, IgG2b and IgG3 (in humans, IgG1,IgG2, IgG3 and IgG4). When antigens are administered to animals, the antibodies produced mostly belong to the classes IgM or IgG. IgG molecules have a molecular weight of about 160,000 daltons and a dimeric structure and are relatively easy to handle. IgM molecules are large with a molecular weight of about 900,000 daltons and occur in the form of a complicated pentameric structure coupled with the joining (J) chain, hence they have the following drawbacks: they are difficult to purify; they tend to agglutinate, hence are difficult to store; they are readily inactivated by partial decomposition in the presence of a protease, hence it is difficult to prepare Fab fragments; and they lose their binding activity in many instances upon chemical modification, for example chemical binding of an anticancer agent or a toxin [J. W. Goding: Monoclonal Antibodies: Principles and Practice, Academic Press, 1986]. As to which are superior in therapeutic effect against cancer, IgG class monoclonal antibodies or IgM class monoclonal antibodies, reference may be made to a detailed study made by Bernstein et al. using an IgG class monoclonal antibody and an IgM class monoclonal antibody to the lymphocyte Thy-1 antigen [Monoclonal Antibodies, edited by R. H. Kennet, T. J. Mckearn and K. B. Bechtol, Plenum Press, 1980, p. 275]. According to the reference an IgG class monoclonal antibody and an IgM class monoclonal antibody comparable in terms of reactivity to Thy-1 antigen-positive lymphocytes, were compared in terms of antitumor effect. While the IgM monoclonal antibody was superior in in vitro complement-dependent antitumor effect, the IgG class monoclonal antibody showed a significant antitumor effect in in vivo antitumor effect in cancer-bearing mice, with no antitumor effect being observed with the IgM class monoclonal antibody. It was further revealed that, as compared with the IgG class monoclonal antibody, the IgM class monoclonal antibody showed a very short half-life in the blood after administration, in an isotope-labeled form, to mice. These results indicate that the monoclonal antibodies to be used clinically in humans should preferably be of the IgG class.

[0006] Gangliosides, a class of glycolipids, are constituents of animal cell membranes. These molecules are composed of a carbohydrate chain, which constitutes a hydrophilic side chain, and sphingosine and a fatty acid, which constitute hydrophobic side chains. It is known that the ganglioside species expressed and the amount thereof differ between cell species, organ species, and animal species, among others. Furthermore, it has been reported that the ganglioside expressed changed quantitatively and qualitatively during the process of cancer development [Cancer Res., 45, 2405 (1985)]. For example, expression of the gangliosides GD2, GD3 and GM2 has been reported in neuroblastoma, lung small cell carcinoma, and melanoma, which are highly malignant neural ectodermal tumors [J. Exp. Med., 155, 1133 (1982); J. Biol. Chem., 257, 12752 (1982); Cancer Res., 47, 225 (1987); ibid., 47, 1098 (1987); ibid., 45, 2642 (1985); Proc. Natl. Acad. Sci. U.S.A., 80, 5392 (1983)].

[0007] GM2, one of the gangliosides that are sialic acid residue containing glycolipids, occurs only in trace amounts in normal cells but is found in increased amounts in cancer cells in lung small cell carcinoma, melanoma, neuroblastoma, etc. Monoclonal antibodies to GM2 are considered to be useful in the treatment of such cancers [Lancer, 4, 786 (1989)]. However, those monoclonal antibodies to GM2 that have so far been reported are of the human IgM class or o: the rat IgM, mouse IgM or mouse IgG class [Cancer Res., 46, 4116 (1986); Proc. Natl. Acad. Sci. U.S.A., 79, 7629 (1982); Cancer Res., 48, 6154 (1988); J. Biol. Chem., 264, 12122 (1989)].

[0008] Anti-ganglioside GM2 monoclonal antibodies, if produced in the form of humanized antibodies, for example chimeric human antibodies or CDR-grafted antibodies, which are not expected to induce anti-mouse immunoglobulin antibody production in the patient's body, produce reduced adverse effects and show a prolonged blood half-life and an enhanced antitumor effector effect. These antibodies are thus expected to be superior in therapeutic effect to the corresponding mouse monoclonal antibodies.

SUMMARY OF THE INVENTION

[0009] It is an object of the present invention to provide humanized antibodies to the ganglioside GM2 (hereinafter, “humanized anti-GM2 antibodies”) which are useful in the treatment of cancers of neural ectodermal origin, among others.

[0010] The present inventors prepared the antibody heavy chain (hereinafter, “H chain”) variable region (hereinafter “VH”) cDNA and light chain (hereinafter, “L chain”) variable region (hereinafter, “VL”) cDNAs from mRNAs isolated from the hybridomas KM750 and KM796, described in EP-A-0 508 472. These hybridomas produce IgG3 class mouse monoclonal antibodies to the ganglioside CM2. VH and VL cDNAs were also prepared from mRNAs isolated from the hybridoma KM603, which produces an IgM class rat monoclonal antibody to the ganglioside GM2. Chimeric human antibody expression vectors were constructed by inserting the cDNA into an expression vector containing human antibody H chain constant region (hereinafter, “CH”) or human antibody L chain constant region (hereinafter, “CL”) encoding sequences. Such vectors were then introduced into animal cells to effect the production of anti-ganglioside GM2 chimeric human antibodies. Among the chimeric antibodies produced, the anti-ganglioside GM2 chimeric human antibody, KM966, was found to react with the ganglioside GM2 and show cytocidal activity. The H chain variable region of KM966 contains an amino acid sequence segment as defined by SEQ ID NO:1 and includes the 1st to 120th amino acids of that sequence and the L chain variable region of KM966 contains an amino acid sequence segment as defined by SEQ ID NO:2 and includes the 1st to 107th amino acids of said sequence. The present invention is based, at least in part, on these findings.

[0011] The present invention thus relates to a humanized antibody reacting with the ganglioside GM2.

BRIEF DESCRIPTION OF THE DRAWINGS

[0012] FIG. 1 illustrates plasmids, pKM796H1 and pKM796L1.

[0013] FIG. 2 illustrates plasmids, pKM750H1 and pKM750L1.

[0014] FIG. 3 illustrates plasmids, pKM603H1 and pKM603L1.

[0015] FIG. 4 shows a construction scheme for a plasmid, pAGE147.

[0016] FIG. 5 shows a construction scheme for a plasmid, pAGE148.

[0017] FIG. 6 shows a construction scheme for plasmids, pChi796HM1 and pChi750HM1.

[0018] FIG. 7 shows a construction scheme for plasmids, pChi796HMS1 and pChi750HMS1.

[0019] FIG. 8 shows a construction scheme for plasmids, pChi796LI1 and pChi750LI1.

[0020] FIG. 9 shows a construction scheme for plasmids, pChi796LM1 and pChi750LM1.

[0021] FIG. 10 shows a construction scheme for plasmids, pChi796LMS1 and pChi750LMS1.

[0022] FIG. 11 shows a construction scheme for plasmids, pChi796H107 and pChi750H107.

[0023] FIG. 12 shows a construction scheme for plasmids, pChi796HL1 and pChi750HL1.

[0024] FIG. 13 shows a construction scheme for a plasmid, pChi603HM1.

[0025] FIG. 14 shows a construction scheme for a plasmid, pChi603HMS1.

[0026] FIG. 15 shows a construction scheme for a plasmid, pChi603LI1.

[0027] FIG. 16 shows a construction scheme for a plasmid, pChi603LM1.

[0028] FIG. 17 shows a construction scheme for a plasmid, pChi603LMS1.

[0029] FIG. 13 shows the electrophoretic patterns in SDS-PAGE (using 4-15% gradient gels) or purified chimeric human anti-GM2 antibodies, KM966 and KM967. The patterns obtained under reducing conditions are shown on the left side and those obtained under nonreducing conditions on the right side. From the left, the lanes include low molecular weight markers, KM967 and KM966 (reducing conditions), and KM967 and KM966 (nonreducing conditions).

[0030] FIG. 19 shows the electrophoretic patterns in SDS-PAGE (using 4-15% gradient gels) of a purified chimeric human anti-GM2 antibody, KM968. The pattern obtained under reducing conditions is shown on the left side and that obtained under nonreducing conditions on the right side. From the left, the lanes include high molecular weight markers, low molecular weight markers, a standard human IgG, KM968 (reducing conditions), the same low molecular weight markers, the standard human IgG, and KM968 (nonreducing conditions)

[0031] FIG. 20 graphically shows the CDC (complement dependent cytotoxicity) activities of KM966 against the human lung small cell carcinoma cell lines SBC-3 and LU-135. The ordinate indicates the cytotoxic activity and the abscissa the concentration of the antibody added. The solid bars indicate the CDC activities of KM-696 and the shaded bars the CDC activities of KM966.

[0032] FIG. 21 graphically shows the CDC activities of KM966 against the human lung squamous cell carcinoma cell line PC-10 and human lung adenocarcinoma cell line RERF-LC-MS. The ordinate indicates the cytotoxicity and the abscissa the concentration of the antibody added. The solid bars indicate the CDC activities of KM-696 and the shaded bars the CDC activities of KM966.

[0033] FIG. 22 graphically shows the CDC activities of KM966 against the human lung large cell carcinoma cell line PC-13 and human neuroblastoma cell line NAGAI. The ordinate indicates the cytotoxic activity and the abscissa the concentration of the antibody added. The solid bars indicate the CDC activities of KM696 and the shaded bars the CDC activities of KM966.

[0034] FIG. 23 graphically shows the CDC activities of KM966 against the human neuroblastoma cell line GOTO and human brain tumor cell line A172. The ordinate indicates the cytotoxic activity and the abscissa the concentration of the antibody added. The solid bars indicate the CDC activities of KM696 and the shaded bars the CDC activities of KM966.

[0035] FIG. 24 graphically shows the ADCC (antibody dependent cell mediated cytotoxicity) activities of KM956 against the human lung small cell carcinoma cell lines SBC-3 and LU-135. The ordinate indicates the cytotoxic activity and the abscissa the concentration of the antibody added. The solid bars indicate the ADCC activities of KM-696 and the shaded bars the ADCC activities of KM966.

[0036] FIG. 25 graphically shows the ADCC activities of KM956 against the human lung squamous carcinoma cell line PC-10 and human lung adenocarcinoma cell line RERF-LC-MS. The ordinate indicates the cytotoxic activity and the abscissa the concentration of the antibody added. The solid bars indicate the ADCC activities of KM-696 and the shaded bars the ADCC activities of KM966.

[0037] FIG. 26 graphically shows the ADCC activities of KM966 against the human lung large cell carcinoma cell line PC-13 and human neuroblastoma cell line NAGAI. The ordinate indicates the cytotoxicity and the abscissa the concentration of the antibody added. The solid bars indicate the ADCC activities of KM-696 and the shaded bars the ADCC activities of KM966.

[0038] FIG. 27 graphically shows the ADCC activities of KM966 against the human neuroblastoma cell line GOTO and human brain tumor cell line A172. The ordinate indicates the cytotoxicity and the abscissa the concentration of the antibody added. The solid bars indicate the ADCC activities of KM-696 and the shaded bars the ADCC activities of KM966.

[0039] FIG. 28 shows a restriction enzyme cleavage map of a 9.3 kb XbaI fragment of the KM50 cell chromosomal DNA.

[0040] FIG. 29 shows a construction scheme for a plasmid, pKMB11.

[0041] FIG. 30 shows a construction scheme for a plasmid, pKMD6.

[0042] FIG. 31 shows a construction scheme for a plasmid, pEPKMA1.

[0043] FIG. 32 shows a construction scheme for a plasmid, pEPKMB1.

[0044] FIG. 33 shows a construction scheme for a plasmid, pAGE501.

[0045] FIG. 34 shows a construction scheme for a plasmid, pAGE109.

[0046] FIG. 35 shows a construction scheme for a plasmid, pAGE502.

[0047] FIG. 36 shows a construction scheme for a plasmid, pAGE503.

[0048] FIG. 37 shows a construction scheme for a plasmid, pSEd1.

[0049] FIG. 38 shows a construction scheme for a plasmid, pSEBd2.

[0050] FIG. 39 shows a construction scheme for a plasmid, pIG1SE1d2.

[0051] FIG. 40 shows a construction scheme for a plasmid, pIG1SE1d3.

[0052] FIG. 41 shows a construction scheme for a plasmid, pIG1SE1d4.

[0053] FIG. 42 shows a construction scheme for a plasmid, pPMOL2.

[0054] FIG. 43 shows a construction scheme for a plasmid, pPMOL3.

[0055] FIG. 44 shows a construction scheme for a plasmid, pchCKA7.

[0056] FIG. 45 shows a construction scheme for a plasmid, pchCKB1.

[0057] FIG. 46 shows a construction scheme for a plasmid, pckCKC1.

[0058] FIG. 47 shows a construction scheme for a plasmid, pChiIgHB2.

[0059] FIG. 48 shows a construction scheme for a plasmid, pChiIgLA1.

[0060] FIG. 49 shows a construction scheme for plasmids, pKM641HA3 and pKM641LA2.

[0061] FIG. 50 shows a construction scheme for a plasmid, pChi641HA1.

[0062] FIG. 51 shows a construction scheme for a plasmid, pKM641HE1.

[0063] FIG. 52 shows a construction scheme for a plasmid, pKM641HF1.

[0064] FIG. 53 shows a construction scheme for a plasmid, pChi641HA1.

[0065] FIG. 54 shows a construction scheme for a plasmid, pChi641HAM1.

[0066] FIG. 55 shows a construction scheme for a DNA, hKM796H.

[0067] FIG. 56 shows a construction scheme for a DNA, hKM796L.

[0068] FIG. 57 shows a construction scheme for a plasmid, phKM796HM1.

[0069] FIG. 58 shows a construction scheme for a plasmid, phKM796HMS1.

[0070] FIG. 59 shows a construction scheme for a plasmid, phKM796LI1.

[0071] FIG. 60 shows a construction scheme for a plasmid, phKM796LM1.

[0072] FIG. 60 shows a construction scheme for a plasmid, phKM796LMS1.

[0073] FIG. 62 shows a construction scheme for a plasmid, phKM796H107.

[0074] FIG. 63 shows a construction scheme for a plasmid, phKM796HL1.

[0075] FIG. 64 shows a construction scheme For a plasmid named pBSA.

[0076] FIG. 65 shows a construction scheme for a plasmid named pBSAE.

[0077] FIG. 66 shows a construction scheme for a plasmid named pBSH-S.

[0078] FIG. 67 shows a construction scheme for a plasmid named pBSK-H.

[0079] FIG. 68 shows a construction scheme for plasmids named pBSH-SA and pBSK-HA.

[0080] FIG. 69 shows a construction scheme for plasmids named pBSH-SAE and pBSK-HAE.

[0081] FIG. 70 shows a construction scheme for plasmids named pBSH-SAEE and pBSK-HAEE.

[0082] FIG. 71 shows a construction scheme for a plasmid named pBSK-HAEESal.

[0083] FIG. 72 shows a construction scheme for a plasmid named pBSX-S.

[0084] FIG. 73 shows a construction scheme for a plasmid named pBSX-SA.

[0085] FIG. 74 shows a construction scheme for a plasmid named pBSSC.

[0086] FIG. 75 shows a construction scheme for a plasmid named pBSMo.

[0087] FIG. 75 shows a construction scheme for a plasmid named pBSMoS.

[0088] FIG. 77 shows a construction scheme for a plasmid named pChiIgLA1S.

[0089] FIG. 78 shows a construction scheme for a plasmid named pMohCK.

[0090] FIG. 79 shows a construction scheme for a plasmid named pBSMoSal.

[0091] FIG. 80 shows a construction scheme for a plasmid named pESMoSalS.

[0092] FIG. 81 shows a construction scheme for a plasmid named pBShCy1.

[0093] FIG. 82 shows a construction scheme for a plasmid named pMohCy1.

[0094] FIG. 83 shows a construction scheme for a plasmid named pMoy1SP.

[0095] FIG. 84 shows a construction scheme for a plasmid named pMoky1SP.

[0096] FIG. 85 shows a construction scheme for a plasmid named pKANTEX93.

[0097] FIG. 86 shows a construction scheme for a plasmid named pBSNA.

[0098] FIG. 87 shows a construction scheme for a plasmid named pBSH3.

[0099] FIG. 88 shows a construction scheme for a plasmid named pBSES.

[0100] FIG. 89 shows a construction scheme for a plasmid named pBSL3.

[0101] FIG. 90 shows a construction scheme for a plasmid named pKANTEX796H.

[0102] FIG. 91 shows a construction scheme for a plasmid named pKANTEX796.

[0103] FIG. 92 shows a construction scheme for a plasmid named pT796.

[0104] FIG. 93 is a graphic representation of transient human anti-GM2 chimera antibody expression by the plasmids pKANTEX796 and pT796. The ordinate donotes the antibody concentration that showed GM2-binding activity, and the abscissa denotes the time after introduction of the plasmid.

[0105] FIG. 94 shows a construction scheme for a plasmid named pBSH10.

[0106] FIG. 95 shows a construction scheme for a plasmid named pBSL16.

[0107] FIG. 96 illustrates a process for mutagenesis by PCR and a process for cloning DNA fragments mutated.

[0108] FIG. 97 shows a construction scheme for a plasmid named pBSLV1+2.

[0109] FIG. 98 shows a construction scheme for a plasmid named pBSLm-28.

[0110] FIG. 99 shows a construction scheme for a plasmid named pBSHSGL.

[0111] FIG. 100 shows a construction scheme for a plasmid named pT796LCDR.

[0112] FIG. 101 shows a construction scheme for plasmids named pT796HLCDR, pT796HLCDRHV2 and pT796HLCDRHV4.

[0113] FIG. 102 shows a construction scheme for a plasmid named pT796HLCDRH10.

[0114] FIG. 103 shows construction scheme for plasmids named pT796HCDR, pT796HCDRHV2, pT796HCDRHV4 and pT796HCDR10.

[0115] FIG. 104 is a graphic representation of the results of human CDR-grafted anti-GM2 antibody activity evaluation in terms of transient expression as obtained using the plasmids pT796, pT796HCDR, pT796HCDRHV2, pT796HCDRHV4 and pT796HCDRH10. The ordinate denotes the plasmid used, and the abscissa denotes the relative activity value with the activity obtained with the chimera antibody being taken as 100%.

[0116] FIG. 105 shows a construction scheme for plasmids named pT796HLCDRLV1, pT796HLCDRLV2, pT796HLCDRLV3, pT796HLCDRLV4, pT796HLCDRLV8, pT796HLCDRLm-2, pT796HLCDRLm-8, pT796HLCDRLm-28 and pT796HLCDRHSGL.

[0117] FIG. 106 is a graphic representation of the results of human CDR-grafted anti-GM2 antibody activity evaluation in terms of transient expression as obtained using the plasmids pT796, pT796HLCDR, pT796HLCDRLV1, pT796HLCDRLV2, pT796HLCDRLV3, pT796HLCDRLV4, pT796HLCDRLV8, pT796HLCDRLm-2, pT796HLCDRLm-8, pT796HLCDRLm-28 and pT796HLCDRHSGL. The ordinate denotes the plasmid used, and the abscissa denotes the relative activity value with the activity obtained wish the chimera antibody being taken as 100%.

[0118] FIG. 107 shows a construction scheme for plasmids named pKANTEX796HLCDRLm-28 and pKANTEX796HLCDRHSGL.

[0119] FIG. 108 shows electrophoretic patterns obtained for human anti-GM2 chimera antibody KM966 and purified human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 by SDS-PAGE (4 to 15% gradient gels used). The patterns shown on the left side are those obtained under reducing conditions, and chose on the right under nonreducing conditions. From the left of each lane, the electrophoretic patterns for high-molecular-weight marker, KM966, KM8966, KM8967, low-molecular-weight marker, KM966, KM8960 and KM8967 are shown in that order.

[0120] FIG. 109 is a graphic representation of the GM2-binding activities of human anti-GM2 chimera antibody KM966 and purified human CDR-grafted anti-GM2 antibodies KM8966 and KM8967. The ordinate denotes the GM2-binding activity, and the abscissa the antibody concentration. The plots ◯, &Circlesolid; and Δ stand for the GM2-binding activities of KM966, KMS966 and KM8967, respectively.

[0121] FIG. 110 is a graphic representation of the reactivities of human anti-GM2 chimera antibody KM966 and purified human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 against various gangliosides. The ordinate denotes the ganglioside species, and the abscissa the binding activity. AcGM2 stands for N-acetyl-GM2, GcGM2 for N-glycolyl-GM2, AcGM3 for N-acetyl-GM3 and GcGM3 for N-glycolyl-GM3. The plots □, ▪ and □ stand for the reactivities of KM966, KM8966 and KM8967, respectively.

[0122] FIG. 111 is a graphic representation of the reactivities of human anti-GM2 chimera antibody KM966 and purified human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 against the human lung small cell carcinoma cell line SBC-3. The ordinate denotes the number of cells, and the abscissa the fluorescence intensity. From the lowermost graph, the reactivities of control, KM8967, KM8966 and KM966 are shown in that order.

[0123] FIG. 112 graphically shows the CDC activities of chimeric human anti-GM2 antibody KM966 and purified human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 against the human lung small cell carcinoma cell line SBC-3. The ordinate indicates the cytotoxic activity and the abscissa the concentration of the antibody. The plots □, ▪ and □ stand for the CDC activities of KM966, KM8966 and KM8967, respectively.

[0124] FIG. 113 graphically shows the ADCC activities of chimeric human anti-GM2 antibody KM966 and purified human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 against the human lung small cell carcinoma cell line SBC-3. The ordinate indicates the cytotoxicity and the abscissa the concentration of the antibody The plots □, ▪ and □ stand for the ADCC activities of KM966, KM8966 and KM8967, respectively.

[0125] FIG. 114 shows a construction scheme for plasmids, pKANTEX796HM1Lm-28, pKANTEX796HM2Lm-28, pKANTEX796HM3Lm-28, pKANTEX796HM31Lm-28 and pKANTEX796HM32Lm-28.

[0126] FIG. 115 shows the electrophoretic patterns in SDS-PAGE (using 4-15% gradient gels) of chimeric human anti-GM2 antibody KM966, human CDR-grafted anti-GM2 antibody KM8966 and human CDR-grafted anti-GM2 antibodies each having different mutation. The pattern obtained under nonreducing conditions is shown on the left side and that obtained under reducing conditions on the right side. M stands for molecular weight markers (from the top, the arrows indicate the molecular weight of 205 Kd, 140 Kd, 83 Kd, 45 Kd, 32.6 Kd, 18 Kd and 7.5 Kd in that order) and 1, 2, 3, 4, 5, 6 and 7 stand for the electrophoretic patterns of KM996, KM8966, M1-28, M2-28, M3-28, M31-28 and M32-28, respectively.

[0127] FIG. 116 graphically shows the CDC activities of chimeric human anti-GM2 antibody KM966, human CDR-grafted anti-GM2 antibody KM8966 and human CDR-grafted anti-GM2 antibodies each having different mutation against the human lung small cell carcinoma cell line SBC-3. The ordinate indicates the cytotoxic activity and the abscissa the concentration of the antibody. The plots □, ▪, □, □, □, □ and □ stand for the CDC activities of KM966, KM8966, M1-28, M2-28, M3-28, M31-28 and M32-28, respectively.

[0128] FIG. 117 shows a construction scheme for plasmids, pKANTEX796HLm-28 No.1, pKANTEX796HM1Lm-28 No.1, pKANTEX796HM2Lm-28 No.1 and pKANTEX796HM3Lm-28 No.1.

[0129] FIG. 118 shows the electrophoretic patterns in SDS-PAGE (using 4-15% gradient gels) of chimeric human anti-GM2 antibody KM966 and human CDR-grafted anti-GM2 antibodies each having different mutation. The pattern obtained under nonreducing conditions is shown on the left side and that obtained under reducing conditions on the right side. M stands for molecular weight markers (from the top, the arrows indicate the molecular weight of 205 Kd, 140 Kd, 83 Kd, 45 Kd, 32.6 Kd, 18 Kd and 7.5 Kd in that order) and 1, 2, 3, 4 and 5 stand for the electrophoretic patterns of KM966, h796H-No.1, M1-No.1, M2-No.1 and M3-No.1, respectively.

[0130] FIG. 119 graphically shows the CDC activities of chimeric human anti-GM2 antibody KM966, human CDR-grafted anti-GM2 antibodies KM8966 and KM8970 and human CDR-grafted anti-GM2 antibodies each having different mutation against the human lung small cell carcinoma cell line SBC-3. The ordinate indicates the cytotoxic activity and the abscissa the concentration of the antibody. The plots □, ▪, □, □, □, □ and □ stand for the CDC activities of KM966, KM8966, KM8970, h796H-No.1, M1-No.1, M2-No.1 and M3-No.1, respectively.

[0131] FIG. 120 graphically shows the GM2-binding activities of chimeric human anti-GM2 antibody KM966 and human CDR-grafted anti-GM2 antibodies KM8969 and KM8970. The ordinate indicates the GM2-binding activity and the abscissa the concentration of the antibody. The plots ◯, &Circlesolid; and ▴ stand for the GM2-binding activities of KM966, KM8969 and KM8970, respectively.

[0132] FIG. 121 graphically shows the reactivities of chimeric human anti-GM2 antibody KM966 and human CDR-grafted anti-GM2 antibodies KM8969 and KM8970 against various gangliosides. The ordinate indicates the ganglioside species and the abscissa the binding activity. AcGM2 stands for N-acetyl-GM2, GcGM2 for N-glycolyl-GM2, AcGM3 for N-acetyl-GM3 and GcGM3 for N-glycolyl-GM3. The plots □, ▪ and □ stand for the GM2-binding activities of KM966, KM8969 and KM8970, respectively.

[0133] FIG. 122 graphically shows the reactivities of chimeric human anti-GM2 antibody KM956 and human CDR-grafted anti-GM2 antibodies KM8969 and KM8970 against the human lung small cell carcinoma cell line SBC-3. The ordinate indicates the number of cells and the abscissa the fluorescence intensity. From the lowermost graph, the reactivities of control, KM966, KM8970 and KM8969 are shown in that order.

[0134] FIG. 123 graphically shows the ADCC activities of chimeric human anti-GM2 antibody KM966 and human CDR-grafted anti-GM2 antibodies KM 8966, KM8969 and KM8970 against the human lung small cell carcinoma cell line SBC-3. The ordinate indicates the cytotoxicity and the abscissa the concentration of the antibody. The plots □, ▪, □ and □ stand for the ADCC activities of KM966, KM8966, KM8969 and KM8970, respectively.

[0135] FIG. 124 graphically shows the CDC activities of chimeric human anti-GM2 antibody KM966 and human CDR-grafted anti-GM2 antibodies KM 8966, KM8969 and KM8970 against the human lung small cell carcinoma cell line SBC-3 obtained when the reaction was carried out for 1 hour and 4 hours after the addition of the human complement. The ordinate indicates the cytotoxicity and the abscissa the concentration of the antibody. The plots □, ▪, □, □, □, □, □ and □ stand for the CDC activities of KM966 (1 hour), KM966 (4 hours), KM8966 (1 hour), KM8966 (4 hours), KM966 (1 hour) KM8969 (4 hours), KM8970 (1 hour) and KM8970 (4 hours).

DETAILED DESCRIPTION OF THE INVENTION

[0136] The present invention relates to humanized antibodies specific for the ganglioside GM2. The antibodies can be of any of the immunoglobulin (Ig) classes, it is preferable, however, that the antibodies be of the IgG type. The term “humanized antibody”, as used herein, includes within its meaning, chimeric human antibodies and CDR-grafted antibodies. Chimeric human antibodies of the invention include the VH and VL of an antibody of an animal other than a human and the CH and CL of a human antibody. The CDR-transplanted antibodies of the invention result from the replacement of CDRs of the VH and VL of a human antibody with those of the VH and VL, respectively, of an antibody of an animal other than a human.

[0137] An example of a chimeric human antibody of the invention is an antibody the VH of which contains an amino acid sequence segment as defined by SEQ ID NO:1, including the 1st to 120th amino acids of that sequence, and the VL of which contains an amino acid sequence segment as defined by SEQ ID NO:2, including the 1st to 107th amino acids of that sequence.

[0138] An example of a CDR-grafted antibody of the invention is an antibody the VH CDRs of which have the amino acid sequences defined by SEQ ID NO:6, SEQ ID NO:7 and SEQ ID NO:8 and the VL CDRs of which have the amino acid sequences defined by SEQ ID NO:9, SEQ ID NO:10 and SEQ ID NO:11.

[0139] The chimeric human antibodies of the invention can be produced in the following manner:

[0140] (1) Preparation of cDNAs Coding for the VH and VL of an Antibody of Nonhuman Animal

[0141] cDNAs coding for the VH and VL of an antibody of a nonhuman animal, for example a mouse anti-GM2 monoclonal antibody, can be prepared as follows.

[0142] mRNAs can be extracted from hybridomas producing the mouse anti-GM2 monoclonal antibody, for example hybridomas producing the mouse anti-GM2 monoclonal antibody KM796, and cDNAs reverse transcribed therefrom. Using the cDNAs, a library can be constructed using phage or plasmid vectors. The recombinant phage or recombinant plasmid containing the cDNA coding for the VH, and the recombinant phage or recombinant plasmid containing the cDNA coding for the VL can be isolated from the library using a constant region portion or a variable region portion of an antibody of a nonhuman animal, for example a mouse antibody, as a probe. The base sequences of the VH-encoding cDNA and VL-encoding cDNA in the recombinant phage or recombinant plasmid can then be determined. Examples of the nonhuman animals include mice, rats, hamsters and monkeys.

[0143] (2) Construction of a Vector for Chimeric Human Antibody Expression

[0144] Expression of chimeric human antibody H chain and L chains can be effected using expression vectors suitable for use in animal cells, inserted into which are the cDNAs coding for the human CH and CL. Any expression vector suitable for use in animal cells can be used, provided that it allows integration and expression of the human antibody constant region-encoding cDNAs. Examples include pAGE107 [Cytotechnology, 3, 133 (1990)], pAGE103 [J. Biochem., 101, 1307 (1987)], pHSG274 [Gene, 27, 223 (1984)], pKCR. [Proc. Natl. Acad. Sci. U.S.A., 78, 1527 (1981)] and pSG1βd2-4 [Cytotechnology, 4, 173 (1990)], among others. Examples of promoters and enhancers suitable for use in such expression vectors include the SV40 early promoter and enhancer [J. Biochem., 101, 1307 (1987)], the Moloney mouse leukemia virus LTR (long terminal repeat) promoter and enhancer [Biochem. Biophys. Res. Commun., 149, 960 (1987)] and the immunoglobulin H chain promoter [Cell, 41, 479 (1985)] and enhancer [Cell, 33, 717 (1983)]. The promoters and enhancers are located in the expression vector in operable linkage with the coding sequences.

[0145] (3) Construction of a Chimeric Human Antibody Expression Vector

[0146] The vector for chimeric human antibody H chain and L chain expression, as obtained in (2), is provided with a cloning site upstream of the human constant region, for insertion of a cDNA coding for the variable region of an antibody of a nonhuman animal. Insertion, at this cloning site, of the cDNA coding for the variable region of a nonhuman animal antibody, using a synthetic DNA comprising a 5′ terminal base sequence of the human antibody constant region and a 3′ terminal base sequence of the variable region of the nonhuman animal antibody and having restriction enzyme sites on both ends, gives a chimeric human antibody expression vector with the cDNA coding for the human antibody constant region and the cDNA coding for the variable region of the nonhuman animal antibody joinedly inserted therein via the synthetic DNA for producing appropriate restriction enzyme sites. The synthetic DNA can be synthesized using a DNA synthesizer based on the 5′ terminal base sequence of the human antibody constant region and the base sequence of said 3′ terminal base sequence of the nonhuman animal antibody variable region.

[0147] (4) Construction of a Chimeric Human Antibody H Chain Expression Vector

[0148] A vector for chimeric human antibody H chain expression is constructed, for example, by excising that portion of the human antibody CH-encoding cDNA which covers from the ApaI site near the 5′ terminus to the 3′ terminus and inserting that portion into an expression vector suitable for use in animal cells. This vector for chimeric human antibody H chain expression is provided with a cloning site for insertion of a cDNA coding for a nonhuman animal VH. cDNA coding for the nonhuman animal VH, excised using an appropriate restriction enzyme, is inserted into the vector at the cloning site using a synthetic DNA comprising that portion of the human antibody CH gene which covers from the 5′ terminus to the ApaI site and the base sequence of a 3′ terminal portion of the nonhuman animal antibody VH gene and having restriction enzyme sites on both ends, to give a chimeric human antibody H chain expression vector which allows no change in the amino acid sequence so VH upon expression thereof and has appropriate restriction enzyme sites.

[0149] (5) Construction of a Chimeric Human Antibody L Chain Expression Vector

[0150] A vector for chimeric human antibody L chain expression is constructed, for example by introducing an EcoRV site into the human antibody CL-encoding cDNA in the vicinity of the 5′ terminus by mutagenesis, excising that portion which covers from the EcoRV site to the 3′ terminus and inserting that portion into a plasmid, such as the plasmid pIg1SE1d4. This vector for chimeric human antibody L chain expression is provided with a cloning site for insertion of the cDNA coding for nonhuman animal VL. The nonhuman animal antibody VL-encoding cDNA, excised with an appropriate restriction enzyme, is inserted into the vector at the cloning site using a synthetic DNA comprising that portion of the human antibody CL gene which covers from the 5′ terminus to the EcoRV site and the base sequence of a 3′ terminal portion of the nonhuman animal antibody VL gene and having restriction enzyme sites on both ends, to give a chimeric human antibody L chain expression vector which allows no change in the amino acid sequence of VL upon expression thereof.

[0151] (6) Introduction of the Chimeric Human Antibody Expression Vectors Into Host Cells

[0152] Introduction of the chimeric human antibody H chain expression vector and chimeric human antibody L chain expression vector into host cells gives a transformant producing the chimeric human antibody. In introducing the vectors into host cells, a splicing signal may be introduced into the chimeric human antibody H chain and L chain expression vectors for mRNA stabilization [Cell, 17, 737 (1979)].

[0153] The chimeric human antibody H chain and L chain vectors can be introduced into host cells, for example, simultaneously by electroporation [JP-A-2-257891 (the term “JP-A” used herein means an unexamined published Japanese patent application.); Cytotechnology, 3, 133 (1990)]. In addition, an expression vector containing genes coding for both the chimeric human antibody H chain and L chain [tandem expression vector] can be introduced into host cells [BIO/TECHNOLOGY, 9, 64 (1991)]. The use of a tandem expression vector is preferred since a higher level of chimeric human antibody expression can be attained thereby, with approximately equal H chain and L chain expression levels.

[0154] An example of a method of producing the CDR-grafted antibodies of the invention is described as follows.

[0155] First, a CDR-grafted antibody expression vector can be constructed by the method of Winter et al. [Nature, 332, 323 (1988)] as follows.

[0156] Three synthetic DNAs are constructed designed so as to comprise the cDNAs coding for three CDR peptides of the VH of a nonhuman animal antibody, for example, peptides having the amino acid sequences defined by SEQ ID NO:5, SEQ ID NO:7 and SEQ ID NO:8, with DNAs coding for amino acid sequences comprising of several amino acids from both ends of the corresponding CDRs of the VH of a human antibody being located at the respective both ends of the cDNAs, DNA synthesis is carried out with a plasmid containing the human antibody VH gene as a template. An example of the human antibody VH gene-containing plasmid is the M13 plasmid containing a human antibody NEW gene-derived sequence [J. Biol. Chem., 253, 585 (1978); Nature, 332, 323 (1988)].

[0157] The DNA obtained is inserted into the vector for chimeric human antibody H chain expression in the same manner as in the construction of the chimeric human antibody expression vector mentioned above to give a CDR-grafted antibody H chain expression vector.

[0158] Similarly, using, as primers, three synthetic DNAs designed to comprise the cDNAs coding for three CDR peptides of the VL of a nonhuman animal antibody, for example, the peptides having the amino acid sequences defined by SEQ ID NO:9, SEQ ID NO:10 and SEQ ID NO:11, with DNAs coding for amino acid sequences comprising several amino acids from both ends of the corresponding CDRs of the human antibody VL being located at the respective both ends of said cDNAs, DNA synthesis is carried out with a human antibody VL gene-containing plasmid as a template. An example of the human antibody VL gene-containing plasmid is the M13 plasmid containing a human myeloma protein (Bence-Jones protein) REI gene-derived sequence [Eur. J. Biochem., 45, 513 (1974); Nature, 332, 323 (1988)].

[0159] By inserting the DNA obtained into a vector for chimeric human L chain expression in the same manner as described in respect of the construction of the chimeric human antibody expression vector, a CDR-grafted antibody L chain expression vector can be constructed.

[0160] It is also possible to construct the CDR-grafted antibody H chain and L chain expression vectors by synthesizing DNAs coding for the peptides having amino acid sequences resulting from replacement of the three CDRs each of the H chain and L chain of a human antibody with the corresponding CDRs of the H chain and L chain of a nonhuman animal antibody and then inserting the DNAs into a vector for chimeric human antibody H chain or L chain expression in the same manner as described in respect of the construction of the chimeric human antibody expression vector mentioned above.

[0161] The CDR-grafted antibody expression vector can be introduced into host cells in the same manner as the chimeric human antibody expression vector to give a transformant producing the CDR-grafted antibody.

[0162] The host cells suited for the introduction thereinto of the chimeric human antibody or CDR-grafted antibody expression vector may be any host cells provided that the chimeric human antibody or CDR-grafted antibody can be expressed therein. Examples include mouse SP2/0-Ag14 cells (ATCC CRL1581; hereinafter, “SP2/0 cells”), mouse P3X63-Ag8.653 cells (ATCC CRL1580), CHO cells deficient in the dihydrofolate reductase gene (hereinafter, “dhfr”) [Urlaub et al.: Proc. Natl. Acad. Sci. U.S.A., 77, 4216 (1980)] and rat YB2/3HL.P2.G11.16Ag.20 cells (ATCC CRL1662; hereinafter, “YB2/0 cells”), with YB2/0 cells being preferred.

[0163] The transformants producing the chimeric human antibody or CDR-grafted antibody are selected by the method disclosed in JP-A-2-257891 using PRMI1640 medium containing G418 and fetal calf serum. A particular example of the chimeric human antibody-producing transformant is the transformant KM966 producing a chimeric human antibody that reacts with the ganglioside GM2. Examples of human CDR-grafted antibody-producing transformants include the transformants KM8966 and KM8967 each producing a human CDR-grafted antibody that reacts with the ganglioside GM2. KM966 has been deposited with the Fermentation Research Institute, Agency of Industrial Science and Technology, of 1-3, Higashi 1-chome, Tsukuba-shi, Ibaraki 305 JAPAN, as of Jul. 15, 1992 under the deposit number FERM BP-3931. KM8966 and KM8967 have also deposited with the above-described institute as of May 23, 1995 under the deposit numbers FERM BP-5105 and FERM BP-5106, respectively.

[0164] When the transformant obtained is cultivated in a medium, the chimeric human antibody or CDR-grafted antibody can be produced and accumulated in the culture fluid. The activity of the chimeric human antibody or CDR-grafted antibody in the medium can be determined by an enzyme-linked immunosorbent assay (ELISA; E. Harlow et al. (ed.) Antibodies—A Laboratory Manual, Cold Spring Harbor Laboratory, 1988). The antibody productivity of the transformant can be increased by utilizing a dhfr amplification system as disclosed in JP-A-2-257891.

[0165] The chimeric human antibody and CDR-grafted antibody can be purified from the culture supernatants obtained as mentioned above using a protein A column (E. Harlow et al. (ed.): Antibodies—A Laboratory Manual, Cold Spring Harbor Laboratory, 1988). As noted above, the chimeric human antibody KM966, which reacts with the ganglioside GM2, is a specific example of the thus-obtained chimeric human antibodies and CDR-grafted antibodies.

[0166] The reactivity of the chimeric human antibody or CDR-grafted antibody of the invention can be checked by ELISA. The molecuar weight of the purified antibody H chain or L chain or whole antibody molecule can be determined by polyacrylamide gel electrophoresis (SDS-PAGE) or Western blotting (E. Harlow et al. (ed.): Antibodies—A Laboratory Manual, Cold Spring Harbor Laboratory, 1988).

[0167] The binding activity of the chimeric human antibody or CDR-grafted antibody that reacts with the ganglioside GM2 of cultured cancer cells can be measured, for example, by the fluorescent antibody technique or by ELISA. The complement dependent cytotoxic activity (CDC activity) and antibody dependent cell mediated cytotoxic activity (ADCC activity) of the chimeric human antibody or CDR-grafted antibody are measured by the methods of Ohta et al. [Cancer Immunol. Immunother., 35, 260 (1993)].

[0168] The humanized antibodies of the invention specifically bind to human cancer cells and exhibit CDC activity and ADCC activity against human cancer cells and therefore are useful in the treatment of human cancers, among others.

[0169] The humanized antibodies according to the present invention can be used alone as an anticancer agent. They may be formulated into an anticancer composition together with at least one pharmaceutically acceptable carrier. For instance, the humanized antibodies are dissolved in physiological saline, an aqueous solution of glucose, lactose or mannitol and the like. The powder of the humanized antibodies for injection can be prepared by lyophilizing the humanized antibodies in accordance with the conventional method and mixing the lyophilized products with sodium chloride. The anticancer composition may further contain additives conventionally used well known in the art of medical preparation, for example, pharmaceutically acceptable salts.

[0170] The humanized antibodies according to the present invention can be administered in the form of the above-described anticancer composition to mammals including human in a dose of 0.2 to 20 mg/kg/day. The dose may vary depending on the age, condition, etc. of patients. The administration of the anticancer composition can be effected by intravenous injection once a day (single administration or consecutive administration) or intermittently one to three times a week or once every two to three weeks.

[0171] The anticancer composition is expected to be useful for treating cancer such as melanoma, neuroblastoma and glioma.

[0172] The following Examples and Reference Examples are further illustrative of the present invention, but are not to be construed to limit the scope of the present invention.

EXAMPLE 1

[0173] Production of chimeric human anti-GM2 antibodies 1. Isolation of mRNAs from hybridoma cells producing the mouse anti-GM2 monoclonal antibody KM-796 or KM-750 and from hybridoma cells producing the rat anti-GM2 monoclonal antibody KM-603

[0174] Using mRNA extraction kit Fast Track (product number K1593-02) manufactured by Invitorogen and following the description of the manual attached to the kit, mRNAs were isolated from 1×103 cells each of the mouse anti-GM2 monoclonal antibody KM-796-producing hybridoma cell line (FERM BP-3340), the mouse anti-GM, monoclonal antibody KM-750-producing hybridoma cell line (FERM BP-3339) and the rat anti-GM2 monoclonal antibody KM-603-producing hybridoma cell line (FERM BP-2636).

[0175] 2. Construction of Monoclonal Antibody KM-796 and KM-750 H Chain and L Chain cDNA Libraries

[0176] Using cDNA Synthesis Kit (product number 27-9260-01) manufactured by Pharmacia and following the manual attached to the kit, cDNA having the EcoRI adapter on both ends was synthesized from 5 μg each of the KM-796- and KM-750-derived mRNAs obtained as described in Paragraph 1 above. About 6 μg of each cDNA product obtained was dissolved in 10 μl of sterilized water and fractionated by agarose gel electrophoresis, and a cDNA fragment- (about 1.8 kb) corresponding to the IgG antibody H chain and a cDNA fragment (about 1.0 kb) corresponding to the L chain were recovered (about 0.1 μg each). Then, 0.1 μg of each cDNA fragment of about 1.8 kb and 0.1 μg of each cDNA fragment of about 1.0 kb were respectively dissolved in 11.5 μl of T4 ligase buffer, together with 1 μg of the Lambda ZAPII vector (cleaved with EcoRI and then treated with calf intestine alkaline phosphatase; product of Stratagene) After addition of 175 units of T4 DNA ligase, each solution was incubated at 12° C. for 24 hours and then at room temperature for 2 hours. A 4-μl portion of each reaction mixture was subjected to packaging into the lambda phage in the conventional manner [Maniatis et al. (ed.) : Molecular Cloning, 2.95 Cold pring Harbor Laboratory, 1989] using Giga Pak Gold (Stratagene), followed bay transfection, in the conventional manner [Maniatis et al. (ed.): Molecular Cloning, 2.95-107, Cold Spring Harbor Laboratory, 1989] of the Escherichia coli strain XL1-Blue [Biotechniques, 5, 376 (1987)] attached to Giga Pak Gold, to give about 4×103 phage clones each as a KM-796 or KM-750 H chain or L chain cDNA library. Then the phage clones of each library were fixed on a nitrocellulose filter in the conventional manner [Maniatis et al. (ed.) Molecular Cloning, 2.112, Cold Spring Harbor Laboratory, 1989].

[0177] 3. Construction of KM-603 H Chain and L Chain cDNA Libraries

[0178] Using 5 μg of the KM-603 mRNA obtained as mentioned above in Paragraph 1 and cDNA Synthesis Kit4 (product number 27-9260-01) manufactured by Pharmacia, cDNA having the EcoRI adapter on both ends was synthesized. About 6 μg of the cDNA produced was dissolved in 10 μl of sterilized water and fractionated by agarose gel electrophoresis. A cDNA fragment (about 2.2 kb) corresponding to the IgG antibody H chain and a cDNA fragment (about 1.0 kb) corresponding to the L chain were recovered (about 0.1 μg each). Then 0.1 μg of the cDNA fragment of about 2.2 kb and 0.1 μg of the cDNA fragment of about 1.0 kb were respectively dissolved in 11.5 μl of T4 ligase buffer, together with 1 μg of the Lambda ZAPII vector (cleaved with EcoRI and then treated with calf intestine alkaline phosphatase; product of Stratagene) and, after addition of 175 units of T4 DNA ligase, the resultant solution was incubated at 12° C. for 24 hours and then at room temperature for 2 hours. A 4-μl portion of each reaction mixture was subjected to packaging into the lambda phage in the conventional manner [Maniatis et al. (ed.): Molecular Cloning, 2.95, Cold Spring Harbor Laboratory, 1989] using Giga Pak Gold (Stratagene), followed by transfection, in the conventional manner [Maniatis et al. (ed.) : Molecular Cloning, 2.95-107, Cold Spring Harbor Laboratory, 1989], of the Escherichia coli strain XL-Blue attached to Giga Pak Gold, whereby about 1×104 phage clones were obtained each as a KM-603 H chain or L chain cDNA library. Then, the phage clones of each library were fixed on a nitrocellulose filter in the conventional manner [Maniatis et al. (ed.) : Molecular Cloning, 2.112, Cold Spring Harbor Laboratory, 1989].

[0179] 4. Cloning of the KM-790 and KM-750 H Chain and L Chain cDNAs

[0180] From among the KM-796 and KM-750 H chain cDNA libraries and L chain cDNA libraries constructed as described above in Paragraph 2, phage clones firmly bound at 65° C. to a probe prepared by labeling a mouse immunoglobulin constant region cDNA [for the H chain, the BamHI-XhoI fragment of the mouse Cy3 cDNA (Wels et al: EMBO J., 3, 2041-206, 1984); for the L chain, the HpaI-XhoI fragment of the mouse CK cDNA (Hieter et at.: Cell, 22, 197-207, 1980)] with 32P were recovered in the conventional manner [Maniatis et al.: Molecular Cloning, 2.108, Cold Spring Harbor Laboratory, 1989]. Then, using a ZAP-cDNA Synthesis Kit (cDNA synthesis kit; product number sc200400) manufactured by Stratagene, phage clones were converted into pBluescript plasmids, and a KM-796 H chain cDNA-containing recombinant plasmid (pKM796H1) and a KM-796 L chain cDNA-containing recombinant plasmid (pKM796L1) (FIG. 1) as well as a KM-750 H chain cDNA-containing recombinant plasmid (pKM750H1) and a KM-750 L chain cDNA-containing recombinant plasmid (pKM750L1) (FIG. 2) were obtained. Cleavage of pKM796H1, pKM750H1, pKM796L1 and pKM750L1 with EcoRI revealed that a cDNA fragment of about 1.8 kb had been inserted into pKM796H1 and pKM750H1 and a cDNA fragment of about 0.9 kb into pKM796L1 and pKM750L1.

[0181] 5. Cloning of KM-603 H Chain and L Chain cDNAs

[0182] Phage clones firmly bound at 65° C. to a probe prepared by labeling a mouse immunoglobulin constant region chromosomal gene [mouse Cμ gene-containing SmaI-KpnI fragment of about 11.5 kb (Kataoka et al.: Proc. Natl. Acad. Sci. U.S.A., 77, 919-923, 1980) and mouse Cκ gene-containing HindIII-BamHI fragment of about 3 kb [(Sakano et al.: Nature, 280, 288, 1979)] with 32P were isolated from the KEM-603 H chain cDNA library and L chain cDNA library constructed as mentioned above in Paragraph 3 in the conventional manner [Maniatis et al. (ed.): Molecular Cloning, 2.108, Cold Spring Harbor Laboratory, 1989]. Then, using ZAP-cDNA Synthesis kit (product number sc200400) manufactured by Stratagene, the phage clones were converted to pBluescript plasmids and a KM-603 H chain cDNA-containing recombinant plasmid, pKM603H1, and a KM-603 L chain cDNA-containing recombinant plasmid, pKM603L1, were obtained (FIG. 3). Cleavage of pKM603H1 and pKM603L1 revealed that pKM603H1 contained a cDNA fragment of about 2.0 kb as inserted therein and pKM603L1 a cDNA fragment of about 0.9 kb as inserted therein.

[0183] 6. Base Sequences of the Variable Regions in the H Chain cDNA and L Chain cDNA

[0184] The base sequences of the variable regions in the H chain cDNA and L chain cDNA obtained as mentioned above in Paragraphs 4 and 5 were determined by the dideoxy method [Maniatis et al. (ed.): Molecular Cloning, 13.42, Cold Spring Harbor Laboratory, 1989] using Sequenase Version 2.0 DNA Sequencing Kit manufactured by United States Biochemical Corporation. All the cDNA had a methionine codon, presumably the initiation codon ATG, at the 5′ terminus and were leader sequence-containing full-length cDNAs. Based on the base sequences of the respective cDNAs, the amino acid sequences of the H chain and L chain of KM!-796, KM-750 and KM-603 were deduced. The amino acid sequence of the KM-795 H chain is shown in SEQ ID NO:1, that of the L chain of KM-796 and KM-750 in SEQ ID NO:2, that of the KM-750 H chain in SEQ ID NO:3, that of the KM-603 H chain in SEQ ID NO:4 and that of the KM-603 L chain in SEQ ID NO:5.

[0185] 7. Construction of KM-796- and KM-750-Derived Chimeric Human Antibody H Chain and L Chain Expression Vectors

[0186] (1) Construction of a Vector, pAGE147, Carrying the Moloney Mouse Leukemia Virus Terminal Repeat Promoter/enhancer

[0187] The plasmid pPMOL1 (2 μg), described in JP-A-1-63394, was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 6 mM 2-mercaptoethanol, 20 units of SmaI was added, and digestion was carried out at 30° C. for 3 hours. Then, sodium chloride was added to a concentration of 50 mM, 20 units of ClaI was added, and digestion was conducted at 37° C. for 2 hours. The reaction mixture was subjected to agarose gel electrophoresis, and a DNA fragment (about 0.6 kb) containing the Moloney mouse leukemia virus terminal repeat promoter/enhancer was recovered.

[0188] Then, the following two synthetic DNAs were synthesized using an automatic DNA synthesizer (model 380A manufactured by Applied Biosystems Co., Ltd.).

XhoI  (9 bases)
|              |
5-′-   TCGA    GGACC   -3′
|              |
3? - |  CCTGGGC  |  -5′
(7 bases)

[0189] The thus-obtained synthetic DNAs (25 picomoles each) were dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.6) containing 10 mM magnesium chloride, 5 mM DTT (dithiothreitol), 0.1 mM EDTA and 0.5 mM adenosine triphosphate (hereinafter, “ATP”), 5 units of T4 DNA kinase was added, and 5′-phosphorylation was carried out at 37° C. for 30 minutes.

[0190] The plasmid pPMOL1-derived ClaI-SmaI fragment (0.6 kb, 0.05 μg) and two 5′-phosphorylated synthetic DNAs (1 picomole each), obtained as described above, and a HindIII linker (5′-pCAAGCTTG-3′; Takara Shuzo) (1 picomole) were dissolved in 30 μl of 66 mM Tris-hydrochloride buffer (pH 7.5) containing 6.6 mM magnesium chloride, 10 mM DTT and 0.1 mM ATP, 200 units of T4 DNA ligase (Takara Shuzo; hereinafter the same shall apply) were added, and ligation was carried out at 12° C. for 16 hours. The resultant DNA fragment was recovered by ethanol precipitation and dissolved in 20 μl of 10 mM. Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride, 100 mM sodium chloride and 6 mM 2-mercaptoethanol, 10 units of HindIII and 10 units of XhoI were added, and digestion was carried out at 37° C. for 2 hours. The reaction was terminated by phenol-chloroform extraction, and the DNA fragment was recovered by ethanol precipitation.

[0191] Separately, 1 μg of the plasmid pAGE107 [Cytotechnology, 3, 133 (1990)] was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 ′ magnesium chloride, 100 mM sodium chloride and 6 mM 2-mercaptoethanol, 10 units of HindIII and 10 units of XhoI were added, and digestion was carried out at 37° C. for 2 hours. The reaction mixture was subjected to agarose gel electrophoresis, and a DNA fragment (about 6.0 kb) containing the G418 resistance gene and ampicillin (hereinafter, “Ap”) resistance gene was recovered.

[0192] The plasmid pAGE107-derived HindIII-XhoI fragment (6.0 kb, 0.3 μg) and plasmid pPMOL1-derived HindIII-XhoI fragment (0.63 kb, 0.01 μg) obtained as mentioned above were dissolved in 20 μl of 66 mM Tris-hydrochloride buffer (pH 7.5) containing 6.6 mM magnesium chloride, 10 mM DTT and 0.1 mM ATP, 200 units of T4 DNA ligase were added, and ligation was carried out at 12° C. for 16 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101, and the plasmid pAGE147 shown in FIG. 4 was obtained.

[0193] (2) Construction of a Vector, pAGE148, Carrying the β-globin 3′ Splicing Signal (SP)

[0194] For introducing the β-globin 3′ splicing signal into the chimeric human antibody expression vector at a site downstream from the antibody constant region gene, a vector (pAGE148), was constructed as follows, which contained the β-globin 3′ splicing signal and the same genes as those in the chimeric human antibody expression vector (except for the human antibody constant region gene).

[0195] Two μg of pSE1-UK1SEd1-3, described in JP-A-2-257851, were added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT. After addition of 10 units of HindIII, digestion was carried out at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation. The precipitate was dissolved in 20 μl of DNA polymerase I buffer, 5 units of Escheichia coli-derived DNA polymerase I Klenow fragment were added, and the 5′ cohesive ends produced by HindIII digestion were rendered blunt by incubation at 22° C. for 30 minutes. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT and 10 units of KpnI were added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT and 10 units of XhoI were added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and two DNA fragments, about 6.67 kb and about 1.98 kb in size, were recovered (about 0.2 μg each).

[0196] Then, 2 μg of pAGE147 obtained in (1) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of KpnI was added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT and 10 units of XhoI were added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.2 μg of a DNA fragment of about 0.66 kb was recovered.

[0197] Then, 0.1 μg of the XhoI-HindIII fragment (about 6.67 kb) of pSE1UK1SEd1-3, as obtained above, 0.1 μg of the KpnI-HindIII fragment (about 1.98 kb), obtained above, and 0.1 μg of the XhoI-KpnI fragment (about 0.66 kb) of pAGE147, as obtained above, were dissolved in a total of 20 μl of T4 ligase buffer. Three hundred fifty units of T4 ligase were added to the solution, and ligation was carried out a; 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101, and the plasmid pAGE148 shown in FIG. 5 was obtained.

[0198] (3) Construction of KM-796- and KM-750-derived Chimeric Human Antibody H Chain Expression Vectors

[0199] First, the cDNA coding for the antibody variable region in the plasmid pKM796H1 or pE750H1 was excised by cleavage at the 5′-terminal EcoRI site and the MaeIII site near the 3′ end of said cDNA and joined, together with a synthetic DNA having the base sequence shown in SEQ ID NO:12, to the chimeric human antibody H chain expression vector pChi641HAM1, as follows (FIG. 6).

[0200] Three μg of pKM796H1 or pKM750H1, obtained in Paragraph 4, were added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. Further, 10 units of EcoRI and 10 units of MaeIII were added, and digestion was effected a, 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 0.43 kb was recovered. Then, 3 μg of pChi641HM1, obtained in Reference Example 2, was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of ApaI were also added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1.0 μg of a DNA fragment of about 9.0 kb was recovered. Then, 0.1 μg of the EcoRI-MaeIII fragment (about 0.43 kb) of pKM796H1 or pKM750H1, as obtained above, 0.1 μg of the EcoRI-ApaI fragment (about 9.0 kb) of pChi641HAM1, as obtained above, and 0.3 μg of a synthetic DNA having the base sequence shown in SEQ ID NO:12 were dissolved in a total of 20 μl of T4 ligase buffer, 350 units of T4 ligase was further added to the solution, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101. In this way, the plasmids pChi796HM1 and pChi750HM1, shown in FIG. 6, were obtained.

[0201] Then, the β-globin 3′ splicing signal was introduced into the plasmids pChi796HM1 and pChi750HM1 by the method described below to construct KM796 and KM-750-derived chimeric human antibody H chain expression vectors (FIG. 7).

[0202] Three μg of pChi796HM1 or pChi750HM1 were added to 30 μl of 33 mM Tris-acetate buffer (pH 7.9) containing 10 mM magnesium acetate, 66 mM potassium acetate, 0.5 mM DTT and 0.01% bovine serum albumin (hereinafter, “BSA”). Ten units of XhoI and 10 units of KpnI were also added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 3.4 kb was recovered. Then, 3 μg of pAGE148 obtained in (2) was added to 30 μl of 33 mM Tris-acetate buffer (pH 7.9) containing 10 mM magnesium acetate, 66 mM potassium acetate, 0.5 mM DTT and 0.01% BSA; 10 units of XhoI and 10 units of KpnI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 8.7 kb was recovered. Then, 0.1 μg of the XhoI-KpnI fragment of pChi796HM1 or pKM750HM1 and 0.1 μg of the XhoI-KpnI fragment of pAGE148 were dissolved in a total of 20 μl of T4 ligase buffer, 350 units of T4 ligase was further added to the solution, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101. The plasmids pChi796HMS1 and pChi750HMS1 shown in FIG. 7 were thus obtained.

[0203] (4) Construction of KM-796- and KM-750-derived Chimeric Human Antibody L Chain Expression Vectors

[0204] First, the cDNA coding for the antibody variable region in the plasmid pKM796L1 or pKM750L1 was excised by cleavage at the 5′-termainal EcoRI site and the Af1III site near the 3′ end of said cDNA and joined, together with a synthetic DNA having the base sequence shown in SEQ ID NO:13, to the chimeric human antibody L chain expression vector pChiIgLA1, as follows (FIG. 8).

[0205] Three μg of pKM795L1 or pKM750L1 were added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. Further, 10 units of EcoRI and 10 units of Af1III were added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 0.39 kb was recovered. Then, 3 μg of pChiIgLA1 obtained in Reference Example 1 was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 nM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of EcoRV were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1 μg of a DNA fragment of about 8.6 kb was recovered.

[0206] Then, 0.1 μg of the EcoRI-Af1III fragment of pKM796L1 or pKM750L1, as obtained above, 0.1 μg of the EcoRI-EcoRV fragment of pChiIgLA1, as obtained above, and 0.3 μg of a synthetic DNA, having the base sequence shown in SEQ ID NO:13, were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was further added to the solution, and ligation was carried our at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101. In this way, the plasmids pChi796LI1 and pChi750LI1 shown in FIG. 8 were obtained.

[0207] Then, the Moloney mouse leukemia virus terminal repeat promoter/enhancer was introduced into the plasmids pChi796LI1 and pChi750LI1 in the following manner (FIG. 9).

[0208] Three μg of pChi796LI1 and pChi750LI1 were added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. Further, 10 units of EcoRI and 10 units of XhoI were added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 8.2 kb was recovered Then, 0.3 μg of the chimric human antibody H chain expression vector pChi641HAM1 obtained in Reference Example 2 was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of XhoI were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 0.6 kb was recovered.

[0209] Then, 0.1 μg of the EcoRI-XhoI fragment of pChi796LI1 or pKM750LI1 as obtained above and 0.1 μg of the EcoRI-XhoI fragment of pChi641HAM1 as obtained above were dissolved in a total of 20 μl of T4 ligase buffer, 350 units of T4 ligase was further added to the solution, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101. In this way, the plasmids pChi796LM1 and pChi750LM1 shown in FIG. 9 were obtained.

[0210] Then, the β-globin 3′ splicing signal was introduced into the plasmids pChi796LM1 and pChi750LM1 in the manner mentioned below to construct KM-795- and KM-750-derived chimeric human antibody L chain expression vectors (FIG. 10).

[0211] Three μg of pChi796LM1 or pChi750LM1 were added to 30 μl of 33 mM Tris-acetate buffer (pH 7.9) containing 10 mM magnesium acetate, 66 mM potassium acetate, 0.5 mM DTT and 0.01% BSA. Further, 10 units of XhoI and 10 units of KpnI were added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 2.0 kb was recovered. Then, 3 μg of pAGE148 obtained in (2) was added to 30 μl of 33 mM Tris-acetate buffer (pH 7.9) containing 10 mM magnesium acetate, 66 mM potassium acetate, 0.5 mM DTT and 0.01 BSA; 10 units of XhoI and 10 units of KpnI were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 8.7 kb was recovered. Then 0.1 μg of the XhoI-KpnI fragment of pChi796LM1 or pKM750LM1 as obtained above and 0.1 μg of the XhoI-KpnI fragment of pAGE148 were dissolved in a total of 20 μl of T4 ligase buffer, 350 units of T4 ligase was further added, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101. In this way, the plasmids pChi796LMS1 and pChi750LMS1 shown in FIG. 10 were obtained.

[0212] 8. Construction of KM-796- and KM-750-derived Chimeric Human Antibody H Chain and L Chain Tandem Expression Vectors

[0213] Tandem expression vectors containing the chimeric human antibody H chain-encoding cDNA and L chain-encoding cDNA on one and the same vector were constructed (FIG. 11 and FIG. 12).

[0214] Three μg of pChi796HMS1 or pChi750HMS1, obtained in Paragraph 7, were added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT. Further, 10 units of MluI and 10 units of SalI were added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis. In each case, about 0.3 μg of a DNA fragment of about 5.9 kb was recovered. Then, 2 μg of pAGE107 described in EP-A-0 405 285 was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT; 10 units of MluI and 10 units of SAlI were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.2 μg of a DNA fragment of about 3.55 kb was recovered. Then, 0.1 μg of the MluI-SalI fragment of pChi796HMS1 or pChi750HMS1 and 0.1 μg of the MluI-SalI fragment of pAGE107 were dissolved in a total of 20 μl of T4 ligase buffer, 350 units of T4 ligase was added, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 to give the plasmid pChi796H107 or pChi750H107 shown in FIG. 11.

[0215] Then, 3 μg of pChi795H107 or pChi750H107 was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT, 10 units of ClaI was further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation. The precipitate was dissolved in 20 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the mixture was incubated at 22° C. for 30 minutes for rendering the cohesive ends formed upon ClaI digestion blunt-ended. The reaction mixture was further subjected to phenol-chloroform extraction and then to ethanol precipitation. To the precipitate were added 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT, and 10 units of MluI. Digestion was carried out at 37° C. for 4 hours and the reaction mixture was fractionated by agarose gel electrophoresis. In each case, about 0.3 μg of a DNA fragment of about 7.5 kb was recovered. Then, 3 μg of pChi796LMS1 or pChi750LMS1 was added to 30 μl of 20 mM Tris-hydrochloride buffer (pH 8.5) containing 10 mM magnesium chloride, 100 mM potassium chloride and 1 mM DTT, 10 units of XhoI was further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation. The precipitate was dissolved in 20 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA. polymerase I Klenow fragment was added, and the mixture was incubated at 22° C. for 30 minutes for rendering the cohesive ends formed upon XhoI digestion blunt-ended. The reaction mixture was further subjected to phenol-chloroform extraction and then to ethanol precipitation. To the precipitate was added 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT as well as 10 units of MluI. Digestion was carried out at 37° C. for 4 hours and the reaction mixture was fractionated by agarose gel electrophoresis. In each case, about 0.3 μg of a DNA fragment of about 9.3 kb was recovered. Then, 0.1 μg of the MluI-ClaI fragment of pChi796H107 or pChi750H107, as obtained above, and 0.1 μg of the MluI-XhoI fragment of pChi796LMS1 or pChi750LMS1, as obtained above, were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was further added, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101, and the plasmid pChi796HL1 or pChi750HL1 shown in FIG. 12 was obtained.

[0216] 9. Construction of a KM-603-derived Chimeric Human Antibody H Chain Expression Vector

[0217] First, the antibody variable region-encoding cDNA of the plasmid pKM603H1 was excised by cleavage at the 5′-terminal EcoRI site and the StyI site near the 3′ end of said cDNA and joined to the chimeric human antibody H chain expression vector pChi641HAM1 together with a synthetic DNA having the base sequence shown in SEQ ID NO:14 in the following manner (FIG. 13).

[0218] Three μg of pKM603H1 obtained in Paragraph 5 were added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, followed by further addition of 10 units of EcoRI and 10 units of StyI. Digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a 0.4-kb DNA fragment was recovered. Then, 3 μg of pChi641HAM1, obtained in Reference Example 2, was added to 30 μl of 10 mM Tris-hydrochloride (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of ApaI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1.0 μg of a DNA fragment of about 9.0 kb was recovered. Then, 0.1 μg of the EcoRI-StyI fragment (about 0.4 kb) of pKM603H1, as obtained above, and 0.1 μg of the EcoRI-ApaI fragment (about 9.0 kb) of pChi641HAM1, as obtained above, were dissolved, together with 0.3 μg of a synthetic DNA having the base sequence shown in SEQ ID NO:14, in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was added to the solution, and ligation was effected at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pChi603HM1 shown in FIG. 13 was obtained.

[0219] Then, a KM-603-derived chimeric human antibody H chain expression vector was constructed by introducing the β-globin 31 splicing signal into the plasmid pChi603HM1 in the following manner (FIG. 14).

[0220] Three μg of pChi603HM1 obtained above were added to 30 μl of 33 mM Tris-acetate buffer (pH 7.9) containing 10 mM magnesium acetate, 66 mM potassium acetate, 0.5 mM DTT and 0.01% BSA. Further, 10 units of XhoI and 10 units of KpnI were added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 3.3 kb was recovered.

[0221] Then, 3 μg of pAGE148 obtained in Paragraph 7 (2) was added to 30 μl of 33 mM Tris-acetate buffer (pH 7.9) containing 10 mM magnesium acetate, 66 mM sodium acetate, 0.5 mM DTT and 0.01 % BSA; 10 units of XhoI and 10 units of KpnI were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 8.7 kb was recovered. Then, 0.1 μg of the XhoI-KpnI fragment of pChi603HM1, as obtained above, and 0.1 μg of the XhoI-KpnI fragment of pAGE148, as obtained above, were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was added to the solution, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pChi603HMS1 shown in FIG. 14 was obtained.

[0222] 10. Construction of a KM-603-derived Chimeric Human Antibody L Chain Expression Vector

[0223] First, the antibody variable region cDNA in the plasmid pKM603L1 was excised by cleavage at the 5′ terminal EcoRI site and the Af1III site near the 3′ end and joined to the chimeric human antibody L chain expression vector pChiIgLA1 together with a synthetic DNA having the base sequence defined by SEQ ID NO:15 (FIG. 15).

[0224] Thus, 3 μg of pKM603L1 obtained in Paragraph 5 was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of AflIII were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 0.4 kb was recovered. Then, 3 μg of pChiIgLA1 obtained in Reference Example 1 was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of EcoRV were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1 μg of a DNA fragment of about 8.6 kb was recovered. Then, 0.1 μg of the EcoRI-Af1III fragment of pKM603L1, as obtained above, 0.1 μg of the EcoRI-EcoRV fragment of pChiIgLA1, as obtained above, and 0.3 μg of a synthetic DNA, having the base sequence defined by SEQ ID NO:15, were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was added to the solution, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pChi603LI1 shown in FIG. 15 was obtained.

[0225] Then, the Moloney mouse leukemia virus terminal repeat promoter/enhancer was introduced into the plasmid pChi603LI1 in the following manner (FIG. 16).

[0226] Thus, 3 μg of pChi603LI1 obtained above was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 10 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of XhoI were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 8.3 kb was recovered. Then, 3 μg of pChi641HAM1 obtained in Reference Example 2 was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of XhoI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 0.6 kb was recovered. Then, 0.1 μg of the EcoRI-XhoI fragment of pChi603LI1, as obtained above, and 0.1 μg of the EcoRI-XhoI fragment of pChi641HAM1, as obtained above, were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was added to the solution, and ligation was effected at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Eschrichia coli HB101 to give the plasmid pChi603LM1 shown in FIG. 16.

[0227] A KM-603-derived chimeric human antibody L chain expression vector was then constructed by introducing the β-globin 3′ splicing signal into the plasmid pChi603LM1, as follows (FIG. 17).

[0228] Thus, 3 μg of pChi603LM1 obtained above was added to 30 μl of 33 mM Tris-acetate buffer (pH 7.9) containing 10 mM magnesium acetate, 66 mM potassium acetate, 0.5 mM DTT and 0.01% BSA, 10 units of XhoI and 10 units of KpnI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 2.0 kb was recovered. Then, 3 μg of pAGE148 obtained in Paragraph 7 (2) was added to 30 μl of 33 mM Tris-acetate buffer (pH 7.9) containing 10 mM magnesium acetate, 66 mM potassium acetate, 0.5 mM DTT and 0.01% BSA, 10 units of XhoI and 10 units of KpnI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a DNA fragment of about 8.7 kb was recovered. Then, 0.1 μg of the XhoI-KpnI fragment of pChi603LM1, as obtained above, and 0.1 μg of the XhoI-KpnI fragment of pAGE148, as obtained above, were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was added to the solution, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 to give the plasmid pChi603LMS1 shown in FIG. 17.

[0229] 11. Expression of the KM-796- and KM-750-derived Chimeric Human Anti-GM2 Antibody in YB2/0 Cells

[0230] The plasmids were introduced into Y32/0 cells by the electroporation method of Miyaji et al. [Cytotechnology, 3, 133-140 (1990)].

[0231] After introduction of 4 μg of pChi750HL1 or pChi796HL1 obtained in Paragraph 8 into 4×106 YB-2/0 (ATCC CRL1581) cells, the cells were suspended in 40 ml of RPMI-1640-FCS(10) [RPMI1640 medium (Nissui Pharmaceutical) containing 10% of FCS, {fraction (1/40)} volume of 7.5% NaHCO3, 3% of 200 mM L-glutamine solution (Gibco) and 0.5% of penicillin-streptomycin solution (Gibco; containing 5,000 units/ml penicillin and 5,000 μg/ml streptomycin)], and the suspension was distributed in 200-μl portions into wells of 96-well microtiter plates. After 24 hours of incubation at 37° C. in a CO2 incubator, G418 (Gibco) was added to a concentration of 0.5 mg/ml and then incubation was continued for 1 to 2 weeks. Transformant colonies appeared, the culture fluid was recovered from each well in which the cells had grown to confluence and an enzyme-linked immunosorbent assay (ELISA) was conducted for anti-GM2 chimeric human antibody activity measurement.

Enzyme-linked Immunosorbent Assay (ELISA)

[0232] In a solution of 5 ng of phosphatidylcholine (Sigma) and 2.5 ng of cholesterol (Sigma) in 2 ml of ethanol was dissolved 2 ng of GM2 (N-acetyl-GM2; Boehringer Mannheim) or some other ganglioside. The solution or dilutions thereof were respectively distributed in 20-μl portions into wells of 96-well microtiter plates (Greiner) and, after air drying, blocking was effected with PBS containing 1% BSA. Each culture supernatant for each transformant, each purified mouse monoclonal antibody solution and each purified chimeric human antibody solution were distributed in 50- to 100-μl portions into the wells and the reaction was allowed to proceed at room temperature for 1 to 2 hours. The wells were then washed with PBS, and 50 to 100 μl of peroxidase-labeled antibody were added thereto followed by reaction at room temperature for 1 to 2 hours. The wells were washed with PBS and an ABTS substrate solution [prepared by dissolving 550 mg of 2,2′-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt in 0.1M citrate buffer (pH 4.2) and adding, just prior to use, hydrogen peroxide to a concentration of 1 μl/ml] was added in 50 - to 100-μl portions to each well for color development, followed by OD415 measurement.

[0233] The clone showing the highest activity in ELISA among the clones obtained gave a chimeric human anti-GM2 antibody content of about 1.0 μg/ml of culture fluid.

[0234] Cells of the clone showing the above-mentioned chimeric human anti-GM2 antibody activity were suspended in RPMI1640-FCS(10) medium containing 0.5 mg/ml G418 and 50 nM methotrexate (hereinafter, “MTX”) to a concentration of 1 to 2×105 cells/ml, and the suspension was distributed in 2-ml portions into wells of 24-well plates. Incubation was performed at 37° C. in a CO2 incubator for 1 to 2 weeks to induce 50 nM MTX-resistant clones. At the time of confluence, the chimeric human anti-GM2 antibody activity in each culture fluid was determined by ELISA. The 50 nM MTX-resistant clone showing the highest activity among the clones obtained showed a chimeric human anti-GM2 antibody content of about 5.0 μg/ml.

[0235] Cells of the above 50 nM MTX-resistant clone were suspended in RPMI1640-FCS(10) medium containing 0.5 mg/ml G418 and 200 nM MTX to a concentration of 1 to 2×105 cells/ml, and the suspension was distributed in 2-ml portions into wells of 24-well plates. Incubation was carried out at 37° C. in a CO2 incubator for 1 to 2 weeks to induce 200 nM MTX-resistant clones. At the time of confluence, each culture fluid was assayed for chimeric human anti-GM2 antibody activity by ELISA. The 200 nM MTX-resistant clone showing the highest activity among the clones obtained had a chimeric human anti-GM2 antibody content of about 10 μg/ml. The 200 nM MTX-resistant clones obtained from pChi750HL1 and pChi796HL1 were named transformants “KM966” (KM-796-derived chimeric human antibody KM966-producing strain) and “KM967” (KM-750-derived chimeric human antibody KM967-producing strain), respectively.

[0236] The following SDS-polyacrylamide gel electrophoresis (SDS-PAGE) confirmed that the above transformants KM966 and KM967 express the respective chimeric human anti-GM2 antibodies.

[0237] The transformants KM966 and KM967 were each suspended in GIT medium (Nippon Pharmaceutical) containing 0.5 mg/ml G418 and 200 nM MTX to a concentration of 1 to 2×105 cells/ml. Each suspension was distributed in 100-ml portions into 175 cm2 flasks (Greiner). Cultivation was performed at 37° C. in a CO2 incubator for 5 to 7 days. At the time of confluence, the culture fluid was recovered. Treatment of about 1 liter of the culture fluid with Affi-Gel Protein A MAPS-II kit (Bio-Rad) gave about 5 mg of a purified chimeric human anti-GM2 antibody for each transformant. About 2 μg of the purified chimeric human anti-GM2 antibody KM966 or KM967 was electrophoresed by the conventional method [(Laemmli: Nature, 227, 680 (1970)] for molecular weight checking. The results are shown in FIG. 18. As shown in FIG. 18, both KM966 and KM967 gave an antibody H chain molecular weight of about 50 kilodaltons and an antibody L chain molecular weight of about 25 kilodaltons under reducing conditions, indicating the correctness in molecular weight of the H chain and L chain expressed. For each of KM966 and KM967, the molecular weight of the chimeric human antibody under nonreducing conditions was about 150 kilodaltons, indicating that the antibody expressed was composed of two H chains and two L chains and was correct in size.

[0238] 12. Expression of KM-603-derived Chimeric Human Anti-GM2 Antibodies in SP2/0 Cells

[0239] A 2-μg portion of the plasmid pChi603HMS1 or pChi603LMS1 obtained in Paragraph 9 was introduced into 4×106 cells of YB2/0 (ATCC CRL1581) by electroporation in the same manner in Paragraph 11. The cells were suspended in 40 ml of RPMI1640-FCS(10) and the suspension was distributed in 200-μl portions into wells of 96-well microtiter plates. After 24 hours of incubation in a CO2 incubator at 37° C., G418 (Gibco) was added to a concentration of 0.5 mg/ml and incubation was continued for 1 to 2 weeks. Transformant colonies appeared. The culture fluid was recovered from confluent wells and the chimeric human anti-GM2 antibody activity was measured by ELISA as described above. The clone showing the highest chimeric human anti-GM2 antibody activity among the clones obtained gave a chimeric human anti-GM2 antibody content of about 0.1 μg/ml of culture fluid.

[0240] Cells of the clone showing the above-mentioned chimeric human anti-GM2 antibody activity were suspended in RPMI1640-FCS(10) medium containing 0.5 mg/ml G418 and 50 nM MTX to a concentration of 1 to 2×105 cells/ml and the suspension was distributed in 2-ml portions into wells of 24-well plates. Clones resistant to 50 nM MTX were induced by incubating in a CO2 incubator at 37° C. for 2 to 3 weeks. When confluence was attained, the culture fluids were subjected to ELISA for chimeric human anti-GM2 antibody activity measurement. Among the 50 nM MTX-resistant clones obtained, the clone showing the highest activity gave a chimeric human anti-GM2 antibody content of about 0.3 μg/ml of culture fluid.

[0241] Cells of the above 50 nM MTX-resistant clone were suspended in RPMI1640-FCS(10) medium containing 0.5 mg/ml G418 and 200 nM MTX to a concentration of 1 to 2×105 cells/ml and the suspensions as distributed in 2-ml portions into well of 24-well plates. Clones resistant to 200 nM MIX were induced by following incubation in a CO2 incubator at 37° C. for 2 to 3 weeks. When confluence was attained, the chimeric human anti-GM2 antibody activity in the culture fluid was measured by ELISA. Among the 200 nM MTX-resistant clones obtained, the clone showing the highest activity gave a chimeric human anti-GM2 antibody content of about 0.5 μg/ml of culture fluid.

[0242] Cells of the above 200 nM MTX-resistant clone were suspended in RPMI 1640-FCS(10) medium containing 0.5 mg/ml G418 and 500 nM MTX to a concentration of 1 to 2×105 cells/ml and the suspension was distributed in 2-ml portions into well of 24-well plates. Clones resistant to 500 nM MTX were induced following incubation in a CO2 incubator at 37° C. for 1 to 2 weeks. When confluence was attained, the chimeric human anti-GM2 antibody activity in the culture fluid was determined by ELISA. Among the 500 nM MTX-resistant clones obtained, the one showing the highest activity gave a chimeric human anti-GM2 antibody content of about 1.0 μg/ml of culture fluid. This 500 M MTX-resistant clone was named “transformant KM968”.

[0243] The following SDS-PAGE confirmed the expression of a chimeric human anti-GM2 antibody in the above transformant KM968.

[0244] Cells of the transformant KM968 were suspended in GIT medium (Nippon Pharmaceutical) containing 0.5 mg/ml G418 and 500 nM MTX to a concentration of 1 to 2×105 cells/ml and the suspension was distributed in 100-ml portions into 175 cm2 flasks (Greiner). Cultivation was conducted in a CO2 incubator at 37° C. for 5 to 7 days and, when confluence was attained, the culture fluid was recovered. Treatment of about 3.0 liters of the culture fluid with Affi-Gel Protein A MAPS-II kit (Bio-Rad) gave about 1 mg of a purified chimeric human anti-GM2 antibody. About 2 μg of this purified chimeric human anti-GM2 antibody KM968 was electrophoresed by the conventional method [(Laemmli: Nature, 227, 680 (1970)] for molecular weight checking. The results are shown in FIG. 19. Under reducing conditions, the molecular weight of the antibody H chain was about 50 kilodaltons and the molecular weight of the antibody L chain was about 25 kilodaltons, thus confirming the expression of the H chain and L chain having the correct molecular weight. Under nonreducing conditions, the molecular weight of the chimeric human antibody was about 150 kilodaltons, confirming that the antibody expressed was composed of two H chains and two L chains and was correct in size.

[0245] 13. Reaction Specificity of the Chimeric Human anti-GM2 antibodies

[0246] The reactivities of the chimeric anti-GM2 antibodies with ganglioside GM1, N-acetyl-GM2 (Boehringer Mannheim), N-glycolyl-GM2, N-acetyl-GM3, N-glycolyl-GM3, GD1a, GD1b (Iatron), GD2, GD3 (Iatron) and GQ1b (Tatron) were examined by the technique of ELISA. The results are shown below in Table 1. GM1 and GD1a were purified from the bovine brain, N-glycolyl-GM2 and N-glycolyl-GM3 from the mouse liver, N-acetyl-GM3 from canine erythrocytes, and GD2 from the cultured cell line IMR32 (ATCC CCL127), by a known method [J. Biol. Chem., 263, 10915 (1988)].

[0247] As shown in Table 1, it was confirmed that the chimeric human anti-GM2 antibodies KM966 and KM967 specifically react with GM2. The reactivity of KM966 was greater than that of KM967, however. On the contrary, KM968 (KM-603-derived chimeric human antibody) did not show any reactivity for GM2.

	TABLE 1
		Binding activity of antibody (OD415)
	Ganglioside	KM966 (5 μg/ml)	KM967 (5 μg/ml)
	GM1	0.105	0.000
	N-Acetyl-GM2	0.870	0.423
	N-Glycolyl-GM2	0.774	0.065
	N-Acetyl-GM3	0.002	0.000
	N-Glycolyl-GM3	0.122	0.001
	GD1a	0.004	0.000
	GD1b	0.002	0.000
	GD2	0.095	0.001
	GD3	0.004	0.000
	GQ1b	0.005	0.000

[0248] 14. Reactivities of the Chimeric Human Anti-GM2 Antibodies KM966 and KM967 with Cancer Cells (Fluorescent Antibody Technique)

[0249] Suspended in PBS were 1×106 cells of cultured human lung small cell carcinoma cell line QC90 [Cancer Res., 49, 2683 (1989)], NCI-H69 (ATCC HTB119), NCI-H128 (ATCC HTB120), SBC-1 (JCRB 0816), SBC-2 (JCRB 0817), SBC-3 (JCRB 0818), SBC-5 (JCRB 0819), RERF-LC-MA (JCRB 0812), Lu-134-A-H (JCR3 0235), Lu-139 (RCB 469), Lu-130 (RCB 465), Lu-135 (RCB 468), Lu-134-B (RCB 467), Lu-140 (RCB 470), PC-6 [Naito et al.: Gann to Kagaku Ryoho (Cancer and Chemotherapy), 5 (suppl.), 89 (1978)], cultured human lung squamous carcinoma cell line PC-1 [Naito et al.: Gann to Kagaku Ryoho, 5 (suppl.), 89 (1978)], PC-10 [Naito et al.: Gann to Kagaku Ryoho, 5 (suppl.), 89 (1978)], Colo16 [Moor et al.: Cancer Res., 35, 2684 (1975)], Calu-1 (ATCC HTB54), SK-LC-4 [Proc. Natl. Acad. Sci. U.S.A., 85, 4441 (1988)], cultured human lung adenocarcinoma cell line PC-7 [Hayata et al.: Hito Gansaibo no Baiyo (Human Cancer Cell Culture), 131 (1975)], PC-9 [Kinjo et al.: Brit. J. Cancer, 39, 15 (1979)], PC-12 (ATCC CRL1721), RERF-LC-MS (JCRB 0081), HLC-1 (RCB 083), cultured human lung large cell carcinoma cell line PC-13 [Oboshi et al.: Tanpakushitsu, Kakusan, Koso (Protein, Nucleic Acid, Enzyme), 23, 697 (1978)], Lu65 (JCRB 0079), CALU-6 (ATCC HTB56), SK-LC-6 [Proc. Natl. Acad. Sci. U.S.A., 85, 4441 (1988)], cultured human neuroblastoma cell line YT-nu [Ishikawa et al.: Acta Path. Jap., 27, 697 (1977)], NAGAI [Ishikawa et al : Acta Path. Jap., 29, 289 (1979)], NB-1 [Ishikawa et al.: Acta Path. Jap., 27, 697 (1977)], IMR32 (ATCC CCL127), GOTO (JCRB 0612), NB-9 (RCB 477), SK-N-MC (ATCC HTB10), cultured human brain tumor (glioma) cell line SK-MG-4 [EMBO J., 6, 2939 (1987)], A172 (ATCC CRL1620), T98G (ATCC CRL1690), U-118MG (ATCC HTBS), cultured human leukemia cell line HSB-2 (ATCC CCL120.1), ATN-1, U-937 (ATCC CRL1593), HPB-ALL [Oboshi et al.: Tanpakushitsu, Kakusan, Koso, 23, 697 (1978)], CCRF-SB (ATCC CCL120), KOPN-K [Hanei et al.: Haigan (Lung Cancer), 25, 524 (1985)], TYH [Haranaka et al.: Int. J. Cancer, 36, 313 (1985)], MOLT-3 (ATCC CRL1552), CCRF-CEM (ATCC CCL119), TALL-1 (JCRB 0086), NALL-1 [Oboshi et al.: Tanpakushitsu, Kakusan, Koso, 23, 697 (1978), CCRF-SB (JCRB 0032), THP-1 (ATCC TIB202), HEL92-1-7 (ATCC TIB180), cultured human maligant melanoma cell line C24.32 (EP-A-0 493686), KHm-3/P [J. Natl. Cancer Inst., 59, 775 (1977)] or G361 (ATCC CRL1424). The suspension was placed in a microtube (Tref) and centrifuged (3,000 rpm, 2 minutes) to wash the cells, 50 μl of KM966 or KM967 (50 μg/ml) was added, the mixture was stirred, and the reaction was allowed to proceed at 4° C. for 1 hour. Then, the cells were washed three times by centrifugation with PBS, 20 μl of fluorescein isocyanate-labeled protein A (30-fold dilution; Boehringer Mannheim Yamanouchi) was added and, after stirring, the reaction was allowed to proceed at 4° C. for 1 hour. Then, the cells were washed three times by centrifugation with PBS, then suspended in PBS and subjected to analysis using flow cytometer EPICS Elite (Coulter). In a control run, the same procedure as described above was followed without adding the chimeric human antibody. The results thus obtained are shown in Table 2. The chimeric human antibody K,M966 reacted with 9 (NCI-H128, SBC-1, SBC-3, SBC-5, Lu-139, Lu-130, Lu-135, Lu-134-B and Lu-140) of the 14 lung small cell carcinoma cell lines, 2 (PC-10 and Calu-1) of the 5 lung squamous carcinoma cell lines, 2 (PC-9 and RERF-LC-MS) of the 5 lung adenocarcinoma cell lines, 2 (PC-13 and SK-LC-6) of the 4 lung large cell carcinoma cell lines, 7 (YT-nu, NAGAI, NB-1, IMR32, GOTO, NB-9 and SK-N-MC) of the 7 neuroblastoma cell lines and 4 (SK-MG-4, A172, T98G and U-118MG) of the 4 brain tumor (glioma) cell lines. On the other hand, the chimeric human antibody KM967 did not react with any of the cultured cell lines. The above results indicate that the chimeric human antibody KM966 is useful in the diagnosis and treatment of brain tumors, peripheral nervous system tumors and lung cancer, among others.

	TABLE 2
		KM966 (%)	KM967 (%)
	Cell line	(50 μg/ml)	(50 μg/ml)
	Lung small cell carcinoma	9/14 (64)	0/14 (0)
	Lung squamous cell	2/5 (40)	0/5 (0)
	carcinoma
	Lung adenocarcinoma	2/5 (40)	0/5 (0)
	Lung large cell carcinoma	2/4 (50)	0/4 (0)
	Neuroblastoma	7/7 (100)	0/7 (0)
	Brain tumor (glioma)	4/4 (100)	0/4 (0)
	Leukemia	0/14 (0)	0/14 (0)
	Malignant melanoma	0/3 (0)	0/3 (0)

[0250] 15. In Vitro Antitumor Activity of the Chimeric Human Anti-GM2 Antibody KM966: Complement Dependent Cytotoxicity (CDC)

[0251] (1) Preparation of Target Cells

[0252] The target cells SBC-3, Lu-135, PC-10, RERF-LC-MS, PC-13, NAGAI, GOTO or A172, cultured in RPMI1640 medium supplemented with 100 FCS, were adjusted to a cell concentration of 5×106 cells/ml, Na251CrO4 was added to a concentration of 100 μCi/5×106 cells, then the reaction was allowed to proceed at 37° C. for 1 hours, and the cells were washed three times with the medium. The cells were then allowed to stand in the medium at 4° C. for 30 minutes for spontaneous dissociation and then, after centrifugation, the medium was added to adjust the cell concentration to 1×106 cells/ml.

[0253] (2) Preparation of the Complement

[0254] Sera from three healthy subjects were combined and used as a complement source.

[0255] (3) CDC Activity Measurement

[0256] The chimeric human anti-GM2 antibody KM966 or mouse anti-GM2 antibody KM696 (FERM BP-3337) was added to wells of 96-well U-bottom plates within the final concentration range of 0.5 to 50 μg/ml and then 5×104 cells/well of the target cells prepared in (1) were added. The reaction was allowed to proceed at room temperature for 30 minutes. After centrifugation, the supernatants were discarded, 150 μl of the human serum obtained in (2) was added to each well (final concentration 15% v/v), and the reaction was allowed to proceed at 37° C. for 1 hour. After centrifugation, the amount of 51Cr in each supernatant was determined using a gamma counter. The amount of spontaneously dissociated 51Cr was determined by adding to the target cells the medium alone in lieu of the antibody and complement solutions and measuring the amount of 51Cr in the supernatant in the same manner as mentioned above. The total amount of dissociated 51Cr was determined by adding to the target cells 5 N sodium hydroxide in lieu of the antibody and complement solutions and measuring the amount of 51Cr in the supernatant in the same manner as mentioned above. The CDC activity was calculated as follows:

[0257] 1 $$ CDC ⁢   ⁢ activity ⁢   ⁢ ( % ) = Amount ⁢   ⁢ of ⁢   51 ⁢ Cr ⁢   ⁢ in sample ⁢   ⁢ supernatant · Amount ⁢   ⁢ of ⁢   51 ⁢ Cr spontaneously ⁢   ⁢ dissociated ⁢   Total ⁢   ⁢ amount ⁢   ⁢ of   51 ⁢ Cr ⁢   ⁢ dissociated · Amount ⁢   ⁢ of ⁢   51 ⁢ Cr spontaneously ⁢   ⁢ dissociated × 100

[0258] The results thus obtained are shown in FIGS. 20 to 23. It was shown that the chimeric human antibody KM966 show CDC activity against all the cells tested.

[0259] 16. In Vitro Antitumor Activity of the Chimeric Human Anti-GM2 Antibody KM966: Antibody Dependent Cell Mediated Cytotoxicity (ADCC)

[0260] (1) Preparation of Target Cells

[0261] The target cells SBC-3, Lu-135, PC-10, RERF-LC-MS, PC-13, NAGAI, GOTO or A172, cultured in RPMI1640 medium supplemented with 10% FCS, were adjusted to a cell concentration of 1×106 cells/ml, Na251CrO4 was added to a concentration of 50 μCi/1×106 cells, then the reaction was allowed to proceed at 37° C. for 1 hour, and the cells were washed three times with the medium. The cells were then allowed to stand in the medium at 4° C. for 30 minutes for spontaneous dissociation and then, after centrifugation, the medium was added to adjust the cell concentration to 2×105 cells/ml.

[0262] (2) Preparation of Effector Cells

[0263] Human venous blood (25 ml) was collected, 0.5 ml of heparin sodium (Takeda Chemical Industries; 1,000 units/ml) was added, and the mixture was gently stirred. This mixture was centrifuged (1,500 to 1,800 g, 30 minutes) using Polymorphprep (Nycomed Pharma AS), the lymphocyte layer was separated and washed three times by centrifugation with RPMI-1640 medium (15,00 to 1,800 g, 15 minutes), and the cells were suspended in RPMI1640 medium supplemented with 10% FCS (5×106 cells/ml) for use as effector cells.

[0264] (3) ADCC Activity Measurement

[0265] The chimeric human anti-GM2 antibody KM966 or mouse anti-GM2 antibody DM696 were added to wells of 96-well U-bottom plates within the final concentration range of 0.05 to 5 μg/ml and then 50 μl (1×104 cells/well) of the target cell suspension prepared in (1) and 100 μl (5×105 cells/well) of the effector cell suspension prepared in (2) were added to each well (the ratio between the effector cells and target cells being 50:1). The reaction was allowed to proceed at 37° C. for 4 hours and, after centrifugation, the amount of 51Cr in each supernatant was measured using a gamma counter. The amount of spontaneously dissociated 51Cr was determined by adding to the target cells the medium alone in lieu of the antibody and effector cells and measuring the amount of 51Cr in the supernatant in the same manner as mentioned above. The total amount of dissociated 51Cr was determined by adding to the target cells 5 N sodium hydroxide in lieu of the antibody and effector cells and measuring the amount of 51Cr in the supernatant in the same manner as mentioned above. The ADCC activity was calculated as follows: 2$\mathrm{ADCC}\mathrm{activity}\left(\%\right)=\frac{\begin{array}{c}\mathrm{Amount}{\mathrm{of}}^{51}\mathrm{Cr}\mathrm{in}\\ \mathrm{sample}\mathrm{supernatant}\end{array}·\begin{array}{c}\mathrm{Amount}{\mathrm{of}}^{51}\mathrm{Cr}\\ \mathrm{spontaneously}\mathrm{dissociated}\end{array}}{\begin{array}{c}\mathrm{Total}\mathrm{amount}\mathrm{of}\\ {}^{51}\mathrm{Cr}\mathrm{dissociated}\end{array}·\begin{array}{c}\mathrm{Amount}{\mathrm{of}}^{51}\mathrm{Cr}\\ \mathrm{spontaneously}\mathrm{dissociated}\end{array}}\times 100$

[0266] The results thus obtained are shown in FIGS. 24 to 27. The chimeric antibody KM966 showed ADCC activity against all the cells whereas the mouse anti-GM2 antibody KMG96 showed no or low ADCC activity. The above results indicate that the chimeric human antibody KM966 is more effective in the treatment of human cancer than the mouse antibody KM-696.

Reference Example 1

Construction of the Vector pChiIgLA1 for Chimeric Human Antibody L Chain Expression

[0267] 1. Isolation of the KM50 Cell-derived Immunoglobulin H Chain Promoter and Enhancer Genes

[0268] (1) Preparation of Chromosomal DNAs from KM50 Cells, P3U1 Cells and Rat Kidney

[0269] Chromosal DNAs were prepared by the conventional method [Maniatis et al. (ed.): Molecular Cloning, 1989, p. 9.16], as follows.

[0270] KM50 cells (1.2×103 cells), P3U1 cells (ATCC CRL1597) (2×108 cells) and a rat kidney sample (frozen at −80° C. and then smashed to a sufficient extent using a wooden hammer) (1.6 g) were suspended in 2 ml of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 150 mM sodium chloride and 10 mM ethylenediaminetetraacetic acid disodium salt (hereinafter, “EDTA”), 0.8 mg of proteinase K (Sigma) and 10 mg of sodium lauryl sulfate (hereinafter, “SDS”), were added to each suspension, and the suspension was incubated at 37° C. for 10 hours. Then, each mixture was extracted once with an equal volume of phenol, twice with an equal volume of chloroform and then once with an equal volume of ether, and dialyzed for 10 hours against 10 mM Tris-hydrochloride buffer (pH 7.5) containing 1 mM EDTA. The DNA solution was recovered from the dialysis tube and ribonuclease A (Sigma) was added to the solution to a final concentration of 20 μg/ml. Each resultant solution was incubated at 37° C. for 6 hours for sufficient decomposition of RNA, 15 mg of SDS and 1 mg of proteinase K were then added and the mixture was incubated at 37° C. for 10 hours. The mixture was then extracted twice with an equal volume of phenol, twice with an equal volume of chloroform and twice with an equal volume of ether and then dialyzed for 10 hours against 10 mM Tris-hydrochloride buffer (pH 7.5) containing 1 mM EDTA. The DNA solution was recovered from the dialysis tube for use as a chromosomal DNA sample. DNA concentration measurement in terms of the absorbance at 260 nm revealed that the yield of chromosomal DNA from 1.2×103 KM50 cells was 1.6 mg, that from 2×108 P3U1 cells 1.5 mg, and that from 1.6 g of rat liver 1.9 mg.

[0271] (2) Identification of the Active-form Immunoglobulin H Chain Gene in KM50 Cells by Southern Blotting

[0272] The KM50 cell, p3U1 cell and rat kidney chromosomal DNAs obtained in (1) (3 μg each) were dissolved in 25 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 15 units of XbaI (Takara Shuzo; hereinafter the restriction enzymes used were products of Takara Shuzo) was added and incubation was carried out at 37° C. for 2 hours for effecting cleavage at the XbaI sites. Each reaction mixture was subjected to agarose gel electrophoresis, then DNA transfer onto a nitrocellulose filter was effected by the method of Southern et al. [J. Mol. Biol., 95, 503 (1975)] and hybridization was carried out by the conventional method [Kameyama et al.: FEBS Letters, 244, 301-306 (1989)] using the mouse JH probe described in the last-cited reference. The KIM50 cell DNA alone gave a band at a site corresponding to about 9.3 kb. Therefore, the immunoglobulin XbaI fragment DNA was considered to code for the active-form immunoglobulin H chain gene in KM50 cells.

[0273] (3) Construction of a KM50 Cell Genomic DNA Library

[0274] A 60-μg portion of the KM50 cell-derived chromosomal DNA obtained in (1) was dissolved in 250 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 150 units of XbaI was added, and incubation was conducted at 37° C. for 2 hours for causing cleavage at the XbaI sites. The reaction mixture was fractionated by agarose gel electrophoresis and a KM50 cell-derived 933 kb DNA fraction sample (about 2 μg) was recovered using, for example, the DEAE paper method (Maniatis et al. (ed.): Molecular Cloning, 1989, p. 6.24]. Separately, 3 μg of Lambda ZAP (Stratagene), for use as the vector, was dissolved in 200 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 50 units of XbaI was added, and the mixture was incubated at 37° C. for 2 hours to effect cleavage at the XbaI sites. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, whereby about 3 μg of DNA was recovered. This DNA was dissolved in 100 μl of 100 mM Tris-hydrochloride buffer (pH 7.5), 1 unit of alkaline phosphatase (Takara Shuzo) was added, dephosphorylation was effected at the restriction enzyme cleavage ends of the vector DNA. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, whereby 2 μg of DNA was recovered. This DNA was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 1 mM EDTA for use as a vector sample. Two tenths μg of the vector DNA sample and 0.2 μg of the KM50 cell-derived 9.3 kb DNA sample were dissolved in 5 μl of T4 ligase buffer, 175 units of T4 ligase (Takara Shuzo) was added, and the mixture was incubated at 4° C. for 3 days. A 2-μl portion of this mixture was packaged into the lambda phage by the conventional method [Maniatis et al. (ed.): Molecular Cloning, 1989, p. 2.95] using Giga Pak Gold (Stratagene), and the packaging mixture was used to transfect Escherichia coli BB4 to give 200,000 phage clones. Among them, 100,000 clones were fixed on a nitrocellulose filter by the conventional method [Maniatis et al. (ed.): Molecular Cloning, 1989, p. 2.112].

[0275] (4) Selection of a Recombinant DNA Containing the Gene For the H Chain Variable Region of an Immunoglobulin Occurring as an Active Form in KM50 Cells (Anti-human Serum Albumin)

[0276] From among the phage library composed of 100,000 clones, as constructed in (3), two clones firmly associable at 65° C. with the 32P-labeled mouse JH probe [labeled by the method of Kameyama et al. [FEBS Letters, 44, 301-306 (1989)]] were isolated. The phage DNA was recovered from them by the conventional method [Maniatis et al. (ed.) : Molecular Cloning, 1989, p. 2.118-2.169], whereupon the 9.3 kb XbaI fragment of the KM50 cell-derived chromosomal DNA was found to have been inserted therein.

[0277] (5) Base Sequence of the Gene for the H Chain Variable Region of the Immunoglobulin Occurring as an Active Form in KM50 Cells (Anti-human Serum Albumin)

[0278] For the two clones obtained in (4), restriction enzyme cleavage maps were prepared by conducting digestion using various restriction enzymes, whereby it was revealed that the same DNA fragment (9.3 kb) had been inserted therein (FIG. 28). Therefore, those portions of this 9.3 kb DNA fragment which were supposed to be coding for the rat immunoglobulin H chain promoter region and variable region were sequenced by the method of Sanger [Sanger et al.: Proc. Natl. Acad. Sci. U.S.A., 74, 5463 (1977); AMERSHAM M13 cloning and sequencing handbook]: In SEQ ID NO:16, the portion containing the octamer sequence such as ATGCAAAT and the TATA box sequence such as TTGAAAA is considered to be the immunoglobulin promoter region.

[0279] 2. Construction of Heterologous Protein Expression Vectors Using the Promoter and Enhancer for the H Chain Variable Region Gene for an Immunoglobulin Occurring as an Active Form in KM50 Cells (Anti-human Serum Albumin)

[0280] (1) Construction of pKMB11

[0281] A 1-μg portion of the 9.3 kb immunoglobulin H chain variable region gene fragment obtained in Paragraph 1 (5) was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units each of BglII and HindIII were added, and the mixture was incubated at 37° C. for 2 hours for causing cleavage at the BglII and HindII sites. The reaction mixture was subjected to agarose gel electrophoresis and 0.01 μg of a DNA fragment containing the 0.8 kb immunoglobulin promoter was recovered. Then, 1 μg of the plasmid pBR322-BglII [Kuwana et al.: FEBS Letters, 219, 360 (1987)] was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of BglII and 10 units of HindIII were added, and the mixture was incubated at 37° C. for 2 hours to effect cleavage at the BglII and HindIII sites. The reaction mixture was subjected to agarose gel electrophoresis and a DNA fragment of about 4.2 kb in size was recovered. The thus-obtained pBR322-BglII-derived DNA fragment (about 4.2 kb, 0.1 μg) and immunoglobulin promoter-containing DNA fragment (0.01 μg) were dissolved in 20 μl of T4 ligase buffer, 175 units of T4 DNA ligase (Takara Shuzo) was added, and the mixture was incubated at 4° C. for 1 day. The reaction mixture was used to transform Escherichia coli HB101 [J. Mol. Biol., 41, 459 (1969)] by the method of Scott et al. [Masaru Shigesada: Saibo Kokagu (Cell Engineering), 2, 616 (1983)] to give an Ap-resistant colony. The recombinant plasmid DNA was recovered from this colony. Plasmid pKMB11, shown in FIG. 29, was thus obtained.

[0282] (2) Construction of pKMD6

[0283] For providing an appropriate restriction enzyme site downstream from the immunoglobulin promoter, the plasmid pKMB11 constructed in (1) was digested at the NcoI site using the nuclease BAL31. Thus, 10 μg of the plasmid pKMB11 was dissolved in 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM potassium chloride, 30 units of NcoI was added, and the mixture was incubated at 37° C. for 2 hours to effect cleavage at the NcoI site. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the whole amount of the DNA fragment was dissolved in 100 μl of BAL31 buffer [20 mM Tris-hydrochloride buffer (pH 8.0) containing 600 mM sodium chloride, 12 mM calcium chloride, 12 mM magnesium chloride and 1 mM EDTA], 0.25 unit of BAL31 [Bethesda Research Laboratories (BRL)]] was added, and digestion was carried out at 37° C. for 5 seconds. The reaction was terminated by extraction with phenol and subjected to chloroform extraction and then to ethanol precipitation, and 1 μg of DNA was recovered. A 0.1-μg portion of this DNA and 0.01 μg of a synthetic DNA linker (SAII) were dissolved in 20 μl of T4 ligase buffer, 175 units of T4 DNA ligase was added, and the mixture was incubated at 4° C. for 1 day. The reaction mixture was used to transform Escherichia coli HB101 by the method of Scott et al. An Ap-resistant colony was obtained and the recombinant plasmid DNA was recovered from this colony to give the plasmid pKMD6 shown in FIG. 30. For this plasmid, the portion of BAL31 digestion was sequenced by the method of Sanger, whereupon deletion was found to the third base (303rd base in SEQ ID NO:16) toward the upstream of the initiation codon ATG for immunoglobulin.

[0284] (3) Construction of pEPKMA!, pEPKMB1 and pAGE501

[0285] The original immunoglobulin promoter and enhancer are positionally separated. Therefore, it was necessary to construct a vector containing the promoter and enhancer connected to each other for use of said vector as a heterologous protein expression vector. Accordingly, the following procedure was followed.

[0286] Thus, 1 μg of the 9.3 kb immunoglobulin H chain variable region gene obtained in Paragraph 1 (5) was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of EcoRV and 10 units of XbaI were added, and the mixture was incubated at 37° C. for 2 hours for causing cleavage at the EcoRV and XbaI sites. The reaction mixture was subjected to agarose gel electrophoresis and 0.1 μg of a DNA fragment (about 1 kb) containing the immunoglobulin enhancer region was recovered. Separately, 1 μg of the plasmid pKMD6 obtained in (2) was dissolved in 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of BglII was added, and the mixture was incubated at 37° C. for 2 hours to effect cleavage at the BglII site. After phenol-chloroform extraction, the DNA was precipitated with ethanol and dissolved in a total of 40 μl of DNA polymerase I buffer, 6 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was allowed to proceed at 16° C. for 90 minutes for rendering the 5′ protruding ends formed upon BglII digestion blunt-ended. The reaction was terminated by extraction with phenol, the mixture was extracted with chloroform and then subjected to ethanol precipitation, the DNA obtained was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride, 10 units of HindIII was added, and the mixture was incubated at 37° C. for 2 hours to effect cleavage at the HindIII site. The reaction mixture was subjected to agarose gel electrophoresis and 0.1 μg of a DNA fragment (about 0.8 kb) containing the immunoglobulin promoter region was recovered. Then, 0.2 μg of the plasmid pUC18 [Messing: Methods in enzymology 101, 20 (1983)] was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of HindIII and 10 units of XbaI were added, and the mixture was incubated at 37° C. for 2 hours for causing cleavage at the HindIII and XbaI sites. The reaction mixture was subjected agarose gel electrophoresis and 0.1 μg of a DNA fragment of about 2.7 kb in size was recovered. The thus-obtained pPKMD6-derived 0.8 kb DNA fragment (0.1 μg), immunoglobulin enhancer region-containing DNA fragment (0.02 μg) and pUC18 (0.1 μg) were dissolved in 20 μl of T4 ligase buffer, 175 units of T4 DNA ligase was added, and the mixture was incubated at 4° C. for 1 day. The reaction mixture was used to transform Escherichia coli HB101 to give an Ap-resistant colony. The recombinant plasmid DNA was recovered from this colony to give pEPKMA1 shown in FIG. 31.

[0287] Then, 1 μg of the plasmid pEPKMA1 was dissolved in 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of XbaI was added, and the mixture was incubated at 37° C. for 2 hours for causing cleavage at the XbaI site. After phenol-chloroform extraction, the resultant DNA fragment was precipitated with ethanol and dissolved in a total of 40 μl of DNA polymerase I buffer, 6 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was allowed to proceed at 16° C. for 90 minutes for rendering the cohesive ends formed upon XbaI digestion blunt-ended. The reaction was terminated by extraction with phenol and, after chloroform extraction, the DNA fragment was recovered by ethanol precipitation. This DNA fragment and a synthetic DNA linker XhoI (Takara Shuzo) (0.01 μg) were dissolved in 20 μl of T4 ligase buffer, 175 units of T4 DNA ligase was added, and the mixture was incubated at 4° C. for 1 day. The reaction mixture was used to transform Escherichia coli HB101 to give an Ap-resistant colony. The recombinant plasmid DNA was recovered from this colony to give pEPKMB1 shown in FIG. 32.

[0288] Then, the SV40 early gene promoter and enhancer regions (hereinafter abbreviated as PSE) of the heterologous gene expression vector pAGE107 for use in animals [Miyaji et al.: Cytotechnology, 3, 133-140 (1990)] were replaced with the KM50-derived immunoglobulin H chain promoter and enhancer (hereinafter abbreviated as PIH) of pEPKMB1 in the following manner.

[0289] One μg of the plasmid pAGE107 was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 150 mM sodium chloride, 10 units of SalI and 10 units of XhoI were added, and the mixture was incubated at 37° C. for 2 hours to effect cleavage at the SalI and I sites. The reaction mixture was subjected to agarose gel electrophoresis and 0.5 μg of a DNA fragment (about 5.95 kb) containing the G418 resistance gene, among others, was recovered. Then, 1 μg of the plasmid pEPKMB1 was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 150 mM sodium chloride, 10 units of SalI and 10 units of XhoI were added, and the mixture was incubated at 37° C. for 2 hours to effect cleavage at the SalI and XhoI sites. The reaction mixture was subjected to agarose gel electrophoresis and 0.1 μg of a DNA fragment (about 1.7 kb) containing the immunoglobulin promoter and enhancer regions was recovered. The thus-obtained pAGE107-derived 5.95 kb DNA fragment (0.1 μg) and immunoglobulin promoter and enhancer region-containing DNA fragment (0.02 μg) were dissolved in 20 μl of T4 ligase buffer, 175 units of T4 DNA ligase was added, and the mixture was incubated at 4° C. for 1 day. The reaction mixture was used to transform Escherichia coli HB101. An Ap-resistant colony was isolated and the recombinant plasmid DNA was recovered therefrom to give pAGE501 shown in FIG. 33.

[0290] (4) Construction of pAGE109

[0291] A plasmid, pAGE109, derived from pAGE106 by deletion of one of the two EcoRI sites in pAGE106 was constructed as follows.

[0292] Thus, 2 μg of the heterologous gene expression vector pAGE106 for use in animal cells as described in EP-A-0 405 285 was added to 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride; 10 units each of EcoRI and SacI were further added, and digestion was conducted at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1.5 μg of a DNA fragment (4.3 kb) resulting from cleavage of pAGE106 with EcoRI and SacI and containing the SV40 early gene promoter and G418 resistance gene was recovered. Then, this DNA fragment was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I large fragment was added, and the reaction was conducted at 16° C. for 2 hours for rendering the 3′ protruding ends formed upon SalI digestion and the 5′ protruding ends formed upon EcoRI digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in 20 μl of T4 ligase buffer; 350 units of T4 DNA ligase was further added to the mixed solution, and ligation was carried out at 4° C. for 4 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 to give the plasmid pAGE109 shown in FIG. 34.

[0293] (5) Construction of pAGE502

[0294] For replacing the SV40 promoter and enhancer of pAGE107 with the immunoglobulin H chain promoter and enhancer, a plasmid named pAGE502 was constructed as follows.

[0295] Two μg of pAGE107 described in EP-A-0 405 285 was added to 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride, 10 units of HindIII was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was conducted at 16° C. for 2 hours for rendering the 5′ protruding ends formed upon HindIII digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of XhoI was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1.5 μg of a DNA fragment (about 5.95 kb), resulting from cleavage of pAGE107 with XhoI and HindIII and containing the G418 resistance gene and Ap resistance, was recovered.

[0296] Two μg of pAGE501 obtained in (3) was added to 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 175 mM sodium chloride, 10 units of SalI was further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was conducted at 16° C. for 2 hours for rendering the 5′ protruding ends formed upon SalI digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of XhoI was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.2 μg of a DNA fragment (1.8 kb) resulting from cleavage of pAGE501 with XhoI and SaI and containing the KM50 cell immunoglobulin H chain promoter and enhancer was recovered.

[0297] Then, 0.1 μg of the HindIII-XhoI fragment (about 5.95 kb) of pAGE107 as obtained above and 0.1 μg of the SalI-XhoI fragment (about 1.8 kb) of pAGE501 were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 DNA ligase was added to the solution, and the mixture was incubated at 4° C for 1 day. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 to give the plasmid pAGE502 shown in FIG. 35.

[0298] (6) Construction of pAGE503

[0299] A plasmid named pAGE503 derived from pAGE502 by deletion of one of the two EcoRI sites was constructed as follows.

[0300] Two μAg of pAGE109 obtained in (4) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride; 10 units of HindIII and 10 units of ClaI were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.2 μg of a DNA fragment (about 1 kb) resulting from cleavage of pAGE109 with ClaI and HindIII and containing the poly-A signal gene for the beta globulin and SV40 early genes was recovered.

[0301] Then, 2 μg of pAGE502 obtained in (5) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride, 10 units of HindIII and 10 units of ClaI were further added, and digestion was conducted at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1 μg of a DNA fragment (about 6.1 kb) resulting from cleavage of pAGE502 with HindIII and ClaI and containing the KMSO cell immunoglobulin H chain promoter and enhancer genes, the Ap resistance gene and the G418 resistance gene was recovered by the DEAE paper method. Then, 0.1 μg of the HindIII-ClaI fragment (about 1 kb) of pAGE109 as obtained above and 0.1 μg of the HindIII-ClaI fragment (about 6.1 kb) of pAGE502 as obtained above were dissolved in a total of 20 μl of T4 ligase buffer, 350 units of T4 DNA ligase was added to the solution, and the mixture was incubated at 4° C. for 1 day. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pAGE503 shown in FIG. 36 was obtained.

[0302] (7) Construction of pSE1d1

[0303] A plasmid named pSE1D1 was constructed by introducing the dhfr gene into pAGE107, as follows.

[0304] Two μg of pAGE107 described in EP-A-0 405 825 was added to 100 μl of 100 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride, 10 units of EcoRI was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was conducted at 16° C. for 2 hours for rendering the 5′ protruding ends formed upon EcoRI digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride; 10 units of HindIII was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1.5 μg of a DNA fragment (about 5.6 kb) resulting from cleavage of pAGE107 with EcoRI and HindIII and containing the G418 resistance gene and Ap resistance gene was recovered.

[0305] Two μg of pSV2-dhfr [Subramani et al.: Mol. Cell. Biol., 1, 854 (1981)] was added to 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of BglII was further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was conducted at 16° C. for 2 hours for rendering the 5′ protruding ends formed upon BglII digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of HindIII was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.2 μg of a pSV2-dhfr DNA fragment (0.76 kb) resulting from cleavage with BglII and HindIII and containing the dehydrofolate reductase (dhfr) gene was recovered.

[0306] Then, 0.1 μg of the HindIII-EcoRI fragment (about 5.6 kb) of pAGE107, as obtained above, and 0.1 g of the BglII-HindIII fragment (about 0.76 kb) of pSV2-dhfr, as obtained above, were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 DNA ligase was added to the solution, and the mixture was incubated at 4° C. for 1 day. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pSE1D1 shown in FIG. 37 was obtained.

[0307] (8) Construction of pSE1d2

[0308] A plasmid named pSE1d2 was constructed by deleting the HindIII cleavage site from pSE1d1, as follows.

[0309] Thus, 2 μg of pSE1d1 obtained in (7) was added to 100 ;1 of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride, 10 units of HindIII was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was conducted at 16° C. for 2 hours for rendering the 5′ protruding ends formed upon HindIII digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in 20 μl of T4 ligase buffer, 350 units of T4 DNA ligase was added to the solution, and the mixture was incubated at 4° C. for 1 day. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pSE1d2 shown in FIG. 38 was obtained.

[0310] (9) Construction of pIg1 SE1d2

[0311] A plasmid named pIg1SE1d2 was constructed by introducing the dhfr gene into pAGE503, as follows.

[0312] Two μg of pAGE503 obtained in (6) was added to 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride, 10 units of ClaI was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was conducted at 16° C. for 2 hours for rendering the 5′ protruding ends formed upon ClaI digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride; 10 units of MluI was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1 μg of a DNA fragment (about 5.4 kb) resulting from cleavage of pAGE503 with ClaI and MluI and containing the KM50 immunoglobulin H chain promoter and enhancer was recovered.

[0313] Then, 2 μg of pSE1d2 obtained in (8) was added to 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of XhoI was further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was conducted at 16° C. for 2 hours for rendering the 5′ protruding ends formed upon XhoI digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of MluI was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1 μg of a DNA fragment (about 3.8 kb) resulting from cleavage of pSE1d2 with XhoI and MluI and containing the dhfr gene was recovered.

[0314] Then, 1 μg of the ClaI-MluI fragment (about 5.4 kb) of pAGE503 as obtained above and 1 μg of the XhoI-MluI fragment (about 3.8 kb) of pSE1d2 as obtained above were dissolved in a total of 20 μl of T4 ligase buffer, 350 units of T4 DNA ligase was added to the solution, and the mixture was incubated at 4° C. for 1 day. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pIg1SE1d2 shown in FIG. 39 was obtained.

[0315] (10) Construction of pIg1SE1d3

[0316] A plasmid named pIg1SE1d3 was constructed by deleting the ApaI cleavage site from pIg1SE1d2, as follows.

[0317] Two μg of pIg!SE1d2 obtained in (9) was added to 100 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride, 10 units of ApaI was further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was carried out at 16° C. for 2 hours for rendering the 3′ protruding ends formed upon ApaI digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in 20 μl of T4 ligase buffer, 350 units of T4 ligase was added to the solution, and ligation was effected at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pIg1SE1d3 shown in FIG. 40 was obtained.

[0318] (11) Construction of pIg1SE1d4

[0319] For providing pIg1SE1d3 with a cloning site between the HindIII cleavage site and EcoRI cleavage site, a plasmid named pIg1SE1d4 was constructed containing the synthetic DNA defined by SEQ ID NO:17 as an insert, as follows.

[0320] Two μAg of pIg1SE1d3 obtained in (10) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride, 10 units each of HindIII and EcoRI were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1 μg of a DNA fragment (about 9.2 kb) resulting from cleavage of pIg1SE1d3 with HindIII and EcoRI and containing the KM50 cell immunoglobulin H chain promoter, enhancer, Ap resistance gene, G418 resistance gene and dhfr gene was recovered.

[0321] Then, 0.1 μg of the HindIII-EcoRI fragment (about 9.2 kb) of pIg1SE1d3 as obtained above and 10 ng of the synthetic DNA (SEQ ID NO:17) were a total of 20 μl of T4 ligase buffer, 350 units of T4 DNA ligase was added to the solution, and the mixture was incubated at 4° C. for 1 day. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pIg1SE1d4 shown in FIG. 41 was obtained.

[0322] 3. Preparation of the Moloney Mouse Leukemia Virus Long Terminal Repeat (Hereinafter Abbreviated as “MoLTR”)

[0323] It is known that MoLTR has promoter and enhancer activity [Kuwana et al.: Biochem. Biophys. Res. Commun., 149, 960 (1987)]. Therefore, for using MoLTR as a promoter and enhancer in vectors for chimeric human antibody expression, a plasmid, pPMOL3, containing MoLTR was constructed as follows.

[0324] Three μg of pPMOL1 described in JP-A-1-63394 was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 7 mM magnesium chloride and 6 mM 2-mercaptoethanol, 10 units of ClaI was further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment was added, and the reaction was carried out at 16° C. for 2 hours for rendering the 5′ protruding ends formed upon ClaI digestion blunt-ended. The reaction was terminated by extraction with phenol, the reaction mixture was subjected to chloroform extraction and then to ethanol precipitation, and 2 μg of a DNA fragment was recovered. This DNA fragment and 0.01 μg of a synthetic DNA linker XhoI (Takara Shuzo) were dissolved in 20 μl of T4 ligase buffer, 175 units of T4 DNA ligase was added, and the mixture was incubated at 4° C. for 1 day. The reaction mixture was used to transform Escherichia coli HB101 and the plasmid pPMOL2 shown in FIG. 42 was obtained. Then, 3 μg of pPMOL2 was added to 30 pl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 7 mM magnesium chloride, 10 mM sodium chloride and 6 mM 2-mercaptoethanol, 10 units of SmaI was further added, and digestion was conducted at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, and 2 μg of a DNA fragment was recovered. This DNA fragment and 0.01 μg of a synthetic DNA linker (EcoRI; Takara Shuzo) were dissolved in 20 μl of T4 ligase buffer, 175 units of T4 DNA ligase was added, and the mixture was incubated at 4° C. for 1 day. The reaction mixture was used to transform Escherichia coli HB101 and the plasmid pPMOL3 shown in FIG. 43 was obtained.

[0325] 4. Cloning of the Human Immunoglobulin IgG1 H Chain Constant Region (Cy1) cDNA and L Chain Constant Region (Cx) cDNA

[0326] (1) Isolation of mRNA From the Chimeric Antibody Producer Cell Line SP2-PC Chimera-1

[0327] Using mRNA extraction kit Fast Track (product number K1593-02) manufactured by Invitrogen, mRNA (6.2 μg) was isolated from 1×108 cells of the chimeric antibody producer cell line SP2-PC Chimera-1 described in FEBS Letters, 244, 301-306 (1989) and capable of producing a chimeric antibody having anti-phosphorylcholine activity.

[0328] (2) Construction of an SP2-PC Chimera-1 cDNA Library and Cloning of the Human Immunoglobulin H Chain Constant Region (Cy1) cDNA and L Chain Constant Region (Cκ) cDNA

[0329] Starting with 2 μg of the mRNA obtained in (1) and using cDNA Synthesis Kit (product number 27-9260-01) manufactured by Pharmacia, EcoRI adapter joining was performed, followed by phosphorylation. The cDNA solution obtained was subjected to phenol-chloroform extraction and then to ethanol precipitation, and 4 μg of cDNA was recovered. This cDNA was dissolved in 20 μl of sterilized water and then fractionated by agarose gel electrophoresis, and about 0.3 μg each of two DNA fragments, about 1.8 kb and about 1.0 kb in size, were recovered.

[0330] Then, 5 μg of the vector pUC18 was added to 100 μl of 100 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 50 units of EcoRI was further added, and digestion was carried out at 37° C. for 4 hours for cleaving the pUC18 DNA at the EcoRI site. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, and about 3 μg of a DNA fragment resulting from cleavage of pUC18 at the EcoRI site thereof was recovered.

[0331] Then, 0.1 μg of the EcoRI fragment (about 2.7 kb) of pUC18 as obtained above and 0.1 μg each of the 1.8 kb and 1.0 kb cDNA fragments prepared from SP2-PC Chimera-1 cells were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 DNA ligase was added to the solution, and ligation was effected at 4° C. for 24 hours.

[0332] The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli LE392. About 3,000 colonies obtained were fixed onto a nitrocellulose filter. From among the strains firmly bound at 65° C. to probes prepared by labeling the human immunoglobulin constant region chromosomal genes (IgG1 H chain constant region Cy1 and L chain constant region Cκ) [Kameyama et al.: FOBS Letters, 2, 301 (1989)] with 32P, a plasmid (pPCVHhCGI1) associable with Cy1 and another (pPCVLhCκ1) associable with CK were isolated.

[0333] (3) Introduction of an EcoRV Site Into the Human Igκ Chain Constant Region Gene

[0334] An EcoRV site was introduced into the human Igκ chain constant region at a site near the 5′ end thereof by site-directed mutagenesis using a kit (catalog number Q6210) manufactured by Promega. The plasmid pPCVLhCK1 (2 μg) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride, 10 units of EcoRI and 10 units of KpnI were further added, and digestion was conducted at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.2 μg of a DNA fragment (about 0.8 kb) resulting from cleavage of pPCVLhCK1 with KpnI and EcoRI and containing the human immunoglobulin L chain constant region gene was recovered.

[0335] Then, 2 μg of pSELECT1 (a kit manufactured by Promega; catalog number Q6210) was added to 30 μl of 10 mM Tris-hydro-chloride buffer (pH 7.5) containing 6 mM magnesium chloride and 50 mM sodium chloride, 10 units each of EcoRI and KpnI were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1 μg of a DNA fragment (about 5.7 kb) resulting from cleavage of pSELECT1 with EcoRI and KpnI was recovered.

[0336] Then, 0.1 μg of the EcoRI-KpnI fragment (about 0.8 kb) of pPCVLhCK1 as obtained above and 0.1 μg of the EcoRI-KpnI fragment (about 5.7 kb) of pSELECT1 as obtained above were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 DNA ligase was added to the solution, and ligation was effected at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli JM109 and the plasmid pchCKA7 shown in FIG. 44 was obtained.

[0337] Then, using pchCKA7 and using the synthetic DNA defined by SEQ ID NO:18 as a mutagenic primer, the sequence covering the 12th base to 14 base from the N terminus of the human immunoglobulin L chain constant region, namely ACC, was converted to GAT and thus an EcoRV site was introduced into that site, to give a plasmid named pchCKB1 (FIG. 45).

[0338] Then, the EcoRV site of pchCKB1 was converted to a HindIII cleavage site in the following manner.

[0339] Thus, 2 μg of the plasmid pchCKB1 was added to 10 μl of 100 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of EcoRI was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the precipitate was dissolved in a total of 40 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived polymerase I Klenow fragment was added, and the reaction was carried out at 37° C. for 30 minutes for rendering the 5′ protruding ends formed upon EcoRI digestion blunt-ended. The reaction mixture was subjected to phenol-chloroform extraction and then ethanol precipitation, the precipitate was dissolved, together with 0.1 μg of a HindIII linker (Takara Shuzo), in 20 μl of T4 ligase buffer; 350 units of T4 ligase was added to the solution, and ligation was effected at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli H3101 and the plasmid pchCKC1 shown in FIG. 46 was obtained.

[0340] 5. Construction of vectors for Chimeric Human Antibody H Chain Expression

[0341] (1) Construction of a Vector to be Used in Constructing Chimeric Human Antibody H Chain Expression Vectors (Vector for Chimeric Human Antibody H Chain Expression)

[0342] The plasmid pIg1SE1d4 obtained in Paragraph 2 (11) (2 μg) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units each of EcoRV and ApaI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1.5 μg of a DNA fragment (about 9.2 kb) resulting from cleavage of pIg1SE1d4 with EcoRV and ApaI was recovered.

[0343] Then, 2 μg of pPCVHhCGI1 obtained in Paragraph 4 (2) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride, 10 units of ApaI and 10 units of SmaI were further added, and digestion was conducted at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.2 μg of a DNA fragment (about 1 kb) resulting from cleavage of pPCVHhCGI1 with ApaI and SmaI and containing the human immunoglobulin H chain constant region gene was recovered.

[0344] Then, 0.1 μg of the EcoRV-ApaI fragment (about 9.2 kb) of pIg1SE1d4 as obtained above and 0.1 μg of the ApaI-SmaI fragment (about 1 kb) of pPCVHhCGI1 as obtained above were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 DNA ligase was added to the solution, and ligation was conducted at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the vector pCHiIgHB2 for chimeric human antibody H chain expression as shown in FIG. 47 was obtained.

[0345] (2) Construction of a Vector to be Used in Constructing Chimeric Human Antibody L Chain Expression Vectors (Vector for Chimeric Human Antibody L Chain Expression)

[0346] The plasmid pIg1SE1d4 obtained in Paragraph 2 (11) (2 μg) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of EcoRV and 10 units of HindIII were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 1.5 μg of a DNA fragment (about 9.2 kb) resulting from cleavage of pIg1SE1d4 with EcoRV and HindIII was recovered.

[0347] Then, 2 μg of pckCKC1 obtained in Paragraph 4 (3) was added to 30 μl of 10 mM Tris-hydrochloride (pH 7.5) containing 6 mM magnesium chloride and 100 mM sodium chloride, 10 units of EcoRV and 10 units of HindIII were further added, and digestion was carried out at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.2 μg of a DNA fragment (about 0.6 kb) resulting from cleavage of pPCVLhCK1 with EcoRV and HindIII and containing the human immunoglobulin L chain constant region gene was recovered.

[0348] Then, 0.1 μg of the EcoRV-HindIII fragment (about 9.2 kb) of pIg1SE1d4 as obtained above and 0.1 μg of the EcoRV-HindIII fragment (about 0.6 kb) of pchCKC1 as obtained above were dissolved in a total of 20 μl of T4 ligase buffer, 350 units of T4 DNA ligase was added to the solution, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the vector pChiIgLA1 for chimeric human antibody L chain expression as shown in FIG. 48 was obtained.

Reference Example 2

[0349] Construction of a Chimeric Human Antibody H Chain Expression Vector, pChi641HA1

[0350] 1. Isolation of mRNA From Mouse Anti-GD3 Monoclonal Antibody KM-641-producing Hybridoma Cells

[0351] Using mRNA extraction kit Fast Track (product number K1593-02) manufactured by Invitrogen, 34 μg of mRNA was isolated from 1×103 mouse anti-GD3 monoclonal antibody KM-641-producing hybridoma cells obtainable as in Reference

Example 1.

[0352] 2 . Construction of a KM-641 H Chain cDNA Library and a KM-641 L Chain cDNA Library

[0353] Using 3 μg of the mRNA obtained in Paragraph 1 and using cDNA synthesis kit ZAP-cDNA Synthesis Kit (product number sc200400) manufactured by Stratagene, cDNA having an EcoRI adapter at the 5′ terminus and cDNA having an XhoI adapter at the 3′ terminus were synthesized. About 6 μg or each cDNA was dissolved in 10 μl of sterilized water and fractionated by agarose gel electrophoresis. In this way, about 0.1 μg of a cDNA fragment having a size of about 1.8 kb and corresponding to the H chain and a cDNA fragment having a size of about 1.0 kb and corresponding to the L chain were recovered. Then, 0.1 μg of the cDNA fragment of about 1.8 kb in size, 0.1 μg of the cDNA fragment of about 1.0 kb in size and 1 μg of Uni-ZAP XR (Stratagene; derived from the Lambda ZAPII vector by cleavage with EcoRI and XhoI, followed by treatment with calf intestine alkaline phosphatase), to be used as the vector, were dissolved in T4 ligase buffer; 175 units of T4 DNA ligase was added, and the mixture was incubated at 12° C. for 10 hours and further at room temperature for 2 hours. A 4-μl portion of the reaction mixture was packaged into the lambda phage by the conventional method [Maniatis et al. (ed.): Molecular Cloning, 1989, p. 2.95] using Giga Pak Gold (Stratagene), followed by transfection of Escherichia coli PLK-F with the packaging mixture by the conventional method [Maniatis et al. (ed.) : Molecular Cloning, 1989, p. 2.95-107]. As an H chain cDNA library and as an L chain cDNA library, about 10,000 phage clones were respectively obtained. The phages were then fixed onto nitrocellulose filters by the conventional method [Maniatis et al. (ed.): Molecular Cloning, 1989, p. 2.112].

[0354] 3. Cloning of the Monoclonal Antibody KM-641 Chain and L Chain cDNAs

[0355] Using probes prepared by labeling a mouse Cy1 gene (mouse immunoglobulin constant region chromosomal gene)-containing EcoRI fragment (about 6.8 kb) [Roeder et al.: Proc. Natl. Acad. Sci. U.S.A., 78, 474 (1981)] and a mouse Cκ gene-containing HindIII-BamHI fragment (about 3 kb) [Sakano et al.: Nature, 280, 288 (1979)] with 32P, one phage clone strongly associable with the former probe at 65° C. and one phage clone strongly associable with the latter probe at 65° C. were isolated from the H chain cDNA library and L chain cDNA library constructed in Paragraph 2 in accordance with the conventional method [Maniatis et al. (ed.): Molecular Cloning, 1989, p. 2.108]. Then, by converting the phage clones to pBluescript plasmids using cDNA synthesis kit ZAP-cDNA Synthesis Kit (product number sc200400) manufactured by Stratagene, a KM-641 H chain cDNA-containing recombinant plasmid, pKM641HA3, and a KM-641 L chain cDNA-containing recombinant plasmid, pKM641LA2, were obtained. Cleavage of pKM641HA3 and pKM641LA2 with EcoRI and XhoI revealed that a cDNA fragment of about 1.6 kb and a cDNA fragment of about 0.9 kb had been inserted therein, respectively (FIG. 49).

[0356] 4. Base Sequences of the Immunoglobulin Variable Regions in the KM-641 H Chain cDNA (pKM641.A3) and KM-641 L Chain cDNA (pKM641LA2)

[0357] The base sequences of the immunoglobulin regions in pKM641HA3 and pKM641LA2 obtained in Paragraph 3 were determined by the dideoxy method [Maniatis et al. (ed.): Molecular Cloning, 1989, p. 13.42] using Sequenase Version 2.0 DNA Sequencing Kit (United States Biochemical Corporation). The results obtained are shown in SEQ ID NO:19 and SEQ ID NO:20. In pKM641LA2, a methionine codon, presumably the initiation codon ATG, was found in the vicinity of the 5′ terminus and the cDNA was a leader sequence-containing full-length one. In pKM641HA3, no methionine initiation codon was found and the leader sequence was partly lacking.

[0358] 5. Construction of a KM-641-derived Chimeric Human Antibody H Chain Expression Vector

[0359] A chimeric human antibody H chain expression vector was constructed by joining the H chain variable region gene obtained by cleaving the plasmid pKM641HA3 at the AluI site near the 5′ terminus of the variable region gene and at the StyI site near the 3′ terminus of the variable region gene to the vector for chimeric human antibody H chain expression as obtained in Reference Example 1 using the synthetic DNAs defined by SEQ ID NO:21 and SEQ ID NO:22 (FIG. 50).

[0360] First, the DNA defined by SEQ ID NO:22 composed of the base sequence from the 3′ terminus of the immunoglobulin H chain variable region in pKM641HA3 to the StyI cleavage site near said 3′terminus and the base sequence from the 5′ terminus of the immunoglobulin H chain constant region in pAGE28 to the ApaI cleavage site near said 5′ terminus and having a StyI cleavage site and an ApaI cleavage site on the respective termini (cf. FIG. 50) was synthesized using a DNA synthesizer. This synthetic DNA was then introduced into the plasmid pKM641HA3 in the following manner.

[0361] Three μg of pKM641HA3 was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of StyI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a 0.41 kb DNA fragment was recovered. Then, 3 μg of pAGE28 [Mizukami et al. : J. Biochem., 101, 1307-1310 (1987)] was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 7 mM magnesium chloride and 6 mM 2-mercaptoethanol, 10 units of EcoRI and 10 units of ApaI were further added, and digestion was carried out at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 2 μg of a 2.45 kb DNA fragment was recovered. Then, 0.1 μg of the EcoRI-StyI fragment (about 0.41 kb) of pKM641HA3, as obtained above, 0.1 μg of the EcoRI-ApaI fragment (about 2.45 kb) of pAGE28, as obtained above, and 0.3 μg of the synthetic DNA, defined by SEQ ID NO:22, were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was added to the solution, and ligation was conducted at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pKM641HE1 shown in FIG. 51 was obtained.

[0362] Since pEM641HE1 had no leader sequence, the following measure was taken to supplement the deficit using the synthetic DNA defined by SEQ ID NO:21.

[0363] pKM641HE1 (3 μg) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 7 mM magnesium chloride and 6 mM 2-mercaptoethanol, 10 units of EcoRI and 10 units of ApaI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.4 μg of a DNA fragment of about 0.42 kb in size was recovered. The EcoRI-ApaI fragment (about 0.42 kb; 0.4 μg) of pKM641HE1 was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 7 mM magnesium chloride, 50 mM sodium chloride and 6 mM 2-mercaptoethanol, 10 units of AluI was further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, and about 0.3 μg of a DNA fragment of about 0.4 kb in size was recovered.

[0364] Then, 0.1 μg of the AluI-ApaI fragment (about 0.4 kb) of pKM641HE1 as obtained above, 0.1 μg of the EcoRI-ApaI fragment (about 2.45 kb) of pAGE28 as obtained above and 0.3 μg of the synthetic DNA defined by SEQ ID NO:21 were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was added to the solution, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pKM641HF1 shown in FIG. 52 was obtained.

[0365] Then, the immunoglobulin H chain variable region of pKM641HF1 was introduced into the vector pChiIgHB2 for chimeric human antibody H chain expression, as follows.

[0366] pKM641HF1 (3 μg) was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 7 mM magnesium chloride and 6 mM 2-mercaptoethanol, 10 units of EcoRI and 10 units of ApaI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.5 μg of a 0.44 kb DNA fragment was recovered. Then, 3 μg of pChiIgHB2 was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 7 mM magnesium chloride and 6 mM 2-mercaptoethanol, 10 units of EcoRI and 10 units of ApaI were further added, and digestion was conducted at 37° C. for 4 hours. The reaction mixture was subjected to phenol-chloroform extraction and about 3 μg of DNA was recovered. Then, 0.1 μg of the EcoRI-ApaI fragment (about 0.44 kb) of pKM641HF1 as obtained above and 0.1 μg of the EcoRI-ApaI fragment (about 10.1 kb) of pChiIgHB2 as obtained above were dissolved in a total of 20 μl of T4 ligase buffer; 350 units of T4 ligase was added to the solution, and ligation was carried out at 4° C. for 24 hours. The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid pChi641HA1 shown in FIG. 53 was obtained.

[0367] Then, the KM50-derived immunoglobulin H chain promoter and enhancer region of pChi641HA1 was replaced with MoLTR, as follows. pChi641HA1 (3 μg) was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of XhoI were further added, and digestion was effected at 37° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.2 μg of a DNA fragment of about 8.8 kb in size was recovered. pPMOL3 (3 μg) obtained in Example 1, Paragraph 2 was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT; 10 units of EcoRI and 10 units of XhoI were further added, and digestion was carried out at 370° C. for 4 hours. The reaction mixture was fractionated by agarose gel electrophoresis and about 0.3 μg of a MoLTR-containing DNA fragment (0.63 kg) was recovered. Then, 0.1 μg of the EcoRI-XhoI fragment of pChi641HA1 and 0.1 μg of the EcoRI-XhoI fragment of pPMOL3 were dissolved in 20 μl of T4 ligase buffer, 175 units of T4 DNA ligase was added, and the mixture was incubated at 4° C. for 1 day. The reaction mixture was used to transform Escherichia coli HB101 and the KM-641-derived chimeric human H chain expression vector pChi641HA1 shown in FIG. 54 was obtained.

EXAMPLE 2

Production of Human CDR-grafted Anti-GM2 Antibodies (1)

[0368] 1. Construction of DNAs Each Coding for Human CDR-grafted anti-GM2 Antibody H Chain Variable Region and Human CDR-grafted anti-GM2 Antibody L Chain Variable Region

[0369] (1) Construction of DNA Coding for Human CDR-grafted Anti-GM2 Antibody H Chain Variable Region

[0370] A DNA coding for a human CDR-grafted anti-GM2 antibody H chain variable region, hKM796H, which contains amino acid sequences of SEQ ID NO:6, SEQ ID NO:7 and SEQ ID NO:8, was constructed in the following manner.

[0371] NEWM [BIO/TECHNOLOGY, 9, 266 (1991)] was used as human antibody H chain variable region-encoding DNA to which each CDR was to be transplanted. DNAs set forth in SEQ ID NO:23 through NO:29 corresponding to NEWM in which each CDR was replaced with amino acid sequences of SEQ ID NO:6, SEQ ID NO:7 and SEQ ID NO:8 were synthesized using an automatic DNA synthesizer (model 380A manufactured by Applied Biosystems Co., Ltd.). The thus-obtained synthetic DNAs (50 picomoles each) were dissolved in 20 μl of 50 mM Tris-hydrochloride buffer (pH 7.6) containing 10 mM magnesium chloride, 5 mM DTT, 0.1 mM EDTA and 0.5 mM ATP, 5 units of T4 polynucleotide kinase was added, and 5′-phosphorylation was carried out at 37° C. for 30 minutes. Ten picomoles each of the resulting phosphorylated synthetic DNAs, which had restriction enzyme sites on both ends, were ligated in the order of SEQ ID NO. (SEQ ID NO:23 through NO:29) using a DNA ligation kit (Takara Shuzo) in accordance with the manufacturer's instruction attached to the kit to obtain a DNA, hKM796H, shown in FIG. 55. The amino acid sequence corresponding to hKM796H is shown in SEQ ID NO:36.

[0372] (2) Construction of DNA coding for human CDR-grafted anti-GM2 antibody L chain variable region

[0373] A DNA coding for a human CDR-grated anti-GM3 antibody L chain variable region, hKM796L, which contains amino acid sequences of SEQ ID NO:9, SEQ ID NO:10 and SEQ ID NO:11, was constructed in the following manner.

[0374] REI [BIO/TECHNOLOGY, 9, 266 (1991)] was used as human antibody L chain variable region-encoding DNA to which each CDR was to be transplanted. DNAs set forth in SEQ ID NO:30 through NO:35 corresponding to REI in which each CDR was replaced with amino acid sequences of SEQ ID NO:9, SEQ ID NO:10 and SEQ ID NO:11 were synthesized using an automatic DNA synthesizer (model 380A manufactured by Applied Biosystems Co., Ltd.). The thus-obtained synthetic DNAs (50 picomoles each) were dissolved in 20 μl of 50 mM Tris-hydrochloride buffer (pH 7.6) containing 10 mM magnesium chloride, 5 mM DTT, 0.1 mM EDTA and 0.5 mM ATP, 5 units of T4 polynucleotide kinase was added, and 5′-phosphorylation was carried out at 37° C. for 30 minutes. Ten picomoles each of the resulting phosphorylated synthetic DNAs, which had restriction enzyme sites on both ends, were ligated in the order of SEQ ID NO. (SEQ ID NO:30 through NO:35) using a DNA ligation kit (Takara Shuzo) in accordance with the manufacturer's instruction attached to the kit to obtain a DNA, hKM796L, shown in FIG. 56. The amino acid sequence corresponding to hKM796L is shown in SEQ ID NO:37.

[0375] 2. Construction of Human CDR-grafted Antibody H Chain Expression Vector and Human CDR-grafted Antibody L Chain Expression Vector

[0376] (1) Construction of Human CDR-grafted Antibody H Chain Expression Vector

[0377] A NotI-ApaI fragment of the DNA coding for human CDR-grafted antibody H chain variable region, obtained in Paragraph 1(1) of Example 2, was ligated to the plasmid pChi796HM1 obtained in Paragraph 7(3) of Example 1, in the following manner (FIG. 57).

[0378] Three μg of pChi796HM1, obtained in Paragraph 7(3) of Example 1, were dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 nM magnesium chloride and 1 mM DTT, 10 units of ApaI were added thereto and the mixture was allowed to react at 37° C. for 1 hour. The resulting mixture was subjected to ethanol precipitation and the thus-obtained precipitate was dissolved in 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. Ten units of NotI were added thereto to allow the mixture to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 2 μg of a DNA fragment of about 9.0 kb. Then, about 0.1 μg of the thus-obtained ApaI-NotI fragment of pChi796HM1 was ligated to 0.5 pmoles of the NotI-ApaI fragment of the DNA coding for human CDR-grafted antibody H chain variable region, obtained in Paragraph 1(1) of Example 2, using a DNA ligation kit (Takara Shuzo). The resulting recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid phKM796HM1 shown in FIG. 57 was obtained.

[0379] Then, a human CDR-grafted antibody H chain expression vector was constructed by introducing β-globin 3′ splicing signal into the plasmid phKM796HM1 in the following manner (FIG. 58).

[0380] Three μg of phKM796HM1 were added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 1 unit of KpnI was added thereto. The mixture was allowed to react at 37° C. for 10 minutes to effect partial digestion. The resulting mixture was subjected to ethanol precipitation and the thus-obtained precipitate was dissolved in 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. After adding 1 unit of XhoI, the mixture was allowed to react at 37° C. for 10 minutes to effect partial digestion. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.2 μg of a DNA fragment of about 2.1 kb. Separately, 3 μg of pAGE148, obtained in Paragraph 7(2) of Example 1, were added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT. Ten units of KpnI were added thereto to allow the mixture to react at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation and the thus-obtained precipitate was dissloved in 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. After adding 10 units of XhoI, the mixture was allowed to react at 37° C. for 1 hour and then fractionated by agarose gel electrophoresis to recover about 1 μg of a DNA fragment of about 8.7 kb. One tenth μg of the thus-obtained XhoI-KpnI fragment of phKM796HM1 was ligated to 0.1 μg of the XhoI-KpoI fragment of pAGE148 using a DNA ligation kit (Takara Shuzo). The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 to obtain the plasmid phKM796HMS1 shown in FIG. 58.

[0381] (2) Construction of Human CDR-grafted Antibody L Chain Expression Vector

[0382] An EcoRI fragment having blunt ends of the DNA coding for human CDR-grafted antibody L chain variable region, obtained in Paragraph 1 (2) of Example 2, was ligated to the chimeric human antibody L chain expression vector pChiIgLA1 in the following manner (FIG. 59).

[0383] Three μg of pChiIgLA1, obtained in Reference Example 1, were added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of EcoRV were added thereto and the mixture was allowed to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 1 μg of a DNA fragment of about 8.6 kb. Then, about 0.1 μg of the thus-obtained EcoRI-EcoRV fragment of pChiIgLA1 was ligated to 0.5 pmoles of the EcoRI fragment having blunt ends derived from the DNA coding for human CDR-grafted antibody L chain variable region, obtained in Paragraph 1(2) of Example 2, using a DNA ligation kit (Takara Shuzo). The resulting recombinant plasmid DNA was used to transform Escherichia coli HB101 and the plasmid phKM796LI1 shown in FIG. 59 was obtained.

[0384] Then, PMO was introduced into the plasmid phKM796LI1 in the following manner (FIG. 60).

[0385] Three μg of phKM796LI1 were added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of EcoRI and 10 units of XhoI were added thereto, and the mixture was allowed to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 1 μg of a DNA fragment of about 8.2 kb. Separately, 3 μg of the chimeric human antibody H chain expression vector pChi641HAM1, obtained in Reference Example 2, were added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. Ten units of EcoRI and 10 units of XhoI were added thereto to allow the mixture to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.3 μg of a DNA fragment of about 0.6 kb. One tenth μg of the thus-obtained EcoRI-XhoI fragment of pChi641HM1 was ligated to 0.1 μg of the EcoRI-XhoI fragment of phKM796LI1 using a DNA ligation kit (Takara Shuzo). The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 to obtain the plasmid phKM796LM1 shown in FIG. 60.

[0386] Then, a human CDR-grafted antibody L chain expression vector was constructed by introducing β-globin 3′ splicing signal into the plasmid phKM796LM1 in the following manner (FIG. 61).

[0387] Three μg of phKM796LM1 were added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of KpnI were added thereto, and the mixture was allowed to react at 37° C. for 1 hour. The resulting mixture was subjected to ethanol precipitation and the thus-obtained precipitate was dissolved in 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. After adding 10 units of XhoI, the mixture was allowed to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.3 μg of a DNA fragment of about 1.6 kb. Separately, 3 μg of pAGE148, obtained in Paragraph 7(2) of Example 1, were added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT. Ten units of KpnI were added thereto to allow the mixture to react at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation and the thus-obtained precipitate was dissloved in 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. After adding 10 units of XhoI, the mixture was allowed to react at 37° C. for 1 hour and then fractionated by agarose gel electrophoresis to recover about 1 μg of a DNA fragment of about 8.7 kb. One tenth μg of the thus-obtained XhoI-KpnI fragment of phKM796LM1 was ligated to 0.1 μg of the XhoI-KpnI fragment of pAGE148 using a DNA ligation kit (Takara Shuzo). The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 to obtain the plasmid phKM796LMS1 shown in FIG. 61.

[0388] 3. Construction of Human CDR-grafted Antibody H Chain and L Chain Tandem Expression Vector

[0389] A tandem expression vector containing both of cDNA coding for human CDR-grafted antibody H chain and cDNA coding for human CDR-grafted antibody L chain was constructed in the following manner (FIG. 62 and FIG. 63).

[0390] Three μg of phKM796HMS1, obtained in Paragraph 2(1) of Example 2, were dissolved in 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 1 unit of SalI was added thereto and the mixture was allowed to react at 37° C. for 10 minutes to effect partial digestion. The resulting mixture was subjected to ethanol precipitation and the thus-obtained precipitate was dissolved in 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT. Ten units of MluI was added thereto to allow the mixture to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.2 μg of a DNA fragment of about 5.9 kb. Then, about 2 μg of pAGE107 as described in EP-A-0 405 285 was added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of MluI and 10 units of SalI were added thereto, and the mixture was allowed to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.2 μg of a DNA fragment of about 3.35 kb. Then, 0.1 μg of the thus-obtained MluI-SalI fragment of phKM796HMS1 was ligated to 0.1 μg of the MluI-SalI fragment of pAGE107 using a DNA ligation kit (Takara Shuzo) The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 to obtain the plasmid phKM796H107 shown in FIG. 62.

[0391] Then, 3 μg of phKM796H107 were added to 30 μl of 10 mM Tris-hydrochloride (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of ClaI was added thereto and the mixture was allowed to react at 37 ° C. for 1 hour. The reaction mixture was subjected to phenol-chloroform extraction and ethanol precipitation. The resulting precipitate was dissolved in 20 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment were added, and the 5′ cohesive ends produced by ClaI digestion were rendered blunt by incubation at 22° C. for 30 minutes. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.2 μg of a DNA fragment of about 3.35 kb. The reaction mixture was also subjected to phenol-chloroform extraction and then to ethanol precipitation. The resulting precipitate was dissolved in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of MluI were added thereto and the mixture was allowed to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.3 μg of a DNA fragment of about 7.5 kb. Separately, 3 μg of phKM796LMS1 were added to 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of XhoI were added and the mixture was allowed to react at 37° C. for 1 hour. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation. The resulting precipitate was dissloved in 20 μl of DNA polymerase I buffer, 5 units of Escherichia coli-derived DNA polymerase I Klenow fragment were added, and the 5′ cohesive ends produced by XhoI digestion were rendered blunt by incubation at 22° C. for 30 minutes. The reaction mixture was subjected to phenol-chloroform extraction followed by ethanol precipitation. The resulting precipitate was added to 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT, 10 units of MluI were added thereto, and the mixture was allowed to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.3 μg of a DNA fragment of about 9.3 kb. Then, 0.1 μg of the thus-obtained Mlu-ClaI fragment of phKM796H107 was ligated to 0.1 μg of the MluI-XhoI fragment of phKM796LMS1 using a DNA ligation kit (Takara Shuzo). The thus-obtained recombinant plasmid DNA was used to transform Escherichia coli HB101 to obtain the plasmid phKM796HL1 shown in FIG. 63.

[0392] 4. Expression of Human CDR-grafted Anti-GM2 Antibody in YB2/0 cells

[0393] The plasmids were introduced into YB2/0 cells by the electroporation method of Miyaji et al. [Cytotechnology, 3, 133 (1990)].

[0394] After introduction of 4 μg of phKM796HL1 obtained in Paragraph 3 of Example 2 into 4×106 YB2/0 (ATCC CRL1581) cells, the cells were suspended in 40 ml of RPMI1640-FCS(10) [RPMI1640 medium (Nissui Pharmaceutical) containing 10% of FCS, ¼ volume of 7.5% NaHCO3, 3% of 200 mM L-glutamine solution (Gibco) and 0.5% of penicillin-streptomycin solution (Gibco; containing 5,000 units/ml penicillin and 5,000 μg/ml streptomycin)], and the suspension was distributed in 200-μl portions into wells of 96-well microtiter plates. After 24 hours of incubation at 37° C. in a CO2 incubator, G418 (Gibco) was added to a concentration of 0.5 mg/ml and then incubation was continued for 1 to 2 weeks. Transformant colonies appeared, the culture fluid was recovered from each well in which the cells had grown to confluence and an enzyme-linked immunosorbent assay (ELISA) described in Paragraph 11 of Example 1 was conducted for anti-GM2 human CDR-grafted antibody activity measurement.

[0395] The clone showing the highest activity in ELISA among the clones obtained gave a human CDR-grafted anti-GM2 antibody content of about 0.1 μg/ml of culture fluid.

[0396] Cells of the clone showing the above-mentioned human CDR-grafted anti-GM2 antibody activity were suspended in RPMI1640-FCS(10) medium containing 0.5 mg/ml G418 and 50 nM MTX to a concentration of 1 to 2×105 cells/ml, and the suspension was distributed in 2-ml portions into wells of 24-well plates. Incubation was performed at 37° C. in a CO2 incubator for 1 to 2 weeks to induce 50 nM MTX-resistant clones. At the time of confluence, the human CDR-grafted anti-GM2 antibody activity in each culture fluid was determined by ELISA The 50 nM MTX-resistant clone showing the highest activity among the clones obtained showed a human CDR-grafted anti-GM2 antibody content of about 1.0 μg/ml.

[0397] Cells of the above 50 nM MTX-resistant clone were suspended in RPMI1640-FCS(10) medium containing 0.5 mg/ml G418 and 200 nM MTX to a concentration of 1 to 2×105 cells/ml, and the suspension was distributed in 2-ml portions into wells of 24-well plates. Incubation was carried out at 37° C. in a CO2 incubator for 1 to 2 weeks to induce 200 nM MTX-resistant clones. At the time of confluence, each culture fluid was assayed for human CDR-grafted anti-GM2 antibody activity by ELISA. The 200 nM MIX-resistant clone showing the highest activity among the clones obtained had a human CDR-grafted anti-GM2 antibody content of about 5.0 μg/ml.

[0398] As described in detail hereinabove, the present invention provides humanized antibodies reacting with the ganglioside GM2.

EXAMPLE 3

Production of Human CDR-grafted Anti-GM2 Antibodies (2)

[0399] 1. Construction of Tandem Cassette Type Humanized Antibody Expression Vector, pKANTEX93

[0400] A tandem cassette type humanized antibody expression vector, pKANTEX93, for the expression of a human CDR-grafted antibody in mammalian cells was constructed based on the plasmid pSE1UK1SEd1-3 described in JP-A-2-257891 by inserting a DNA fragment coding for a human CDR-grafted anti-GM2 antibody H chain variable region and a DNA fragment coding for a human CDR-grafted anti-GM2 antibody L chain variable region into said plasmid upstream of the human antibody γ 1H chain constant region cDNA and human antibody κ L chain constant region cDNA, respectively, in the following manner.

[0401] (1) Modification of ApaI and EcoRI Restriction Enzyme Sites Occurring in Rabbit β-globin Gene Splicing and Poly A Signals

[0402] For making it possible to construct a human CDR-grafted antibody expression vector by inserting human CDR-grafted antibody variable regions cassette-wise in the form of NotI-ApaI (H chain) and EcoRI-SplI (L chain) restriction fragments into a vector for human CDR-grafted antibody expression, the ApaI and EcoRI restriction sites occurring in the rabbit β-globin gene splicing and poly A signals of the plasmid pSE1UK1SEd1-3 were modified in the following manner.

[0403] Three μg of the plasmid pBluescript SK(−) (Stratagene) was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the digestion reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, and the 3′ cohesive ends resulting from ApaI digestion were rendered blunt using DNA Blunting Kit (Takara Shuzo), followed by ligation using DNA Ligation Kit (Takara Shuzo). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101. Thus was obtained a plasmid, pBSA, shown in FIG. 64. Furthermore, 3 μg of the plasmid pBSA thus obtained was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of the restriction enzyme EcoRI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, and the 5′ cohesive ends resulting from SRI digestion were rendered blunt using DNA Blunting Kit (Takara Shuzo), followed by ligation using DNA Ligation Kit (Takara Shuzo). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101. Thus was obtained the plasmid pBSAE shown in FIG. 65.

[0404] Then, 3 μg of the thus-obtained plasmid pBSAE was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT, 10 units of the restriction enzyme HindIII (Takara Shuzo) was further added, and the reaction was allowed to proceed a; 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 20 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, and the solution was divided into two 10-μl portions. To one portion, 10 units of the restriction enzyme SacII (Toyobo) was further added and, to the other, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. Both the reaction mixtures were fractionated by agarose gel electrophoresis, whereby about 0.3 μg each of a HindIII-SacII fragment (about 2.96 kb) and a KpnI-HindIII fragment (about 2.96 kb) were recovered.

[0405] Then, 3 μg of the plasmid pSE1UK1SEd1-3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SacII (Toyobo) and 10 units of the restriction enzyme KpnI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 50 mM sodium chloride and 1 mM DTT, 10 units of the restriction enzyme HindIII (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 0.2 μg each of a HindIII-SacII fragment (about 2.42 kb) and a KpnI-HindIII fragment (about 1.98 kb) were recovered.

[0406] Then, 0.1 μg of the thus-obtained HindIII-SacII fragment of pSE1UK1SEd1-3 and 0.1 μg of the above HindIII-SacII fragment of pBSA7 were dissolved in a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101 and, as a result, a plasmid, pBSH-S, shown in FIG. 66 was obtained. Furthermore, 0.1 μg of the above-mentioned KpnI-HindIII fragment of pSE1UK1SEd1-3 and 0.1 μg of the above-mentioned KpnI-HindIII fragment of pBSAE were dissolved in a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSK-H shown in FIG. 67 was obtained.

[0407] Then, 3 μg each of the thus-obtained plasmids pBSH-S and pBSK-H were respectively added to 10-μl portions of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added to each mixture, and the reaction was allowed to proceed at 37° C. for 1 hour. Both the reaction mixtures were subjected to ethanol precipitation. With each precipitate, the 3′ cohesive ends resulting from ANTI digestion were rendered blunt using DNA Blunting Kit (Takara Shuzo), followed by ligation using DNA Ligation Kit (Takara Shuzo). The thus-obtained recombinant DNA solution were used to transform Escherichia coli HB101, and the plasmids pBSH-SA and pBSK-HA shown in FIG. 68 were obtained.

[0408] Then, 5 μg each of the thus-obtained plasmids pBSH-SA and pBSK-HA were respectively added to 10-μl portions of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 1 unit of the restriction enzyme EcoRI (Takara Shuzo) was further added to each mixture, and the reaction was allowed to proceed at 37° C. for 10 minutes for partial digestion. Both the reaction mixtures were subjected to ethanol precipitation. With each precipitate, the 5′ cohesive ends resulting from EcoRI digestion were rendered blunt using DNA Blunting Kit (Takara Shuzo), followed by fractionation by agarose gel electrophoresis, whereby about 0.5 μg each of a fragment about 5.38 kb in length and a fragment about 4.94 kb in length were recovered. The thus-recovered fragments (0.1 μg each) were each dissolved in a total of 20 μl of sterilized water and subjected to ligation treatment using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant DNA solutions were respectively used to transform Escherichia coli HB101, and the plasmids pBSH-SAE and pBSK-HAE shown in FIG. 69 were obtained.

[0409] Then, 3 μg each of the thus-obtained plasmids pBSH-SAE and pBSK-HAE were respectively added to 10-μl portions of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride, 100 mM sodium chloride and 1 mM DTT, 10 units of the restriction enzyme EcoRI (Takara Shuzo) was further added to each mixture, and the reaction was allowed to proceed at 37° C. for 1 hour. Both the reaction mixtures were subjected to ethanol precipitation. With each precipitate, the 5′ cohesive ends resulting from EcoRI digestion were rendered blunt using DNA Blunting Kit (Takara Shuzo), followed by ligation using DNA Ligation Kit (Takara Shuzo). The thus-obtained recombinant plasmid DNA solutions were each used to transform Escherichia coli H101, and two plasmids, pBSH-SAEE and pBSK-HAEE, shown in FIG. 70 were obtained. Ten μg each of the thus-obtained plasmids were subjected to sequencing reaction according to the instructions attached to AutoRead Sequencing Kit (Pharmacia Biotech), followed by base sequence determination by electrophoresis on A.L.F. DNA Sequencer (Pharmacia Biotech), whereby it was confirmed that both the ApaI and EcoRI sites had disappeared as a result of the above modification.

[0410] (2) SalI Restriction Site Introduction Downstream from Rabbit β-globin Gene Splicing and Poly A Signals and SV40 Early Gene Poly A Signal

[0411] For making it possible to exchange the antibody H chain and L chain expression promoters of the human CDR-grafted antibody expression vector each for an arbitrary promoter, a SalI restriction site was introduced into the plasmid pSE1UK1SEd1-3 downstream from the rabbit β-globin gene splicing and poly A signals and from the SV40 early gene poly A signal in the following manner.

[0412] Three μg of the plasmid pBSK-HAEE obtained in Paragraph 1 (1) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme NaeI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 20 μl of 50 mM Tris-hydrochloride buffer (pH 9.0) containing 1 mM magnesium chloride, 1 unit of alkaline phosphatase (E. coli C75, Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour for dephosphorylation at the 5′ termini. The reaction mixture was further subjected to phenol-chloroform extraction and then to ethanol precipitation, and the precipitate was dissolved in 20 μl of 10 mM Tris-hydrochloride buffer (pH 8.0) containing 1 mM disodium ethylenediaminetetraacetate (hereinafter briefly referred to as TE buffer). One μl of said reaction solution and 0.1 μg of a phosphorylated SalI linker (Takara Shuzo) were added to sterilized water to make a total volume of 20 μl, followed by ligation treatment using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and a plasmid, pBSK-HAEESal, shown in FIG. 71 was obtained. Ten μg of the plasmid thus obtained was subjected to sequencing reaction according to the instructions attached to AutoRead Sequencing Kit (Pharmacia Biotech), followed by electrophoresis on A.L.F. DNA Sequencer (Pharmacia Biotech) for base sequence determination, whereby it was confirmed that one SalI restriction site had been introduced downstream from the rabbit β-globin gene splicing and poly A signals and from the SV40 early gene poly A signal.

[0413] (3) Modification of ApaI Restriction Site Occurring in Poly A Signal of Herpes Simplex Virus Thymidine Kinase (Hereinafter Referred to as HSVtk) Gene

[0414] The ApaI restriction site occurring in the HSVtk gene poly A signal downstream from the Tn5 kanamycin phosphotransferase gene of the plasmid pSE1UK1Sd1-3 was modified in the following manner.

[0415] Three μg of the plasmid pBSA obtained in Paragraph 1 (1) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SacII (Toyobo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of a SacII-XhoI fragment (about 2.96 kb) was recovered.

[0416] Then, 5 μg of the plasmid pSE1UK1SEd1-3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SacII (Toyobo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was fractionated by agarose gel electrophoresis, whereby about 1 μg of a SacII-XhoI fragment (about 4.25 kb) was recovered.

[0417] Then, 0.1 μg of the above SacII-XhoI fragment of pBSA and the above SacII-XhoI fragment of pSE1UK1SEd1-3 were added to a total of 20 μl of sterilized water, followed by ligation using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSX-S shown in FIG. 72 was obtained.

[0418] Then, 3 μg of the thus-obtained plasmid pBSX-S was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the 3′ cohesive ends resulting from ApaI digestion were rendered blunt using DNA Blunting Kit (Takara Shuzo) and then ligation was carried out using DNA Ligation Kit (Takara Shuzo). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and a plasmid pBSX-SA, shown in FIG. 73 was obtained. Ten μg of the thus-obtained plasmid was subjected to sequencing reaction according to the instructions attached to AutoRead Sequencing Kit (Pharmacia Biotech), followed by electrophoresis on A.L.F. DNA Sequencer (Pharmacia Biotech) for base sequence determination, whereby it was confirmed that the ApaI restriction site in the HSVtk gene poly A signal had disappeared.

[0419] (4) Construction of Human CDR-grafted Antibody L Chain Expression Unit

[0420] A plasmid, pMohCK, containing a human antibody κ L chain constant region cDNA downstream from the promoter/enhancer of the Moloney mouse leukemia virus long terminal repeat and having a human CDR-grafted antibody L chain expression unit allowing cassette-wise insertion thereinto of a human CDR-grafted antibody L chain variable region was constructed in the following manner.

[0421] Three 1 μg of the plasmid pBluescript SK(−) (Stratagene) was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chlorides 1 mM DTT, 10 units of the restriction enzyme SacI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ClaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, and the cohesive ends resulting from SacI and ClaI digestion were rendered blunt using DNA Blunting Kit (Takara Shuzo), followed by fractionation by agarose gel electrophoresis, whereby about 1 μg of a DNA fragment about 2.96 kb in length was recovered. A 0.1-μg portion of the DNA fragment recovered was added to a total of 20 μl of sterilized water and subjected to ligation reaction using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech) The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSSC shown in FIG. 74 was obtained.

[0422] Then, 3 μg of the thus-obtained plasmid pBSSC was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of a KpnI-XhoI fragment (about 2.96 kb) was recovered.

[0423] Then, 5 μg of the plasmid pAGE147 obtained in Paragraph 7 (1) of Example 1 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was fractionated by agarose gel electrophoresis, whereby about 0.3 μg of a KpnI-XhoI fragment (about 0.66 kb) containing the Moloney mouse leukemia virus long terminal repeat promoter/enhancer was recovered.

[0424] Then, 0.1 μg of the KpnI-XhoI fragment of pBSSC and 0.1 μg of the KpnI-XhoI fragment of pAGE147 each obtained as mentioned above were dissolved in a total of 20 μl of sterilized water and subjected to ligation using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSMo shown in FIG. 75 was obtained.

[0425] Then, 3 μg of the above plasmid pBSMo was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme HindIII (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of a =I-HindIII fragment (about 3.62 kb) was recovered.

[0426] Then, synthetic DNAs respectively having the base sequences shown in SEQ ID No:38 and SEQ ID No:39 were synthesized using an automatic DNA synthesizer (Applied Biosystems model 380A). To 15 μl of sterilized water were added 0.3 μg each of the thus-obtained synthetic DNAs, and the mixture was heated at 65° C. for 5 minutes. The reaction mixture was allowed to stand at room temperature for 30 minutes and then 2 μl of 10-fold concentrated buffer [500 mM Tris-hydrochloride (pH 7.6), 100 mM magnesium chloride, 50 mM DTT] and 2 μl of 10 mM ATP were added, 10 units of T4 polynucleotide kinase was further added, and the reaction was allowed to proceed at 37° C. for 30 minutes for phosphorylation of the 5′ termini. To a total of 20 μl of sterilized water were added 0.1 μg of the above KpnI-HindIII fragment (3.66 kb) derived from the plasmid pBSMo and 0.05 μg of the phosphorylated synthetic DNA pair, and ligation was effected using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSMoS shown in FIG. 76 was obtained. Ten 4 g of the plasmid thus obtained was subjected to sequencing reaction according to the instructions attached to AutoRead Sequencing Kit (Pharmacia Biotech), followed by electrophoresis on A.L.F. DNA Sequencer (Pharmacia Biotech) for base sequence determination, whereby it was confirmed that the synthetic DNA pair had been introduced as desired.

[0427] Then, 3 μg of the plasmid pChiIgLA1 described in JP-A-5-304989 was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and EcoRV (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of an EcoRI-EcoRV fragment (about 9.70 kb) was recovered.

[0428] Then, synthetic DNAs respectively having the base sequences shown in SEQ ID NO:40 and SEQ ID NO:41 were synthesized using an automatic DNA synthesizer (Applied Biosystems model 380A). To 15 μl of sterilized water were added 0.3 μg each of the thus-obtained synthetic DNAs, and the mixture was heated at 65° C. for 5 minutes. The reaction mixture was allowed to stand at room temperature for 30 minutes. Then, 2 μl of 10-fold concentrated buffer [500 mM Tris-hydrochloride (pH 7.6), 100 mM magnesium chloride, 50 mM DTT] and 2 μl of 10 mM ATP were added, 10 units of T4 polynucleotide kinase was further added, and the reaction was v owed to proceed at 37° C. for 30 minutes for phosphorylation of the 5′ termini. To a total of 20 μl of sterilized water were added 0.1 μg of the above EcoRI-EcoRV fragment (9.70 kb) derived from the plasmid pChiIgLA1 and 0.05 μg of the phsophorylated synthetic DNA, and ligation was effected using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech) The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pChiIgLA1S shown in FIG. 77 was obtained.

[0429] Then, 3 μg of the plasmid pBSMoS obtained in the above manner was dissolved in 10 μl of 20 mM Tris-hydrochloride buffer (pH 8.5) containing 100 mM potassium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme HpaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme EcoRI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of an HpaI-EcoRI fragment (about 3.66 kb) was recovered.

[0430] Then, 10 μg of the plasmid pChiIgLA1S obtained as mentioned above was dissolved in 10 μl of 20 mM Tris-acetate buffer (pH 7.9) containing 50 mM potassium acetate, 10 mM magnesium acetate, 1 mM DTT and 100 μg/ml BSA, 10 units of the restriction enzyme NlaIV (New England BioLabs) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme EcoRI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 0.3 μg of an NlaIV-EcoRI fragment (about 0.41 kb) was recovered.

[0431] Then, 0.1 μg of the above HpaI-EcoRI fragment of pBSMoS and 0.1 μg of the above NlaIV-EcoRI fragment of pChiIgLA1S were added to a total of 20 μl of sterilized water, and ligation was effected using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pMohCκ shown in FIG. 78 was obtained.

[0432] (5) Construction of Human CDR-grafted Antibody H Chain Expression Unit

[0433] A plasmid, pMohCγ1, containing a human antibody γ1 H chain constant region cDNA downstream from the promoter/enhancer of the Moloney mouse leukemia virus long terminal repeat and having a human CDR-grafted antibody H chain expression unit allowing cassette-wise insertion thereinto of a human CDR-grafted antibody H chain variable region was constructed in the following manner.

[0434] Three μg of the plasmid pBSMo obtained in Paragraph 1 (4) of Example 3 was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 30 mM sodium acetate buffer (pH 5.0) containing 100 mM sodium chloride, 1 mM zinc acetate and 10% glycerol, 10 units of Mung bean nuclease (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 10 minutes. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, the cohesive ends of the precipitate were rendered blunt using DNA Blunting Kit (Takara Shuzo) and ligation was effected using DNA Ligation Kit (Takara Shuzo). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSMoSal shown in FIG. 79 was obtained. A 10-μg portion of the plasmid obtained was subjected to sequencing reaction according to the instructions attached to AutoRead Sequencing Kit (Pharmacia Biotech), followed by electrophoresis on A.L.F. DNA Sequencer (Pharmacia Biotech) for base sequence determination, whereby it was confirmed that the XhoI restriction site upstream of the Moloney mouse leukemia virus long terminal repeat promoter/enhancer had disappeared.

[0435] Then, 3 μg of the plasmid pBSMoSal obtained as mentioned above was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme HindIII (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of a KpnI-HindIII fragment (about 3.66 kb) was recovered.

[0436] Then, synthetic DNAs respectively having the base sequences shown in SEQ ID NO:42 and SEQ ID NO:43 were synthesized using an automatic DNA synthesizer (Applied Biosystems model 380A). To 15 μl of sterilized water were added 0.3 μg each of the thus-obtained synthetic DNAs, and the mixture was heated at 65° C. for 5 minutes. The reaction mixture was allowed to stand at room temperature for 30 minutes. Then, 2 μl of 10-fold concentrated buffer [500 mM Tris-hydrochloride (pH 7.6), 100 mM magnesium chloride, 50 mM DTT] and 2 μl of 10 mM ATP were added, 10 units of T4 polynucleotide kinase was further added, and the reaction was allowed to proceed at 37° C. for 30 minutes for phosphorylation of the 5′ termini. To a total of 20 μl of sterilized water were added 0.1 μg of the above KpnI-HindIII fragment (3.66 kb) derived from the plasmid pBSMoSa1 and 0.05 μg of the phosphorylated synthetic DNA, and ligation was effected using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSMoSalS shown in FIG. 80 was obtained. A 10-μg portion of the thus-obtained plasmid was subjected to sequencing reaction according to the instructions attached to AutoRead Sequencing Kit (Pharmacia Biotech), followed by electrophoresis on A.L.F. DNA Sequencer (Pharmacia Biotech), for base sequence determination whereby it was confirmed that the synthetic DNA had been introduced as desired.

[0437] Then, 10 μg of the plasmid pChiIgHB2 described in JP-A-5-304989 was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme Eco52I (Toyobo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 30 nM sodium acetate buffer (pH 5.0) containing 100 mM sodium chloride, 1 mM zinc acetate and 10% glycerol, 10 units of Mung bean nuclease (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 10 minutes. The reaction mixture was subjected to phenol-chloroform extraction and then to ethanol precipitation, and the cohesive ends were rendered blunt using DNA Blunting Kit (Takara Shuzo). After ethanol precipitation, the precipitate was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 0.7 μg of ApaI-blunt end fragment (about 0.99 kb) was recovered.

[0438] Then, 3 μg of the plasmid pBluescript SK(−) (Stratagene) was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 33 mM Tris-acetate buffer (pH 7.9) containing 10 mM magnesium acetate, 66 mM potassium acetate, 0.5 mM DTT and 100 μg/ml BSA, 10 units of the restriction enzyme SmaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of an ApaI-SmaI fragment (about 3.0 kb) was recovered.

[0439] Then, 0.1 μg of the ApaI-blunt end fragment of pCHiIgHB2 and 0.1 μg of the ApaI-SmaI fragment of pBluescript SK(−), each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBShCγl shown in FIG. 81 was obtained.

[0440] Then, 5 μg of the above plasmid pBShCγ1 was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SpeI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of an ApaI-SpeI fragment (about 1.0 kb) was recovered.

[0441] Then, 3 μg of the plasmid pBSMoSalS obtained as mentioned above was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SpeI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of an ApaI-SpeI fragment (about 3.66 kb) was recovered.

[0442] Then, 0.1 μg of the ApaI-SpeI fragment of pBShCγ1 and 0.1 μg of the ApaI-SpeI fragment of pBSMoSa1S, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pMohCγ1 shown in FIG. 82 was obtained.

[0443] (6) Construction of Tandem Cassette Type Humanized Antibody Expression Vector pKANTEX93

[0444] A tandem cassette type humanized antibody expression vector, pKANTEX93, was constructed using the various plasmids obtained in Paragraphs (1) through (5) of Example 3 in the following manner.

[0445] Three μg of the plasmid pBSH-SAEE obtained in Paragraph 1 (1) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme HindIII (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SalI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of a HindIII-SalI fragment (about 5.42 kb) was recovered.

[0446] Then, 5 μg of the plasmid pBSK-HAEE obtained in Paragraph 1 (1) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (ph 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme HindIII (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 0.8 μg of a KpnI-HindIII fragment (about 1.98 kb) containing the rabbit β-globin gene splicing and poly A signals, the SV40 early gene poly A signal and the SV40 early gene promoter was recovered.

[0447] Then, 5 μg of the plasmid pMohCγ1 obtained in Paragraph 1 (5) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SalI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 0.8 μg of a human CDR-grafted antibody H chain expression unit-containing KpnI-SalI fragment (about 1.66 kb) was recovered.

[0448] Then, 0.1 μg of the HindIII-SalI fragment of pBSH-SAEE, 0.1 μg of the KpnI-HindIII fragment of pBSK-HAEE and 0.1 μg of the KpnI-SalI fragment of pMohCγ1, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated together using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pMoγ1SP shown in FIG. 83 was obtained.

[0449] The, 3 μg of the above plasmid pMoγ1SP was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SalI (Takara Shuzo) and 10 units of the restriction enzyme XhoI were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of a SalI-XhoI fragment (about 9.06 kb) was recovered.

[0450] Then, 5 μg of the plasmid pBSK-HAEESal obtained in Paragraph 1 (2) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SalI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 0.7 μg of a KpnI-SalI fragment (about 1.37 kb) containing the rabbit β-globin gene splicing and poly signals and the SV40 early gene poly A signal was recovered.

[0451] Then, 5 μg of the plasmid pMohCκ obtained in Paragraph 1 (4) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 0.7 μg of a human CDR-grafted antibody L chain expression unit-containing KpnI-XhoI fragment (about 1.06 kb) was recovered.

[0452] Then, 0.1 μg of the SalI-XhoI fragment of pMoγ1SP, 0.1 μg of the KpnI-SalI fragment of pBSK-HAEESal and 0.1 μg of the KpnI-XhoI fragment of pMohCκ, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated together using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech) The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pMoκγ1SP shown in FIG. 84 was obtained.

[0453] Then, 3 μg of the above plasmid pMoκγ1SP was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 1 units of the restriction enzyme SacII (Toyobo) was further added, and the reaction was allowed to proceed at 37° C. for 10 minutes for partial digestion. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.2 μg of a SacII-XhoI fragment (about 8.49 kb) was recovered.

[0454] Then, 3 μg of the plasmid pBSX-SA obtained in Paragraph 1 (4) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SacII (Toyobo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of a SacII-XhoI fragment (about 4.25 kb) was recovered.

[0455] Then, 0.1 μg of the SacII-XhoI fragment of pMoκγ1SP and 0.1 μg of the SacII-XhoI fragment of pBSX-SA, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pKANTEX93 shown in FIG. 85 was obtained.

[0456] 2. Expression of Human Anti-GM2 Chimera Antibody Using Humanized Antibody Expression Vector pKANTEX93

[0457] Human anti-GM2 chimera antibody expression was effected using the humanized antibody expression vector pKANTEX93 mentioned above in Paragraph 1 of Example 3 in the following manner.

[0458] (1) Construction of Plasmid pBSH3 Containing Mouse Anti-GM2 Antibody KM796 H Chain Variable Region cDNA

[0459] Three μg of the plasmid pBluescript SK(−) (Stratagene) was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units each of the restriction enzymes SacII (Toyobo) and KpnI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, and the precipitate was subjected to blunting treatment for rendering blunt the 3′ cohesive ends resulting from the restriction enzyme digestion using DNA Blunting Kit (Takara Shuzo) and then to fractionation by agarose gel electrophoresis, and about 1 μg of a DNA fragment about 2.95 kb in size was recovered.

[0460] Then, synthetic DNAs respectively having the base sequences shown in SEQ ID NO:44 and SEQ ID NO:45 were synthesized using an automatic DNA synthesizer (Applied Biosystems model 380A). To 15 μl of sterilized water were added 0.3 μg each of the synthetic DNAs obtained, and the mixture was heated at 65° C. for 5 minutes. The reaction mixture was allowed to stand at room temperature for 30 minutes and then 2 μl of 10-fold concentrated buffer [500 mM Tris-hydrochloride (pH 7.6), 100 mM magnesium chloride, 50 mM DTT] and 2 μl of 10 mM ATP were added, 10 units of T4 polynucleotide kinase was further added, and the reaction was allowed to proceed at 37° C. for 30 minutes for phosphorylating the 5′ termini. To a total of 20 μl of sterilized water were added 0.1 μg of the DNA fragment (2.95 kb) derived from the plasmid pBluescript SK(−) and 0.05 μg of the phosphorylated synthetic DNA, each obtained as mentioned above, followed by ligation to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSNA shown in FIG. 86 was obtained. Ten μg of the plasmid obtained was subjected to sequencing reaction treatment according to the instructions attached to AutoRead Sequencing Kit (Pharmacia Biotech), followed by electrophoresis on A.L.F. DNA Sequencer (Pharmacia Biotech) for base sequence determination, whereby it was confirmed that the synthetic DNA had been introduced as desired.

[0461] Then, 3 μg of the plasmid pBSNA obtained as mentioned above was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% Triton X-100, 10 units of the restriction enzyme NotI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of a DNA fragment about 2.95 kb in size was recovered.

[0462] Then, 10 μg of the plasmid pChi796HM1 obtained in Paragraph 7 (3) of Example 1 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% Triton X-100, 10 units of the restriction enzyme NotI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.3 μg of a DNA fragment about 0.45 kb in size was recovered.

[0463] Then, 0.1 μg of the Apa-NotI fragment of pBSNA and 0.1 μg of the Apa-NotI fragment of pChi796HM1, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech) The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSH3 shown in FIG. 87 was obtained.

[0464] (2) Construction of Plasmid pBSL3 Containing Mouse Anti-GM2 Antibody KM796 L Chain Variable Region cDNA

[0465] Three μg of the plasmid pBluescript SK(−) (Stratagene) was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme KpnI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, and the precipitate was subjected to blunting treatment for rendering blunt the 3′ cohesive ends resulting from KpnI digestion using DNA Blunting Kit (Takara Shuzo) and then to ethanol precipitation, the precipitate was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme SacI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of a DNA fragment about 2.95 kb in size was recovered.

[0466] Then, synthetic DNAs respectively having the base sequences shown in SEQ ID NO:46 and SEQ ID NO:47 were synthesized using an automatic DNA synthesizer (Applied Biosystems model 380A). To 15 μl of sterilized water were added 0.3 μg each of the synthetic DNAs obtained, and the mixture was heated at 65° C. for 5 minutes. The reaction mixture was allowed to stand at room temperature for 30 minutes. Then, 2 μl of 10-fold concentrated buffer [500 mM Tris-hydrochloride (pH 7.5), 100 mM magnesium chloride, 50 mM DTT] and 2 μl of 10 mM ATP were added, 10 units of T4 polynucleotide kinase was further added, and the reaction was allowed to proceed at 37° C. for 30 minutes for phosphorylating the 5′ termini. The, 0.1 μg of the DNA fragment (2.95 kb) derived from the plasmid pBluescript SK(−) and 0.05 μg of the phosphorylated synthetic DNA, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSES shown in FIG. 88 was obtained. Ten μg of the plasmid obtained was subjected to sequencing reaction treatment according to the instructions attached to AutoRead Sequencing Kit (Pharmacia Biotech) followed by electrophoresis on A.L.F. DNA Sequencer (Pharmacia Biotech) for base sequence determination, whereby it was confirmed that the synthetic DNA had been introduced as desired.

[0467] Then, 3 μg of the plasmid pBSES obtained as mentioned above was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of a DNA fragment about 2.95 kb in size was recovered.

[0468] The, 5 μg of the plasmid pKM796L1 obtained in Paragraph 4 of Example 1 was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and AflIII (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.3 μg of an EcoRI-AflIII fragment about 0.39 kb in size was recovered.

[0469] The, synthetic DNAs respectively having the base sequences shown in SEQ ID NO:48 and SEQ ID NO:49 were synthesized using an automatic DNA synthesizer (Applied Biosystems model 380A) To 15 μl of sterilized water were added 0.3 μg each of the synthetic DNAs obtained, and the mixture was heated at 65° C. for 5 minutes. The reaction mixture was allowed to stand at room temperature for 30 minutes. Then, 2 μl of 10-fold concentrated buffer [500 mM Tris-hydrochloride (pH 7.6), 100 mM magnesium chloride, 50 mM DTT] and 2 μl of 10 mM ATP were added, 10 units of T4 polynucleotide kinase was further added, and the reaction was allowed to proceed at 37° C. for 30 minutes for phosphorylating the 5′ termini.

[0470] Then, 0.1 μg of the pBSES-derived EcoRI-SplI fragment (2.95 kb), 0.1 μg of the pK4796L1-derived EcoRI-AflIII fragment and 0.05 μg of the phosphorylated synthetic DNA, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated together using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSL3 shown in FIG. 89 was obtained. Ten μg of the plasmid obtained was subjected to sequencing reaction treatment according to the instructions attached to AutoRead Sequencing Kit (Pharmacia Biotech), followed by electrophoresis on A.L.F. DNA Sequencer (Pharmacia Biotech) for base sequence determination, whereby it was confirmed that the synthetic DNA had been introduced as desired.

[0471] (3) Construction of Human Anti-GM2 Chimera Antibody Expression Vector pKANTEX796

[0472] An human anti-GM2 chimera antibody expression vector, pKANTEX796, was constructed using the plasmid pKANTEX93 obtained in Paragraph 1 of Example 3 and the plasmids pBSH3 and pBSL3 respectively obtained in Paragraph 2 (1) and (2) of Example 3, in the following manner.

[0473] Three μg of the plasmid pBSH3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme AzaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% Triton X-100, 10 units of the restriction enzyme NotI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.3 μg of an ApaI-NotI fragment about 0.46 kb in size was recovered.

[0474] Then, 3 μg of the plasmid pKANTEX93 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% Triton X-100, 10 units of the restriction enzyme NotI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, whereby about 1 μg of an ApaI-NotI fragment about 12.75 kb in size was recovered.

[0475] Then, 0.1 μg of the pBSH3-derived ApaI-NotI fragment and 0.1 μg of the pKANTEX93-derived ApaI-NotI fragment, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pKANTEX796H shown in FIG. 90 was obtained.

[0476] Then, 3 μg of the plasmid pBSL3 was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.3 μg of an EcoRI-SplI fragment about 0.4 kb in size was recovered.

[0477] Then, 3 μg of the plasmid pKANTEX796H was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of an EcoRI-SplI fragment about 13.20 kb in size was recovered.

[0478] Then, 0.1 μg off the pBSL3-derived EcoRI-SplI fragment and 0.1 μg of the pKANTEX796H-derived EcoRI-SplI fragment, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pKANTEX796 shown in FIG. 91 was obtained.

[0479] (4) Expression of Human Anti-GM2 Chimera Antibody by pKANTEX796

[0480] According to the procedure described in Paragraph 11 of Example 1, pKANTEX796 was introduced into YB2/0 (ATCC CRL 1581) cells and, as a result of selection by means of G418 (0.5 mg/ml) and MTX (200 nM), a cell line capable of producing about 1 to 2 μg/ml of human anti-GM2 chimera antibody was obtained. It was confirmed that efficient and stable humanized antibody expression is possible using expression vector pKANTEX93.

[0481] 3. Transient Humanized Antibody Expression in COS-7 (ATCC CRL 1651) Cells

[0482] For enabling more rapid activity evaluation of various versions of human CDR-grafted anti-GM2 antibody, transient expression of human anti-GM2 chimera antibody expression was caused in COS-7 cells by the Lipofectamine method using pKANTEX796 and a variant thereof in the following manner.

[0483] (1) Construction of Variant of pKANTEX796

[0484] Since transient antibody expression in animal cells is dependent on the copy number of an expression vector introduced, it was supposed that an expression vector smaller in size would show a higher expression efficiency. Therefore, a smaller humanized antiobdy expression vector, pT796, was constructed by deleting a region supposedly having no effect on humanized antibody expression from pKANTEX796 in the following manner.

[0485] Thus, 3 μg of the plasmid pKANTEX796 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme HindIII (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme MluI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, and the 5′ cohesive ends resulting from the restriction enzyme digestion were rendered blunt using DNA Blunting Kit (Takara Shuzo) The reaction mixture was fractionated by agarose gel electrophoresis and about 1 μg of a DNA fragment about 9.60 kb in size was recovered. A 0.1-μg portion of the thus-recovered DNA fragment was added to a total of 20 μl of sterilized water and subjected to ligation treatment using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pT796 shown in FIG. 92 was obtained.

[0486] (2) Transient Expression of Human anti-GM2 Chimera Antibody Using pKANTEX796 and pT796

[0487] A 1×105 cells/ml suspension of COS-7 cells was distributed in 2-ml portions into wells of a 6-well plate (Falcon) and cultured overnight at 37° C. Two μg of pKANTEX796 or pT796 was added to 100 μl of OPTI-MEM medium (Gibco), a solution prepared by adding 10 μl of LIPOFECTAMINE reagent (Gibco) to 100 μl of OPTI-MEM medium (Gibco) was further added, and the reaction was allowed to proceed at room temperature for 40 minutes to cause DNA-liposome complex formation. The COS-7 cells cultured overnight were washed twice with 2 ml of OPTI-MEM medium (Gibco), the complex-containing solution was added, and the cells were cultured at 37° C. for 7 hours. Then, the solution was removed, 2 ml of DMEM medium (Gibco) containing 10% FCS was added to each well, and the cells were cultured at 37° C. After 24 hours, 48 hours, 72 hours, 96 hours and 120 hours of cultivation, the culture supernatant was recovered and, after concentration procedure as necessary, evaluated for human anti-GM2 chimera antibody activity in the culture supernatant by the ELISA method described in Paragraph 11 of Example 1. The results are shown in FIG. 93. As shown in FIG. 93, higher levels of transient human anti-GM2 chimera antibody expression was observed with pT796 as compared with pKANTEX796. For pT796, the level of expression was highest at 72 to 96 hours, the concentration being about 30 ng/ml (in terms of GM2 binding activity). The above results indicate that construction of a pKANTEX93-derived vector having a reduced size and introduction thereof into COS-7 cells make it possible to make activity evaluation of expression vector-derived humanized antibodies in a transient expression system. Furthermore, for close activity comparison of various versions of human CDR-grafted anti-GM2 antibody as mentioned hereinafter, the ELISA method described below under (3) was used to determine antibody concentrations in transient expression culture supernatants.

[0488] (3) Determination by ELISA of Humanized Antibody Concentrations in Transient Expression Culture Supernatants

[0489] A solution prepared by 400-fold dilution of goat anti-human IgG (γ chain) antibody (Igaku Seibutugaku Kenkyusho) with PBS was distributed in 50-μl portions into wells of a 96-well microtiter plate and allowed to stand overnight at 4° C. for binding to the wells. After removing the antiobdy solution, blocking was effected with 100 μl of PBS containing 1% BSA at 37° C. for 1 hour. Fifty μl of a transient expression culture supernatant or purified human anti-GM2 chimera antibody was added thereto and allowed to react at room temperature for 1 hour. Thereafter, the solution was removed, the wells were washed with PBS, and 50 μl of a solution prepared by 500-fold dilution of peroxidase-labeled mouse anti-human κ L chain antibody (Zymet) with PBS was added and allowed to react at room temperature for 1 hour. After washing with PBS, 50 μl of an ABTS substrate solution [prepared by dissolving 550 mg of 2,2′-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt in 0.1 M citrate buffer (pH 4.2) and adding, just before use, 1 μl/ml of hydrogen peroxide] was added for causing color development, and the OD415 was measured.

[0490] 4. Production of Human CDR-grafted Anti-GM2 Antibody Using Humanized Antibody Expression Vector pKANTEX93

[0491] A human CDR-grafted anti-GM2 antibody higher in GM2-binding activity than the human CDR-grafted anti-GM2 antibody described in Example 2 was produced in the following manner.

[0492] (1) Modification of Human CDR-grafted Anti-GM2 Antibody H Chain Variable Region Described in Paragraph 1 (1) of Example 2

[0493] DNAs coding for some versions of the human CDR-grafted anti-GM2 antibody H chain variable region described in Example 2 as derived by replacing several amino acids in regions other than the CDR (framework; hereinafter referred to as FR) with original mouse antibody amino acids were constructed in the following manner. Based on a computer model for the variable region of mouse KM796, those amino acid residues that were expected to contribute to restoration of antigen-binding activity as a result of mutation were selected as the amino acid residues to be mutated.

[0494] First, DNAs respectively having the base sequences of SEQ ID NO:50 and SEQ ID NO:51 were synthesized using an automatic DNA synthesize (Applied Biosystems model 380A).

[0495] Then, a version (version 2) of human CDR-grafted antibody H chain variable region shown in SEQ ID NO:52 and having mutation in positions 78 (threonine in lieu of glutamine), 79 (alanine in lieu of phenylalanine) and 80 (tyrosine in lieu of serine) was constructed in the same manner as in Paragraph 1 (1) of Example 2 using a synthetic DNA of SEQ ID NO:50 in lieu of the synthetic DNA of SEQ ID NO:27.

[0496] Then, another version (version 4) of human CDR-grafted antibody H chain variable region shown in SEQ ID NO:53 and having mutations in positions 27 (tyrosine in lieu of phenylalanine), 30 (threonine in lieu of serine), 40 (serine in lieu of praline) and 41 (histidine in lieu of praline) was constructed in the same manner as in Paragraph 1 (1) of Example 2 using a synthetic DNA of SEQ ID NO:51 in lieu of the synthetic DNA of SEQ ID NO:25.

[0497] (2) Construction of Human CDR-grafted Anti-GM2 Antibody H Chain Variable Region Using Known Common Human Antibody H Chain Variable Region

[0498] According to Kabat et al. (Kabat E. A. et al., “Sequences of Proteins of Immunological Interest”, US Dept. of Health and Human Services, 1991), known human antiobdy H chain variable regions are classifiable into subgroups I to III (HSG I to III) based on the homology of their FR regions, and common sequences have been identified for respective subgroups. Therefore, a human CDR-grafted anti-GM2 antibody H chain variable region was constructed based on those common sequences. First, for selecting common sequences to serve as the base, the homology was examined between the FR of the mouse KM796 H chain variable region and the common sequence FR of the human antibody H chain variable region of each subgroup (Table 3).

TABLE 3
Homology (%) between mouse KM796 H chain variable region
FR and human antibody H chain variable region common
sequence FR
HSG I	HSG II	HSG III
72.1	52.9	58.6

[0499] As a result, it was confirmed that subgroup I shows the greatest similarity. Thus, based on the common sequences of subgroup I, a human CDR-grafted anti-GM2 antibody H chain variable region was constructed by the PCR method in the following manner.

[0500] Synthetic DNAs respectively having the base sequences of SEQ ID NO:54 through SEQ ID NO:59 were synthesized using an automatic DNA synthesizer (Applied Systems model 380A). The DNAs synthesized were added, each to a final concentration of 0.1 μM, to 50 μl of 10 mM Tris-hydrochloride buffer (pH 8.3) containing 50 mM potassium chloride, 1.5 mM magnesium chloride, 0.001% gelatin, 200 μM dNTP, 0.5 μM M13 primer RV (Takara Shuzo), 0.5 μM M13 primer M4 (Takara Shuzo) and 2 units of TaKaRa Taq DNA polymerase, the mixture was covered with 50 μl of mineral oil, a DNA thermal cycler (Perkin Elmer model PJ480) was loaded with the mixture, and 30 PCR cycles (2 minutes at 94° C., 2 minutes at 55° C. and 2 minutes at 72° C. per cycle) were conducted. The reaction mixture was purified using QIAquick PCR Purification Kit (Qiagen) and then made into a solution in 30 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% Triton X-100, 10 units of the restriction enzyme NotI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.2 μg of an ApaI-NotI fragment about 0.44 kb in size was recovered.

[0501] Then, 3 μg of the plasmid pBSH3 obtained in Paragraph 2 (1) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% Triton X-100, 10 units of the restriction enzyme NotI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agrose gel electrophoresis, and about 1 μg of an ApaI-NotI fragment about 2.95 kb in size was recovered. Then, 0.1 μg of the ApaI-NotI fragment of the human CDR-grafted anti-GM2 antibody H chain variable region and 0.1 μg of the ApaI-NotI fragment of pBSH3, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101. Plasmid DNAs were prepared from 10 transformant clones and their base sequences were determined. As a result, a plasmid, pBSH10, shown in FIG. 94 and having the desired base sequence was obtained. The amino acid sequence and base sequence of the human CDR-grafted anti-GM2 antibody H chain variable region contained in pBSH10 are shown in SEQ ID NO:60. In the amino acid sequence of the thus-constructed human CDR-grafted anti-GM2 antibody H chain variable region, arginine in position 67, alanine in position 72, serine in position 84 and arginine in position 98 in the FR as selected based on a computer model for the variable region are replaced with lysine, valine, histidine and threonine, respectively, that are found in the mouse KM796 H chain variable region. This is for the purpose of retaining the antigen-binding capacity of mouse KM796.

[0502] (3) Modification of Human CDR-grafted Anti-GM2 Antibody L Chain Variable Region Described in Paragraph 1 (2) of Example 2

[0503] First, a DNA having the base sequence of SEQ ID NO:61 was synthesized using an automatic DNA synthesizer (Applied Biosystems model 380A), and a human CDR-grafted anti-GM2 antibody L chain variable region cDNA with a 3′ terminus capable of pairing with the restriction enzyme SplI was constructed by following the same reaction procedure as in Paragraph 1 (2) of Example 2 using the synthetic DNA in lieu of the synthetic DNA of SEQ ID NO:35.

[0504] Then, 3 μg of the plasmid pBSL3 obtained in Paragraph 2 (2) of Example 3 was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of an EcoRI-SplI fragment about 2.95 kb in size was recovered.

[0505] Then, 0.1 μg of the EcoRI-SplI fragment of the human CDR-grafted anti-GM2 antiobdy L chain variable region obtained as mentioned above and 0.1 μg of the above EcoRI-SplI fragment of pBSL3 were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSL16 shown in FIG. 95 was obtained.

[0506] Then, DNAs coding for certain versions of the human CDR-grafted anti-GM2 antibody L chain variable region contained in the above plasmid pBSL16 were constructed by replacing a certain number of amino acids in the FR with original mouse antibody amino acids by mutagenesis by means of PCR in the following manner (FIG. 96). Based on a computer model for the variable region of mouse KM796, those amino acid residues that were expected to contribute to restoration of antigen-binding activity as a result of mutation were selected as the amino acid residues to be mutated.

[0507] Antisense and sense DNA primers for introducing mutations were synthesized using an automatic DNA synthesizer (Applied Biosystems model 380A). A first PCR reaction was conducted in the same manner as in Paragraph 4 (2) of Example 3 using a final concentration each of 0.5 μM of M13 primer RV (Takara Shuzo) and the antisense DNA primer and of M13 primer M4 (Takara Shuzo) and the sense DNA primer, with 1 ng of pBSL16 as the template. Each reaction mixture was purified using QIAquick PCR Purification Kit (Qiagen) with elution with 20 μl of 10 mM Tris-hydrochloride (pH 8.0). Using 5 μl of each eluate, a second PCR reaction was conducted in the same manner as in Paragraph 4 (2) of Example 3. The reaction mixture was purified using QIAaquick PCR Purification Kit (Qiagen) and then made into a solution in 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.2 μg of an EcoRI-SplI fragment (about 0.39 kb) of each mutant version of the human CDR-grafted anti-GM2 antiobdy L chain variable region was recovered.

[0508] Then, 0.1 μg of the above EcoRI-SplI fragment of each mutant version of the human CDR-grafted anti-GM2 antibody L chain variable region and 0.1 μg of the EcoRI-SplI fragment of pBSL3 were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and a plasmid DNA was prepared from a transformant clone, and the base sequence of said plasmid was determined. In this way, plasmids respectively containing a base sequence having a desired mutation or mutations were obtained.

[0509] Thus, a plasmid, pBSLV1, containing version 1, shown in SEQ ID NO:64, of the human CDR-grafted anti-GM2 antibody L chain variable region was obtained following the above procedure using the synthetic DNA of SEQ ID NO:62 as the mutant antisense primer and the synthetic DNA of SEQ ID NO:63 as the mutant sense primer. In the amino acid sequence of the version 1 human CDR-grafted anti-GM2 antibody L chain variable region, the amino acid valine in position 15 in the FR is replaced with proline that is found in the mouse KM796 L chain variable region. This is for the purpose of retaining the antigen-binding capacity of mouse KM796.

[0510] A plasmid, pBSLV2, containing version 2, shown in SEQ ID NO:67, of the human CDR-grafted anti-GM2 antibody L chain variable region was obtained following the above procedure using the synthetic DNA of SEQ ID NO:65 as the mutant antisense primer and the synthetic DNA of SEQ ID NO:66 as the mutant sense primer. In the amino acid sequence of the version 2 human CDR-grafted anti-GM2 antibody L chain variable region, the amino acid leucine in positions 46 in the FR is replaced with tryptophan that is found in the mouse KM796 L chain variable region. This is for the purpose of retaining the antigen-binding capacity of mouse KM796.

[0511] A plasmid, pBSLV3, containing version 3, shown in SEQ ID NO:70, of the human CDR-grafted anti-GM2 antibody L chain variable region was obtained following the above procedure using the synthetic DNA of SEQ ID NO:68 as the mutant antisense primer and the synthetic DNA of SEQ ID NO:69 as the mutant sense primer. In the amino acid sequence of the version 3 human CDR-grafted anti-GM2 antibody L chain variable region, praline in position 79 and isoleucine in position 82 in the FR are both replaced with alanine that is found in the mouse KM796 L chain variable region. This is for the purpose of retaining the antigen-binding capacity of mouse KM796.

[0512] Then, a plasmid, pBSLV1+2, containing a human CDR-grafted anti-GM2 antibody L chain variable region having both the version 1 and version 2 mutations was constructed in the following manner.

[0513] Three μg of the plasmid pBSLV1 obtained as mentioned above was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and HindIII (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.2 μg of an EcoRI-HindIII fragment about 0.20 kb in size was recovered.

[0514] Then, 3 μg of the plasmid pBSLV2 obtained as mentioned above was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and HindIII (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of an EcoRI-HindIII fragment about 3.2 kb in size was recovered. Then, 0.1 μg of the EcoRI-HindIII fragment of pBSLV1 and 0.1 μg of the EcoRI-HindIII fragment of pBSLV2, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSLV1+2 shown in FIG. 97 was obtained.

[0515] Then, the PCR reaction procedure mentioned above was followed using 1 ng of the plasmid pBSLV1+2 obtained as mentioned above as the template, a synthetic DNA having the base sequence of SEQ ID NO:71 as the mutant antisense primer and a synthetic DNA having the base sequence of SEQ ID NO:72 as the mutant sense primer, whereby a plasmid, pBSLV4, containing a version-4 human CDR-grafted anti-GM2 antibody L chain variable region set forth in SEQ ID NO:73 was obtained. In the amino acid sequence of the version 4 human CDR-grafted anti-GM2 antibody L chain variable region, valine in position 15, leucine in position 46, aspartic acid in position 69, phenylalanine in position 70 and threonine in position 71 in the FR are replaced with proline, tryptophan, serine, tyrosine and serine, respectively, that are found in the mouse KM796 L chain variable region. This is for the purpose of retaining the antigen-binding capacity of mouse 14796.

[0516] Then, the PCR reaction procedure mentioned above was followed using 1 ng of the plasmid pBSLV1+2 obtained as mentioned above as the template, a synthetic DNA having the base sequence of SEQ ID NO:74 as the mutant antisense primer and a synthetic DNA having the base sequence of SEQ ID NO:75 as the mutant sense primer, whereby a plasmid, pBSLV8, containing a version 8 human CDR-grafted anti-GM2 antibody L chain variable region set forth in SEQ ID NO:76 was obtained. in the amino acid sequence of the version 8 human CDR-grafted anti-GM2 antibody L chain variable region, valine in position 15, leucine in position 46, aspartic acid in position 69, phenylalanine in position 70, threonine in position 71, serine in position 76, leucine in position 77 and glutamine in position 78 in the FR are replaced with praline, tryptophan, serine, tyrosine, serine, arginine, methionine and glutamic acid, respectively, that are found in the mouse KM796 L chain variable region. This is for the purpose of retaining the antigen-binding capacity of mouse KM796

[0517] Then, the PCR reaction procedure mentioned above was followed using 1 ng of the plasmid pBSLV4 obtained as mentioned above as the template, a synthetic DNA having the base sequence of SEQ ID NO:77 as the mutant antisense primer and a synthetic DNA having the base sequence of SEQ ID NO:78 as the mutant sense primer, whereby a plasmid, pBSLm-2, containing a version Lm-2 human CDR-grafted anti-GM2 antibody L chain variable region set forth in SEQ ID NO:79 was obtained. In the amino acid sequence of the version Lm-2 human CDR-grafted anti-GM2 antibody L chain variable region, valine in position 15, tyrosine in position 35, leucine in position 46, aspartic acid in position 69, phenylalanine in position 70 and threonine in position 71 in the FR are replaced with proline, phenylalanine, tryptophan, serine, tyrosine and serine, respectively, that are found in the mouse KM796 L chain variable region. This is for the purpose of retaining the antigen-binding capacity of mouse KM796.

[0518] Then, the PCR reaction procedure mentioned above was followed using 1 ng of the plasmid pBSLV4 obtained as mentioned above as the template, a synthetic DNA having the base sequence of SEQ ID NO:80 as the mutant antisense primer and a synthetic DNA having the base sequence of SEQ ID NO:81 as the mutant sense primer, whereby a plasmid, pBSLm-8, containing a version Lm-8 human CDR-grafted anti-GM2 antiobdy L chain variable region set forth in SEQ ID NO:82 was obtained. In the amino acid sequence of the version Lm-8 human CDR-grafted anti-GM2 antibody L chain variable region, valine in position 15, leucine in posit8ion 46, aspartic acid in position 69, phenylalanine in position 70, threonine in position 71, phenylalanine in position 72 and serine in position 76 in the FR are replaced with proline, tryptophan, serine, tyrosine, serine, leucine and arginine, respectively, that are found in the mouse M796 L chain variable region. This is for the purpose of retaining the antigen-binding capacity of mouse KM796.

[0519] Then, a plasmid, pBSLm-28, containing a human CDR-grafted anti-GM2 antibody L chain variable region having- both the version Lm-2 and version Lm-8 mutations was constructed in the following manner.

[0520] Three μg of the plasmid pBSLm-2 obtained as mentioned above was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme EcoRI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units of the restriction enzyme XbaI (Takara Shuzo) was further added, and the reaction as allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.2 μg of an EcoRI-XbaI fragment about 0.24 kb in size was recovered.

[0521] Then, 3 μg of the plasmid pBSLm-8 obtained as mentioned above was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme EcoRI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 50 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units of the restriction enzyme XbaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of an EcoRI-XbaI fragment about 3.16 kb in size was recovered.

[0522] Then, 0.1 μg of the EcoRI-XbaI fragment of pBSLm-2 and 0.1 μg of the EcoRI-XbaI fragment of pBSLm-8, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pBSLm-28 shown in FIG. 98 was obtained. The version Lm-28 human CDR-grafted anti-GM2 antibody L chain variable region contained in the plasmid pBSLm-28 is shown in SEQ ID NO:83. In the amino acid sequence of the version Lm-28 human CDR-grafted anti-GM2 antibody L chain variable region thus constructed, valine in position 15, tyrosine in position 35, leucine in position 46, aspartic acid in position 69, phenylalanine in position 70, threonine in position 71, phenylalanine in position 72 and serine in position 76 are replaced with proline, phenylalanine, tryptophan, serine, tyrosine, serine, leucine and arginine, respectively, that are found in the mouse KM796 L chain variable region. This is for the intended purpose of retaining the antigen-binding capacity of mouse KM796.

[0523] (4) Construction of Human CDR-grafted Anti-GM2 Antibody L Chain Variable Region Using Known Common Human Antibody L Chain Variable Region

[0524] According to Kabat et al. (Kabat E. A. et al., “Sequences of Proteins of Immunological Interest”, US Dept. of Health and Human Services, 1991), known human antibody L chain variable regions are classifiable into subgroups I to IV based on the homology of their FR regions, and common sequences have been identified for respective subgroups. Therefore, a human CDR-grafted anti-GM2 antibody L chain variable region was constructed based on those common sequences. First, for selecting common sequences to serve as the base, the homology was examined between the FR of the mouse KM796 L chain variable region and the common sequence FR of the human antibody L chain variable region of each subgroup (Table 4).

TABLE 4
Homology (%) between mouse KM796 L chain variable region FR
and human antibody L chain variable region common sequence FR
	HSG I	HSG II	HSG III	HSG IV
	70.0	65.0	68.8	67.5

[0525] As a result, it was confirmed that subgroup I shows the greatest similarity. Thus based on the common sequence of subgroup I, a human CDR-grafted anti-GM2 antibody L chain variable region was constructed by the PCR method in the following manner.

[0526] Synthetic DNAs respectively having the base sequences of SEQ ID NO:84 through SEQ ID NO:89 were synthesized using an automatic DNA synthesizer (Applied Systems model 380A) The DNAs synthesized were added, each to a final concentration of 0.1 μM, to 50 μl of 10 mM Tris-hydrochloride buffer (pH 8.3) containing 50 mM potassium chloride, 1.5 mM magnesium chloride, 0.001% gelatin, 200 μM dNTP, 0.5 μM M13 primer RV (Takara Shuzo), 0.5 μM M13 primer M4 (Takara Shuzo) and 2 units of TaKaRa Taq DNA polymerase. The mixture was covered with 50 μl of mineral oil, a DNA thermal cycler (Perkin Elmer model PJ480) was loaded with the mixture, and 30 PCR cycles (2 minutes at 94° C., 2 minutes at 55° C. and 2 minutes at 72° C. per cycle) were conducted. The reaction mixture was purified using QIAquick PCR Purification Kit (Qiagen) and then made into a solution in 30 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.2 μg of an EcoRI-SplI fragment about 0.39 kb in size was recovered.

[0527] Then, 0.1 μg of the above EcoRI-SplI fragment of the human CDR-grafted anti-GM2 antibody L chain variable region and 0.1 μg of the EcoRI-SplI fragment of pBSL3 were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech) The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101. Plasmid DNAs were prepared from 10 transformant clones and their base sequences were determined. As a result, a plasmid, pBSHSGL, shown in FIG. 99 and having the desired base sequence was obtained. The amino acid sequence and base sequence of the human CDR-grafted anti-GM2 antibody L chain variable region contained in pBSHSGL are shown in SEQ ID NO:90. In the amino acid sequence of the thus-constructed human CDR-grafted anti-GM2 antibody L chain variable region, methionine in position 4, leucine in position 11, valine in position 15, tyrosine in position 35, alanine in position 42, leucine in position 46, aspartic acid in position 69, phenylalanine in position 70, threonine in position 71, leucine in position 77 and valine in position 103 in the FR as selected based on a computer model for the variable region are replaced with leucine, methionine, proline, phenylalanine, serine, tryptophan, serine, tyrosine, serine, methionine and leucine, respectively, that are found in the mouse K/796 L chain variable region. This is for the intended purpose of retaining the antigen-binding capacity of mouse MK796.

[0528] (5) Activity Evaluation of Mutant Versions of Human CDR-grafted Anti-GM2 Antibody in Terms of Transient Expression

[0529] Various mutant version human CDR-grafted anti-GM2 antibodies composed of the human CDR-grafted anti-GM2 antibody H chain and L chain variable regions constructed in Paragraphs (1) through (4) of Example 3 and having varying mutations were evaluated for activity in terms of transient expression in the following manner.

[0530] First, for evaluating the human CDR-grafted anti-GM2 antibody H chain variable regions having varying mutations, expression vectors, pT796HCDRHV2, pT796HCDRHV4 and pT796HCDRH10, were constructed by replacing the mouse H chain variable region of the human anti-GM2 chimera antibody transient expression vector pT796 obtained in Paragraph 3 (1) of Example 3 with the human CDR-grafted anti-GM2 antibody H chain variable regions having varying mutations, in the following manner. For comparison, an expression vector, pT796HCDR was constructed by replacing the mouse H chain variable region of pT796 with the human CDR-grafted anti-GM2 antibody H chain variable region obtained in Paragraph 1 (1) of Example 1.

[0531] Three μg of the plasmid pT796 was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of an EcoRI-SplI fragment about 9.20 kb in size was recovered.

[0532] Then, 3 μg of the plasmid pBSL16 obtained in Paragraph 4 (3) of Example 3 was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.3 μg of an EcoRI-SplI fragment about 0.39 kb in size was recovered.

[0533] Then, 0.1 μg of the EcoRI-SplI fragment of pT796 and 0.1 μg of the EcoRI-SplI fragment of pBSL16, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HE101, and the plasmid pT796LCDR shown in FIG. 100 was obtained.

[0534] Then, 3 μg of the above plasmid pT796LCDR was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% Triton X-100, 10 units of the restriction enzyme NotI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of an ApaI-NotI fragment about 9.11 kb in size was recovered.

[0535] Then, 0.1 μg of the human CDR-grafted anti-GM2 antibody H chain variable region obtained in Paragraph 1 (1) of Example 2 or the mutant version 2 or 4 human CDR-grafted anti-GM2 antibody H chain variable region obtained in Paragraph 4 (1) of Example 3 and 0.1 μg of the ApaI-NotI fragment of pT796LCDR were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). Each recombinant plasmid DNA solution thus obtained was used to transform Escherichia coli HB101. The plasmids pT796HLCDR, pT796HLCDRHV2 and pT796HLCDRHV4 shown in FIG. 101 were obtained.

[0536] Then, 3 μg of the plasmid pBSH10 obtained in Paragraph 4 (2) of Example 3 was added to 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme ApaI (Takara Shuzo) was further added, and the restriction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% Triton X-100, 10 units of the restriction enzyme NotI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.3 μg of an ApaI-NotI fragment about 0.44 kb in size was recovered.

[0537] Then, 0.1 μg of the ApaI-NotI fragment of pBSM10 and 0.1 μg of the ApaI-NotI fragment of pT796LCDR were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). The thus-obtained recombinant plasmid DNA solution was used to transform Escherichia coli HB101, and the plasmid pT796HLCDRH10 shown in FIG. 102 was obtained.

[0538] Then, 3 μg each of the plasmids pT796HLCDR, pT796HLCDRHV2, pT796HLCDRHV4 and pT796HLCDRH10 were respectively added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. Each reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of an EcoRI-SplI fragment about 9.15 kb in size was recovered.

[0539] Then, 5 μg of the plasmid pBSL3 obtained in Paragraph 2 (2) of Example 3 was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 0.4 μg of an EcoRI-SplI fragment about 0.39 kb in size was recovered.

[0540] Then, 0.1 μg of the EcoRI-SplI fragment of each of pT796HLCDR, pT796HLCDRHV2, pT796HLCDRHV4 and pT796HLCDRH10 and 0.1 μg of the EcoRI-SplI fragment of pBSL3 were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go DNA Ligase (Pharmacia Biotech). Each recombinant plasmid DNA solution thus obtained was used to transform Escherichia coli HB101. In this way, the plasmids pT796HCDR, pT796HCDRHV2, pT796HCDRHV4 and pT796HCDRH10 shown in FIG. 103 were obtained.

[0541] Then, 2 μg each of the plasmids pT796HCDR, pT796HCDRHV2, pT796HCDRHV4 and pT796HCDRH10 thus obtained were used for transient human CDR-grafted anti-GM2 antibody expression and for culture supernatant human CDR-grafted anti-GM2 antibody activity evaluation by the procedures described in Paragraph 3 (2) and (3) of Example 3, respectively. After introduction of each plasmid, the culture supernatant was recovered at 72 hours, and the GM2-binding activity and antibody concentration in the culture supernatant were determined by ELISA and the relative activity was calculated with the activity of the positive control chimera antibody taken as 100%. The results are shown in FIG. 104.

[0542] The results revealed that the amino acid mutations alone in mutant versions 2 and 4 have little influence on the restoration of the antigen-binding activity of the human CDR-grafted anti-GM2 antibody but that the use of the pBSH10-derived human CDR-grafted antibody H chain variable region constructed based on the known human antiobdy H chain variable region common sequence, contributes to the restoration of the antigen-binding activity.

[0543] In view of the above results, the human CDR-grafted anti-GM2 antibody H chain variable region constructed based on the known human antibody H chain variable region common sequence as shown in SEQ ID NO:60 was selected as a novel human CDR-grafted anti-GM2 antibody H chain variable region.

[0544] Then, for evaluating the human CDR-grafted anti-GM2 antibody L chain variable regions having varying variations, expression vectors, pT796HLCDRLV1, pT796HLCDRLV2, pT796HLCDRLV3, pT796HLCDRLV4, pT796HLCDRLV8, pT796HLCDRLm-2, PT796HLCDRLm-8, pT796HLCDRLm-28 and pT796HLCDRHSGL, were constructed in the following manner by replacing the mouse L chain variable region of the vector pT796HCDRH10 for transient human CDR-grafted anti-GM2 antibody expression obtained as mentioned above with the human CDR-grafted anti-GM2 antibody L chain variable regions having varying mutations.

[0545] Thus, 3 μg of the plasmid pT796HCDRH10 was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of an EcoRI-SplI fragment about 9.15 kb in size was recovered.

[0546] Then, 3 μg of the plasmid pBSLV1, pBSLV2, pBSLV3, pBSLV4, pBSLV8, pBSLm-2, pBSLm-8, pBSLm-28 or pBSHSGL obtained in Paragraph (3) or (4) of Example 3 was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA, 10 units each of the restriction enzymes EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were further added, and the reaction was allowed to proceed at 37° C. for 1 hour. Each reaction mixture was fractionated by agarose gel electrophoresis, and about 0.3 μg of an EcoRI-SplI fragment about 0.39 kb in size was recovered.

[0547] Then, 0.1 μg of the EcoRI-SplI fragment of the pT796HCDRH10 and 0.1 μg of the EcoRI-SplI fragment of each mutant version human CDR-grafted anti-GM2 antibody L chain variable region were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). Each recombinant plasmid DNA solution thus obtained was used to transform Escherichia coli HB101. In this way, the plasmids pT796HLCDRLV1, pT796HLCDRLV2, pT796HLCDRLV3, pT796HLCDRLV4, pT796HLCDRLV8, pT796HLCDRLm-2, pT796HLCDRLm-8, pT796HLCDRLm-28 and pT796HLCDRHSGL were obtained as shown in FIG. 105.

[0548] Then, 2 μg each of the thus-obtained plasmids pT796HLCDRLV1, pT796HLCDRLV2, pT796HLCDRLV3, pT796HLCDRLV4, pT796HLCDRLV8, pT796HLCDRLm-2, pT796HLCDRLm-8, pT796HLCDRLm-28 and pT796HLCDRHSGL and of the plasmid pT796HLCDR described in Example 2 and capable of expressing human CDR-grafted anti-GM2 antibody were used for transient human CDR-grafted anti-GM2 antibody expression and for culture supernatant human CDR-grafted anti-GM2 antibody activity evaluation by the procedures described in Paragraph 3 (2) and (3) of Example 3, respectively. After introduction of each plasmid, the culture supernatant was recovered at 72 hours, and the GM2-binding activity and antibody concentration in the culture supernatant were determined by ELISA and the relative activity was calculated with the activity of the positive control chimera antibody taken as 100%. The results are shown in FIG. 106.

[0549] The results revealed that the amino acid mutations alone in mutant versions 1, 2, 3, 4 and 8 have little influence on the restoration of the antigen-binding activity of the human CDR-grafted anti-GM2 antibody but that the amino acid mutations in mutant versions Lm-2 and Lm-8 contributes to the restoration of the antigen-binding activity. Furthermore, version Lm-28 having both the amino acid mutations of Lm-2 and Lm-8 showed a high level of antigen-biding activity almost comparable to that of the chimera antibody, revealing that those amino acids mutated in producing Lm-28 were very important from the antigen-binding activity viewpoint.

[0550] In view of the above results, the version Lm-28 human CDR-grafted anti-G-M2 antibody L chain variable region shown in SEQ ID NO:83 was selected as a first novel human CDR-grafted anti-GM2 antibody L chain variable region.

[0551] It was further revealed that the antigen-binding activity can be restored when the pBSHSGL-derived human CDR-grafted anti-GM2 antibody L chain variable region, namely the human CDR-grafted anti-GM2 antibody L chain variable region constructed based on the known human antibody L chain variable region common sequence, is used.

[0552] In view of the above result, the human CDR-grafted anti-GM2 antibody L chain variable region constructed based on the known human antibody L chain variable region common sequence as set forth in SEQ ID NO:90 was selected as a second novel human CDR-grafted anti-GM2 antibody L chain variable region.

[0553] It is to be noted that in those human CDR-grafted anti-GM2 antibody L chain variable regions that showed high binding activity against GM2, certain amino acid residues which cannot be specified by deduction from known human CDR-grafted antibody production examples have been replaced with amino acids found in the mouse L chain variable region. Thus, obviously, it was very important, in human CDR-grafted anti-GM2 antibody production, to identify these amino acid residues.

[0554] Furthermore, the fact that the human CDR-grafted anti-GM2 antibodies having those human CDR-grafted anti-GM2 antibody H chain and L chain variable regions based on the known human antibody variable region common sequences showed high antigen binding activity is proof of the usefulness of the present process in human CDR-grafted antibody production.

[0555] (6) Acquisition of Cell Lines for Stable Production of Human CDR-grafted Anti-GM2 Antibodies

[0556] Based on the results of Paragraph 4 (5) of Example 3, two cell lines, KM8966 and KM8967, capable of stably expressing KM8966, which has the amino acid sequence set forth in SEQ ID NO:60 as the H chain variable region and the amino acid sequence set forth in SEQ ID NO:83 as the L chain variable region, and KM8967, which has the amino acid sequence set forth in SEQ ID NO:60 as the H chain variable region and the amino acid sequence set forth in SEQ ID NO:90 as the L chain variable region, respectively as human CDR-grafted anti-GM2 antibodies having higher antigen-binding activity than the human CDR-grafted anti-GM2 antibody described in Example 2 were obtained in the following manner.

[0557] Three μg each of the plasmids pT796HLCDRLm-28 and pT796HLCDRHSGL obtained in Paragraph 4 (5) of Example 3 were respectively added to 10 μl of 20 mM Tris-hydrochloride buffer (pH 8.5) containing 100 mM potassium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme BamHI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. Each reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. Each reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of a BamHI-XhoI fragment about 4.93 kb in size was recovered.

[0558] The transformants KM8966 and KM8967 have been deposited with National Institute of Bioscience and Human-Technology, Agency of Industrial Science and Technology as of May 23, 1995 under the deposit numbers FERM BP-5105, and FERM BP-5107, respectively.

[0559] Then, 3 μg of the plasmid pKANTEX93 obtained in Paragraph 1 of Example 3 was added to 10 μl of 20 mM Tris-hydrochloride buffer (pH 8.5) containing 100 mM potassium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme BamHI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was subjected to ethanol precipitation, the precipitate was added to 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT, 10 units of the restriction enzyme XhoI (Takara Shuzo) was further added, and the reaction was allowed to proceed at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis, and about 1 μg of a BamHI-XhoI fragment about 8.68 kb in size was recovered.

[0560] Then, 0.1 μg of the BamHI-XhoI fragment of pT796HLCDRLm-28 or pT796HLCDRHSGL and 0.1 μg of the BamHI-XhoI fragment of pKANTEX93, each obtained as mentioned above, were added to a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). Each recombinant plasmid DNA solution thus obtained was used to transform Escherichia coli HE101. In this way, the plasmids pKANTEX796HLCDRLm-28 and pKANTEX796HLCDRHSGL shown in FIG. 107 were obtained Then, 4 μg each of the above plasmids pKANTEX796HLCDRLm-28 and pKANTEX796HLCDRHSGL were respectively used to transform YB2/0 (ATCC CRL 1581) cells according to the procedure described in Paragraph 11 of Example 1 and, after final selection using G418 (0.5 mg/ml) and MTX (200 nM), a transformant cell line, KM8966, capable of producing about 40 μg/ml of KM8966, i.e. the pKANTEX796HLCDRLm-28-derived human CDR-grafted anti-GM2 antibody, and a transformant cell line, KM8967, capable of producing about 30 μg/ml of KM8967, i.e. the pKANTEX796HLCDRHSGL-derived human CDR-grafted anti-GM2 antibody, were obtained.

[0561] (7) Purification of Human CDR-grafted Anti-GM2 Antibodies KM8966 and KM8967

[0562] The transformant cell lines KM8966 and 8967 obtained in Paragraph 4 (6) of Example 3 were respectively suspended in GIT medium (Nippon Pharmaceutical) containing 0.5 mg/ml G418 and 200 nM MTX and, according to the procedure of Paragraph 11 of Example 1, 18 mg of purified human CDR-grafted anti-GM2 antibody KM8966 and 12 mg of purified KM8967 were obtained each from about 0.5 liter of culture fluid. Three μg each of the purified human CDR-grafted anti-GM2 antibodies obtained and the human anti-GM2 chimera antibody KM966 were subjected to electrophoresis by the known method [Laemli, Nature, 227, 680 (1979)] for molecular weight determination. The results are shown in FIG. 108. As shown, under reducing conditions, both antibody H chains showed a molecular weight of about 50 kilodaltons and both antibody L chains showed a molecular weight of about 25 kilodaltons. Expression of H and L chains of correct molecular weights was thus confirmed. Under non-reducing conditions, both human CDR-grafted anti-GM2 antibodies showed a molecular weight of about 150 kilodaltons and it was thus confirmed that antibodies each composed of two H chains and two L chains and having a correct size had been expressed. Furthermore, the H and L chains of each human CDR-grafted anti-GM2 antibody were analyzed for N-terminal amino acid sequence by automatic Edman degradation using a protein sequencer (Applied Biosystems model 470A), whereby an amino acid sequence deducible from the base sequence of the variable region DNA constructed was revealed.

[0563] 5. In vitro Reactivity of Human CDR-grafted Anti-GM2 Antibodies KM8966 and KM8967 against GM2

[0564] The human anti-GM2 chimera antibody K966 and the purified human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 were tested for reactivity against GM2 by ELISA as described in Paragraph 11 of Example 1. The results are shown in FIG. 109. GM2 (N-acetyl-GM4) used was purified from cultured cell line HPB-ALL [Oboshi et al., Tanpakushitsu, Kakusan & Koso (Protein, Nucleic acid & Enzyme), 23, 697 (1978)] in accordance with the known method [J. Biol. Chem., 263, 10915 (1988)]. As shown, it was found that the purified human CDR-grafted anti-GM2 antibody KM8966 exerted the binding activity comparable to that of the human anti-GM2 chimera antibody KM966. On the other hand, the binding activity of purified human CDR-grafted anti-GM2 antibody KM8967 was about ¼ to ⅕ of that of the human anti-GM2 chimera antibody KM966.

[0565] 6. Reaction Specificity of Human CDR-grafted Anti-GM2 Antibodies KM8966 and KM8967

[0566] The human anti-GM2 chimera antibody KM966 and the human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 were tested for reactivity against the gangliosides GM1, N-acetyl-GM2, N-glycolyl-GM2, N-acetyl-GM3, N-glycolyl-GM3, GD1a, GD1b (Iatron), GD2, GD3 (Iatron) and GOx3 (Iatron) by ELISA as described in Paragraph 11 of Example 1. The results are shown in FIG. 110. GM1 and GD1a were purified from bovine brain, N-acetyl-GM2 from cultured cell line HPB-ALL [Oboshi et al., Tanpakushitsu, Kakusan & Koso (Protein, Nucleic acid & Enzyme), 23, 697 (1978)], N-glycolyl-GM2 and N-glycolyl-GM3 from mouse liver, N-acetyl-GM3 canine erythrocytes, and GD2 from cultured cell line IMR32 (ATCC CCL127), respectively by the per se known method [J. Biol. Chem., 263, 10915 (1988)]. Each antibody was used in a concentration of 10 μg/ml.

[0567] As shown in FIG. 110, it was confirmed that the human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 react specifically with GM2 (N-acetyl-GM2 and N-glycolyl-GM2) like the human anti-GM2 chimera antibody KM966.

[0568] 7. Reactivity of Human CDR-grafted Anti-GM2 Antibodies KM8966 and KM8967 Against Cancer Cells

[0569] The human lung small cell carcinoma culture cell line SBC-3 (JCRB 0818) (1×106 cells) was suspended in PBS, the suspension was placed in a microtube (TREF) and centrifuged (1200 rpm, 2 minutes). To the thus-washed cells was added 50 μl (50 μg/ml) of the human anti-GM2 chimera antibody KM966 or the purified human CDR-grafted anti-GM2 antibody KM8966 or KM8967, followed by stirring and 1 hour of standing at 4° C. After the above reaction step, the cells were washed three times with PBS, each time followed by centrifugation. Then, 20 μl of fluorescein isocyanate-labeled protein A (30-fold dilution, Boehringer Mannheim) was added and, after stirring, the reaction was allowed to proceed at 4° C. for 1 hour. Thereafter, the cells were washed three times with PBS, each time followed by centrifugation, then further suspended in PBS and subjected to analysis using a flow cytometer, EPICS Elite (Coulter). In a control run, the above procedure was followed without addition of the human CDR-grafted anti-GM2 antibody and analyzed. The results are shown in FIG. 111. It was found that the purified human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 strongly reacted with the human lung small cell carcinoma culture cell line SBC-3 like the human anti-GM2 chimera antibody KM966.

[0570] 8. In vitro Antitumor Activity of Human CDR-grafted Anti-GM2 Antibodies KM8966 and KM8967: Complement Dependent Cytotoxicity (CDC)

[0571] (1) Preparation of Target Cells

[0572] The target cells SBC-3, cultured in RPMI1640 medium supplemented with 10% FCS, were adjusted to a cell concentration of 5×106 cells/500 μl, 3.7 MBq of Na2 CrO4 (Daiichi Pure Chemicals Co., Ltd.) was added thereto. Then, the reaction was allowed to proceed at 37° C. for 1 hour, and the cells were washed three times with the medium. The cells were then allowed to stand in the medium at 4° C. for 30 minutes and, after centrifugation, the medium was added to adjust the cell concentration to 1×105 cells/ml.

[0573] (2) Preparation of the Complement

[0574] Sera from healthy subjects were combined and used as a complement source.

[0575] (3) CDC Activity Measurement

[0576] The chimeric human anti-GM2 antibody KM966 or purified human CDR-grafted anti-GM2 antibody KM8966 or KM8967 was added to wells of 96-well U-bottom plates within the final concentration range of 0.05 to 50 μg/ml and then 50 μl (5×104 cells/well) of the target cells prepared in (1) were added to each well. The reaction was allowed to proceed at room temperature for 1 hour. After centrifugation, the supernatants were discarded, the human complement obtained in (2) was added to each well to give a final concentration of 15% v/v, and the reaction was allowed to proceed at 37° C. for 1 hour. After centrifugation, the amount of 51Cr in each supernatant was determined using a gamma counter. The amount of spontaneously dissociated 51Cr was determined by adding to the target cells the medium alone in stead of the antibody and complement solutions and measuring the amount of 51Cr in the supernatant in the same manner as mentioned above. The total amount of dissociated 51Cr was determined by adding to the target cells 1 N hydrochloric acid in stead of the antibody and complement solutions and measuring the amount of 51Cr in the supernatant in the same manner as mentioned above. The CDC activity was calculated as follows: 3$\mathrm{CDC}\mathrm{activity}\left(\%\right)=\frac{\begin{array}{c}\mathrm{Amount}{\mathrm{of}}^{51}\mathrm{Cr}\mathrm{in}\\ \mathrm{sample}\mathrm{supernatant}\end{array}·\begin{array}{c}\mathrm{Amount}{\mathrm{of}}^{51}\mathrm{Cr}\\ \mathrm{spontaneously}\mathrm{dissociated}\end{array}}{\begin{array}{c}\mathrm{Total}\mathrm{amount}\mathrm{of}\\ {}^{51}\mathrm{Cr}\mathrm{dissociated}\end{array}·\begin{array}{c}\mathrm{Amount}{\mathrm{of}}^{51}\mathrm{Cr}\\ \mathrm{spontaneously}\mathrm{dissociated}\end{array}}\times 100$

[0577] The results thus obtained are shown in FIG. 112. It was shown that CDC activity of the human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 was lower than that of the chimeric human anti-GM2 antibody KM966.

[0578] 9. In vitro Antitumor Activity of Human CDR-grafted Anti-GM2 Antibodies KM8966 and KM8967: Antibody Dependent Cell Mediated Cytotoxicity (ADCC)

[0579] (1) Preparation of Target Cells

[0580] The target cells SBC-3 cultured in RPMI1640 medium supplemented with 10% FCS were adjusted to a cell concentration of 1×106 cells/500 μl, 3.7 MBq of Na2 CrO4 (Daiichi Pure Chemicals Co., Ltd.) was added thereto. Then, the reaction was allowed to proceed at 37° C. for 1 hour and the cells were washed three times with the medium. The cells were then allowed to stand in the medium at 4° C. for 30 minutes and then, after centrifugation, the medium was added to adjust the cell concentration to 2×105 cells/ml.

[0581] (2) Preparation of Effector Cells

[0582] Human venous blood (50 ml) was collected, 0.5 ml of heparin sodium (Takeda Chemical Industries; 1,000 units/ml) was added, and the mixture was gently stirred. This mixture was overlaid on Polymorphprep (Nycomed) and centrifuged to separate the lymphocyte layer (PBMC). The resulting lymphocytes were washed three times by centrifugation with RPMI1640 medium supplemented with 10% FCS, and the cells were suspended in the medium (5×106 cells/ml) for use as effector cells.

[0583] (3) ADCC Activity Measurement

[0584] The chimeric human anti-GM2 antibody KM966 or purified human CDR-grafted anti-GM2 antibodies KM8966 or KM8967 was added to wells of 96-well U-bottom plates within the final concentration range of 0.05 to 50 μg/ml and then 50 μl (1×104 cells/well) of the target cell suspension prepared in (1) and 100 μl (5×105 cells/well) of the effector cell suspension prepared in (2) were added to each well. The reaction was allowed to proceed at 37° C. for 4 hours and, after centrifugation, the amount of 51Cr in each supernatant was measured using a gamma counter. The amount of spontaneously dissociated 51Cr was determined by adding to the target cells the medium alone in lieu of the antibody and effector cells and measuring the amount of 51Cr in the supernatant in the same manner as mentioned above. The total amount of dissociated 51Cr was determined by adding to the target cells 1 N hydrochloric acid in lieu of the antibody and effector cells and measuring the amount of 51Cr in the supernatant in the same manner as mentioned above. The ADCC activity was calculated as follows: 4$\mathrm{ADCC}\mathrm{activity}\left(\%\right)=\frac{\begin{array}{c}\mathrm{Amount}{\mathrm{of}}^{51}\mathrm{Cr}\mathrm{in}\\ \mathrm{sample}\mathrm{supernatant}\end{array}·\begin{array}{c}\mathrm{Amount}{\mathrm{of}}^{51}\mathrm{Cr}\\ \mathrm{spontaneously}\mathrm{dissociated}\end{array}}{\begin{array}{c}\mathrm{Total}\mathrm{amount}\mathrm{of}\\ {}^{51}\mathrm{Cr}\mathrm{dissociated}\end{array}·\begin{array}{c}\mathrm{Amount}{\mathrm{of}}^{51}\mathrm{Cr}\\ \mathrm{spontaneously}\mathrm{dissociated}\end{array}}\times 100$

[0585] The results thus obtained are shown in FIG. 113. The human CDR-grafted anti-GM2 antibody KM8966 showed ADCC activity comparable to that of the chimeric human anti-GM2 antibody KM966, whereas the human CDR-grafted anti-GM2 antibody KM8967 showed ADCC activity slightly lower than that of the chimeric human anti-GM2 antibody KM966.

EXAMPLE 4

Production of Human CDR-grafted Anti-GM2 Antibodies

[0586] The human CDR-grafted anti-GM2 antibodies KM8966 and KM8967 showed antigen binding activity (ELISA), binding specificity and ADCC activity comparable to those of the chimeric human anti-GM2 antibody KM966, while its CDC activity was lower than that of the chimeric antibody. In order to improve the CDC activity, human CDR-grafted anti-GM2 antibodies were produced in the following manner.

[0587] 1. Modification of Human CDR-grafted Anti-GM2 Antibody KM966 H Chain Variable Region

[0588] Among the human CDR-grafted anti-GM2 antibodies prepared in Example 3, the antibody KM8966 showing higher CDC activity was subjected to amino acid mutation at the H chain variable region (SEQ ID NO:60) in order to improve CDC activity. The amino acids to be mutated were selected at random with reference to the results of various mutation obtained in Example 3 and a computer model for the variable region of mouse KM796. Mutations were introduced by PCR method using as a template 1 ng of the plasmid pBSH10 containing the human CDR-grafted anti-GM2 antibody H chain variable region obtained in Paragraph 4 (2) of Example 3 and using as a primer antisense and sense synthetic DNA containing mutations described in Paragraph 4 (3) of Example 3.

[0589] The reaction was carried out in the same manner as described in Paragraph 4 (3) of Example 3 using the synthetic DNA of SEQ ID NO:91 as the mutant antisense primer and the synthetic DNA of SEQ ID NO:92 as the mutant sense primer to obtain the plasmid pBSHM1 containing version HM1, shown in SEQ ID NO:93, of the human CDR-grafted anti-GM2 antibody H chain variable region. In the amino acid sequence of the version HM1, arginine in position 38, alanine in position 40, glutamine in position 43 and glycine in position 44 in the FR shown in SEQ ID NO:60 were replaced with lysine, serine, lysine and serine, respectively, that are found in the mouse KM796 H chain variable region. The plasmid pBSHM2 containing version HM2, shown in SEQ ID NO:96, of the human CDR-grafted anti-GM2 antibody H chain variable region was obtained following the reaction described in Paragraph 4 (3) of Example 3 using the synthetic DNA of SEQ ID NO:94 as the mutant antisense primer and the synthetic DNA of SEQ ID NO:95 as the mutant sense primer. In the amino acid sequence of the version HM2, arginine in position 38 and alanine in position 40 in the FR shown in SEQ ID NO:60 were replaced with lysine and serine, respectively, that are found in the mouse M796 H chain variable region.

[0590] The plasmid pBSHM3 containing version HM3, shown in SEQ ID NO:99, of the human CDR-grafted anti-GM2 antibody H chain variable region was obtained following the reaction described in Paragraph 4 (3) of Example 3 using the synthetic DNA of SEQ ID NO:97 as the mutant antisense primer and the synthetic DNA of SEQ ID NO:98 as the mutant sense primer. In the amino acid sequence of the version HM3, valine in position 68 and isoleucine in position 70 in the FR shown in SEQ ID NO:60 were replaced with alanine and leucine, respectively, that are found in the mouse KM796 H chain variable region.

[0591] The plasmid pBSHM31 containing version HM31, shown in SEQ ID NO:100, of the human CDR-grafted anti-GM2 antibody H chain variable region was obtained following the reaction described in Paragraph 4 (3) of Example 3 using 1 ng of the plasmid pBSHM3 as the template, the synthetic DNA of SEQ ID NO:91 as the mutant antisense primer and the synthetic DNA of SEQ ID NO:92 as the mutant sense primer. In the amino acid sequence of the version HM31, arginine in position 38, alanine in position 40, glutamine in position 43 and glycine in position 44 in the FR of the version HM3 were replaced with lysine, serine, lysine and serine, respectively, that are found in the mouse KM796 H chain variable region.

[0592] Further, the plasmid pBSHM32 containing version HM32, shown in SEQ ID NO:101, of the human CDR-grafted anti-GM2 antibody H chain variable region was obtained following the reaction described in Paragraph 4 (3) of Example 3 using 1 ng of the plasmid pBSHM3 as the template, the synthetic DNA of SEQ ID NO:94 as the mutant antisense primer and the synthetic DNA of SEQ ID NO:95 as the mutant sense primer. In the amino acid sequence of the version HM32, arginine in position 38 and alanine in position 40 in the FR of the version HM3 were replaced with lysine and serine, respectively, that are found in the mouse KM796 H chain variable region.

[0593] 2. Evaluation of CDC Activity of Various Human CDR-grafted Anti-GM2 Antibodies Having Mutations in the Human CDR-grafted Anti-GM2 Antibody H Chain Variable Region

[0594] (1) Construction of Expression Vectors

[0595] Expression vectors for various human CDR-grafted anti-GM2 antibodies containing the H chain variable region of human CDR-grafted anti-GM2 antibodies having various mutations obtained in Paragraph 1 of Example 4 and the L chain variable region of KM8966 (SEQ ID NO:83) were prepared in the following manner.

[0596] 3 μg each of the plasmids pBSHM1, pBSHM2, pBSHM3, pBSHM31 and pBSHM32 obtained in Paragraph 1 of Example 4 were dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of ApaI (Takara Shuzo) were added thereto and the mixture was allowed to react at 37° C. for 1 hour. The resulting mixture was subjected to ethanol precipitation and the thus-obtained precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% of Triton X-100. 10 units of NotI (Takara Shuzo) were further added thereto to allow the mixture to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.2 μg of the ApaI-NotI fragment of about 0.44 kb.

[0597] Then, 3 μg of the plasmid pKANTEX796HLCDRLm-28 obtained in Paragraph 4 (5) of Example 3 was dissolved in 10 μl of 10 mM Tris-hydrochloride buffer (pH 7.5) containing 10 mM magnesium chloride and 1 mM DTT, 10 units of ApaI (Takara Shuzo) were added thereto and the mixture was allowed to react at 37° C. for 1 hour. The resulting mixture was subjected to ethanol precipitation and the thus-obtained precipitate was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT, 100 μg/ml BSA and 0.01% of Triton X-100. 10 units of NotI (Takara Shuzo) were added thereto to allow the mixture to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 1 μg of the ApaI-NotI fragment of about 13.14 kb.

[0598] About 0.1 μg each of the thus-obtained ApaI-NotI fragment of pBSHM1, pBSHM2, pBSHM3, pBSHM31 and pBSHM32 and 0.1 μg of the ApaI-NotI fragment of pKANTEX796HLCDRLm-28 were added in a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). Each of the resulting recombinant plasmid DNA solutions was used to transform Escherichia coli HB101 and plasmids, pKANTEX796HM1Lm-28, pKANTEX796HM2Lm-28, pKANTEX796HM3Lm-28, pKANTEX796HM31Lm-28 and pKANTEX796HM32Lm-28 shown in FIG. 114 were obtained.

[0599] (2) Expression of Mutant Versions of Human CDR-grafted Anti-GM2 Antibodies

[0600] 4 μg each of the plasmids pKANTEX796HM1Lm-28, pKANTEX796HM2Lm-28, pKANTEX796HM3Lm-28, pKANTEX796HM31Lm-28 and pKANTEX796HM32Lm-28 obtained in Paragraph 2 (1) of Example 4 were used to transform YB2/0 cells (ATCC CRL 1581) in accordance with the method as described in Paragraph 11 of Example 1. The cells were ultimately selected using G418 (0.5 mg/ml) and MTX (200 nM) to obtain about 2 to 5 μg/ml of transformants capable of producing human CDR-grafted anti-GM2 antibodies derived from the corresponding expression vectors.

[0601] (3) Purification of Mutant Versions of Human CDR-grafted Anti-GM2 Antibodies

[0602] Cells of each transformant obtained in Paragraph 2 (2) of Example 4 were suspended in GIT medium (Nihon Pharmaceutical) containing 0.5 mg/ml G418 and 200 nM MTX and about 1 to 3 mg of purified human CDR-grafted anti-GM2 antibodies were obtained from about 0.6 liter of the culture broth in accordance with the method described in Paragraph 11 of Example 1. The human CDR-grafted anti-GM2 antibodies derived from the plasmids pKANTEX796HM1Lm-28, pKANTEX796HM2Lm-28, pKANTEX796HM3Lm-28,. pKANTEX796HM31Lm-28 and pKANTEX796HM32Lm-28 are hereinafter referred to as M1-28, M2-28, M3-28, M31-28 and M32-28, respectively. 4 μg each of the purified human CDR-grafted anti-GM2 antibodies, the human CDR-grafted anti-GM2 antibody KM8966 and the chimeric human anti-GM2 antibody KM966 were electrophoresed by the conventional method [Laemmli: Nature, 227, 680 (1970)] for molecular weight checking. The results are shown in FIG. 115. As shown in FIG. 115, under reducing conditions, the molecular weight of the antibody H chain was about 50 KDa and the molecular weight of the antibody L chain was about 25 KDa, thus confirming the expression of the H chain and L chain having the correct molecular weight. Under nonreducing conditions, the molecular weight of the human CDR-grafted anti-GM2 antibodies was about 150 KDa, confirming that the antibody expressed was composed of two H chains and two L chains and was correct in size. The N-terminal amino acid sequence of the H and L chains of each purified human CDR-grafted anti-GM2 antibodies was examined by automatic Edman degradation using a protein sequencer (Applied Biosystems model 470A). As a result, it was confirmed that the amino acid sequence was consistent with that deduced from the synthesized variable region DNA sequence.

[0603] (4) CDC Activity of Mutant Versions of Human CDR-grafted Anti-GM2 Antibodies

[0604] CDC activity of the mutant versions of the human CDR-grafted anti-GM2 antibodies obtained in Paragraph 2 (3) of Example 4, the human CDR-grafted anti-GM2 antibody KM8966 and the chimeric human anti-GM2 antibody KM966 was measured in accordance with the method described in Paragraph 8 of Example 3. The results are shown in FIG. 116. As shown in FIG. 116, it was found that, among the mutant versions of the human CDR-grafted anti-GM2 antibodies, the human CDR-grafted anti-2GM antibody M2-28 derived from the plasmid pKANTEX796HM2Lm-28 showed the highest CDC activity which was higher than that of the human CDR-grafted anti-GM2 antibody KM8966 prepared in Example 3. This result indicates that the mutated amino acid residues in positions 38 and 40 in the FR of the version HM2 among the various mutant versions prepared in Paragraph 1 of Example 4 play an important role for improving CDC activity. It was assumed from the computer model for the variable region of mouse KM796 that the mutation of the amino acid residues in positions 38 and 40 in the FR of the version HM2 would influence on the entire structure of the variable region since these amino acid residues are located at the site which interacts with the L chain variable region. Recent study of the production of human CDR-grafted antibody reveals that the amino acid residues which affect the structure of the antibody varies in each antibody. No method for precisely predicting such amino acid residues has been established and the above results provide a significant finding for the production of the human CDR-grafted antibody.

[0605] The human CDR-grafted anti-GM2 antibody M2-28 derived from the plasmid pKANTEX796HM2Lm-28 was designated as KM8970 and the antibody KM8970-producing trasformant KM8970 has been deposited with National Institute of Bioscience and Human-Technology, Agency of Industrial Science and Technology as of under the deposit number FERM BP-.

[0606] 3. Modification of Human CDR-grafted Anti-GM2 Antibody KM8966 L Chain Variable Region

[0607] The human CDR-grafted anti-GM2 antibody KM8966 prepared in Example 3 was subjected to amino acid mutation in the L chain variable region (SEQ ID NO:83) to improve CDC activity. As an amino acid to be mutated, serine residue in position 59 was selected based on the results of various mutations obtained in Example 3 which suggested that it was important to support the structure of CDR2 for the human CDR-grafted anti-GM2 antibody activity. Mutations were introduced by PCR method using as a template 1 ng of the plasmid pBSLm-28 containing the human CDR-grafted anti-GM2 antibody L chain variable region obtained in Paragraph 4 (3) of Example 3 and using as a primer antisense and sense synthetic DNA containing mutations described in Paragraph 4 (3) of Example 3.

[0608] The reaction was carried out in the same manner as described in Paragraph 4 (3) of Example 3 using the synthetic DNA of SEQ ID NO:102 as the mutant antisense primer and the synthetic DNA of SEQ ID NO:103 as the mutant sense primer to obtain the plasmid pBSLm-28 No.1, containing version Lm-23 No.1, shown in SEQ ID NO:104, of the human CDR-grafted anti-GM2 antibody L chain variable region. In the amino acid sequence of the version Lm-28 No.1, serine in position 59 in the FR shown in SEQ ID NO:83 was replaced with alanine that is found in the mouse KM796 L chain variable region.

[0609] 4. Evaluation of CDC Activity of Human CDR-grafted Anti-GM2 Antibody Having New Mutation in Human CDR-grafted Anti-GM2 Antibody L Chain Variable Region

[0610] (1) Construction of Expression Vectors

[0611] Expression vectors for the human CDR-grafted anti-GM2 antibody containing the human CDR-grafted anti-GM2 antibody L chain variable region having the mutation obtained in Paragraph 3 of Example 4 and the human CDR-grafted anti-GM2 antibody H chain variable region were obtained in the following manner.

[0612] 6 μg of the plasmid pBSLm-28 No.1 obtained in Paragraph 3 of Example 4 was dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride, 1 mM DTT and 100 μg/ml BSA. 10 units each of EcoRI (Takara Shuzo) and SplI (Takara Shuzo) were added thereto to allow the mixture to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 0.4 μg of the EcoRI-SplI fragment of about 0.39 kb.

[0613] Then, 3 μg each of the plasmid pKANTEX796HLCDRLm-28 obtained in Paragraph 4 (5) of Example 3 and the plasmids pKANTEX796HM1Lm-28, pKANTEX796HM2Lm-28 and pKANTEX796HM3Lm-28 obtained in Paragraph 2 (1) of Example 4 were dissolved in 10 μl of 50 mM Tris-hydrochloride buffer (pH 7.5) containing 100 mM sodium chloride, 10 mM magnesium chloride and 1 mM DTT and 100 μg/ml BSA, 10 units each of EcoRI (Takara Shuzo) and SplI were added thereto and the mixture was allowed to react at 37° C. for 1 hour. The reaction mixture was fractionated by agarose gel electrophoresis to recover about 1 μg of the EcoRI-SplI fragment of about 13.19 kb.

[0614] A 0.1 μg portion each of the thus-obtained EcoRI-SplI fragment of pBSLm-28 No.1 and 0.1 μg of the-EcoRI-SplI of pKANTEX796HLCDRLm-28, pKANTEX796HM1Lm-28, pKANTEX796HM2Lm-28 and pKANTEX796HM3Lm-28 were added in a total of 20 μl of sterilized water and ligated to each other using Ready-To-Go T4 DNA Ligase (Pharmacia Biotech). Each of the resulting recombinant plasmid DNA solutions was used to transform Escherichia coli HB101 and the plasmids pKANTEX796HLm-28 No.1, pKANTEX796HM1Lm-28 No.1, pKANTEX796HM2Lm-28 No.1 and pKANTEX796HM3Lm-28 No.1 shown in FIG. 117 were obtained.

[0615] (2) Expression of Human CDR-grafted Anti-GM2 Antibodies Having Mutations in the L Chain Variable Region

[0616] 4 μg each of the plasmids pKANTEX796HLm-28 No.1, pKANTEX796HM1Lm-28 No.1, pKANTEX796HM2Lm-28 No.1 and pKANTEX796HM3Lm-28 No.1 obtained in Paragraph 4 (1) of Example 4 was used to transform YB2/0 cells (ATCC CRL 1581) in accordance with the method as described in Paragraph 11 of Example 1. The cells were ultimately selected using G418 (0.5 mg/ml) and MTX (200 nM) to obtain about 2 to 5 μg/ml of transformants capable of producing human CDR-grafted anti-GM2 antibodies derived from the corresponding expression vectors.

[0617] (3) Purification of Human CDR-grafted Anti-GM2 Antibodies Having Mutation in the L Chain Variable Region

[0618] Cells of each transformant obtained in Paragraph 4 (2) of Example 4 were suspended in GIT medium (Nihon Pharmaceutical) containing 0.5 mg/ml G418 and 200 nM MTX and about 1 to 3 mg of purified human CDR-grafted anti-GM2 antibodies were obtained from about 0.5 liter of the culture broth in accordance with the method described in Paragraph 11 of Example 1. The human CDR-grafted anti-GM2 antibodies derived from the plasmids pKANTEX796HLm-28 No.1, pKANTEX796HM1Lm-28 No.1, pKANTEX796HM2Lm-28 No.1 and pKANTEX796HM3Lm-28 No.1 are hereinafter referred to as h796H-No.1, M1-No.1, M2-No.1 and M3-No.1, respectively. 4 μg each of the purified human CDR-grafted anti-GM2 antibodies and the chimeric human anti-GM2 antibody KM965 was electrophoresed by the conventional method [Laemmli: Nature, 227, 680 (1970)] for molecular weight checking. The results are shown in FIG. 118. As shown in FIG. 118, under reducing conditions, the molecular weight of the antibody H chain was about 50 KDa and the molecular weight of the antibody L chain was about 25 KDa, thus confirming the expression of the H chain and L chain having the correct molecular weight. Under nonreducing conditions, the molecular weight of the human CDR-grafted anti-GM2 antibodies was about 150 KDa, confirming that the antibody expressed was composed of two H chains and two L chains and was correct in size. The N-terminal amino acid sequence of the H and L chains of each purified human CDR-grafted anti-GM2 antibodies was examined by automatic Edman degradation using a protein sequencer (Applied Biosystems model 470A). As a result, it was confirmed that the amino acid sequence was consistent with that deduced from the synthesized variable region DNA sequence.

[0619] (4) CDC Activity of Human CDR-grafted Anti-GM2 Antibodies Having Mutation in the L Chain Variable Region

[0620] CDC activity of the human CDR-grafted anti-GM2 antibodies having mutation in the L chain variable region obtained in Paragraph 4 (3) of Example 4, the human CDR-grafted anti-GM2 antibody KM8970 obtained in Paragraph 2 of Example 4, the human CDR-grafted anti-GM2 antibody KM8966 and the chimeric human anti-GM2 antibody KM966 was measured in accordance with the method described in Paragraph 8 of Example 3. The results are shown in FIG. 119. Comparing CDC activity of KM8966 with that of h796H-No.1, it was found that h796H-No.1 in which mutation was introduced into only the L chain variable region showed improved CDC activity. Among the mutant antibodies having mutations in both of the L chain variable region and the H chain variable region, M2-No.1 having mutation in the human CDR-grafted anti-GM2 antibody KM8970 H and L chain variable region showed the highest CDC activity, which was comparable to or higher than that of KM8970. These results indicates that the mutated amino acid residue in position 59 in the FR of the L chain variable region prepared in Paragraph 3 of Example 4 played an important role for improving its CDC activity and it interacted with the mutated amino acid residue in the H chain variable region of KM8970 for improving its CDC activity. It was not assumed from the computer model for the variable region of mouse KM796 that the mutation of the amino acid residue in position 59 in the FR of the version Lm-28 No.1 would be involved in direct action with antigen GM2 and interaction with each CDR residue. However, the above results suggested that the amino acid residue at position 59 in the FR was quite important for maintaining the entire structure of the variable region. There has been no report on the production of a humanized antibody which shows improved biological activity by modifying the above-described amino acid residue in the L chain variable region. Further, since it is impossible to predict in theory the. amino acid residue to be mutated using, for example, the computer model, the above findings will provide an important indication for the production of human CDR-grafted antibody.

[0621] The human CDR-grafted anti-GM2 antibody M2-No.1 derived from the plasmid pKANTEX796HM2Lm-28 No.1 was designated as KM8969 and the antibody KM8969-producing trasformant KM8969 has been deposited with National Institute of Bioscience and Human-Technology, Agency of Industrial Science and Technology as of under the deposit number FERM BP-.

[0622] 5. vitro Reactivity of Human CDR-grafted Anti-GM2 Antibodies KM8969 and KM8970 with GM2

[0623] Reactivities of the chimeric human anti-GM2 antibody KM966 and the human CDR-grafted anti-GM2 antibodies KM8969 and KM8970 with GM2 were measured in accordance with the method described in Paragraph 5 of Example 3. The results are shown in FIG. 120. As shown in FIG. 120, the human CDR-grafted anti-GM2 antibodies KM8969 and DM8970 showed binding activity comparable to that of the chimeric human anti-GM2 antibody KM966.

[0624] 6. Reaction Specificity of Human CDR-grafted Anti-GM2 Antibodies KM8969 and KM8970

[0625] The chimeric human anti-GM2 antibody KM966 and the human CDR-grafted anti-GM, antibodies KM8969 and KM8970 were examined for reactivity with various gangliosides in accordance with the method described in Paragraph 6 of Example 3. The results are shown in FIG. 121. As shown in FIG. 121, it was found that the human CDR-grafted anti-GM2 antibodies KM8969 and KM8970 specifically reacted with GM2 (N-acetyl G2M and N-glycolyl GM2) like the chimeric human anti-GM2 antibody KM966.

[0626] 7. Reactivity of Human CDR-grafted Anti-GM2 Antibodies KM8969 and KM8970 with Cancer Cells

[0627] The chimeric human anti-GM2 antibody KM966 and the human CDR-grafted anti-GM2 antibodies KM8969 and KM8970 were examined for reactivity with the human lung small cell carcinoma cell line SBC-3 (JCRB 0818) using fluorescein isocyanate-labeled rabbit anti-human IgG antibody (Dako) as a second antibody in accordance with the method described in Paragraph 7 of Example 3. The results are shown in FIG. 122. As shown in FIG. 122, the human CDR-grafted anti-GM2 antibodies KM8969 and KM8970 strongly reacted with the human lung small cell carcinoma cell line SBC-3 like the chimeric human anti-GM2 antibody KM966.

[0628] 8. In vitro Antitumor Effect of Human CDR-grafted Anti-GM2 Antibodies KM8969 and KM8970: Antibody Dependent Cell Mediated Cytotoxicity (ADCC)

[0629] The chimeric human anti-GM2 antibody KM966 and the human CDR-grafted anti-GM2 antibodies KM8966, KM8969 and KM8970 were examined for ADCC activity against the human lung small cell carcinoma cell line SBC-3 (JCRB 0818) in accordance with the method described in Paragraph 9 of Example 3. The results are shown in FIG. 123. As shown in FIG. 123, the human CDR-grafted anti-GM2 antibodies KM8969 and KM8970 showed ADCC activity comparable to that of the chimeric human anti-GM2 antibody KM966.

[0630] 9. In vitro Antitumor Effect of Humanized Anti-GM2 Antibodies: Comparison of CDC Activities

[0631] The CDC activities of the humanized anti-GM2 antibodies KM966, KM8966, KM8969 and KM8970 prepared in Examples 3 and 4 were measured with prolonging the reaction period. Namely, after the human complement was added, the reaction was carried out for 4 hours in accordance with the method described in Paragraph 8 of Example 3. The results are shown in FIG. 124. As shown in FIG. 124, the human CDR-grafted anti-GM2 antibodies KM8966, KM8969 and KM8970 showed the CDC activity comparable to that of the chimeric human anti-GM2 antibody KM966 when the antibody concentration was not less that 5 μg/ml. These results suggest that the human anti-GM2 antibodies prepared in Examples are useful for human cancer treatment.

[0632] As described above, the production of humanized anti-GM2 antibodies and the results of their activities as described in Examples show that the thus established humanized anti-GM2 antibodies are useful for human cancer treatment.

[0633] While the invention has been described in detail and with reference to specific examples thereof, it will be apparent to one skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope thereof.

Claims

1. A humanized antibody specific for the ganglioside GM2.

2. The humanized antibody as claimed in claim 1, wherein said antibody belongs to the class IgG.

3. The humanized antibody as claimed in claim 1, wherein said antibody is a chimeric human antibody.

4. The humanized antibody as claimed in claim 3, wherein said antibody comprises the H chain variable region and L chain variable region of a nonhuman animal antibody and the H chain constant region and L chain constant region of a human antibody.

5. The humanized antibody as claimed in claim 1, wherein said H chain variable region has an amino acid sequence corresponding to amino acids 1-120 of SEQ ID NO:1 and wherein said L chain variable region has an amino acid sequence corresponding to amino acids 1-107 of SEQ ID NO:2.

6. The humanized antibody as claimed in claim 1, wherein said antibody is a CDR-grafted antibody.

7. The humanized antibody as claimed in claim 6, wherein said antibody results from replacement of the complementarity determining regions (CDRs) of the H chain variable region and L chain variable region of a human antibody with the CDRs of the H chain variable region and L chain variable region of a nonhuman animal antibody.

8. The humanized antibody as claimed in claim 7, wherein the CDRs of the H chain variable region have amino acid sequences as set forth in SEQ ID NO:6, SEQ ID NO:7 and SEQ ID NO:8 and wherein the CDRs of the L chain variable region have amino acid sequences as set forth in SEQ ID NO:9, SEQ ID NO:10 and SEQ ID NO:11.

9. The humanized antibody as claimed in claim 5, wherein said antibody is KM966.

10. A transformant that produces the antibody KM966 of claim 9.

11. The transformant as claimed in claim 10, wherein said transformant is KM966 (FERM BP-3931).

12. A pharmaceutical composition comprising the humanized antibody of claim 1 and a pharmaceutically acceptable carrier.

13. A method of producing a cytocidal effect comprising contacting a cell containing GM2 with the antibody according to claim 1 under conditions such that said antibody binds to said GM2 and said cytocidal effect is thereby produced.

14. The method according to claim 13 wherein said cell is a human cell.

15. The method according to claim 14 wherein said cell is a cancer cell.

16. The method according to claim 15 wherein said cell is a neural ectodermal cell.

17. The method according to claim 15 wherein said cell is a lung carcinoma cell.

18. A DNA sequence encoding the humanized antibody according to claim 1.

19. An expression vector comprising the DNA sequence according to claim 18 operably linked to a promoter.

20. A cell comprising the expression vector according to claim 19.

21. The cell according to claim 20 wherein said cell is an animal cell.

22. A method of making a humanized antibody specific for the ganglioside GM2 comprising culturing the cell according to claim 20 under conditions such that said DNA sequence is expressed and said humanized antibody thereby produced.

23. The humanized antibody as claimed in claim 5, wherein said antibody is KM8966.

24. The humanized antibody as claimed in claim 5, wherein said antibody is KM8967.

25. A transformant that produced the antibody KMS966 of claim 23.

26. A transformant that produces the antibody KM8967 of claim 24.

27. The transformant as claimed in claim 25, wherein said transformant is KM8960 (FERM BP-5105).

28. The transformant as claimed in claim 26, wherein said transformant is KM8967 (FERM BP-5106).

29. The humanized antibody as claimed in claim 5, wherein said antibody is KM8969.

30. The humanized antibody as claimed in claim 5, wherein said antibody is DM8970.

31. A transformant that produced the antibody KM8969 of claim 29.

32. A transformant that produced the antibody KM8970 of claim 30.

33. The transformant as claimed in claim 31, wherein said transformant is KM8969 (FERM BP-5527).

34. The transformant as claimed in claim 32, wherein said transformant is KM8970 (FERM BP-5528).