Bispecific antibodies that bind to CD3

Abstract

The purpose of the present invention is to provide a bispecific antibody that binds to CD3, or a fragment of the bispecific antibody. The present invention pertains to: a bispecific antibody that binds to CD3, or a fragment of the bispecific antibody; DNA encoding the bispecific antibody or the bispecific antibody fragment; a vector including the DNA; a hybridoma and a transformant strain producing the bispecific antibody or the bispecific antibody fragment; a method for producing the bispecific antibody or the bispecific antibody fragment; a therapeutic and diagnostic agent containing the bispecific antibody or the bispecific antibody fragment; a therapeutic and diagnostic method using the bispecific antibody or the bispecific antibody fragment; and a detection or assay reagent containing the bispecific antibody or the bispecific antibody fragment.

Description

[Technical Field]

[0001] The present invention relates to bispecific antibodies that bind to CD3 or bispecific antibody fragments thereof, DNA encoding the bispecific antibodies or bispecific antibody fragments, vectors containing the DNA, hybridomas and transformants that produce the bispecific antibodies or bispecific antibody fragments, methods for producing the bispecific antibodies or bispecific antibody fragments, therapeutic and / or diagnostic agents comprising the bispecific antibodies or bispecific antibody fragments, therapeutic and / or diagnostic methods using the bispecific antibodies or bispecific antibody fragments, and detection or measurement reagents comprising the bispecific antibodies or bispecific antibody fragments. [Background technology]

[0002] Antibody drugs approved to date for cancer treatment are known to have a variety of mechanisms of action (Non-Patent Document 1). Representative examples include neutralizing activity, which inhibits the binding of ligands such as growth factors to their receptors; agonistic activity, which activates the bound receptor; and effector functions possessed by IgG antibody molecules, such as antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC). Biomarker analysis of clinical trials of rituximab and trastuzumab suggests that ADCC is an important mechanism for the clinical use of antibody drugs (Non-Patent Documents 2 and 3).

[0003] ADCC activity is a mechanism of cytotoxicity that occurs when natural killer cells (NK cells) and other cells recognize the Fc of IgG antibodies bound to membrane antigens on the surface of cancer cells via FcγRIIIA (CD16a), a type of Fc receptor (Non-patent Document 1).

[0004] The IgG1 subclass in humans and the IgG2a subclass in mice are capable of strongly binding to Fc receptors and inducing ADCC activity. Therefore, ADCC activity can be enhanced by artificially modifying the Fc of human IgG1-type antibody drugs to increase their binding ability to FcγRIIIA.

[0005] It is known that ADCC activity can be enhanced by modifying the amino acids in Fc or the sugar chains of the N-linked complex sugar chains that bind to Fc (Non-Patent Document 4). In particular, the technique of enhancing ADCC activity by sugar chain modification has been applied to approved antibody drugs such as mogamulizumab (Non-Patent Document 5) and obinutuzumab (Non-Patent Document 6).

[0006] Unlike natural antibodies, bispecific antibodies are artificially engineered antibody molecules that are capable of binding to two different antigens, and many molecular forms have been reported (Non-Patent Document 7). One example of their application to medicine is a technology that binds to an antigen on a cancer cell and CD3 on the surface of a T cell, cross-linking the two and thereby damaging the cancer cell.

[0007] The T cell-mediated cytotoxicity that occurs when CD3 / cancer antigen bispecific antibodies bind to CD3 on T cells is referred to here as ADTC activity (Antibody-dependent T-cell-mediated cytotoxicity).

[0008] Examples of CD3 / cancer antigen bispecific antibodies include catumaxomab (Non-Patent Document 8), an IgG bispecific antibody against CD3 and the cancer antigen EpCAM, bispecific T cell engager [BiTE (registered trademark)] (Non-Patent Document 9), and RG7802, an IgG bispecific antibody against CD3 and the cancer antigen CEA.

[0009] Many other molecular forms exist (Non-Patent Document 7), but these anti-CD3 bispecific antibodies generally have only ADTC activity and either no ADCC activity or a suppressed ADCC activity.

[0010] For example, BiTE® is a compound consisting of two different antibody fragments called single chain Fvs (scFvs) that bind to CD3 and a cancer antigen, linked via a peptide linker, and does not contain an Fc domain, and therefore does not have an Fc-mediated mechanism of action such as ADCC activity [Figure 1(C)]. RG7802 also has an Fc domain, but the P329G mutation in the Fc domain eliminates its ability to bind to Fc receptors, and it also does not exhibit ADCC activity (Non-Patent Document 10).

[0011] An exception is catumaxomab, an anti-CD3 bispecific antibody that has a hybrid mouse IgG2a / rat IgG2b Fc that binds to Fc receptors, thereby inducing ADCC activity (Non-Patent Document 8). However, rodent-derived Fc has a lower ability to induce ADCC activity than human-derived Fc, which is used in conventional antibody drugs (Non-Patent Document 11).

[0012] The reason why no studies have been conducted on combining anti-CD3 bispecific antibodies with Fc that has strong Fc receptor binding ability is thought to be because of concerns about the side effects that may result from T cell activation caused by CD3 clustering (aggregation on the cell membrane) via the Fc receptor.

[0013] For example, the anti-CD3 antibody OKT3, which is used to suppress rejection during organ transplantation, causes the side effect of cytokine release syndrome, which is thought to be caused by clustering of CD3 by leukocytes with Fc receptors via the mouse IgG2a Fc of OKT3 (Non-patent Document 12).

[0014] Catumaxomab is a bispecific antibody against CD3 and EpCAM that has the ability to bind to Fc receptors. The serious side effects associated with cytokine release syndrome that occur after administration of catumaxomab are thought to be caused by activation of T cells via Kupffer cells in the liver that have Fc receptors (Non-patent Document 13).

[0015] Furthermore, as a format for bispecific antibodies having Fc, a heterodimer structure that has a monovalent CD3-binding site is generally used to avoid concerns about side effects due to nonspecific activation of CD3 (Patent Documents 1 to 4).

[0016] Anti-CD3 bispecific antibodies have also been produced using anti-CD3 antibodies with low affinity for CD3. D =1.0×10 -7 RG7802, which is a mAb-specific antibody, is currently under clinical development (Non-Patent Documents 14 and 15). RG7802 has an Fc region that does not have the ability to bind to Fc receptors, and therefore does not have effector activity such as ADCC activity (Non-Patent Document 15). [Prior art documents] [Patent documents]

[0017] [Patent Document 1] International Publication No. 2011 / 028952 [Patent Document 2] International Publication No. 2014 / 151910 [Patent Document 3] International Publication No. 2015 / 048272 [Patent Document 4] International Publication No. 2014 / 054804 [Non-patent literature]

[0018] [Non-Patent Document 1] Carter P. Nat Rev Cancer. 1: 118-129, 2001 [Non-licensed document 2] Cartron G, Dacheux L, Salles G, et al. Blood. 99: 754-758, 2002 [Non-licensed document 3] Weng WK, Levy R. J Clin Oncol. 21: 3940-3947, 2003

Non-licensed Document 4

Non-licensed Document 5

Non-licensed Document 6

Non-licensed Document 7

Non-licensed literature 9

Non-licensed literature 10

Non-licensed Document 11

Non-licensed Document 12

[0019] Various anti-CD3 bispecific antibodies that bind to CD3 and specific antigens have been known to date, but they have not achieved sufficient efficacy, and there remains the issue of serious side effects due to excessive immune responses such as cytokine release syndrome.

[0020] Furthermore, heterodimeric CD3 / cancer antigen bispecific antibodies that exhibit ADCC activity or high ADCC activity, and their cytotoxic activity, have not been thoroughly investigated, and no anti-CD3 bispecific antibodies are known that have high cytotoxic activity and suppress the induction of excessive cytokine production.

[0021] On the other hand, although it is known that antibodies containing an Fc region with enhanced affinity for Fc receptors by amino acid residue-modified Fc or glycosylated Fc contribute to the high ADCC activity of monoclonal antibodies, there is no motivation to enhance the ADCC activity of anti-CD3 bispecific antibodies, which still have the issue of serious side effects, by using an Fc region with enhanced affinity for Fc receptors by amino acid residue-modified Fc or glycosylated Fc, thereby enhancing the immune response induced by the bispecific antibody.

[0022] An object of the present invention is to provide an anti-CD3 bispecific antibody that inhibits the induction of cytokine production and a bispecific antibody that specifically binds to CD3 and a disease-associated antigen. The present invention also aims to provide a bispecific antibody comprising an antigen-binding domain that binds to CD3 and an antigen-binding domain that binds to a disease-related antigen, a bispecific antibody fragment, DNA encoding the bispecific antibody or the bispecific antibody fragment, a vector comprising the DNA, hybridomas and transformants that produce the bispecific antibody or the bispecific antibody fragment, a method for producing the bispecific antibody or the bispecific antibody fragment, therapeutic and diagnostic agents comprising the bispecific antibody or the bispecific antibody fragment, therapeutic and diagnostic methods using the bispecific antibody or the bispecific antibody fragment, and a detection or measurement reagent comprising the bispecific antibody or the bispecific antibody fragment, a CD3-binding domain with reduced affinity for CD3 for producing an anti-CD3 bispecific antibody, and a CD3-binding domain with reduced affinity for CD3 for suppressing cytokine production induction by an anti-CD3 bispecific antibody. [Means for solving the problem]

[0023] The present invention relates to the following: 1. An anti-CD3 bispecific antibody or a bispecific antibody fragment thereof, comprising an Fc region capable of binding to an Fc receptor, a single CD3-binding domain with reduced affinity for CD3 bound to the C-terminus of the Fc region, and further comprising a disease-related antigen-binding domain. 2. An anti-CD3 bispecific antibody or a bispecific antibody fragment thereof, which comprises an Fc region capable of binding to an Fc receptor, has one CD3-binding domain bound to the C-terminus of the Fc region, and further comprises a disease-related antigen-binding domain, and which has a reduced ability to induce cytokine production in the presence of CD3-positive T cells and disease-related antigen-positive cells, compared to an anti-CD3 bispecific antibody that uses the anti-CD3 monoclonal antibody SP34. 3. The dissociation constant (K D ) is 6 x 10 -8 3. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to 1 or 2 above. 4. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to any one of 1 to 3 above, wherein the dissociation constant of the CD3-binding domain with CD3 is greater than that of the comparative anti-CD3 monoclonal antibody SP34 or KM14. 5. An anti-CD3 bispecific antibody or a bispecific antibody fragment thereof described in any one of 1 to 4 above, wherein the amino acid sequence of the CD3-binding domain has 90% or more homology with the amino acid sequence of the CD3-binding domain of the comparative anti-CD3 monoclonal antibody SP34 or KM14, and the affinity for CD3 is reduced by 10% or more compared to the anti-CD3 monoclonal antibody SP34 or KM14. 6. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of 1 to 5 above, which comprises one or two of the disease-related antigen-binding domains. 7. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of 1 to 6 above, wherein the CD3-binding domain and the disease-related antigen-binding domain are each any one selected from scFv, Fab, and VHH. 8. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of 1 to 7 above, wherein the CD3-binding domain and / or the disease-related antigen-binding domain is linked to the Fc region via a linker. 9. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof described in any one of 1 to 8 above, wherein the CD3-binding domain comprises a heavy chain variable region (abbreviated as VH) comprising complementarity determining regions (abbreviated as CDRs) 1 to 3 of the antibody heavy chain, and a light chain variable region (abbreviated as VL) comprising CDRs 1 to 3 of the antibody light chain. 10. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of 1 to 9 above, wherein the CD3-binding domain is an scFv. 11. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to any one of 1 to 10 above, wherein the amino acid sequences of CDR1 to 3 (HCDR1 to 3) of VH and CDR1 to 3 (LCDR1 to 3) of VL of the CD3-binding domain have 90% or more homology to the amino acid sequences of HCDR1 to 3 and LCDR1 to 3 of any one selected from the following (a) to (h): (a) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 118 to 120, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 121 to 123, respectively. (b) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 83 to 85, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 86 to 88, respectively. (c) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 132, 96, and 97, respectively. (d) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 133, 96, and 97, respectively. (e) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 134, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (f) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 135, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (g) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 136, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (h) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 137, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. 12. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof described in any one of 1 to 11 above, wherein the amino acid sequences of VH and VL of the CD3-binding domain have 80% or more homology to the amino acid sequences of VH and VL of any one selected from the following (aa) to (rr): (aa) VH comprising the amino acid sequence represented by SEQ ID NO: 124, and VL comprising the amino acid sequence represented by SEQ ID NO: 125 (bb) VH comprising the amino acid sequence represented by SEQ ID NO: 115, and VL comprising the amino acid sequence represented by SEQ ID NO: 116 (cc) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 126 (dd) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 127 (ee) VH comprising the amino acid sequence represented by SEQ ID NO: 128, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ff) VH comprising the amino acid sequence represented by SEQ ID NO: 129, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (gg) VH comprising the amino acid sequence represented by SEQ ID NO: 130, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (hh) VH comprising the amino acid sequence represented by SEQ ID NO: 131, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ii) VH comprising the amino acid sequence represented by SEQ ID NO: 159, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (jj) VH comprising the amino acid sequence represented by SEQ ID NO: 160, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (kk) VH comprising the amino acid sequence represented by SEQ ID NO: 161, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (ll) VH comprising the amino acid sequence represented by SEQ ID NO: 162, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (mm) VH comprising the amino acid sequence represented by SEQ ID NO: 168, and VL comprising the amino acid sequence represented by SEQ ID NO: 180 (nn) VH comprising the amino acid sequence represented by SEQ ID NO: 169, and VL comprising the amino acid sequence represented by SEQ ID NO: 181 (oo) VH comprising the amino acid sequence represented by SEQ ID NO: 170, and VL comprising the amino acid sequence represented by SEQ ID NO: 182 (pp) VH comprising the amino acid sequence represented by SEQ ID NO: 171, and VL comprising the amino acid sequence represented by SEQ ID NO: 183 (qq) VH comprising the amino acid sequence represented by SEQ ID NO: 172, and VL comprising the amino acid sequence represented by SEQ ID NO: 184 (rr) VH comprising the amino acid sequence represented by SEQ ID NO: 173, and VL comprising the amino acid sequence represented by SEQ ID NO: 185 13. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of 1 to 12 above, wherein the amino acid sequences of HCDR1 to 3 and LCDR1 to 3 of the CD3-binding domain are any one selected from the following (a) to (h): (a) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 118 to 120, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 121 to 123, respectively. (b) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 83 to 85, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 86 to 88, respectively. (c) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 132, 96, and 97, respectively. (d) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 133, 96, and 97, respectively. (e) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 134, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (f) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 135, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (g) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 136, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (h) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 137, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. 14. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of 1 to 13 above, wherein the amino acid sequences of VH and VL of the CD3-binding domain are any one selected from the following (aa) to (rr): (aa) VH comprising the amino acid sequence represented by SEQ ID NO: 124, and VL comprising the amino acid sequence represented by SEQ ID NO: 125 (bb) VH comprising the amino acid sequence represented by SEQ ID NO: 115, and VL comprising the amino acid sequence represented by SEQ ID NO: 116 (cc) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 126 (dd) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 127 (ee) VH comprising the amino acid sequence represented by SEQ ID NO: 128, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ff) VH comprising the amino acid sequence represented by SEQ ID NO: 129, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (gg) VH comprising the amino acid sequence represented by SEQ ID NO: 130, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (hh) VH comprising the amino acid sequence represented by SEQ ID NO: 131, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ii) VH comprising the amino acid sequence represented by SEQ ID NO: 159, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (jj) VH comprising the amino acid sequence represented by SEQ ID NO: 160, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (kk) VH comprising the amino acid sequence represented by SEQ ID NO: 161, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (ll) VH comprising the amino acid sequence represented by SEQ ID NO: 162, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (mm) VH comprising the amino acid sequence represented by SEQ ID NO: 168, and VL comprising the amino acid sequence represented by SEQ ID NO: 180 (nn) VH comprising the amino acid sequence represented by SEQ ID NO: 169, and VL comprising the amino acid sequence represented by SEQ ID NO: 181 (oo) VH comprising the amino acid sequence represented by SEQ ID NO: 170, and VL comprising the amino acid sequence represented by SEQ ID NO: 182 (pp) VH comprising the amino acid sequence represented by SEQ ID NO: 171, and VL comprising the amino acid sequence represented by SEQ ID NO: 183 (qq) VH comprising the amino acid sequence represented by SEQ ID NO: 172, and VL comprising the amino acid sequence represented by SEQ ID NO: 184 (rr) VH comprising the amino acid sequence represented by SEQ ID NO: 173, and VL comprising the amino acid sequence represented by SEQ ID NO: 185 15. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of 1 to 14 above, wherein the Fc region has enhanced binding activity to an Fc receptor. 16. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to 15 above, wherein the Fc region with enhanced binding activity to an Fc receptor is an Fc region comprising an amino acid residue modification and / or a sugar chain modification. 17. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof described in 15 or 16 above, wherein the Fc region with enhanced binding activity to an Fc receptor is an Fc region that comprises both amino acid residue modifications and glycosylation modifications. 18. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to 16 or 17 above, wherein the amino acid residue modifications include at least one amino acid residue modification that enhances the binding activity to an Fc receptor. 19. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof described in 16 or 17, wherein the sugar chain modification is a sugar chain modification in which fucose that α1,6-links to N-acetylglucosamine at the reducing end of the N-linked sugar chain that binds to Asn at EU numbering position 297 in the Fc region is deleted. 20. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to any one of 1 to 19 above, wherein the disease-related antigen-binding domains comprise two Fabs, and the C-terminus of the heavy chain (VH-CH1) of each Fab is linked to the Fc region directly or via a linker. 21. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to any one of 1 to 19 above, wherein the disease-related antigen-binding domain comprises one Fab, and the C-terminus of the heavy chain (VH-CH1) and light chain (VL-CL) of the Fab are linked to the Fc region directly or via a linker. 22. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to any one of 1 to 19 or 21 above, which comprises one Fab as the disease-related antigen-binding domain, and the CD3-binding domain is bound directly or via a linker to the Fc chain that binds to the heavy chain (VH-CH1) of the Fab. 23. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to any one of 1 to 19 or 21 above, which comprises one Fab as the disease-related antigen-binding domain, and the CD3-binding domain is bound directly or via a linker to the Fc chain that binds to the light chain (VL-CL) of the Fab. 24. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of 20 to 23 above, wherein the linker is a hinge or a modified version thereof. 25. A DNA encoding the anti-CD3 bispecific antibody or the bispecific antibody fragment according to any one of 1 to 24 above. 26. A recombinant vector containing the DNA according to 25 above. 27. A transformant obtained by introducing the recombinant vector according to 26 above into a host cell. 28. A method for producing an anti-CD3 bispecific antibody or a bispecific antibody fragment described in any one of 1 to 24 above, comprising culturing the transformant described in 27 above in a medium, producing and accumulating the anti-CD3 bispecific antibody or the bispecific antibody fragment described in any one of 1 to 24 above in the culture, and collecting the anti-CD3 bispecific antibody or the bispecific antibody fragment from the culture. 29. A therapeutic and / or diagnostic agent for a disease associated with at least one of CD3 and the disease-related antigen, comprising as an active ingredient the anti-CD3 bispecific antibody or bispecific antibody fragment according to any one of 1 to 24 above. 30. The therapeutic and / or diagnostic agent according to 29 above, wherein the disease associated with at least one of CD3 and the disease-associated antigen is cancer. 31. A method for treating and / or diagnosing a disease associated with at least one of CD3 and the disease-associated antigen, using the anti-CD3 bispecific antibody or the bispecific antibody fragment according to any one of 1 to 24 above. 32. The therapeutic and / or diagnostic method according to 31, wherein the disease associated with at least one of CD3 and the disease-associated antigen is cancer. 33. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to any one of 1 to 24 above, for use in the treatment and / or diagnosis of a disease in which at least one of CD3 and the disease-associated antigen is associated. 34. The anti-CD3 bispecific antibody or the bispecific antibody fragment according to 33, wherein the disease associated with at least one of CD3 and the disease-associated antigen is cancer. 35. Use of the anti-CD3 bispecific antibody or the bispecific antibody fragment according to any one of 1 to 24 above for the manufacture of a therapeutic and / or diagnostic agent for a disease involving at least one of CD3 and the disease-associated antigen. 36. The use according to 35, wherein the disease associated with at least one of CD3 and the disease-associated antigen is cancer. 37. A reagent for detecting or measuring at least one of CD3 and a disease-related antigen, comprising the anti-CD3 bispecific antibody or the bispecific antibody fragment according to any one of 1 to 24 above. 38. A method for producing an anti-CD3 bispecific antibody or a bispecific antibody fragment thereof, comprising an Fc region capable of binding to an Fc receptor and a disease-related antigen-binding domain, characterized by attaching a CD3-binding domain with reduced affinity for CD3 to the Fc region. 39. A method for suppressing cytokine production induction by an anti-CD3 bispecific antibody or a bispecific antibody fragment thereof, comprising an Fc region capable of binding to an Fc receptor and a disease-related antigen-binding domain, characterized by using a CD3-binding domain with reduced affinity for CD3. 40. The dissociation constant (K D ) is 6 x 10 -8 40. The method according to claim 38 or 39, wherein 41. The method according to any one of 38 to 40 above, wherein the dissociation constant of the CD3-binding domain with CD3 is greater than that of the comparative anti-CD3 monoclonal antibody SP34 or KM14. 42. The method described in any one of 38 to 41 above, wherein the amino acid sequence of the CD3-binding domain has 90% or more homology with the amino acid sequence of the CD3-binding domain of the control anti-CD3 monoclonal antibody SP34 or KM14, and has a 10% or more reduced affinity compared to the anti-CD3 antibody SP34 or KM14. 43. The method according to any one of 38 to 42 above, wherein the anti-CD3 bispecific antibody or the bispecific antibody fragment comprises one or two disease-related antigen-binding domains. 44. The method according to any one of 38 to 43 above, wherein the CD3-binding domain and the disease-related antigen-binding domain are each any one selected from scFv, Fab, and VHH. 45. The method according to any one of 38 to 44 above, wherein the CD3-binding domain and / or the disease-related antigen-binding domain is linked to an Fc region via a linker. 46. ​​The method according to any one of 38 to 45, wherein the CD3-binding domain comprises a VH comprising CDRs 1 to 3 of an antibody heavy chain and a VL comprising CDRs 1 to 3 of an antibody light chain. 47. The method according to any one of 38 to 46, wherein the CD3-binding domain is an scFv. 48. The method according to any one of 38 to 47 above, wherein the amino acid sequences of CDR1 to 3 (HCDR1 to 3) of VH and CDR1 to 3 (LCDR1 to 3) of VL of the CD3-binding domain have 90% or more homology to the amino acid sequences of HCDR1 to 3 and LCDR1 to 3 of any one selected from the following (a) to (h): (a) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 118 to 120, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 121 to 123, respectively. (b) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 83 to 85, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 86 to 88, respectively. (c) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 132, 96, and 97, respectively. (d) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 133, 96, and 97, respectively. (e) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 134, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (f) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 135, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (g) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 136, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (h) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 137, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. 49. The method described in any one of 38 to 48, wherein the amino acid sequences of VH and VL of the CD3-binding domain have 80% or more homology to the amino acid sequences of VH and VL of any one selected from the following (aa) to (rr): (aa) VH comprising the amino acid sequence represented by SEQ ID NO: 124, and VL comprising the amino acid sequence represented by SEQ ID NO: 125 (bb) VH comprising the amino acid sequence represented by SEQ ID NO: 115, and VL comprising the amino acid sequence represented by SEQ ID NO: 116 (cc) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 126 (dd) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 127 (ee) VH comprising the amino acid sequence represented by SEQ ID NO: 128, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ff) VH comprising the amino acid sequence represented by SEQ ID NO: 129, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (gg) VH comprising the amino acid sequence represented by SEQ ID NO: 130, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (hh) VH comprising the amino acid sequence represented by SEQ ID NO: 131, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ii) VH comprising the amino acid sequence represented by SEQ ID NO: 159, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (jj) VH comprising the amino acid sequence represented by SEQ ID NO: 160, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (kk) VH comprising the amino acid sequence represented by SEQ ID NO: 161, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (ll) VH comprising the amino acid sequence represented by SEQ ID NO: 162, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (mm) VH comprising the amino acid sequence represented by SEQ ID NO: 168, and VL comprising the amino acid sequence represented by SEQ ID NO: 180 (nn) VH comprising the amino acid sequence represented by SEQ ID NO: 169, and VL comprising the amino acid sequence represented by SEQ ID NO: 181 (oo) VH comprising the amino acid sequence represented by SEQ ID NO: 170, and VL comprising the amino acid sequence represented by SEQ ID NO: 182 (pp) VH comprising the amino acid sequence represented by SEQ ID NO: 171, and VL comprising the amino acid sequence represented by SEQ ID NO: 183 (qq) VH comprising the amino acid sequence represented by SEQ ID NO: 172, and VL comprising the amino acid sequence represented by SEQ ID NO: 184 (rr) VH comprising the amino acid sequence represented by SEQ ID NO: 173, and VL comprising the amino acid sequence represented by SEQ ID NO: 185 50. The method according to any one of 38 to 49 above, wherein the amino acid sequences of HCDR1 to 3 and LCDR1 to 3 of the CD3-binding domain are any one selected from the following (a) to (h): (a) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 118 to 120, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 121 to 123, respectively. (b) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 83 to 85, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 86 to 88, respectively. (c) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 132, 96, and 97, respectively. (d) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 133, 96, and 97, respectively. (e) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 134, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (f) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 135, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (g) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 136, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (h) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 137, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. 51. The method according to any one of 38 to 50, wherein the amino acid sequences of VH and VL of the CD3-binding domain are any one selected from the following (aa) to (rr): (aa) VH comprising the amino acid sequence represented by SEQ ID NO: 124, and VL comprising the amino acid sequence represented by SEQ ID NO: 125 (bb) VH comprising the amino acid sequence represented by SEQ ID NO: 115, and VL comprising the amino acid sequence represented by SEQ ID NO: 116 (cc) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 126 (dd) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 127 (ee) VH comprising the amino acid sequence represented by SEQ ID NO: 128, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ff) VH comprising the amino acid sequence represented by SEQ ID NO: 129, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (gg) VH comprising the amino acid sequence represented by SEQ ID NO: 130, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (hh) VH comprising the amino acid sequence represented by SEQ ID NO: 131, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ii) VH comprising the amino acid sequence represented by SEQ ID NO: 159, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (jj) VH comprising the amino acid sequence represented by SEQ ID NO: 160, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (kk) VH comprising the amino acid sequence represented by SEQ ID NO: 161, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (ll) VH comprising the amino acid sequence represented by SEQ ID NO: 162, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (mm) VH comprising the amino acid sequence represented by SEQ ID NO: 168, and VL comprising the amino acid sequence represented by SEQ ID NO: 180 (nn) VH comprising the amino acid sequence represented by SEQ ID NO: 169, and VL comprising the amino acid sequence represented by SEQ ID NO: 181 (oo) VH comprising the amino acid sequence represented by SEQ ID NO: 170, and VL comprising the amino acid sequence represented by SEQ ID NO: 182 (pp) VH comprising the amino acid sequence represented by SEQ ID NO: 171, and VL comprising the amino acid sequence represented by SEQ ID NO: 183 (qq) VH comprising the amino acid sequence represented by SEQ ID NO: 172, and VL comprising the amino acid sequence represented by SEQ ID NO: 184 (rr) VH comprising the amino acid sequence represented by SEQ ID NO: 173, and VL comprising the amino acid sequence represented by SEQ ID NO: 185 52. The method according to any one of 38 to 51 above, wherein the Fc region is an Fc region with enhanced binding activity to an Fc receptor. 53. A CD3-binding domain with reduced affinity for CD3 for producing an anti-CD3 bispecific antibody or bispecific antibody fragment comprising an Fc region and a disease-related antigen-binding domain. 54. A CD3-binding domain with reduced affinity for CD3, for suppressing cytokine production induction by an anti-CD3 bispecific antibody or bispecific antibody fragment comprising an Fc region and a disease-related antigen-binding domain. 55. Use of a CD3-binding domain with reduced affinity for CD3 for producing an anti-CD3 bispecific antibody or bispecific antibody fragment comprising an Fc region and a disease-related antigen-binding domain. [Effects of the Invention]

[0024] The present invention provides bispecific antibodies comprising an antigen-binding domain that binds to CD3 and an antigen-binding domain that binds to a disease-related antigen, bispecific antibody fragments, DNA encoding the bispecific antibodies or the bispecific antibody fragments, vectors comprising the DNA, hybridomas and transformants that produce the bispecific antibodies or the bispecific antibody fragments, methods for producing the bispecific antibodies or the bispecific antibody fragments, therapeutic and diagnostic agents comprising the bispecific antibodies or the bispecific antibody fragments, therapeutic and diagnostic methods using the bispecific antibodies or the bispecific antibody fragments, detection or measurement reagents comprising the bispecific antibodies or the bispecific antibody fragments, as well as CD3-binding domains with reduced affinity for CD3 for producing anti-CD3 bispecific antibodies, and CD3-binding domains with reduced affinity for CD3 for suppressing cytokine release from anti-CD3 bispecific antibodies.

[0025] The bispecific antibodies or specific antibody fragments of the present invention comprise an Fc region capable of binding to an Fc receptor, a single CD3-binding domain with reduced affinity for CD3 bound to the C-terminus of the Fc region, and a disease-related binding domain, thereby possessing disease-related antigen-specific cytotoxic activity through ADCC activity and ADTC activity (antibody-dependent T-cell-mediated cytotoxicity). Therefore, the compositions of the present invention can be used to treat various diseases by targeting antigens expressed by cells associated with those diseases. [Brief explanation of the drawings]

[0026] [Figure 1]Figures 1(A) to 1(C) show schematic diagrams of the molecular forms of antibody drugs. Figure 1(A) shows a general IgG antibody, Figure 1(B) shows the CD3 / EpCAM bispecific antibody Catumaxomab, and Figure 1(C) shows the bispecific T-cell Engager [BiTE®] Blinatumomab. [Figure 2] Figures 2(A) and (B) show schematic diagrams of the molecular forms of an IgG1 monovalent antibody (Figure 2(A)) and an IgG4PE(R409K) monovalent antibody (Figure 2(B)), along with the amino acid alterations contained in each molecular form. In each monovalent antibody, the heavy chain is the first polypeptide, and the light chain-Fc fusion protein is the second polypeptide. [Figure 3] Figures 3(A) and (B) show schematic diagrams of bispecific antibody molecules in which an anti-CD3 scFv is attached to the C-terminus of the Fc region on the first polypeptide (H chain, VH-CH1-hinge-Fc) of the IgG1 monovalent antibody of Figure 3(A) or the IgG4PE(R409K) monovalent antibody of Figure 3(B), and show the sites of amino acid modifications. [Figure 4] Figure 4 shows the results of measuring the binding of IgG antibodies, monovalent antibodies, and bispecific antibody molecules to HER2 on the cell membrane, using the binding inhibition of fluorescently labeled Trastuzumab as an indicator. The horizontal axis shows the concentration of the test antibody molecule, and the vertical axis shows the relative value (%) of the mean fluorescence intensity of the cells. The stronger the binding affinity of the molecule to HER2, the smaller the value on the vertical axis becomes in a concentration-dependent manner when it inhibits the binding of fluorescently labeled Trastuzumab. [Figure 5] Figure 5 shows the results of measuring the binding of IgG, monovalent, and bispecific antibody molecules to CD3 on a membrane using a fluorescently labeled secondary antibody. The vertical axis shows the relative value of the mean fluorescence intensity of the cells (baseline value is 1) on a logarithmic scale. This shows that molecules containing anti-CD3 scFv bind to T cells. [Figure 6]Figure 6 shows the results of measuring the cytotoxic activity against BT-20 cells using a real-time cell analyzer. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates a smaller number of cells adhering to the well, i.e., a higher cytotoxic activity. Compared with an antibody with only ADCC activity and an antibody with only ADTC activity, mixing them in half the amount shows a higher cytotoxic activity. [Figure 7] Figure 7 shows the results of measuring the cytotoxic activity against BT-20 cells using a real-time cell analyzer. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. The lower the cell index value, the higher the cytotoxic activity. This shows that an antibody that has both ADCC and ADTC activities in a single molecule has higher cytotoxic activity than a mixture of half an antibody with only ADCC activity and half an antibody with only ADTC activity. [Figure 8] Figures 8(A) to (E) show the results of measuring cytotoxic activity against MCF-7 cells using a real-time cell analyzer. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. The lower the cell index value, the higher the cytotoxic activity. Under conditions where antibodies with only ADCC activity and antibodies with only ADTC activity show almost no cytotoxic activity, antibodies with both ADCC and ADTC activities in a single molecule show high cytotoxic activity. [Figure 9] Figure 9 shows the results of measuring the binding of CD3 / HER2 bispecific antibody molecules to HER2 on a membrane, using the inhibition of binding by fluorescently labeled Trastuzumab as an indicator. The horizontal axis shows the concentration of the test antibody molecule, and the vertical axis shows the relative value of the mean fluorescence intensity of the cells (reference value is 1). The stronger the binding affinity of a molecule to HER2, the smaller the value on the vertical axis becomes in a concentration-dependent manner due to the inhibition of binding of fluorescently labeled Trastuzumab. This shows that although the test molecules have different linker moieties connecting the C-terminus of the Fc region to the N-terminus of the anti-CD3 scFv, their binding activity to HER2 remains unchanged. [Figure 10] Figure 10 shows the results of measuring the binding of IgG antibodies, monovalent antibodies, and bispecific antibody molecules to CD3 on a membrane using a fluorescently labeled secondary antibody. The vertical axis shows the relative value of the mean fluorescence intensity of the cells (with the reference value set to 1) on a logarithmic scale. The test molecules have different linker moieties connecting the C-terminus of the Fc region and the N-terminus of the anti-CD3 scFv, but the binding activity to CD3 does not change significantly. [Figure 11] Figure 11 shows the results of measuring cytotoxic activity against BT-20 cells using a real-time cell analyzer. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. Although the test molecules have different linker moieties connecting the C-terminus of the Fc region and the N-terminus of the anti-CD3 scFv, they show roughly the same level of cytotoxic activity. [Figure 12] Figure 12 shows the results of measuring cytotoxic activity against BT-20 cells using a real-time cell analyzer. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. Although the test molecules have different linker moieties connecting the C-terminus of the Fc region and the N-terminus of the anti-CD3 scFv, they show roughly the same level of cytotoxic activity. [Figure 13] Figure 13 shows the results of measuring cytotoxic activity against MKN-7 cells using a real-time cell analyzer. The horizontal axis shows the time after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. [Figure 14] Figure 14 shows the results of measuring cytotoxic activity against MKN-45 cells using a real-time cell analyzer. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. [Figure 15]Figure 15 shows the results of measuring the cytotoxic activity of a known anti-CD3 bispecific antibody against BT-20 cells and the cytokine concentrations in the culture supernatants of the control antibody (left column) and 4D5_mvG1_scT3a(DF) (right column), plotted on the same graph. In each graph, the horizontal axis represents the test substance concentration, the left vertical axis represents cytotoxic activity (%), and the right vertical axis represents the cytokine concentration (pg / mL). Cytotoxic activity is represented by black squares (■) and a solid line, and cytokine concentrations are represented by diamonds and dotted lines. [Figure 16] Figures 16(A) to (F) show molecular types designed to verify the synergistic increase in cytotoxic activity due to the combination of ADCC activity and ADTC activity. Figure 16(A) shows a molecular form in which an anti-CD3 scFv is linked to the C-terminus of the Fc region of the first polypeptide (H chain) of a monovalent antibody. Figure 16(B) shows a molecular form in which an anti-CD3 scFv is linked to the C-terminus of one heavy chain of a bivalent antibody. Figure 16(C) shows the same molecular form as Figure 16(A) but with a sugar chain having α1,6 fucose added. Figure 16(D) shows a molecular form in which an anti-CD3 scFv is linked to the C-terminus of the Fc region of the second polypeptide (light chain-Fc fusion protein, or VL-CL-hinge-Fc) of a monovalent antibody. Figure 16(E) shows a molecular form in which the cancer antigen Fab and the anti-CD3 scFv are both located on the N-terminus of the Fc region. Figure 16(F) shows a molecular form in which one CD3 scFv is linked to the C-terminus of each of the first and second polypeptides of a monovalent antibody. [Figure 17] Figures 17(A) and (B) show the results of measuring the cytotoxic activity of CD3 / HER2 bispecific antibodies with bivalent (Figure 17(A)) and monovalent (Figure 17(B)) cancer antigens against BT-20 cells using a real-time cell analyzer to verify molecular type. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. [Figure 18]Figures 18(A) and (B) show the results of measuring the cytotoxic activity of a bispecific CD3 / GM2 antibody with a bivalent cancer antigen on SBC-3 cells using a real-time cell analyzer to verify the molecular type. Figure 18(A) shows a negative control molecule that uses an anti-DNP antibody in the cancer antigen variable region, and Figure 18(B) shows a molecule that uses an anti-GM2 antibody. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. [Figure 19] Figure 19(A) shows the results of measuring the cytotoxic activity of anti-CD3 / HER2 bispecific antibodies and other antibodies containing α1,6 fucose against BT-20 cells using a real-time cell analyzer to verify the molecular type. Figure 19(B) shows the results of comparing the cytotoxic activity of anti-CD3 / HER2 bispecific antibodies and other antibodies containing α1,6 fucose against BT-20 cells with that of anti-CD3 / HER2 bispecific antibodies and other antibodies not containing α1,6 fucose. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. [Figure 20] Figure 20(A) shows the results of measuring the cytotoxic activity of molecules such as those in which anti-CD3 scFv is linked to the C-terminus of the second polypeptide of a monovalent antibody (light chain-Fc fusion protein or VL-CL-hinge-Fc) against BT-20 cells using a real-time cell analyzer to verify the molecular type. Figure 20(B) shows the results of comparing the cytotoxic activity of molecules such as those in which anti-CD3 scFv is linked to the C-terminus of the second polypeptide against BT-20 cells with that of a molecule in which anti-CD3 scFv is linked to the C-terminus of the first polypeptide of a monovalent antibody (VH-CH1-hinge-Fc). The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. [Figure 21]Figure 21 shows the results of measuring the cytotoxic activity of a negative control anti-CD3 bispecific antibody that does not bind to HER2, and a HER2 / CD3 bispecific antibody against BT-20 cells using a real-time cell analyzer to verify the molecular type. The horizontal axis represents the time (h) after the start of measurement, and the vertical axis represents the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. Target+PBMC represents no test substance, and 0.5% Triton X-100 represents 100% cytotoxic activity. Measurements were conducted with the test substance at three final concentrations: 50 nM, 5 nM, and 0.5 nM. * and ** represent values ​​close to each other, so the respective test substances are indicated below. [Figure 22A] Figure 22A shows the results of measuring the cytotoxic activity of an anti-CD3 bispecific antibody that does not bind to HER2 against BT-20 cells using a real-time cell analyzer to verify the molecular type. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. The lower the cell index value, the higher the cytotoxic activity. Target+PBMC indicates no test substance, and 0.5% TritonX-100 indicates 100% cytotoxic activity. Because the variable region 4D5mut does not bind to HER2, the cytotoxic activity observed in this measurement is nonspecific activity. [Figure 22B] Figure 22B shows the results of measuring the cytotoxic activity of an anti-CD3 bispecific antibody that does not bind to HER2 against BT-20 cells using a real-time cell analyzer to verify the molecular type. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. Target+PBMC indicates no test substance, and 0.5% Triton X-100 indicates 100% cytotoxic activity. Because the variable region 4D5mut does not bind to HER2, the cytotoxic activity observed in this measurement is nonspecific activity. [Figure 22C]Figure 22C shows the results of measuring the cytotoxic activity of an anti-CD3 bispecific antibody that does not bind to HER2 against BT-20 cells using a real-time cell analyzer to verify the molecular type. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. Target+PBMC indicates no test substance, and 0.5% Triton X-100 indicates 100% cytotoxic activity. Because the variable region 4D5mut does not bind to HER2, the cytotoxic activity observed in this measurement is nonspecific activity. [Figure 22D] Figure 22D shows a sensorgram of the binding activity of anti-CD3 bispecific antibodies to human CD3D&E proteins measured by surface plasmon resonance (SPR). The names of the test antibodies and their approximate KD values ​​are shown in the figure. [Figure 22E] Figure 22E shows a sensorgram of the binding activity of anti-CD3 bispecific antibodies to human CD3D&E proteins measured by surface plasmon resonance (SPR). The names of the test antibodies and their approximate KD values ​​are shown in the figure. [Figure 22F] Figure 22F shows a sensorgram of the binding activity of anti-CD3 bispecific antibodies to human CD3D&E proteins measured by surface plasmon resonance (SPR). The names of the test antibodies and their approximate KD values ​​are shown in the figure. [Figure 22G] FIG. 22G is an enlarged view of the dotted line portion of FIG. 22F. [Figure 23] Figures 23(A) and (B) show the results of measuring the cytotoxic activity of two types of anti-CD3 bispecific antibodies that do not bind to HER2 against BT-20 cells using a real-time cell analyzer to verify their molecular type. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. Since neither the 4D5mut variable region nor DNP2 binds to HER2, the cytotoxic activity observed in this measurement is nonspecific activity. [Figure 24A] Figure 24A shows the measurement of the cytotoxic activity of anti-CD3 / HER2 or a negative control bispecific antibody using a high-affinity anti-CD3 scFv. The cytotoxic activity against BT-20 cells was measured using a real-time cell analyzer. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. [Figure 24B] Figure 24B shows the measurement of nonspecific cytokine production in anti-CD3 / HER2 or a negative control bispecific antibody using a high-affinity anti-CD3 scFv. Cytokine production was measured when human PBMCs were mixed with the test antibody. The horizontal axis shows the test antibody concentration, and the vertical axis shows the cytokine concentration in the culture supernatant. Legends and abbreviations are shown at the bottom of the figure. [Figure 25] Figures 25(A) and (B) show examples of amino acid modifications designed to be introduced into the CDRs of anti-CD3 monoclonal antibody clone SP34. Figure 25(A) shows a design for introducing a single amino acid modification into the VL CDR or VH CDR of anti-CD3 monoclonal antibody clone SP34. The leftmost column of the table lists the sequential number of the modification, and the second column from the left lists the amino acid residue number of the modified residue according to Kabat numbering, along with the amino acid residues before and after the modified amino acid residue, indicated by single letters. The amino acid residues shown in bold in the amino acid sequence of each CDR in the table indicate the modified amino acid residue. Figure 25(B) shows an example of a combination of two or more CDR modified residues in the CDRs of anti-CD3 monoclonal antibody clone SP34. [Figure 26] Figures 26(A) and (C) show examples of VL and VH FR sequence designs of humanized antibodies using modified CDRs of the anti-CD3 monoclonal antibody clone SP34. Figures 26(B) and (D) show examples of VH and VL FR sequence designs of humanized antibodies using modified CDRs of the anti-CD3 monoclonal antibody clone SP34. [Figure 27]Figure 27 shows the results of measuring the binding activity of a CD3 / HER2 bispecific antibody using a CDR-modified version of the anti-CD3 monoclonal antibody clone SP34 by surface plasmon resonance (SPR). From left to right, the name of the bispecific antibody tested, the mutation site, the number of cycles in the measurement, ka (M-1s-1), kd (s-1), and KD (M) are shown. In the mutation column, "-" indicates that no mutation was introduced, and "Protein not obtained" indicates that the bispecific antibody tested could not be obtained as a purified protein. [Figure 28] Figures 28(A)-(C) show the results of measuring the cytotoxic activity of CD3 / HER2 bispecific antibodies using CDR-modified versions of anti-CD3 monoclonal antibody clone SP34 against BT-20 cells using a real-time cell analyzer. The horizontal axis represents the time (h) after the start of measurement, and the vertical axis represents the cell index value of the well to which the test antibody was added. A lower cell index value indicates higher cytotoxic activity. Target+PBMC represents no test substance, and 0.5% Triton X-100 represents 100% cytotoxic activity. Figures 28(B) and 28(C) are enlarged views of the boxed areas in Figure 28(A). Figure 28(B) shows the activity of 4D5_mvG1_scSP34(H04')(DF), and Figure 28(C) shows the activity of 4D5_mvG1_scSP34(H05')(DF) compared with other antibody groups. [Figure 29] Figures 29(A) to (D) show the results of measuring cytokine production during cytotoxicity of BT-20 cells by a CD3 / HER2 bispecific antibody using a CDR-modified version of the anti-CD3 monoclonal antibody clone SP34. The figures show the results of measuring cytokine concentrations (IL-2, IL-6, IFN-γ, TNF-α) in the culture supernatant by flow cytometry. The vertical axis shows cytokine concentration (pg / mL), and the horizontal axis shows antibody concentration (nM). [Figure 30]Figure 30 shows the results of measuring nonspecific cytokine production (IL-2, IL-6, IFN-γ, and TNF-α) of a CD3 / HER2 bispecific antibody using a CDR-modified version of the anti-CD3 monoclonal antibody clone SP34. "Medium" indicates PBMC and medium alone, while "vehicle" indicates the addition of an equal volume of the buffer (citrate buffer) in which the test substance was dissolved. The vertical axis indicates cytokine concentration (pg / mL), and the horizontal axis indicates the type and concentration (μg / mL) of antibody. The names of the test substances in the figure are abbreviated, and the corresponding names are indicated at the bottom. [Figure 31] Figure 31 shows an example of the amino acid alterations designed to be introduced into the CDRs of the anti-CD3 monoclonal antibody clone KM14. The upper part of the figure shows the alterations introduced into the VL CDR, and the lower part shows the alterations introduced into the VH CDR. The alteration serial numbers are shown on the left of the table, and the amino acid residue numbers of the altered residues according to Kabat numbering are shown on the right. The residues shown in bold are the residues resulting from the alteration. [Figure 32] Figure 32 shows the results of Biacore measurement of the binding activity of a bispecific antibody containing an anti-CD3 sc Fv derived from a modified sequence of the anti-CD3 monoclonal antibody clone KM14 to human CD3 protein [KD(M) Biacore] and the results of measurement of the cytotoxic activity against BT-20 cells using a real-time cell analyzer (Specific cytotoxicity). The number of "+" symbols next to the cytotoxic activity indicates the strength of the activity. [Figure 33] Figures 33(A) and (B) show the results of measurements of the cytotoxic activity of a bispecific antibody, whose anti-CD3 scFv contains a modified sequence of the anti-CD3 monoclonal antibody clone KM14, against BT-20 cells, using a real-time cell analyzer. The horizontal axis shows the time (h) after the start of measurement, and the vertical axis shows the cell index value of the well to which the test antibody was added. A lower cell index value indicates a higher cytotoxic activity. T+P represents wells without the test substance, and 0.5% Triton X-100 represents the line indicating 100% cell damage. Figures 33(A) and 33(B) are from experiments performed under identical conditions, but on different measurement plates. [Figure 34]Figures 34(A) and (B) show the results of flow cytometry analysis of cytokine production during cytotoxicity of BT-20 cells by a bispecific antibody containing an anti-CD3 scFv derived from a modified sequence of the anti-CD3 monoclonal antibody clone KM14. The test substance concentration was 10 nM. Figure 34(A) shows the concentration of INF-γ in the culture supernatant, and Figure 34(B) shows the concentration of IL-6. LLOQ indicates the limit of quantitation specified by the kit. [Figure 35] Figures 35(A) and (B) show the results of measuring the cytotoxic activity of a bispecific antibody containing a modified sequence of clone KM14 as the anti-CD3 scFv against BT-20 cells using a real-time cell analyzer. The cell index value of a well to which no test substance was added was set to 0%, and the cell index value of a well to which 0.5% Triton X-100 was added was set to 100%, and the cytotoxic activity at a specific time was calculated and shown. Figures 35(A) and 35(B) show the cytotoxic activity 48 and 120 hours after the addition of the test substance, respectively. The vertical axis shows cytotoxic activity (%), and the horizontal axis shows the type of each antibody. [Figure 36] 36(A) to (C) are graphs showing the concentration-dependent cytotoxic activity of CD3 / CCR4 bispecific antibodies and the like against PEER cells. The vertical axis shows cytotoxic activity (%), and the horizontal axis shows antibody concentration (nM). [Figure 37] Figures 37(A) and (B) show the concentration-dependent cytokine (IL-2 and IFN-γ) concentrations in the culture supernatants upon cytotoxicity of PEER cells with CD3 / CCR4 bispecific antibodies and other antibodies. The vertical axis shows cytokine concentration (pg / mL), and the horizontal axis shows antibody concentration (nM). [Figure 38] Figures 38(A) and (B) show the concentration-dependent cytotoxic activity of a CD3 / CD123 bispecific antibody against MOLM13 cells (Figure 38(A)), and the cytokine concentration in the culture supernatant (Figure 38(B)). The vertical axis shows cytotoxic activity (%), and the horizontal axis shows antibody concentration (nM). [Figure 39]Figures 39(A) and (B) show the concentration-dependent cytotoxic activity of CD3 / CD123 bispecific antibodies and the cytokine concentrations in the culture supernatants at that time. Figure 39(A) shows the cytotoxic activity against MOLM13 cells, and Figure 39(B) shows the cytotoxic activity of the same antibodies against OCI-AML3 cells. The vertical axis shows cytotoxic activity (%), and the horizontal axis shows antibody concentration (nM). [Figure 40] Figures 40(A) to 40(D) ​​show the concentration-dependent cytokine concentrations in the culture supernatants during cytotoxicity of a CD3 / CD123 bispecific antibody. Figure 40(A) shows the IL-2 concentration in the culture supernatants during cytotoxicity of MOLM13 cells, and Figure 40(B) shows the IFN-γ concentration. Figure 40(C) shows the IL-2 concentration in the culture supernatants during cytotoxicity of OCI-AML3 cells by the same antibody, and Figure 40(D) ​​shows the IFN-γ concentration. The vertical axis shows cytokine concentration (pg / mL), and the horizontal axis shows antibody concentration (nM). [Figure 41] Figures 41(A) to 41(C) show the cytotoxic activity of a bispecific antibody containing an anti-CD3 scFv derived from a modified sequence of the anti-CD3 monoclonal antibody clone KM14 against BT-20 cells [Figure 41(A)], and the results of flow cytometry analysis of cytokine production upon cytotoxicity [Figures 41(B) and (C)]. The vertical axis of Figure 41(A) shows the name of the bispecific antibody, and the horizontal axis shows cytotoxic activity (%). The vertical axis of Figures 41(B) and (C) shows the name of the bispecific antibody, and the horizontal axis shows cytokine concentration (pg / mL), respectively. Figure 41(B) shows the concentration of INF-γ in the culture supernatant, and Figure 41(C) shows the concentration of IL-6. LLOQ indicates the limit of quantitation specified by the kit. [Figure 42] FIG. 42 illustrates design examples and structural locations of amino acid modifications introduced into the CH3 region and anti-CD3 scFv for the purpose of improving the productivity and physicochemical stability of bispecific antibodies. [Figure 43]Figure 43 shows modifications introduced into an anti-CD3 bispecific antibody designed to improve productivity and physicochemical stability, listing each element, including modifications in the CH3 portion, scFv clones, the order of VH and VL of the scFv, and amino acid modifications introduced into the scFv. [Figure 44] Figure 44 shows the results of surface plasmon resonance (SPR) measurements of the binding activity to soluble CD3 for bispecific antibodies with modifications introduced into Fc(CH3), bispecific antibodies with VL-Linker-VH anti-CD3 scFv, and bispecific antibodies with modifications introduced into anti-CD3 scFv. The left column lists the names of the bispecific antibodies tested and their KD (M). [Figure 45A] Figure 45A shows the results of measuring the cytotoxic activity against BT-20 of a bispecific antibody with modifications introduced into the Fc(CH3), a bispecific antibody with a VL-Linker-VH type anti-CD3 scFv, and a bispecific antibody with modifications introduced into the anti-CD3 scFv. The activity value in wells to which no test substance was added is set to 0%, and the activity value in wells to which 0.5% Triton X-100 was added is set to 100%, and the cytotoxic activity after 48 hours is shown. The vertical axis shows cytotoxic activity (%), and the horizontal axis shows each bispecific antibody. [Figure 45B] Figure 45B shows the results of measuring the cytotoxic activity against BT-20 of a bispecific antibody with modifications introduced into the Fc(CH3), a bispecific antibody with a VL-Linker-VH type anti-CD3 scFv, and a bispecific antibody with modifications introduced into the anti-CD3 scFv. The activity value in wells to which no test substance was added is set to 0%, and the activity value in wells to which 0.5% Triton X-100 was added is set to 100%, and the cytotoxic activity after 48 hours is shown. The vertical axis shows each bispecific antibody, and the horizontal axis shows cytotoxic activity (%). [Figure 45C]Figure 45C shows the results of measuring the cytotoxic activity against BT-20 of a bispecific antibody with modifications introduced into the Fc(CH3), a bispecific antibody with a VL-Linker-VH type anti-CD3 scFv, and a bispecific antibody with modifications introduced into the anti-CD3 scFv. The activity value in wells without the test substance was set to 0%, and the activity value in wells with 0.5% Triton X-100 was set to 100%, and the cytotoxic activity after 48 hours is shown. The vertical axis shows each bispecific antibody, and the horizontal axis shows cytotoxic activity (%). [Figure 45D] Figure 45D shows the results of measuring the cytotoxic activity against BT-20 of a bispecific antibody with modifications introduced into the Fc(CH3), a bispecific antibody with a VL-Linker-VH anti-CD3 scFv, and a bispecific antibody with modifications introduced into the anti-CD3 scFv. The activity value in wells without the test substance was set to 0%, and the activity value in wells with 0.5% Triton X-100 was set to 100%, and the cytotoxic activity after 48 hours is shown. The vertical axis shows each bispecific antibody, and the horizontal axis shows cytotoxic activity (%). [Figure 46A] Figure 46A shows the results of flow cytometry analysis of cytokine production during BT-20 cytotoxicity in a bispecific antibody with modifications introduced into Fc(CH3), a bispecific antibody with a VL-Linker-VH anti-CD3 scFv, and a bispecific antibody with modifications introduced into the anti-CD3 scFv. Figure 46(A) shows the concentration of INF-γ in the culture supernatant. [Figure 46B] Figure 46B shows the results of flow cytometry analysis of cytokine production during BT-20 cytotoxicity in a bispecific antibody with modifications introduced into Fc(CH3), a bispecific antibody with a VL-Linker-VH anti-CD3 scFv, and a bispecific antibody with modifications introduced into the anti-CD3 scFv. Figure 46B shows the concentration of IL-6 in the culture supernatant. [Figure 46C]Figure 46C shows the results of flow cytometry analysis of cytokine production during BT-20 cytotoxicity in bispecific antibodies with modifications introduced into the Fc(CH3) and anti-CD3 scFv. Figure 46C shows the INF-γ concentration in the culture supernatant. [Figure 46D] Figure 46D shows the results of flow cytometry analysis of cytokine production during BT-20 cytotoxicity in bispecific antibodies with modifications introduced into the Fc(CH3) and anti-CD3 scFv. Figure 46(D) shows the concentration of IL-6 in the culture supernatant. [Figure 46E] Figure 46E shows the results of flow cytometry analysis of cytokine production during BT-20 cytotoxicity in a bispecific antibody with modifications introduced into Fc(CH3), a bispecific antibody with a VL-Linker-VH anti-CD3 scFv, and a bispecific antibody with modifications introduced into the anti-CD3 scFv. Figure 46(E) shows the INF-γ concentration in the culture supernatant. [Figure 46F] Figure 46F shows the results of flow cytometry analysis of cytokine production during BT-20 cytotoxicity in a bispecific antibody with modifications introduced into Fc(CH3), a bispecific antibody with a VL-Linker-VH anti-CD3 scFv, and a bispecific antibody with modifications introduced into the anti-CD3 scFv. Figure 46(F) shows the concentration of IL-6 in the culture supernatant. DETAILED DESCRIPTION OF THE INVENTION

[0027] The present invention relates to a bispecific antibody comprising an antigen-binding domain that binds to CD3 and an antigen-binding domain that binds to a disease-associated antigen, or a bispecific antibody fragment thereof (hereinafter also referred to as the bispecific antibody of the present invention or the bispecific antibody fragment).

[0028] Examples of bispecific antibodies of the present invention include bispecific antibodies comprising a CD3-binding domain with reduced affinity for CD3, bispecific antibodies comprising an antigen-binding domain that binds to CD3 and an antigen-binding domain that binds to a disease-associated antigen, bispecific antibodies comprising an Fc region capable of binding to an Fc receptor and comprising a CD3-binding domain with reduced affinity for CD3 and an antigen-binding domain that binds to a disease-associated antigen, and bispecific antibodies comprising an Fc region with enhanced affinity for an Fc receptor and comprising a CD3-binding domain with reduced affinity for CD3 and an antigen-binding domain that binds to a disease-associated antigen, etc. By binding to both CD3 and a disease-associated antigen, the bispecific antibodies of the present invention can cytotoxicize cells expressing the disease-associated antigen in a T cell- and / or NK cell-dependent manner without inducing excessive cytokine production.

[0029] In the present invention, CD3 is used synonymously with CD3E, T3E, and T-cell surface glycoprotein CD3 epsilon chain. Examples of CD3 include human CD3 containing the amino acid sequence set forth in GenBank accession No. NP_000724.1 or SEQ ID NO: 138 at NCBI (http: / / www.ncbi.nlm.nih.gov / ) and monkey CD3 containing the amino acid sequence set forth in GenBank accession No. NP_001270544.1 or SEQ ID NO: 139. Further examples include polypeptides having the function of CD3, each of which has an amino acid sequence in which one or more amino acids have been deleted, substituted, or added to the amino acid sequence set forth in SEQ ID NO: 138, GenBank accession No. NP_000724.1, SEQ ID NO: 139, or GenBank accession No. NP_001270544.1.

[0030] The CD3 of the present invention also encompasses polypeptides comprising an amino acid sequence that is typically 70% or more, preferably 80% or more, and more preferably 90% or more identical to the amino acid sequence shown in SEQ ID NO: 138, GenBank accession No. NP_000724.1, SEQ ID NO: 139, or GenBank accession No. NP_001270544.1, and most preferably 95% or more, 96% or more, 97% or more, 98% or more, or 99% or more identical thereto, and that have the function of CD3.

[0031] A polypeptide having an amino acid sequence in which one or more amino acid residues have been deleted, substituted, or added in the amino acid sequence shown in SEQ ID NO: 138, GenBank accession No. NP_000724.1 or SEQ ID NO: 139, GenBank accession No. NP_001270544.1 can be produced by site-directed mutagenesis [Molecular Cloning, A Laboratory Manual, Second Edition, Cold Spring Harbor Laboratory Press (1989); Current Protocols in Molecular Biology, John Wiley & Sons (1987-1997); Nucleic Acids Research, 10, 6487 (1982); Proc. Natl. Acad. Sci. USA, 79, 6409 (1982); Gene, 34, 315 (1985); Nucleic Acids Research, 13, 4431 (1985); Proceedings of the National Academy of Sciences in USA, 82, 488 (1985)] or the like, can be obtained by introducing site-specific mutations into DNA encoding the amino acid sequence shown in SEQ ID NO: 138, GenBank accession No. NP_000724.1 or SEQ ID NO: 139, GenBank accession No. NP_001270544.1. The number of amino acids to be deleted, substituted, or added is not particularly limited, but is preferably one to several tens of amino acids, for example, one to twenty, and more preferably one to several amino acids, for example, one to five.

[0032] Examples of genes encoding CD3 include the nucleotide sequence of human CD3 shown in SEQ ID NO: 141 or GenBank accession No. NM_000733.3, and the nucleotide sequence of monkey CD3 shown in SEQ ID NO: 142 or GenBank accession No. NM_001283615.1.

[0033] Furthermore, a gene comprising a DNA encoding a polypeptide having the function of CD3, which comprises a nucleotide sequence in which one or more nucleotides are deleted, substituted or added in the nucleotide sequence shown in SEQ ID NO: 141, GenBank accession No. NM_000733.3 or SEQ ID NO: 142, GenBank accession No. NM_001283615.1, for example; A gene comprising a DNA encoding a polypeptide having the function of CD3, which comprises a nucleotide sequence having preferably 60% or more homology, more preferably 80% or more homology, and even more preferably 95% or more homology to the nucleotide sequence shown in SEQ ID NO: 141, GenBank accession No. NM_000733.3 or SEQ ID NO: 142, GenBank accession No. NM_001283615.1; and Genes encoding CD3 of the present invention also include genes consisting of DNA that hybridizes under stringent conditions with DNA consisting of the nucleotide sequence shown in SEQ ID NO: 141, GenBank accession No. NM_000733.3 or SEQ ID NO: 142, GenBank accession No. NM_001283615.1, and which contain DNA encoding a polypeptide having the function of CD3.

[0034] DNA that hybridizes under stringent conditions refers to hybridizable DNA obtained by colony hybridization, plaque hybridization, Southern blot hybridization, DNA microarray, or the like, using, for example, DNA having the base sequence shown in SEQ ID NO: 141, GenBank accession No. NM_000733.3 or SEQ ID NO: 142, GenBank accession No. NM_001283615.1 as a probe.

[0035] Specifically, the DNA can be identified by hybridizing a filter or slide glass onto which DNA derived from hybridized colonies or plaques, or a PCR product or oligo DNA having the sequence, is immobilized in the presence of 0.7 to 1.0 mol / L sodium chloride at 65°C [Molecular Cloning, A Laboratory Manual, Second Edition, Cold Spring Harbor Laboratory Press (1989), Current Protocols in Molecular Biology, John Wiley & Sons (1987-1997), DNA Cloning 1: Core Techniques, A Practical Approach, Second Edition, Oxford University (1995)], followed by washing the filter or slide glass at 65°C using 0.1 to 2x SSC solution (1x SSC solution consists of 150 mmol / L sodium chloride and 15 mmol / L sodium citrate). Examples of hybridizable DNA include DNA that has preferably 60% or more homology, more preferably 80% or more homology, and even more preferably 95% or more homology to the base sequence shown in SEQ ID NO: 141, GenBank accession No. NM_000733.3 or SEQ ID NO: 142, GenBank accession No. NM_001283615.1.

[0036] Genetic polymorphisms are often found in the nucleotide sequences of genes encoding proteins in eukaryotes. Genes used in the present invention that have small mutations in their nucleotide sequences due to such polymorphisms are also included in the CD3-encoding genes of the present invention.

[0037] Unless otherwise specified, the homology values ​​in the present invention may be values ​​calculated using a homology search program known to those skilled in the art. For base sequences, examples include values ​​calculated using default parameters in BLAST [J. Mol. Biol., 215, 403 (1990)], and for amino acid sequences, examples include values ​​calculated using default parameters in BLAST2 [Nucleic Acids Research, 25, 3389 (1997), Genome Research, 7, 649 (1997), http: / / www.ncbi.nlm.nih.gov / Education / BLASTinfo / information3.html].

[0038] The default parameters are: G (Cost to open gap) is 5 for nucleotide sequences and 11 for amino acid sequences; -E (Cost to extend gap) is 2 for nucleotide sequences and 1 for amino acid sequences; -q (Penalty for nucleotide mismatch) is -3; -r (reward for nucleotide match) is 1; -e (expect value) is 10; -W (wordsize) is 11 residues for nucleotide sequences and 3 residues for amino acid sequences; -y[Dropoff(X) for blast extensions in bits] is 20 for blastn and 7 for programs other than blastn; -X (X dropoff value for gapped alignment in bits) is 15; and -Z (final X dropoff value for gapped alignment in bits) is 50 for blastn and 25 for programs other than blastn (http: / / www.ncbi.nlm.nih.gov / blast / html / blastcgihelp.html).

[0039] A polypeptide consisting of a partial sequence of the amino acid sequence of CD3 can be produced by methods known to those skilled in the art. For example, a polypeptide consisting of a partial sequence of CD3 can be produced by deleting a portion of the DNA encoding the amino acid sequence shown in SEQ ID NO: 138, GenBank accession No. NP_000724.1 or SEQ ID NO: 139, GenBank accession No. NP_001270544.1, and culturing a transformant into which an expression vector containing the deleted DNA has been introduced.

[0040] Furthermore, based on the polypeptide or DNA produced by the above method, a polypeptide having an amino acid sequence in which one or more amino acids are deleted, substituted, or added in a partial sequence of the amino acid sequence shown in, for example, SEQ ID NO: 138, GenBank accession No. NP_000724.1 or SEQ ID NO: 139, GenBank accession No. NP_001270544.1 can be obtained by a method similar to that described above.

[0041] Furthermore, a polypeptide consisting of a partial sequence of the amino acid sequence of CD3, or a polypeptide having an amino acid sequence in which one or more amino acids are deleted, substituted, or added in a partial sequence of the amino acid sequence of CD3, can also be produced by chemical synthesis methods such as the fluorenylmethyloxycarbonyl (Fmoc) method and the t-butyloxycarbonyl (tBoc) method.

[0042] Examples of the extracellular region of CD3 in the present invention include regions predicted from the amino acid sequence of human CD3 shown in GenBank accession No. NP_000724.1 using known transmembrane region prediction programs such as SOSUI (http: / / sosui.proteome.bio.tuat.ac.jp / sosuiframe0.html), TMHMM ver.2 (http: / / www.cbs.dtu.dk / services / TMHMM-2.0 / ), or ExPASy Proteomics Server (http: / / Ca.expasy.org / ). Specifically, examples of the extracellular region of CD3 include the amino acid sequence shown in SEQ ID NO: 140 or positions 22 to 126 of GenBank accession No. NP_000724.1.

[0043] The function of CD3 is to directly associate with the T cell receptor (TCR) to form a subunit structure of a protein complex, recognize antigen peptides bound to the major histocompatibility complex (MHC), and participate in the transmission of intracellular signals. Cells that express CD3 include T cells, NKT cells, γδT cells, and their precursor cells, as well as mature thymocytes.

[0044] The disease-associated antigen in the present invention may be any antigen that is involved in a disease such as cancer, immune disease, allergic disease, autoimmune disease, central nervous system disease, or cardiovascular disease, and examples thereof include cytokines, chemokines, growth factors and their receptors, and cluster of differentiation (hereinafter referred to as CD) antigens.

[0045] Examples of cytokine or growth factor receptors include receptors for interferon (hereinafter referred to as IFN)-α, IFN-β, IFN-γ, interleukin (hereinafter referred to as IL)-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-12, IL-13, IL-15, IL-17, IL-18, IL-21, IL-23, IL-27, granulocyte colony-stimulating factor (G-CSF), granulocyte / macrophage colony-stimulating factor (GM-CSF), or macrophage colony-stimulating factor (M-CSF).

[0046] Examples of chemokine receptors include receptors for SLC, ELC, I-309, TARC, MDC, MIP-3α, and CTACK.

[0047] Examples of growth factor receptors include receptors for epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), angiopoietin, fibroblast growth factor (FGF), hepatocyte growth factor (HGF), platelet-derived growth factor (PDGF), insulin-like growth factor (IGF), erythropoietin (EPO), TGFβ, ephrin, angiopoietin, Frizzled ligand, and SDF-1.

[0048] As CD antigens, there are CD1a, CD1c (BDCA1), CD1d, CD2, CD3, CD4, CD5, CD6, CD7, CD8, CD10, CD11a, CD11b, CD11c, CD14, CD16, CD18, CD19, CD20, CD21, CD22, CD23, CD24, CD25, CD26 (DPP-4), CD27, CD28, CD30, CD32, CD34, CD37, CD38, CD39, CD40, CD43, CD44, CD45, CD47, CD49, CD51, CD52, CD53, CD54, CD55, CD56, CD57, CD59, CD62E, CD62L, CD62P, CD64, CD66a (CEACAM1), CD66b (NCA-95), CD66c (NCA-50 / 90), CD66d (CGM1), CD66e (CEA), CD66f (PSG), CD68, CD69, CD70, CD72, CD73, CD74, CD75, CD76, CD77, CD78, CD79a, CD79b, CD80 (B7.1), CD81, CD82, CD83, CD84 (SLAMF5), CD85a (ILT-5), CD85b (ILT8), CD85c (LIR8), CD85d (ILT4), CD85f (ILT11), CD85g (ILT7), CD85h (ILT1), CD85i (LIR6a), CD85j (ILT2), CD85k (ILT3), CD85m (ILT10), CD86 (B7.2), CD87, CD89, CD94 (NKG2), CD95 (Fas), CD98, CD103, CD107a (LAMP1), CD114 (G-CSFR), CD115 (M-CSFR), CD116 (GM-CSFR), CD117 (SCF-R), CD119 (IFNGR1), CD121a (IL-IR1), CD122 (IL-2Rb), CD123 (IL-3Ra), CD124 (IL-4Ra), CD125 (IL-5Ra), CDI26 (IL-6Ra), CD127 (IL-7Ra), CD134 (OX40), CD135 (FLT3), CD137 (4-1BB), CD138 (Syndecan-1), CD140 (PDGFR), CD146 (MUC18), CD147 (EMMRRIN), CD152 (CTLA-4), CD158a (KIR2DL1), CD158b1 (KIR2DL2), CD158b2 (KIRDL3),CD158c(KIR2DS6)、CD158d(KIR2DL4)、CD158e1(KIR3DL1)、CD158e2(KIR3DS1)、CD158f(KIR2DL5)、CD158g(KIR2DS5)、CD158h(KIR2DS1)、CD158i(KIR2DS4)、CD158j(KIR2DS2)、CD158k(KIR3DL2)、CD159a(NKG2A)、CD159c(NKG2C)、CD161(NKRP1A)、CD162(PSGL-1)、CD163、CD169(SIGLEC1)、CD178(FasL)、CD183(CXCR3)、CD184(CXCR4)、CD185(CXCR5)、CD191(CCR1)、CD193(CCR3)、CD194(CCR4)、CD195(CCR5)、CD196(CCR6)、CD197(CCR7)、CD198(CCR8)、CD199(CCR9)、CD200(OX2)、CD206(MMR)、CD207(Langerin),CD209(DC-SIGN)、CD212(IL-12Rβ1)、CD213a1(IL-13Ra1)、CD213a2(IL-13Ra2)、CD215(IL-15RA)、CD217(IL-17R)、CD218a(IL-18Ra)、CD218b(IL-18Rβ)、CD223(LAG3)、CD226(DNAM-1)、CD229(SLAMF3)、CD252(OX40L)、CD269(BCMA)、CD272(BTLA)、CD274(PD-L1)、CD276(B7H3)、CD278(ICOS)、CD279(PD-1)、CD281(TLR1)、CD282(TLR2)、CD283(TLR3)、CD284(TLR4)、CD286(TLR6)、CD288(TLR8)、CD289(TLR9)、CD294(CRTH2)、CD301(MGL)、CD302(DCL1)、CD303(BDCA2)、CD304(BDCA4)、CD317(BST2)、CD324(E-cadherin)、CD326(EpCAM)、CD357(GITR)、CD358(DR6)、CD360(IL-21R)、CD365(TIM-1)、CD366(TIM-3)、CD369(DECTIN-1)、CD370(CLEC9A)、Human leukocyte antigen (HLA)-Class II and HLA-I are examples.

[0049] Other antigens for antibodies related to cancer, immune diseases, allergic diseases, autoimmune diseases, central nervous system or circulatory system diseases, etc. include, for example, gangliosides GM1, GM2, GD2, GD3, Lewis X, Lewis Y, glypican-3, claudin, ASCT-2, CD3, CD4, CD40, CD40 ligand, B7 family molecules (e.g., CD80, CD86, CD274, B7-DC, B7-H2, B7-H3, or B7-H4), ligands for B7 family molecules (e.g., CD28, CTLA-4, ICOS, PD-1, or BTLA), OX-40, OX-40 ligand, CD137, tumor necrosis factor (TNF) receptor family molecules (e.g., DR3, DR4, DR5, BAFFR, LIGHT, TNFR1, or TNFR2), and TNF-related apoptosis-inducing ligand. receptor (TRAIL) family molecules, receptor family of TRAIL family molecules (e.g., TRAIL-R1, TRAIL-R2, TRAIL-R3, or TRAIL-R4), receptor activator of nuclear factor kappa B ligand (RANK), RANK ligand, CD25, folate receptor, mesothelin, SIGLEC8, cytokine and chemokine receptors [e.g., IL-1RII, IL-12Rβ2, IL-17RB, IL-23R, IL-27Rα, IL-31R, IL-33Rα, IL-36R, transforming growth factor (TGF)βRII, CCR2, CCR10, CXCR1, CXCR2], NK cell receptors (e.g., NKG2D, E4BP4, NKp30, NKp44, NKp46, AhR), T cell receptors (e.g., TCRα / β, TCR These include Vβ11, TCRγ / δ, TSLPR, SLAM, SLAMF6, LAP, GARP, SR-A1, CD200R, DCR3, TIGIT), B cell receptors (e.g., BLYS, APRIL, TSLPR), and dendritic cell receptors (e.g., FCER1A, TLR7, CADM1, XCR1, BTLA, SIRPA, DCIR, TROP2, AXL, SIGLEC6, SIGLEC15, CX3CR1, S100A8, S100A9, ASGR1).

[0050] The antibody of the present invention is a protein derived from genes (referred to as "antibody genes") encoding all or part of the heavy chain variable and constant regions and light chain variable and constant regions that constitute immunoglobulins. The antibody of the present invention includes antibodies and antibody fragments of any immunoglobulin class and subclass.

[0051] The heavy chain (H chain) is the polypeptide with the larger molecular weight of the two types of polypeptides that make up an immunoglobulin molecule. The heavy chain determines the class and subclass of the antibody. IgA, IgD, IgE, IgG, and IgM each have α, δ, ε, γ, and μ chains as heavy chains, and the constant regions of the heavy chains are characterized by different amino acid sequences. The light chain (L chain) is the polypeptide with the smaller molecular weight of the two types of polypeptides that make up an immunoglobulin molecule. In human antibodies, there are two types of light chains: κ chains and λ chains.

[0052] The variable region (V region) usually refers to a highly diverse region present in the amino acid sequence at the N-terminus of an immunoglobulin. The portion other than the variable region has a structure with little diversity, and is therefore called the constant region (C region). The variable regions of the heavy and light chains associate to form the antigen-binding site, which determines the binding characteristics of the antibody to the antigen.

[0053] In the heavy chain of a human antibody, the variable region corresponds to the amino acid sequence from position 1 to position 117 in the EU index of Kabat et al. (Kabat et al., Sequences of proteins of immunological interest, 1991 Fifth edition) (hereinafter simply referred to as the EU index), and the constant region corresponds to the amino acid sequence from position 118 onwards. In the light chain of a human antibody, the variable region corresponds to the amino acid sequence from position 1 to position 107 in the Kabat numbering, and the constant region corresponds to the amino acid sequence from position 108 onwards. Hereinafter, the heavy chain variable region and the light chain variable region are abbreviated as VH and VL, respectively.

[0054] The antigen-binding site is the site in an antibody that recognizes and binds to an antigen and forms a three-dimensional structure complementary to the antigenic determinant (epitope). The antigen-binding site generates strong intermolecular interactions with the antigenic determinant. The antigen-binding site is composed of VH and VL, which contain at least three complementarity-determining regions (CDRs). In the case of human antibodies, VH and VL each have three CDRs. These CDRs are referred to as CDR1, CDR2, and CDR3, respectively, from the N-terminus.

[0055] The heavy chain constant region and light chain constant region of the constant region are designated as CH and CL, respectively. CH is classified into heavy chain subclasses: α chain, δ chain, ε chain, γ chain, and μ chain. CH is composed of a CH1 domain, hinge domain, CH2 domain, and CH3 domain arranged in order from the N-terminus, and the CH2 domain and CH3 domain together are called the Fc region. On the other hand, CL is classified into two subclasses called Cλ chain and Cκ chain.

[0056] The anti-CD3 antibody of the present invention refers to a monoclonal antibody that specifically recognizes and binds to the extracellular domain of CD3. The antibody of the present invention also encompasses polyclonal and oligoclonal antibodies.

[0057] In the present invention, the binding of a bispecific antibody or a bispecific antibody fragment to CD3 or a disease-related antigen can be confirmed by, for example, a known immunological detection method, preferably a fluorescent cell staining method, to confirm the binding of the antibody to cells expressing CD3 or a disease-related antigen. Also, known immunological detection methods [Monoclonal Antibodies - Principles and Practice, Third Edition, Academic Press (1996), Antibodies - A Laboratory Manual, Cold Spring Harbor Laboratory (1988), Monoclonal Antibody Experiment Manual, Kodansha Scientific (1987)], etc. can also be used in combination.

[0058] Monoclonal antibodies are antibodies secreted by antibody-producing cells that maintain monoclonality and recognize a single epitope (also called an antigenic determinant). Monoclonal antibody molecules have the same amino acid sequence (primary structure) and have a single structure. Polyclonal antibodies are a group of antibody molecules secreted by antibody-producing cells of different clones. Oligoclonal antibodies are a group of antibody molecules that are a mixture of multiple different monoclonal antibodies.

[0059] An epitope refers to a structural site of an antigen that an antibody recognizes and binds to. Examples of epitopes include a single amino acid sequence that a monoclonal antibody recognizes and binds to, a three-dimensional structure consisting of an amino acid sequence, an amino acid sequence to which a sugar chain is bound, and a three-dimensional structure consisting of an amino acid sequence to which a sugar chain is bound.

[0060] The monoclonal antibodies of the present invention include antibodies produced by hybridomas and recombinant antibodies produced by transformants transformed with an expression vector containing an antibody gene.

[0061] Hybridomas can be prepared, for example, by preparing an antigen, obtaining antibody-producing cells with antigen specificity from an animal immunized with the antigen, and then fusing the antibody-producing cells with myeloma cells. The hybridomas are cultured or administered to an animal to cause the hybridomas to become ascites tumors, and the culture medium or ascites is separated and purified to obtain the desired monoclonal antibody. Any animal can be immunized with the antigen as long as it is capable of producing hybridomas, but mice, rats, hamsters, rabbits, and the like are preferred. Alternatively, antibody-producing cells can be obtained from such an immunized animal, immunized in vitro, and then fused with myeloma cells to produce hybridomas.

[0062] Examples of recombinant antibodies of the present invention include antibodies produced by genetic recombination techniques, such as recombinant mouse antibodies, recombinant rat antibodies, recombinant hamster antibodies, recombinant rabbit antibodies, human chimeric antibodies (also called chimeric antibodies), humanized antibodies (also called CDR-grafted antibodies), and human antibodies. For recombinant antibodies, the animal species from which the heavy and light chain variable and constant regions are derived can be determined depending on the target animal species and purpose. For example, if the target animal species is human, the variable regions can be derived from humans or non-human animals such as mice, and the constant regions and linkers can be derived from humans.

[0063] A chimeric antibody refers to an antibody consisting of the VH and VL of an antibody from an animal other than a human (non-human animal) and the CH and CL of a human antibody. Any non-human animal can be used as long as it is possible to produce a hybridoma, such as a mouse, rat, hamster, or rabbit. Chimeric antibodies can be produced by obtaining cDNA encoding VH and VL from a hybridoma derived from a non-human animal that produces a monoclonal antibody, inserting the cDNA into an expression vector for animal cells containing DNA encoding the CH and CL of a human antibody to construct a chimeric antibody expression vector, and then introducing the vector into animal cells for expression.

[0064] A humanized antibody refers to an antibody in which the CDRs of the VH and VL of a non-human animal antibody have been grafted onto the corresponding CDRs of the VH and VL of a human antibody. The regions other than the CDRs of the VH and VL are referred to as framework regions (hereinafter referred to as FR). A humanized antibody can be produced by constructing a cDNA encoding a VH amino acid sequence consisting of the amino acid sequence of the CDR of the VH of a non-human animal antibody and the amino acid sequence of the FR of the VH of any human antibody, and a cDNA encoding a VL amino acid sequence consisting of the amino acid sequence of the CDR of the VL of a non-human animal antibody and the amino acid sequence of the FR of the VL of any human antibody, and inserting each into an expression vector for animal cells containing DNA encoding the CH and CL of a human antibody to construct a humanized antibody expression vector, which is then introduced into animal cells for expression.

[0065] Human antibodies originally refer to antibodies that naturally occur in the human body, but also include human antibody phage libraries produced through recent advances in genetic engineering, cell engineering, and developmental engineering technologies, and antibodies obtained from human antibody-producing transgenic animals.

[0066] Antibodies naturally occurring in the human body can be obtained, for example, by infecting human peripheral blood lymphocytes with EB virus or the like to immortalize them, cloning them, culturing the lymphocytes that produce the antibody, and purifying the antibody from the culture supernatant.

[0067] A human antibody phage library is a library in which antibody fragments such as Fab and scFv are expressed on the surface of phages by inserting antibody genes prepared from human B cells into phage genes. Phages expressing antibody fragments with the desired antigen-binding activity on their surface can be recovered from the library using their binding activity toward an antigen-immobilized substrate as an indicator. These antibody fragments can then be converted into human antibody molecules consisting of two complete heavy chains and two complete light chains using genetic engineering techniques.

[0068] A human antibody-producing transgenic animal refers to an animal in which a human antibody gene has been incorporated into its cells. Specifically, for example, a human antibody-producing transgenic mouse can be produced by introducing a human antibody gene into mouse ES cells, transplanting the ES cells into an early mouse embryo, and then allowing the mouse to develop. Human antibodies derived from a human antibody-producing transgenic animal can be prepared by obtaining and culturing hybridomas using a hybridoma production method commonly used for non-human animals, and then producing and accumulating antibodies in the culture supernatant.

[0069] The CH of the recombinant antibody may be any human immunoglobulin, but is preferably of the human immunoglobulin G (hIgG) class. Furthermore, any of the subclasses of the hIgG class, such as hIgG1, hIgG2, hIgG3, and hIgG4, may be used. The CL of the recombinant antibody may be any human immunoglobulin, and may be either the κ or λ class.

[0070] In the present invention, an antibody fragment refers to a protein that contains an antigen-binding site and has antigen-binding activity to the antigen, such as Fab, Fab', F(ab')2, scFv, diabody, dsFv, VHH, or a peptide containing CDR.

[0071] Fab is an antibody fragment with a molecular weight of approximately 50,000 that has antigen-binding activity and is obtained by treating an IgG antibody with the protease papain (cleaved at the 224th amino acid residue of the H chain). Approximately half of the N-terminal H chain is linked to the entire L chain via a disulfide bond (SS bond). The polypeptide chain of Fab containing VH and CH1 is referred to as the Fab heavy chain (H chain) or VH-CH1. The polypeptide chain of Fab containing VL and CL is referred to as the Fab light chain (L chain) or VL-CL.

[0072] F(ab')2 is an antibody fragment obtained by treating IgG with the protease pepsin (cleaved at the 234th amino acid residue of the H chain). It has a molecular weight of approximately 100,000 and is slightly larger than Fab fragments linked via disulfide bonds in the hinge region, and has antigen-binding activity.

[0073] Fab' is an antibody fragment with a molecular weight of approximately 50,000 that has antigen-binding activity and is obtained by cleaving the SS bond in the hinge region of the above-mentioned F(ab')2.

[0074] scFv is a VH-P-VL or VL-P-VH polypeptide in which one VH and one VL are linked using an appropriate peptide linker (P) of 12 or more residues, and is an antibody fragment that has antigen-binding activity.

[0075] Diabodies are antibody fragments formed by dimerization of scFvs with the same or different antigen-binding specificities, and have bivalent antigen-binding activity for the same antigen or specific antigen-binding activity for different antigens.

[0076] dsFv refers to polypeptides in which one amino acid residue in each of VH and VL is substituted with a cysteine ​​residue, and the cysteine ​​residues are linked via an S—S bond.

[0077] VHH (also called nanobody) refers to the heavy chain variable region of a VHH antibody, which can bind to an antigen in the absence of other polypeptides.

[0078] VHH antibodies are present in camelids such as alpacas and cartilaginous fish such as sharks, and are composed only of heavy chains, lacking light chains and CH1.

[0079] CDR-containing peptides comprise at least one region of the VH or VL CDR. Peptides containing multiple CDRs can be produced by linking the CDRs directly or via an appropriate peptide linker. CDR-containing peptides can be produced by constructing DNA encoding the VH and VL CDRs of the bispecific antibody of the present invention, inserting the DNA into a prokaryotic or eukaryotic expression vector, and introducing the expression vector into a prokaryotic or eukaryotic organism for expression. CDR-containing peptides can also be produced by chemical synthesis methods such as the Fmoc method or the tBoc method.

[0080] In the present invention, a bispecific antibody fragment is a bispecific antibody fragment that essentially consists of a partial structure of a bispecific antibody and has antigen-binding activity against two types of antigens.

[0081] The present invention also encompasses fusion proteins in which the bispecific antibody fragment of the present invention is linked to an Fc domain, Fc fusion proteins (also known as immunoadhesins) in which the Fc domain is linked to a naturally occurring ligand or receptor, and Fc fusion proteins in which multiple Fc domains are fused. Furthermore, Fc domains containing amino acid residue modifications for the purposes of enhancing or deleting antibody effector activity, stabilizing the antibody, and controlling its half-life in blood can also be used in the bispecific antibodies of the present invention.

[0082] The bispecific antibody of the present invention refers to a polypeptide or protein having two antigen-binding domains with different specificities. Each antigen-binding domain of a bispecific antibody may bind to a different epitope of a single antigen or may bind to different antigens.

[0083] In the present invention, an antigen-binding domain is a partial structure that has the function of specifically recognizing and binding to an antigen. The antigen-binding domain of the present invention may be in any form, such as an antibody, an antibody fragment thereof, a ligand, a receptor, or a polypeptide that can be prepared by genetic engineering, such as a naturally occurring interacting molecule, a protein molecule or a fragment thereof, or a conjugate of the protein molecule with a small molecule or a natural product.

[0084] The two antigen-binding domains of the bispecific antibody of the present invention are a CD3-binding domain and a disease-associated antigen-binding domain. When the bispecific antibody contains multiple disease-associated antigen-binding domains, the domains may bind to the same antigen or different antigens.

[0085] The bispecific antibodies of the present invention comprise an Fc region capable of binding to an Fc receptor, a CD3-binding domain, and a disease-associated antigen-binding domain, and the C-terminus of the Fc region is linked to the CD3-binding domain directly or via a linker.

[0086] In the present invention, the CD3-binding domain refers to an antigen-binding domain that binds to CD3.

[0087] In the present invention, a disease-associated antigen-binding domain refers to an antigen-binding domain that binds to a disease-associated antigen.

[0088] The CD3-binding domain of the present invention may be any domain that specifically recognizes and binds to CD3, and examples thereof include a CD3-binding domain comprising a CDR sequence derived from an anti-CD3 antibody, a CD3-binding domain comprising VH and VL derived from an anti-CD3 antibody, etc. A CD3 antigen-binding domain comprising CDR sequences or VH and VL derived from an anti-CD3 antibody is preferably an scFv.

[0089] In the present invention, the disease-associated antigen-binding domain may be any domain that specifically recognizes and binds to a disease-associated antigen, but is preferably a variable region consisting of antibody Fab or VH and VL.

[0090] In the present invention, the binding of a polypeptide, an antibody or an antibody fragment thereof, or a bispecific antibody or a bispecific antibody fragment thereof to CD3 and / or a disease-related antigen can be confirmed by, for example, a known immunological detection method, preferably a fluorescent cell staining method, to confirm the binding of the antibody to cells expressing the antigen to be evaluated. Also, known immunological detection methods [Monoclonal Antibodies - Principles and Practice, Third Edition, Academic Press (1996), Antibodies - A Laboratory Manual, Cold Spring Harbor Laboratory (1988), Monoclonal Antibody Experiment Manual, Kodansha Scientific (1987)], etc. can also be used in combination.

[0091] Examples of bispecific antibodies or bispecific antibody fragments of the present invention include anti-CD3 bispecific antibodies that have a reduced ability to induce cytokine production in the presence of CD3-positive T cells and disease-related antigen-positive cells, compared to anti-CD3 bispecific antibodies whose CD3-binding domain includes a CD3-binding domain derived from anti-CD3 monoclonal antibody SP34 (U.S. Pat. No. 10,066,015). Such bispecific antibodies or bispecific antibody fragments are preferred because they cause fewer side effects associated with cytokine production, such as chills, nausea, fatigue, headache, fever, tachycardia, and / or blood pressure fluctuations, and are less likely to cause cytokine release syndrome.

[0092] In the present invention, the ability to induce cytokine production refers to the activity of a bispecific antibody or bispecific antibody fragment of the present invention to induce cytokine production by T cells, target cells, NK cells, etc., upon binding to CD3 on T cells, an Fc receptor on NK cells, and / or a disease-related antigen on target cells.

[0093] Cytokines whose production is preferably reduced may be any cytokine that is associated with the side effects caused by excessive or unnecessary cytokine production, as described above, and examples include inflammatory cytokines such as interferon-γ (IFN-γ), tumor necrosis factor-α (TNF-α), and interleukin-6 (IL-6), as well as IL-2, IL-4, and IL-10.

[0094] The bispecific antibody or bispecific antibody fragment of the present invention preferably binds to CD3 and a disease-associated antigen expressed on different cells.

[0095] The bispecific antibody or bispecific antibody fragment of the present invention is preferably one that induces cell death of target cells by binding to CD3 on T cells and a disease-associated antigen on target cells, and more preferably one that specifically damages only target cells in the presence of CD3-positive T cells and target cells, but does not exhibit cytotoxic activity in the absence of CD3-positive cells or target cells.

[0096] The mechanism of the cytotoxic activity of the bispecific antibody or bispecific antibody fragment of the present invention includes ADCC activity, CDC activity, ADCP activity, and ADTC activity.

[0097] Specifically, examples of the bispecific antibody or bispecific antibody fragment of the present invention include a bispecific antibody or bispecific antibody fragment that specifically induces cytotoxicity and / or cell death of disease-associated antigen-positive cells when bound to both CD3-positive cells and disease-associated antigen-positive cells.

[0098] The number of binding domains for a given antigen in a single bispecific antibody molecule is called the binding valency. For example, in the present invention, if a single bispecific antibody molecule has two antigen-binding domains that bind to CD3 and two antigen-binding site domains that bind to a disease-specific antigen, the bispecific antibody binds to both CD3 and the disease-specific antigen with two valencies.

[0099] Furthermore, bispecific antibodies of the present invention also include antibodies comprising multiple antigen-binding domains linked via suitable linkers, such as linkers comprising immunoglobulin domains or fragments thereof.

[0100] The immunoglobulin domain used as a linker in the present invention is a peptide of approximately 100 amino acid residues with an amino acid sequence similar to that of an immunoglobulin and containing at least two cysteine ​​residues, as its minimum unit. In the present invention, the immunoglobulin domain also encompasses polypeptides containing multiple immunoglobulin domains of the above minimum unit. Examples of immunoglobulin domains include VH, CH1, CH2, and CH3 of the immunoglobulin heavy chain, and VL and CL of the immunoglobulin light chain.

[0101] The animal species of the immunoglobulin is not particularly limited, but is preferably human. The subclass of the constant region of the immunoglobulin heavy chain may be any of IgD, IgM, IgG1, IgG2, IgG3, IgG4, IgA1, IgA2, and IgE, preferably IgG-derived or IgM-derived. The subclass of the constant region of the immunoglobulin light chain may be either κ or λ.

[0102] Immunoglobulin domains are also present in proteins other than immunoglobulins, including, for example, immunoglobulin domains contained in proteins belonging to the immunoglobulin superfamily, such as major histocompatibility complexes (MHC), CD1, B7, and T cell receptors (TCR). Any immunoglobulin domain can be used in the bispecific antibodies of the present invention.

[0103] In the case of human IgG, CH1 refers to the region having the amino acid sequence from positions 118 to 215 in the EU index. Similarly, CH2 refers to the region having the amino acid sequence from positions 231 to 340 in the EU index of Kabat et al., and CH3 refers to the region having the amino acid sequence from positions 341 to 447 in the EU index of Kabat et al. Between CH1 and CH2 is a highly flexible amino acid region called the hinge region (hereinafter sometimes referred to as the hinge). The hinge region refers to the region having the amino acid sequence from positions 216 to 230 in the EU index of Kabat et al.

[0104] In the case of a human antibody κ chain, CL refers to the region having the amino acid sequence from positions 108 to 214 indicated by Kabat numbering, and in the case of a λ chain, CL refers to the region having the amino acid sequence from positions 108 to 215 indicated by Kabat numbering.

[0105] The Fc region of the bispecific antibody or bispecific antibody fragment of the present invention may be an Fc region derived from an antibody of any animal species, but is preferably an Fc region derived from a human, and particularly preferably an Fc region derived from IgG1.

[0106] In the present invention, an "Fc region capable of binding to an Fc receptor" refers to an Fc region that has the ability to bind to various Fc receptors to an extent sufficient to exert effector activity such as ADCC activity via the Fc receptor. Specific examples include the Fc region of human IgG1, the Fc region of human IgG3, and the Fc region of mouse IgG2a. Furthermore, Fc regions with enhanced Fc receptor-binding ability are also included in the Fc region capable of binding to an Fc receptor.

[0107] The "Fc region capable of binding to an Fc receptor" possessed by the bispecific antibody or bispecific antibody fragment of the present invention is preferably an Fc region capable of binding to FcγR, more preferably an Fc region capable of binding to FcγRIIIA.

[0108] Furthermore, the "Fc region capable of binding to an Fc receptor" possessed by the bispecific antibody or bispecific antibody fragment of the present invention is preferably an "Fc region with enhanced affinity for an Fc receptor."

[0109] An Fc region with enhanced affinity for an Fc receptor may be any Fc region that has enhanced affinity for an Fc receptor compared to the Fc region of a naturally occurring antibody, and examples include an Fc region with modified glycosylation and an Fc region with modified amino acid residues.

[0110] Examples of Fc regions with modified glycosylation include Fc regions with reduced or deleted α1,6 fucose addition, or Fc regions with amino acid modifications (modifications with natural amino acid residues or non-natural amino acid residues) that enhance affinity for Fc receptors.

[0111] Examples of Fc regions containing amino acid residue modifications include Fc regions containing at least one amino acid residue modification selected from the following amino acid modifications in the human IgG1 constant region: P247I, A339D, F243L, R292P, Y300L, P396L, T393A, H433P, S239D, S298A, A330L, I332E, E333A, and K334A, and the amino acid modifications described in Current Opinion in Biotechnology 2009, 20: 685-691.

[0112] In the following, amino acid residue modifications are indicated in the following order: [single-letter code of the amino acid residue before modification], [amino acid position indicated by the EU index], and [single-letter code of the amino acid residue after modification].

[0113] The Fc region or antibody constant region is preferably of the IgG class, and portions of their amino acid sequences may be deleted, added, substituted, and / or inserted. Furthermore, all or partial fragments of the amino acid sequence consisting of CH1, hinge, CH2, and CH3 of the IgG heavy chain may be used in appropriate combination. Furthermore, these amino acid sequences may be partially deleted or their order may be reversed.

[0114] The IgG subclass of the Fc region or antibody constant region used in the bispecific antibody of the present invention is not particularly limited, and may be an Fc region or constant region derived from any of the subclasses IgG1, IgG2, IgG3, and IgG4, although IgG1 is preferred.

[0115] In the case of an IgG1 Fc region or heavy chain constant region, examples include an Fc region or heavy chain constant region comprising a deletion of amino acid residues 216 to 220 in the hinge region and at least one alteration selected from the amino acid residue alterations C220S, H435R, and Y436F; an Fc region or constant region comprising a deletion of amino acid residues 216 to 220 in the hinge region and the amino acid residue alteration C220S; and an Fc region or heavy chain constant region comprising at least one alteration selected from S354C, T366W, Y349C, T366S, L368A, and Y407V.

[0116] The two polypeptide chains constituting the Fc region of the present invention are each referred to as an Fc polypeptide chain, or simply as an Fc chain.

[0117] In the bispecific antibody of the present invention, one CD3-binding domain is linked to any one Fc chain in the Fc region, either directly or via a linker. Thus, the Fc region of the present invention is composed of an "Fc chain" and an "Fc chain with a CD3-binding domain linked to the C-terminus of the Fc chain."

[0118] Furthermore, examples of heavy chain constant regions comprising an Fc region in the present invention include heavy chain constant regions consisting of a "CH polypeptide chain" and a "CH chain having a CD3 domain bound to the C-terminus of the CH polypeptide chain," and constant regions consisting of a "CH polypeptide chain (also referred to as CH1-Fc)" and a "polypeptide chain in which a CL is fused to an Fc chain (hereinafter abbreviated as CL-Fc)," in which a CD3-binding domain is bound to the C-terminus of one of the polypeptide chains. The CD3-binding domain may be bound to either of the two polypeptide chains constituting the above-mentioned Fc region or constant region.

[0119] Alternatively, they can be prepared by linking VH or VL, VH-CH or VL-CL, which constitute the CD3-binding domain, to each of the polypeptide chains contained in the Fc region or heavy chain constant region.

[0120] As described above, the Fc region or constant region contained in the bispecific antibody of the present invention forms a heterodimer because a CD3-binding domain is bound to one of the two polypeptide chains constituting each region. Any modification such as addition, deletion, or substitution may be included as long as this heterodimer structure can be formed.

[0121] Specifically, when two Fc or CH polypeptide chains are involved, heterodimers can be formed by adding appropriate amino acid residue substitutions to the CH3 domain. For example, by adding S354C and T366W amino acid residue substitutions to one of the two Fc polypeptide chains and Y349C, T366S, L368A, and Y407V amino acid residue substitutions to the other, a heterodimer of the Fc or CH domain can be formed [Knobs into Holes modification (Nature Biotechnology, vol. 16: 677-681, 1998)]. This modification is also referred to as a KIH modification.

[0122] When KIH modifications are made to the Fc region of a bispecific antibody of the present invention, Knobs and Holes modifications may be made to either of the two polypeptide chains constituting the Fc region, regardless of which of the two polypeptide chains contains the CD3-binding domain. KIH modifications that do not involve disulfide bond formation in the CH3 region (S354C and Y349C) can also be used. In this case, the amino acid residue T366W is substituted in one of the two Fc polypeptide chains, and the amino acid residues T366S, L368A, and Y407V are substituted in the other.

[0123] Examples of specific amino acid sequences of an Fc region into which the above-mentioned KIH modification has been introduced include, but are not limited to, an Fc region comprising a polypeptide comprising the amino acid sequence set forth in SEQ ID NO: 147 and a polypeptide comprising the amino acid sequence set forth in SEQ ID NO: 148; an Fc region comprising a polypeptide comprising the amino acid sequence set forth in SEQ ID NO: 149 and a polypeptide comprising the amino acid sequence set forth in SEQ ID NO: 150; an Fc region comprising a polypeptide comprising the amino acid sequence set forth in SEQ ID NO: 151 and a polypeptide comprising the amino acid sequence set forth in SEQ ID NO: 152; and an Fc region comprising a polypeptide comprising the amino acid sequence set forth in SEQ ID NO: 153 and a polypeptide comprising the amino acid sequence set forth in SEQ ID NO: 154.

[0124] Furthermore, examples of CH regions consisting of two heterogeneous CH chains (hereinafter referred to as heterogeneous CH regions) include heterogeneous CH regions comprising a CH polypeptide chain and a CL-Fc. More preferred are: a CH polypeptide chain comprising C220S, and a CH region comprising a CL-Fc with amino acid residues 216-220 deleted and C214S; a CH polypeptide comprising C220S, and a CL-Fc with amino acid residues 216-220 deleted, C214S, and H435R; and a CH region comprising a CH polypeptide comprising C220S, and a CL-Fc with amino acid residues 216-220 deleted, C214S, H435R, and Y436F.

[0125] The CD3-binding domain of the present invention may be a single chain or a multimer consisting of multiple polypeptide chains, so long as it has the ability to bind to an antigen against CD3. Examples of CD3 antigen-binding domains include CD3-binding domains comprising three or six CDR sequences of an antibody against CD3, or CD3-binding domains comprising the VH and VL of an antibody against CD3, with Fab, Fab', scFv, dsFv, and VHH being preferred, and Fab and scFv being more preferred. Ligand molecules and receptor molecules for cell surface antigens can also be used similarly.

[0126] When the CD3-binding domain of the present invention is an scFv, scFvs are known to be formed by binding VH, a linker, and VL in that order from the N-terminus (also referred to as VH-linker-VL type or HL type), and to bind VL, a linker, and VH in that order (VL-linker-VH type or LH type), and either of these can be used for the scFv of the present invention.

[0127] When the CD3-binding domain of the present invention is an scFv, modifications may be made to the scFv framework. Examples of modifications include modifications to increase physicochemical stability, modifications to improve productivity, and modifications of amino acids at the interacting surfaces of scFv VH and VL. Specific examples include a modification in which the 44th amino acid (according to Kabat numbering) of scFv VH is substituted with Cys and the 100th amino acid (according to Kabat numbering) of VL is substituted with Cys (CC modification), and a modification in which the 44th amino acid of VH is substituted with Ser and the 100th amino acid of VL is substituted with Glu (SE modification).

[0128] The affinity of the CD3-binding domain of the present invention for CD3 is preferably attenuated. The affinity of the CD3-binding domain is attenuated, for example, by reducing the dissociation constant (K D ) is 6 x 10 -8 The dissociation constant (K D ) is 7 x 10 -8 More preferably, it is 8×10 or more. -8 More preferably, it is 9×10 or more. -8 Furthermore, the CD3-binding domain having reduced affinity for CD3 means that the dissociation constant for CD3 is greater than that of the anti-CD3 monoclonal antibody SP34 or KM14, for example.

[0129] The dissociation constant of the CD3-binding domain of the present invention can be calculated by surface plasmon resonance (SPR) using, for example, Biacore T100 (GE Healthcare).

[0130] In one embodiment, the CD3-binding domain of the present invention has a CDR or VH / VL amino acid sequence that is 80% or more, 85% or more, 86% or more, 87% or more, 88% or more, 89% or more, 90% or more, 91% or more, 92% or more, 93% or more, 94% or more, 95% or more, 96% or more, 97% or more, 98% or more, or 99% or more homologous to the CDR or VH / VL amino acid sequence of the control anti-CD3 monoclonal antibody SP34 or KM14, and the affinity for CD3 is preferably reduced by 10% or more, more preferably 20% or more, and even more preferably 30% or more, compared to SP34 or KM14.

[0131] In one embodiment, the CD3-binding domain of the present invention has an amino acid sequence of CDR or VH / VL that is 80% or more, 85% or more, 86% or more, 87% or more, 88% or more, 89% or more, 90% or more, 91% or more, 92% or more, 93% or more, 94% or more, 95% or more, 96% or more, 97% or more, 98% or more, or 99% or more homologous to the amino acid sequence of the anti-CD3 monoclonal antibody SP34, which is used as a comparison control, and has an affinity that is reduced by 70% or more, preferably 80% or more, more preferably 85% or more, and even more preferably 90% or more, compared to SP34.

[0132] Furthermore, one embodiment of the CD3-binding domain of the present invention includes those in which the amino acid sequence of the CDR or VH / VL has 80% or more, 85% or more, 86% or more, 87% or more, 88% or more, 89% or more, 90% or more, 91% or more, 92% or more, 93% or more, 94% or more, 95% or more, 96% or more, 97% or more, 98% or more, or 99% or more homology to the amino acid sequence of the anti-CD3 monoclonal antibody KM14, which is used as a comparison control, and the affinity is preferably reduced by 10% or more, more preferably 15% or more, even more preferably 30% or more, and even more preferably 40% or more compared to KM14.

[0133] In the present invention, the term "affinity is reduced by A% or more" means that the dissociation constant of the control CD3-binding domain for CD3 is reduced by K D0, where K is the dissociation constant of the test CD3-binding domain with CD3. D1 When this is done, the ratio K D0 / K D1 This means that the resistance is (100-A)% or less.

[0134] In the present invention, an antibody comprising the amino acid sequences of VH and VL of the anti-CD3 monoclonal antibody SP34, or the amino acid sequences of the six CDRs of SP34, is referred to as an SP34 clone. In the present invention, an antibody comprising the amino acid sequences of VH and VL of the anti-CD3 monoclonal antibody KM14, or the amino acid sequences of the six CDRs of KM14, is referred to as a KM14 clone.

[0135] In the present invention, the linker may have any molecular structure that can link the Fc region and the CD3-binding domain, or the Fc region and the disease-related antigen-binding domain, and examples thereof include immunoglobulin domains or fragments thereof and peptide chains, with peptide chains being preferred. Examples of the amino acid sequence of the peptide chain include the so-called GS linker consisting of SGGGG or SGGGG repeats, linkers consisting of EAAAK or repeats thereof, linkers consisting of the amino acid sequence PAPAP, and sequences derived from constant regions such as the hinge region and CH1 domain of antibodies, or modified sequences thereof.

[0136] The bispecific antibody or bispecific antibody fragment of the present invention may have one disease-related antigen-binding domain or two or more disease-related antigen-binding domains.

[0137] The disease-related antigen-binding domain may be bound to the Fc region directly or via a linker. Furthermore, it may be bound to the C-terminus or N-terminus of the Fc region, with the C-terminus being preferred. When one CD3-binding domain is bound to the C-terminus of the Fc region, the CD3-binding domain may be bound to either of the two polypeptide chains that make up the Fc region.

[0138] The disease-related antigen-binding domain of the present invention may be a single chain or a multimer consisting of multiple polypeptide chains, so long as it has antigen-binding ability to a disease-related antigen. Examples of disease-related antigen-binding domains include antigen-binding domains comprising three or six CDR sequences of an antibody against a disease-related antigen, or antigen-binding domains comprising VH and VL of an antibody against a disease-related antigen, and include Fab, Fab', scFv, dsFv, and VHH. However, an antigen-binding domain consisting of VH and VL or Fab is preferred. Ligand molecules or receptor molecules for cell surface antigens can also be used.

[0139] As an antibody against a disease-associated antigen, a monoclonal antibody is preferred, and any antibody with any specificity that can be genetically modified, such as a monoclonal antibody known to specifically recognize and bind to an antigen associated with each disease or a monoclonal antibody already on the market as an antibody pharmaceutical, can be used as the disease-associated antigen-binding domain contained in the anti-CD3 bispecific antibody of the present invention.

[0140] Antibodies or bispecific antibody fragments of the present invention that have one or more amino acid residues deleted, added, substituted, or inserted in the amino acid sequence constituting the bispecific antibody or bispecific antibody fragment thereof and that have activity similar to that of the above-mentioned antibodies or antibody fragments thereof are also encompassed by the bispecific antibody or bispecific antibody fragment of the present invention.

[0141] The number of amino acids to be deleted, substituted, inserted, and / or added is one or more and is not particularly limited, but is a number that can be deleted, substituted, inserted, or added by well-known techniques such as site-directed mutagenesis described in Molecular Cloning, The Second Edition, Cold Spring Harbor Laboratory Press (1989), Current Protocols in Molecular Biology, John Wiley & Sons (1987-1997), Nucleic Acids Research, 10, 6487 (1982), Proc. Natl. Acad. Sci., USA, 79, 6409 (1982), Gene, 34, 315 (1985), Nucleic Acids Research, 13, 4431 (1985), Proc. Natl. Acad. Sci. USA, 82, 488 (1985), etc. For example, the number is usually 1 to several tens, preferably 1 to 20, more preferably 1 to 10, and even more preferably 1 to 5.

[0142] The phrase "one or more amino acid residues have been deleted, substituted, inserted, or added" in the amino acid sequence of the bispecific antibody of the present invention means the following: "one or more amino acid residues have been deleted, substituted, inserted, or added" in any one or more amino acid sequences within the same sequence. Furthermore, deletion, substitution, insertion, and addition may occur simultaneously, and the substituted, inserted, or added amino acid residues may be either naturally occurring or non-naturally occurring.

[0143] Examples of naturally occurring amino acid residues include L-alanine, L-asparagine, L-aspartic acid, L-glutamine, L-glutamic acid, glycine, L-histidine, L-isoleucine, L-leucine, L-lysine, L-arginine, L-methionine, L-phenylalanine, L-proline, L-serine, L-threonine, L-tryptophan, L-tyrosine, L-valine, and L-cysteine.

[0144] Preferred examples of amino acid residues that can be substituted for each other are shown below. Amino acid residues that belong to the same group can be substituted for each other.

[0145] Group A: leucine, isoleucine, norleucine, valine, norvaline, alanine, 2-aminobutanoic acid, methionine, O-methylserine, t-butylglycine, t-butylalanine, cyclohexylalanine Group B: aspartic acid, glutamic acid, isoaspartic acid, isoglutamic acid, 2-aminoadipic acid, 2-aminosuberic acid Group C: asparagine, glutamine D group: lysine, arginine, ornithine, 2,4-diaminobutanoic acid, 2,3-diaminopropionic acid Group E: proline, 3-hydroxyproline, 4-hydroxyproline Group F: serine, threonine, homoserine Group G: phenylalanine, tyrosine

[0146] The bispecific antibodies of the present invention also include antibodies containing any post-translationally modified amino acid residues, such as deletion of lysine residues at the C-terminus of the heavy chain (lysine clipping) and substitution of glutamine residues at the N-terminus of the polypeptide with pyroglutamine (pyroGlu) [Beck et al., Analytical Chemistry, 85, 715-736 (2013)].

[0147] Specific examples of the bispecific antibody or bispecific antibody fragment of the present invention include any one bispecific antibody or bispecific antibody fragment selected from the group consisting of the following (1) to (3): (1) An anti-CD3 bispecific antibody or a bispecific antibody fragment thereof, comprising an Fc region capable of binding to an Fc receptor, a CD3-binding domain bound to the C-terminus of the Fc region, and one or more disease-related antigen-binding domains bound to the N-terminus of the Fc region. (2) An anti-CD3 bispecific antibody or a bispecific antibody fragment thereof, comprising an Fc region capable of binding to an Fc receptor, a CD3-binding domain bound to the C-terminus of the Fc region, and one or more disease-related antigen-binding domains bound to the C-terminus of the Fc region. (3) An anti-CD3 bispecific antibody or a bispecific antibody fragment thereof, comprising an Fc region capable of binding to an Fc receptor, a CD3-binding domain bound to the C-terminus of the Fc region, one or more disease-related antigen-binding domains at the C-terminus of the Fc region, and one or more disease-related antigen-binding domains at the N-terminus of the Fc region.

[0148] One embodiment of the present invention is an anti-CD3 bispecific antibody or bispecific antibody fragment thereof described in (1) above, in which the Fc region forms a heterodimer in which a Knobs into Holes modification (Nature Biotechnology. vol. 16: 677-681, 1998) (also referred to as a KIH modification) has been introduced into the CH3 portion, or a modification obtained by excluding the modifications for forming disulfide bonds (S354C and Y349C) from the KIH modification, and a CD3-binding domain is linked via a linker to either C-terminus of the heterodimer, and a disease-related antigen-binding domain is linked to one or both N-termini of the heterodimer.

[0149] In one embodiment of the present invention, there is provided an anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to (1) above, in which CL and CH1 are bound to the N-terminal ends of two polypeptide chains of the Fc region, respectively, to form a heterodimer consisting of CH1-Fc and CL-Fc, and a CD3-binding domain is bound to the C-terminus of either CH1-Fc or CL-Fc via a linker, (a) a bispecific antibody or a bispecific antibody fragment thereof in which VH and VL are bound to the N-terminus of CH1-Fc and CL-Fc, respectively, to form a single disease-related antigen-binding domain; (b) a bispecific antibody or a bispecific antibody fragment in which VL and VH are linked to the N-terminus of CH1-Fc and CL-Fc, respectively, to form a single disease-related antigen-binding domain; and (c) Bispecific antibodies or bispecific antibody fragments thereof, in which Fab, VHH, and / or scFv are linked to the N-terminus of CH1-Fc and / or CL-Fc and contain one or more disease-related antigen-binding domains.

[0150] In particular, it is preferred that VL and VH, or VH and VL, respectively, are linked to the N-terminus of CH1-Fc and CL-Fc, and it is even more preferred that VH and VL, respectively, are linked to the N-terminus of CH1-Fc and CL-Fc.

[0151] Furthermore, the amino acid sequences of CH1-Fc and CL-Fc preferably contain amino acid modifications that allow heterodimer formation when expressed in cells, specifically, deletion of residues 216-220 and modification of amino acid residues C214S, H435R, and Y436F in CL-Fc, and modification of amino acid residue C220S in CH1-Fc. These modifications may be introduced together with the above-mentioned KIH modifications or modifications excluding modifications for forming disulfide bonds from the KIH modifications.

[0152] Examples of disease-related antigen-binding domains include variable regions consisting of Fab, VH, and VL, VHH, and scFv. Specifically, it is preferred that Fab is bound to both N-termini of the heterodimer as a disease-related antigen-binding domain. Either of the VH-CH1 or VL-CL polypeptide chains forming the Fab may be bound to the N-terminus of the Fc region, but it is more preferred that VH-CH1 is bound.

[0153] One embodiment of the present invention is an anti-CD3 bispecific antibody or bispecific antibody fragment thereof described in (2) above, in which CL and CH1 are bound to the N-terminus of each of the two polypeptide chains of the Fc region to form a heterodimer of CH1-Fc and CL-Fc, and a CD3-binding domain is bound via a linker to the C-terminus of either CH1-Fc or CL-Fc, and in which Fab, VHH, and / or scFv are bound as disease-related antigen-binding domains to the C-terminus of CH1-Fc and / or CL-Fc.

[0154] In this case, the disease-related antigen-binding domain may be bound to CH1-Fc and / or CL-Fc directly, or via a linker, or further to the C-terminus of the CD3-binding domain.

[0155] It is preferable that the amino acid sequence of the heterodimer has been modified so that the heterodimer is formed when expressed in cells. Specifically, it is desirable that CL-Fc has a deletion of amino acid residues 216-220 and amino acid residue modifications of C214S, H435R, and Y436F, and that CH1-Fc has an amino acid residue modification of C220S.

[0156] Another embodiment of the present invention is the anti-CD3 bispecific antibody or bispecific antibody fragment thereof described in (2) above, which comprises Knobs-into-Holes modifications or KIH modifications other than S354C and Y349C in the Fc region, and which comprises one or more disease-associated antigen-binding domains at the C-terminus of the heterodimer forming the Fc region. Examples of the disease-associated antigen-binding domain include Fab, antibody variable regions consisting of VH and VL, VHH, and scFv. The disease-associated antigen-binding domain may be bound to CH1-Fc and / or CL-Fc directly, via a linker, or further C-terminally to the CD3-binding domain.

[0157] One embodiment of the present invention is an anti-CD3 bispecific antibody or bispecific antibody fragment thereof described in (3) above, in which CL and CH1 are bound to the N-terminus of each of the two polypeptide chains of the Fc region to form a heterodimer consisting of CH1-Fc and CL-Fc, and a CD3-binding domain is bound to the C-terminus of either CH1-Fc or CL-Fc via a linker, and which has one or more disease-related antigen-binding domains at the N-terminus of CH1-Fc and CL-Fc, and further has one or more disease-related antigen-binding domains at the C-terminus of CH1-Fc and CL-Fc.

[0158] The disease-related antigen-binding domain bound to the N-terminus of CH1-Fc and / or CL-Fc may be formed by VH and VL (or VL and VH, respectively) binding to the N-terminus of CH1-Fc and CL-Fc, respectively, to form a single disease-related antigen-binding domain, or by Fab, VHH, and / or scFv binding to the N-terminus of at least one of CH1-Fc and CL-Fc to form at least one disease-related antigen-binding domain.

[0159] Furthermore, the disease-related antigen-binding domain preferably comprises a VH and a VL (or a VL and a VH) linked to the N-terminus of CH1-Fc and CL-Fc, respectively, to form a single disease-related antigen-binding domain, and more preferably comprises a VH and a VL linked to the N-terminus of CH1-Fc and CL-Fc, respectively.

[0160] The disease-related antigen-binding domain attached to the C-terminus of CH1-Fc and / or CL-Fc may be attached to these polypeptide chains directly, via a linker, or via a CD3-binding domain. Examples of such disease-related antigen-binding domains include Fab, VHH, and scFv.

[0161] Furthermore, it is preferable that the amino acid sequences of CH1-Fc and CL-Fc have been modified with amino acids that will form a heterodimer when expressed in cells. Specifically, it is preferable that CL-Fc has a deletion of amino acid residues 216-220 and has amino acid residue modifications of C214S, H435R, and Y436F, and that CH1-Fc has an amino acid residue modification of C220S.

[0162] Another embodiment of the present invention is an anti-CD3 bispecific antibody or bispecific antibody fragment thereof described in (3) above, which has Knobs into Holes modifications or KIH modifications other than S354C and Y349C in the Fc region, and which has one or more disease-related antigen-binding domains on the N-terminus of the Fc region heterodimer and further has one or more disease-related antigen-binding domains on the C-terminus of the Fc region heterodimer.

[0163] Examples of disease-related antigen-binding domains include Fab, VHH, and scFv. When Fab is bound to the N-terminus of the heterodimer as the disease-related antigen-binding domain, it does not matter which of the VH-CH1 or VL-CL polypeptide chains forming the Fab is bound to the N-terminus of the Fc domain, but it is more preferable that VH-CH1 is bound.

[0164] Furthermore, the disease-related antigen-binding domain bound to the C-terminus of the heterodimer may be bound directly to the C-terminus of the heterodimer, may be bound via a linker, or may be bound further to the C-terminus of the CD3-binding domain.

[0165] In any of the above embodiments (1) to (3), an Fc region with enhanced affinity for an Fc receptor can also be used as an Fc region capable of binding to an Fc receptor.

[0166] In any of the above embodiments (1) to (3), the VH and VL constituting the CD3-binding domain are linked to each of the two polypeptide chains constituting the Fc domain, thereby producing an anti-CD3 bispecific antibody or a bispecific antibody fragment thereof in which one CD3-binding domain is linked to the C-terminus of the Fc domain.

[0167] The anti-CD3 bispecific antibodies or bispecific antibody fragments of the present invention can also be produced using known heterodimer production techniques, such as CrossMAb technology, BEAT technology, XmAb technology, ART-Ig technology, and Azymetric technology (all described in Nat Rev Drug Discov. 18:585-608, 2019), as well as other heterodimer production techniques described in Nat Rev Drug Discov. 18:585-608, 2019.

[0168] A further preferred embodiment of the present invention includes a bispecific antibody or a bispecific antibody fragment in which the CD3-binding domain is an scFv.

[0169] In a further preferred embodiment of the present invention, the dissociation constant (K D ) is 6 x 10 -8 That's it, 7 x 10 -8 That's it, 8 x 10 -8 or more, or 9 x 10 -8 The bispecific antibody or bispecific antibody fragment described above is an example.

[0170] One embodiment of the present invention includes a bispecific antibody or a bispecific antibody fragment thereof, in which the dissociation constant of the CD3-binding domain with CD3 is greater than that of the anti-CD3 monoclonal antibody SP34 or KM14, which is used as a comparison control.

[0171] One embodiment of the present invention is a bispecific antibody or a bispecific antibody fragment thereof, in which the amino acid sequence of the CDR or VH / VL of the CD3-binding domain has 90% or more homology to the amino acid sequence of the CDR or VH / VL of the CD3-binding domain of the control anti-CD3 monoclonal antibody SP34 or KM14, and which has a 10% or more reduction in affinity compared to the anti-CD3 antibody SP34 or KM14.

[0172] A further preferred embodiment of the present invention is a bispecific antibody that has ADCC activity and ADTC activity against cells expressing a disease-associated antigen, and suppresses the induction of cytokine production upon cytotoxicity of cells expressing the disease-associated antigen.

[0173] A further preferred embodiment of the bispecific antibody or bispecific antibody fragment thereof is one in which the amino acid sequence of the linker between the Fc region and the CD3-binding domain is selected from (SGGGG)n (n = 1 to 3), (EAAAK)n (n = 1 to 3), and PAPAP.

[0174] Specific examples of CD3-binding domains of the present invention include CD3-binding domains in which the amino acid sequences of CDR1 to 3 (HCDR1 to 3) of VH and CDR1 to 3 (LCDR1 to 3) of VL of the CD3-binding domain have 90% or more homology to the amino acid sequences of HCDR1 to 3 and LCDR1 to 3 of any one selected from the following (a) to (h): (a) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 118 to 120, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 121 to 123, respectively. (b) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 83 to 85, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 86 to 88, respectively. (c) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 132, 96, and 97, respectively. (d) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 133, 96, and 97, respectively. (e) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 134, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (f) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 135, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (g) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 136, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (h) HCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 98, 137, and 100, respectively, and LCDRs 1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively.

[0175] Specific examples of CD3-binding domains of the present invention include CD3-binding domains whose VH and VL amino acid sequences have 80% or more homology to any one of the VH and VL amino acid sequences selected from the following (aa) to (rr): (aa) VH comprising the amino acid sequence represented by SEQ ID NO: 124, and VL comprising the amino acid sequence represented by SEQ ID NO: 125 (bb) VH comprising the amino acid sequence represented by SEQ ID NO: 115, and VL comprising the amino acid sequence represented by SEQ ID NO: 116 (cc) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 126 (dd) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 127 (ee) VH comprising the amino acid sequence represented by SEQ ID NO: 128, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ff) VH comprising the amino acid sequence represented by SEQ ID NO: 129, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (gg) VH comprising the amino acid sequence represented by SEQ ID NO: 130, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (hh) VH comprising the amino acid sequence represented by SEQ ID NO: 131, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ii) VH comprising the amino acid sequence represented by SEQ ID NO: 159, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (jj) VH comprising the amino acid sequence represented by SEQ ID NO: 160, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (kk) VH comprising the amino acid sequence represented by SEQ ID NO: 161, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (ll) VH comprising the amino acid sequence represented by SEQ ID NO: 162, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (mm) VH comprising the amino acid sequence represented by SEQ ID NO: 168, and VL comprising the amino acid sequence represented by SEQ ID NO: 180 (nn) VH comprising the amino acid sequence represented by SEQ ID NO: 169, and VL comprising the amino acid sequence represented by SEQ ID NO: 181 (oo) VH comprising the amino acid sequence represented by SEQ ID NO: 170, and VL comprising the amino acid sequence represented by SEQ ID NO: 182 (pp) VH comprising the amino acid sequence represented by SEQ ID NO: 171, and VL comprising the amino acid sequence represented by SEQ ID NO: 183 (qq) VH comprising the amino acid sequence represented by SEQ ID NO: 172, and VL comprising the amino acid sequence represented by SEQ ID NO: 184 (rr) VH comprising the amino acid sequence represented by SEQ ID NO: 173, and VL comprising the amino acid sequence represented by SEQ ID NO: 185

[0176] CD3-binding domains of the present invention include CD3-binding domains that bind to CD3 in competition with a CD3-binding domain comprising any one selected from the above (a) to (h), (aa) to (rr), and the anti-CD3 monoclonal antibodies SP34 and KM14; CD3-binding domains that bind to an epitope on CD3 to which a CD3-binding domain comprising any one selected from the above (a) to (h), (aa) to (rr), and the anti-CD3 monoclonal antibodies SP34 and KM14 binds; and CD3-binding domains that bind to an epitope contained in the epitope to which a CD3-binding domain comprising any one selected from the above (a) to (h), (aa) to (rr), and the anti-CD3 monoclonal antibodies SP34 and KM14 binds.

[0177] An example of an epitope present on a CD3 molecule to which the CD3-binding domain of the present invention binds is the amino acid sequence of human CD3 from the N-terminus to amino acid number 27. Since both anti-CD3 monoclonal antibodies SP34 and KM14 bind to this epitope, preferred CD3 epitopes of the present invention include epitopes containing at least one amino acid residue contained in the amino acid sequence of CD3 from the N-terminus to amino acid number 27, and epitopes having a three-dimensional structure consisting of this amino acid sequence. Antibodies that bind competitively with the anti-CD3 monoclonal antibodies SP34 and / or KM14 and antibodies that bind to the epitopes to which these antibodies bind can be obtained by performing an antibody binding assay using the amino acid sequence of amino acids 1 to 27 of the amino acid sequence of human CD3. Specific examples of the anti-CD3 bispecific antibodies of the present invention include bispecific antibodies comprising the above-mentioned CD3-binding domain.

[0178] The CD3-binding domain of the present invention can also be used for the purpose of producing an anti-CD3 bispecific antibody with reduced cytokine production induction ability, and for the purpose of suppressing the excessive cytokine production induction of an anti-CD3 bispecific antibody.

[0179] The bispecific antibodies or bispecific antibody fragments of the present invention also include bispecific antibodies or bispecific antibody fragments that have effector activity.

[0180] Effector activity refers to antibody-dependent cellular cytotoxicity activity mediated by the Fc region of an antibody, and includes, for example, ADCC activity, complement-dependent cytotoxicity activity (CDC activity), antibody-dependent cellular phagocytosis activity (ADCP activity) by phagocytes such as macrophages and dendritic cells, and the opsonization effect.

[0181] In the present invention, ADCC activity and CDC activity can be measured using known measurement methods [Cancer Immunol. Immunother., 36, 373 (1993)].

[0182] ADCC activity refers to the activity of an antibody bound to an antigen on a target cell, activating immune cells (such as natural killer cells) by binding to the Fc receptor of the immune cell via the Fc region of the antibody, thereby damaging the target cell.

[0183] Fc receptors (FcRs) bind to the Fc region of antibodies, inducing various effector activities upon antibody binding. Each FcR corresponds to a subclass of antibody, with IgG, IgE, IgA, and IgM specifically binding to FcγR, FcεR, FcαR, and FcμR, respectively. Furthermore, there are subtypes of FcγR: FcγRI (CD64), FcγRII (CD32), and FcγRIII (CD16). Each subtype has isoforms: FcγRIA, FcγRIB, FcγRIC, FcγRIIA, FcγRIIB, FcγRIIC, FcγRIIIA, and FcγRIIIB. These different FcγRs are present on different cell types [Annu. Rev. Immunol. 9:457-492(1991)]. In humans, FcγRIIIB is specifically expressed on neutrophils, whereas FcγRIIIA is expressed on monocytes, natural killer cells (NK cells), macrophages, and some T cells. NK cell-dependent ADCC activity is induced via antibody binding to FcγRIIIA.

[0184] CDC activity refers to the activity of antibodies bound to antigens on target cells, activating a series of cascades (complement activation pathways) consisting of complement-related proteins in the blood, resulting in target cell damage. Protein fragments generated by complement activation induce the migration and activation of immune cells. The CDC cascade begins with C1q binding to the Fc region, followed by binding to two serine proteases, C1r and C1s, to form the C1 complex.

[0185] The CDC activity or ADCC activity of the bispecific antibody or bispecific antibody fragment of the present invention against antigen-expressing cells can be evaluated by known measurement methods [Cancer Immunol. Immunother., 36, 373 (1993)].

[0186] Known methods for controlling the effector activity of bispecific antibodies of the present invention include a method of controlling the amount of fucose (also known as core fucose) α-1,6-linked to N-acetylglucosamine (GlcNAc) present at the reducing end of an N-linked complex glycan that is linked to asparagine (Asn) at position 297 in the Fc region (constant region consisting of CH2 and CH3 domains) of the antibody (WO 2005 / 035586, WO 2002 / 31140, and WO 00 / 61739), and a method of controlling the effector activity by modifying amino acid residues in the Fc region of the antibody (WO 00 / 42072).

[0187] The ADCC activity of a bispecific antibody can be increased or decreased by controlling the amount of core fucose bound to the N-linked complex glycans bound to the Fc of the antibody. For example, a bispecific antibody with high ADCC can be obtained by expressing the bispecific antibody using a host cell lacking the α1,6-fucosyltransferase gene, which can reduce the content of core fucose bound to the N-linked complex glycans bound to the Fc of the antibody. On the other hand, a bispecific antibody with low ADCC activity can be obtained by expressing the antibody using a host cell into which the α1,6-fucosyltransferase gene has been introduced, which can increase the content of fucose bound to the N-linked complex glycans bound to the Fc of the bispecific antibody.

[0188] Furthermore, ADCC activity and CDC activity can be increased or decreased by modifying amino acid residues in the Fc region of a bispecific antibody. For example, the CDC activity of a bispecific antibody can be increased by using the amino acid sequence of the Fc region described in U.S. Patent Application Publication No. 2007 / 0148165. Furthermore, ADCC activity or CDC activity can be increased or decreased by making amino acid modifications described in U.S. Patent No. 6,737,056, U.S. Patent No. 7,297,775, U.S. Patent No. 7,317,091, or the like.

[0189] Furthermore, by combining the above methods, a bispecific antibody with controlled effector activity may be obtained.

[0190] The stability of the bispecific antibody of the present invention can be evaluated by measuring the amount of aggregates (oligomers) formed during the purification process or in samples stored under certain conditions. That is, a reduction in the amount of aggregates under the same conditions is evaluated as improved antibody stability. The amount of aggregates can be measured by separating aggregated and non-aggregated antibodies using an appropriate chromatography method, including gel filtration chromatography.

[0191] The productivity of the bispecific antibodies of the present invention can be evaluated by measuring the amount of antibody produced in the culture medium from the antibody-producing cells. More specifically, the productivity can be evaluated by measuring the amount of antibody contained in the culture supernatant obtained by removing the producing cells from the culture medium using an appropriate method such as HPLC or ELISA.

[0192] The bispecific antibody or bispecific antibody fragment of the present invention includes antibody derivatives in which a radioisotope, a low-molecular-weight drug, a high-molecular-weight drug, a protein, an antibody drug, or the like is chemically or genetically bound to the bispecific antibody or bispecific antibody fragment of the present invention.

[0193] The antibody derivatives of the present invention can be produced by chemically binding a radioisotope, a low-molecular-weight drug, a high-molecular-weight drug, an immunostimulant, a protein, or an antibody drug to the N-terminus or C-terminus of the H chain or L chain of the bispecific antibody or bispecific antibody fragment of the present invention, to an appropriate substituent or side chain in the antibody or its antibody fragment, or to a sugar chain in the antibody or its antibody fragment [Introduction to Antibody Engineering, Chijin Shokan (1994)].

[0194] Furthermore, the antibody derivatives of the present invention can be produced by genetic engineering techniques, in which DNA encoding the bispecific antibody or bispecific antibody fragment of the present invention is linked to DNA encoding a desired protein or antibody drug, inserted into an expression vector, and the expression vector is introduced into an appropriate host cell for expression.

[0195] Examples of radioisotopes include: 111 In, 131 I, 125 I, 90 Y, 64 Cu, 99 Tc, 77 Lu or 211 Radioisotopes can be directly bound to antibodies using the chloramine T method or the like. Alternatively, a substance that chelates radioisotopes can be bound to the antibody. Examples of chelating agents include 1-isothiocyanatobenzyl-3-methyldiethylenetriaminepentaacetic acid (MX-DTPA).

[0196] Examples of low-molecular-weight drugs include anticancer drugs such as alkylating agents, nitrosoureas, antimetabolites, antibiotics, plant alkaloids, topoisomerase inhibitors, hormone therapy agents, hormone antagonists, aromatase inhibitors, P-glycoprotein inhibitors, platinum complex derivatives, M-phase inhibitors, and kinase inhibitors [Clinical Oncology, Cancer and Chemotherapy, Inc. (1996)], steroid drugs such as hydrocortisone and prednisone, non-steroid drugs such as aspirin and indomethacin, immunomodulators such as gold thiomalate and penicillamine, immunosuppressants such as cyclophosphamide and azathioprine, and antihistamines such as chlorpheniramine maleate and clemacitin [Inflammation and Anti-inflammatory Therapy, Ishiyaku Publishing Co., Ltd. (1982)].

[0197] Examples of anticancer drugs include amifostine (Ethyol), cisplatin, dacarbazine (DTIC), dactinomycin, mechlorethamine (nitrogen mustard), streptozocin, cyclophosphamide, ifosfamide, carmustine (BCNU), lomustine (CCNU), doxorubicin (Adriamycin), epirubicin, gemcitabine (Gemzar), daunorubicin, procarbazine, mitomatin, and the like. Isin, cytarabine, etoposide, 5-fluorouracil, fluorouracil, vinblastine, vincristine, bleomycin, daunomycin, peplomycin, estramustine, paclitaxel (Taxol), docetaxel (Taxotere), aldesleukin, asparaginase, busulfan, carboplatin, oxaliplatin, nedaplatin, cladribine, camptothecin, 10-hydroxy-7-ene Til-camptothecin (SN38), floxuridine, fludarabine, hydroxyurea, idarubicin, mesna, irinotecan (CPT-11), nogitecan, mitoxantrone, topotecan, leuprolide, megestrol, melphalan, mercaptopurine, hydroxycarbamide, plicamycin, mitotane, pegaspargase, pentostatin, pipobroman, streptozocin, tamoxifen, goserelin , leuprorenin, flutamide, teniposide, testolactone, thioguanine, thiotepa, uracil mustard, vinorelbine, chlorambucil, hydrocortisone, prednisolone, methylprednisolone, vindesine, nimustine, semustine, capecitabine, tomudex, azacitidine, UFT, oxaloplatin, gefitinib (Iressa), imatinib (STI571), erlotinib, FMS-like tyrosine kinase 3 (Flt3) inhibitors, vascular endothelial growth factor receptor (VEGFR) inhibitors, fibroblast growth factor receptor (FGFR) inhibitors, and epidermal growth factor inhibitors such as Tarcevareceptor (EGFR) inhibitors, radicicol, 17-allylamino-17-demethoxygeldanamycin, rapamycin, amsacrine, all-trans retinoic acid, thalidomide, lenalidomide, anastrozole, fadrozole, letrozole, exemestane, gold thiomalate, D-penicillamine, bucillamine, azathioprine, mizoribine, cyclosporine, rapamycin, hydrocortisone, bexarotene (targretin), tamoxifen, dexamethasone, progestins, estrogens, Anas Examples of such antihistamines include Trozole (Arimidex), leuprin, aspirin, indomethacin, celecoxib, azathioprine, penicillamine, gold thiomalate, chlorpheniramine maleate, chlorpheniramine, clemacitin, tretinoin, bexarotene, arsenic, bortezomib, allopurinol, calicheamicin, ibritumomab tiuxetan, targretin, ozogamicin, clarithromycin, leucovorin, ketoconazole, aminoglutethimide, suramin, methotrexate, or maytansinoid or a derivative thereof.

[0198] Methods for binding a low molecular weight drug to the bispecific antibody or bispecific antibody fragment of the present invention include, for example, a method of binding the drug to the amino group of the antibody via glutaraldehyde, or a method of binding the amino group of the drug to the carboxyl group of the antibody via a water-soluble carbodiimide.

[0199] Examples of polymeric drugs include polyethylene glycol (PEG), albumin, dextran, polyoxyethylene, styrene-maleic acid copolymer, polyvinylpyrrolidone, pyran copolymer, and hydroxypropyl methacrylamide. By conjugating these polymeric compounds to the bispecific antibody or antibody fragment of the present invention, effects such as (1) improved stability against various chemical, physical, or biological factors, (2) significant extension of the blood half-life, or (3) elimination of immunogenicity or suppression of antibody production can be expected [Bioconjugate Pharmaceuticals, Hirokawa Shoten (1993)].

[0200] For example, methods for linking PEG to the bispecific antibody of the present invention include reacting with a PEGylation reagent [Bioconjugate Pharmaceuticals, Hirokawa Shoten (1993)]. Examples of PEGylation reagents include a lysine ε-amino group modifying agent (Japanese Patent Laid-Open Publication No. 61-178926), an aspartic acid and glutamic acid carboxyl group modifying agent (Japanese Patent Laid-Open Publication No. 56-23587), or an arginine guanidino group modifying agent (Japanese Patent Laid-Open Publication No. 2-117920).

[0201] The immunostimulant may be a natural product known as an immunoadjuvant, and specific examples of immunostimulating agents include β(1→3) glucan (e.g., lentinan or schizophyllan) or α-galactosylceramide (KRN7000).

[0202] Examples of proteins include cytokines or growth factors that activate immunocompetent cells such as NK cells, macrophages, or neutrophils, or toxin proteins.

[0203] Examples of cytokines or growth factors include interferon (hereinafter referred to as IFN)-α, IFN-β, IFN-γ, interleukin (hereinafter referred to as IL)-2, IL-5, IL-6, IL-10, IL-12, IL-15, IL-18, IL-21, IL-23, tumor necrosis factor (TNF)-α, TNF-β, granulocyte colony-stimulating factor (G-CSF), granulocyte / macrophage colony-stimulating factor (GM-CSF), and macrophage colony-stimulating factor (M-CSF).

[0204] Examples of toxin proteins include ricin, diphtheria toxin, and ONTAK, and also include protein toxins in which mutations have been introduced into the protein to regulate toxicity.

[0205] Fusion antibodies with proteins or antibody drugs can be produced by linking a cDNA encoding the protein to a cDNA encoding the bispecific antibody or antibody fragment of the present invention to construct DNA encoding the fusion antibody, inserting the DNA into an expression vector for prokaryotes or eukaryotes, and introducing the expression vector into a prokaryote or eukaryote to express it.

[0206] When the antibody derivatives are used as detection methods, quantification methods, detection reagents, quantification reagents, or diagnostic agents, agents that bind to the bispecific antibodies or antibody fragments of the present invention include labels used in conventional immunological detection or measurement methods. Examples of labels include enzymes such as alkaline phosphatase, peroxidase, or luciferase, luminescent substances such as acridinium esters or lophine, and fluorescent substances such as fluorescein isothiocyanate (FITC), tetramethylrhodamine isothiocyanate (RITC), Alexa® Fluor 488, and R-phycoerythrin (R-PE).

[0207] The present invention encompasses bispecific antibodies and bispecific antibody fragments having cytotoxic activity such as CDC or ADCC. The CDC or ADCC activity of the bispecific antibodies or bispecific antibody fragments of the present invention against antigen-expressing cells can be assessed by known assay methods [Cancer Immunol. Immunother., 36, 373 (1993)].

[0208] The present invention also relates to a composition comprising a bispecific antibody or a bispecific antibody fragment that specifically recognizes and binds to CD3 and a disease-related antigen, or a therapeutic agent for a disease associated with at least one of CD3 and a disease-related antigen, preferably a disease associated with cells expressing the disease-related antigen, which comprises the bispecific antibody or the bispecific antibody fragment as an active ingredient.

[0209] The disease associated with at least one of CD3 and a disease-associated antigen may be any disease associated with at least one of CD3 and a disease-associated antigen, and examples thereof include malignant tumors and cancers.

[0210] Therapeutic agents containing the bispecific antibody or bispecific antibody fragment of the present invention, or derivatives thereof, may contain only the bispecific antibody or bispecific antibody fragment, or derivatives thereof, as the active ingredient, but are usually preferably provided as pharmaceutical formulations prepared by mixing them with one or more pharmacologically acceptable carriers and using any method known in the technical field of pharmaceuticals.

[0211] The route of administration is preferably the most effective for treatment, and examples thereof include oral administration, or parenteral administration such as oral, intratracheal, rectal, subcutaneous, intramuscular, or intravenous administration, with intravenous administration being preferred.

[0212] Examples of dosage forms include sprays, capsules, tablets, powders, granules, syrups, emulsions, suppositories, injections, ointments, and tapes.

[0213] The dosage or frequency of administration varies depending on the desired therapeutic effect, administration method, treatment period, age, body weight, etc., but is usually 10 μg / kg to 10 mg / kg per day for adults.

[0214] The present invention further relates to a reagent for immunologically detecting or measuring at least one of CD3 and a disease-associated antigen, which comprises the bispecific antibody or the bispecific antibody fragment of the present invention, or a diagnostic agent for a disease in which CD3 and at least one of the disease-associated antigen is involved, preferably a disease in which cells expressing the disease-associated antigen are involved.The present invention also relates to a method for immunologically detecting or measuring at least one of CD3 and at least one of the disease-associated antigen, a method for treating a disease in which CD3 and at least one of the disease-associated antigen is involved, preferably a disease in which cells expressing the disease-associated antigen are involved, or a method for diagnosing a disease in which CD3 and at least one of the disease-associated antigen is involved, preferably a disease in which cells expressing the disease-associated antigen are involved, using the bispecific antibody or the bispecific antibody fragment of the present invention.

[0215] In the present invention, methods for detecting or measuring the amount of at least one of CD3 and disease-associated antigens include any known method, such as immunological detection or measurement methods.

[0216] Immunological detection or measurement methods are methods that use labeled antigens or antibodies to detect or measure the amount of antibodies or antigens, including, for example, radioimmunoassay (RIA), enzyme-linked immunosorbent assay (EIA or ELISA), fluorescent immunoassay (FIA), luminescent immunoassay, Western blotting, and physicochemical techniques.

[0217] By using the bispecific antibody or bispecific antibody fragment of the present invention to detect or measure cells expressing at least one of CD3 and a disease-associated antigen, it is possible to diagnose diseases involving at least one of CD3 and a disease-associated antigen, preferably diseases involving cells expressing a disease-associated antigen.

[0218] Cells expressing at least one of CD3 and a disease-related antigen can be detected by known immunological detection methods, such as immunoprecipitation, immunocytostaining, immunohistostaining, or fluorescent antibody staining. Also included is fluorescent antibody staining using the FMAT8100HTS system (Applied Biosystems).

[0219] In the present invention, biological samples to be used for detecting or measuring at least one of CD3 and a disease-related antigen are not particularly limited, as long as they may contain cells expressing at least one of CD3 and a disease-related antigen, such as tissue cells, blood, plasma, serum, pancreatic juice, urine, feces, tissue fluid, or culture medium.

[0220] A diagnostic agent containing the bispecific antibody of the present invention, the bispecific antibody fragment, or a derivative thereof may contain a reagent for carrying out an antigen-antibody reaction and a reagent for detecting the reaction, depending on the intended diagnostic method. Examples of the reagent for carrying out the antigen-antibody reaction include a buffer and a salt.

[0221] Examples of detection reagents include reagents used in conventional immunological detection or measurement methods, such as a labeled secondary antibody that binds to the bispecific antibody or the bispecific antibody fragment, or a derivative thereof, or a substrate corresponding to the label.

[0222] The following specifically describes a method for producing a bispecific antibody of the present invention, a method for evaluating the activity of the bispecific antibody or the bispecific antibody fragment, and a method for treating and diagnosing a disease using the bispecific antibody or the bispecific antibody fragment.

[0223] 1. How to produce monoclonal antibodies The method for producing monoclonal antibodies of the present invention comprises the following steps: (1) purification of an antigen to be used as an immunogen and / or preparation of cells overexpressing the antigen on their surface, (2) immunizing an animal with the antigen, collecting blood, testing the antibody titer, determining the timing for removing the spleen, etc., and preparing antibody-producing cells, (3) preparation of myeloma cells, (4) cell fusion of antibody-producing cells with myeloma cells, (5) selection of hybridomas producing the desired antibody, (6) isolation (cloning) of monoclonal cells from the hybridomas, (7) optionally culturing hybridomas or raising animals transplanted with hybridomas to produce monoclonal antibodies in large quantities, and (8) examining the physiological activity and antigen-binding specificity of the monoclonal antibodies produced in this manner, or testing their properties as labeling reagents.

[0224] Below, the methods for producing monoclonal antibodies that bind to CD3 and disease-related antigens, which are used to produce bispecific antibodies that bind to CD3 and disease-related antigens in the present invention, are described in detail along the steps described above. The methods for producing the antibodies are not limited to these, and for example, antibody-producing cells other than spleen cells and myelomas can also be used.

[0225] (1) Antigen purification Cells expressing a disease-related antigen or CD3 can be obtained by introducing an expression vector containing a cDNA encoding the full-length or partial length of the disease-related antigen or CD3 into Escherichia coli, yeast, insect cells, animal cells, or the like. Alternatively, CD3 and / or a disease-related antigen can be purified from various cultured human tumor cells or human tissues expressing large amounts of at least one of CD3 and / or a disease-related antigen and used as an antigen. Alternatively, the cultured tumor cells or tissues can be used directly as antigens. Furthermore, synthetic peptides containing a partial sequence of CD3 or a disease-related antigen can be prepared by chemical synthesis methods such as the Fmoc method or the tBoc method and used as antigens.

[0226] The CD3 or disease-associated antigen used in the present invention can be produced by expressing DNA encoding the CD3 or disease-associated antigen in host cells using methods described in Molecular Cloning, A Laboratory Manual, Second Edition, Cold Spring Harbor Laboratory Press (1989) or Current Protocols in Molecular Biology, John Wiley & Sons (1987-1997), for example, by the following method.

[0227] A recombinant vector is prepared by inserting a full-length cDNA containing a portion encoding CD3 or a disease-associated antigen downstream of a promoter in an appropriate expression vector. Instead of the full-length cDNA, a DNA fragment of an appropriate length containing a portion encoding a polypeptide, prepared based on the full-length cDNA, may be used. The resulting recombinant vector can then be introduced into host cells compatible with the expression vector to obtain transformants that produce CD3 or a disease-associated antigen.

[0228] Any expression vector can be used as long as it is capable of autonomous replication in the host cell to be used or of integration into the chromosome and contains a suitable promoter at a position where it can transcribe the DNA encoding CD3 or a disease-associated antigen.

[0229] Any host cells can be used as long as they are capable of expressing the target gene, such as microorganisms belonging to the genus Escherichia, such as E. coli, yeast, insect cells, or animal cells.

[0230] When using prokaryotes such as E. coli as host cells, the recombinant vector is preferably a vector capable of autonomous replication in prokaryotes and containing a promoter, a ribosome binding sequence, DNA containing a portion encoding CD3 or a disease-related antigen, and a transcription termination sequence. Although the recombinant vector does not necessarily require a transcription termination sequence, it is preferable to position the transcription termination sequence immediately downstream of the structural gene. Furthermore, the recombinant vector may also contain a gene that controls the promoter.

[0231] As the recombinant vector, it is preferable to use a plasmid in which the distance between the Shine-Dalgarno sequence, which is a ribosome binding sequence, and the initiation codon is adjusted to an appropriate distance (for example, 6 to 18 bases).

[0232] Furthermore, bases in the base sequence of the DNA encoding the CD3 or disease-related antigen can be substituted to provide codons optimal for expression in the host, thereby improving the production rate of the desired CD3 or disease-related antigen.

[0233] Any expression vector can be used as long as it can function in the host cell to be used. Examples of such vectors include pBTrp2, pBTac1, and pBTac2 (all manufactured by Roche Diagnostics), pKK233-2 (manufactured by Pharmacia), pSE280 (manufactured by Invitrogen), pGEMEX-1 (manufactured by Promega), pQE-8 (manufactured by Qiagen), pKYP10 (Japanese Patent Application Laid-Open No. 58-110600), pKYP200 [Agricultural Biological Chemistry, 48, 669 (1984)], pLSA1 [Agric. Biol. Chem., 53, 277 (1989)], pGEL1 [Proc. Natl. Acad. Sci. USA, 82, 4306 (1985)], and pBluescript II. SK(-) (Stratagene), pTrs30 [prepared from E. coli JM109 / pTrS30 (FERM BP-5407)], pTrs32 [prepared from E. coli JM109 / pTrS32 (FERM BP-5408)], pGHA2 [prepared from E. coli IGHA2 (FERM BP-400), Japanese Patent Application Laid-Open No. 60-221091], pGKA2 [prepared from E. coli IGKA2 (FERM BP-6798), Japanese Patent Application Laid-Open No. 60-221091], pTerm2 (U.S. Pat. Nos. 4,686,191, 4,939,094, and 5,160,735), pSupex, pUB110, pTP5, pC194, and pEG400 [J. Bacteriol., 172, 2392 (1990)], pGEX (Pharmacia), pET system (Novagen), or pME18SFL3 (Toyobo).

[0234] The promoter may be any that functions in the host cell to be used. Examples include promoters derived from E. coli or phages, such as the trp promoter (Ptrp), lac promoter, PL promoter, PR promoter, or T7 promoter. Other examples include artificially designed and modified promoters, such as a tandem promoter with two Ptrp promoters in tandem, the tac promoter, the lacT7 promoter, or the letI promoter.

[0235] Examples of host cells include E. coli XL1-Blue, E. coli XL2-Blue, E. coli DH1, E. coli MC1000, E. coli KY3276, E. coli W1485, E. coli JM109, E. coli HB101, E. coli No. 49, E. coli W3110, E. coli NY49, and E. coli DH5α.

[0236] Any method for introducing a recombinant vector into a host cell can be used as long as it is a method for introducing DNA into the host cell to be used, such as a method using calcium ions [Proc. Natl. Acad. Sci. USA, 69, 2110(1972); Gene, 17, 107(1982); Molecular & General Genetics, 168, 111(1979)].

[0237] When animal cells are used as hosts, any expression vector that functions in animal cells can be used, such as pcDNAI (Invitrogen), pcDM8 (Funakoshi), pAGE107 [JP Patent Publication No. 3-22979; Cytotechnology, 3, 133 (1990)], pAS3-3 (JP Patent Publication No. 2-227075), pCDM8 [Nature, 329, 840 (1987)], pcDNAI / Amp (Invitrogen), pcDNA3.1 (Invitrogen), pREP4 (Invitrogen), pAGE103 [J. Biochemistry, 101, 1307 (1987)], pAGE210, pME18SFL3, or pKANTEX93 (WO 97 / 10354).

[0238] Any promoter that can function in animal cells can be used, including, for example, the promoter of the immediate early (IE) gene of cytomegalovirus (CMV), the SV40 early promoter, a retrovirus promoter, a metallothionein promoter, a heat shock promoter, an SRα promoter, or a promoter or enhancer of Moloney murine leukemia virus. The enhancer of the IE gene of human CMV may also be used together with the promoter.

[0239] Examples of host cells include human Burkitt's lymphoma cells Namalwa, African green monkey kidney-derived cells COS, Chinese hamster ovary-derived cells CHO, and human leukemia cells HBT5637 (Japanese Patent Laid-Open Publication No. 63-000299).

[0240] Any method for introducing DNA into animal cells can be used to introduce a recombinant vector into a host cell, such as electroporation [Cytotechnology, 3, 133 (1990)], calcium phosphate method (Japanese Patent Laid-Open Publication No. 2-227075), or lipofection [Proc. Natl. Acad. Sci. USA, 84, 7413 (1987)].

[0241] CD3 or a disease-related antigen can be produced by culturing in a medium a transformant derived from a microorganism or animal cell harboring a recombinant vector incorporating DNA encoding CD3 or a disease-related antigen, thereby producing and accumulating at least one of the CD3 and the disease-related antigen in the culture, and then collecting the CD3 or the disease-related antigen from the culture. The method for culturing the transformant in a medium can be performed according to a conventional method used for culturing hosts.

[0242] When expressed in eukaryotic cells, CD3 or disease-associated antigens can be obtained that have sugar or glycosylated chains attached.

[0243] When culturing a microorganism transformed with a recombinant vector using an inducible promoter, an inducer may be added to the medium as needed. For example, isopropyl-β-D-thiogalactopyranoside may be added to the medium when culturing a microorganism transformed with a recombinant vector using the lac promoter, and indoleacrylic acid may be added to the medium when culturing a microorganism transformed with a recombinant vector using the trp promoter.

[0244] Examples of media for culturing transformants obtained using animal cells as hosts include commonly used RPMI 1640 medium [The Journal of the American Medical Association, 199, 519 (1967)], Eagle's MEM medium [Science, 122, 501 (1952)], Dulbecco's Modified MEM medium [Virology, 8, 396 (1959)], 199 medium [Proc. Soc. Exp. Biol. Med., 73, 1 (1950)], Iscove's Modified Dulbecco's Medium (IMDM) medium, and media supplemented with fetal bovine serum (FBS) or other ingredients. Culture is typically performed for 1 to 7 days under conditions such as pH 6-8, 30-40°C, and 5% CO2. Antibiotics such as kanamycin and penicillin may be added to the medium during culture, if necessary.

[0245] In addition to direct expression, methods such as secretory production and fusion protein expression can be used to express genes encoding CD3 or disease-associated antigens [Molecular Cloning, A Laboratory Manual, Second Edition, Cold Spring Harbor Laboratory Press (1989)]. Methods for producing CD3 or disease-associated antigens include, for example, production within host cells, secretion outside host cells, and production on the outer membrane of host cells. An appropriate method can be selected by changing the host cells used and the structure of the CD3 or disease-associated antigen to be produced.

[0246] For example, an antigen fusion protein can be produced by preparing DNA in which DNA encoding the amino acid sequence of the extracellular region is linked to DNA encoding the Fc region of an antibody, DNA encoding glutathione S-transferase (GST), DNA encoding a FLAG tag, or DNA encoding a histidine tag, and expressing and purifying the DNA. Specific examples include Fc fusion proteins in which the extracellular region of CD3 or a disease-associated antigen is linked to the Fc region of human IgG, and fusion proteins of the extracellular region of CD3 or a disease-associated antigen with glutathione S-transferase (GST).

[0247] When CD3 or a disease-associated antigen is produced within a host cell or on its outer membrane, CD3 or a disease-associated antigen can be actively secreted outside the host cell by using the method of Paulson et al. [J. Biol. Chem., 264, 17619 (1989)], the method of Rowe et al. [Proc. Natl. Acad. Sci., USA, 86, 8227 (1989), Genes Develop., 4, 1288 (1990)], the method described in Japanese Patent Application Laid-Open No. 05-336963, or International Publication No. 94 / 23021. The amount of CD3 or a disease-associated antigen produced can also be increased by using a gene amplification system using a dihydrofolate reductase gene or the like (Japanese Patent Application Laid-Open No. 2-227075).

[0248] The produced CD3 or disease-associated antigen can be isolated and purified, for example, as follows.

[0249] If CD3 or a disease-related antigen is expressed intracellularly in a dissolved state, the cells are collected by centrifugation after the culture is completed, suspended in an aqueous buffer, and disrupted using an ultrasonicator, French press, Manton-Gaulin homogenizer, or Dynomill to obtain a cell-free extract. The supernatant obtained by centrifugation of the cell-free extract can be purified by conventional protein isolation and purification methods, such as solvent extraction, salting out with ammonium sulfate, desalting, precipitation with organic solvents, anion exchange chromatography using resins such as diethylaminoethyl (DEAE)-Sepharose or DIAION HPA-75 (Mitsubishi Chemical), cation exchange chromatography using resins such as S-Sepharose FF (Pharmacia), hydrophobic chromatography using resins such as butyl-Sepharose or phenyl-Sepharose, gel filtration using molecular sieves, affinity chromatography, chromatofocusing, or electrophoresis such as isoelectric focusing, either alone or in combination.

[0250] If CD3 or a disease-related antigen is expressed as an insoluble body within the cells, the cells are collected and disrupted as described above, and centrifuged to collect the insoluble CD3 or disease-related antigen as a precipitate fraction. The collected insoluble CD3 or disease-related antigen is solubilized with a protein denaturant. The solubilized solution is diluted or dialyzed to restore the CD3 or disease-related antigen to its normal three-dimensional structure, after which the purified polypeptide protein can be obtained by the same isolation and purification method as described above.

[0251] When CD3 or a disease-related antigen, or a derivative thereof such as a glycosylated form thereof, is secreted extracellularly, the CD3 or disease-related antigen, or a derivative thereof such as a glycosylated form thereof, can be recovered from the culture supernatant. The culture supernatant can be treated by a technique such as centrifugation as described above to obtain a soluble fraction, and purified protein can be obtained from the soluble fraction by the same isolation and purification method as described above.

[0252] The CD3 or disease-related antigen used in the present invention can also be produced by chemical synthesis methods such as the Fmoc or tBoc method, using a peptide synthesizer manufactured by, for example, Advanced ChemTech, Perkin-Elmer, Pharmacia, Protein Technology Instruments, Synthecel-Vega, Perceptive, or Shimadzu Corporation.

[0253] (2) Preparation of antibody-producing cells Animals such as mice, rats, hamsters, rabbits, cattle, or alpacas are immunized with the antigen obtained in (1), and antibody-producing cells are collected from the animal's spleen, lymph nodes, or peripheral blood. Examples of animals that can be immunized include transgenic mice that produce human-derived antibodies, as described in Tomizuka et al., Proc Natl Acad Sci USA, 97, 722, (2000), and conditional knockout mice for CD3 or disease-related antigens to enhance immunogenicity.

[0254] Immunization is performed by administering the antigen together with an appropriate adjuvant, such as Freund's complete adjuvant or aluminum hydroxide gel and Bordetella pertussis vaccine. The immunogen may be administered by subcutaneous, intraperitoneal, intravenous, intradermal, intramuscular, or footpad injection, with intraperitoneal, footpad, or intravenous injection being preferred. When the antigen is a partial peptide, it is conjugated with a carrier protein such as BSA (bovine serum albumin) or KLH (keyhole limpet hemocyanin) and used as the immunogen.

[0255] After the first administration, antigen is administered 5 to 10 times every 1 to 2 weeks. Blood is collected from the retinal venous plexus 3 to 7 days after each administration, and the serum antibody titer is measured using enzyme immunoassay [Antibodies - A Laboratory Manual, Cold Spring Harbor Laboratory (1988)] or similar. If animals whose serum shows sufficient antibody titers against the antigen used for immunization are used as a source of antibody-producing cells for fusion, the effectiveness of subsequent procedures can be enhanced.

[0256] Three to seven days after the final administration of antigen, tissues containing antibody-producing cells, such as the spleen, are removed from the immunized animal, and the antibody-producing cells are collected. Antibody-producing cells are plasma cells and their precursor cells, lymphocytes. They can be obtained from any part of the individual, typically the spleen, lymph nodes, bone marrow, tonsils, peripheral blood, or an appropriate combination of these. However, spleen cells are most commonly used. When using spleen cells, the spleen is disaggregated and then centrifuged, followed by removal of red blood cells to obtain antibody-producing cells for fusion.

[0257] (3) Myeloma preparation process Myeloma cells derived from mammals such as mice, rats, guinea pigs, hamsters, rabbits, or humans that do not produce autoantibodies can be used. Generally, however, established cell lines obtained from mice are used, such as the 8-azaguanine-resistant mouse (BALB / c-derived) myeloma cell lines P3-X63Ag8-U1 (P3-U1) [Current Topics in Microbiology and Immunology, 18, 1 (1978)], P3-NS1 / 1-Ag41 (NS-1) [European J. Immunology, 6, 511 (1976)], SP2 / 0-Ag14 (SP-2) [Nature, 276, 269 (1978)], and P3-X63-Ag8653 (653) [J. Immunology, 123, 1548 (1978)]. (1979)] or P3-X63-Ag8(X63) [Nature, 256, 495(1975)] is used. The cell line is subcultured in an appropriate medium, such as 8-azaguanine medium [RPMI-1640 medium supplemented with glutamine, 2-mercaptoethanol, gentamicin, FCS, and 8-azaguanine], Iscove's Modified Dulbecco's Medium (hereinafter referred to as "IMDM"), or Dulbecco's Modified Eagle Medium (hereinafter referred to as "DMEM"). Three to four days before cell fusion, the above cell line is subcultured in normal medium (e.g., DMEM medium containing 10% FCS), and on the day of fusion, 2 x 10 7 Ensure that there are at least 100 cells.

[0258] (4)Cell fusion The antibody-producing cells for fusion obtained in (2) and the myeloma cells obtained in (3) are thoroughly washed with Minimum Essential Medium (MEM) or PBS (1.83 g disodium phosphate, 0.21 g monopotassium phosphate, 7.65 g sodium chloride, 1 L distilled water, pH 7.2), mixed at a ratio of 5:1 to 10:1 antibody-producing cells for fusion to myeloma cells, centrifuged, and the supernatant removed. After thoroughly loosening the precipitated cell clumps, a mixture of polyethylene glycol-1000 (PEG-1000), MEM medium, and dimethyl sulfoxide is added with stirring at 37°C. 1–2 mL of MEM medium is added several times every 1–2 minutes, followed by the addition of MEM medium until the total volume reaches 50 mL. After centrifugation, the supernatant is removed, and the precipitated cell clumps are gently loosened and gently suspended in HAT medium (normal medium supplemented with hypoxanthine, thymidine, and aminopterin). This suspension is cultured in a 5% CO2 incubator at 37°C for 7 to 14 days.

[0259] Cell fusion can also be performed using the following method. Spleen cells and myeloma cells are thoroughly washed with serum-free medium (e.g., DMEM) or phosphate-buffered saline (hereinafter referred to as "phosphate buffer"), mixed so that the spleen cells and myeloma cells have a cell number ratio of approximately 5:1–10:1, and centrifuged. The supernatant is removed, and the precipitated cell clumps are thoroughly loosened. Then, 1 mL of serum-free medium containing 50% (w / v) polyethylene glycol (molecular weight 1000–4000) is added dropwise while stirring. Then, 10 mL of serum-free medium is slowly added, followed by centrifugation. The supernatant is again discarded, and the precipitated cells are suspended in an appropriate amount of normal medium containing hypoxanthine-aminopterin-thymidine (HAT) solution and human interleukin-2 (IL-2) (hereinafter referred to as HAT medium). The suspension is then dispensed into each well of a culture plate (hereinafter referred to as "plate") and cultured at 37°C in the presence of 5% carbon dioxide for approximately two weeks. Supplement HAT medium as needed during the incubation.

[0260] (5) Selection of hybridoma population If the myeloma cells used for fusion are 8-azaguanine-resistant, i.e., hypoxanthine guanine phosphoribosyltransferase (HGPRT)-deficient, the unfused myeloma cells and the fused cells between myeloma cells cannot survive in HAT medium. On the other hand, fused cells between antibody-producing cells and hybridomas between antibody-producing cells and myeloma cells can survive in HAT medium, but the fused cells between antibody-producing cells eventually reach the end of their lifespan. Therefore, by continuing to culture in HAT medium, only hybridomas between antibody-producing cells and myeloma cells will survive, and hybridomas can be obtained.

[0261] For hybridomas that have grown into colonies, the medium is changed from HAT medium to a medium lacking aminopterin (hereinafter referred to as HT medium). A portion of the culture supernatant is then collected, and antibody-producing hybridomas can be selected using the antibody titer measurement method described below. Various known techniques for measuring antibody titers include radioisotope immunoassay (RIA), enzyme-linked immunosorbent assay (ELISA), fluorescent antibody assay, and passive hemagglutination assay. From the viewpoints of detection sensitivity, speed, accuracy, and the possibility of automating the procedure, RIA or ELISA are preferred.

[0262] Hybridomas that are found to produce the desired antibody by measuring the antibody titer are transferred to another plate and cloned. Examples of cloning methods include the limiting dilution method, in which the cells are diluted and cultured so that one cell is contained in each well of the plate, the soft agar method, in which the cells are cultured in a soft agar medium and colonies are recovered, the method of isolating a single cell using a micromanipulator, and the method of isolating a single cell using a cell sorter.

[0263] For wells in which an antibody titer is observed, cloning is repeated 2 to 4 times, for example, by limiting dilution, and those in which a stable antibody titer is observed are selected as hybridoma strains that produce monoclonal antibodies against CD3 or disease-related antigens.

[0264] (6) Preparation of monoclonal antibodies The monoclonal antibody-producing hybridomas obtained in (5) are injected intraperitoneally into pristane-treated mice or nude mice (8-10 weeks old) for two weeks. The hybridomas develop into ascites tumors within 10-21 days. Ascites fluid is collected from these mice, centrifuged to remove solids, salted out with 40-50% ammonium sulfate, and purified using caprylic acid precipitation, DEAE-Sepharose column, protein A column, or gel filtration column. The IgG or IgM fractions are collected and used as purified monoclonal antibodies. Alternatively, the hybridomas can be grown in the peritoneal cavity of the same strain of mice (e.g., BALB / c or Nu / Nu mice), rats, guinea pigs, hamsters, or rabbits to obtain ascites containing large amounts of monoclonal antibodies that bind to CD3 or disease-related antigens.

[0265] The monoclonal antibody-producing hybridomas obtained in (5) are cultured in RPMI 1640 medium supplemented with 10% FBS, the supernatant removed by centrifugation, and the cells are suspended in GIT medium or Hybridoma SFM medium supplemented with 5% Daigo GF21, and cultured for 3 to 7 days by flask culture, spinner culture, or back culture. The resulting cell suspension is centrifuged, and the supernatant is purified using a Protein A or Protein G column to collect the IgG fraction, yielding purified monoclonal antibodies. A convenient method for purification is to use a commercially available monoclonal antibody purification kit (e.g., MAbTrap GII kit; Amersham Pharmacia Biotech).

[0266] The antibody subclass is determined by enzyme immunoassay using a subclass typing kit. Protein quantity is quantified by the Lowry method and absorbance at 280 nm [1.4 (OD 280 ) = immunoglobulin 1 mg / mL].

[0267] (7) Binding assay of monoclonal antibodies against CD3 or disease-associated antigens The anti-CD3 bispecific antibody of the present invention comprises a CD3-binding domain with reduced affinity for CD3, and the CD3-binding domain can be established by using the anti-CD3 monoclonal antibody SP34 or KM14 as a control in the above-mentioned binding assay. For example, a desired CD3-binding domain can be established by screening an anti-CD3 monoclonal antibody library for antibodies with affinity 10% or more lower than that of anti-CD3 monoclonal antibody SP34 or KM14. Alternatively, a desired CD3-binding domain can be established by preparing an antibody library in which random mutations have been introduced into the amino acid sequences of the CDR or FR regions based on the amino acid sequences of the CDR or VH / VL of anti-CD3 monoclonal antibody SP34 or KM14, and screening the library for antibodies with affinity 10% or more lower than that of anti-CD3 monoclonal antibody SP34 or KM14. The binding activity of a monoclonal antibody to CD3 or a disease-associated antigen can be measured by a binding assay system such as the Ouchterlony method, ELISA, RIA, flow cytometry (FCM), or surface plasmon resonance (SPR).

[0268] Although the Ouchterlony method is simple, it requires a concentration step when the antibody concentration is low.On the other hand, when using the ELISA or RIA method, the culture supernatant is directly reacted with an antigen-adsorbed solid phase, and antibodies corresponding to various immunoglobulin isotypes and subclasses are used as secondary antibodies, which allows the isotype and subclass of the antibody to be identified and the binding activity of the antibody to be measured.

[0269] In a specific example, purified or partially purified recombinant CD3 or a disease-related antigen is adsorbed onto the solid surface of a 96-well ELISA plate. The unadsorbed solid surface is then blocked with a protein unrelated to the antigen, such as bovine serum albumin (BSA). After washing the ELISA plate with phosphate buffered saline (PBS) and PBS containing 0.05% Tween 20 (Tween-PBS), serially diluted primary antibodies (e.g., mouse serum or culture supernatant) are added to bind to the antigen immobilized on the plate. Next, an anti-immunoglobulin antibody labeled with biotin, an enzyme (horse radish peroxidase [HRP], alkaline phosphatase [ALP], etc.), a chemiluminescent substance, or a radioactive compound is added as the secondary antibody, allowing the secondary antibody to react with the plate-bound primary antibody. After thorough washing with Tween-PBS, the secondary antibody is reacted according to the secondary antibody's labeling substance to select monoclonal antibodies that specifically react with the target antigen.

[0270] The FCM method can measure the binding activity of an antibody to an antigen-expressing cell [Cancer Immunol. Immunother., 36, 373(1993)]. The binding of an antibody to a membrane protein antigen expressed on the cell membrane means that the antibody recognizes and binds to the three-dimensional structure of the naturally occurring antigen.

[0271] Examples of SPR include kinetic analysis using Biacore. For example, a Biacore T100 is used to measure the binding kinetics between an antigen and a test substance, and the results are analyzed using the instrument's included analysis software. A specific example of the procedure involves immobilizing an anti-mouse IgG antibody to a CM5 sensor chip using the amine coupling method, then injecting an appropriate amount of a test substance, such as hybridoma culture supernatant or a purified monoclonal antibody, to allow binding. Then, multiple known concentrations of the antigen are injected to measure binding and dissociation. The resulting data are then analyzed using the instrument's included software based on a 1:1 binding model to obtain various parameters. Alternatively, CD3 or a disease-related antigen is immobilized on a sensor chip using, for example, the amine coupling method, and then multiple known concentrations of purified monoclonal antibody are injected to measure binding and dissociation. The resulting data are then analyzed using the instrument's included software based on a 1:1 binding model or a bivalent binding model to obtain various parameters.

[0272] Furthermore, in the present invention, antibodies that bind to CD3 or disease-related antigens in competition with antibodies against CD3 or disease-related antigens can be selected by reacting the test antibody in the presence of the test antibody in the above-mentioned binding assay system. That is, by screening for antibodies whose binding to the antigen is inhibited when the test antibody is added, antibodies that compete with the antibody obtained above for binding to CD3 or disease-related antigens can be obtained.

[0273] (8) Identification of epitopes for monoclonal antibodies against CD3 or disease-related antigens In the present invention, the epitope recognized and bound by an antibody can be identified as follows.

[0274] For example, if a partially deleted antigen, a mutant in which different amino acid residues are altered between species, or a mutant in which a specific domain is altered is prepared, and the reactivity of an antibody against the deleted antigen or mutant is reduced, it becomes clear that the deleted or altered amino acid site is the epitope of the antibody. Such partially deleted antigens and mutants can be obtained as secreted proteins using appropriate host cells, such as E. coli, yeast, plant cells, or mammalian cells, or can be expressed on the cell membrane of the host cell to prepare antigen-expressing cells. In the case of membrane-type antigens, it is preferable to express them on the cell membrane of the host cell so that the antigen's three-dimensional structure is maintained. Alternatively, synthetic peptides mimicking the primary or three-dimensional structure of the antigen can be prepared and antibody reactivity can be confirmed. Synthetic peptides can be prepared by using known peptide synthesis techniques to prepare various partial peptides of the molecule.

[0275] For example, chimeric proteins can be prepared by appropriately combining the domains constituting the extracellular regions of human and mouse CD3 or disease-related antigens, and antibody reactivity with the chimeric proteins can be confirmed to identify the epitope of the antibody. Subsequently, various oligopeptides of the corresponding portions or variants of the peptides can be synthesized using oligopeptide synthesis techniques well known to those skilled in the art, and the reactivity of antibodies with the peptides can be confirmed to identify the epitope. A convenient method for obtaining a wide variety of oligopeptides is to use commercially available kits [e.g., the SPOTs kit (Genosys Biotechnologies) and a series of multipin peptide synthesis kits (Chiron) using the multipin synthesis method].

[0276] Antibodies that bind to the same epitope as an antibody that binds to CD3 or a disease-related antigen can be obtained by identifying the epitope of the antibody obtained in the above-mentioned binding assay system, and preparing a partial synthetic peptide of the epitope, a synthetic peptide mimicking the three-dimensional structure of the epitope, or a recombinant form of the epitope, and then immunizing the subject.

[0277] For example, if the epitope is a membrane protein, a recombinant fusion protein can be produced in which the entire extracellular region or a portion of the extracellular domain is linked to an appropriate tag, such as a FLAG tag, a histidine tag, a GST protein, or an antibody Fc region, and the recombinant protein can be used for immunization, thereby enabling more efficient production of antibodies specific to the epitope.

[0278] 2. Production of recombinant antibodies Examples of recombinant antibody production are outlined in PJ Delves, ANTIBODY PRODUCTION ESSENTIAL TECHNIQUES, 1997 Wiley; P. Shepherd and C. Dean, Monoclonal Antibodies, 2000 Oxford University Press; and JW Goding, Monoclonal Antibodies: principles and practice, 1993 Academy Press. The following describes methods for producing chimeric, humanized, and human antibodies. Genetically recombinant mouse, rat, hamster, and rabbit antibodies can also be produced using similar methods.

[0279] (1) Isolation of cDNA encoding the V region of a monoclonal antibody from a hybridoma cDNA encoding VH and VL of a monoclonal antibody can be obtained, for example, as follows.

[0280] First, mRNA is extracted from a hybridoma producing a monoclonal antibody, and cDNA is synthesized. The synthesized cDNA is then cloned into a vector such as a phage or plasmid to prepare a cDNA library. From the library, recombinant phages or recombinant plasmids carrying cDNA encoding VH or VL are isolated using DNA encoding the C or V region of the antibody as a probe. The entire nucleotide sequence of VH or VL in the isolated recombinant phage or recombinant plasmid is determined, and the entire amino acid sequence of VH or VL is deduced from the nucleotide sequence.

[0281] Non-human animals used to prepare hybridomas include mice, rats, hamsters, rabbits, etc., but any animal can be used as long as it is possible to prepare hybridomas.

[0282] Total RNA from hybridomas can be prepared by the guanidine thiocyanate-cesium trifluoroacetate method [Methods in Enzymol., 154, 3 (1987)] or by using a kit such as the RNA Easy Kit (Qiagen).

[0283] To prepare mRNA from total RNA, oligo(dT)-immobilized cellulose column method [Molecular Cloning, A Laboratory Manual, Second Edition, Cold Spring Harbor Laboratory Press (1989)] or Oligo-dT30 <super>Kits such as the mRNA Purification Kit (Takara Bio) are used. Alternatively, mRNA can be prepared using kits such as the Fast Track mRNA Isolation Kit (Invitrogen) or QuickPrep mRNA Purification Kit (Pharmacia).

[0284] cDNA synthesis and cDNA library construction can be performed using known methods [Molecular Cloning, A Laboratory Manual, Second Edition, Cold Spring Harbor Laboratory Press (1989), Current Protocols in Molecular Biology, Supplement 1, John Wiley & Sons (1987-1997)] or kits such as SuperScript Plasmid System for cDNA Synthesis and Plasmid Cloning (Invitrogen) or ZAP-cDNA Synthesis Kit (Stratagene).

[0285] When preparing a cDNA library, any vector can be used to incorporate cDNA synthesized using mRNA extracted from a hybridoma as a template, as long as it can incorporate said cDNA.

[0286] For example, ZAP Express [Strategies, 5, 58 (1992)], pBluescript II SK(+) [Nucleic Acids Research, 17, 9494 (1989)], λZAPII (Stratagene), λgt10, λgt11 [DNA Cloning: A Practical Approach, I, 49 (1985)], Lambda BlueMid (Clontech), λExCell, pT7T3-18U (Pharmacia), pcD2 [Mol. Cell. Biol., 3, 280 (1983)], or pUC18 [Gene, 33, 103 (1985)] may be used.

[0287] Any E. coli strain capable of introducing, expressing, and maintaining a cDNA library constructed using a phage or plasmid vector can be used for the introduction of the cDNA library, such as XL1-Blue MRF' [Strategies, 5, 81 (1992)], C600 [Genetics, 39, 440 (1954)], Y1088, Y1090 [Science, 222, 778 (1983)], NM522 [J. Mol. Biol., 166, 1 (1983)], K802 [J. Mol. Biol., 16, 118 (1966)], or JM105 [Gene, 38, 275 (1985)].

[0288] To select cDNA clones encoding the VH or VL of non-human antibodies from a cDNA library, methods such as colony hybridization using isotope- or fluorescently-labeled probes or plaque hybridization [Molecular Cloning, A Laboratory Manual, Second Edition, Cold Spring Harbor Laboratory Press (1989)] are used.

[0289] Alternatively, cDNA encoding VH or VL can be prepared by preparing primers and using cDNA synthesized from mRNA or a cDNA library as a template, and performing the Polymerase Chain Reaction method (hereinafter referred to as PCR; Molecular Cloning, A Laboratory Manual, Second Edition, Cold Spring Harbor Laboratory Press (1989); Current Protocols in Molecular Biology, Supplement 1, John Wiley & Sons (1987-1997)).

[0290] The selected cDNA is digested with an appropriate restriction enzyme or the like, and then cloned into a plasmid such as pBluescript SK(-) (Stratagene), and the nucleotide sequence of the cDNA is determined by a commonly used nucleotide sequence analysis method, for example, by the dideoxy method [Proc. Natl. Acad. Sci. USA, 74, 5463 (1977)], followed by analysis using an automatic nucleotide sequence analyzer such as an ALF DNA sequencer (Pharmacia).

[0291] The entire amino acid sequences of VH and VL are deduced from the determined total nucleotide sequence and compared with the entire amino acid sequences of VH and VL of known antibodies [Sequences of Proteins of Immunological Interest, U.S. Dept. Health and Human Services (1991)] to confirm whether the obtained cDNA encodes the complete amino acid sequences of each of the antibody VH and VL, including the secretory signal sequence.

[0292] The complete amino acid sequences of the VH and VL of the antibody, including the secretory signal sequence, can be determined by comparing them with the complete amino acid sequences of the VH and VL of known antibodies [Sequences of Proteins of Immunological Interest, US Dept. Health and Human Services (1991)]. The length and N-terminal amino acid sequence of the secretory signal sequence can be deduced, and the subgroup to which they belong can be identified.

[0293] In addition, the amino acid sequences of each CDR of VH and VL can be predicted by comparing them with the amino acid sequences of VH and VL of known antibodies [Sequences of Proteins of Immunological Interest, US Dept. Health and Human Services (1991)].

[0294] Furthermore, the complete amino acid sequences of the obtained VH and VL can be subjected to a homology search using any database, such as SWISS-PROT or PIR-Protein, using the BLAST method [J. Mol. Biol., 215, 403 (1990)], to confirm whether the complete amino acid sequences of the VH and VL are novel.

[0295] (2) Construction of recombinant antibody expression vectors A recombinant antibody expression vector can be constructed by cloning DNA encoding at least one of the CH and CL of a human antibody into an expression vector for animal cells.

[0296] The C region of a human antibody can be the CH and CL of any human antibody, such as the CH of the γ1 subclass and the CL of the κ class. Although cDNA is used to encode the CH and CL of a human antibody, chromosomal DNA consisting of exons and introns can also be used.

[0297] Any expression vector for animal cells can be used as long as it can incorporate and express a gene encoding the C region of a human antibody. Examples of such vectors include pAGE107 [Cytotechnol., 3, 133 (1990)], pAGE103 [J. Biochem., 101, 1307 (1987)], pHSG274 [Gene, 27, 223 (1984)], pKCR [Proc. Natl. Acad. Sci. USA, 78, 1527 (1981)], pSG1bd2-4 [Cytotechnol., 4, 173 (1990)], and pSE1UK1Sed1-3 [Cytotechnol., 13, 79 (1990)]. (1993)], INPEP4 (Biogen-IDEC), N5KG1val (U.S. Pat. No. 6,001,358), N5KG4PE R409K (described in International Publication No. 2006 / 033386), N5KG2 vector (described in International Publication No. 2003 / 033538), transposon vector (WO 2010 / 143698), etc. can be used.

[0298] Examples of promoters and enhancers that can be used in animal cell expression vectors include the SV40 early promoter [J. Biochem., 101, 1307 (1987)], Moloney murine leukemia virus LTR [Biochem. Biophys. Res. Commun., 149, 960 (1987)], the CMV promoter (U.S. Pat. No. 5,168,062), or the immunoglobulin heavy chain promoter [Cell, 41, 479 (1985)] and enhancer [Cell, 33, 717 (1983)].

[0299] For expression of recombinant antibodies, a vector carrying both the antibody H-chain and L-chain genes (tandem vector) [J. Immunol. Methods, 167, 271 (1994)] is used from the viewpoints of ease of vector construction, ease of introduction into animal cells, and balance of the expression levels of the antibody H-chain and L-chain in the cells. However, multiple vectors carrying separate antibody H-chain and L-chain genes (separate vectors) can also be used in combination.

[0300] Examples of tandem recombinant antibody expression vectors include pKANTEX93 (WO 97 / 10354), pEE18 [Hybridoma, 17, 559 (1998)], N5KG1val (U.S. Pat. No. 6,001,358), N5KG4PE R409K (described in WO 2006 / 033386), N5KG2 vector (described in WO 2003 / 033538), and Tol2 transposon vector (WO 2010 / 143698).

[0301] (3) Construction of chimeric antibody expression vector A chimeric antibody expression vector can be constructed by cloning the cDNA encoding the VH or VL of the non-human antibody obtained in (1) upstream of the gene encoding the CH or CL of a human antibody in the recombinant antibody expression vector obtained in (2).

[0302] First, to link the 3'-end of the cDNA encoding the VH or VL of a non-human antibody to the 5'-end of the CH or CL of a human antibody, VH and VL cDNAs are prepared so that the nucleotide sequence of the linking portion encodes the appropriate amino acids and is an appropriate restriction enzyme recognition sequence. Next, the prepared VH and VL cDNAs are cloned upstream of the genes encoding the CH or CL of a human antibody in the recombinant antibody expression vector obtained in (2) so that they are expressed in the appropriate form, thereby constructing a chimeric antibody expression vector.

[0303] Alternatively, a chimeric antibody expression vector can be constructed by amplifying the cDNA encoding the VH or VL of a non-human antibody by PCR using synthetic DNA containing appropriate restriction enzyme recognition sequences at both ends, and cloning the amplified cDNA into the recombinant antibody expression vector obtained in (2).

[0304] (4) Preparation of cDNA encoding the V region of a humanized antibody cDNA encoding the VH or VL of a humanized antibody can be prepared as follows: First, the amino acid sequence of the framework region (hereinafter referred to as FR) of the VH or VL of a human antibody into which the amino acid sequence of the CDR of the VH or VL of the non-human antibody obtained in (1) is grafted is selected.

[0305] The amino acid sequence of the FR to be selected can be any sequence derived from a human antibody. For example, the amino acid sequences of the FR of human antibodies registered in databases such as the Protein Data Bank, or the consensus amino acid sequences of each subgroup of FRs of human antibodies [Sequences of Proteins of Immunological Interest, US Dept. Health and Human Services (1991)], etc., can be used. To minimize a decrease in antibody binding activity, a human FR amino acid sequence that has as high a homology (60% or higher) as possible with the amino acid sequence of the FR of the VH or VL of the original non-human antibody is selected.

[0306] Next, the amino acid sequences of the CDRs of the original non-human antibody are grafted onto the amino acid sequences of the FRs of the VH or VL of the selected human antibody, respectively, to design the amino acid sequences of the VH or VL of the humanized antibody. The designed amino acid sequences are converted into DNA sequences taking into account the codon usage frequency found in the nucleotide sequences of antibody genes [Sequences of Proteins of Immunological Interest, U.S. Dept. Health and Human Services (1991)], to design the cDNA sequences of the VH or VL of the humanized antibody, respectively.

[0307] Based on the designed cDNA sequence, several synthetic DNAs each about 100 to 150 bases long are synthesized and used in PCR. In this case, from the viewpoints of PCR reaction efficiency and the length of DNA that can be synthesized, preferably 4 to 6 synthetic DNAs are designed for each of the H chain and L chain. Alternatively, synthetic DNAs of the full-length variable regions can be synthesized and used.

[0308] Furthermore, by introducing appropriate restriction enzyme recognition sequences into the 5' ends of the synthetic DNA at both ends, cDNA encoding the VH or VL of a humanized antibody can be easily cloned into the recombinant antibody expression vector obtained in (2). After the PCR reaction, the amplified products are cloned into a plasmid such as pBluescript SK(-) (Stratagene), and the nucleotide sequence is determined using a method similar to that described in (1), to obtain a plasmid containing a DNA sequence encoding the amino acid sequence of the VH or VL of the desired humanized antibody.

[0309] (5) Modification of the amino acid sequence of the V region of a humanized antibody When only the CDRs of the VH and VL of a non-human antibody are grafted onto the FRs of the VH and VL of a human antibody, the antigen-binding activity of the resulting humanized antibody is reduced compared to the original non-human antibody [BIO / TECHNOLOGY, 9, 266(1991)]. Therefore, the reduced antigen-binding activity of a humanized antibody can be increased by identifying, from the amino acid sequences of the FRs of the VH and VL of a human antibody, amino acid residues that are directly involved in binding to the antigen, amino acid residues that interact with amino acid residues in the CDRs, and amino acid residues that maintain the three-dimensional structure of the antibody.

[0310] To identify the amino acid residues in FRs involved in antigen-binding activity, the three-dimensional structure of an antibody can be constructed and analyzed using X-ray crystallography [J. Mol. Biol., 112, 535 (1977)] or computer modeling [Protein Engineering, 7, 1501 (1994)]. Alternatively, modified humanized antibodies with the required antigen-binding activity can be obtained by repeatedly preparing several variants of each antibody and examining the correlation with each variant's antigen-binding activity through trial and error.

[0311] Amino acid residues in the FRs of VH and VL of a human antibody can be modified by PCR using synthetic DNA for modification as described in (4). The nucleotide sequence of the amplified product after PCR is determined by the method described in (1) to confirm that the desired modification has been made.

[0312] (6) Construction of humanized antibody expression vector A humanized antibody expression vector can be constructed by cloning the cDNA encoding the VH or VL of the constructed humanized antibody upstream of the gene encoding the CH or CL of the human antibody in the recombinant antibody expression vector obtained in (2).

[0313] For example, by introducing appropriate restriction enzyme recognition sequences into the 5' ends of the synthetic DNAs located at both ends of the synthetic DNAs used to construct the VH or VL of the humanized antibody obtained in (4) and (5), they are cloned upstream of the genes encoding the CH or CL of the human antibody in the recombinant antibody expression vector obtained in (2) so that they can be expressed in an appropriate form.

[0314] (7) Construction of human antibody expression vectors When a hybridoma producing a monoclonal antibody is established using an animal that produces a human antibody as an immunized animal, the amino acid sequences and cDNA sequences of the VH and VL of the human antibody can be obtained in (1). Therefore, a human antibody expression vector can be constructed by cloning the gene encoding the VH or VL of the human antibody obtained in (1) upstream of the gene encoding the CH or CL of the human antibody in the recombinant antibody expression vector obtained in (2).

[0315] (8) Transient expression of recombinant antibodies The recombinant antibody expression vectors obtained in (3), (6), and (7), or modified expression vectors thereof, can be used to transiently express recombinant antibodies, and the antigen-binding activity of the resulting various recombinant antibodies can be efficiently evaluated.

[0316] Any host cells capable of expressing recombinant antibodies can be used as the host cells into which the expression vector is introduced, such as COS-7 cells [American Type Culture Collection (ATCC) No. CRL1651]. The expression vector can be introduced into COS-7 cells using the DEAE-dextran method [Methods in Nucleic Acids Res., CRC Press (1991)] or lipofection [Proc. Natl. Acad. Sci. USA, 84, 7413 (1987)].

[0317] After introduction of the expression vector, the expression level and antigen-binding activity of the recombinant antibody in the culture supernatant are measured using enzyme immunoassay [Monoclonal Antibodies - Principles and practice, Third Edition, Academic Press (1996), Antibodies - A Laboratory Manual, Cold Spring Harbor Laboratory (1988), Monoclonal Antibody Experiment Manual, Kodansha Scientific (1987)] or the like.

[0318] (9) Obtaining stable expression strains of recombinant antibodies and preparing recombinant antibodies By introducing the recombinant antibody expression vectors obtained in (3), (6) and (7) into suitable host cells, transformants that stably express the recombinant antibody can be obtained.

[0319] Examples of methods for introducing an expression vector into a host cell include the electroporation method [JP Patent Publication No. 2-257891, Cytotechnology, 3, 133 (1990)], the calcium ion method, the electroporation method, the spheroplast method, the lithium acetate method, the calcium phosphate method, the lipofection method, etc. Furthermore, examples of methods for introducing a gene into an animal, as described below, include the microinjection method, methods for introducing a gene into ES cells using electroporation or lipofection, and the nuclear transfer method.

[0320] Any host cells capable of expressing recombinant antibodies can be used as the host cells into which recombinant antibody expression vectors are introduced. Examples include mouse SP2 / 0-Ag14 cells (ATCC CRL1581), mouse P3X63-Ag8.653 cells (ATCC CRL1580), Chinese hamster CHO-K1 cells (ATCC CCL-61), DUKXB11 (ATCC CCL-9096), Pro-5 cells (ATCC CCL-1781), CHO-S cells (Life Technologies, Cat No. 11619), dihydrofolate reductase gene (dhfr)-deficient CHO cells (CHO / DG44 cells) [Proc. Natl. Acad. Sci. USA, 77, 4216 (1980)], and lectin-resistant Lec13 cells [Somatic Cell and Molecular Genetics, 12, 55 (1986)], CHO cells lacking the α1,6-fucosyltransferase gene (WO 2005 / 035586, WO 02 / 31140), rat YB2 / 3HL.P2.G11.16Ag.20 cells (ATCC number: CRL1662), etc. are used.

[0321] Furthermore, host cells can also be used that have reduced or absent activity of proteins such as enzymes involved in the synthesis of the intracellular sugar nucleotide GDP-fucose, proteins such as enzymes involved in glycosylation in which the 1-position of fucose is α-linked to the 6-position of N-acetylglucosamine at the reducing end of an N-glycosidically linked complex glycan, or proteins involved in the transport of the intracellular sugar nucleotide GDP-fucose to the Golgi apparatus, such as CHO cells deficient in the α1,6-fucosyltransferase gene (WO 2005 / 035586, WO 02 / 31140).

[0322] After introduction of the expression vector, transformants that stably express the recombinant antibody are selected by culturing them in an animal cell culture medium containing a drug such as G418 sulfate (hereinafter referred to as G418) (Japanese Patent Publication No. 2-257891).

[0323] Media for animal cell culture include RPMI1640 medium (Invitrogen), GIT medium (Nihon Pharmaceutical Co., Ltd.), EX-CELL301 medium (JRH), EX-CELL302 medium (JRH), EX-CELL325 medium (JRH), IMDM medium (Invitrogen), Hybridoma SFM medium (Invitrogen), and media containing various additives such as FBS. The resulting transformed strain is cultured in the medium, resulting in the expression and accumulation of the recombinant antibody in the culture supernatant. The expression level and antigen-binding activity of the recombinant antibody in the culture supernatant can be measured by ELISA or other methods. The expression level of the recombinant antibody produced by the transformant can also be increased using a DHFR amplification system (Japanese Patent Publication No. 2-257891).

[0324] Recombinant antibodies can be purified from the culture supernatant of transformed strains using a protein A column [Monoclonal Antibodies - Principles and practice, Third Edition, Academic Press (1996), Antibodies - A Laboratory Manual, Cold Spring Harbor Laboratory (1988)]. Alternatively, they can be purified by a combination of methods commonly used for protein purification, such as gel filtration, ion exchange chromatography, and ultrafiltration.

[0325] The molecular weight of the H chain, L chain, or entire antibody molecule of a purified recombinant antibody can be measured using polyacrylamide gel electrophoresis [Nature, 227, 680 (1970)] or Western blotting [Monoclonal Antibodies - Principles and practice, Third edition, Academic Press (1996), Antibodies - A Laboratory Manual, Cold Spring Harbor Laboratory (1988)].

[0326] 3. Design of bispecific antibodies The bispecific antibodies of the present invention can be produced by designing a CD3-binding domain, an Fc region or a constant region containing an Fc region, and a disease-related antigen-binding domain, and then linking them to create a bispecific antibody.

[0327] 3-1.Design of CD3-binding domain The desired CD3-binding domain can be obtained using the method described in 1 above, and the cDNA sequence encoding the amino acid sequence of the CDR or variable region contained in each binding domain can be determined using the method described in 2 above.

[0328] 3-2. Design of Fc region or constant region containing Fc region In the present invention, an Fc region or a heavy chain constant region comprising an Fc region has one CD3-binding domain bound to the C-terminus. Thus, in each of the two Fc chains comprised in the Fc region or the two CH chains comprised in the heavy chain constant region, one CD3-binding domain such as scFv or dsFv may be bound to one of the polypeptides, or one CD3-binding domain may be bound to each of the two polypeptides by binding the VH or VL of the CD3-binding domain, respectively.

[0329] To enable the formation of a heterodimer structure between an Fc polypeptide chain or CH1-Fc linked to the amino acid sequence of the CD3-binding domain and an Fc polypeptide chain or CL-Fc, an amino acid sequence is prepared in which the amino acid residues S354C and T366W have been substituted in one CH3 domain, and an amino acid sequence in which the amino acid residues Y349C, T366S, L368A, and Y407V have been substituted in the other CH3 domain.

[0330] Furthermore, in the case of a hetero CH region formed by CL-Fc and CH1-Fc linked to the amino acid sequence of a CD3-binding domain, an amino acid sequence is prepared in which the amino acid residue C220S is added to CH1-Fc, and an amino acid sequence is prepared in which the amino acid residues 216-220 are deleted and the amino acid residue C214S is added to CL-Fc. To form a hetero CH region more efficiently, the amino acid residues H435R and Y436F can be further added to CL-Fc, or an appropriate combination of the amino acid residue substitutions in the CH3 domain can be used.

[0331] 3-3. Design of disease-related antigen-binding domains The amino acid sequence of a binding domain for a desired disease-related antigen is linked to each polypeptide chain of the Fc region or CH region prepared in 3-2. For example, by linking the amino acid sequences of VH / VL, VH-CH1 / VL-Cκ, scFv, or VHH of a monoclonal antibody that specifically binds to a disease-related antigen to the N-terminus of the first or second polypeptide, respectively, anti-CD3 bispecific antibodies with monovalent or divalent disease-related antigen-binding domains can be prepared. Furthermore, by linking the amino acid sequence of scFv or VHH to the C-terminus of the first or second polypeptide, bispecific antibodies with monovalent disease-related antigen-binding domains at the C-terminus, or bispecific antibodies with bivalent or trivalent disease-related antigen-binding domains combined with the N-terminal valency, can be prepared. When the disease-related antigen-binding domains are bivalent or trivalent, the respective antigen-binding domains can be prepared to the same or different epitopes. The bispecific antibody has bispecificity or trispecificity with respect to antibodies other than CD3.

[0332] 4. Preparation of Bispecific Antibodies The above design allows the production of anti-CD3 bispecific antibodies in various molecular formats, and some typical examples are shown below. Hereinafter, unless otherwise specified, VH / VL, scFv, and Fab refer to the VH / VL, scFv, and Fab of antibodies specific to disease-associated antigens. 4-1.An anti-CD3 bispecific antibody comprising a monovalent disease-associated antigen-binding domain at the N-terminus of the Fc region and a monovalent anti-CD3-binding domain at the C-terminus of the Fc region. VH-CH1-Fc-CD3scFv and VL-CL-Fc VL-CH1-Fc-CD3scFv and VH-CL-Fc VH-CH1-Fc and VL-CL-Fc-CD3scFv VL-CH1-Fc and VH-CL-Fc-CD3scFv VH-CH1-Fc-CD3VH and VL-CL-Fc-CD3VL VL-CH1-Fc-CD3VH and VH-CL-Fc-CD3VL VH-CH1-Fc-CD3VL and VL-CL-Fc-CD3VH VL-CH1-Fc-CD3VL and VH-CL-Fc-CD3VH 4-2.An anti-CD3 bispecific antibody comprising a divalent disease-associated antigen-binding domain at the N-terminus of the Fc region and a monovalent anti-CD3-binding domain at the C-terminus of the Fc region. Fab1-Fc-CD3scFv and Fab2-Fc Fab1-Fc and Fab2-Fc-CD3scFv scFv1-CH1-Fc-CD3scFv and scFv2-CL-Fc scFv1-CH1-Fc and scFv2-CL-Fc-CD3scFv scFv1-CH1-Fc-CD3scFv and scFv2-CL-Fc scFv1-CH1-Fc and scFv2-CL-Fc-CD3scFv Bispecific antibodies are produced by substituting the amino acid residues required for heterodimer formation into the above amino acid sequence, and then introducing a vector containing a cDNA encoding the amino acid sequence into animal cells and expressing it. Furthermore, bispecific antibodies lacking α1,6-fucose can also be produced by introducing the vector into FUT8KO cells and expressing them. Furthermore, Fc amino acid residue substitutions can be added to produce bispecific antibodies with high ADCC activity. Protein expression and purification can be performed using the method described in Section 2 above.

[0333] 5. Evaluation of the activity of the bispecific antibody or antibody fragment thereof of the present invention The activity of a purified bispecific antibody or bispecific antibody fragment can be evaluated as follows.

[0334] The binding activity of the bispecific antibody of the present invention toward a cell line expressing at least one of CD3 and a disease-associated antigen can be measured using the binding assay system described above in 1.(7).

[0335] The CDC activity or ADCC activity against cells expressing at least one of CD3 and a disease-associated antigen can be measured by a known method [Cancer Immunol. Immunother., 36, 373 (1993)].

[0336] The cytotoxic activity of the bispecific antibodies of the present invention can be measured by the following method. For example, target cells expressing a disease-associated antigen and PBMCs are seeded on a dedicated plate, the bispecific antibody is added, and the cells are cultured for a certain period of time. The changes in cell number are analyzed using a Real-time Cell Analyzer xCELLigence (ACEA Biosciences) or the like. Alternatively, fluorescently stained target cells and unstained PBMCs are seeded on a 96-well plate, the bispecific antibody is added, the cells are cultured for a certain period of time, dead cell staining is performed, and only viable target cells are counted by flow cytometry to calculate the cytotoxic activity.

[0337] The ability of bispecific antibodies of the present invention to induce cytokine production can be measured by the following method: For example, target cells expressing a disease-associated antigen and PBMCs are seeded into a 96-well plate, and the bispecific antibodies are added and cultured for a certain period of time, after which the cytokine concentration in the culture supernatant is measured using ELISA, the Bio-Plex Suspension Array System, or the BD Cytometric Bead Array (CBA) kit.

[0338] 6. Method for treating diseases using the bispecific antibody or antibody fragment thereof of the present invention The bispecific antibody or bispecific antibody fragment of the present invention can be used to treat diseases associated with at least one of CD3 and a disease-associated antigen, preferably diseases associated with cells expressing the disease-associated antigen, such as malignant tumors and cancers.

[0339] Examples of malignant tumors and cancers include colon cancer, colorectal cancer, lung cancer, breast cancer, glioma, malignant melanoma, thyroid cancer, renal cell carcinoma, leukemia, lymphoma, T-cell lymphoma, gastric cancer, pancreatic cancer, cervical cancer, endometrial cancer, ovarian cancer, bile duct cancer, esophageal cancer, liver cancer, head and neck cancer, squamous cell carcinoma, skin cancer, urinary tract cancer, bladder cancer, prostate cancer, choriocarcinoma, pharyngeal cancer, laryngeal cancer, pleurima, male embryonal tumor, endometrial hyperplasia, endometriosis, embryonal tumor, fibrosarcoma, Kaposi's sarcoma, hemangioma, cavernous hemangioma, hemangioblastoma, retinoblastoma, astrocytoma, neurofibroma, oligodendroglioma, medulloblastoma, neuroblastoma, glioma, rhabdomyosarcoma, glioblastoma, osteogenic sarcoma, leiomyosarcoma, and Wilms' tumor.

[0340] Therapeutic agents containing the bispecific antibody or bispecific antibody fragment of the present invention, or derivatives thereof may contain only the antibody or antibody fragment, or derivatives thereof, as the active ingredient, but are usually provided as pharmaceutical formulations prepared by mixing the antibody or antibody fragment, or derivatives thereof, with one or more pharmacologically acceptable carriers and using methods known in the technical field of pharmaceuticals.

[0341] Examples of administration routes include oral administration and parenteral administration such as oral, respiratory, rectal, subcutaneous, intramuscular, or intravenous administration. Examples of administration forms include sprays, capsules, tablets, powders, granules, syrups, emulsions, suppositories, injections, ointments, and tapes. Various formulations can be manufactured by conventional methods using commonly used excipients, fillers, binders, wetting agents, disintegrants, surfactants, lubricants, dispersants, buffers, preservatives, solubilizers, antiseptics, colorants, flavorings, stabilizers, and the like.

[0342] Examples of excipients include lactose, fructose, glucose, corn starch, sorbitol, crystalline cellulose, sterilized water, ethanol, glycerol, physiological saline, buffer solutions, etc. Examples of disintegrants include starch, sodium alginate, gelatin, calcium carbonate, calcium citrate, dextrin, magnesium carbonate, and synthetic magnesium silicate.

[0343] Examples of binders include methyl cellulose or a salt thereof, ethyl cellulose, gum arabic, gelatin, hydroxypropyl cellulose, and polyvinylpyrrolidone. Examples of lubricants include talc, magnesium stearate, polyethylene glycol, and hydrogenated vegetable oil.

[0344] Examples of stabilizers include amino acids such as arginine, histidine, lysine, and methionine, human serum albumin, gelatin, dextran 40, methylcellulose, sodium sulfite, and sodium metasulfite.

[0345] Other additives include, for example, syrup, petrolatum, glycerin, ethanol, propylene glycol, citric acid, sodium chloride, sodium nitrite, and sodium phosphate.

[0346] Examples of formulations suitable for oral administration include emulsions, syrups, capsules, tablets, powders, and granules.

[0347] Liquid preparations such as emulsions or syrups are produced using additives such as water, sugars such as sucrose, sorbitol, or fructose, glycols such as polyethylene glycol or propylene glycol, oils such as sesame oil, olive oil, or soybean oil, preservatives such as p-hydroxybenzoic acid esters, or flavors such as strawberry flavor or peppermint.

[0348] Capsules, tablets, powders, granules, etc. are produced using additives such as excipients such as lactose, glucose, sucrose, or mannitol; disintegrating agents such as starch or sodium alginate; lubricants such as magnesium stearate or talc; binders such as polyvinyl alcohol, hydroxypropyl cellulose, or gelatin; surfactants such as fatty acid esters; or plasticizers such as glycerin.

[0349] Examples of formulations suitable for parenteral administration include injections, suppositories, sprays, etc. Injections are prepared using a carrier such as a salt solution, a glucose solution, or a mixture of both.

[0350] Suppositories are prepared using carriers such as cocoa butter, hydrogenated fats, or carboxylic acids. Sprays are prepared using carriers that do not irritate the recipient's oral and respiratory mucosa and disperse the monoclonal antibody or its antibody fragment of the present invention into fine particles, facilitating absorption. Examples of carriers include lactose and glycerin. They can also be prepared as aerosols or dry powders. Furthermore, the above-mentioned parenteral preparations can also contain the ingredients exemplified as additives in preparations suitable for oral administration.

[0351] The effective amount of the bispecific antibody of the present invention administered in combination with a suitable diluent and a pharmacologically acceptable carrier is 0.0001 mg to 100 mg per kg of body weight per dose, administered every 2 days to 8 weeks.

[0352] 7. Method for diagnosing diseases using the bispecific antibody or bispecific antibody fragment of the present invention By using the bispecific antibody or bispecific antibody fragment of the present invention to detect or measure cells expressing at least one of CD3 and a disease-associated antigen, it is possible to diagnose diseases involving CD3 and at least one of the disease-associated antigens, preferably diseases involving cells expressing a disease-associated antigen-binding domain.

[0353] Malignant tumors or cancers, which are diseases associated with at least one of CD3 and disease-associated antigens, can be diagnosed, for example, by detecting or measuring at least one of CD3 and disease-associated antigens as follows.

[0354] First, biological samples collected from the bodies of multiple healthy individuals are subjected to detection or measurement of at least one of CD3 and disease-related antigens using the bispecific antibody of the present invention, or a bispecific antibody fragment, or a derivative thereof, and the immunological techniques described below, to examine the amount of at least one of CD3 and disease-related antigens present in the biological samples from the healthy individuals.

[0355] Next, the amount of CD3 and / or disease-related antigen present in the biological sample of the subject is also measured, and the amount is compared with that of healthy individuals.For example, if the amount of disease-related antigen present in the subject is increased compared with that of healthy individuals, the subject is diagnosed with a disease such as cancer.The diagnosis of other diseases related to at least one of CD3 and disease-related antigen can also be performed in the same manner.

[0356] Immunological techniques are methods that use labeled antigens or antibodies to detect or measure the amount of antibodies or antigens, such as radioactive-labeled immunosorbent assays, enzyme immunoassays, fluorescent immunoassays, luminescent immunoassays, Western blotting, and physicochemical techniques.

[0357] An example of a radioactive substance-labeled immunoantibody method is a method in which the bispecific antibody of the present invention or a bispecific antibody fragment is reacted with an antigen or cells expressing the antigen, and then a radiolabeled anti-immunoglobulin antibody or binding fragment is reacted therewith, followed by measurement using a scintillation counter or the like.

[0358] Examples of enzyme immunoassays include methods such as sandwich ELISA, in which a bispecific antibody or a bispecific antibody fragment of the present invention is reacted with an antigen or a cell expressing the antigen, followed by a labeled anti-immunoglobulin antibody or a binding fragment thereof, and then the resulting colored dye is measured with an absorptiometer.

[0359] The label used in the enzyme immunoassay can be a known enzyme label [Enzyme Immunoassay, Igaku-Shoin (1987)], such as alkaline phosphatase label, peroxidase label, luciferase label, or biotin label.

[0360] Sandwich ELISA involves binding an antibody to a solid phase, trapping the antigen to be detected or measured, and then reacting the trapped antigen with a second antibody. In this ELISA, two antibodies or antibody fragments that bind to the antigen to be detected or measured, each with a different antigen-binding site, are prepared. The first antibody or antibody fragment is pre-adsorbed onto a plate (e.g., a 96-well plate), and the second antibody or antibody fragment is then labeled with a fluorescent substance such as FITC, an enzyme such as peroxidase, or biotin. Cells or their lysates isolated from a living body, tissue or their lysates, cell culture supernatant, serum, pleural effusion, ascites, or ocular fluid are then reacted with the plate to which the antibody is adsorbed, followed by the addition of a labeled antibody or antibody fragment, and a detection reaction according to the labeled substance is performed. The antigen concentration in the test sample is calculated from a calibration curve prepared by serially diluting a known concentration of antigen.

[0361] The antibodies used in sandwich ELISA may be either polyclonal or monoclonal antibodies, or antibody fragments such as Fab, Fab', or F(ab)2. The combination of two antibodies used in sandwich ELISA may be a combination of monoclonal antibodies or antibody fragments thereof that bind to different epitopes, or a combination of a polyclonal antibody and a monoclonal antibody or antibody fragments thereof.

[0362] Fluorescence immunoassays are performed using methods described in literature such as "Monoclonal Antibodies - Principles and Practice, Third Edition, Academic Press (1996) and "Monoclonal Antibody Experiment Manual," Kodansha Scientific (1987)." Labels used in fluorescence immunoassays include known fluorescent labels such as "Fluorescent Antibody Method," Soft Sciences (1983). For example, FITC or RITC may be used.

[0363] Luminescence immunoassays are performed, for example, by methods described in Bioluminescence and Chemiluminescence: Clinical Tests 42, Hirokawa Shoten (1998). Labels used in luminescence immunoassays include known luminescent labels, such as acridinium esters or lophine.

[0364] In Western blotting, antigens or cells expressing antigens are fractionated by SDS (sodium dodecyl sulfate)-PAGE [Antibodies—A Laboratory Manual, Cold Spring Harbor Laboratory (1988)], the gel is then blotted onto a polyvinylidene fluoride (PVDF) or nitrocellulose membrane, and the membrane is reacted with an antibody or antibody fragment that binds to the antigen, followed by reaction with an anti-IgG antibody or antibody fragment labeled with a fluorescent substance such as FITC, an enzyme such as peroxidase, or biotin, and the label is visualized for measurement. An example is shown below.

[0365] First, cells or tissues expressing a polypeptide having the desired amino acid sequence are lysed, and 0.1–30 μg of protein per lane is electrophoresed by SDS-PAGE under reducing conditions. The electrophoresed protein is then transferred to a PVDF membrane and blocked by incubation in PBS containing 1–10% BSA (hereinafter referred to as BSA-PBS) at room temperature for 30 minutes. The membrane is then reacted with a bispecific antibody of the present invention, washed with PBS containing 0.05–0.1% Tween-20 (Tween-PBS), and incubated with peroxidase-labeled goat anti-mouse IgG for 2 hours at room temperature. After washing with Tween-PBS, the antigen is detected by detecting the antibody-bound band using ECL Western Blotting Detection Reagents (Amersham) or similar. Antibodies used for Western blotting detection are those capable of binding to polypeptides that do not retain their native conformation.

[0366] As a physicochemical technique, for example, the bispecific antibody or bispecific antibody fragment of the present invention is bound to at least one of the antigens CD3 and a disease-related antigen to form aggregates, and the aggregates are then detected. Other physicochemical techniques that can be used include the capillary method, one-dimensional immunodiffusion, immunoturbidimetry, and latex immunoturbidimetry [Clinical Test Methods Summary, Kanehara Publishing (1998)].

[0367] In latex immunoturbidimetry, a carrier such as polystyrene latex with a particle size of approximately 0.1 to 1 μm that has been sensitized with an antibody or antigen is used. When an antigen-antibody reaction occurs with the corresponding antigen or antibody, scattered light in the reaction solution increases and transmitted light decreases. This change is detected as absorbance or integrating sphere turbidity, allowing the antigen concentration in the test sample to be measured.

[0368] On the other hand, cells expressing at least one of CD3 and a disease-related antigen can be detected or measured using known immunological detection methods, but immunoprecipitation, immune cell staining, immunohistochemical staining, or fluorescent antibody staining are preferably used.

[0369] In the immunoprecipitation method, cells expressing at least one of CD3 and a disease-related antigen are reacted with the bispecific antibody of the present invention or its antibody fragment, and then a carrier capable of specifically binding to immunoglobulin, such as protein G Sepharose, is added to precipitate the antigen-antibody complex.

[0370] Alternatively, the following method can also be used. First, the bispecific antibody or bispecific antibody fragment of the present invention is immobilized on a 96-well ELISA plate, followed by blocking with BSA-PBS. Next, the BSA-PBS is discarded and the plate is thoroughly washed with PBS, after which a lysate of cells or tissues expressing at least one of CD3 and a disease-related antigen is reacted. After thorough washing, the immunoprecipitates are extracted from the plate using SDS-PAGE sample buffer and detected by Western blotting as described above.

[0371] In immunocytostaining or immunohistostaining, cells or tissues expressing an antigen are treated, optionally with a detergent or methanol to improve antibody permeability, and then reacted with the bispecific antibody of the present invention. The bispecific antibody is then further reacted with an anti-immunoglobulin antibody or its binding fragment that has been fluorescently labeled (e.g., FITC), enzyme-labeled (e.g., peroxidase), or biotin-labeled, and the label is visualized and then examined under a microscope. Detection can also be performed by fluorescent antibody staining, in which cells are reacted with a fluorescently labeled antibody and analyzed using a flow cytometer [Monoclonal Antibodies - Principles and Practice, Third Edition, Academic Press (1996); Monoclonal Antibody Experimental Manual, Kodansha Scientific (1987)]. In particular, the bispecific antibody or bispecific antibody fragment of the present invention can detect at least one of CD3 and disease-related antigens expressed on the cell membrane by fluorescent antibody staining.

[0372] Furthermore, when fluorescent antibody staining methods such as the FMAT8100HTS system (Applied Biosystems) are used, the amount of antigen or antibody can be measured without separating the formed antibody-antigen complex from the free antibody or antigen that is not involved in the formation of the antibody-antigen complex. [Example]

[0373] The present invention will be explained below by giving specific examples, but the present invention is not limited to these examples.

[0374] [Example 1] Construction of animal cell expression vectors for anti-HER2 monovalent antibody and CD3 / HER2 bispecific antibody (1) Construction of an expression vector for anti-HER2 monovalent antibody in animal cells Expression vectors for animal cells (referred to as pKANTEX93mvG1 and pKANTEX93mvG4PE(R409K), respectively) of an IgG1-type anti-HER2 monovalent antibody and an IgG4PE(R409K)-type HER2 monovalent antibody (referred to as mvG1 and mvG4PE(R409K), respectively, having the structures shown in Figure 2(A)) were prepared according to the method described in WO 2014 / 054804.

[0375] Monovalent antibodies are heterodimers consisting of two different polypeptides, and the polypeptide containing the VH is referred to as the first polypeptide (VH-CH1-hinge-Fc, also referred to as polypeptide 1), and the polypeptide containing the VL is referred to as the second polypeptide (VL-CL-hinge-Fc, also referred to as polypeptide 2).

[0376] Table 1 shows the amino acid sequence and nucleotide sequence of the CH1-Hinge-Fc portion of the first polypeptide used to prepare each monovalent antibody.

[0377] [Table 1]

[0378] The IgG4PE(R409K) type shown in Table 1 has an IgG4 heavy chain constant region in which the Ser residue at position 228 is substituted with Pro, the Leu residue at position 235 with Glu, and the Arg residue at position 409 with Lys, and does not have ADCC activity.

[0379] Furthermore, the amino acid and nucleotide sequences of the VH and VL of the above-mentioned anti-Her2 monovalent antibody were the same as those of the anti-Her2 antibody Trastuzumab (described in Proc. Natl. Acad. Sci. USA 89: 4285, 1992).

[0380] (2) Construction of animal cell expression vector for CD3 / HER2 bispecific antibody To prepare heterodimeric CD3 / HER2 bispecific antibodies (Figure 2(B)), based on the above-mentioned IgG1 anti-HER2 monovalent antibody and IgG4PE(R409K) anti-HER2 monovalent antibody, an anti-CD3 scFv was linked via a linker to the C-terminus of the first polypeptide of each monovalent antibody, an animal cell expression vector was prepared by the method described below.

[0381] Using restriction enzyme sites, the nucleotide sequence encoding a portion of the C-terminal portion of the H-chain constant region of the first polypeptide of the IgG1-type anti-HER2 monovalent antibody expression vector pKANTEX93mvG1 prepared in (1) above was replaced with a nucleotide sequence encoding a portion of the C-terminal portion of the H-chain constant region + peptide linker + anti-CD3 scFv, thereby obtaining the IgG1-type CD3 / HER2 bispecific antibody expression vector pKANTEX93mvG1_scT3a.

[0382] The amino acid sequence of the peptide linker used was [SerGlyGlyGlyGly (SEQ ID NO: 143)]. The amino acid sequence and nucleotide sequence of the anti-CD3 scFv used were those described in US Pat. No. 7,575,923 B2. The amino acid sequence of the anti-CD3 scFv is shown in SEQ ID NO: 9. This sequence is hereinafter referred to as scT3a.

[0383] Using a similar method, the IgG4PE(R409K) CD3 / HER2 bispecific antibody expression vector pKANTEX93mvG4PE(R409K)_scT3a was obtained from the IgG4PE(R409K) anti-HER2 monovalent antibody expression vector pKANTEX93mvG4PE(R409K) prepared in (1) above.

[0384] The antibody expression vectors prepared in (1) above and in this section were both prepared using the PureYield Plasmid Midiprep System (Promega) with Escherichia coli Competent Quick DH5α (TOYOBO) as the host, and were used in subsequent experiments.

[0385] [Example 2] Expression and purification of anti-HER2 monovalent antibody and CD3 / HER2 bispecific antibody Using the various antibody expression vectors prepared in Example 1, the high ADCC IgG1-type anti-HER2 monovalent antibody 4D5_mvG1(DF), the IgG4PE(R409K)-type HER2 antibody 4D5_mvG4 PE(R409K)(F), the high ADCC IgG1-type CD3 / HER2 bispecific antibody 4D5_mvG1_scT3a(DF), and the IgG4PE(R409K)-type CD3 / HER2 bispecific antibody 4D5_mvG4PE(R409K)_scT3a(F) were obtained according to the methods described below.

[0386] Table 2 shows the name of each antibody, the corresponding expression vector, the host cells used, the Fc glycan structure, and the presence or absence of anti-CD3 scFv.

[0387] [Table 2]

[0388] In the table, CHO / DG44 refers to CHO / DG44 cells derived from Chinese Hamster Ovary (Somatic Cell Mol Genet 12; 555, 1986), and FUT8KO CHO refers to CHO / DG44 cells in which the α1,6 fucosyltransferase gene (FUT8) has been knocked out (U.S. Patent No. 6,946,292). 4×10 6 8 μg of each expression vector was added to each of the above host cells, and gene transfer was performed by electroporation (Cytotechnology. 3: 133, 1990). After gene transfer, each cell was cultured for 2 days in IMDM medium (GIBCO) containing 10% dialyzed fetal bovine serum (dFBS), and then further cultured for approximately 2 weeks in the above medium containing 0.5 mg / mL G418 sulfate to obtain G418-resistant clones. The G418-resistant clones were cultured in Excell 302 medium (SAFC Biosciences), and the culture supernatant containing the expressed protein was collected.

[0389] One liter of culture supernatant was passed through a column packed with Protein A (MabSelect SuRe: GE Healthcare) at a flow rate of 0.5-1.0 mL / min. The column was washed twice with 10 mL of phosphate buffered saline (PBS), and the antibody was eluted with 0.1 M citrate buffer (pH 3.9). The resulting eluate was immediately neutralized with 2 M Tris-HCl buffer (pH 8.0) and then dialyzed against 10 mM citric acid, 150 mM sodium chloride (pH 6.0) to replace the buffer.

[0390] The antibody solution was further subjected to gel filtration chromatography to remove multimers or degradation products. A Superdex 200 10 / 300 GL column (GE Healthcare) was connected to an AKTA explorer 10S (GE Healthcare), and the flow path was replaced with 10 mM citric acid and 150 mM sodium chloride (pH 6.0). The prepared antibody solution was added to the column and passed through at a flow rate of 0.5 mL / min, and 0.5 mL fractions were collected using a fraction collector. The OD of each collected fraction was 1.0 mL. 280 The fractions showing the main peak were collected as fractions containing the target antibody monomer, and the antibody solution was sterilized through a 0.22 μm filter and stored. SDS-PAGE analysis of the obtained antibody samples showed the expected molecular weight size, confirming that the target antibody had been obtained. As a control sample, Trastuzumab (DF), an antibody in which the α1,6 fucose in the N-linked glycan of the anti-HER2 antibody Trastuzumab had been removed, was prepared according to the method described in Clin Cancer Res. 13: 1875-1882, 2007, and used in some experiments.

[0391] Furthermore, as a negative control, an IgG1-type anti-2,4-dinitrophenyl (DNP) antibody was prepared in accordance with the methods described in Examples 1 and 2 using a DNA fragment (clone name DNP2) encoding the anti-2,4-dinitrophenyl (DNP) IgG1 antibody described in Clin Cancer Res. 11(8): 3126-3135, 2005, and was used in some experiments.

[0392] [Example 3] Evaluation of antigen-binding activity of anti-HER2 monovalent antibody and CD3 / HER2 bispecific antibody (1) Evaluation of HER2 binding activity The HER2 binding activity of the various antibodies obtained in Example 2 was evaluated by a binding inhibition experiment against Trastuzumab. First, Trastuzumab (Roche) was labeled with Alexa 488 using an Alexa Fluor 488 Antibody Labeling Kit (Life Technologies) to produce Trastuzumab_Alx. Next, binding inhibition analysis was performed using a flow cytometer using the HER2-positive human breast cancer cell line SK-BR-3 (ATCC HTB-30).

[0393] Trastuzumab_Alx was added at a final concentration of 2.0 μg / mL, and unlabeled IgG1 antibody, Trastuzumab, various anti-HER2 monovalent antibodies, or CD3 / HER2 bispecific antibodies at final concentrations of 105, 21, 4.2, and 0.84 nM were added and incubated for 1.5 hours at 4°C. Human serum-derived IgG1 / kappa (Millipore) was added at the same concentration as the negative control. After washing, the mean fluorescent intensity (MFI) of the antibody bound to SK-BR-3 cells was measured using a Cytomics FC 500 MPL flow cytometer (Beckman Coulter). The results are shown in Figure 4.

[0394] As shown in Figure 4, Trastuzumab, various anti-HER2 monovalent antibodies, and various CD3 / HER2 bispecific antibodies inhibited the binding of Trastuzumab_Alx in an antibody concentration-dependent manner, confirming their Her2-specific binding activity. The four anti-HER2 monovalent antibodies and CD3 / HER2 bispecific antibodies were found to have similar levels of binding activity to HER2.

[0395] (2) Evaluation of CD3 binding activity The CD3 binding activity of the various antibodies obtained in Example 2 was evaluated based on their binding activity to commercially available frozen T cells (AllCells). Each antibody was added to the T cells at a final concentration of 10 μg / mL and incubated at 4°C for 1 hour. After washing, Goat anti-Human IgG F(ab)2-FITC (Acris Antibodies) diluted to 30 μg / mL was added and incubated at 4°C for 1 hour.

[0396] As a positive control, 30 μg / mL of anti-CD3 antibody OKT-3 (Abcam) was used as the primary antibody, and Goat anti-mouse IgG F(ab)2-FITC (Dako) was used as the secondary antibody. After washing, the MFI of the antibody bound to T cells was measured using a Cytomics FC 500 MPL flow cytometer (Beckman Coulter). The results are shown in Figure 5.

[0397] As shown in Figure 5, among the antibodies evaluated, the CD3 / HER2 bispecific antibodies 4D5_mvG1_scT3a(DF) and 4D5_mvG4PE(R409K)_scT3a(F) both bound to T cells to a similar extent. On the other hand, Trastuzumab and the anti-HER2 monovalent antibodies 4D5_mvG1(DF) and 4D5_mvG4PE(R409K)(F) did not bind to T cells. From the above, 4D5_mvG1_scT3a(DF) and 4D5_mvG4PE ( R409K)_scT3a(F) was found to be able to bind to CD3 with similar strength.

[0398] [Example 4] Evaluation of the cytotoxic activity of anti-HER2 monovalent antibody and CD3 / HER2 bispecific antibody The cytotoxic activity of the various antibodies prepared in Example 2, including both Fc-mediated ADCC activity and CD3-mediated ADTC activity, was measured using the Real-time Cell Analyzer xCELLigence (ACEA Biosciences), which detects the amount of cancer cells attached to a plate over time, as described below. This utilizes the fact that changes in electrical resistance occur as the amount of cells on the plate changes. In other words, the cell index value obtained by this measurement shows a positive correlation with the amount of cells on the plate.

[0399] It is known that the main effector cells for ADCC activity are NK cells that express the Fc receptor FcγRIIIA, and the main effector cells for ADTC activity are T cells that express CD3. Therefore, by using human peripheral blood mononuclear cells (PBMCs) that contain both of these as effectors and HER2-positive cancer cell lines as targets, the cytotoxic activity, including both ADCC activity and ADTC activity, of various antibodies can be measured.

[0400] First, RPMI1640 medium (supplemented with 10% FBS) was added to an E-plate VIEW 16 (ACEA Biosciences) for baseline correction. HER2-positive breast cancer cell line BT-20 cells (ATCC HTB-19) or HER2-low expressing MCF-7 cells (ATCC HTB-22) were detached from the flask with 0.02% EDTA (Nacalai Tesque), washed with PBS, and then plated at 2.0 × 10 cells / ml in RPMI1640 (10% FBS, 10 μg / mL gentamicin). 5 The solution was diluted to 100 cells / mL and added to the plate at 50 μL / well.

[0401] The plate was incubated in a CO2 incubator for 30 minutes, and then cell index measurements were initiated. Approximately 20 hours later, 1 × 10 PBMCs were collected from peripheral blood of healthy volunteers by density gradient method. 6 The antibody was diluted to 1000 cells / mL and added to the plate at 50 μL / well. Then, 50 μL / well of each antibody, adjusted to a concentration four times the final concentration, was added and measurements were restarted. The final data was normalized to the cancer cell mass value immediately before antibody addition. The results are shown in Figures 6 to 8.

[0402] Figure 6 shows the cell index for BT-20 cells alone, BT-20 cells plus PBMCs, or BT-20 cells plus the IgG1-type anti-HER2 monovalent antibody 4D5_mvG1(DF) (final concentration 50 nM), the IgG4PE(R409K)-type CD3 / HER2 bispecific antibody 4D5_mvG4_scT3a(F) (final concentration 50 nM), or a mixture of these (a mixture of each at a final concentration of 25 nM).

[0403] In Figure 6, the difference between the amount of cancer cells when target cells and PBMCs are added and the amount of cancer cells when the antibody is further added represents the cytotoxic activity of the antibody. As shown in Figure 6, 4D5_mvG1(DF) and 4D5_mvG4PE(R409K)_scT3a(F) each have moderate cytotoxic activity, and a mixture of these 50% antibodies has a higher cytotoxic activity than either 4D5_mvG1(DF) or 4D5_mvG4PE(R409K)_scT3a(F) alone.

[0404] As shown in Table 2, the IgG1 anti-HER2 monovalent antibody 4D5_mvG1(DF) has an IgG1 constant region to which an N-linked glycan lacking α1,6 fucose is attached, and as described in WO 2014 / 054804, it has high ADCC activity mediated by FcγRIIIA-binding activity. However, since it does not have a CD3-binding site, it does not have ADTC activity.

[0405] As shown in Table 2, the IgG4PE(R409K) CD3 / HER2 bispecific antibody 4D5_mvG4(R409K)_scT3a(F) has an IgG4-type constant region with α1,6 fucose added, and contains amino acid residue modifications that reduce effector activity. Therefore, it has no or very low ability to induce ADCC activity, but its molecular structure possesses CD3-mediated ADCC activity.

[0406] The cytotoxic activity of the group to which these two antibodies were added at 25 nM each was higher than the cytotoxic activity of either the group to which 4D5_mvG1(DF) alone or 4D5_mvG4PE(R409K)_scT3a(F) alone was added at 50 nM, demonstrating that mixing molecules with only ADCC activity and molecules with only ADTC activity exhibits synergistic cytotoxic activity.

[0407] Therefore, these results indicate that acting half the amount of an antibody with only ADCC activity or half the amount of an antibody with only ADTC activity on cancer cells simultaneously has a greater antitumor effect than acting half the amount of an antibody with only ADCC activity or half the amount of an antibody with only ADTC activity on cancer cells alone.

[0408] Furthermore, the cytotoxic activities of both 4D5_mvG1(DF) and 4D5_mvG4PE(R409K)_scT3a(F) were moderate, and the activity of one of them was not saturated at the same concentration, indicating that the effective concentration ranges for ADCC activity and ADTC activity in this experiment were not significantly different.

[0409] Figure 7 shows the cell index for BT-20 cells alone, BT-20 cells plus PBMCs, and a mixture of 4D5_mvG1(DF) and 4D5_mvG4_scT3a(F) (0.3 nM each), or the IgG1-type CD3 / HER2 bispecific antibody 4D5_mvG1_scT3a(DF) (0.6 nM).

[0410] The results of comparing the cytotoxic activity of a mixture of 4D5_mvG1(DF) and 4D5_mvG4PE(R409K)_scT3a(F) (0.3 nM each) with that of 4D5_mvG1_scT3a(DF) (0.6 nM) are shown in Figure 7.

[0411] As shown in Figure 7, 4D5_mvG1_scT3a(DF) exhibited a stronger antitumor effect than the simultaneous addition of 4D5_mvG1(DF) and 4D5_mvG4PE(R409K)_scT3a(F).

[0412] As shown in Table 2, 4D5_mvG1_scT3a(DF) has a molecular structure that possesses high ADCC activity due to the IgG1-type constant region bound to an N-linked glycan lacking α1,6 fucose, as well as CD3-mediated ADTC activity.

[0413] Therefore, the high antitumor effect achieved by combining a molecule having only ADCC activity and a molecule having only ADTC activity, as shown in Figure 6, was further enhanced by conferring both ADCC and ADTC activities to the same molecule, demonstrating a greater synergistic effect.

[0414] Figures 8(A) to (E) show the cell index for MCF-7 cells alone, MCF-7 cells plus PBMCs, or MCF-7 cells plus the anti-HER2 antibody Trastuzumab (DF), the IgG1-type anti-HER2 monovalent antibody 4D5_mvG1 (DF), the IgG4PE(R409K)-type CD3 / HER2 bispecific antibody 4D5_mvG4PE(R409K)_scT3a (F), or the IgG1-type CD3 / HER2 bispecific antibody 4D5_mvG1_scT3a (DF) (all at a final concentration of 50 nM).

[0415] As shown in Figure 8(A) to (E), Trastuzumab (DF), 4D5_mvG1 (DF), and 4D5_mvG4PE(R409K)_scT3a (F) all showed almost no cytotoxic activity against MCF-7 cells, whereas 4D5_mvG1_scT3a (DF) showed extremely high cytotoxic activity.

[0416] These results demonstrate that an antibody possessing both ADCC and ADTC activities in a single molecule can exert an extremely potent antitumor effect against cancer cells for which antibodies possessing only ADCC activity or only ADTC activity show little cytotoxicity.

[0417] [Example 5] Examination of linker sequences of CD3 / HER2 bispecific antibodies In this section, we prepared bispecific antibodies of the CD3 / HER2 bispecific antibody 4D5_mvG1_scT3a(DF) by varying the type and length of the peptide linker sequence connecting the C-terminus of CH3 of the first polypeptide and the anti-CD3 scFv, and examined the effect of the peptide linker on the biological activity.

[0418] Table 3 shows the names of the CD3 / HER2 bispecific antibodies prepared, as well as the amino acid sequences and structures of the peptide linkers.

[0419] [Table 3]

[0420] Among the antibodies listed in Table 3, H1 and H3 are bispecific antibodies each having one or three turns of α-helix, and Pro having a peptide linker containing proline. The peptide linkers of these antibodies are all [SerGlyGlyGlyGly (SEQ ID NO: 143)]. n (n is 1 to 3) It has a rigid structure compared to the linker (Chen X, et al. Adv. Drug Deliv. Rev. 65: 1357-1369, 2013). Therefore, by comparing the biological activities of various bispecific antibodies with the above-mentioned different peptide linkers, it is possible to determine the effect of the rigidity and length of the linker moiety on the cytotoxic activity of CD3 / HER2 bispecific antibodies.

[0421] Using the EcoT22I-BamHI restriction enzyme site of the bispecific antibody expression vector pKANTEX93mvG1_scT3a prepared in Example 1, the gene fragment comprising a nucleotide sequence encoding the C-terminus of the first polypeptide-[SerGlyGlyGlyGly (SEQ ID NO: 143)] linker-anti-CD3 scFv was replaced with a gene fragment comprising a nucleotide sequence encoding the C-terminus of the first polypeptide-a peptide linker listed in Table 3-anti-CD3 scFv, thereby preparing animal cell expression vectors for the various bispecific antibodies listed in Table 3.

[0422] These expression vectors were introduced into FUT8KO CHO cells to express and purify the proteins using the method described in Example 2. Of the bispecific antibodies produced, the purified proteins obtained for G1, G2, H1, H3, and Pro showed the predicted molecular weight size in SDS-PAGE analysis, confirming that the desired antibodies were obtained.

[0423] The HER2 and CD3 binding activities of the various bispecific antibodies prepared were measured using the method described in Example 3. The results are shown in Figures 9 and 10. Figure 9 shows that 4D5_mvG1_scT3a(DF), G1, G2, H1, H3, and Pro all have equivalent HER2 binding activities, and Figure 10 shows that, although there was some variability between the antibodies, all of the antibodies had largely similar CD3 binding activities.

[0424] The cytotoxic activity of the bispecific antibody against BT-20 cells was measured by the method described in Example 3 (final concentration: 0.617 nM). Controls included target cells alone, or target cells and PBMCs alone (no antibody added). The results are shown in Figures 11 and 12.

[0425] As shown in Figures 11 and 12, the CD3 / HER2 bispecific antibodies G1, G2, H1, H3, and Pro all exhibited cytotoxic activity equivalent to or greater than that of 4D5_mvG1_scT3a(DF). These results demonstrate that the high cytotoxic activity of the IgG1-type CD3 / HER2 bispecific antibody 4D5_mvG1_scT3a(DF) is independent of the structure of the linker moiety.

[0426] [Example 6] Cytotoxic activity of CD3 / HER2 bispecific antibodies against gastric cancer cell lines Genes encoding the amino acid sequences of each polypeptide shown in Table 4 were inserted into antibody expression vectors prepared by modifying Promega's pCI vector. For example, the antibody designated DNP2 indicates a negative control antibody that uses the variable region of an antibody that binds to DNP as the cancer antigen variable region. The (F) and (DF) at the end of the antibody name stand for fucosylated and defucosylated, respectively, and indicate the presence or absence of α1,6 fucose attached to the sugar chains of the Fc region.

[0427] The gene sequences were inserted downstream of the CAG promoter for the heavy chain (first polypeptide) and the CMV promoter for the light chain (second polypeptide). Transient expression was performed using this vector and the Freestyle CHO Expression kit (Thermo Fisher Scientific), CHO-S cells, or suspension CHO cells with the α1,6 fucosyltransferase gene (FUT8) knockout (hereafter referred to as CHO(FUT8 KO)). Alternatively, similar antibodies could be produced using the Expi293 Expression System (Thermo Fisher Scientific) and Expi293F cells or Expi293F cells with FUT8 knockout (hereafter referred to as Expi293F(FUT8 KO)).

[0428] Antibody molecules with α1,6 fucose were produced using CHO-S cells or Expi293F cells, and antibody molecules lacking α1,6 fucose were produced using CHO(Fut8 KO) cells or Expi293F(FUT8 KO) cells.

[0429] The cell culture medium was centrifuged and the culture supernatant was collected through a 0.2 μm filter (Thermo Scientific). Crudely purified antibodies were obtained from the culture supernatant by affinity purification using MabSelect SuRe (GE Healthcare). Specifically, the resin packed in the column was equilibrated with PBS, and the culture supernatant was applied to the column. After washing once with PBS containing 0.1% Triton X-114 (final concentration) and once with PBS, the antibody was eluted with elution buffer (100 mM citric acid-NaOH, pH 3.9 or 20 mM acetic acid-NaOH, 50 mM NaCl, pH 3.9).

[0430] The resulting antibody solution was neutralized by adding 1 / 10 volume of neutralization buffer (1 M phosphate-NaOH, pH 7.0) and concentrated by ultrafiltration using Amicon Ultra-4 Centrifugal Filter Units (Millipore). Because the crudely purified antibody contains a certain proportion of aggregates, gel filtration chromatography was performed to separate only the monomers using an AKTA chromatography system and a Superdex 200 Increase 10 300 GL column (both GE Healthcare).

[0431] The mobile phase was citrate buffer (10 mM citric acid-NaOH, 150 mM NaCl, pH 6.0). Only the monomer fraction was separated, and the absorbance A280 was measured using a Nanodrop (Thermo Fisher Scientific) to determine and adjust the concentration of the antibody solution.

[0432] [Table 4]

[0433] The cytotoxic activity of the bispecific antibodies against the Her2-positive human gastric cancer cell line MKN-7 cells (RIKEN BRC RCB0999) was measured using the method described in Example 3 (final concentration: 50 nM). Controls included target cells alone, or target cells and PBMCs alone (no antibody added). The results are shown in Figure 13. None of the antibodies with DNP2 variable regions exhibited cytotoxic activity.

[0434] On the other hand, antibodies with only ADCC activity (4D5_mvG1(DF) and 4D5_IgG1(DF)) exhibited cytotoxic activity, but the activity was weaker than that of an antibody with only ADTC activity (4D5_mvG4PE(R409K)_scT3a(F)). A 50 nM mixture of 4D5_mvG1(DF) and 4D5_mvG4PE(R409K)_scT3a(F) showed higher activity than the above antibody groups. Furthermore, 4D5_mvG1_scT3a(DF), which can exert both ADCC and ADTC activities in a single molecule, showed the highest cytotoxic activity of all antibody groups.

[0435] Similarly, cytotoxic activity was measured against the Her2-positive human gastric cancer cell line MKN-45 cells (JCRB Cell Bank JCRB0254), and the results are shown in Figure 14. None of the negative control antibodies with DNP2 variable regions exhibited cytotoxic activity. Of the antibodies with 4D5 variable regions, only 4D5_mvG1_scT3a(DF) exhibited clear cytotoxic activity. These results demonstrate that this technology exerts high cytotoxic activity due to a synergistic effect even against HER2-positive cell lines other than BT-20 cells.

[0436] [Example 7] Cytotoxic activity of CD3 / HER2 bispecific antibody and cytokine production in culture supernatant To compare the properties of the bispecific antibodies produced in the present invention with those of representative anti-CD3 bispecific antibodies described in the Background Art, a control molecule was prepared and the cytotoxic activity and cytokine production in the culture supernatant upon cytotoxicity were measured. A known anti-CD3 bispecific antibody (having the amino acid sequences set forth in SEQ ID NOS: 160, 359, and 399 described in WO 2016 / 071004) was prepared as the control molecule. Hereinafter, this antibody will also be referred to as CD3 / HER2 Fab×scFv.

[0437] An expression vector was prepared in the same manner as in Example 6, and the antibody was expressed using the Expi293 Expression System (Thermo Scientific). Subsequent purification procedures were carried out in accordance with the method described in Example 6.

[0438] The cytotoxic activity of the prepared control bispecific antibody and 4D5_mvG1_scT3a(DF) against BT-20 cells was measured by the method described in Example 4. The cell index value for wells to which no test substance was added was defined as 0% activity, and the cell index value for wells to which 0.5% Triton X-100 was added instead of the test substance was defined as 100% activity. The cytotoxic activity value of each antibody was calculated from the cell index value on Day 2 (48 hours after antibody addition).

[0439] Culture supernatants were collected at the same time points, and cytokine (IFN-γ, TNF-α, IL-6) concentrations were measured using the BD Cytometric Bead Array (CBA) Human Th1 / Th2 Cytokine Kit II (BD Pharmingen). Cytotoxic activity and cytokine concentrations at each concentration were plotted on the same graph, as shown in Figure 15. The maximum cytotoxic activity was comparable for 4D5_mvG1_scT3a(DF) and the control antibody. However, the level of cytokine production differed significantly between the two. Specifically, the control antibody exhibited a concentration-dependent increase in cytotoxic activity along with an increase in cytokine production, demonstrating a small discrepancy between the concentration ranges at which cytotoxic activity was observed and those at which cytokine production was induced. In contrast, 4D5_mvG1_scT3a(DF) showed almost no increase in cytokine production associated with an increase in cytotoxic activity.

[0440] [Example 8] Design of various molecular forms of CD3 / HER2 bispecific antibodies Examples 1 to 7 show the process for producing a heterodimeric CD3 / HER2 bispecific antibody (Figure 2(B)), which is based on an IgG1-type anti-HER2 monovalent antibody and has an anti-CD3 scFv linked to the C-terminus of its first polypeptide, and the results of its activity evaluation. These results demonstrate, for the first time, that a single molecule exhibits both ADCC and ADTC activities, thereby exerting synergistically high cytotoxic activity while maintaining low cytokine production induction ability.

[0441] To further investigate molecular types that can achieve such synergistic effects between ADCC and ADTC activities, the molecular types shown in Figures 16(A) to 16(F) were designed. Figure 16(A) shows the molecule used in the previous examples. Figure 16(B) shows a molecule in which an anti-CD3 scFv is linked to the C-terminus of one heavy chain of a bivalent antibody, making the cancer antigen Fab bivalent. To create a heterogeneous heavy chain, a Knobs-into-Hole modification (Nature Biotechnology. Vol. 16: 677-681, 1998) was introduced into the CH3 portion.

[0442] Figure 16(C) shows a molecule with the same amino acid sequence as Figure 16(A), but with a sugar chain containing α1,6 fucose added by using conventional CHO cells for expression. Figure 16(D) shows a molecule in which an anti-CD3 scFv is linked to the C-terminus of the second polypeptide (VL-CL-hinge-Fc) of a monovalent antibody. Figure 16(E) shows a molecule in which both the cancer antigen Fab and the anti-CD3 scFv are located at the N-terminus. Similar to Figure 16(B), a knobs-into-holes modification was introduced to create a heterogeneous heavy chain. Figure 16(F) shows a molecule in which one anti-CD3 scFv is linked to the C-terminus of each of the first and second polypeptides of a monovalent antibody, resulting in two anti-CD3 scFvs in one molecule.

[0443] [Example 9] Preparation of various molecules and evaluation of cytotoxic activity (1) Preparation of designed bispecific antibody molecules The molecules designed in Example 8 were produced in the same manner as in Example 6. Table 5 shows the antibodies produced and their amino acid sequences.

[0444] [Table 5]

[0445] The "molecule type" in Table 5 corresponds to Figures 16(B) to (F). Molecules without scFv, which were used as negative controls when evaluating each molecular type, are classified as the same molecular type as molecules with scFv.

[0446] The "antibody name" in Table 5 indicates molecular characteristics in the following order: (1) the name of the antibody clone on the cancer antigen side, (2) the origin and structure of the antibody subclass in the constant region, (3) the CD3 clone name, and (4) the presence or absence of fucose. The "mv" in antibody names such as mvG1 and mvG4PE(R409K) indicates that the antibody is monovalently bound to the cancer antigen side. The "G1" in antibody names such as mvG1 and mvG4PE(R409K) indicates that the antibody has an IgG1-derived Fc, and "G4PE(R409K)" indicates that the antibody has an IgG4-derived Fc with the addition of S228P, L235E, and R409K mutations.

[0447] In Table 5, (F) indicates the presence of fucose, and (DF) indicates the absence of fucose. Antibodies marked with (DF) were expressed in CHO (FUT8 KO), and those marked with (F) were expressed in CHO-S. The "sc" in antibody names such as scT3a and scSP34 is an abbreviation for scFv, and scSP34 refers to an scFv that uses VH and VL derived from SP34. (LC) indicates that the scFv is bound to a second polypeptide consisting of CL-Fc, as shown in Figure 14(D).

[0448] SP34 in molecular form (E) is an anti-CD3 monoclonal antibody described in the literature (EMBO J. 1985. 4(2):337-344; J. Immunol. 1986, 137(4):1097-100; J. Exp. Med. 1991, 174:319-326; J. Immunol. 1991, 147(9):3047-52), and its amino acid sequence is published as SEQ ID NO: 5 and SEQ ID NO: 10 in U.S. Pat. No. 10,066,015, among others.

[0449] SP34(H05') refers to the sequence of an anti-CD3 antibody in which the affinity for CD3 has been reduced by amino acid alterations in the variable region sequence of SP34 (prepared in Example 11 described below). In Table 5, "4D5mut" refers to the variable region of an antibody in which the binding to HER2 has been completely eliminated by introducing alterations of two amino acid residues into the CDR of the heavy chain variable region of Tratsuzumab.

[0450] (2) Cancer antigen binding valency and cytotoxic activity of CD3 / HER2 and CD3 / GM2 bispecific antibodies The cytotoxic activity of the molecular type (B) shown in Table 5 against BT-20 cells was measured by the method described in Example 4. 4D5_IgG4PE(R409K)(F), 4D5_IgG1(DF), 4D5_IgG4PE(R409K)_scT3a(F), and 4D5_IgG1_scT3a(DF) were each added to a final concentration of 50 nM. For comparison, the cytotoxic activity of molecular type (A) 4D5_mvG1_scT3a(DF) and 4D5_mvG4PE(R409K)_scT3a(F) was also measured under the same conditions. The results are shown in Figures 17(A) and (B).

[0451] As shown in Figure 17(A), 4D5_IgG4PE(R409K)(F) has no effector activity, and no cytotoxic activity was observed solely through binding to HER2. As shown in Figure 17(B), both 4D5_IgG4PE(R409K)_scT3a(F) and 4D5_IgG1(DF) were observed to exhibit cytotoxic activity, with 4D5_IgG1_scT3a(DF) demonstrating higher cytotoxic activity than 4D5_IgG4PE(R409K)_scT3a(F). The activity of 4D5_IgG1_scT3a(DF), in which the Fab of the cancer antigen-binding domain is bivalent, was similar to that of 4D5_mvG1_scT3a(DF), in which the Fab of the cancer antigen-binding domain is monovalent.

[0452] These findings demonstrate that the synergistic cytotoxicity effect due to the ADTC and ADCC activities of a single antibody molecule is observed not only when the cancer antigen-binding domain is monovalent, but also when it is bivalent.

[0453] Similarly, the cytotoxic activity of the molecule (B) anti-cancer specific antigen (TSA) antibody as an anti-ganglioside GM2 monoclonal antibody against the human small cell lung cancer cell line SBC-3 cells (JCRB Cell Bank JCRB0818) was measured. 8962_IgG1(DF), 8962_mvG1_scT3a(DF), and 8962_IgG1_scT3a(DF) were added to a final concentration of 50 nM. The results are shown in Figure 18(A) and (B).

[0454] As a negative control for the anti-TSA antibody variable region, the activity of an antibody molecule containing the variable region of an anti-DNP antibody is shown in Figure 18(A), and the activity of the test substance, an antibody molecule containing the variable region of an anti-GM2 monoclonal antibody, is shown in Figure 18(B). Under these conditions, no cytotoxic activity was observed with the bivalent ADCC antibody 8962_IgG1(DF), but high cytotoxic activity was observed with the bivalent 8962_IgG1_scT3a(DF) which has both ADCC and ADTC activities.

[0455] (3) Cytotoxicity of CD3 / HER2 bispecific antibodies with or without α1,6 fucose The cytotoxic activity of the molecular type (C) shown in Table 5 against BT-20 cells was measured by the method described in Example 4. 4D5_mvG4PE(R409K)(F), 4D5_mvG1(F), 4D5_mvG4PE(R409K)_scT3a(F), and 4D5_mvG1_scT3a(F) were each added to a final concentration of 50 nM.

[0456] The results are shown in Figure 19(A). MV G4PE(R409K)(F) did not show cytotoxic activity (data not shown). MV G4PE(R409K)_scT3a(F) showed weak cytotoxic activity. Furthermore, 4D5_mvG1_scT3a(F) was found to have higher cytotoxic activity than the two aforementioned molecules. Even when α1,6 fucose was present, it was revealed that the activity was increased by exerting both ADCC activity and ADTC activity in a single molecule.

[0457] However, as shown in Figure 19(B), when compared with molecules from which α1,6 fucose had been removed, the highly ADCC antibody 4D5_mvG1(DF) had stronger cytotoxic activity than 4D5_mvG1_scT3a(F), with 4D5_mvG1_scT3a(DF) having even stronger cytotoxic activity. These results demonstrate that the synergistic increase in cytotoxic activity due to the exertion of ADCC and ADTC in a single molecule is a general phenomenon that is independent of the presence or absence of α1,6 fucose. Furthermore, it was revealed that the exertion of high ADCC and ATDC activity in a single molecule due to the removal of α1,6 fucose results in higher cytotoxic activity.

[0458] (4) Fusion site of anti-CD3 scFv and cytotoxic activity of CD3 / HER2 bispecific antibodies The cytotoxic activity of the molecular type (D) shown in Table 5 against BT-20 cells was measured using the method described in Example 4. 4D5_mvG4PE(R409K), 4D5_mvG1(DF), 4D5_mvG4PE(R409K)_scT3a(LC)(F), and 4D5_mvG1_scT3a(LC)(DF) were each added to a final concentration of 50 nM. The results are shown in Figure 20(A).

[0459] 4D5_mvG4PE(R409K)_scT3a(LC)(F) showed cytotoxic activity comparable to that of 4D5_mvG1(DF), while 4D5_mvG1_scT3a(LC)(DF) showed higher cytotoxic activity. The activity of molecular forms (A) and (D) was also compared in the same manner, and the results are shown in Figure 20(B). 4D5_mvG4PE(R409K)_scT3a (LC) (F )and 4D5_mvG4PE(R409K)_scT3a(F), and 4D5_mvG1_scT3a (LC) (DF )and When compared with the combination of 4D5_mvG1_scT3a(DF), both combinations showed almost the same level of cytotoxic activity.

[0460] In other words, it was revealed that addition of the anti-CD3 scFv to the C-terminus of either the first polypeptide or the second polypeptide does not affect the ADTC activity of the anti-CD3 bispecific antibody, nor the synergistic effect of the ADTC and ADCC activities, nor the increase in either activity.

[0461] The cytotoxic activity of the molecular type (E) shown in Table 5 against BT-20 cells was measured by the method described in Example 4. 4D5_scT3a_IgG1(DF) and 4D5mut_scT3a_IgG1(DF) were added to final concentrations of 50 nM, 5 nM, and 0.5 nM, respectively. The results are shown in Figure 21.

[0462] 4D5_scT3a_IgG1(DF) showed high cytotoxic activity at all concentrations. As mentioned above, 4D5mut_scT3a_IgG1(DF) does not bind to HER2, but showed nonspecific cytotoxic activity at all concentrations.

[0463] The activity of 4D5mut_scSP34_IgG1(DF) and 4D5mut_scSP34(H05')_IgG1(DF) was measured in a similar manner, and the results compared with 4D5mut_scT3a_IgG1(DF) are shown in Figures 22A to 22C. Both antibodies exhibited nonspecific cytotoxic activity, but 4D5mut_scSP34(H05')_IgG1(DF) and 4D5mut_scSP34_IgG1(DF) exhibited higher cytotoxic activity than 4D5mut_scT3a_IgG1(DF).

[0464] The binding of these antibodies to CD3 was measured using Biacore (GE Healthcare). Specifically, anti-tetra His mouse antibody (QIAGEN) was immobilized on a CM5 sensor chip at approximately 8000 RU using an amine coupling kit (GE Healthcare). 10 mM sodium acetate (pH 4.5) was used as the coupling buffer. His-tagged human CD3D&E protein (Sino Biological) dissolved in HBS-EP(+) buffer at 5 μg / mL was used as the ligand and captured at 5 μL / min for 120 seconds.

[0465] Next, a bispecific antibody solution, diluted two-fold in five serial dilutions starting from 10 μg / mL, was added as the analyte at a flow rate of 30 μL / min. The binding reaction between each antibody and the analyte was measured for 2 minutes, and the dissociation reaction for 5 minutes. Measurements were performed using single-cycle kinetics. The resulting sensorgrams were analyzed using Bia Evaluation Software (GE Healthcare), and the kinetic constants of each antibody were calculated.

[0466] The results are shown in Figures 22D to 22G. D The value is 1.26 x 10 -8 M, K of 4D5mut_scSP34(H05')_IgG1(DF) D The value is 1.08 x 10 -7 M. K of 4D5mut_scT3a_IgG1(DF) D Although the value could not be calculated, binding was observed on the sensorgram, and the binding-dissociation equilibrium state was reached at approximately 6 RU at the highest concentration added. 4D5mut_scSP34(H05')_IgG1(DF) reached the binding-dissociation equilibrium state at approximately 100 RU at the highest concentration added, so the K D The value is larger than this, K D >1.08×10 -7 It can be determined that it is M.

[0467] That is, binding of the anti-CD3 scFv to the N-terminus, rather than the C-terminus, of the Fc region or constant region results in the affinity of the CD3-binding domain being K D >1.08×10 -7 It was shown that even when anti-CD3 scFv with a weak M was used, undesirable nonspecific cytotoxic activity unrelated to specific binding to disease-associated antigens was observed.

[0468] (5) Valency and cytotoxic activity of anti-CD3 scFv of CD3 / HER2 bispecific antibodies The cytotoxic activity of the molecular type (F) shown in Table 5 against BT-20 cells was measured by the method described in Example 4. As a positive control for cytotoxic activity, 4D 5_ mvG1_scSP34(H05') (DF) was added to a final concentration of 50 nM. 4D5mut_mvG1_scT3a(DF) and DNP2_mvG1_(scT3a)2(DF) were added to a final concentration of 50 nM, 5 nM, and 0.5 nM. The results are shown in Figures 23(A) and (B), respectively.

[0469] Although none of these molecules bind to HER2, 4D5mut_mvG1_scT3a (DF) showed no cytotoxic activity at any concentration, whereas DNP2_mvG1_(scT3a)2(DF) showed concentration-dependent cytotoxic activity at 50 nM and 5 nM, indicating that the bivalent anti-CD3 scFv results in nonspecific cytotoxicity.

[0470] The results of (4) and (5) above demonstrate that a single anti-CD3 bispecific antibody molecule having both ADCC and ADTC activities exhibits synergistically high cytotoxic activity and that, in order to prevent the nonspecific cytotoxic activity of the anti-CD3 bispecific antibody, it is necessary for the CD3-binding domain to be added monovalently to the C-terminus of the Fc of the anti-CD3 bispecific antibody.

[0471] [Example 10] Preparation of a CD3 / HER2 bispecific antibody with a high affinity anti-CD3 scFv and evaluation of its cytotoxic activity and cytokine production The influence of the affinity of anti-CD3 scFv on the activity of the bispecific antibody was examined. The anti-CD3 scFv was an scFv having the amino acid sequence shown in SEQ ID NO: 73 (VH and VL of clone I2C described in U.S. Patent Publication No. 2011 / 0275787), K D The value is 1 x 10 -8 Using the same procedure as in Example 6, DNP2_mvG1_scI2C(DF) and 4D5_mvG1_scI2C(DF) having the molecular type (A) structure shown in Table 5 were prepared using the same procedure as in Example 6. The cytotoxic activity of these molecules (final concentration 50 nM) against BT-20 cells was measured using the method described in Example 4. The results are shown in Figure 24A.

[0472] As shown in Figure 24A, for molecules using scT3a as the anti-CD3 scFv (prepared in Example 6), no cytotoxic activity was observed for DNP2_mvG1_scT3a(DF), which does not specifically bind to HER2, but cytotoxic activity was observed for 4D5_mvG1_scT3a(DF), which specifically binds to HER2. On the other hand, for molecules using clone scI2C as the anti-CD3 scFv, high cytotoxic activity was observed for both DNP2_mvG1_scI2C(DF) and 4D5_mvG1_scI2C(DF), regardless of whether or not HER2-specific binding was present. Therefore, it was revealed that nonspecific cytotoxic activity occurs when clone scI2C is used as the anti-CD3 scFv, even with the same molecular type.

[0473] Next, human PBMCs were incubated with 4D5_mvG1_scT3a(DF) or 4D5_mvG1_scI2C(DF) and the amount of cytokines produced was measured. 6 The cells / 100 μL / tube) and antibodies (final concentrations of 100, 10, and 1 μg / mL) were mixed with X-VIVO® 15 serum-free hematopoietic cell medium and cultured at 37°C for 24 hours. After culture, the supernatant was collected and the cytokine levels were measured according to the method described in Example 7. The results are shown in Figure 24B.

[0474] As shown in Figure 24B, in the 4D5_mvG1_scT3a(DF)-treated group, increased cytokine production was observed at 100 μg / mL, but almost no cytokine production was observed at 10 μg / mL or 1 μg / mL. On the other hand, significant cytokine production was observed even at low concentrations with 4D5_mvG1_scI2C(DF). These results demonstrate that the use of the I2C clone, which has a high affinity as an anti-CD3 scFv, results in nonspecific cytotoxic activity and significant cytokine production.

[0475] [Example 11] Preparation and activity evaluation of CD3 / HER2 bispecific antibodies with adjusted affinity of the anti-CD3 scFv portion (1) Design of amino acid modifications Example 10 demonstrated the importance of appropriately controlling the affinity of the anti-CD3 scFv portion. Using SP34, the anti-CD3 monoclonal antibody shown in Example 9(1), an attempt was made to reduce its affinity to an appropriate range by introducing amino acid modifications into the CDR. The nucleotide sequence, amino acid sequence, and CDR amino acid sequence of scSP34 used as the scFv are shown in SEQ ID NOs: 74, 75, and 76 to 81, respectively.

[0476] A model antibody structure was created from the amino acid sequence of SP34 by three-dimensional structural modeling using Molecular Operating Environment 2016.08 (Molsys). Amino acid residues present on the surface and likely contributing to binding were selected as candidate residues for modification, and various amino acid-altered CDRs were designed. Figure 25(A) shows the design for introducing a single amino acid modification, and Figure 25(B) shows the design for introducing two or more amino acid residues. The bolded amino acid residues in the amino acid sequences in the table are the modified amino acid residues. Furthermore, the amino acid sequence of a humanized antibody with modified SP34 CDRs was designed using a procedure similar to that described in Example 15(1) of International Patent Publication No. 2019 / 017401. The amino acid sequences of the VL and VH frameworks are shown in Figures 26(A), (C), and (B), (D), respectively.

[0477] (2) Antibody production and affinity measurement A portion of the amino acid sequence designed in (1) was actually produced as a HER2 / CD3 bispecific antibody in the same manner as described in Example 6, and its binding activity was examined. Expi293F cells were used as the expressing cells. Affinity measurements were performed according to the method described in Example 9(4). The results are shown in Figure 27. Compared to scSP34, an anti-CD3 scFv produced from the parent clone SP34, the affinity was improved by introducing one or more amino acid residue modifications. D =10 -7 It was revealed that the temperature can be reduced to the order of .

[0478] (3) Cytotoxic activity and cytokine production during cytotoxicity HER2 / CD3 bispecific antibodies 4D5_mvG1_scSP34(H04')(DF) and 4D5_mvG1_scSP34(H05')(DF), which have the scSP34 CDR variants scSP34(H04') and scSP34(H05') discovered in (2), were produced using FUT8KO CHO cells. The cytotoxic activity of these antibodies against BT-20 cells (final antibody concentration: 400 pM) and cytokine production upon cytotoxicity were measured in accordance with the method described in Example 7. The results are shown in Figures 28(A)-(C) and 29(A)-(D), respectively.

[0479] As shown in Figures 28(A) to 28(C), the cytotoxic activity of 4D5_mvG1_scT3a(DF) was slightly higher than that of 4D5_mvG1(DF), but the difference was not significant. On the other hand, 4D5_mvG1_scSP34(H04')(DF) and 4D5_mvG1_scSP34(H05')(DF) were found to have high cytotoxic activity, although their activity was slightly lower than that of 4D5_mvG1_scSP34(DF).

[0480] As shown in Figure 29(A)-(D), the amount of cytokines in the culture supernatant was extremely high in 4D5_mvG1_scSP34(DF), significantly lower in 4D5_mvG1_scSP34(H04')(DF), and even lower in 4D5_mvG1_scSP34(H05')(DF). In particular, IL-2 and IFN-γ were hardly produced in the 4D5_mvG1_scSP34(H04')(DF) and 4D5_mvG1_scSP34(H05')(DF)-treated groups under the experimental conditions used in this study.

[0481] These results demonstrate that cytotoxic activity and cytokine production can be controlled by adjusting affinity through the introduction of amino acid residue modifications into anti-CD3 scFv. The amino acid sequence of the affinity-adjusted anti-CD3 scFv scSP34(H04') is shown in SEQ ID NO: 117, the amino acid sequences of the heavy chain CDRs 1 to 3 in SEQ ID NOs: 118 to 120, the amino acid sequences of the light chain CDRs 1 to ...

Claims

1. An anti-CD3 bispecific antibody or a bispecific antibody fragment thereof, comprising an Fc region having the ability to bind to an Fc receptor, one CD3-binding domain having reduced affinity for CD3 bound to the C-terminus of the Fc region, and further comprising a disease-related antigen-binding domain, wherein the CD3-binding domain is composed of complementarity determining regions (CDRs) 1 to 3 (HCDRs) 1 to 3 of an antibody heavy chain variable region (VH) and a light chain variable region (LCDs) 1 to 3 of an antibody light chain variable region (LCDs). an anti-CD3 bispecific antibody or a bispecific antibody fragment thereof, comprising CDR1 to 3 (LCDR1 to 3) of a CD3 region (VL), wherein the amino acid sequences of the HCDR1 to 3 and the LCDR1 to 3 are any one selected from the following (a) to (h), and the bispecific antibody or the bispecific antibody fragment thereof satisfies at least one requirement selected from the following requirements (A) and (B): (a) HCDR1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 118 to 120, respectively, and LCDR1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 121 to 123, respectively. (b) HCDR1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 83 to 85, respectively, and LCDR1 to 3 containing the amino acid sequences represented by SEQ ID NOs: 86 to 88, respectively. (c) HCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 132, 96, 97, respectively. (d) HCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 98 to 100, respectively, and LCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 133, 96, 97, respectively. (e) HCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 98, 134, 100, respectively, and LCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (f) HCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 98, 135, 100, respectively, and LCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (g) HCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 98, 136, 100, respectively, and LCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (h) HCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 98, 137, 100, respectively, and LCDR1 to 3 comprising the amino acid sequences represented by SEQ ID NOs: 95 to 97, respectively. (A) The dissociation constant (K D ) of the CD3-binding domain for CD3 is 6×10 −8 or more. (B) The dissociation constant of the CD3-binding domain with CD3 is greater than that of the comparative anti-CD3 monoclonal antibodies SP34 and KM14.

2. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to claim 1, which has a reduced ability to induce cytokine production in the presence of CD3-positive T cells and disease-related antigen-positive cells, compared to an anti-CD3 bispecific antibody using the anti-CD3 monoclonal antibody SP34.

3. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to claim 1 or 2, which satisfies at least one requirement selected from the group consisting of the following requirements (D) to (F) and (H): (D) comprising one or two of the disease-related antigen-binding domains. (E) The CD3-binding domain is any one selected from scFv and Fab, and the disease-related antigen-binding domain is any one selected from scFv, Fab, and VHH. (F) The CD3-binding domain and / or the disease-associated antigen-binding domain are linked to the Fc region via a linker. (H) The CD3 binding domain is an scFv.

4. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of claims 1 to 3, wherein the VH and VL amino acid sequences of the CD3-binding domain are any one of the VH and VL amino acid sequences selected from the following (aa) to (ss): (aa) VH comprising the amino acid sequence represented by SEQ ID NO: 124, and VL comprising the amino acid sequence represented by SEQ ID NO: 125 (bb) VH comprising the amino acid sequence represented by SEQ ID NO: 115, and VL comprising the amino acid sequence represented by SEQ ID NO: 116 (cc) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 126 (dd) VH comprising the amino acid sequence represented by SEQ ID NO: 94, and VL comprising the amino acid sequence represented by SEQ ID NO: 127 (ee) VH comprising the amino acid sequence represented by SEQ ID NO: 128, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ff) VH comprising the amino acid sequence represented by SEQ ID NO: 129, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (gg) VH comprising the amino acid sequence represented by SEQ ID NO: 130, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (hh) VH comprising the amino acid sequence represented by SEQ ID NO: 131, and VL comprising the amino acid sequence represented by SEQ ID NO: 93 (ii) VH comprising the amino acid sequence represented by SEQ ID NO: 159, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (jj) VH comprising the amino acid sequence represented by SEQ ID NO: 160, and VL comprising the amino acid sequence represented by SEQ ID NO: 165 (kk) VH comprising the amino acid sequence represented by SEQ ID NO: 161, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (11) VH comprising the amino acid sequence represented by SEQ ID NO: 162, and VL comprising the amino acid sequence represented by SEQ ID NO: 166 (mm) VH comprising the amino acid sequence represented by SEQ ID NO: 168, and VL comprising the amino acid sequence represented by SEQ ID NO: 180 (nn) VH comprising the amino acid sequence represented by SEQ ID NO: 169, and VL comprising the amino acid sequence represented by SEQ ID NO: 181 (oo) VH comprising the amino acid sequence represented by SEQ ID NO: 170, and VL comprising the amino acid sequence represented by SEQ ID NO: 182 (pp) VH comprising the amino acid sequence represented by SEQ ID NO: 171, and VL comprising the amino acid sequence represented by SEQ ID NO: 183 (qq) VH comprising the amino acid sequence represented by SEQ ID NO: 172, and VL comprising the amino acid sequence represented by SEQ ID NO: 184 (rr) VH comprising the amino acid sequence represented by SEQ ID NO: 173, and VL comprising the amino acid sequence represented by SEQ ID NO: 185 (ss) an amino acid sequence having 80% or more homology with each of the VH and VL amino acid sequences selected from the above (aa) to (rr).

5. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of claims 1 to 4, wherein the Fc region has enhanced binding activity to an Fc receptor.

6. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to claim 5, wherein the Fc region with enhanced binding activity to an Fc receptor is an Fc region that comprises an amino acid residue modification and / or a sugar chain modification.

7. the amino acid residue modification comprises at least one amino acid residue modification that enhances the binding activity to an Fc receptor, or the sugar chain modification is a sugar chain modification in which α1,6-linked fucose to N-acetylglucosamine at the reducing end of the N-linked sugar chain bound to Asn at position 297 (EU numbering) in the Fc region is deleted; The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to claim 6.

8. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of claims 1 to 7, comprising two Fabs as the disease-related antigen-binding domains, and wherein the C-terminus of the heavy chain (VH-CH1) of each Fab is linked to the Fc region directly or via a linker.

9. The anti-CD3 bispecific antibody or the bispecific antibody fragment thereof according to any one of claims 1 to 7, which satisfies at least one requirement selected from the group consisting of the following requirements (I) to (K): (I) The disease-related antigen-binding domain comprises one Fab, and the C-terminus of the heavy chain (VH-CH1) and light chain (VL-CL) of the Fab are linked to the Fc region directly or via a linker. (J) The disease-related antigen-binding domain comprises one Fab, and the CD3-binding domain is bound directly or via a linker to the Fc chain on the side that binds to the heavy chain (VH-CH1) of the Fab. (K) The disease-associated antigen-binding domain comprises one Fab, and the CD3-binding domain is bound directly or via a linker to the Fc chain on the side that binds to the light chain (VL-CL) of the Fab.

10. The anti-CD3 bispecific antibody or bispecific antibody fragment thereof according to claim 8 or 9, wherein the linker is a hinge or a modified version thereof.

11. A DNA encoding the anti-CD3 bispecific antibody or the bispecific antibody fragment according to any one of claims 1 to 10.

12. A recombinant vector containing the DNA of claim 11.

13. A transformant obtained by introducing the recombinant vector according to claim 12 into a host cell.

14. A method for producing the anti-CD3 bispecific antibody or the bispecific antibody fragment according to any one of claims 1 to 10, comprising culturing the transformant according to claim 13 in a medium, producing and accumulating the anti-CD3 bispecific antibody or the bispecific antibody fragment according to any one of claims 1 to 10 in the culture, and collecting the anti-CD3 bispecific antibody or the bispecific antibody fragment from the culture.

15. A therapeutic and / or diagnostic agent for a disease associated with at least one of CD3 and the disease-related antigen, comprising the anti-CD3 bispecific antibody or the bispecific antibody fragment according to any one of claims 1 to 10 as an active ingredient.

16. The therapeutic and / or diagnostic agent according to claim 15, wherein the disease associated with at least one of CD3 and the disease-associated antigen is cancer.

17. A reagent for detecting or measuring at least one of CD3 and a disease-related antigen, comprising the anti-CD3 bispecific antibody or the bispecific antibody fragment according to any one of claims 1 to 10.

Images