Anti-cd3 antibody, molecule comprising the antibody, production methods thereof, use of said molecule, cells and uses thereof, polynucleotides, vectors, and pharmaceutical compositions

Abstract

This invention provides a novel antibody that binds to human CD3, or a molecule having antigen-binding properties containing the antibody. The provision of novel antibodies that bind to human CD3, molecules that contain such antibodies and have antigen-binding properties, and pharmaceutical compositions that contain such antibodies or molecules as active ingredients and have cytotoxic activity.

Description

Anti-CD3 Antibody, Molecule Containing the Antibody, and Method for Producing the Same The present invention relates to a novel antibody that binds to human CD3 and cynomolgus monkey CD3, and a molecule containing the antibody. 1) Like any other monoclonal antibody, an anti-CD3 monoclonal antibody functions by highly recognizing its target molecule. An anti-CD3 antibody recognizes only a single antigenic determinant (epitope) on the CD3 molecule as the target. The most widely used and best-characterized specific monoclonal antibody in the CD3 complex is OKT3. 2) OKT3 is a mouse-derived anti-human CD3 monoclonal antibody (Non-Patent Document 1). To prevent allograft rejection of organs, the administration of an anti-CD3 monoclonal antibody, which binds to the TCR complex on human T cells, and the method of preventing transplant organ rejection by inhibiting the activation and proliferation of T cells have been used for a long time (Non-Patent Documents 2-4). OKT3 is the first anti-CD3 antibody used for such treatment. OKT3 has a strong immunosuppressive effect. On the other hand, its clinical use is hindered due to serious side effects related to its immunogenicity and potential ability to promote cell division (Non-Patent Documents 5-8). 3) OKT3 induces T cell proliferation and cytokine production in vitro, and releases a large amount of cytokines in vivo, causing cytokine syndrome (Non-Patent Document 5). This is because OKT3 is a bivalent IgG, which produces cross-linking of T cells and Fcγ receptor-expressing cells, resulting in T cell proliferation (Non-Patent Document 8). Also, since OKT3 is a mouse antibody, it is known that heterophilic antibodies such as human anti-mouse antibodies (HAMA) are produced after long-term administration (Non-Patent Document 7). Reports on the therapeutic application and side effects of anti-CD3 antibodies are summarized as follows (Patent Document 1). 4) To solve such problems, scFv of OKT3 (Non-Patent Document 9), humanized OKT3 (Non-Patent Document 10), etc. have been prepared. Also, as a further application example of OKT3, a bispecific antibody in which a single chain of OKT3 with T cell activation ability is combined with a single chain of an antibody against a target antigen expressed on the surface of cancer cells has been reported (Patent Document 2, Non-Patent Document 11). 5) Multispecific antibodies using anti-CD3 antibodies described in the current art are expected to have great therapeutic potential for the treatment of malignant diseases. For example, it is known that TROP2 is overexpressed in various epithelial cancer types (Non-Patent Documents 12-16). There has been no report yet on a bispecific antibody in which an antigen-binding fragment of a human TROP2-specific antibody and an antigen-binding fragment of an anti-CD3 antibody are genetically linked and expressed. 6) OKT3 reacts with CD3 of chimpanzees, but does not react with CD3 from other primates such as cynomolgus monkeys (Non-Patent Document 17). UCHT-1, which is a monoclonal anti-CD3 antibody, also reacts with CD3 from chimpanzees, but does not react with CD3 from cynomolgus monkeys (Non-Patent Document 18). On the other hand, there are also examples of monoclonal antibodies that recognize cynomolgus monkey antigens but do not recognize their human counterparts. One example of this group is FN-18, a monoclonal antibody directed against CD3 from cynomolgus monkeys (Non-Patent Document 19). 7) The limitation of OKT3 and a series of antibodies modified from OKT3 is that it only specifically binds to human CD3. Such a limitation may become a major obstacle in the development of therapeutic agents for treating human diseases. This is because in order to obtain market approval for the development of candidate drugs, it is necessary to conduct pre-clinical trials. In pre-clinical trials, it is expected to use animals, especially cynomolgus monkeys, which are higher primates, for the trials. Therefore, in the case of candidate drugs using anti-CD3 antibodies, it is highly desirable to use anti-CD3 antibodies that can bind to CD3 of both humans and cynomolgus monkeys. 8) Antibodies that bind to both humans and cynomolgus monkeys have been reported in Patent Documents 3, 4, and Non-Patent Document 20. Also, bispecific antibodies in which a single chain of such anti-CD3 antibodies binds to a single chain of an antibody against a target antigen expressed on the surface of cancer cells have been reported (Patent Document 5, Non-Patent Document 21). However, in order to utilize multispecific antibodies and multispecific molecules that may be adaptable to a rich variety of cancer targets, anti-CD3 antibodies that bind to different antigenic determinants and bind to both humans and cynomolgus monkeys are desired. Prior Art Documents Patent Documents Patent Document 1 International Publication No. 2012 / 162067 Patent Document 2 International Publication No. 2007 / 108152 Patent Document 3 US Patent Publication No. 8236308 Specification Patent Document 4 International Publication No. 2008 / 119567 Patent Document 5 International Publication No. 2015 / 026892 Non-Patent Documents Non-Patent Document 1: Salmeron A. et al., J. Immunol. (1991) 147, 3047-3052 Non-Patent Document 2: Cosmi AB. et al., Transplantation (1981) 32, 535-539 Non-Patent Document 3: Gilbert EM. et al., Am. J. Med. (1987) 82, 202-206 Non-Patent Document 4: Thistlethwaite JR. et al., Transplanation (1987) 43, 176-184 Non-Patent Document 5: Abramowicz D. et al., Transplanation (1989) 47, 606-608 Non-Patent Document 6: Toussaint D. et al., Transplanation (1989) 48, 524-526 Non-Patent Document 7: Thistlethwaite, JR. et al., Am. J. Kidney Dis. (1988) 11, 112-119 Non-Patent Document 8: Meuer, SC. et al., Eur. J. Immunol. (1986) 136, 4106-4112 Non-Patent Document 9: George AJ. et al., J. Immunol. (1994) 152(4), 1802-11 Non-Patent Document 10: Woodle ES. et al., J Immunol. (1992) 148(9), 2756-63 Non-Patent Document 11: Yankelevich M.et al., Pediatr. Blood Cancer (2012) 59(7), 1198-1205 Non-Patent Document 12: Qhmachi T. et al., Clin. Cancer Res. (2006) 12(18), 3857-3863 Non-Patent Document 13: Muhlmann G., et al., J. Clin. Pathol. (2009) 62(2), 152-158 Non-Patent Document 14: Fong D., et al., Br. J. Cancer (2000) 99(8), 1290-1295. Non-Patent Document 15: Fong D. et al., Mod. Pathol. (2000) 21(2)(2000), 186-191 Non-Patent Document 16: Ning S., et al., Neurol. Sci. (2013) 34(10), 1745-1750 Non-Patent Document 17: Sandusky et al., J.Med. Primatol.(1986)15, 441-451 Non-Patent Document 18 http: / / www.nhpreagents.org / NHP / clonelist.aspx?ID=77 Non-Patent Document 19 Uda et al., J.Med.Primatol.(2001)30, 141-147 Non-Patent Document 20 Conrad ML.et al., Cytometry A.(2007)71(11), 925-33 Non-Patent Document 21 Lum LG. et al., BioDrugs(2011)25(6), 365-379. Summary of the Invention Problems to be Solved by the Invention An object of the present invention is to provide a novel antibody that binds to human CD3 and cynomolgus monkey CD3 or an antigen-binding fragment of the antibody (hereinafter, also referred to as an antibody, etc.), a molecule containing the antibody, etc. and one or two or more other antibodies or antigen-binding fragments of the antibody, a molecule that is multispecific, a pharmaceutical composition containing the antibody, etc. or the molecule as an active ingredient and having cytotoxic activity, etc. Means for Solving the Problems The inventors of the present invention conducted in-depth research to solve the above-mentioned problems, created a novel anti-CD3 antibody and a molecule containing the antibody, and completed the present invention. That is, the present invention includes the following inventions. (1) An antibody or an antigen-binding fragment of the antibody, characterized in that: the heavy chain sequences respectively include: CDRH1 containing the amino acid sequence shown in SEQ ID NO: 26, CDRH2 containing the amino acid sequence shown in SEQ ID NO: 98, and CDRH3 containing the amino acid sequence shown in SEQ ID NO: 28; the light chain sequences respectively include: CDRL1 containing the amino acid sequence shown in SEQ ID NO: 29, CDRL2 containing the amino acid sequence shown in SEQ ID NO: 99, and CDRL3 containing the amino acid sequence shown in SEQ ID NO: 31; and binds to human CD3 and cynomolgus monkey CD3. (2) The antibody or the antigen-binding fragment of the antibody according to the above (1), characterized in that the first X of the above CDRH2 aa is selected from the group consisting of (A, E, G, H, I, L, T, V, R, S), and the second X aa is S, or the first X aa is N, and the second X aais selected from the group consisting of (E, R, F, Y, L, V, I, K, T), and the X of the foregoing CDRL2 aa is selected from the group consisting of (Q, A, G, S, N, D), and binds to human CD3 and cynomolgus monkey CD3. (3) The antibody or antigen-binding fragment of the antibody as described in the foregoing (1) or (2), wherein the first X of the foregoing CDRH2 aa is selected from the group consisting of (R, S), the second X aa is S, and the X of the foregoing CDRL2 aa is selected from the group consisting of (Q, A, G, S, N, D), and binds to human CD3 and cynomolgus monkey CD3. (4) The antibody or antigen-binding fragment of the antibody as described in the foregoing (1), wherein the heavy chain sequence comprises a variable region having CDRH1, CDRH2, and CDRH3, the foregoing CDRH1 is composed of the amino acid sequence shown in SEQ ID NO: 26, the foregoing CDRH2 is composed of the amino acid sequence shown in SEQ ID NO: 27, the foregoing CDRH3 is composed of the amino acid sequence shown in SEQ ID NO: 28; and the light chain sequence comprises a variable region having CDRL1, CDRL2, CDRL3, the foregoing CDRL1 is composed of the amino acid sequence shown in SEQ ID NO: 29, the foregoing CDRL2 is composed of the amino acid sequence shown in SEQ ID NO: 30, the foregoing CDRL3 is composed of the amino acid sequence shown in SEQ ID NO: 31; and binds to human CD3 and cynomolgus monkey CD3. (5) The antibody or antigen-binding fragment of the antibody as described in any one of the foregoing (1) to (4), wherein the heavy chain variable region sequence comprises the amino acid sequence shown in SEQ ID NO: 100. (6) The antibody or antigen-binding fragment thereof as described in the foregoing (5), wherein the first X of the amino acid sequence shown in SEQ ID NO: 100 aa is selected from the group consisting of (A, E, G, H, I, L, T, V, R, S), and the second X aa is S, or the first X aa is N, and the second X aa is selected from the group consisting of (E, R, F, Y, L, V, I, K, T). (7) The antibody or antigen-binding fragment thereof as described in the foregoing (5), wherein the first X of the amino acid sequence shown in SEQ ID NO: 100 aais selected from the group consisting of (R, S), and the second X aa is S. (8) The antibody or antigen-binding fragment thereof according to any one of the foregoing (1) to (7), wherein the light chain variable region comprises the amino acid sequence shown in any one of SEQ ID NOs: 101, 102, and 103. (9) The antibody or antigen-binding fragment thereof according to the foregoing (8), wherein the X of the amino acid sequence shown in any one of SEQ ID NOs: 101, 102, and 103 aa is selected from the group consisting of (Q, A, G, S, N, D). (10) The antibody or antigen-binding fragment of the antibody according to the foregoing (5), wherein the heavy chain variable region sequence comprises the amino acid sequence shown in SEQ ID NO: 16. (11) The antibody or antigen-binding fragment of the antibody according to the foregoing (8), wherein the light chain variable region sequence comprises the amino acid sequence shown in any one of SEQ ID NOs: 17, 20, and 23. (12) The antibody or antigen-binding fragment thereof according to the foregoing (1) or (2), which comprises: a heavy chain variable region comprising the amino acid sequence shown in SEQ ID NO: 100, and a light chain variable region comprising the amino acid sequence shown in any one of SEQ ID NOs: 101, 102, and 103, the first X of the amino acid sequence shown in SEQ ID NO: 100 aa is selected from the group consisting of (A, E, G, H, I, L, T, V, R, S), and the second X aa is S, or the first X aa is N, and the second X aa is selected from the group consisting of (E, R, F, Y, L, V, I, K, T), the X of the amino acid sequence shown in any one of SEQ ID NOs: 101, 102, and 103 aa is selected from the group consisting of (Q, A, G, S, N, D). (13) The antibody or antigen-binding fragment thereof according to the foregoing (12), wherein the first X of SEQ ID NO: 100 aa is selected from the group consisting of (R, S), the second X aa is S, and the X of the amino acid sequence shown in any one of SEQ ID NOs: 101, 102, and 103 aais selected from the group consisting of (Q, A, G, S, N, D). (14) The antibody or antigen-binding fragment of the antibody according to the foregoing (13) is: an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 60 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 60; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 64 and a light chain variable region comprising amino acid residues 135 to 241 of SEQ ID NO: 64; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 66 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 66; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 68 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 68; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 70 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 70; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 72 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 72; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 74 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 74; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 76 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 76; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 78 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 78; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 80 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 80; an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 82 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 82; or an antibody or antigen-binding fragment of the antibody containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 84 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 84.(15) The antibody or an antigen-binding fragment thereof as described in the foregoing (1), (4), (5), (8), (10), or (11), which comprises a heavy-chain variable region containing the amino acid sequence shown in SEQ ID NO: 16, a linker, and a light-chain variable region containing the amino acid sequence shown in any one of SEQ ID NOs: 17, 20, and 23. (16) The antibody or its antigen-binding fragment according to any one of the foregoing (1) to (15), wherein the variable regions are linked in the order of the heavy-chain variable region and the light-chain variable region, or in the order of the light-chain variable region and the heavy-chain variable region, from the amino-terminal side of the antibody, and optionally: i) having a linker between the two variable regions; ii) having a glycine residue at the amino terminus of the variable region on the amino-terminal side; iii) having a linker, a FLAG tag, and / or a His tag attached to the carboxyl terminus of the variable region on the carboxyl-terminal side. (17) The antigen or the antigen-binding fragment of the antibody according to the foregoing (16), which comprises: an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 60; an amino acid sequence containing the amino acid residues at positions 2 to 241 of SEQ ID NO: 64; an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 66; an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 68; an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 70; an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 72; an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 74; an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 76; an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 78; an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 80; an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 82; or an amino acid sequence containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 84.(18) The antigen or antigen-binding fragment of the antibody as described in (16) above, which comprises: an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 19, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 22, an amino acid sequence comprising amino acid residues 2 to 267 of SEQ ID NO: 25, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 60, an amino acid sequence comprising amino acid residues 2 to 267 of SEQ ID NO: 64, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 66, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 68, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 70, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 72, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 74, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 76, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 78, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 80, an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 82, or an amino acid sequence comprising amino acid residues 2 to 269 of SEQ ID NO: 84. (19) An antibody or antigen-binding fragment of the antibody, which comprises an amino acid sequence and binds to human CD3 and cynomolgus monkey CD3, wherein the amino acid sequence is encoded by a nucleotide sequence contained in a polynucleotide that hybridizes under stringent conditions to the complementary strand of the following polynucleotide: a polynucleotide encoding the amino acid sequence contained in the antibody or antigen-binding fragment of the antibody as described in any one of (14) to (18) above. (20) An antibody or antigen-binding fragment of the antibody, which comprises a heavy chain and a light chain, wherein the heavy chain comprises an amino acid sequence that is at least 90% identical to the amino acid sequence of the heavy chain contained in the antibody or antigen-binding fragment of the antibody as described in any one of (14) to (18) above, and the light chain comprises an amino acid sequence that is at least 70% identical to the amino acid sequence of the light chain contained in the antibody or antigen-binding fragment of the antibody as described in any one of (14) to (18) above; and binds to human CD3 and cynomolgus monkey CD3.(21) An antibody or an antigen-binding fragment thereof, which comprises a heavy chain and a light chain, wherein the heavy chain comprises: an amino acid sequence in which 1 to several amino acids are substituted, deleted or added in the amino acid sequence of the heavy chain contained in the antibody or the antigen-binding fragment thereof described in any one of (14) to (18) above; and the light chain comprises: an amino acid sequence in which 1 to several amino acids are substituted, deleted or added in the amino acid sequence of the light chain contained in the antibody or the antigen-binding fragment thereof described in any one of (14) to (18) above; and binds to human CD3 and cynomolgus monkey CD3. (22) An antibody or an antigen-binding fragment thereof, which binds to the same site on human CD3 to which the antibody or the antigen-binding fragment thereof described in any one of (14) to (18) above binds, and binds to cynomolgus monkey CD3. (23) An antibody or an antigen-binding fragment thereof, which competes for binding to human CD3 with the antibody or the antigen-binding fragment thereof described in any one of (14) to (18) above, and binds to cynomolgus monkey CD3. (24) The antibody or the antigen-binding fragment thereof described in (22) above, wherein the site on human CD3 to which the antibody or the antigen-binding fragment thereof binds is composed of 7 or more amino acids among serine (Ser) at position 55, glutamic acid (Glu) at position 56, leucine (Leu) at position 58, tryptophan (Trp) at position 59, asparagine (Asn) at position 65, isoleucine (Ile) at position 66, serine (Ser) at position 77, aspartic acid (Asp) at position 78, arginine (Arg) at position 101, glycine (Gly) at position 101, serine (Ser) at position 103, lysine (Lys) at position 104, and proline (Pro) at position 105 in the amino acid sequence shown in SEQ ID NO: 1. (25) The antibody or the antigen-binding fragment thereof described in any one of (1 to 17), (19) to (24) above, which is IgG. (26) The antibody or the antigen-binding fragment thereof described in any one of (1) to (23) above, which is selected from the group consisting of Fab, F(ab)', Fv, scFv and sdAb. (27) The antibody or the antigen-binding fragment thereof described in any one of (1) to (17), (19) to (25) above, which is a humanized antibody or a human antibody comprising a human immunoglobulin constant region. (28) A polynucleotide, which comprises a nucleotide sequence encoding the amino acid sequence of the antibody or the antigen-binding fragment thereof described in any one of (1) to (27) above.(29) The polynucleotide as described in (27) above, which comprises a nucleotide sequence encoding the following amino acid sequences: an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 19, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 22, an amino acid sequence comprising the amino acid residues at positions 2 to 241 of SEQ ID NO: 25, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 60, an amino acid sequence comprising the amino acid residues at positions 2 to 241 of SEQ ID NO: 64, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 66, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 68, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 70, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 72, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 74, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 76, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 78, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 80, an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 82, or an amino acid sequence comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 84. (30) A vector which comprises the polynucleotide as described in any one of (28) or (29) above. (31) A cell which comprises the polynucleotide as described in any one of (28) or (29) above or the vector as described in (30) above, or produces the antibody as described in any one of (1) to (27) above or an antigen-binding fragment of the antibody. (32) A method for producing an antibody or an antigen-binding fragment of the antibody which binds to human CD3 and cynomolgus monkey CD3, which comprises: culturing the cell as described in (31) above; and recovering from the culture an antibody or an antigen-binding fragment of the antibody which binds to human CD3. (33) An antibody or an antigen-binding fragment of the antibody which binds to human CD3 and cynomolgus monkey CD3, which is obtained by the method as described in (32) above. (34) A pharmaceutical composition which contains, as an active ingredient, the antibody as described in any one of (1) to (27), (33) above or an antigen-binding fragment of the antibody. (35) A molecule having antigen-binding properties, which comprises the antibody as described in any one of (1) to (27), (33) above or an antigen-binding fragment of the antibody. (36) The molecule as described in (35) above, which is multispecific.(37) A molecule as described in (35) or (36) above, which comprises an antibody or an antigen-binding fragment of the antibody as described in any one of (1) to (27), (33) above, and one or more additional antibodies or antigen-binding fragments of the antibody. (38) A molecule as described in (37) above, wherein the antigen-binding fragment of the additional antibody is Fab, F(ab)’, Fv, scFv, or sdAb. (39) A molecule as described in (38) above, which comprises Fc. (40) A molecule as described in any one of (37) to (39) above, wherein the additional antibody is a humanized antibody or a human antibody comprising a human immunoglobulin constant region. (41) A molecule as described in any one of (37) to (38) above, wherein the additional antibody or the antigen-binding fragment of the antibody is bound to the antibody or the antigen-binding fragment of the antibody as described in any one of (1) to (27), (33) above by a linker or is not bound by a linker. (42) A molecule as described in (41) above, wherein the carboxyl terminus of the amino acid sequence of the additional antibody or the antigen-binding fragment of the antibody is bound to the linker, and further the carboxyl terminus of the amino acid sequence of the linker is bound to the antibody or the antigen-binding fragment of the antibody as described in any one of (1) to (27), (33) above. (43) A molecule as described in (42) above, which comprises: the carboxyl terminus of the amino acid sequence of the additional antibody or the antigen-binding fragment of the antibody is bound to the linker, and further comprises the amino acid sequence formed by binding the carboxyl terminus of the amino acid sequence of the linker to the following antibody or the antigen-binding fragment of the antibody: an antibody or the antigen-binding fragment of the antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 19, an antibody or the antigen-binding fragment of the antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 22, or an antibody or the antigen-binding fragment of the antibody comprising the amino acid residues at positions 2 to 241 of SEQ ID NO: 25. (44) A molecule as described in any one of (35) to (42) above, in the antibody or the antigen-binding fragment of the antibody as described in any one of (1) to (27), (33) above, the variable regions are bound in the order of the heavy chain variable region and the light chain variable region from the amino-terminal side of the antibody, or in the order of the light chain variable region and the heavy chain variable region, optionally: i) having a linker between the two variable regions; ii) having a glycine residue at the amino terminus of the variable region on the amino-terminal side; iii) at the carboxyl terminus of the variable region on the carboxyl-terminal side, a linker, a FLAG tag, and / or a His tag are bound. In applicable forms, it includes hybrid-type and dual-type bispecific molecules.(45) A molecule as described in the foregoing (44), which comprises the additional antibody or an antigen-binding fragment of the antibody, and an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 19 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 241 of SEQ ID NO: 25 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 60 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 241 of SEQ ID NO: 64 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 66 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 68 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 70 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 72 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 74 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 76 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 78 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 80 or an antigen-binding fragment of the antibody, an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 82 or an antigen-binding fragment of the antibody, or an antibody comprising the amino acid residues at positions 2 to 243 of SEQ ID NO: 84 or an antigen-binding fragment of the antibody. In applicable forms, it comprises hybrid-type and dual-type bispecific molecules. (46) A molecule as described in any one of the foregoing (36) to (45), wherein the additional antibody is an anti-cancer target antibody. (47) A molecule as described in any one of the foregoing (36) to (46), which is bispecific. (48) A molecule as described in any one of the foregoing (35) to (47), which is a polypeptide. (49) A polynucleotide, which comprises a nucleotide sequence encoding the amino acid sequence of the molecule as described in the foregoing (48). (50) A vector, which contains the polynucleotide as described in the foregoing (49). (51) A cell, which produces the polynucleotide as described in the foregoing (49) or the vector as described in the foregoing (50), or the molecule as described in the foregoing (48). (52) A method for manufacturing a molecule that binds to human CD3 and cynomolgus monkey CD3, which comprises the steps of culturing the cell as described in the foregoing (51), and recovering from the culture a molecule that binds to human CD3. (53) A molecule that binds to human CD3 and cynomolgus monkey CD3, which is obtained by the method as described in the foregoing (52).(54) A pharmaceutical composition containing, as an active ingredient, a molecule described in any one of the foregoing (35) to (48), (53). (55) The pharmaceutical composition described in the foregoing (54), characterized by redirecting T cells to target cells and inducing cytotoxicity against the target cells. Effects of the Invention According to the present invention, a novel anti-CD3 antibody that binds to human CD3 and binds to cynomolgus monkey CD3 or an antigen-binding fragment of the antibody, and a novel molecule containing the antibody or the like and having antigen-binding properties can be obtained. Furthermore, a novel pharmaceutical composition containing such an antibody or the like or a molecule as an active ingredient can be obtained. The antibody or the like or the molecule has T cell-dependent cytotoxic activity and is useful as a therapeutic agent or prophylactic agent for various diseases, such as cancer. Modes for Carrying Out the Invention 1. Definitions In the present invention, "gene" means a nucleotide or its complementary strand contained in a base sequence encoding an amino acid of a protein. For example, it includes polynucleotides, oligonucleotides, DNA, mRNA, cDNA, cRNA, etc., which are nucleotides or their complementary strands contained in a base sequence encoding an amino acid of a protein, within the meaning of "gene". The gene is a nucleotide of single-stranded, double-stranded or triple-stranded or more, a combination of a DNA strand and an RNA strand, a mixture of ribonucleotides (RNA) and deoxyribonucleotides (DNA) on a single-stranded nucleotide chain, and a double-stranded or triple-stranded or more nucleotide containing such a nucleotide chain are also included within the meaning of "gene". In the present invention, the base sequence and the nucleotide sequence have the same meaning. In the present invention, "polynucleotide", "nucleic acid" and "nucleic acid molecule" have the same meaning. For example, DNA, RNA, probes, oligonucleotides, primers, etc. are also included within the meaning of "polynucleotide". The polynucleotide includes polynucleotides of single-stranded, double-stranded or triple-stranded or more chains, a combination of a DNA strand and an RNA strand, a mixture of ribonucleotides (RNA) and deoxyribonucleotides (DNA) on a single-stranded polynucleotide chain, and a combination of double-stranded or triple-stranded or more chains containing the polynucleotide chain are also included within the meaning of "polynucleotide". In the present invention, "polypeptide", "peptide" and "protein" have the same meaning. In the present invention, "antigen" is sometimes used in the sense of "immunogen". In the present invention, "cell" also includes various cells derived from an animal individual, subcultured cells, primary cultured cells, cell lines, recombinant cells, and microorganisms. In the present invention, "antibody" has the same meaning as immunoglobulin. However, in the case of the "antibody" referred to as an anti-CD3 antibody in the present invention, "antibody" is used in the sense of an immunoglobulin having a constant region and a variable region. The antibody is a natural or immunoglobulin produced by partial or total synthesis, but is not particularly limited. The anti-CD3 antibody of the present invention is included in the "molecule" described later. The structure of a basic four-chain antibody is composed of two identical light (L) chains and two identical heavy (H) chains. The light chain is bound to the heavy chain by one covalent disulfide bond. The two heavy chains are bound to each other by one or a plurality of disulfide bonds according to the isotype of the heavy chain. Each light chain and heavy chain have regular intervals and have intra-chain disulfide bonds. In the heavy chain and the light chain, there are a constant region where the amino acid sequences show very high similarity and a variable region where the similarity of the amino acid sequences is low. The light chain is connected to the constant region (CL) and has a variable region (VL) at the amino terminus. The heavy chain is connected to three constant regions (CH1 / CH2 / CH3) and has a variable region (VH) at the amino terminus. VL and VH are paired, and CL is aligned with the first constant region (CH1) of the heavy chain. VL and VH are paired to form a single antigen-binding site. Regarding the constant region of the antibody of the present invention, although not particularly limited, for the antibody of the present invention used for treating or preventing human diseases, it is preferably a human antibody. Examples of the heavy chain constant region of a human antibody include, for example, Cγ1, Cγ2, Cγ3, Cγ4, Cμ, Cδ, Cα1, Cα2, Cε, etc. Examples of the light chain constant region of a human antibody include, for example, Cκ, Cλ, etc. Fab is composed of CH1 of the heavy chain and VH following it, and CL of the light chain and VL following it. VH and VL contain complementarity-determining regions (CDRs). Fc is the carboxyl-terminal region of the constant region of the heavy chain, contains CH2 and CH3, and is a dimer. The Fc of the present invention can be Fc of a natural sequence or a variant Fc with a mutation added to the natural sequence. The variable region is composed of a region having extremely high variability called a hypervariable region (HVR) and a relatively invariant region called a framework region (FR) separated by this region. The variable regions of natural heavy and light chains contain four FRs connected by three hypervariable regions. The hypervariable regions of each chain are kept very close together with the hypervariable regions of the other chain by the FRs, which helps to form the antigen-binding site of the antibody. In the heavy chain and light chain of an antibody molecule, there are known to be three Complementarity Determining Regions (CDRs) each. The Complementarity Determining Region is also known as the hypervariable region. It is a region with extremely high variability in the primary structure within the variable regions of the heavy chain and light chain of an antibody. On the primary structure of the polypeptide chains of the heavy chain and light chain, generally, they are separated at three positions each. In the present invention, regarding the Complementarity Determining Region of an antibody, the Complementarity Determining Region of the heavy chain is labeled as CDRH1, CDRH2, CDRH3 from the amino-terminal side of the heavy chain amino acid sequence, and the Complementarity Determining Region of the light chain is labeled as CDRL1, CDRL2, CDRL3 from the amino-terminal side of the light chain amino acid sequence. These regions are close to each other in the three-dimensional structure and determine the specificity for the antigen to which they bind. In the present invention, the positions and lengths of the CDRs are determined by the definition of IMGT (Developmental and Comparative Immunology 27 (2003) 55-77). The Framework Region (FR) is the variable region other than the CDR residues. Generally, the variable region has four FRs, namely FR1, FR2, FR3, and FR4. The positions of the CDR and FR are various definitions well-known in the art. For example, in addition to IMGT, they can also be determined according to the definitions such as Kabat, Chothia, AbM, contact, etc. In the present invention, the "antigen-binding fragment of an antibody" means a partial fragment of an antibody having binding activity to an antigen, which is composed of the heavy chain variable region and the light chain variable region. Examples of the "antigen-binding fragment of an antibody" include, but are not limited to, antigen-binding fragments such as Fab, F(ab’) 2 , scFv, Fab’, Fv, single-domain antibody(sdAb), etc. The antigen-binding fragment of the antibody can be not only obtained by treating the full-length molecule of the antibody protein with enzymes such as papain and pepsin, but also a recombinant protein produced using recombinant genes in appropriate host cells. In the present invention, the "site" to which the antibody binds, that is, the "site" recognized by the antibody, means a partial peptide or partial higher-order structure on the antigen to which the antibody binds or recognizes. In the present invention, this site is also referred to as an antigenic determinant site or a binding site of an antibody. In the present invention, an "antibody variant" means an amino acid sequence formed by amino acid substitution, deletion, addition (including insertion in the case of addition) (hereinafter collectively referred to as "mutation") in the amino acid sequence of the original antibody, and a polypeptide that binds to CD3 of the present invention. The number of mutated amino acids in the antibody variant is 1 to 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 15, 20, 25, 30, 40, or 50. The antibody variant is also included in the "antibody" of the present invention. In the present invention, "several" in "1 to several" means 2 to 10. In this specification, "molecule" includes molecules such as the above-mentioned antibody and antigen-binding fragments of the antibody, and further includes multispecific molecules formed by an antibody or a plurality of antigen-binding fragments derived therefrom. In this specification, "multispecific molecule", "multispecific (molecule)", and "multispecific molecule" have the same meaning, and any molecule that can bind to a plurality of different antigenic determinant sites on one molecule and / or can bind to different antigenic determinant sites on two or more molecules is acceptable, and there is no particular limitation. Regarding the multispecific molecule, it also includes an antibody containing a heavy chain variable region (VH) and a light chain variable region (VL). In such a multispecific molecule, it includes a full-length antibody molecule having two or more different types of heavy chains and light chains, that is, an IgG-type multispecific molecule, and a molecule formed by antigen-binding fragments having two or more types of VL and VH, that is, a molecule derived from a combination of Fab, Fab', Fv, scFv, sdAb, etc., that is, including tandem scFv, diabody, single-chain diabody, triabody, etc., but is not limited thereto. Molecules generated by genetically or chemically linking a protein having antigen-binding properties without an immunoglobulin backbone to other antigen-binding fragments are also included in the multispecific molecule. Regarding the activities and properties of the anti-CD3 antibody of the present invention, an antigen-binding fragment of the antibody, or the multispecific molecule of the present invention, for example, biological activities, physicochemical properties, etc. can be cited. Specifically, various biological activities, binding activities to antigens or antigenic determinant sites, stability during production or storage, thermal stability, etc. can be cited. In the present invention, "hybridization under stringent conditions" is carried out at 65°C in a solution containing 5×SSC, followed by 20 minutes at 65°C in an aqueous solution containing 2×SSC - 0.1% SDS, 20 minutes at 65°C in an aqueous solution containing 0.5×SSC - 0.1% SDS, and 20 minutes at 65°C in an aqueous solution containing 0.2×SSC - 0.1% SDS, and hybridization is carried out under each washing condition or equivalent conditions. SSC means an aqueous solution of 150 mM NaCl - 15 mM sodium citrate, and n×SSC means SSC at n-fold concentration. In the present invention, "cytotoxic" refers to any means that causes pathological changes to cells, including not only direct trauma, but also all damage to cell structures or functions such as DNA cleavage or dimer formation of bases, chromosome cleavage, damage to the cell division apparatus, and reduction of various enzyme activities. In the present invention, "cytotoxic activity" means that which causes the above-mentioned cytotoxicity. In the present invention, "antibody-dependent cellular cytotoxicity activity" refers to "antibody dependent cellular cytotoxicity (ADCC) activity", and also refers to the activity of NK cells to damage target cells such as tumor cells through antibodies. In the present invention, "cytotoxic activity caused by T cell redirection" means causing the above-mentioned cytotoxicity by a multi-specific molecule containing an anti-tumor antigen antibody and an anti-CD3 antibody. That is, the anti-tumor antigen antibody binds to the target tumor cell, the anti-CD3 antibody binds to the T cell, the distance between the target tumor cell and the T cell is shortened, and cytotoxicity is induced by T cell activation. This molecule can be included in a pharmaceutical composition. 2. Antigen protein CD3 (CD3 complex) In the present invention, "CD3" is used in the same sense as the CD3 protein. CD3 is expressed on T cells as part of a multi-molecular T cell receptor complex and is a complex of five types of polypeptides, namely γ chain, δ chain, ε chain, ζ chain, and η chain (molecular weights are 25,000 - 28,000, 21,000, 20,000, 16,000, and 22,000 in sequence). CD3 used in the present invention can be purified from animal tissues (including body fluids), cells derived from the tissues, or cell cultures thereof and prepared by isolation, gene recombination, in vitro translation, chemical synthesis, etc. The nucleotide sequence of the cDNA encoding human CD3ε is registered in GenBank under the accession number: NM_000733.3. The nucleotide sequence of the cDNA encoding cynomolgus monkey CD3 is registered in GenBank under the accession number: NM_001283615.1. The amino acid sequence of human CD3ε is set forth in SEQ ID NO: 1 of the Sequence Listing. The cDNA of CD3ε can be obtained by, for example, a so-called PCR method of polymerase chain reaction (hereinafter referred to as "PCR") (Saiki, R. K., et al., Science (1988) 239, 487-49) using a primer that specifically amplifies the cDNA of CD3ε with a cDNA library of an organ expressing mRNA of CD3ε as a template. Furthermore, the nucleotide sequence of a protein having an amino acid sequence in which one to several amino acids are substituted, deleted, or added in the amino acid sequence encoded by CD3 and having biological activity equivalent to that of CD3 is also included in the nucleotide sequence of the CD3 gene. Also, a protein having an amino acid sequence in which one or several amino acids are substituted, deleted, or added in the amino acid sequence of CD3 and having biological activity equivalent to that of CD3 is also included in CD3. Furthermore, a polynucleotide that hybridizes under stringent conditions with a nucleotide sequence complementary to the nucleotide sequence encoding human or cynomolgus monkey CD3ε and encodes a protein having biological activity equivalent to that of CD3ε is also included in the cDNA of CD3ε. Furthermore, a splicing variant transcribed from the human or cynomolgus monkey CD3ε gene locus or a polynucleotide that hybridizes with it under stringent conditions and encodes a protein having biological activity equivalent to that of CD3ε is also included in the cDNA of CD3ε. 3. Anti-CD3 antibody (3-1) Classification of antibodies The anti-CD3 antibody of the present invention and an antigen-binding fragment of the antibody (hereinafter, also referred to as the antibody of the present invention, etc.) can be either a monoclonal antibody or a polyclonal antibody. Regarding the monoclonal antibody of the present invention, antibodies derived from non-human animals (non-human animal antibodies), human antibodies, chimerized antibodies (also referred to as "chimeric antibodies"), humanized antibodies, etc. can be mentioned. Regarding non-human animal antibodies, antibodies derived from vertebrates such as mammals and birds can be mentioned. Regarding antibodies derived from mammals, antibodies derived from rodents such as mouse antibodies and rat antibodies can be mentioned. Regarding antibodies derived from birds, chicken antibodies can be mentioned. Regarding the anti-human CD3 rat monoclonal antibody, C3-147 (Example 1)-7) of the present invention can be mentioned. Examples of chimeric antibodies include antibodies formed by combining the variable regions derived from non-human animal antibodies with the constant regions of human antibodies (human immunoglobulins), but are not limited thereto. Examples of humanized antibodies include those in which the CDRs in the variable regions of non-human animal antibodies are transplanted into human antibodies (variable regions of human immunoglobulins), those in which, in addition to the CDRs, partial framework region sequences of non-human animal antibodies are transplanted into human antibodies, and those in which one or more amino acids derived from non-human animal antibodies among these are replaced with human-type amino acids, etc., but are not limited thereto. Examples of the CDRs in the variable regions of non-human animal antibodies include CDRH1 to CDRH3 in the heavy chain variable region and CDRL1 to CDRL3 in the light chain variable region of the rat anti-CD3 antibody C3E-7000 derived in the present invention, and those in which the amino acid sequences of these CDRs have one or two amino acids replaced with other amino acids, etc. Examples of human antibodies include those that preferably bind to human CD3, more preferably those that bind to both human CD3 and cynomolgus monkey CD3, and are not particularly limited. Examples also include human antibodies that bind to the same sites as the humanized antibodies of the present invention. For example, human antibodies that bind to the same site as C3E-7034 can be exemplified. Preferred antibodies and the like of the present invention bind to human CD3. Furthermore, more preferred antibodies and the like of the present invention have binding activity to cynomolgus monkey CD3. If the antibodies and the like of the present invention bind to human CD3 and also bind to cynomolgus monkey CD3, they can be antibodies composed of parts derived from a plurality of different antibodies. Examples include those in which the heavy chain and / or light chain are exchanged between a plurality of different antibodies, those in which the entire lengths of the heavy chain and / or light chain are exchanged, those in which only the variable regions are exchanged or only the constant regions are exchanged, those in which all of the CDRs are exchanged or only a part of the CDRs is exchanged, etc. The heavy chain variable region and the light chain variable region of the chimeric antibody can be derived from different antibodies of the present invention. The CDRH1 to CDRH3 and CDRL1 to CDRL3 in the variable regions of the heavy chain and light chain of the humanized antibody can be derived from two or more antibodies of the present invention. The CDRH1 to CDRH3 and CDRL1 to CDRL3 in the variable regions of the heavy chain and light chain of the human antibody can be a combination of CDRs possessed by two or more antibodies of the present invention. The isotype of the monoclonal antibody of the present invention is not particularly limited, and examples include IgG such as IgG1, IgG2, IgG3, IgG4; IgA such as IgM, IgA1, IgA2; IgD; IgE, etc. The isotype and subtype of the monoclonal antibody can be determined, for example, by the Ouchterlony method, the Enzyme-Linked ImmunoSorbent Assay (ELISA) method, the Radio Immuoassay (RIA) method, etc., and commercially available identification kits (such as Rat Immunoglobulin Isotyping ELISA Kit (BD Pharmingen)) can also be used. (3-2) Binding specificity of anti-CD3 antibody The antibody of the present invention, etc. recognizes CD3. That is, the antibody of the present invention, etc. binds to CD3. The antibody of the present invention, etc. preferably binds to human CD3, cynomolgus monkey CD3, etc., and more preferably binds to human CD3 and cynomolgus monkey CD3. More specifically, the antibody, antigen-binding fragment thereof, and variable region of the present invention bind to the Ig-like domain in the extracellular region of the ε chain (Figure 1, SEQ ID NO: 1) present in the human CD3 complex. Furthermore, it also binds to the Ig-like domain in the extracellular region of the ε chain present in the cynomolgus monkey CD3 complex. The antigenic determinant site in the extracellular region of the ε chain (Figure 1, SEQ ID NO: 1) present in the human CD3 complex to which the antibody of the present invention, etc. binds includes the following amino acids: Ser55, Glu56, Leu58, Trp59, Asn65, Ile66, Ser77, Asp78, Arg101, Gly102, Ser103, Lys104, and Pro105. The antibody of the present invention, etc. can preferably maintain binding to human CD3 by binding to an antigenic determinant region containing at least 7 amino acids selected from these 13 amino acids. When the antibody is adjacent to the above amino acids at a distance within 4 Å, such an antibody can be judged to have the same antigenic determinant specificity as the antibody of the present invention, etc. On the other hand, among the amino acids of the above antigenic determinant, Arg101, Gly102, Ser103, Lys104, and Pro105 are also antigenic determinant residues that interact with the known anti-CD3 antibodies OKT3 or UCHT1 (Lars Kjer-Nielsen et al., PNAS (2004) (Kelly L Arnett et al., PNAS (2004)). However, although OKT3 and UCHT1 bind to human CD3, they do not bind to cynomolgus monkey CD3. In the present invention, "recognition", that is, "binding", means not non-specifically adsorbing and binding. As for the criterion for determining whether recognition, that is, whether binding occurs, for example, the dissociation constant (hereinafter referred to as "KD") can be cited. The KD value of the preferred antibody of the present invention and the like for CD3 is 1×10 -5 M or less, 5×10 -6 M or less, 2×10 -6 M or less or 1×10 -6 M or less. The binding between the antigen and the antibody in the present invention can be measured or determined by using an in-vivo molecular interaction analysis system such as the SPR method, the BLI method, or the ELISA method, the RIA method, etc. The binding between the antigen expressed on the cell surface and the antibody can be measured by using a flow cytometry method, etc. The SPR method (Surface Plasmon Resonance analysis method) is used as an analysis method for obtaining the dissociation constant (KD value) and other affinity indicators by measuring the association rate constant (Ka value) and the dissociation rate constant (Kd value) analyzed by chemical kinetics (kinetics). Examples of the instrument used for SPR analysis include Biacore (trademark) (manufactured by GE HEALTHCARE), ProteOn (trademark) (manufactured by BioRad), SPR-Navi (trademark) (manufactured by BioNavis), Spreeta (trademark) (manufactured by Texas Instruments), SPRi-PlexII (trademark) (manufactured by HORIBA), Autolab SPR (trademark) (manufactured by Metrohm), etc. The BLI method (BioLayer Interferometry) is a method for measuring the in-vivo molecular interaction using biofilm layer interference. Examples of the instrument used for interaction analysis using the BLI method include the Octet system (manufactured by Pall ForteBio), etc. The ELISA method is a method for detecting and quantifying the target antigen or antibody contained in a sample solution while capturing it with a specific antibody or antigen, and using an enzyme reaction. An antigen or antibody labeled with an enzyme is incorporated into the reaction system to detect the enzyme activity. For the detection of enzyme activity, a substrate with an absorbance spectrum change according to the reaction is used, and quantification is performed by measuring the absorbance. Cell-ELISA is a method of detecting and quantifying by using an enzyme reaction while supplementing each cell with the measurement target on the cell surface. The RIA method (Radio Immunoassay) labels an antibody with a radioactive substance and quantifies the antibody by measuring the radioactivity from the antibody. Flow cytometry is a technique that disperses fine cells in a fluid, forms the fluid into a thin stream, and optically analyzes each cell. An antibody labeled with a fluorescent dye binds to a cell surface antigen through an antigen-antibody reaction, and the antigen-binding property of the antibody is measured by measuring the number of cells bound to the antibody using fluorescence intensity. As described above, antibodies that bind to human CD3 and cynomolgus monkey CD3, etc., can be preferably supplied for various tests related to efficacy or safety in primates, particularly cynomolgus monkeys, which are essential for non-clinical development (pre-clinical development) of pharmaceuticals. Also, antibodies that bind to human CD3 and cynomolgus monkey CD3, etc., have cytotoxic activity and are useful for the treatment or prevention of diseases such as cancer in humans and cynomolgus monkeys, either alone or as the molecules of the present invention. The pharmaceutical composition is described below. Furthermore, since the antibodies of the present invention that bind to human CD3 and cynomolgus monkey CD3 do not bind to mouse CD3, they can be used in cells, tissues, and individuals of mice into which the human CD3 gene has been introduced (including transgenic animals, gene knockout animals, knock-in animals), and various analyses, immunohistochemistry, etc. using such antibodies can be carried out without being affected by mouse CD3 as the host. It is preferable for research and non-clinical development of pharmaceuticals, veterinary drugs, diagnostic drugs, etc. that use mice and contain such antibodies. (3-3) Monoclonal antibody The present invention provides monoclonal antibodies. Monoclonal antibodies include monoclonal antibodies derived from non-human animals such as rat antibodies, mouse antibodies, rabbit antibodies, chicken antibodies, fish antibodies, chimerized antibodies, humanized antibodies, human antibodies, antigen-binding fragments thereof, antibody variants thereof, modified forms thereof, etc. For example, C3-147 obtained by the method described in Example 1 is a rat monoclonal antibody against CD3. The nucleotide sequence of the DNA encoding the heavy chain variable region of C3-147 is described in Sequence ID No. 6 (Figure 14) of the Sequence Listing, and the amino acid sequence is described in Sequence ID No. 7 (Figure 15). The nucleotide sequence of the DNA encoding the light chain variable region of C3-147 is described in Sequence ID No. 8 (Figure 16) of the Sequence Listing, and the amino acid sequence is described in Sequence ID No. 9 (Figure 17). Preferably, the antibody variants of the present invention can be made to have reduced sensitivity to protein degradation or oxidation, maintain, improve or reduce and inhibit changes in biological activity and function, improve or regulate antigen-binding ability, or confer physicochemical properties or functional properties. It is known that a protein changes in its function or activity due to changes in specific amino acid side chains on its surface. Such examples include deamidation of the aspartic acid side chain, isomerization of the asparagine side chain, etc. In order to prevent such changes in amino acid side chains, substitution with other amino acids is included in the scope of the antibody variants of the present invention. Examples of the antibody variants of the present invention include antibodies having an amino acid sequence with conserved amino acid substitutions in the amino acid sequence possessed by the antibody. Conserved amino acid substitutions are substitutions that occur within an amino acid group related to the amino acid side chain. Suitable amino acid groups are as follows: acidic group = aspartic acid, glutamic acid; basic group = lysine, arginine, histidine; non-polar group = alanine, valine, leucine, isoleucine, proline, phenylalanine, methionine, tryptophan; and uncharged polar family = glycine, aspartic acid, glutamine, cysteine, serine, threonine, tyrosine. Other suitable amino acid groups are as follows: aliphatic hydroxyl group = serine and threonine; amide group-containing group = aspartic acid and glutamine; aliphatic group = alanine, valine, leucine and isoleucine; and aromatic group = phenylalanine, tryptophan and tyrosine. The amino acid substitution in the antibody variant is preferably carried out within a range that does not reduce the antigen-binding activity originally possessed by the antibody. Antibody variants having an amino acid sequence with conserved amino acid substitutions and / or other mutations in the amino acid sequence possessed by the above-mentioned (1) antibody containing C3-147 of the present invention or its antigen-binding fragment, and mouse antibodies, rat antibodies, chimeric antibodies, humanized antibodies, human antibodies, their antigen-binding fragments, and those containing them, in which any of the amino acid sequences of CDRH1 to CRDH3 and CDRL1 to CDRL3 contained in the above-mentioned (1) antibody containing C3-147 or its antigen-binding fragment have an amino acid sequence with conserved amino acid substitutions and / or other mutations, are also included in the present invention. (3-4) Antigen-binding fragment of anti-CD3 antibody Providing an antigen-binding fragment of the anti-CD3 antibody of the present invention is one aspect of the present invention. The antigen-binding fragment of an antibody means a fragment or its modification that retains at least the antigen-binding property among the functions possessed by the antibody. As the functions of the antibody, generally, antigen-binding activity, activity of regulating antigen activity, antibody-dependent cytotoxic activity, and complement-dependent cytotoxic activity can be cited. Regarding the functions of the antibody and multi-specific molecules containing the antibody of the present invention, for example, redirection of T cells, activation of T cells, and cytotoxic activity of cancer cells using activated T cells can be cited. Regarding the antigen-binding fragment of an antibody, any fragment of the antibody that retains at least the antigen-binding property among the activities possessed by the antibody is sufficient, and it is not particularly limited. For example, Fab, Fab’, F(ab’) 2 、Fv, single-chain Fv (scFv) in which the Fvs of the heavy chain and light chain are linked by an appropriate linker, single-domain antibody (sdAb), etc. can be cited, but it is not limited to these. Molecules such as scFv having a linker portion and containing portions other than the antigen-binding fragment of the antibody of the present invention are also included in the meaning of the antigen-binding fragment of the antibody of the present invention. Molecules in which one to several or more amino acids at the amino terminus and / or carboxy terminus of the antibody protein are deleted and at least a part of the functions possessed by the antibody are retained are also included in the meaning of the antigen-binding fragment of the antibody. Such modified forms of the antigen-binding fragment of the antibody are also included in the antibody or its antigen-binding fragment, or its modification (described later) of the present invention. One aspect of the antigen-binding fragment of the antibody of the present invention is scFv. ScFv is obtained by linking the variable region of the heavy chain and the variable region of the light chain of the antibody with a polypeptide linker (Pluckthun A. The Pharmacology of Monoclonal Antibodies 113, edited by Rosenberg and Moore, Springer Verlag, New York, 269-315 (1994), Nature Biotechnology (2005), 23, 1126-1136). Also, tandem scFv produced by binding two scFvs with a polypeptide linker can be used as a bispecific molecule. Further, trispecific antibodies and the like composed of three or more scFvs can also be used as multi-specific molecules. The antibody of the present invention can be an antibody having a single heavy chain variable region and no light chain sequence. Such an antibody is called a single domain antibody (sdAb) or a nanobody, and it has been reported that the antigen-binding ability is retained (Muyldemans S. et al., Protein Eng., (1994) 7(9), 1129-35, Hamers-Casterman C. et al., Nature (1993) 363(6428), 446-448). These antibodies are also included in the meaning of the antigen-binding fragments of the antibodies in the present invention. Furthermore, in the present invention, a single chain immunoglobulin is included which links the full-length sequences of the heavy chain and light chain of an antibody using an appropriate linker (Lee, H-S, et al., Molecular Immunology (1999) 36, 61-71; Shirrmann, T. et al., mAbs (2010), 2(1), 1-4). Such a single chain immunoglobulin may retain a structure and activity similar to that of an antibody which is originally a tetramer by dimerization. (3-5) Molecule having antigen-binding property The molecule of the present invention comprises the anti-CD3 antibody of the present invention or its antigen-binding fragment. The molecule of the present invention may further comprise the following: a signal sequence; a tag for purification, etc.; Gly at the amino terminus; a drug linker part of an ADC; an albumin-binding polypeptide; a polymer such as PEG; an antibody other than the anti-CD3 antibody, its antigen-binding fragment; a protein having antigen-binding property without an immunoglobulin backbone; a compound (part) having an anti-cancer effect, cytotoxic activity, other pharmacological activities, etc. The molecule of the present invention binds to human CD3 and cynomolgus monkey CD3. The molecule of the present invention comprises a multi-specific molecule described below. The molecule of the present invention can be in a form introduced into cells such as CAR-T or a form presented on the cell surface. (3-6) Multi-specific molecule, bispecific molecule The multispecific molecule of the present invention is a molecule having two or more antigen-binding sites. That is, a molecule that can bind to two or more different antigenic determinants on one molecule or to different antigenic determinants on two or more molecules, and includes a plurality of different antigen-binding fragments. Such multispecific molecules include IgG-type multispecific molecules, multispecific molecules having two or more types of variable regions, such as antibody fragments like tandem scFv, single-chain diabody, diabody, and triabody, antibody fragments linked by covalent or non-covalent bonds, but are not limited thereto. The multispecific molecule may include Fc. The multispecific molecule of the present invention includes the anti-CD3 antibody of the present invention or an antigen-binding fragment of the antibody. The multispecific molecule of the present invention includes the antibody of the present invention, etc., and one or two or more additional antibodies or antigen-binding fragments of the antibody. Examples of the antigen-binding fragment of the additional antibody include Fab, F(ab)', Fv, scFv, sdAb. The multispecific molecule of the present invention can specifically bind to CD3, or can further bind to a target such as an Fc receptor on an effector cell. As a preferred example of the multispecific molecule of the present invention, a bispecific molecule can be mentioned. "Bispecific" means capable of binding to two different antigenic determinants on the same molecule or to different antigenic determinants on two molecules, and includes an antibody or antigen-binding fragment having such bispecificity. The bispecific molecule of the present invention binds to CD3 and binds to an antigenic determinant of another antigen that does not have CD3. More specifically, the bispecific molecule (i) binds to a certain antigenic determinant (antigenic determinant 1) on CD3, and (ii) binds to an antigenic determinant (antigenic determinant 2) different from the antigenic determinant 1 on CD3 or binds to an antigenic determinant (antigenic determinant 3) of an antigen other than CD3. For example, in the tandem scFv type bispecific molecule represented by BiTE, the antigen-binding site of the heavy chain variable region of the first antibody and the antigen-binding site of the light chain variable region of the first antibody are linked by a linker or directly bound without a linker to form a first polypeptide. Also, the antigen-binding site of the heavy chain variable region of the second antibody and the antigen-binding site of the light chain variable region of the second antibody are linked by a linker or directly bound without a linker to form a second polypeptide. The first polypeptide and the second polypeptide are linked by a linker or directly bound without a linker. Also, the first polypeptide and the second polypeptide can be bound via another molecule. In a bispecific molecule of the diabody type, the antigen-binding site of the heavy-chain variable region of the first antibody and the antigen-binding site of the light-chain variable region of the second antibody are linked by a linker or directly bound without a linker. Also, the antigen-binding site of the light-chain variable region of the first antibody and the antigen-binding site of the heavy-chain variable region of the second antibody are linked by a linker or directly bound without a linker. Further, a bispecific molecule in which the diabody-type bispecific molecule is further dimerized can be produced. Also, one or both chains of a single chain of a diabody-type bispecific molecule and Fc can be linked by a linker (diabody-Fc type bispecific molecule). In a bispecific molecule of the dual scFv type, two scFvs that bind to different antigenic determinant sites are each linked to one of the Fcs of the dimer by a linker or directly bound without a linker. Alternatively, two types of scFvs that bind to different antigenic determinant sites are each linked to CH and CL by a linker, and then each of the Fcs of the dimer is linked to the linker. Hereinafter, the bispecific molecule of the dual scFv type will also be referred to as a dual-type bispecific molecule, or simply as a dual type. In a bispecific molecule of the IgG type, two Fabs that bind to different antigenic determinant sites are each linked to one of the Fcs of the dimer by a linker or directly bound without a linker. Hereinafter, the bispecific molecule of the IgG type will also be referred to as a full-size antibody (FSA) type bispecific molecule, or simply as an FSA type. Alternatively, as the bispecific molecule of the present invention, a Fab and an scFv that bind to different antigenic determinant sites can be a bispecific antibody in which each is linked to one of the Fcs of the dimer by a linker or directly bound without a linker. Alternatively, it can be a bispecific molecule in which, for one of the Fcs of the dimer, the Fab of the first antibody and the scFv of the second antibody are bound through a linker. Hereinafter, such a bispecific molecule will also be referred to as a hybrid-type bispecific molecule, or a hybrid type. The scFv and Fab contained in the bispecific molecule of the present invention are preferably scFv and Fab of a humanized antibody or a human antibody, and the Fc is preferably Fc of a human antibody. In the variable regions contained in the bispecific molecule of the present invention, the heavy chain variable region and the light chain variable region can be combined in this order from the amino-terminal side of the antibody, or the light chain variable region and the heavy chain variable region can be combined in this order. An arbitrary linker is provided between the two variable regions. Further, a glycine residue may be present (optionally) at the amino terminus of the variable region on the amino-terminal side. In the tandem scFv type bispecific molecule, the carboxyl terminus of the variable region on the carboxyl-terminal side may be bound to a linker, a FLAG tag, and / or a His tag (optionally). As one of the suitable embodiments, an example of the combination from the amino terminus is the heavy chain variable region, the first linker, the light chain variable region, the second linker, the FLAG tag, and the His tag. The linker also includes a single-chain polypeptide or single-chain oligopeptide, or a synthetic product such as PEG, nucleotide, sugar chain, compound, etc. Otherwise, any molecule that binds two polypeptides may be used without particular limitation, and it is possible to use a known linker. Regarding the length of the linker, for example, in the case of a peptide linker, it is 5 to 30 amino acids. In the case where the bispecific molecule contains a plurality of linkers, peptide linkers of all the same length may be used, or peptide linkers of different lengths may be used. Regarding the peptide linker, for example, the repetition of (Gly・Gly・Gly・Gly・Ser) may be exemplified, and one to several amino acid residues different from Gly and Ser may be added thereto. (3-7) Humanized anti-CD3 antibody In one aspect, the present invention provides a humanized anti-CD3 antibody or an antigen-binding fragment thereof. Examples of the humanized antibody of the present invention include an antibody that incorporates only the complementarity determining region (CDR) into a human-derived antibody (Nature (1986) 321, 522-525); an antibody in which, by the CDR transplantation method, in addition to the CDR sequence, a part of the framework amino acid residues are also transplanted into a human antibody (International Patent Publication No. WO1990 / 007861). The heavy chain variable region contained in the suitable humanized anti-CD3 antibody of the present invention or the antigen-binding fragment of the antibody retains: CDRH1 (GVTFNYYG) consisting of the amino acid sequence shown by SEQ ID NO: 26 (Figure 24), CDRH2 (ITX aa X aa GGRI) (wherein the first X aa and the second X aaare each an arbitrary natural amino acid residue. Hereinafter, the first X of CDRH2 will also be aa described as X1, and the second X aa described as X 1 、X 2 。), and CDRH3 (TLDGRDGWVAY) consisting of the amino acid sequence shown in SEQ ID NO: 28 (Figure 26). Furthermore, the light chain variable region contained in the suitable humanized anti-CD3 antibody or the antigen-binding fragment of the antibody of the present invention has: CDRL1 (TGNIGSNY) consisting of the amino acid sequence shown in SEQ ID NO: 29 (Figure 27), CDRL2 (RX aa D) (wherein X aa is an arbitrary natural amino acid residue. Hereinafter, the X of CDRL2 will also be aa described as X 3 。), and CDRL3 (QSYSSGFI) consisting of the amino acid sequence shown in SEQ ID NO: 31 (Figure 29). In the above CDRH2 (ITX 1 X 2 GGRI), preferably X 1 is selected from the group consisting of (A, E, G, H, I, L, T, V, R, S), and X 2 is S; or X 1 is N, and X 2 is selected from the group consisting of (E, R, F, Y, L, V, I, K, T), and in the above CDRL2 (RX 3 D), preferably X 3 is selected from the group consisting of (Q, A, G, S, N, D). In the above CDRH2 (ITX 1 X 2GGRI), more preferably X 1 is selected from the group consisting of (R, S), and X 2 is S, in the above CDRL2 (RX 3 D), more preferably X 3 is selected from the group consisting of (Q, A, G, S, N, D). For an example of the heavy chain variable region contained in such a humanized anti-CD3 antibody of the present invention or an antigen-binding fragment of the antibody, there may be mentioned a heavy chain variable region comprising the amino acid residues shown in SEQ ID NO: 100 (Figure 114). Further, for an example of the light chain variable region contained in a humanized anti-CD3 antibody suitable for the present invention or an antigen-binding fragment of the antibody, there may be mentioned a light chain variable region comprising the amino acid residues shown in SEQ ID NO: 101 (Figure 115), SEQ ID NO: 102 (Figure 116), and SEQ ID NO: 103 (Figure 117). For a specific example of the heavy chain variable region contained in a humanized anti-CD3 antibody suitable for the present invention or an antigen-binding fragment of the antibody, there may be mentioned a heavy chain variable region having CDRH1 (GVTFNYYG) consisting of the amino acid sequence shown in SEQ ID NO: 26 (Figure 24), CDRH2 (ITNSGGRI) consisting of the amino acid sequence shown in SEQ ID NO: 27 (Figure 25), and CDRH3 (TLDGRDGWVAY) consisting of the amino acid sequence shown in SEQ ID NO: 28 (Figure 26). Further, for a specific example of the light chain variable region contained in a humanized anti-CD3 antibody suitable for the present invention or an antigen-binding fragment of the antibody, there may be mentioned a light chain variable region having CDRL1 (TGNIGSNY) consisting of the amino acid sequence shown in SEQ ID NO: 29 (Figure 27), CDRL2 (RDD) consisting of the amino acid sequence shown in SEQ ID NO: 30 (Figure 28), and CDRL3 (QSYSSGFI) consisting of the amino acid sequence shown in SEQ ID NO: 31 (Figure 29). In the present invention, the positions and lengths of the CDRs are determined by the definition of IMGT (Developmental and Comparative Immunology 27 (2003) 55-77). For a specific example of the heavy chain variable region of the present invention, there may be mentioned an amino acid sequence comprising the amino acid residues shown in SEQ ID NO: 16. As a specific example of the light chain variable region of the present invention, an amino acid sequence containing the amino acid residues shown in SEQ ID NOs: 17, 20, and 23 can be cited. Examples of suitable specific examples included in the humanized anti-CD3 antibody or the antigen-binding fragment of the antibody of the present invention include the following antibodies or antigen-binding fragments of the antibodies: antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 60 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 60; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 64 and a light chain variable region comprising amino acid residues 135 to 241 of SEQ ID NO: 64; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 66 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 66; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 68 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 68; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 70 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 70; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 72 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 72; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 74 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 74; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 76 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 76; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 78 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 78; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 80 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 80; antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 82 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 82; or antibodies or antigen-binding fragments of the antibodies containing a heavy chain variable region comprising amino acid residues 2 to 119 of SEQ ID NO: 84 and a light chain variable region comprising amino acid residues 135 to 243 of SEQ ID NO: 84. Furthermore, examples of the specific examples included in the humanized anti-CD3 antibody or the antigen-binding fragment of the antibody of the present invention may include an antibody or an antigen-binding fragment of the antibody containing a heavy chain variable region containing the amino acid sequence shown in SEQ ID NO: 16, a linker, and a light chain variable region containing the amino acid sequence shown in any one of SEQ ID NO: 17, 20, and 23. The variable regions contained in the antibody or the antigen-binding fragment of the antibody of the present invention may be combined in the order of the heavy chain variable region and the light chain variable region from the amino terminal side of the antibody, or may be combined in the order of the light chain variable region and the heavy chain variable region. In addition, a glycine residue may be present at the amino terminal of the variable region, and a linker, FLAG, and His tag may be bound to the carboxyl terminal of the variable region. Examples of suitable specific examples included in the humanized anti-CD3 antibody or the antigen-binding fragment of the antibody of the present invention may include: an antibody or an antigen-binding fragment of the antibody containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 19, an antibody or an antigen-binding fragment of the antibody containing the amino acid residues at positions 2 to 243 of SEQ ID NO: 22, and an antibody or an antigen-binding fragment of the antibody containing the amino acid residues at positions 2 to 241 of SEQ ID NO: 25. As for more suitable specific examples included in the humanized anti-CD3 antibody or the antigen-binding fragment of the antibody of the present invention, examples include: an antibody (clone ID: C3E-7078) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 60 (Figure 68) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7085) comprising the amino acid residues at positions 1 to 241 of SEQ ID NO: 64 (Figure 72) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7086) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 66 (Figure 74) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7087) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 68 (Figure 76) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7088) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 70 (Figure 78) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7089) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 72 (Figure 80) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7090) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 74 (Figure 82) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7091) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 76 (Figure 84) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7092) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 78 (Figure 86) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7093) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 80 (Figure 88) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7094) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 82 (Figure 90) or an antigen-binding fragment of the antibody; an antibody (clone ID: C3E-7095) comprising the amino acid residues at positions 1 to 243 of SEQ ID NO: 84 (Figure 92) or an antigen-binding fragment of the antibody. In a preferred embodiment of the humanized anti-CD3 antibody or the antigen-binding fragment of the antibody of the present invention, which is bound in the order of the heavy chain variable region, the first linker, and the light chain variable region from the amino acid terminal side, and further has a second linker, a FLAG tag, and a His tag bound to the carboxyl terminus of the light chain variable region, there may be mentioned the antigen described in the above (16) or the antigen-binding fragment of the antibody, which comprises: an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 22, an amino acid sequence comprising the amino acid residues 2 to 267 of SEQ ID NO: 25, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 60, an amino acid sequence comprising the amino acid residues 2 to 267 of SEQ ID NO: 64, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 66, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 68, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 70, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 72, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 74, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 76, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 78, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 80, an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 82, or an amino acid sequence comprising the amino acid residues 2 to 269 of SEQ ID NO: 84. For the antibody and the like of the present invention, the amino acid sequence of the heavy chain variable region and / or the amino acid sequence of the light chain variable region may have 70%, 71%, 72%, 73%, 74%, 75%, 76%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% or more identity with the amino acid sequence of the heavy chain variable region and / or the amino acid sequence of the light chain variable region contained in the anti-CD3 antibody or the antigen-binding fragment of the antibody of the present invention described above, and may be a molecule such as an antibody that binds to human CD3 and cynomolgus monkey CD3. For the antibody and the like of the present invention, a mutation may be introduced into the above humanized anti-CD3 antibody or the antigen-binding fragment of the antibody to optimize the binding ability to CD3. Specific methods for introducing mutations include random mutagenesis using the error-prone PCR method, site-specific amino acid mutation introduction using an NNK library, site-specific mutation introduction using structural information, and combinations thereof. For the antibodies and the like of the present invention, in order to reduce the effector activity of the antibody, it may be one that reduces ADCC and CDC activities by substituting the constant region. To avoid cytotoxicity to normal human CD3-expressing cells, it is desired that the effector activity of the antibody is low. It is known that the effector activity varies depending on the antibody subtype. It can be seen that IgG4 has low ADCC and CDC activities, and IgG2 has the characteristics of having CDC activity but low ADCC activity. Due to these characteristics, it is possible to produce antibodies with reduced ADCC and CDC activities by replacing the constant region of IgG1 with the constant regions of IgG2 and 4. Also, by referring to IgG2 and 4 and replacing a part of the sequence of the constant region of IgG1, IgG1 antibodies with reduced ADCC and CDC activities can be produced. As an example, according to Marjan Hezareh et. al. Journal of Virology, 75(24): 12161-12168 (2001), when the leucine residues at positions 234 and 235 of IgG1 (the numbers are based on the EU index of Kabat et al.) are each replaced with alanine residues, it shows reduced ADCC and CDC activities. (3-8) Antibodies that bind to the same site and also bind to cynomolgus monkey CD3 Antibodies that "bind to the same site and also bind to cynomolgus monkey CD3" such as the antibodies provided by the present invention are also included in the antibodies and the like of the present invention. An antibody that "binds to the same site as a certain antibody" means another antibody that binds to the site on the antigen molecule recognized by that antibody. If a second antibody binds to a partial peptide or partial three-dimensional structure on the antigen molecule bound by the first antibody, it can be determined that the first antibody and the second antibody bind to the same site. Also, by confirming that the second antibody competes with the first antibody for binding to the antigen, that is, the second antibody inhibits the binding of the first antibody to the antigen, even if the peptide sequence or three-dimensional structure of the specific binding site is not determined, it can be determined that the first antibody and the second antibody bind to the same site. Furthermore, when the first antibody and the second antibody bind to the same site and the first antibody has an effect of a characteristic of an aspect of the antibody of the present invention such as cytotoxic activity, the possibility that the second antibody also has the same activity is extremely high. Therefore, if a second anti-CD3 antibody binds to the binding site of the first anti-CD3 antibody and the second anti-CD3 antibody binds to cynomolgus monkey CD3, it can be determined that the first antibody and the second antibody bind to the same site on the CD3 protein. Antibodies or antigen-binding fragments of the antibody that bind to the same site on human CD3 that binds to the anti-CD3 antibody of the present invention or an antigen-binding fragment of the antibody and also bind to cynomolgus monkey CD3 are also included in the antibodies and the like of the present invention. Furthermore, if it can be confirmed that the second anti-CD3 antibody competes with the first anti-CD3 antibody for binding to the CD3 protein and the second anti-CD3 antibody binds to cynomolgus monkey CD3, it can be determined that the first antibody and the second antibody bind to the same site on the CD3 protein. Antibodies or antigen-binding fragments thereof that compete with the anti-CD3 antibody of the present invention or the antigen-binding fragment of the antibody for binding to human CD3 and bind to cynomolgus monkey CD3 are also included in the antibodies of the present invention and the like. The binding site of the antibody can be determined by methods well known to those of ordinary skill in the art such as immunological tests. For example, a series of peptides prepared by appropriately truncating the amino acid sequence of the antigen from the carboxyl terminus or the amino terminus are made, the reactivity of the antibody with these peptides is studied, and after roughly determining the recognition site, short peptides are further synthesized and the reactivity of the antibody with these peptides is studied, whereby the binding site can be determined. Antigen fragment peptides can be prepared using techniques such as gene recombination and peptide synthesis. The antibodies and the like of the present invention recognize and bind to the region that forms the three-dimensional structure in CD3, that is, the Ig-like domain. The binding site (epitope) of the antibody is determined by using X-ray crystallography and specifying the amino acid residues on the antigen adjacent to the antibody. (3-9) Modifications of anti-CD3 antibodies or their antigen-binding fragments The present invention provides a modified form of an antibody or its antigen-binding fragment. The modified form of the antibody or its antigen-binding fragment of the present invention means a chemical or biological modification of the antibody or its antigen-binding fragment of the present invention. Chemical modifications include bonding of chemical moieties to the amino acid backbone, chemical modifications of N-linked or O-linked carbohydrate chains, and the like. Biological modifications include post-translational modifications (e.g., addition of N-linked or O-linked sugar chains, processing of the amino-terminal region or carboxyl-terminal region, deamidation, isomerization of aspartic acid, oxidation of methionine), those expressed by using prokaryotic host cells and having a methionine residue added to the amino terminus, and the like. Furthermore, in order to make the detection or isolation of the antibody or antigen of the present invention possible, labeled substances such as enzyme labels, fluorescent labels, and affinity labels are also included in the meaning of such modifications. Such modified forms of the antibody or its antigen-binding fragment of the present invention are useful for improving the stability and blood retention of the antibody or its antigen-binding fragment of the present invention, reducing antigenicity, and detecting or isolating the antibody or antigen. Examples of the chemical moieties included in the chemical modification include water-soluble polymers such as polyethylene glycol, ethylene glycol / propanediol copolymer, carboxymethyl cellulose, dextran, and polyvinyl alcohol. In terms of biological modifiers, examples include those modified by enzyme treatment, cell treatment, etc., fusions with other peptides such as tags added by genetic recombination, and those prepared using cells expressing endogenous or exogenous glycosylation modifying enzymes as hosts, etc. This modification can be applied to any position on the antibody or its antigen-binding fragment, or at a desired position, and the same or two or more different modifications can be applied at one or two or more positions. In the present invention, the meaning of "modified form of the antigen-binding fragment of an antibody" also includes "fragment of the modified form of an antibody". Also, for example, it is known that lysine residues at the carboxyl terminus of the heavy chain are missing in antibodies produced by mammalian cultured cells (Journal of Chromatography A, 705: 129-134 (1995)), and it is also known that two amino acid residues, glycine and lysine, at the carboxyl terminus of the same heavy chain are missing, and the proline residue at the new carboxyl terminus is amidated (Analytical Biochemistry, 360: 75-83 (2007)). Furthermore, it is known that during the production of antibodies, the glutamine or glutamic acid residue at the amino terminus of the heavy chain or light chain of the antibody is modified by pyroglutamylation, and the antibodies of the present invention can have such modifications (International Patent Publication WO2013 / 147153). However, these deletions and modifications of the heavy chain sequence have little effect on the antigen-binding ability and effector antigen-binding properties (such as complement activation, antibody-dependent cell cytotoxicity, etc.) of the antibody. Therefore, the present invention also includes antibodies and antigen-binding fragments of the antibody that have been modified, deletion mutants with one or two amino acid deletions at the carboxyl terminus of the heavy chain, and amidated deletion mutants thereof (for example, the heavy chain with the proline residue at the carboxyl terminus amidated), antibodies with pyroglutamylation of the amino-terminal residue of the heavy chain or light chain, etc. (collectively referred to as "deletion mutants"). However, as long as all or part of the antigen-binding ability is retained, the deletion mutants at the carboxyl termini of the heavy chain and light chain of the antibodies of the present invention are not limited to the above-described type limitations. When the antibodies of the present invention include two or more chains (such as heavy chains), the two or more chains (such as heavy chains) can be any of the heavy chains selected from the group consisting of the full length and the above-described deletion mutants, or any combination of two or more thereof. The quantitative ratio or molecular number ratio of each deletion mutant can be affected by factors such as the type of mammalian cultured cells and culture conditions that produce the antibodies of the present invention, but for the deletion mutants of the antibodies of the present invention, examples include cases where one amino acid residue at the carboxyl terminus is missing in both of the two heavy chains. All of these are also included in the meaning of the antibody variants, antigen-binding fragments of the antibody, or modified forms thereof of the present invention. Furthermore, by adjusting the sugar chain modification (glycosylation, deglucosylation, etc.) that binds to the antibody of the present invention, the antibody-dependent cytotoxic activity can be enhanced. Regarding the technology for adjusting the sugar chain modification of antibodies, International Patent Publications WO99 / 54342, WO00 / 61739, WO02 / 31140, etc. are known, but are not limited thereto. The antibody of the present invention and the antigen-binding fragment of the antibody also include the antibody and the antigen-binding fragment thereof with the adjusted sugar chain modification. In the present invention, the scope of "antibody or its antigen-binding fragment" also includes "deletion mutants", "modified forms", and mixtures thereof of "antibody or its antigen-binding fragment". Furthermore, the scope of "antibody or its antigen-binding fragment" included in the antigen-binding molecules, multispecific molecules, bispecific molecules, etc. of the present invention described in (3-5) and (3-6) also includes "deletion mutants", "modified forms", and mixtures thereof of "antibody or its antigen-binding fragment". 4. Production of Antibodies (4-1) Method Using Hybridoma As an anti-CD3 antibody according to one aspect of the present invention, it can be prepared, for example, according to the method of Kohler and Milstein (Kohler and Milstein, Nature (1975) 256, p. 495-497, Kennet, R. ed., Monoclonal Antibodies, p. 365-367, Plenum Press, N.Y. (1980)). The antibody-producing cells of the anti-CD3 antibody are isolated from the spleen of an animal immunized with CD3 protein, and a hybridoma is established by fusing the cells with myeloma cells. A monoclonal antibody is obtained from the culture of the hybridoma. (4-1-1) Preparation of Antigen The antigen for producing the anti-CD3 antibody can be obtained according to the preparation method of natural or recombinant CD3 protein (human CD3εγ single-chain antigen), etc. Examples of antigens that can be prepared in this way include CD3 protein or CD3 protein fragments, or any amino acid sequence thereof, derivatives with an attached carrier, etc. (hereinafter, collectively referred to as "CD3"). Natural CD3, for example, can be purified and isolated from cells derived from human tissues or the culture of such cells. The recombinant human CD3εγ single-chain antigen is prepared by introducing a gene containing a base sequence encoding the amino acid sequence of the human CD3εγ single-chain antigen into a host cell and recovering the antigen from the culture of the cell. Furthermore, CD3 obtained by cell-free protein synthesis of a gene contained in a base sequence encoding the amino acid sequence of the CD3 antigen using an in vitro translation system is also included in the "CD3 antigen" of the present invention. (4-1-2) Production of anti-CD3 monoclonal antibody The production of monoclonal antibodies generally involves the following steps. (a) Step of preparing an antigen, (b) Step of preparing antibody-producing cells, (c) Step of preparing myeloma cells (hereinafter referred to as "myeloma"), (d) Step of fusing antibody-producing cells with myeloma, (e) Step of screening a group of hybridomas producing the target antibody, and (f) Step of obtaining a single cell line (cloning). If necessary, further steps such as (g) culturing the hybridoma, raising the animal transplanted with the hybridoma, etc., and (h) measuring and determining the biological activity of the monoclonal antibody are performed. Hereinafter, the method for producing monoclonal antibodies will be described in detail according to the above steps, but the method for producing the antibody is not limited thereto, and antibody-producing cells and myelomas other than spleen cells can also be used, for example. (a) Step of preparing an antigen The CD3 protein of the present invention is prepared by, for example, animal tissues (including body fluids), cells derived from the tissues, purification and isolation from cell cultures thereof, gene recombination, cell-free protein synthesis, chemical synthesis, etc. (b) Step of preparing antibody-producing cells The antigen obtained in step (a) is mixed with Freund's complete or incomplete adjuvant, or an adjuvant such as potash alum, and used as an immunogen to immunize experimental animals. Experimental animals can be used without problems in animals used in well-known hybridoma production methods. Specifically, for example, mice, rats, sheep, goats, cows, horses, etc. can be used. However, from the viewpoint of the ease of obtaining myeloma cells that fuse with the isolated antibody-producing cells, it is preferable to use mice or rats as the immunized animals. Moreover, the strains of mice and rats actually used are not particularly limited. In the case of mice, for example, A, AKR, BALB / c, BALB / cAnNCrj, BDP, BA, CE, C3H, 57BL, C57BL, C57L, DBA, FL, HTH, HT1, LP, NZB, NZW, RF, R III, SJL, SWR, WB, 129, etc. can be used. In the case of rats, for example, Wistar, Low, Lewis, Sprague-Dawley, ACI, BN, Fischer, etc. can be used. These mice and rats can be obtained, for example, from experimental animal breeding and selling companies such as Japan Claire and Japan Charles River. Among them, considering the fusion suitability with myeloma cells described later, it is particularly preferable to use the BALB / c strain as the immunized animal for mice, and the Wistar and Low strains as the immunized animals for rats. Furthermore, considering the homology between humans and mice with respect to the antigen, it is preferable to use autoimmune disease mice in which the in-vivo mechanism for removing autoantibodies is reduced. Also, the age of these mice or rats at the time of immunization is preferably 5 to 12 weeks old, more preferably 6 to 8 weeks old. When immunizing an animal with CD3 protein, methods such as those described in Weir, D.M., Handbook of Experimental Immunology Vol.I.II.III., Blackwell Scientific Publications, Oxford (1987); Kabat, E.A. and Mayer, M.M., Experimental Immunochemistry, Charles C Thomas Publisher Spigfield, Illinois (1964), etc. can be used. Regarding the method for measuring antibody titer, examples include immunoassays such as the RIA method and the ELISA method, but it is not limited to these methods. Spleen cells isolated from immunized animals or antibody-producing cells derived from lymphocytes can be prepared according to well-known methods such as those described in Kohler et al., Nature (1975) 256, 495,; Kohler et al., Eur. J. Immunol. (1977) 6, 511,; Milstein et al., Nature (1977), 266, 550,; Walsh, Nature (1977) 266, 495, etc. In the case of spleen cells, a general method can be employed in which the spleen is finely minced, the cells are filtered through a stainless steel mesh, and then suspended in Eagle minimum essential medium (MEM) or the like to isolate the antibody-producing cells. (c) Steps for preparing myeloma cells The myeloma cells used for cell fusion are not particularly limited and can be appropriately selected from known cell lines for use. Considering the convenience of selecting hybridomas from fused cells, it is preferable to use a (hypoxanthine-guanine phosphoribosyl transferase (HGPRT))-deficient strain for which the selection procedure has been established, such as X63-Ag8 (X63), NS1-ANS / 1 (NS1), P3X63-Ag8.Ul (P3Ul), X63-Ag8.653 (X63.653), SP2 / 0-Ag14 (SP2 / 0), MPC11-45.6TG1.7 (45.6TG), FO, S149 / 5XXO, BU.1, etc. derived from mice, 210.RSY3.Ag.1.2.3 (Y3), etc. derived from rats, U266AR (SKO-007), GM1500・GTG-A12 (GM1500), UC729-6, LICR-LOW-HMy2 (HMy2), 8226AR / NIP4-1 (NP41), etc. derived from humans. These HGPRT-deficient strains can be obtained from, for example, the American Type Culture Collection (ATCC), etc. These cell lines are subcultured in an appropriate medium, such as an 8-azaguanine medium [a medium obtained by adding 8-azaguanine to a medium containing glutamine, 2-mercaptoethanol, gentamicin, and fetal bovine serum (hereinafter referred to as "FCS") in RPMI-1640 medium], Iscove's Modified Dulbecco's Medium (hereinafter referred to as "IMDM"), or Dulbecco's Modified Eagle Medium (hereinafter referred to as "DMEM"), but are subcultured in a normal medium [for example, ASF104 medium (manufactured by Ajinomoto Co., Inc.) containing 10% FCS] 3 to 4 days before cell fusion, and ensure that there are 2×10 7 cells as described above on the day of fusion. (d) Step of fusing antibody-producing cells with myeloma cells The fusion of antibody-producing cells with myeloma cells is carried out according to known methods (Weir, D.M., Handbook of Experimental Immunology Vol. I. II. III., Blackwell Scientific Publications, Oxford (1987); Kabat, E.A. and Mayer, M.M., Experimental Immunochemistry, Charles C Thomas Publisher Spigfield, Illinois (1964), etc.), and can be carried out under conditions where the survival rate of the cells is not extremely reduced. For example, a chemical method of mixing antibody-producing cells and myeloma cells in a high-concentration polymer solution such as polyethylene glycol, a physical method using electrical stimulation, etc. can be used. (e) Step of screening a fusion tumor population that produces the antibody of interest The method for selecting the fusion tumors obtained by cell fusion is not particularly limited, but usually the HAT (hypoxanthine aminopterin thymidine) selection method (Kohler et al., Nature (1975) 256, 495; Milstein et al., Nature (1977) 266, 550) is used. This method is effective in the case of obtaining fusion tumors using myeloma cells that are HGPRT-deficient strains and cannot survive with aminopterin. That is, by culturing unfused cells and fusion tumors in HAT medium, only the fusion tumors that are resistant to aminopterin are selectively survived and can be proliferated. (f) Step of obtaining a single cell line (cloning) As for the cloning method of the fusion tumors, known methods such as the methyl cellulose method, the soft agarose method, the limiting dilution method, etc. (for example, Barbara, B.M. and Stanley, M.S.: Selected Methods in Cellular Immunology, W.H. Freeman and Company, San Francisco (1980)) can be used, and the limiting dilution method is preferably used. (g) Step of culturing the fusion tumors, step of raising the animals transplanted with the fusion tumors By culturing the selected fusion tumors, monoclonal antibodies can be produced. Preferably, after cloning the desired fusion tumors, they are supplied for the production of antibodies. The monoclonal antibody producing the hybridoma can be recovered from the culture of the hybridoma. Further, it can also be recovered as a recombinant antibody from the culture of cells into which the monoclonal antibody gene has been introduced. Furthermore, by injecting the hybridoma into the abdominal cavity of syngeneic mice (for example, the above-mentioned BALB / cAnNCrj), or Nu / Nu mice, and proliferating the hybridoma, it can also be recovered from its ascites. (h) Measurement steps and determination steps for the biological activity of monoclonal antibodies Various biological tests can be selected and applied according to the purpose. (4-2) Cellular immunization method By using cells expressing native CD3, cells expressing recombinant CD3 or its fragments, etc. as immunogens, anti-CD3 antibodies can be prepared using the aforementioned hybridoma method. Examples of cells expressing native CD3 include human thymocytes, T lymphocytes, etc. The CD3-expressing cells are used at 1×10 5 to 1×10 9 cells for the primary immunization, preferably 1×10 6 to 1×10 8 cells, more preferably 0.5 to 2×10 7 cells, even more preferably 1×10 7 cells. However, the number of cells supplied for immunization can be changed according to the expression level of CD3. This immunogen is generally administered into the abdominal cavity, and can also be administered intradermally, etc. For the method of producing hybridomas, the method described in (4-1-2) can be applied. (4-3) DNA immunization method The anti-CD3 antibody of the present invention can also use the DNA immunization method. By introducing an antigen-expressing plasmid gene into an animal individual such as a mouse or a rat, and inducing the immune response to the antigen by expressing the antigen in the individual. For the gene introduction method, there are methods such as directly injecting the plasmid into the muscle, intravenous injection of an introduction reagent such as liposome or polyethylenimine, using a viral vector, using a gene gun to shoot gold particles attached with the plasmid, and the hydrodynamic method of rapidly injecting a large amount of plasmid solution intravenously. Examples of the rat anti-human CD3 antibody established in this way include C3-147. The amino acid sequence of the light chain variable region of C3-147 is shown in SEQ ID NO: 9 (Figure 17) in the sequence listing. Also, the amino acid sequence of the heavy chain variable region of C3-147 is shown in SEQ ID NO: 7 (Figure 15) in the sequence listing. (4-4) Gene Recombination The antibody of the present invention can be prepared by introducing the nucleotides contained in the nucleotide sequence encoding its heavy chain amino acid sequence (heavy chain nucleotides) and the nucleotides contained in the nucleotide sequence encoding its light chain amino acid sequence (light chain nucleotides), or a vector into which the heavy chain nucleotides are inserted and a vector into which the light chain nucleotides are inserted into a host cell, culturing the cell, and then recovering the antibody from the culture. The heavy chain nucleotides and the light chain nucleotides can be inserted into one vector. As for the host cell, prokaryotic cells or eukaryotic cells can be used. When using eukaryotic cells as the host, animal cells, plant cells, or eukaryotic microorganisms can be used. Examples of animal cells include cells derived from mammals, such as kidney cells from human fetuses, HEK293F cells (Subedi GP et al., J Vis Exp. (2015) 106), COS cells derived from monkey kidneys (Gluzman, Y. Cell (1981) 23, 175-182, ATCC CRL-1650), mouse fibroblasts NIH3T3 (ATCC No. CRL-1658), Chinese hamster ovary cells (CHO cells, ATCC CCL-61), and its dihydrofolate reductase-deficient strain (CHOdhfr-: Urlaub, G. and Chasin, L.A. PNAS (1980) 77, 4126-4220), cells derived from birds such as chickens, and cells derived from insects. In addition, cells that have been modified to change the sugar chain structure to improve the biological activity of the antibody can also be used as the host. For example, among the N-glycosidic bond complex-type sugar chains that bind to the Fc region of the antibody, by using CHO cells in which the ratio of the sugar chain with fucose not bound to N-acetylglucosamine at the reducing end of the sugar chain is changed to 20% or more, an antibody with increased ADCC activity and CDC activity can be prepared (International Patent Publication WO02 / 31140). Examples of eukaryotic microorganisms include yeast. Examples of prokaryotic cells include Escherichia coli and Bacillus subtilis. Regarding the signal peptide used for secreting the antibody of the present invention (monoclonal antibodies, rat antibodies, mouse antibodies, chimeric antibodies, humanized antibodies, human antibodies, etc. derived from various animals), it is not limited to the secretion signals of antibodies of the same species, type, and subtype as the antibody, and the secretion signal of the antibody itself. Any secretion signal of other types or subtypes of antibodies, or the secretion signal of proteins derived from other eukaryotic or prokaryotic species can be selected and utilized. Antibodies and the like secreted containing a signal peptide are also included in the antibodies and the like of the present invention or the molecules of the present invention. (4-5) Methods for Designing and Preparing Humanized Antibodies For humanized antibodies, examples include antibodies in which only the CDRs of non-human animal antibodies are incorporated into antibodies derived from humans (see Nature (1986) 321, p. 522-525), antibodies in which, using the CDR grafting method, amino acid residues of a part of the framework of the sequence added to the CDRs are also grafted onto human antibodies (see WO90 / 07861, US Patent No. 6972323), antibodies in which one or more amino acids of non-human animal antibodies in either of them are substituted with human-type amino acids, etc., but are not limited to them. (4-6) Methods for Preparing Human Antibodies For the antibodies of the present invention, human antibodies can be further exemplified. A human anti-CD3 antibody means an anti-CD3 antibody composed of the amino acid sequence of an antibody derived from a human. A human anti-CD3 antibody can be obtained by a method using a human antibody-producing mouse containing a human genomic DNA fragment having the genes of the heavy chain and light chain of a human antibody (see Tomizuka, K. et al., Nature Genetics (1997) 16, 133-143,; Kuroiwa, Y. et.al., Nuc. Acids Res. (1998) 26, 3447-3448; Yoshida, H. et.al., Animal Cell Technology: Basic and Applied Aspects vol. 10, 69-73 (Kitagawa, Y., Matuda, T. and Iijima, S. eds.), Kluwer Academic Publishers, 1999.; Tomizuka, K. et.al., Proc. Natl. Acad. Sci. USA (2000) 97, 722-727, etc.). Specifically, such a human antibody-producing animal can be prepared by disrupting the loci of the endogenous immunoglobulin heavy chain and light chain of a non-human mammal and introducing the loci of the human immunoglobulin heavy chain and light chain by a yeast artificial chromosome (YAC) vector or the like. Also, by genetic recombination technology, eukaryotic cells are transformed with the cDNA encoding each of the heavy chain and light chain of such a human antibody, preferably a vector containing the cDNA, and this antibody can be obtained from the culture supernatant by culturing the transformed cells that produce the recombinant human monoclonal antibody. Among them, as the host, for example, eukaryotic cells can be used, and mammalian cells such as HEK293F cells and CHO cells are preferably used. Also, methods for obtaining phage-displayed human antibodies derived from a human antibody library are also known. For example, a phage display method can be used in which the variable region of a human antibody is expressed as an scFv on the phage surface, and phages that bind to an antigen are selected. By analyzing the genes of the phages selected by binding to the antigen, the DNA sequence encoding the variable region of the human antibody that binds to the antigen can be determined. If the DNA sequence of the scFv that binds to the antigen is known, an expression vector having that sequence can be prepared, introduced into an appropriate host, and expressed to obtain a human antibody (WO92 / 01047, WO92 / 20791, WO93 / 06213, WO93 / 11236, WO93 / 19172, WO95 / 01438, WO95 / 15388, Annu.Rev.Immunol (1994) 12, 433-455). (4-7) Method for preparing an antigen-binding fragment of an antibody Methods for preparing scFv are well known in the art (for example, see U.S. Patent No. 4,946,778, U.S. Patent No. 5,260,203, U.S. Patent No. 5,091,513, U.S. Patent No. 5,455,030, etc.). In this scFv, the heavy chain variable region and the light chain variable region are linked by a linker, preferably a polypeptide linker, in a manner that does not form a conjugate (Huston, J.S. et al., PNAS (1988), 85, 5879-5883). The heavy chain variable region and the light chain variable region in the scFv can be derived from the same antibody or from different antibodies. As for the polypeptide linker that links the variable regions, for example, any single-chain peptide consisting of 5 to 30 residues can be used. The DNA encoding the scFv is amplified by the PCR method using, as templates, the DNA encoding the heavy chain or the heavy chain variable region of the aforementioned antibody and the DNA encoding the light chain or the light chain variable region, and the DNA portions encoding all or the desired amino acid sequences in those sequences, and using specific primers at both ends thereof. Then, it is obtained by combining the DNA encoding the polypeptide linker portion and primers that link to the heavy chain and the light chain, respectively, at both ends thereof. Alternatively, it can also be obtained by fully synthesizing the DNA encoding the entire region of the scFv. Using the DNA encoding the scFv, an expression vector containing the DNA and a host cell transformed with the expression vector can be prepared according to a conventional method. Also, by culturing the host cell, the scFv can be recovered from the culture according to a conventional method. Other antigen-binding fragments of antibodies can also be obtained by introducing the gene encoding the antigen-binding fragment obtained by the aforementioned method into a cell and recovering the antigen-binding fragment from the culture of the cell. The antibodies and the like of the present invention can be multimerized to enhance the affinity for antigens. Regarding the multimerized antibodies, they can be one type of antibody or a plurality of antibodies that recognize a plurality of antigenic determinant sites of the same antigen. Regarding the method of multimerizing antibodies, examples include the binding of the IgG CH3 domain to two scFvs, the binding to streptavidin, the introduction of a helix-turn-helix motif, etc. The antibodies and the like of the present invention can be a mixture of a plurality of types of anti-CD3 antibodies with different amino acid sequences, that is, they can be polyclonal antibodies. Regarding polyclonal antibodies, for example, a mixture of a plurality of types of antibodies in which part or all of the CDR sets are different can be cited. Such polyclonal antibodies can be obtained by mixing and culturing the antibody-producing cells of different antibodies and recovering them from the culture (WO2004 / 061104). Also, individually prepared antibodies can be mixed. Furthermore, an antiserum, which is one aspect of polyclonal antibodies, can be prepared by immunizing an animal with a desired antigen and recovering the serum from the animal according to a conventional method. Regarding the modified products of antibodies, antibodies conjugated with various molecules such as polyethylene glycol (PEG) can also be used. The antibodies and the like of the present invention can be a complex (Immunoconjugate) in which these antibodies are linked to other molecules by a linker. Regarding an example of an antibody-drug conjugate in which the antibody is conjugated with a radioactive substance or a compound having a pharmacological action (drug), ADC (Antibody-Drug Conjugate) can be cited (Methods Mol Biol. (2013) 1045: 1-27). The antibodies and the like of the present invention can further link these antibodies to other functional polypeptides. Regarding an example of such an antibody-peptide complex, a complex of the antibody and an albumin-binding polypeptide can be cited (Protein Eng Des Sel. (2012)(2): 81-8). (4-8) Purification of antibodies and antigen-binding fragments of antibodies The obtained antibodies and antigen-binding fragments of antibodies can be uniformly purified in a manner that does not contain anything other than the antibodies and the like. The separation and purification of antibodies and antigen-binding fragments of antibodies can be carried out using the separation and purification methods commonly used for proteins. For example, if chromatography columns, filtration, ultrafiltration, salting out, dialysis, preparative polyacrylamide gel electrophoresis, isoelectric focusing electrophoresis, etc. are appropriately selected and combined, the antibodies can be separated and purified, but are not limited to these. Regarding a preferred separation and purification method, for example, an expression vector is prepared by attaching a DNA sequence encoding a His tag or a FLAG tag to the carboxyl terminus of the antibody variable region. The cells are transformed with this vector, and the cells are further cultured to express the antibody and the antigen-binding fragment of the antibody. After the culture is completed, the culture supernatant is extracted and purified by metal affinity chromatography using Ni, Co, etc., an anti-FLAG tag antibody column, gel filtration, ion exchange chromatography, etc. Antibodies and antigen-binding fragments of antibodies that are expressed and contain an amino acid sequence encoding a tag such as a His tag or a FLAG tag are also included in the antibodies or molecules of the present invention. (4-9) Multispecific molecules, bispecific molecules. Examples of the bispecific molecules and multispecific molecules of the present invention include methods of transient expression by introducing an expression plasmid into a host cell; methods of stable expression by selecting stable expression cell lines by drug selection after introducing a plasmid into a host cell; cell-free synthesis methods; methods of chemically binding respective antibodies or antigen-binding fragments produced by the above methods using a synthetic peptide linker. Regarding bispecific molecules using the antibody variable region, there are methods such as binding two single-chain antibodies (scFv) with a peptide linker (tandem scFv), forming a dimer (bispecific antibody) by exchanging domains of two antibodies with different specificities without covalent bonding, exchanging domains of two antibodies with different specificities and single-strandizing (single-chain bispecific antibody), forming a dimer by single-strandizing a bispecific antibody without covalent bonding (TandAb, US Patent US7129330), etc. The present invention also provides the antibody of the present invention or an antigen-binding fragment of the antibody, or a gene encoding a modification of an antigen, etc., a recombinant vector into which the gene is inserted, a cell into which the gene or the vector is introduced, and other cells that produce the antibody of the present invention. 5. Pharmaceutical composition. The present invention provides an anti-CD3 antibody, an antigen-binding fragment thereof or a modified form thereof, and / or a molecule of the present invention containing them, for example, a pharmaceutical composition containing a multispecific molecule. In the present invention, the treatment and / or prevention of a disease includes prevention of the onset of the disease, inhibition or obstruction of deterioration or progression, alleviation of one or more symptoms presented by an individual suffering from the disease, inhibition or remission of deterioration or progression, treatment or prevention of secondary diseases, etc., but is not limited thereto. Regarding the disease, for example, in the case of the aforementioned molecule, cancer can be mentioned. In the pharmaceutical composition of the present invention, it may contain an anti-CD3 antibody or an antigen-binding fragment of the antibody in a therapeutically or prophylactically effective amount and a pharmaceutically acceptable diluent, carrier, solubilizer, emulsifier, preservative, and / or adjuvant. "An amount effective for treatment or prevention" means an amount that exerts a therapeutic or preventive effect on a specific disease, administration form, and administration route. In the pharmaceutical composition of the present invention, substances (hereinafter referred to as "substances for pharmaceutical preparations") that can change, maintain, or keep the pH, osmotic pressure, viscosity, transparency, color, isotonicity, sterility, stability, solubility, sustained release property, absorbability, permeability, dosage form, strength, physical properties, shape, etc. of the composition or the antibody contained therein may be included. Regarding the substances for pharmaceutical preparations, as long as they are pharmacologically acceptable substances, there is no particular limitation. For example, non-toxic or low-toxicity is a property that substances for pharmaceutical preparations preferably possess. Regarding the substances for pharmaceutical preparations, for example, the following can be mentioned, but are not limited thereto: amino acids such as glycine, alanine, glutamine, aspartic acid, histidine, arginine, or lysine, antibacterial agents, antioxidants such as ascorbic acid, sodium sulfate, or sodium bisulfite, buffers such as phosphoric acid, citric acid, boric acid buffer, sodium bicarbonate, Tris-hydrochloric acid (Tris-Hcl) solution, fillers such as mannitol or glycine, chelating agents such as ethylenediaminetetraacetic acid (EDTA), complexing agents such as caffeine, polyvinylpyrrolidone, β-cyclodextrin, or hydroxypropyl-β-cyclodextrin, bulking agents such as glucose, mannose, or dextrin, monosaccharides, disaccharides, or other carbohydrates such as glucose, mannose, or dextrin, coloring agents, flavoring agents, diluents, emulsifiers, or hydrophilic polymers such as polyvinylpyrrolidone, low molecular weight polypeptides, salt-forming counter ions, preservatives such as benzalkonium chloride, benzoic acid, salicylic acid, thimerosal, phenethyl alcohol, methyl p-hydroxybenzoate, propyl p-hydroxybenzoate, chlorhexidine, sorbic acid, or hydrogen peroxide, solvents such as glycerol, propylene glycol, or polyethylene glycol, sugar alcohols such as mannitol or sorbitol, suspending agents, polyethyleneglycol (PEG), sorbitan esters, polysorbates such as polysorbate 20 or polysorbate 80, surfactants such as triton, tromethamine, lecithin, or cholesterol, stability enhancers such as sucrose or sorbitol, elasticity enhancers such as sodium chloride, potassium chloride, mannitol, or sorbitol, delivery agents, diluents, excipients, and / or pharmaceutical adjuvants. The addition amount of these substances for pharmaceutical preparations is 0.001 to 1000 times, preferably 0.01 to 100 times, more preferably 0.1 to 10 times, relative to the weight of the anti-CD3 antibody, its antigen-binding fragment, or its modified form, or the molecule of the present invention, for example, the multispecific molecule. The pharmaceutical compositions containing the anti-CD3 antibody of the present invention, its antigen-binding fragment or its modified form, or the molecule of the present invention, for example, immunoliposomes in which a multispecific molecule is contained in liposomes, antibody-modified forms formed by binding an antibody to a liposome (U.S. Patent No. 6,214,388, etc.) are also included in the pharmaceutical compositions of the present invention. Excipients or carriers are usually liquids or solids, and are not particularly limited as long as they are substances used in water for injection, physiological saline, artificial cerebrospinal fluid, or other preparations for oral or parenteral administration. Examples of physiological saline include neutral ones and those containing serum albumin. Examples of buffers include Tris buffer for adjusting the final pH of the pharmaceutical composition to 7.0 to 8.5, acetate buffer for adjusting the pH to 4.0 to 5.5, citrate buffer for adjusting the pH to 5.0 to 8.0, histidine buffer for adjusting the pH to 5.0 to 8.0, etc. The pharmaceutical composition of the present invention is in the form of a solid, liquid, suspension, etc. Lyophilized preparations can be cited. Excipients such as sucrose can be used to form the lyophilized preparations. Regarding the administration route of the pharmaceutical composition of the present invention, it can be any of enteral administration, local administration, and parenteral administration, and a suitable administration route can be selected according to the disease to be treated. Specifically, intravenous administration, intraarterial administration, intramuscular administration, intradermal administration, subcutaneous administration, intraperitoneal administration, transdermal administration, intraosseous administration, intra-articular administration, etc. can be cited. The composition of the pharmaceutical composition can be determined according to the administration method, the CD3 protein-binding affinity of the antibody, etc. For the dosage of the anti-CD3 antibody, its antigen-binding fragment or its modified form, or the molecule of the present invention, for example, a multispecific molecule of the present invention, it can be appropriately determined according to the type of individual, the type of disease, symptoms, gender, age, previous illnesses, the CD3 protein-binding affinity or its biological activity of the antibody, and other factors, but it is usually 0.01 to 1000 mg / kg, preferably 0.1 to 100 mg / kg, and is administered once every 1 to 180 days, or twice or more a day. Examples of the form of the pharmaceutical composition include injections (including lyophilized preparations, drip preparations), suppositories, nasal absorption preparations, transdermal absorption preparations, sublingual preparations, capsules, tablets, ointments, granules, aerosols, pills, powders, suspensions, emulsions, eye drops, implantable preparations, etc. The present invention provides that an anti-CD3 antibody, an antigen-binding fragment thereof, a modified form thereof, and / or a molecule of the present invention containing them, such as a multispecific molecule (hereinafter referred to as an anti-CD3 antibody, etc.) can be used in combination with other agents. The anti-CD3 antibody, etc., or a pharmaceutical composition containing the same as an active ingredient, can be administered simultaneously or separately with a pharmaceutical composition containing other agents, that is, an agent other than the anti-CD3 antibody, etc. as an active ingredient. For example, after administering other agents, a pharmaceutical composition containing an anti-CD3 antibody, etc. as an active ingredient can be administered, or after administering a pharmaceutical composition containing an anti-CD3 antibody, etc. as an active ingredient, other agents can be administered, or a pharmaceutical composition containing an anti-CD3 antibody, etc. as an active ingredient and other agents can be administered simultaneously. In either case, where both the anti-CD3 antibody, etc. and other agents are contained as active ingredients in a single pharmaceutical composition, or where the two active ingredients are contained in a plurality of pharmaceutical compositions respectively, it is referred to as a "pharmaceutical composition containing an anti-CD3 antibody, etc. and other agents" in the present invention. In the present invention, the "pharmaceutical composition" has the same meaning as the "pharmaceutical composition for administering an anti-CD3 antibody, etc. in combination with other agents". In the present invention, "administering in combination" of an anti-CD3 antibody, etc. and other agents means incorporating an anti-CD3 antibody, etc. and other agents into the body of the recipient within a certain period. The anti-CD3 antibody, etc. and other agents can be administered as a preparation contained in a single formulation, or can be formulated separately and administered individually. In the case of being formulated separately, the administration time is not particularly limited, and they can be administered simultaneously, or can be administered at different times or on different days at intervals. When the anti-CD3 antibody, etc. and other agents are administered at different times or on different days respectively, the administration order is not particularly limited. Usually, since each formulation is administered according to its respective administration method, the number of administrations of them may be the same or different. Also, in the case of being formulated separately, the administration method (administration route) of each formulation can be the same or can be administered by different administration methods (administration routes). Moreover, it is not necessary for the anti-CD3 antibody, etc. and other agents to coexist in the body at the same time, as long as they are taken into the body within a certain period (for example, one month, preferably one week, more preferably several days, and even more preferably one day), and it is also possible that the active ingredient of the other agent has disappeared from the body at the time of administering either one. Regarding the administration form of "pharmaceutical compositions administered in combination with anti-CD3 antibodies and other agents", for example, 1) administration of a single preparation containing anti-CD3 antibodies and other agents, 2) simultaneous administration of two preparations obtained by separately formulating anti-CD3 antibodies and other agents through the same administration route, 3) administration of two preparations obtained by separately formulating anti-CD3 antibodies and other agents with a time interval difference between the same administration routes, 4) simultaneous administration of two preparations obtained by separately formulating anti-CD3 antibodies and other agents through different administration routes, 5) administration of two preparations obtained by separately formulating anti-CD3 antibodies and other agents with a time interval difference between different administration routes, etc. The dosage, administration interval, administration form, preparation, etc. of "pharmaceutical compositions administered in combination with anti-CD3 antibodies and other agents" are similar to those of pharmaceutical compositions containing anti-CD3 antibodies, but are not limited thereto. In the case where the pharmaceutical composition becomes two different preparations, it can be a kit containing them. In the present invention, the "combination" of anti-CD3 antibodies and other agents means those "administered in combination" of anti-CD3 antibodies and other agents. The combination or pharmaceutical composition of the present invention can further use other medicines. The present invention also provides a method for treating or preventing a disease related to CD3, the use of the antibody of the present invention for preparing a therapeutic or prophylactic pharmaceutical composition for modulating the disease, and the use of the antibody of the present invention for treating or preventing the disease. A therapeutic or prophylactic kit containing the antibody of the present invention is also included in the present invention. [Examples] Hereinafter, the present invention will be described more specifically, but the present invention is not limited thereto. Also, in the following examples, unless otherwise specifically stated, each operation regarding genetic manipulation is carried out by the method described in "Molecular Cloning" (written by Sambrook, J., Fritsch, E.F. and Maniatis, T., published by Cold Spring Harbor Laboratory Press in 1989), or in the case of using commercially available reagents or kits, they are used according to the instructions of the commercial products. (Example 1) Preparation of Rat Anti-Human CD3 Antibody 1) - Construction of a human CD3εδ expression vector Using the Gateway Vector Conversion System (Thermo Fisher Scientific), a control vector pcDNA3.1-DEST that is changed to a destination vector was produced. The cDNA encoding the human CD3ε protein (NCBI Reference Sequence: NP_000724.1) shown in FIG. 1 (SEQ ID NO: 1) was purchased from Sino Biological, and was cloned into the pcDNA3.1-DEST vector using Gateway LR Clonase Enzyme mix (Thermo Fisher Scientific) to construct hCD3ε-pcDNA3.1. The cDNA encoding the human CD3δ protein (NP_000723.1) shown in FIG. 2 (SEQ ID NO: 2) was amplified by PCR using cDNA derived from human T cells as a template according to a method well-known to those of ordinary skill in the art, and the expression vector hCD3δ-pcDNA3.1 was constructed by cloning into pcDNA3.1(+) (Thermo Fisher Scientific). For the large-scale preparation of each expression vector, the Endofree Plasmid Giga Kit (QIAGEN) was used. 1) - 2 Immunization Female WKY / Izm rats (Japan SLC) were used for immunization. First, the calves of both legs of the rats were pretreated with hyaluronidase (SIGMA-ALDRICH), and the hCD3ε-pcDNA3.1 and hCD3δ-pcDNA3.1 expression vectors prepared in Example 1) - 1 were intramuscularly injected at the same site. Then, using ECM830 (BTX), in vivo electroporation was performed at the same site using two needle electrodes. After in vivo electroporation repeated approximately once every two weeks, the lymph nodes or spleens of the rats were collected and used for hybridoma production. 1)-3 Preparation of hybridomas Lymph node cells or spleen cells were electrofused with mouse myeloma SP2 / 0-ag14 cells (ATCC, No. CRL-1581) using an LF301 Cell Fusion Unit (BEX) and diluted and cultured in ClonaCell-HY Selection Medium D (StemCell Technologies). Monoclonal hybridomas were prepared by recovering the emerging hybridoma colonies. Each recovered hybridoma strain was cultured in ClonaCell-HY Selection Medium E (StemCell Technologies), and the obtained hybridoma culture supernatant was used to screen for hybridomas producing anti-human CD3 antibodies. 1)-4 Antibody screening using Cell-ELISA 1)-4-1 Preparation of antigen gene-expressing cells for Cell-ELISA HEK293α cells (a stable expression cell line derived from HEK293 expressing integrin αv and integrin β3) were prepared at 7.5×10 5 cells / mL in DMEM medium containing 10% FBS. To this, hCD3ε-pcDNA3.1 and hCD3δ-pcDNA3.1, or pcDNA3.1-DEST as a control, were introduced according to the transfection procedure using Lipofectamine 2000 (Thermo Fisher Scientific). 100 μL of each was dispensed into a 96-well plate (Corning) and cultured overnight at 37°C and 5% CO 2 in a medium containing 10% FBS. The obtained transfected cells were used directly in the attached state for Cell-ELISA. 1)-4-2 Cell-ELISA: After removing the culture supernatant of HEK293α cells into which the expression vector prepared in Example 1)-4-1 was introduced, the culture supernatant of the hybridoma was added to HEK293α cells into which hCD3ε-pcDNA3.1 and hCD3δ-pcDNA3.1, or pcDNA3.1-DEST were introduced, and left standing at 4°C for 1 hour. After washing the cells in the wells once with PBS containing 5% FBS, anti-rat IgG, HRP-Linked Whole Ab Goat (GE Healthcare Bioscience) diluted 500-fold with PBS containing 5% FBS was added and left standing at 4°C for 1 hour. After washing the cells in the wells twice with PBS containing 5% FBS, OPD chromogenic solution (o-phenylenediamine dihydrochloride (Wako Pure Chemical Industries, Ltd.) and H 2 O 2 dissolved to 0.4 mg / mL, 0.6% (v / v)) was added at 100 μL / well. The color reaction was carried out while occasionally stirring, and after stopping the color reaction by adding 1 M HCl at 100 μL / well, the absorbance at 490 nm was measured with a plate reader (ENVISION: PerkinElmer). In order to select hybridomas that produce antibodies that bind to human CD3 expressed on the cell membrane surface, compared with HEK293α cells transfected with the control pcDNA3.1-DEST, hybridomas that produced culture supernatants showing higher absorbance in HEK293α cells transfected with the hCD3ε-pcDNA3.1 and hCD3δ-pcDNA3.1 expression vectors were selected as positive for anti-human CD3 antibody production. 1)-5 Screening of antibodies using human T cell activation: The activation of T cells caused by the anti-CD3 antibody produced by the obtained hybridomas was evaluated using the detection of the CD69 activation marker as an index. Jurkat cells (ATCC, No. TIB-152), a human T cell line, were prepared at 5×10 6The concentration of cells / mL, 100 μL was inoculated into each well of a 96-well plate. In Jurkat cells from which the supernatant was removed by centrifugation, the culture supernatant of the hybridoma that was positive for anti-human CD3 antibody selected by Cell-ELISA of Examples 1)-4 or rat IgG isotype control antibody (R&D Systams) was added at a final concentration of 5 μg / mL, and left standing at 37°C for 30 minutes. Thereafter, cross-linked goat anti-rat IgG Fcγ Fragment specific (JACKSON IMMUNORESEARCH) was added at a final concentration of 10 μg / well, and cultured overnight under the conditions of 37°C and 5% CO 2 The next day, the supernatant was removed, the cells in the wells were washed once with PBS containing 5% FBS, and then 20 μL / well of PE mouse anti-human CD69 antibody (BD Bioscience) was added and left standing at 4°C for 30 minutes. After the cells in the wells were washed twice with PBS containing 5% FBS, they were resuspended in PBS containing 5% FBS, and detected using a flow cytometer (FC500: Beckman Coulter). Data analysis was performed using Flowjo (Treestar). A histogram of PE fluorescence intensity was created, and the hybridoma of the sample in which the histogram of the fluorescence intensity of PE produced was shifted toward the strong fluorescence intensity side relative to the fluorescence intensity histogram of the rat IgG isotype control antibody was selected as a hybridoma that was positive for human T cell activation ability and produced anti-human CD3 antibody. 1)-6 Screening for selective binding to human or monkey CD3 by flow cytometry 1)-6-1 Preparation of human antigen gene-expressing cells Lenti-X 293T cells (TAKARA, Cat#632180) were inoculated into a 225 cm 2 flask to a density of 5.3×10 4 cells / cm 2 and cultured overnight in DMEM medium containing 10% FBS under the conditions of 37°C and 5% CO 2 The next day, hCD3ε-pcDNA3.1 and hCD3δ-pcDNA3.1, or pcDNA3.1-DEST as a control were each introduced into Lenti-X 293T cells using Lipofectamine 2000, and cultured at 37°C and 5% CO 2Further culture overnight under the conditions. On the next day, the expression vector was introduced into Lenti-X 293T cells, which were treated with TrypLE Express (Thermo Fisher Scientific), washed with DMEM containing 10% FBS, and then adjusted to a concentration of 5×10 6 cells / mL with PBS containing 5% FBS. The obtained cell suspension was used for flow cytometry analysis. 1)-6-2 Flow cytometry analysis of the binding to human CD3 Further confirm the binding specificity to human CD3 of the antibodies produced by the hybridomas determined to be positive for human T cell activation ability in Example 1)-5 by flow cytometry. The Lenti-X 293T cell suspension prepared in Example 1)-6-1 was inoculated into a 96-well U-bottom microplate at 100 μL / well, and the supernatant was removed after centrifugation. The hybridoma culture supernatant was added to and suspended in Lenti-X 293T cells transfected with hCD3ε-pcDNA3.1 and hCD3δ-pcDNA3.1 and Lenti-X 293T cells transfected with pcDNA3.1-DEST, and left standing at 4°C for 1 hour. After washing once with PBS containing 5% FBS, it was added with and suspended in an anti-rat IgG FITC conjugate (SIGMA) diluted 500-fold with PBS containing 5% FBS, and left standing at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, it was resuspended in PBS containing 5% FBS with 2 μg / ml 7-aminoactinomycin D (Molecular Probes) and detected by flow cytometry. Data analysis was performed using Flowjo. After excluding the 7-aminoactinomycin D-positive dead cells by gating, a histogram of the FITC fluorescence intensity of live cells was made. Select the hybridomas of the samples in which the histograms of the fluorescence intensity of hCD3ε-pcDNA3.1 and hCD3δ-pcDNA3.1 transfected Lenti-X 293T cells are shifted to the strong fluorescence intensity side compared to the histogram of the fluorescence intensity of pcDNA3.1-DEST transfected Lenti-X 293T cells as the control, as the hybridomas producing antibodies that bind to human CD3. 1) - Construction of cynomolgus monkey CD3εδ expression vector The cDNAs encoding cynomolgus monkey CD3ε protein (NCBI reference sequence: NP_001270544.1) and cynomolgus monkey CD3δ protein (NCBI reference sequence: NP_001274617.1) were amplified by PCR using cDNA derived from cynomolgus monkey T cells as a template by those with ordinary knowledge in this technical field according to known methods, and the expression vectors cynoCD3ε-pcDNA3.1 and cynoCD3δ-pcDNA3.1 were constructed by cloning into pcDNA3.1(+) (Thermo Fisher Scientific). 1) - Preparation of cynomolgus monkey antigen gene-expressing cells Lenti-X 293T cells were inoculated into a 225 cm 4 flask at a density of 5.3×10 2 cells / cm 2 and cultured overnight at 37°C and 5% CO 2 in DMEM medium containing 10% FBS. The next day, cynoCD3ε-pcDNA3.1 and cynoCD3δ-pcDNA3.1, or pcDNA3.1-DEST as a control, were each introduced into Lenti-X 293T cells using Lipofectamine 2000 and further cultured overnight at 37°C and 5% CO 2 . The next day, the cells transfected with the expression vectors were treated with TrypLE Express, washed with DMEM containing 10% FBS, and then adjusted to a concentration of 5×10 6 cells / mL with PBS containing 5% FBS. The obtained cell suspension was used for flow cytometry analysis. 1)-6-5 Analytical method for binding to cynomolgus monkey CD3 by flow cytometry Using flow cytometry, the binding specificity of the antibody produced by the hybridoma, which was determined to bind to human CD3 in Example 1)-6-2 and was judged to be a hybridoma producing an antibody, to cynomolgus monkey CD3 was further confirmed. The Lenti-X293T cell suspension prepared in Example 1)-6-4 was inoculated into a 96-well U-bottom microplate at 100 μL / well, and after centrifugation, the supernatant was removed. The hybridoma culture supernatant was added to each of the Lenti-X293T cells transfected with cynoCD3ε-pcDNA3.1 and cynoCD3δ-pcDNA3.1 and the Lenti-X293T cells transfected with pcDNA3.1-DEST, and the cells were suspended and left standing at 4°C for 1 hour. After washing once with PBS containing 5% FBS, the cells were suspended by adding an anti-rat IgG FITC conjugate diluted 500-fold with PBS containing 5% FBS and left standing at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, the cells were resuspended in PBS containing 5% FBS with 2 μg / ml 7-aminoactinomycin D, and detection was performed using a flow cytometer. Data analysis was performed using Flowjo. After excluding dead cells that were positive for 7-aminoactinomycin D with a gate, a histogram of the FITC fluorescence intensity of live cells was created. Hybridomas that produced samples in which the histograms of the fluorescence intensity of the Lenti-X293T cells transfected with cynoCD3ε-pcDNA3.1 and cynoCD3δ-pcDNA3.1 were shifted toward the side of stronger fluorescence intensity compared to the histogram of the fluorescence intensity of the Lenti-X293T cells transfected with pcDNA3.1-DEST as a control group were selected as hybridomas producing antibodies that bind to cynomolgus monkey CD3. 1)-6-6 Preparation of human CD3δ gene-expressing cells The Lenti-X293T cells were inoculated into a 225 cm 2 flask to make it 5.3×10 4 cells / cm 2 , and cultured overnight at 37°C and 5% CO 2 in DMEM medium containing 10% FBS. The next day, hCD3δ-pcDNA3.1 or pcDNA3.1-DEST as a control was introduced into the Lenti-X293T cells using Lipofectamine 2000, and the cells were cultured at 37°C and 5% CO 2Further culture overnight under the following conditions. The next day, introduce the expression vector into Lenti-X 293T cells, treat with TrypLE Express, wash the cells with DMEM containing 10% FBS, and then prepare them into a concentration of 5×10 6 cells / mL with PBS containing 5% FBS. Use the obtained cell suspension for flow cytometry analysis. 1)-6-7 Flow cytometry analysis of the binding to human CD3δ Further confirm the binding specificity to human CD3δ of the antibodies produced by the fusion tumors that were determined to bind to monkey CD3 in Example 1)-6-5 by flow cytometry. Inoculate 100 μL / well of the Lenti-X 293T cell suspension prepared in Example 1)-6-6 into a 96-well U-bottom microplate, and remove the supernatant after centrifugation. Add the fusion tumor culture supernatant to each of the Lenti-X 293T cells transfected with hCD3δ-pcDNA3.1 and the Lenti-X 293T cells transfected with pcDNA3.1-DEST to suspend, and let stand at 4°C for 1 hour. After washing once with PBS containing 5% FBS, add anti-rat IgG FITC conjugate diluted 500-fold with PBS containing 5% FBS to suspend, and let stand at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, resuspend in PBS containing 5% FBS with 2 μg / ml 7-aminoactinomycin D, and detect with a flow cytometer. Data analysis is performed with Flowjo. After excluding the 7-aminoactinomycin D-positive dead cells with a gate, a histogram of the FITC fluorescence intensity of live cells is made. For the fluorescence intensity histogram of the pcDNA3.1-DEST-transfected Lenti-X 293T cells as a control, exclude the fusion tumors of the samples that produce a histogram shifted towards the stronger fluorescence intensity side compared to the histogram of the hCD3δ-pcDNA3.1-transfected Lenti-X 293T cells as the fusion tumors producing antibodies that bind to human CD3δ. 1)-6-8 Flow cytometry analysis of the binding to monkey T cell lines Further confirm the binding specificity to monkey T cell lines of the antibodies produced by the fusion tumors that were not excluded in Example 1)-6-7 by flow cytometry. For the cynomolgus monkey T cell line HSC-F (JCRB cell bank, No. JCRB1164), prepare it into 5×10 6The concentration of cells / mL was 100 μL seeded in each well of a 96-well plate. To the HSC-F cells from which the supernatant had been removed by centrifugation, the culture supernatant of the antibody-producing hybridoma not excluded in Examples 1)-5-7, or a rat IgG isotype control antibody was added, and the mixture was left standing at 4°C for 1 hour. Thereafter, the supernatant was removed, the cells in the wells were washed once with PBS containing 5% FBS, and then suspended by adding an anti-rat IgG FITC conjugate diluted 500-fold with PBS containing 5% FBS, and left standing at 4°C for 1 hour. After washing 3 times with PBS containing 5% FBS, the cells were resuspended in PBS containing 5% FBS and 2 μg / ml 7-aminoactinomycin D, and detected by flow cytometry. Data analysis was performed using Flowjo. After excluding the 7-aminoactinomycin D-positive dead cells by gating, a histogram of the FITC fluorescence intensity of viable cells was made, and a hybridoma of the sample in which the histogram of the FITC fluorescence intensity was shifted toward the strong fluorescence intensity side relative to the histogram of the fluorescence intensity of the rat IgG isotype control antibody was selected as an antibody-producing hybridoma that also binds to the monkey T cell line. 1)-7 Isotype determination of the antibody Among the rat anti-CD3 antibody-producing hybridomas obtained in Examples 1)-6, C3-147, which binds to human and monkey CD3ε, also binds to the monkey T cell line, and is suggested to have a high human T cell activation ability, was selected, and the antibody isotype was identified. The isotype was determined using a Rat Immunoglobulin Isotyping ELISA Kit (BD Pharmingen). As a result, it was confirmed that the isotype of the rat anti-CD3 monoclonal antibody C3-147 is IgG2b, λ chain. (Example 2) Study on the binding of the rat anti-CD3 monoclonal antibody (C3-147) to human CD3 2)-1 Preparation of the monoclonal antibody from the hybridoma supernatant 2)-1-1 Culture of the hybridoma producing C3-147 The rat anti-CD3 monoclonal antibody was purified from the hybridoma culture supernatant. First, the C3-147-producing hybridoma was proliferated to a sufficient amount using ClonaCell-HY Selection Medium E (StemCell Technologies), then 20% Ultra Low IgG FBS (Thermo Fisher Scientific) was added, and the culture medium was exchanged with Hybridoma SFM (Thermo Fisher Scientific) containing 5 μg / mL gentamicin (Thermo Fisher Scientific), and cultured for 7 days. The culture supernatant was recovered and sterilized by passing through a 0.22 μm filter (Corning). 2)-1-2 Purification: The antibody was purified from the culture supernatant of the hybridoma prepared in Example 2)-1-1 by Protein G affinity chromatography. The antibody was adsorbed onto a Protein G column (GE Healthcare Bioscience), and after washing the column with PBS, it was eluted with a 0.1 M glycine / hydrochloric acid aqueous solution (pH 2.7). 1 M Tris-HCl (pH 9.0) was added to the eluate, and after adjusting the pH to 7.0 - 7.5, while performing buffer replacement with PBS using a Centrifugal UF Filter Device VIVASPIN20 (cut-off molecular weight UF30K, Sartorius), the antibody was concentrated, and the antibody concentration was adjusted to 2 mg / mL. Finally, it was filtered through a Minisart-Plus filter (Sartorius) to obtain a purified sample. 2)-2 Binding of Rat Anti-CD3 Antibody (C3-147) to Human Single-Chain Antigen 2)-2-1 Preparation of Human CD3εγ Single-Chain Antigen The amino acid sequences encoding CD3ε or CD3γ are the sequences used in the crystal structure (PDB ID: 1SY6) of the OTK3-human CD3εγ single-chain antigen complex published in the protein database. The linker connecting the carboxyl terminus of CD3ε and the amino terminus of CD3γ is a peptide linker consisting of 26 amino acids identical to those reported in the reference (Kim, K.S. et al., (2000) J. Mol. Biol. 302, 899-916). The gene encoding the human CD3εγ single-chain antigen shown in Figure 3 (Sequence ID No. 4) was synthesized (GENEART) with restriction enzyme sites BamHI and HindIII added to the 5' end and 3' end, respectively. A fragment of approximately 4.8 kb obtained by digesting plasmid pQE80L (Qiagen) with restriction enzymes BamHI and HindIII, and a fragment of approximately 0.6 kb obtained by digesting the human CD3εγ single-chain antigen gene with BamHI and HindIII were ligated using Ligation high (Toyobo) to prepare the plasmid pQE80L-scCD3εγ for expression in E. coli. The amino acid sequence of the generated scCD3εγ is shown in Figure 4 (Sequence ID No. 5). The expression E. coli BL21(DE3) was transformed with the expression plasmid pQE80L-scCD3εγ, and the obtained clone was inoculated into 1 L of MagicMedia (Invitrogen) / Ultrayield flask (Thomson) and cultured with shaking at 30°C and 250 rpm for 21 hours. The cultured cells were harvested, and cell disruption and freeze-thaw cycles were repeated using an ultrasonic homogenizer in the presence of Tris buffer containing 1% Triton solution. Finally, the cells were centrifuged at 4°C and 15,000 rpm for 15 minutes to recover the inclusion body. The refolding from the inclusion body to purification was carried out according to the method of the reference (Kjer-Nielsen et al. (2004) PNAS vol. 101, no. 20, 7675-7680). However, the anti-CD3 antibody used for the antibody column was not 2C11 used in the literature, but the mouse anti-CD monoclonal antibody OKT3 (Sgro, Toxicology 105 (1995), 23-29, Orthoclone, Janssen-Cilag).2)-2-2 Binding of SPR to the human CD3εγ single-chain antigen The binding of antibody and antigen was measured using a Biacore 3000 (GE Healthcare Bioscience) by a capture method in which the antibody was used as a ligand to capture and immobilize the anti-mouse IgG antibody, and the antigen was used as an analyte for measurement. The antigen used was the human CD3εγ prepared in 2)-2-1. The anti-mouse IgG antibody (Mouse Antibody Capture Kit, GE Healthcare Bioscience) was covalently bound to the sensor chip CM5 (GE Healthcare Bioscience) at approximately 11000 RU by an amine coupling method. It was also fixed to the reference cell in the same manner. HBS-EP + (10 mM HEPES pH 7.4, 0.15 M NaCl, 3 mM EDTA, 0.05% surfactant P20) was used as the running buffer. On the chip where the anti-mouse IgG antibody was immobilized, the antibody was added as a ligand and captured for about 1 minute, and then 100 nM antigen was added at a flow rate of 30 μl / min for 120 seconds to monitor the binding of the antigen. As the regeneration solution, 10 mM glycine-HCl pH 1.7 was added at a flow rate of 10 μl / min for 3 minutes. As a result, the binding signal after 120 seconds for C3-147 was 34 RU. 2)-3 Confirmation of the antigen-binding site of the rat anti-CD3 antibody (C3-147) using SPR The method for confirming the antigen-binding site was carried out according to 2)-2-2. As a result, a binding signal of 34 RU was obtained for C3-147. On the other hand, as a comparative example, the antigen-binding site of SP34 (BD Pharmingen), a well-known anti-CD3 antibody, was similarly confirmed. No binding signal of 34 RU was observed for SP34. Therefore, the binding site on the CD3εγ surface recognized by C3-147 was shown to be different from that of SP34. (Example 3) Determination of the nucleotide sequence of the cDNA encoding the variable region of the rat anti-CD3 antibody (C3-147) The nucleotide sequence of the cDNA encoding the variable region of the rat anti-CD3 antibody (C3-147) was determined using the following method. 3) -1 cDNA synthesis The cell lysate of the fusion tumor producing rat anti-CD3 antibody (C3-147) (50 mM Tris-HCl (pH 7.5), 250 mM LiCl, 5 mM EDTA (pH 8), 0.5% lithium dodecyl sulfate (LiDS), 2.5 mM dithiothreitol (DTT)) was mixed with magnetic beads of the Dynabeads mRNA DIRECT Kit (Thermo Fisher Scientific) bound to oligo dT25 to bind mRNA to the magnetic beads. Next, the magnetic beads were washed with mRNA wash solution A (10 mM Tris-HCl (pH 7.5), 0.15 M LiCl, 1 mM EDTA, 0.1% LiDS, 0.1% Triton X-100) and cDNA synthesis solution (50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl2, 5 mM DTT, 0.5 mM dNTP, 0.2% Triton X-100, 1.2 units of RNase inhibitor (Thermo Fisher Scientific)) once, and then cDNA synthesis was performed with the cDNA synthesis solution supplemented with 12 units of SuperScriptIII reverse transcriptase (Thermo Fisher Scientific). Then, after washing with the 3'-tail addition reaction solution (50 mM potassium phosphate, 4 mM MgCl2, 0.5 mM dGTP, 0.2% Triton X-100, 1.2 unit RNase inhibitor), 48 units of terminal transferase was added, and the 3'-tail addition reaction was performed with the reaction solution supplemented with the recombinant (Roche). 3) Amplification and sequencing of gene fragments of rat immunoglobulin heavy and light chain variable regions After washing the magnetic beads with TE solution (10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 0.1% Triton X-100), the 5'-RACE PCR method was used to amplify the genes of rat immunoglobulin heavy and light chains. That is, the magnetic beads were transferred to the PCR reaction solution (0.2 μM primer, 0.2 mM dNTP, 0.25 unit PrimeSTAR HS DNA Polymerase (TAKARA)), and the reaction of 94°C for 30 seconds - 68°C for 90 seconds was carried out for 35 cycles. The primer sets used are as follows. PCR primer set for heavy chain gene amplification Sense primer Nhe-polyC-S 5'-GCTAGCGCTACCGGACTCAGATCCCCCCCCCCCCCDN-3' Figure 5 (SEQ ID NO: 50) First antisense primer rIgγ-AS1 5'-TCACTGAGCTGGTGAGAGTGTAGAGCCC-3' Figure 6 (SEQ ID NO: 51) Second antisense primer rIgγ-AS2 5'-TCACCGAGCTGCTGAGGGTGTAGAGCCC-3' Figure 7 (SEQ ID NO: 52) PCR primer set for light chain gene amplification Sense primer Nhe-polyC-S2 5'-GCTAGCGCTACCGGACTCAGATCCCCCCCCCCCCCDN-3' Figure 8 (SEQ ID NO: 53) First antisense primer rIgL-AS1 5'-TTCCACATCACTCGGGTAGAAATCAG-3' Figure 9 (SEQ ID NO: 54) Second antisense primer rIgγ-AS2 5'-TAACACCAGGGTAGAAATCTGTCACCAT-3' Figure 10 (SEQ ID NO: 55) For the fragments amplified by the above PCR reaction, sequence analysis of the base sequence was performed. The primers used are as follows.Sense primer for heavy chain sequencing rIgγ-seq 5’-CTGGCTCAGGGAAATAGCC-3’ Figure 11 (SEQ ID NO: 56), antisense primer for light chain sequencing rIgL-seq1 5’-TCCCTGGAGCTCCTCAGT-3’ Figure 12 (SEQ ID NO: 57), antisense primer for light chain sequencing rIgL-seq2 5’-GCCTTGTCAGTCTTGAGC-3’ Figure 13 (SEQ ID NO: 58). Sequence analysis was performed using a gene sequence analysis device (“ABI PRISM 3700 DNA Analyzer; Applied Biosystems” or “Applied Biosystems 3730xl Analyzer; Applied Biosystems”). The sequencing reaction was performed using the Dye Terminator Cycle Sequencing System with AmpliTaq DNA polymerase (Life Technologies) and GeneAmp 9700 (Applied Biosystems). The nucleotide sequence of the C3-147 heavy chain variable region determined by sequence analysis was recorded in Figure 14 (SEQ ID NO: 6), the amino acid sequence was recorded in Figure 15 (SEQ ID NO: 7), the nucleotide sequence of the C3-147 light chain variable region was recorded in Figure 16 (SEQ ID NO: 8), and the amino acid sequence was recorded in Figure 17 (SEQ ID NO: 9). (Example 4) Production of rat anti-CD3 scFv (C3E-7000) and its humanized form (C3E-7034) 4)-1 Production of rat anti-CD3 antibody (C3-147) scFv 4)-1-1 Construction of the rat antibody CD3 scFv expression vector (pC3E-7000) A DNA fragment with additional sequences of 15 bases before and after the DNA sequence encoding the linker inserted between the variable heavy chain (VH) and variable light chain (VL) of C3-147, the sense oligonucleotide of Figure 18 (SEQ ID NO: 10), and the synthetic antisense oligonucleotide of Figure 19 (SEQ ID NO: 11) (Sigma Aldrich, custom oligonucleotide synthesis service) were adjusted to 100 pmol / μL and then 20 μL of each was mixed. By standing at 96 °C for 10 minutes, 70 °C for 2 minutes, 60 °C for 2 minutes, 40 °C for 2 minutes, and 30 °C for 2 minutes, the two were annealed to produce a linker fragment inserted between VH and VL. Next, in a manner of adding the human IgG heavy chain message sequence to the vector backbone derived from the animal cell expression vector pcDNA-3.3TOPO (Thermo Scientific), using the In-Fusion HD cloning kit, the DNA fragment amplified by PCR, the DNA fragment amplified by PCR of the VH of the rat anti-CD3 antibody C3-147 shown in Figure 14 (SEQ ID NO: 6), the linker fragment inserted between VH and VL, and the DNA fragment amplified by PCR in which the DNA sequence encoding the FLAG-His tag was added to the carboxyl terminus in the region containing VL of C3-147 shown in Figure 16 (SEQ ID NO: 8) were combined to produce the scFv expression vector pC3E-7000 containing the nucleotide sequence of Figure 20 (SEQ ID NO: 14) in the ORF. 4)-1-2 Expression and purification of rat anti-CD3 scFv (C3E-7000) Expi293F cells (Thermo Scientific) were subcultured and cultured according to the manual. The above scFv expression vector was introduced into Expi293F cells in the logarithmic growth phase for transient expression. After filtration, it was used for purification. Purification was carried out in a two-step procedure using Ni affinity chromatography with His Trap excel (GE Healthcare) and gel filtration with Superdex 200 increase (GE Healthcare). The peak corresponding to the molecular weight of the scFv monomer was recovered as the purified protein sample. During purification, the AKTA chromatography system was used and all steps were carried out at 4 °C. The buffer for the purified protein was HBSor (25 mM histidine / 5% sorbitol, pH 5.0). The purified protein sample was supplied to SEC for analysis. After determining the purity and concentration, it was used for various analyses. The amino acid sequence of C3E-7000 is shown in Figure 21 (SEQ ID NO: 15). 4)-2 Humanization of rat anti-CD3 scFv (C3E-7000) 4) Humanization design of anti-CD3 antibody - Molecular simulation of the variable region of the rat antibody was performed using the commercially available protein three-dimensional structure analysis program Discovery Studio 3.5 (Dassault Systemes) according to the well-known method of homology modeling (Methods in Enzymology, 203, 121-153, (1991)). Humanization was carried out according to the generally well-known method of CDR grafting (Proc. Natl. Acad. Sci. USA 86, 10029-10033 (1989)). The recipient antibody was selected based on the amino acid identity within the framework region of the human subgroup consensus sequence or Germline sequence determined by KABAT et al. (Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service National Institutes of Health, Bethesda, MD. (1991)) or the predicted scores of immunogenicity and physical properties, etc. Also, with reference to the criteria proposed by Queen et al. (PNAS (1989) 86, 10029-10033), etc., the back mutations were selected using the three-dimensional structure model constructed by the above method. 4) Design of the humanized amino acid sequence of C3E-7000 Based on the method described in Example 4)-2-1, the amino acid sequence of C3E-7034, which is a humanized form of C3E-7000, was designed using the γ3 and λ6 of the human subgroup consensus sequence as receptors. In the amino acid sequence of C3-147VH shown in Fig. 15 (SEQ ID NO: 7), the arginine at amino acid position 16 was replaced with glycine, the alanine at amino acid position 17 was replaced with serine, the lysine at amino acid position 19 was replaced with arginine, the valine at amino acid position 23 was replaced with alanine, the valine at amino acid position 24 was replaced with alanine, the serine at amino acid position 88 was replaced with alanine, and the threonine at amino acid position 93 was replaced with valine. The amino acid sequence of C3E-7034VH with this design is described in Fig. 22 (SEQ ID NO: 16). In the amino acid sequence of C3-147VL shown in Fig. 17 (SEQ ID NO: 9), the glutamine at amino acid position 1 was replaced with asparagine, the valine at amino acid position 3 was replaced with methionine, the asparagine at amino acid position 8 was replaced with histidine, the threonine at amino acid position 12 was replaced with glutamine, the asparagine at amino acid position 13 was replaced with serine, the leucine at amino acid position 14 was replaced with proline, the threonine at amino acid position 16 was replaced with lysine, the glutamic acid at amino acid position 19 was replaced with threonine, the leucine at amino acid position 20 was replaced with isoleucine, the arginine at amino acid position 43 was replaced with serine, the leucine at amino acid position 75 was replaced with serine, the asparagine at amino acid position 79 was replaced with serine, the valine at amino acid position 81 was replaced with leucine, and the glutamine at amino acid position 82 was replaced with lysine. The amino acid sequence of C3E-7034VL with this design is described in Fig. 23 (SEQ ID NO: 17). The CDR sequences of C3E-7000 and C3E-7034 in the IMGT CDR definition are shown in Fig. 24 (SEQ ID NO: 26) for CDR-H1, Fig. 25 (SEQ ID NO: 27) for CDR-H2, Fig. 26 (SEQ ID NO: 28) for CDR-H3, Fig. 27 (SEQ ID NO: 29) for CDR-L1, Fig. 28 (SEQ ID NO: 30) for CDR-L2, and Fig. 29 (SEQ ID NO: 31) for CDR-L3, respectively. 4)-2-3 Modifications of humanized anti-CD3 scFv C3E-7034 To generate variants that maintain cross-reactivity with cynomolgus CD3ε while having differences in binding activity and cytotoxic activity, the amino acids in the framework region of the VL of C3E-7034 were designed in the same manner as the method described in 4)-2-1. A variant was designed in which the VL of the scFv (a sequence with 4 mutations of A2S, S8P, V13A, and F80L inserted into IGLV1-40*01) was replaced. 4)-2-3-1 Amino acid sequence design of C3E-7035 The amino acid sequence of C3E-7035, which is designed to be a variant of C3E-7034. In the light chain of C3E-7034 shown in Fig. 23 (SEQ ID NO: 17), the asparagine at amino acid position 1 in the variable region was replaced with glutamic acid, the phenylalanine at amino acid position 2 was replaced with alanine, the methionine at amino acid position 3 was replaced with valine, the histidine at amino acid position 8 was replaced with serine, the glutamic acid at amino acid position 12 was replaced with glycine, the serine at amino acid position 13 was replaced with valine, the lysine at amino acid position 16 was replaced with glutamine, the threonine at amino acid position 17 was replaced with arginine, the histidine at amino acid position 40 was replaced with leucine, the glutamic acid at amino acid position 41 was replaced with proline, the serine at amino acid position 43 was replaced with threonine, the serine at amino acid position 44 was replaced with alanine, the threonine at amino acid position 46 was replaced with lysine, the threonine at amino acid position 47 was replaced with leucine, the isoleucine at amino acid position 48 was replaced with leucine, the aspartic acid at amino acid position 57 was replaced with serine, the serine at amino acid position 60 was replaced with proline, the isoleucine at amino acid position 67 was replaced with lysine, the aspartic acid at amino acid position 68 was deleted, the arginine at amino acid position 69 was deleted, the serine at amino acid position 71 was replaced with glycine, the lysine at amino acid position 72 was replaced with threonine, the threonine at amino acid position 77 was replaced with alanine, the serine at amino acid position 79 was replaced with threonine, the asparagine at amino acid position 80 was replaced with glycine, the leucine at amino acid position 81 was replaced with phenylalanine, the lysine at amino acid position 82 was replaced with glutamine, the threonine at amino acid position 83 was replaced with alanine, and the phenylalanine at amino acid position 90 was replaced with tyrosine. The amino acid sequence of the light chain of C3E-7035 is described in Fig. 30 (SEQ ID NO: 20). The full-length sequence of C3E-7035 with methionine and alanine inserted immediately in front of the variable region of the light chain is described in Fig. 31 (SEQ ID NO: 22). 4) Amino acid sequence design of C3E-7036 The amino acid sequence of C3E-7036, which is designed to be a variant of C3E-7034. Along with the histidine at amino acid position 8 in the variable region of the light chain of C3E-7034 shown in Figure 23 (Sequence ID No. 17) being replaced by serine, the glutamic acid at amino acid position 12 being replaced by glycine, the serine at amino acid position 13 being replaced by valine, the lysine at amino acid position 16 being replaced by glutamine, the threonine at amino acid position 17 being replaced by arginine, the lysine at amino acid position 23 being replaced by threonine, the arginine at amino acid position 24 being replaced by glycine, the histidine at amino acid position 40 being replaced by leucine, the glutamic acid at amino acid position 41 being replaced by proline, the serine at amino acid position 43 being replaced by threonine, the serine at amino acid position 44 being replaced by alanine, the threonine at amino acid position 46 being replaced by lysine, the threonine at amino acid position 47 being replaced by leucine, the isoleucine at amino acid position 48 being replaced by leucine, the aspartic acid at amino acid position 57 being replaced by serine, the serine at amino acid position 60 being replaced by proline, the isoleucine at amino acid position 67 being replaced by lysine, deletion of the aspartic acid at amino acid position 68, deletion of the arginine at amino acid position 69, the serine at amino acid position 71 being replaced by glycine, the lysine at amino acid position 72 being replaced by threonine, the threonine at amino acid position 77 being replaced by alanine, the serine at amino acid position 79 being replaced by threonine, the asparagine at amino acid position 80 being replaced by glycine, the leucine at amino acid position 81 being replaced by phenylalanine, the lysine at amino acid position 82 being replaced by glutamine, the threonine at amino acid position 83 being replaced by alanine, the phenylalanine at amino acid position 90 being replaced by tyrosine. The amino acid sequence of the C3E-7036 light chain with this design is recorded in Figure 32 (Sequence ID No. 23), and the full-length amino acid sequence of C3E-7036 is recorded in Figure 33 (Sequence ID No. 25). 4) - Amino acid sequence design of 2-3-3 CDR variants. For the purpose of removing the deamination site present in CDRH2 of C3E-7034 (Figure 25, Sequence ID No. 27), C3E-7078 in which aspartic acid at amino acid number 53 of the heavy chain variable region of C3E-7034 shown in Figure 22 (Sequence ID No. 16) is substituted with arginine, and C3E-7079 in which serine is substituted, were designed. Also, C3E-7085 in which aspartic acid at amino acid number 53 of the heavy chain variable region of C3E-7036 is substituted with arginine was designed. The full-length amino acid sequence of C3E-7078 is described in Figure 68 (Sequence ID No. 60), the full-length amino acid sequence of C3E-7079 is described in Figure 70 (Sequence ID No. 62), and the full-length amino acid sequence of C3E-7085 is described in Figure 72 (Sequence ID No. 64). Furthermore, for the purpose of reducing the human CD3 affinity of C3E-7078, C3E-7086 in which aspartic acid at amino acid number 52 of the light chain variable region of C3E-7078 is substituted with glycine, C3E-7087 in which glutamine is substituted, C3E-7088 in which aspartic acid is substituted, C3E-7089 in which serine is substituted, and C3E-7090 in which alanine is substituted were designed. Similarly, for the purpose of reducing the human CD3 affinity of C3E-7079, C3E-7091 in which aspartic acid at amino acid number 52 of the light chain variable region of C3E-7079 is substituted with glycine, C3E-7092 in which glutamine is substituted, C3E-7093 in which aspartic acid is substituted, C3E-7094 in which serine is substituted, and C3E-7095 in which alanine is substituted were designed. The full-length amino acid sequence of C3E-7086 is described in Figure 74 (Sequence ID No. 66), the full-length amino acid sequence of C3E-7087 is described in Figure 76 (Sequence ID No. 68), the full-length amino acid sequence of C3E-7088 is described in Figure 78 (Sequence ID No. 70), the full-length amino acid sequence of C3E-7089 is described in Figure 80 (Sequence ID No. 72), the full-length amino acid sequence of C3E-7090 is described in Figure 82 (Sequence ID No. 74), the full-length amino acid sequence of C3E-7091 is described in Figure 84 (Sequence ID No. 76), the full-length amino acid sequence of C3E-7092 is described in Figure 86 (Sequence ID No. 78), the full-length amino acid sequence of C3E-7093 is described in Figure 88 (Sequence ID No. 80), the full-length amino acid sequence of C3E-7094 is described in Figure 90 (Sequence ID No. 82), and the full-length amino acid sequence of C3E-7095 is described in Figure 92 (Sequence ID No. 84). 4) - Preparation of humanized anti-CD3 scFv (C3E-7034, C3E-7035, C3E-7036) 4)-3-1 Construction of the humanized anti-CD3 scFv (C3E-7034) expression vector pC3E-7034 Synthesize a DNA fragment (GENEART) of the scFv DNA sequence that connects the region containing the C3E-7034 light chain shown in Fig. 23 (SEQ ID NO: 17) by a flexible linker of 15 amino acids separated at the carboxyl terminus of the C3E-7034 heavy chain shown in Fig. 22 (SEQ ID NO: 16), and a DNA fragment containing additional sequences of 15 bases before and after. Using this as a template, amplify C3E-7034 and the region containing the additional sequences before and after by PCR to obtain an inserted DNA fragment. Also, using the expression vector pC3E-7000 prepared in Example 4)-1-1 as a template, amplify the vector region excluding the scFv region by PCR to obtain a vector fragment. The respective DNA fragments are combined using the In-Fusion HD cloning kit (CLONTECH) to produce the humanized anti-CD3 scFv expression vector pC3E-7034 containing the nucleotide sequence of Fig. 34 (SEQ ID NO: 18) in the ORF. 4)-3-2 Construction of the humanized anti-CD3 scFv (C3E-7035) expression vector pC3E-7035 Synthesize a DNA fragment (GENEART) of the scFv DNA sequence that connects the C3E-7035 light chain shown in Fig. 30 (SEQ ID NO: 20) by a flexible linker composed of 17 amino acids separated at the carboxyl terminus of the C3E-7034 heavy chain shown in Fig. 22 (SEQ ID NO: 16), and a DNA fragment containing additional sequences of 15 bases before and after. In the same manner as in Example 4)-3-1, construct the C3E-7035 expression vector containing the nucleotide sequence of Fig. 35 (SEQ ID NO: 21) in the ORF. Name the obtained expression vector "pC3E-7035". 4)-3-2 Construction of the humanized anti-CD3 scFv (C3E-7036) expression vector pC3E-7036 Synthesize a DNA fragment (GENEART) of the scFv DNA sequence that connects the C3E-7036 light chain shown in Fig. 32 (SEQ ID NO: 23) by a flexible linker composed of 15 amino acids separated at the carboxyl terminus of the C3E-7034 heavy chain shown in Fig. 22 (SEQ ID NO: 16), and a DNA fragment containing additional sequences of 15 bases before and after. In the same manner as in Example 4)-3-1, construct the C3E-7036 expression vector containing the nucleotide sequence of Fig. 36 (SEQ ID NO: 24) in the ORF. Name the obtained expression vector "pC3E-7036". 4)-3-4 Construction of expression vectors for humanized anti-CD3 scFv with 4)-3-4CDR mutations Using the nucleotide sequence of C3E-7034 shown in Fig. 34 (SEQ ID NO: 18) in the ORF of pC3E-7034 as a template and primers with the base sequences shown in Figs. 93 and 94 (SEQ ID NO: 85, 86), site-specific mutagenesis using PCR was performed to produce an expression vector C3E-7078 in the ORF containing the nucleotide sequence of C3E-7078, in which the aspartic acid at amino acid position 53 in the heavy chain variable region of C3E-7034 was replaced with arginine. The obtained expression vector was named "pC3E-7078". Similarly, using pC3E-7034 as a template and Figs. 95 and 96 (SEQ ID NO: 87, 88) as primers, site-specific mutagenesis using PCR was performed to produce an expression vector C3E-7079 in the ORF containing the nucleotide sequence of C3E-7079, in which the aspartic acid at amino acid position 53 in the heavy chain variable region of C3E-7034 was replaced with serine. The obtained expression vector was named "pC3E-7079". Similarly, using pC3E-7036 as a template and Figs. 93 and 94 (SEQ ID NO: 85, 86) as primers, site-specific mutagenesis using PCR was performed to produce an expression vector C3E-7085 in the ORF containing the nucleotide sequence of C3E-7085, in which the aspartic acid at amino acid position 53 in the heavy chain variable region of C3E-7036 was replaced with arginine. The obtained expression vector was named "pC3E-7085". Expression vectors in the ORF containing the nucleotide sequences of C3E-7086, in which the aspartic acid at amino acid position 52 in the light chain variable region of C3E-7078 was replaced with glycine, C3E-7087 in which glutamine was replaced, C3E-7088 in which aspartic acid was replaced, C3E-7089 in which serine was replaced, C3E-7090 in which alanine was replaced, and expression vectors in the ORF containing the nucleotide sequences of C3E-7091, in which the aspartic acid at amino acid position 52 in the light chain variable region of C3E-7079 was replaced with glycine, C3E-7092 in which glutamine was replaced, C3E-7093 in which aspartic acid was replaced, C3E-7094 in which serine was replaced, C3E-7095 in which alanine was replaced were also produced in the same manner. A list of the produced vector names, templates, and primers is shown together in Table 1, and the primer list is shown together in Fig. 105. Table 1 4)-3-5 Expression and purification of humanized anti-CD3 scFv The expression and purification of C3E-7034, C3E-7035, and C3E-7036 were carried out in the same manner as in Example 4)-1-2. 4)-3-6 CDR modification of humanized anti-CD3 scFv and expression and purification of C3E-7078, C3E-7079, C3E-7085, C3E-7086, C3E-7087, C3E-7088, C3E-7089, C3E-7090, C3E-7091, C3E-7092, C3E-7093, C3E-7094, C3E-7095 of each CDR variant were performed in the same manner as in Example 4)-1-2. (Example 5) Crystallographic analysis of humanized anti-CD3 scFv (C3E-7034) 5)-1 Preparation of humanized anti-CD3 scFv (C3E-7034)-human CD3εγ single-chain antigen complex The CD3εγ prepared in Example 2)-2-1 and C3E-7034 prepared in Example 4)-3-1 were mixed at a molar ratio of 1:2, and the buffer was replaced with 10 mM Tris HCl (pH 7.5), 50 mM NaCl in AmiconUltra15 MWCO 10 kilo (Millipore) and concentrated to 3.5 mg / mL. This was purified by gel filtration chromatography using Superdex200 10 / 300GL (GE Healthcare), and the fraction of the complex was concentrated to approximately 4.0 mg / mL using AmiconUltra15 MWCO 10 kilo (Millipore). 5)-2 Crystallization The obtained CD3εγ and C3E-7034 complex was crystallized by the vapor diffusion method. An equal amount of the precipitant solution (0.1 M MES monohydrate (pH 6.5), 1.6 M ammonium sulfate, 10% v / v 1,4- butane) was added to 0.5 μL of the protein solution, and the solution was collected in a sealed container containing 0.05 mL of the precipitant solution so that the two solutions did not contact each other, and left standing at 25°C. After 1 month, rod-shaped crystals of 0.1 mm × 0.05 mm × 0.05 mm were obtained. 5) - 3 X-ray crystallographic analysis and identification of antigenic determinants The obtained crystals were immersed in perfluoropolyether PFO-X175 / 08 (Hampton Research), and then frozen with liquid nitrogen. X-ray diffraction data were collected using the beamline BL41XU (SPring-8, Hyogo). From the obtained diffraction images, the diffraction intensities were digitized using the software imosflm (CCP4: Collaborative Computational Project No.4) to obtain the crystal structure factors. The crystal was hexagonal, with a space group of P62. The unit cell of the crystal had a = 193.54 Å, b = 193.54 Å, and c = 43.88 Å. Using the obtained structure factors and the three-dimensional structure coordinates of the homology model, the molecular replacement method was performed to determine the phase. The software phaser (CCP4: Collaborative Computational Project No.4) was used for the calculation. The crystal contained 1 complex in the asymmetric unit. The software Refmac5 (CCP4: Collaborative Computational Project No.4) was used for the refinement of the structure, and the software coot was used for the model correction. This operation was repeated, and at a resolution of 3.3 Å, the final R value was 22.1% and the free R value was 27.0%. The final model included amino acid residues 1-108 of the light chain region of C3E-7034 (Figure 23, Sequence ID No. 17), amino acid residues 1-118 of the heavy chain region of C3E-7034 (Figure 22, Sequence ID No. 16), amino acid residues 33-67 and 71-118 of the CD3ε region (Figure 1, Sequence ID No. 1), and amino acid residues 23-103 of the CD3γ region (Figure 37, Sequence ID No. 3). The amino acid residues 68-70 of the CD3ε region (Figure 1, Sequence ID No. 1), and the amino-terminal region (amino acid residue 1), linker portion (amino acid residues 120-134), and carboxyl-terminal region (amino acid residues 243-269) of C3E-7034 (Figure 38, Sequence ID No. 19) were not modeled because their respective electron densities were unclear. The ribbon model and surface of the entire complex are shown in Figure 39. The interaction between CD3ε and the light and heavy chains of C3E-7034 is presented in Figure 40. Panel A shows a figure in which the amino acid residues of CD3ε within a distance of 4 Å from the light chain variable region of C3E-7034 are represented by thick stick models, and the other amino acid residues are represented by thin stick models.The amino acid residues of CD3ε within a distance of 4 Å from the variable light chain of C3E-7034, which are Ser55, Glu56, Arg101, Gly102, Ser103, Lys104, and Pro105 with residue names and residue numbers marked within the square in the figure, each amino acid number corresponding to Sequence Identification Number 1 in the sequence listing. Group B is a figure in which the amino acid residues of CD3ε within a distance of 4 Å from the variable heavy chain of C3E-7034 are represented by thick rod models, and the other amino acids are represented by thin rod models. In the figure, the residues Ser55, Glu56, Leu58, Trp59, Asn65, Ile66, Ser77, Asp78, and Arg101 with residue names and residue numbers marked within the square are amino acids of CD3ε, and each amino acid number corresponds to Sequence Identification Number 1 in the sequence listing. The distance within 4 Å from C3E-7034 is interpreted as follows: the amino acid residues of the antigenic determinant site of CD3ε for C3E-7034 are Ser55, Glu56, Leu58, Trp59, Asn65, Ile66, Ser77, Asp78, Arg101, Gly102, Ser103, Lys104, and Pro105. Among the antigenic determinant sites of CD3ε for C3E-7034, Arg101, Gly102, Ser103, Lys104, and Pro105 are also common antigenic determinant residues for CD3ε of OKT3 and UCHT1 (Kjer-Nielsen et al., PNAS 101 (2004), p. 7675-80; Arnett et al., PNAS 101 (2004), p. 16268-73). Also, C3E-7034 was demonstrated not to interact with the amino acid numbers 22 to 48 in Sequence Identification Number 1 corresponding to the antigenic determinant sites of anti-CD3 antibodies such as I2C and H2C described in WO2008 / 119565A2. In Figure 41, the interacting residues are presented on the sequence of CD3ε. The message sequence of CD3ε is shown in italics, and the amino acids within a distance of 4 Å from C3E-7034 are underlined. (Example 6) Preparation of Humanized OKT3 scFv 6)-1 Construction of the OKT3 scFv expression vector pC3E-3000 Using the same method as in Example 4)-1-1, the scFv of the mouse anti-CD3 monoclonal antibody OKT3 (Sgro, Toxicology 105 (1995), 23-29, Orthoclone, Janssen-Cilag) was prepared and introduced into an animal cell expression vector derived from pcDNA3.3. The obtained expression vector was named "pC3E-3000". 6) Design of the Amino Acid Sequence of Humanized 2 OKT3 scFv (C3E-3000) Based on the method described in Example 4)-2-1, the amino acid sequence of C3E-3007, which is a humanized form of OKT3, was designed using the human subgroup consensus sequences gamma1 and kappa4 as receptors. Also, considering the effects on immunogenicity scoring and physical properties, the amino acids of kappa1 were introduced at some positions. In the heavy chain of OKT3 shown in Fig. 42 (SEQ ID NO: 36), the glutamine at amino acid position 5 following the variable region was replaced with valine, the leucine at amino acid position 11 was replaced with serine, the alanine at amino acid position 12 was replaced with lysine, the arginine at amino acid position 13 was replaced with lysine, the methionine at amino acid position 20 was replaced with valine, the lysine at amino acid position 38 was replaced with arginine, the arginine at amino acid position 40 was replaced with alanine, the isoleucine at amino acid position 48 was replaced with methionine, the lysine at amino acid position 67 was replaced with arginine, the alanine at amino acid position 68 was replaced with valine, the leucine at amino acid position 70 was replaced with isoleucine, the threonine at amino acid position 72 was replaced with alanine, the serine at amino acid position 76 was replaced with threonine, the glutamine at amino acid position 82 was replaced with glutamic acid, the threonine at amino acid position 87 was replaced with arginine, the serine at amino acid position 91 was replaced with threonine, the threonine at amino acid position 114 was replaced with leucine, and the leucine at amino acid position 115 was replaced with valine. The amino acid sequence of the designed C3E-3007 heavy chain is shown in Fig. 43 (SEQ ID NO: 38).Among the variable regions of the light chain of OKT3 shown in FIG. 44 (SEQ ID NO: 37), the valine at amino acid position 3 is replaced with glutamine, the leucine at amino acid position 4 is replaced with methionine, the alanine at amino acid position 9 is replaced with serine, the isoleucine at amino acid position 10 is replaced with serine, the methionine at amino acid position 11 is replaced with leucine, the serine at amino acid position 12 is replaced with alanine, the alanine at amino acid position 13 is replaced with valine, the proline at amino acid position 15 is replaced with leucine, the lysine at amino acid position 18 is replaced with arginine, the valine at amino acid position 19 is replaced with alanine, the methionine at amino acid position 21 is replaced with isoleucine, the serine at amino acid position 39 is replaced with proline, the threonine at amino acid position 41 is replaced with lysine, the serine at amino acid position 42 is replaced with alanine, the alanine at amino acid position 59 is replaced with aspartic acid, the histidine at amino acid position 60 is replaced with arginine, the arginine at amino acid position 62 is replaced with serine, the serine at amino acid position 69 is replaced with aspartic acid, the tyrosine at amino acid position 70 is replaced with phenylalanine, the serine at amino acid position 71 is replaced with threonine, the glycine at amino acid position 76 is replaced with serine, the methionine at amino acid position 77 is replaced with leucine, the glutamic acid at amino acid position 78 is replaced with glutamine, the alanine at amino acid position 82 is replaced with valine, the serine at amino acid position 99 is replaced with glutamine, and the leucine at amino acid position 103 is replaced with valine. The amino acid sequence of the designed C3E-3007 light chain is shown in FIG. 45 (SEQ ID NO: 39). 6)-3 Construction of the expression vector pC3E-3007 for humanized OKT3 scFv (C3E-3007) A DNA fragment (GENEART) containing the scFv DNA sequence linked to the C3E-3007 light chain region shown in FIG. 45 (SEQ ID NO: 39) and an additional sequence of 15 bases before and after, with a 15-amino acid flexible linker intervening at the carboxyl terminus of the C3E-3007 heavy chain synthesized in FIG. 43 (SEQ ID NO: 38). In the same method as in Example 4)-3-1, a C3E-3007 expression vector containing the nucleotide sequence of FIG. 46 (SEQ ID NO: 34) in the ORF was constructed. The obtained expression vector was named "pC3E-3007". 6)-4 Expression and purification of humanized OKT3 scFv (C3E-3007) The expression and purification of C3E-3007 were carried out in the same method as in Example 4)-1-2. The amino acid sequence of C3E-3007 is shown in FIG. 47 (SEQ ID NO: 35). (Example 7) In vitro activity of humanized anti-CD3 scFv 7)-1 Study on the binding of humanized anti-CD3 scFv (C3E-3007, C3E-7034, C3E-7035, C3E-7036) to human CD3 7)-1-1 Study on the binding of humanized anti-CD3 scFv (C3E-3007, C3E-7034, C3E-7035, C3E-7036) to human CD3 by flow cytometry. Prepare at an appropriate concentration with PBS containing 5% FBS of commercially available human PBMC (CTL Corporation), add LIVE / DEAD Fixable Near-IR Dead Cell Stain Kit (Thermo Fisher Scientific) and anti-CD19 antibody (Beckman Coulter), and let stand at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, prepare at a concentration of 1×10 6 cells / mL with PBS containing 5% FBS, inoculate 100 μL / well into a 96-well U-bottom microplate, and remove the supernatant after centrifugation. Add 100 μL / well of humanized anti-CD3 scFv (C3E-3007, C3E-7034, C3E-7035, C3E-7036) diluted with PBS containing 5% FBS, and let stand at 4°C for 60 minutes. After washing twice with PBS containing 5% FBS, add 30 μL / well of Penta-His Alexa Fluor 488 (QIAGEN) diluted with PBS containing 5% FBS, and let stand at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, resuspend with PBS containing 5% FBS and detect with a flow cytometer (FACSCanto (trademark) II: BD). Data analysis is performed with Flowjo (Treestar), calculate the mean fluorescence intensity (MFI) of Alexa Fluor 488 in the fraction excluding dead cells and CD19-positive cells, subtract the MFI value of the sample without antibody addition from the MFI value of the scFv-added sample, and calculate the relative value (rMFI) of the MFI value. As shown in Figure 48, humanized anti-CD3 scFv shows binding to human CD3. 7)-1-2 Study on the binding of humanized anti-CD3 scFv (C3E-3007, C3E-7034, C3E-7035, C3E-7036) to human CD3 The affinity of humanized anti-CD3 scFv (C3E-3007, C3E-7034, C3E-7035, C3E-7036) for CD3 was determined by surface plasmon resonance using a BIAcore T-200 (GE Healthcare). Five different concentrations of scFv were flowed over CD3 immobilized on the sensor chip, and Rmax was estimated from the obtained response. The antibody concentration at 1 / 2 Rmax was taken as the dissociation constant of scFv for CD3. As a result, the dissociation constants of the scFv for CD3 were 400, 4.5, 22, and 25 nM, respectively. 7)-2 Study on the binding of humanized anti-CD3 scFv (C3E-3007, C3E-7034, C3E-7036) to cynomolgus monkey CD3 7)-2-1 Preparation of cynomolgus monkey PBMC PBMC was collected from cynomolgus monkey blood by the usual method using SepMate (STEMCELL) and Lymphocyte Separation Solution (Nacalai Tesque). 7)-2-2 Study on the binding of humanized anti-CD3 scFv (C3E-3007, C3E-7034, C3E-7035, C3E-7036) to cynomolgus monkey CD3 by flow cytometry The cynomolgus monkey PBMC obtained in Example 7)-2-1 was adjusted to an appropriate concentration with PBS containing 5% FBS, and staining and analysis were performed in the same manner as in Example 7)-1-1. As shown in Figure 49, humanized anti-CD3 scFv (C3E-7034, C3E-7035, C3E-7036) was shown to bind to cynomolgus monkey CD3. 7) T cell activation by humanized anti-CD3 scFv (C3E-3007, C3E-7034) Human peripheral blood mononuclear cells (PBMC) were isolated from the fresh buffy coat of random donors by density gradient separation using Lympholyte-H (Cedarlane). Humanized anti-CD3 scFv (C3E-3007, C3E-7034) diluted to 100 nM and anti-His antibody (Qiagen) at the same concentration were mixed in equal amounts with LR10 (FBS RPMI1640 (Thermo Fisher Scientific) containing 10% ultra low IgG). Human or monkey PBMC were prepared at 2×10 5 cells and mixed in equal amounts with humanized anti-CD3 scFv mixed with anti-His antibody in a 96-well U-bottom microplate and cultured at 37 °C for 24 hours under 5% CO 2 conditions. After the reaction, centrifugation was performed, and sorter buffer (HBSS (-) (Thermo Fisher Scientific), 0.1% BSA (Sigma-Aldrich), 0.1% sodium azide (Sigma-Aldrich)) was added. After centrifugation, LIVE / DEAD Fixable Near-IR Dead Cell Stain Kit (Thermo Fisher Scientific) was added to the cells and left standing at 4 °C for 20 minutes. After washing with screening buffer, PE-labeled anti-CD69 antibody (Becton, Dickinson) and FITC-labeled anti-CD8 antibody (Becton, Dickinson) diluted with screening buffer were added and left standing at 4 °C for 20 minutes. After washing with screening buffer, the cells were resuspended in PBS (Wako Pure Chemical Industries) containing 1% paraformaldehyde and detected using a flow cytometer (FACSCanto II: Becton, Dickinson). Data analysis was performed using Flowjo (Treestar), and the ratio of the CD8-highly expressing and PE-highly expressing fractions was calculated as a percentage of the parent population (% of parents) from the fraction excluding dead cells. As shown in Figure 50, the humanized anti-CD3 scFv was shown to activate human and monkey CD8-highly expressing cells. 7)-4 Comparison of the binding of CDR-modified humanized anti-CD3 scFv to human and cynomolgus monkey CD3 by flow cytometry. Human PBMC obtained in Example 7)-1-1 and cynomolgus monkey PBMC obtained in 7)-2-1 were adjusted to an appropriate concentration with PBS containing 5% FBS, stained in the same manner as in Example 7)-1-1, and analyzed. As shown in FIG. 106, the binding of CDR-modified humanized anti-CD3 scFv to human and monkey CD3 was confirmed. (Example 8) Preparation of humanized anti-TROP2 scFv 8)-1 Construction of the HT1-11 scFv expression vector pHT1-11scFv. A DNA fragment (GENEART) encoding the amino acid sequence of HT1-11 scFv shown in FIG. 51 (SEQ ID NO: 41) was synthesized. Using the In-Fusion HD PCR cloning kit (Clontech), a DNA fragment synthesized from a vector derived from pcDNA-3.3TOPO (Thermo Scientific) was inserted to construct the humanized anti-TROP2 scFv expression vector pHT1-11scFv containing the nucleotide sequence shown in FIG. 52 (SEQ ID NO: 40) in the ORF. 8)-2 Expression and purification of HT1-11 scFv. The expression and purification of HT1-11 scFv were carried out in the same manner as in Example 4)-1-2. (Example 9) Evaluation of the binding of humanized anti-TROP2 scFv (HT1-11scFv) to human TROP2 by flow cytometry. The pharyngeal squamous cell carcinoma cell line FaDu (ATCC) and the pancreatic cancer cell line HPAF-II (ATCC) were adjusted to an appropriate concentration with PBS containing 5% FBS, and the LIVE / DEAD Fixable Near-IR Dead Cell Stain Kit was added, and the cells were left standing at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, the cells were adjusted to 1×10 6The cells were inoculated at a concentration of cells / mL into a 96-well U-bottom microplate at 100 μL / well, and the supernatant was removed after centrifugation. Humanized anti-TROP2 scFv (HT1-11scFv) diluted with PBS containing 5% FBS was added at 100 μL / well and left to stand at 4°C for 60 minutes. After washing twice with PBS containing 5% FBS, 30 μL / well of Penta-His Alexa Fluor 488 diluted with PBS containing 5% FBS was added and left to stand at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, the cells were resuspended with PBS containing 5% FBS and detected using a flow cytometer (FACSCanto™ II). Data analysis was performed using Flowjo, and the mean fluorescence intensity (MFI) of Alexa Fluor 488 in the fraction of dead cells removed was calculated. The MFI value of the sample without antibody addition was subtracted from the MFI value of the scFv-added sample to calculate the relative value (rMFI) of the MFI value. As shown in Figure 53, the humanized anti-TROP2 scFv binds to human TROP2. (Example 10) Preparation of anti-TROP2-CD3 bispecific molecule 10)-1 Construction of anti-TROP2-CD3 bispecific molecule expression vector 10)-1-1 Construction of anti-TROP2-CD3 bispecific molecule (T2C-0001) expression vector pHT1-11scFv / C3E-7034 Using the pHT1-11scFv prepared in Example 8)-1 as a template, primers designed to link a part of the HT1-11scFv to the human antibody heavy chain message sequence on the 5' side and the linker between the scFvs on the 3' side were used for PCR to obtain an inserted DNA fragment. Also, using the expression vector pC3E-7034 prepared in 4)-3-1 as a template, primers for the message sequence and the amino-terminal sequence encoding the anti-CD3 scFv were used for PCR to obtain a vector DNA fragment containing the entire region of the vector containing the anti-CD3 scFv. The respective DNA fragments were combined using the In-Fusion HD cloning kit (CLONTECH) to prepare an anti-TROP2-CD3 bispecific molecule expression vector pT2C-0001 containing the nucleotide sequence of Figure 54 (SEQ ID NO: 42) in the ORF. 10)-1-2 Construction of anti-TROP2-CD3 bispecific molecule (T2C-0003) expression vector HT1-11scFv / C3E-3007 The anti-TROP2-CD3 bispecific molecule expression vector containing the nucleotide sequence of Figure 55 (SEQ ID NO: 44) in the ORF was constructed in the same manner as in 10)-1-1. However, pC3E-3007 was used as the template when preparing the vector fragment. The obtained expression vector was named "pT2C-0003". 10)-1-3 Construction of the expression vector for the bispecific molecule (T2C-0005) of HT1-11scFv / C3E-7035 In the same manner as in Example 10)-1-1, an expression vector for the anti-TROP2-CD3 bispecific molecule containing the nucleotide sequence of FIG. 56 (SEQ ID NO: 46) in the ORF was constructed. However, pC3E-7035 was used as the template when preparing the vector fragment. The obtained expression vector was named "pT2C-0005". 10)-1-4 Construction of the expression vector for the bispecific molecule (T2C-0006) of HT1-11scFv / C3E-7036 In the same manner as in Example 10)-1-1, an expression vector for the anti-TROP2-CD3 bispecific molecule containing the nucleotide sequence of FIG. 57 (SEQ ID NO: 48) in the ORF was constructed. However, pC3E-7036 was used as the template when preparing the vector fragment. The obtained expression vector was named "pT2C-0006". 10)-2 Expression and purification of the anti-TROP2-CD3 bispecific molecule The expression and purification of T2C-0001, T2C-0003, T2C-0005, and T2C-0006 were carried out in the same manner as in Example 4)-1-2. The amino acid sequence of T2C-0001 is shown in FIG. 58 (SEQ ID NO: 43). The amino acid sequence of T2C-0003 is shown in FIG. 59 (SEQ ID NO: 45). The amino acid sequence of T2C-0005 is shown in FIG. 60 (SEQ ID NO: 47). The amino acid sequence of T2C-0006 is shown in FIG. 61 (SEQ ID NO: 49). (Example 11) In vitro activity evaluation of the anti-TROP2-CD3 bispecific molecule 11)-1 Binding activity evaluation using SPR 11)-1-1 Binding of anti-TROP2-CD3 bispecific molecule to TROP2 The binding of the bispecific molecule to TROP2 was measured by a capture method using BIAcore 3000 (GE Healthcare Bioscience) with the antigen captured by anti-human IgG antibody and the bispecific molecule as the analyte. The antigen used was recombinant human TROP-2 / human IgG F fusion (R&D Systems). The anti-human IgG (Fc) antibody (Human Antibody Capture Kit, GE Healthcare Bioscience) was covalently bound to the sensor chip CM5 (GE Healthcare Bioscience) at approximately 2000 RU by amine coupling method. The same was fixed to the reference cell. HBS-P (10 mM HEPES pH 7.4, 0.15 M NaCl, 0.005% surfactant P20) was used as the running buffer. The bispecific molecule was prepared by 2-fold dilution from 200 nM to 1 nM. After adding 1 μg / ml of the antigen to the chip immobilized with the anti-human IgG (Fc) antibody for about 30 seconds, each concentration of the bispecific molecule was added at a flow rate of 30 μl / min for 300 seconds, and then the dissociation phase was monitored for 600 seconds. As the regeneration solution, 3 M magnesium chloride solution was added for 30 seconds. For data analysis, the binding rate constant ka, dissociation rate constant kd, and dissociation constant (KD; KD = kd / ka) were calculated using the 1:1 binding model of the analysis software (BIAevaluation software, version 4.1.1). 11)-1-2 Binding of the anti-TROP2-CD3 bispecific molecule to the human CD3εγ single-chain antigen The binding of the bispecific molecule to the CD3εγ antigen was measured by a method using BIAcore 3000 (GE Healthcare Bioscience) to measure the antibody as an analyte for the immobilized antigen. As the antigen, the human CD3εγ single-chain antigen prepared in 2)-2-1 was used, and about 100 RU was covalently bound to the sensor chip CM5 (GE Healthcare Bioscience) by amine coupling method. No antigen protein was added to the reference unit and only immobilization treatment was performed. HBS-P (10 mM HEPES pH 7.4, 0.15 M NaCl, 0.005% surfactant P20) was used as the running buffer. The bispecific molecule was prepared by serial two-fold dilution from the highest concentration of 1 μM to 4 nM, or from 200 nM to 1 nM. Each concentration of the bispecific molecule was added to the chip immobilized with the antigen at a flow rate of 10 μl / min for 25 minutes, and the binding amount was detected. As the regeneration solution, 10 mM glycine hydrochloride solution pH 1.5 was added for 30 seconds. In data analysis, the dissociation constant KD was calculated from the binding amount in each concentration using analysis software (BIAevaluation software, version 4.1.1). The results are shown in Tables 2 and 3. Table 2 Table 3 11)-2 Evaluation of binding activity by flow cytometry 11)-2-1 Binding of the anti-TROP2-CD3 bispecific molecule to TROP2 Using the same cancer cell line as in Example 9, staining and analysis were performed in the same manner. As shown in Fig. 62, the anti-TROP2-CD3 bispecific molecule binds to TROP2. 11)-2-2 Binding of the anti-TROP2-CD3 bispecific molecule to the human CD3εγ single-chain antigen Staining and analysis were performed in the same manner as in Example 7)-1-1. As shown in Fig. 63, the anti-TROP2-CD3 bispecific molecule binds to the human CD3εγ single-chain antigen. 11)-2-3 Binding of anti-TROP2-CD3 bispecific molecules to cynomolgus monkey CD3 antigen The cynomolgus monkey PBMCs obtained in Example 7)-2-1 were adjusted to an appropriate concentration with PBS containing 5% FBS, and staining and analysis were performed in the same manner as in Example 7)-1-1. As shown in Figure 64, the anti-TROP2-CD3 bispecific molecules (T2C-0001, T2C-0005, T2C-0006) bind to the cynomolgus monkey CD3 antigen. 11)-3 Evaluation of cytotoxic activity of anti-TROP2-CD3 bispecific molecules 11)-3-1 Analysis of TROP2 expression in target cells The pharyngeal squamous cell carcinoma cell line FaDu (ATCC), pancreatic cancer cell line HPAF-II (ATCC), and human lung cancer cell line Calu-6 were adjusted to an appropriate concentration with PBS containing 5% FBS, and the LIVE / DEAD Fixable Dead Cell Stain Kit (Thermo Fisher Scientific) was added and left standing at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, they were resuspended with PBS containing 5% FBS and adjusted to a concentration of 2×10 6 cells / mL, and 100 μL / well was inoculated into a 96-well U-bottom microplate. After centrifugation, the supernatant was removed. 25 μL / well of anti-TROP2 Alexa Fluor 488 antibody (eBioscience) and isotype control antibody (eBioscience) diluted with PBS containing 5% FBS were added and left standing at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, they were resuspended with PBS containing 5% FBS and detected with a flow cytometer (Cytomics FC500, Beckman Coulter). Data analysis was performed with Flowjo (Treestar), and the geometric mean fluorescence intensity (geometric MFI) of Alexa Fluor 488 in the dead cell removal fraction was calculated. As shown in Figures 65 (A, B, C), TROP2 expression was observed in FaDu and HPAF-II, but not in Calu-6. 11)-3-2 Preparation of target cells FaDu, HPAF-II, and Calu-6 were adjusted to 2×10 6Concentration in cells / mL, 100 μL of Chromium-51 Radionuclide (PerkinElmer) was added to each 1 mL of the cell suspension, and cultured at 37 °C and 5% CO 2 for 2 hours under the conditions of. After washing twice with RPMI1640 medium containing 10% FBS, it was resuspended in RPMI1640 medium containing 10% FBS to 2×10 5 cells / mL and used as target cells. 11)-3-3 Preparation of effector cells Commercially available frozen PBMC (Cellular Technology Limited) was thawed at 37 °C and transferred to a solution of RPMI1640 medium containing 10% FBS and added with Anti-aggregate Wash reagent (Cellular Technology Limited). After washing twice, it was prepared into 1×10 6 cells / mL as effector cells. 11)-3-4 Cytotoxicity test FaDu, HPAF-II, and Calu-6 obtained in Example 11)-3-2 were added to a 96-well U-bottom microplate at 50 μL / well. Various anti-TROP2-CD3 bispecific molecules prepared at each concentration were added thereto at 50 μL / well, and the effector cells prepared in 11)-3-1-3 were added at 100 μL / well. After centrifugation at room temperature at 1000 rpm for 1 minute, it was incubated at 37 °C and 5% CO 2Under the above conditions, culture for 20 - 24 hours. Recover 50 μL of the supernatant in a solid scintillation counting plate (LumaPlate) (PerkinElmer), dry at 50 °C for about 2 hours, and then measure with a plate reader (TopCount: PerkinElmer). The cell lysis rate is calculated by the following formula. Cell lysis rate (%) = (A - B) / (C - B) × 100 A: Count of the sample well B: Average value of the background (antibody non-added well) count (n = 3). When adding the antibody, add 50 μL of the test medium. Otherwise, perform the same operations as the sample well. C: Average value of the maximum release (well in which the target cells are lysed with a surfactant) count (n = 3). When adding the antibody, add 50 μL of the test medium. Add 100 μL of the surfactant, and similar to the sample well, aliquot 50 μL to the solid scintillation counting plate and perform the measurement. As shown in Fig. 66 (A, B, C), the cytotoxic activities of various TROP2-CD3 bispecific molecules against FaDu and HPAF-II are shown. On the other hand, no cytotoxic activity against Calu-6 was observed. (Example 12) Preparation of anti-Axl-CD3 bispecific molecule 12)-1 Construction of anti-Axl-CD3 bispecific molecule expression vector 12)-1-1 Construction of expression vector for bispecific molecule (AXC-0001) of 11D5-T3scFv / C3E-7034 The sequence of glycine was added to the N-terminus of h#11D5-T3H (shown in Figure 12 of the specification of European Patent Application Publication No. 2270053), and the sequence of h#11D5-T3L (shown in Figure 6 of the specification of European Patent Application Publication No. 2270053). The anti-Axl single-chain antibody linked through a polypeptide linker composed of three repeats of (GGGGS) was designed for the amino acid sequence of 11D5-T3scFv, and the nucleotide sequence encoding this was synthesized (GENEART, Thermo Fisher Science). Using this as a template, a DNA fragment with a part of the human antibody heavy chain message sequence added to the 5' side and a linker connecting scFvs added to the 3' side was obtained by PCR using the designed primers. Also, using the expression vector pC3E-7034 prepared in 4)-3-1 as a template, the entire region of the vector containing the anti-CD3 scFv was amplified by PCR using primers composed of the human antibody heavy chain message sequence and the nucleotide sequence encoding the anti-CD3 scFv, obtaining a vector DNA fragment. The respective DNA fragments were combined using the In-Fusion HD cloning kit (CLONTECH) to prepare an anti-Axl-CD3 bispecific molecule expression vector pAXC-0001 containing the nucleotide sequence shown in Figure 97 (sequence identification number 89) in the ORF. 12)-1-2 Construction of expression vector for bispecific molecule (AXC-0002) of 11D5-T3scFv / C3E-7036 The anti-Axl-CD3 bispecific molecule expression vector containing the nucleotide sequence shown in Figure 99 (sequence identification number 91) in the ORF was constructed in the same manner as in Example 10)-1-1. However, pC3E-7036 was used as a template when preparing the vector fragment. The obtained expression vector was named "pAXC-0002". 12)-2 Expression and purification of anti-Axl-CD3 bispecific molecule The expression and purification of AXC-0001 and AXC-0002 were carried out in the same manner as in Example 4)-1-2. The amino acid sequence of AXC-0001 is shown in Figure 98 (sequence identification number 90). The amino acid sequence of AXC-0002 is shown in Figure 100 (sequence identification number 92). (Example 13) In vitro activity evaluation of anti-Axl-CD3 bispecific molecule 13)-1 Analysis of Axl expression in target cells Human lung cancer cell line A549 (ATCC), human pancreatic cancer cell lines PANC-1 (ATCC), MIA PaCa-2 (ATCC), human myeloma cell line U266B1 (ATCC), and mantle cell lymphoma cell line Jeko-1 (ATCC) were adjusted to an appropriate concentration with PBS containing 5% FBS, and the LIVE / DEAD Fixable Dead Cell Stain Kit (Thermo Fisher Scientific) was added, followed by incubation at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, they were resuspended in PBS containing 5% FBS to a concentration of 2×10 6 cells / mL, and 100 μL / well was seeded into a 96-well U-bottom microplate. After centrifugation, the supernatant was removed. Anti-Axl antibody (RD-Systems) and Isotype Control antibody (RD-Systems) diluted with PBS containing 5% FBS were added at 25 μL / well, and the mixture was incubated at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, Alexa Fluor 488 anti-mouse IgG antibody (Thermo Fisher Scientific) diluted with PBS containing 5% FBS was added at 25 μL / well, and the mixture was incubated at 4°C for 30 minutes. After washing twice with PBS containing 5% FBS, they were resuspended in PBS containing 5% FBS and analyzed using a flow cytometer (Cytomics FC500, Beckman Coulter). Data analysis was performed using Flowjo (Treestar), and the geometric mean fluorescence intensity (geometric MFI) of Alexa Fluor 488 in the dead cell removal fraction was calculated. As shown in Figures 107 (A, B, C, D, E), Axl expression was observed in A549, PANC-1, and MIA PaCa-2, but not in U266B1 and Jeko-1. 13)-2 Preparation of target cells A549, PANC-1, MIA PaCa-2, U266B1, and Jeko-1 were adjusted to a concentration of 2×10 6 cells / mL with RPMI 1640 medium (Thermo Fisher Scientific) containing 10% FBS. 100 μL of Chromium-51 Radionuclide (PerkinElmer) was added to each 1 mL of the cell suspension, and the mixture was incubated at 37°C and 5% CO 2Cultured for 2 hours under the conditions. After washing twice with RPMI1640 medium containing 10% FBS, resuspend with RPMI1640 medium containing 10% FBS to 2×10 5 cells / mL as target cells. 13)-3 Preparation of effector cells Thaw commercially available frozen PBMC (Cellular Technology Limited) at 37°C, transfer to a solution of RPMI1640 medium containing 10% FBS supplemented with Anti-aggregate Wash reagent (Cellular Technology Limited), wash twice, and then prepare with RPMI1640 medium containing 10% FBS to 1×10 6 cells / mL as effector cells. 13)-4 Cytotoxicity assay Add the A549, PANC-1, MIA PaCa-2, U266B1, and Jeko-1 obtained in Example 13)-2 to a 96-well U-bottom microplate at 50 μL / well. Add various anti-Axl-CD3 bispecific molecules prepared at each concentration to it at 50 μL / well, add the effector cells prepared in 13)-3 to it at 100 μL / well, centrifuge at 1000 rpm for 1 minute at room temperature, and then culture at 37°C and 5% CO 2 for 20 - 24 hours under the conditions. Recover 50 μL of the supernatant into a solid scintillation counting plate (PerkinElmer), dry at 50°C for about 2 hours, and then measure with a plate reader (TopCount: PerkinElmer). The cell lysis rate is calculated by the following formula. Cell lysis rate (%) = (A - B) / (C - B)×100 A: Count of the sample well. B: Average value of the background (antibody non-added well) count (n = 3). When adding the antibody, add 50 μL of the test medium. Otherwise, perform the same operations as the sample well. C: Average value of the maximum release (the well where the target cells are lysed with a surfactant) count (n = 3). When adding the antibody, add 50 μL of the test medium. The surfactant is added at 100 μL, and 50 μL is aliquoted into the solid scintillation counting plate and measured in the same manner as the sample well. As shown in FIGS. 108 (A, B, C, D, E), various anti-Axl-CD3 bispecific molecules exhibited cytotoxic activities against A549, PANC-1, and MIA PaCa-2. On the other hand, no cytotoxic activities against U266B1 and Jeko-1 were observed. (Example 14) Preparation of an expression vector for an anti-HLA-A2 / MAGEC1-CD3 bispecific molecule 14)-1 Construction of an expression vector for an anti-HLA-A2 / MAGEC1-CD3 bispecific molecule 14)-1-1 Construction of an expression vector for the MAG-032scFv / C3E-7034 bispecific molecule (MGC-0001) The MAG-032 scFv that specifically binds to HLA-A2 / MAGEC1 was obtained from a human antibody phage library. Using the nucleotide sequence encoding the amino acid sequence of MAG-032 scFv as a template, by PCR with designed primers, a part of the human antibody heavy chain message sequence was added to the 5' side, a linker connecting scFvs, and a nucleotide sequence encoding a part of C3E-7034 were added to the 3' side to obtain an inserted DNA fragment containing the nucleotide sequence encoding the amino acid sequence of MAG-032 scFv. Also, using the expression vector pC3E-7034 prepared in 4)-3-1 as a template, and using primers composed of a human antibody heavy chain message sequence and a nucleotide sequence encoding the amino-terminal sequence of anti-CD3 scFv, the entire region of the vector containing anti-CD3 scFv was amplified by PCR to obtain a vector DNA fragment. The respective DNA fragments were combined using the In-Fusion HD cloning kit (CLONTECH) to prepare an anti-HLA-A2 / MAGEC1-CD3 bispecific molecule expression vector pMGC-0001 containing the nucleotide sequence shown in Figure 101 (SEQ ID NO: 93) in the ORF. 14)-1-2 Construction of an expression vector for the MAG-032scFv / C3E-7036 bispecific molecule (MGC-0002) In the same manner as in Example 14)-1-1, an anti-HLA-A2 / MAGEC1-CD3 bispecific molecule expression vector containing the nucleotide sequence shown in Figure 103 (SEQ ID NO: 95) in the ORF was constructed. However, pC3E-7036 was used as the template for preparing the vector fragment. The obtained expression vector was named "pMGC-0002". 14)-2 Expression and purification of the anti-HLA-A2 / MAGEC1-CD3 bispecific molecule The expression and purification of MGC-0001 and MGC-0002 were carried out in the same manner as in Example 4)-1-2. The amino acid sequence of MGC-0001 is shown in Figure 102 (SEQ ID NO: 94). The amino acid sequence of MGC-0002 is shown in Figure 104 (SEQ ID NO: 96). (Example 15) In vitro activity evaluation of the anti-HLA-A2 / MAGEC1-CD3 bispecific molecule 15)-1 Analysis of the expression of HLA-A2 / MAGEC1 in target cells Human lymphoblastoid cell line T2 (ATCC) cells were adjusted to an appropriate concentration with AIM-V medium (Thermo Fisher Scientific) containing 20% FBS, and MAGEC1 peptide (Sigma Genosys) of SEQ ID NO: 97 or DMSO was added, followed by incubation at 37 °C for 4 hours. After washing twice with AIM-V medium containing 20% FBS, the cells were adjusted to an appropriate concentration with PBS containing 5% FBS, and LIVE / DEAD Fixable Dead Cell Stain Kit (Thermo Fisher Scientific) was added, followed by standing at 4 °C for 30 minutes. After washing twice with PBS containing 5% FBS, the cells were adjusted to a concentration of 2×10 6 cells / mL with PBS containing 5% FBS, and 100 μL / well was seeded into a 96-well U-bottom microplate. After centrifugation, the supernatant was removed. Anti-HLA-A2 / MAGEC1 antibody (MAG032 scFv) diluted with PBS containing 5% FBS was added at 25 μL / well, followed by standing at 4 °C for 30 minutes. After washing twice with PBS containing 5% FBS, the cells were diluted with PBS containing 5% FBS. Penta-His Alexa Fluor 488 (QIAGEN) was added at 25 μL / well, followed by standing at 4 °C for 30 minutes. After washing twice with PBS containing 5% FBS, the cells were resuspended with PBS containing 5% FBS and detected with a flow cytometer (Cytomics FC500, Beckman Coulter). Data analysis was performed with Flowjo (Treestar), and the geometric mean fluorescence intensity (geometric MFI) of Alexa Fluor 488 in the dead cell removal fraction was calculated. As shown in FIGS. 109 (A, B), the expression of HLA-A2 / MAGEC1 was observed in T2 cells to which MAGEC1 peptide was added, and no expression was observed in T2 cells to which DMSO was added. 15)-2 Preparation of target cells T2 cells were adjusted to an appropriate concentration with AIM-V medium (Thermo Fisher Scientific) containing 20% FBS, and MAGEC1 peptide or DMSO was added, followed by incubation at 37 °C for 4 hours. The cells were adjusted to 2×10 6Concentration in cells / mL, 100 μL of Chromium-51 Radionuclide (PerkinElmer) was added to each 1 mL of the cell suspension, and cultured at 37 °C and 5% CO 2 for 2 hours under the condition of. After washing twice with RPMI1640 medium containing 10% FBS, it was resuspended with RPMI1640 medium containing 10% FBS to 2×10 5 cells / mL as target cells. 15)-3 Preparation of effector cells Thawed commercially available frozen PBMC (Cellular Technology Limited) at 37 °C, transferred it to a solution of RPMI1640 medium containing 10% FBS supplemented with Anti-aggregate Wash reagent (Cellular Technology Limited), washed twice, and then prepared with RPMI1640 medium containing 10% FBS to 1×10 6 cells / mL as effector cells. 15)-4 Cytotoxicity assay The T2 cells obtained in Example 15)-2 were added to a 96-well U-bottom microplate at 50 μL / well. Various anti-HLA-A2 / MAGEC1-CD3 bispecific molecules prepared at each concentration were added thereto at 50 μL / well, and the effector cells prepared in 15)-3 were added at 100 μL / well. After centrifugation at room temperature at 1000 rpm for 1 minute, it was incubated at 37 °C and 5% CO 2Cultivate for 20 - 24 hours under the following conditions. Recover 50 μL of the supernatant in a solid scintillation counting plate (PerkinElmer), dry at 50°C for approximately 2 hours, and then measure with a plate reader (TopCount: PerkinElmer). The cell lysis rate is calculated by the following formula. Cell lysis rate (%) = (A - B) / (C - B) × 100 A: Count of the sample well. B: Average value of the background (antibody non-added well) count (n = 3). When adding the antibody, add 50 μL of the test medium. Otherwise, perform the same operations as the sample well. C: Average value of the maximum release (the well where the target cells are lysed with a surfactant) count (n = 3). Add 50 μL of the test medium when adding the antibody. Add 100 μL of the surfactant, and like the sample well, aliquot 50 μL to the solid scintillation counting plate and perform the measurement. As shown in FIGS. 110(A, B), various anti-HLA-A2 / MAGEC1-CD3 bispecific molecules exhibit cytotoxic activity against T2 cells to which MAGEC1 peptide is added. On the other hand, no cytotoxic activity against T2 cells to which DMSO is added is observed. Industrial applicability The humanized anti-CD3 antibody of the present invention binds to human CD3 and also binds to cynomolgus monkey CD3, and is suitable for non-clinical trials for pharmaceutical development. No [Non-keyword text of sequence listing] Sequence identifier 1: Amino acid sequence of human CD3ε Sequence identifier 2: Amino acid sequence of human CD3δ Sequence identifier 3: Amino acid sequence of human CD3γ Sequence identifier 4: Nucleotide sequence encoding the human CD3εγ single-chain antigen Sequence identifier 5: Amino acid sequence of His-scCD3 antigen Sequence identifier 6: Nucleotide sequence encoding the heavy chain variable region of C3-147 Sequence identifier 7: (C3-147_VH AA): Amino acid sequence of the heavy chain variable region of C3-147 Sequence identifier 8: (C3-147_VL DNA): Nucleotide sequence encoding the light chain variable region of C3-147 Sequence identifier 9: (C3-147_VL AA): Amino acid sequence of the light chain variable region of C3-147 Sequence identifier 10: (G4S linker sense): Sequence identifier 11: (G4S linker antisense): Sequence identifier 12: (C3E-7000_VH AA): Amino acid sequence of the heavy chain variable region of C3E-7000 Sequence identifier 13: (C3E-7000_VL AA): Amino acid sequence of the light chain variable region of C3E-7000 Sequence identifier 14: (C3E-7000 ORF): Nucleotide sequence encoding C3E-7000 Sequence identifier 15: (C3E-7000 AA): Amino acid sequence of C3E-7000 Sequence identifier 16: (C3E-7034_VH AA): Amino acid sequence of the heavy chain variable region of C3E-7034 Sequence identifier 17: Amino acid sequence of the light chain variable region of C3E-7034 Sequence identifier 18: (C3E-7034 ORF): Nucleotide sequence encoding C3E-7034 Sequence identifier 19: (C3E_7034 AA): Amino acid sequence of C3E-7034 Sequence identifier 20: (C3E-7035_VL AA): Amino acid sequence of the light chain variable region of C3E-7035 Sequence identifier 21: (C3E-7035 ORF): Nucleotide sequence encoding C3E-7035 Sequence identifier 22: (C3E_7035 AA): Amino acid sequence of C3E-7035 Sequence identifier 23: (C3E-7036_VL_AA): Amino acid sequence of the light chain variable region of C3E-7036 Sequence identifier 24: (C3E-7036 ORF): Nucleotide sequence encoding C3E-7036 Sequence identifier 25: (C3E_7036 AA): Amino acid sequence of C3E-7036 Sequence identifier 26: (7000_CDR-H1): Amino acid sequence of CDR-H1 of the C3E-7000 series Sequence identifier 27: (7000_CDR-H2): Amino acid sequence of CDR-H2 of the C3E-7000 seriesSequence ID No. 28: (7000_CDR-H3): Amino acid sequence of CDR-H3 of the C3E-7000 series Sequence ID No. 29: (7000_CDR-L1): Amino acid sequence of CDR-L1 of the C3E-7000 series Sequence ID No. 30: (7000_CDR-L2): Amino acid sequence of CDR-L2 of the C3E-7000 series Sequence ID No. 31: (7000_CDR-L3): Amino acid sequence of CDR-L3 of the C3E-7000 series Sequence ID No. 32: (C3E-3000 ORF): Nucleotide sequence encoding OKT3 scFv Sequence ID No. 33: (C3E-3000 AA): Amino acid sequence of OKT3 scFv Sequence ID No. 34: (C3E-3007 ORF): Nucleotide sequence encoding C3E-3007 scFv Sequence ID No. 35: (C3E-3007 AA): Amino acid sequence of C3E-3007 scFv Sequence ID No. 36: (C3E-3000_VH AA): Amino acid sequence of the heavy chain variable region of OKT3 Sequence ID No. 37: (C3E-3000_VL AA): Amino acid sequence of the light chain variable region of OKT3 Sequence ID No. 38: (C3E-3007_VH AA): Amino acid sequence of the heavy chain variable region of C3E-3007 Sequence ID No. 39: (C3E-3007_VL AA): Amino acid sequence of the light chain variable region of C3E-3007 Sequence ID No. 40: (HT1-11 ORF): Nucleotide sequence encoding HT1-11 scFv Sequence ID No. 41: (HT1-11 AA): Amino acid sequence of HT1-11 scFv Sequence ID No. 42: (T2C-0001 ORF): ORF nucleotide sequence encoding T2C-0001 Sequence ID No. 43: (T2C-0001 AA): Amino acid sequence of T2C-0001 Sequence ID No. 44: (T2C-0003 ORF): ORF nucleotide sequence encoding T2C-0003 Sequence ID No. 45: (T2C-0003 AA): Amino acid sequence of T2C-0003 Sequence ID No. 46: (T2C-0005 ORF): ORF nucleotide sequence encoding T2C-0005 Sequence ID No. 47: (T2C-0005 AA): Amino acid sequence of T2C-0005 Sequence ID No. 48: (T2C-0006 ORF): ORF nucleotide sequence encoding T2C-0006 Sequence ID No. 49: (T2C-0006 AA): Amino acid sequence of T2C-0006 Sequence ID No. 50: Amino acid sequence of the sense primer for heavy chain gene amplification Sequence ID No. 51: Nucleotide sequence of the first antisense primer for heavy chain gene amplificationSequence ID No. 52: Nucleotide sequence of the second antisense primer for heavy chain gene amplification Sequence ID No. 53: Nucleotide sequence of the sense primer for light chain gene amplification Sequence ID No. 54: Nucleotide sequence of the first antisense primer for light chain gene amplification Sequence ID No. 55: Nucleotide sequence of the second antisense primer for light chain gene amplification Sequence ID No. 56: Nucleotide sequence of the sense primer for heavy chain sequencing Sequence ID No. 57: Nucleotide sequence of antisense primer 1 for light chain sequencing Sequence ID No. 58: Nucleotide sequence of antisense primer 2 for light chain sequencing Sequence ID No. 59: ORF nucleotide sequence encoding C3E-7078 Sequence ID No. 60: Amino acid sequence of C3E-7078 Sequence ID No. 61: ORF nucleotide sequence encoding C3E-7079 Sequence ID No. 62: Amino acid sequence of C3E-7079 Sequence ID No. 63: ORF nucleotide sequence encoding C3E-7085 Sequence ID No. 64: Amino acid sequence of C3E-7085 Sequence ID No. 65: ORF nucleotide sequence encoding C3E-7086 Sequence ID No. 66: Amino acid sequence of C3E-7086 Sequence ID No. 67: ORF nucleotide sequence encoding C3E-7087 Sequence ID No. 68: Amino acid sequence of C3E-7087 Sequence ID No. 69: ORF nucleotide sequence encoding C3E-7088 Sequence ID No. 70: Amino acid sequence of C3E-7088 Sequence ID No. 71: ORF nucleotide sequence encoding C3E-7089 Sequence ID No. 72: Amino acid sequence of C3E-7089 Sequence ID No. 73: ORF nucleotide sequence encoding C3E-7090 Sequence ID No. 74: Amino acid sequence of C3E-7090 Sequence ID No. 75: ORF nucleotide sequence encoding C3E-7091 Sequence ID No. 76: Amino acid sequence of C3E-7091 Sequence ID No. 77: ORF nucleotide sequence encoding C3E-7092 Sequence ID No. 78: Amino acid sequence of C3E-7092 Sequence ID No. 79: ORF nucleotide sequence encoding C3E-7093 Sequence ID No. 80: Amino acid sequence of C3E-7093 Sequence ID No. 81: ORF nucleotide sequence encoding C3E-7094 Sequence ID No. 82: Amino acid sequence of C3E-7094 Sequence ID No. 83: ORF nucleotide sequence encoding C3E-7095 Sequence ID No. 84: Amino acid sequence of C3E-7095 Sequence ID No. 85: Nucleotide sequence of HN53R Fw Sequence ID No. 86: Nucleotide sequence of HN53R Rv Sequence ID No. 87: Nucleotide sequence of HN53S Fw Sequence ID No. 88: Nucleotide sequence of HN53S Rv Sequence ID No. 89: (AXC-0001 ORF): ORF nucleotide sequence encoding AXC-0001Sequence ID No. 90: (AXC-0001 AA): Amino acid sequence of AXC-0001 Sequence ID No. 91: (AXC-0002 ORF): ORF nucleotide sequence encoding AXC-0002 Sequence ID No. 92: (AXC-0002 AA): Amino acid sequence of AXC-0002 Sequence ID No. 93: (MGC-0001 ORF): ORF nucleotide sequence encoding MGC-0001 Sequence ID No. 94: (MGC-0001 AA): Amino acid sequence of MGC-0001 Sequence ID No. 95: (MGC-0002 ORF): ORF nucleotide sequence encoding MGC-0002 Sequence ID No. 96: (MGC-0002 AA): Amino acid sequence of MGC-0002 Sequence ID No. 97: (MAGEC1 peptide): Amino acid sequence of MAGEC1 peptide Sequence ID No. 98: (CDRH2 of variant): Amino acid sequence of CDR variant CDRH2 Sequence ID No. 99: (CDRL2 of variant): Amino acid sequence of CDR variant CDRL2 Sequence ID No. 100: (VH of C3E-7034 variant): Amino acid sequence of the variable heavy chain of the CDR variant of C3E-7034 Sequence ID No. 101: (VL of C3E-7034 variant): Amino acid sequence of the variable light chain of the CDR variant of C3E-7034 Sequence ID No. 102: (VL of C3E-7035 variant): Amino acid sequence of the variable light chain of the CDR variant of C3E-7035 Sequence ID No. 103: (VL of C3E-7036 variant): Amino acid sequence of the variable light chain of the CDR variant of C3E-7036 Figure 1 is a figure showing the amino acid sequence of human CD3ε. Figure 2 is a figure showing the amino acid sequence of human CD3δ. Figure 3 is a figure showing the nucleotide sequence encoding the human CD3εγ single-chain antigen. Figure 4 is a figure showing the amino acid sequence of the human CD3εγ single-chain antigen. Figure 5 is a figure showing the nucleotide sequence of the sense primer Nhe-polyC-S for heavy chain gene amplification. Figure 6 is a figure showing the nucleotide sequence of the first anti-sense primer rIgγ-AS1 for heavy chain gene amplification. Figure 7 is a figure showing the nucleotide sequence of the second anti-sense primer rIgγ-AS2 for heavy chain gene amplification. Figure 8 is a figure showing the nucleotide sequence of the sense primer Nhe-polyC-S2 for light chain gene amplification. Figure 9 is a figure showing the nucleotide sequence of the first anti-sense primer rIgL-AS1 for light chain gene amplification. Figure 10 is a figure showing the nucleotide sequence of the second anti-sense primer rIgL-AS2 for light chain gene amplification. Figure 11 is a figure showing the nucleotide sequence of the sense primer rIgγ-seq for heavy chain sequencing. Figure 12 is a figure showing the nucleotide sequence of the anti-sense primer 1rIgL-seq1 for light chain sequencing. Figure 13 is a figure showing the nucleotide sequence of the anti-sense primer 2rIgL-seq2 for light chain sequencing. Figure 14 is a figure showing the nucleotide sequence encoding the heavy chain variable region of C3-147. Figure 15 is a figure showing the amino acid sequence of the heavy chain variable region of C3-147. Figure 16 is a figure showing the nucleotide sequence encoding the light chain variable region of C3-147. Figure 17 is a figure showing the amino acid sequence of the light chain variable region of C3-147. Figure 18 is a figure showing the oligonucleotide sequence of the sense strand of the G4S linker. Figure 19 is a figure showing the oligonucleotide sequence of the anti-sense strand of the G4S linker. Figure 20 is a figure showing the nucleotide sequence encoding C3E-7000. Figure 21 is a figure showing the amino acid sequence of C3E-7000. Figure 22 is a figure showing the amino acid sequence of the heavy chain variable region of C3E-7034. Figure 23 is a figure showing the amino acid sequence of the light chain variable region of C3E-7034. Figure 24 is a figure showing the amino acid sequence of CDR-H1 of C3E-7000. Figure 25 is a figure showing the amino acid sequence of CDR-H2 of C3E-7000. Figure 26 is a figure showing the amino acid sequence of CDR-H3 of C3E-7000. Figure 27 is a figure showing the amino acid sequence of CDR-L1 of C3E-7000. Figure 28 is a figure showing the amino acid sequence of CDR-L2 of C3E-7000. Figure 29 is a figure showing the amino acid sequence of CDR-L3 of C3E-7000. Figure 30 is a figure showing the amino acid sequence of the light chain variable region of C3E-7035. Figure 31 is a figure showing the amino acid sequence of C3E-7035.Figure 32 is a figure showing the amino acid sequence of the light chain variable region of C3E-7036. Figure 33 is a figure showing the amino acid sequence of C3E-7036. Figure 34 is a figure showing the nucleotide sequence encoding C3E-7034. Figure 35 is a figure showing the nucleotide sequence encoding C3E-7035. Figure 36 is a figure showing the nucleotide sequence encoding C3E-7036. Figure 37 is a figure showing the amino acid sequence of human CD3γ. Figure 38 is a figure showing the amino acid sequence of C3E-7034. Figure 39 is a figure showing the complex structure of CD3εγ and C3E-7034. Figure 40 is a figure showing the interaction of the heavy and light chains of CD3εγ and C3E-7034. A is a figure showing the amino acid residues of CD3ε within 4 Å of the light chain variable region of C3E-7034 and CD3ε itself in a thick stick model, and other amino acids in a thin stick model. B is a figure showing the amino acid residues of CD3ε within 4 Å of the heavy chain variable region of C3E-7034 and CD3ε itself in a thick stick model, and other amino acids in a thin stick model. Figure 41 is a figure showing the sequence interaction sites of the heavy and light chain variable regions of human CD3ε and C3E-7034. Figure 42 is a figure showing the amino acid sequence of the heavy chain variable region of OKT3. Figure 43 is a figure showing the amino acid sequence of the heavy chain variable region of C3E-3007. Figure 44 is a figure showing the amino acid sequence of the light chain variable region of OKT3. Figure 45 is a figure showing the amino acid sequence of the light chain variable region of C3E-3007. Figure 46 is a figure showing the nucleotide sequence encoding C3E-3007 scFv. Figure 47 is a figure showing the amino acid sequence of C3E-3007 scFv. Figure 48 is a figure showing the binding of C3E-3007, C3E-7034, C3E-7035, C3E-7036, which are humanized anti-CD3 scFvs, to human CD3 (PBMC). Figure 49 is a figure showing the binding of C3E-3007, C3E-7034, C3E-7035, C3E-7036, which are humanized anti-CD3 scFvs, to cynomolgus monkey CD3 (PBMC). In Figure 50, A is a figure showing the activation of human CD8-positive cells by C3E-7034 and C3E-3007, which are humanized anti-CD3 scFvs. B is a figure showing the activation of cynomolgus monkey CD8-positive cells by C3E-7034 and C3E-3007, which are humanized anti-CD3 scFvs. Figure 51 is a figure showing the amino acid sequence of HT1-11 scFv. Figure 52 is a figure showing the nucleotide sequence encoding HT1-11 scFv. Figure 53 is a figure showing the binding of HT1-11 scFv to a human TROP2-positive cell line. A is a figure showing the binding to the pharyngeal squamous cell carcinoma cell line FaDu.Figure B shows the binding to the pancreatic cancer cell line HPAF-II. Figure 54 shows the ORF nucleotide sequence encoding T2C-0001. Figure 55 shows the ORF nucleotide sequence encoding T2C-0003. Figure 56 shows the ORF nucleotide sequence encoding T2C-0005. Figure 57 shows the ORF nucleotide sequence encoding T2C-0006. Figure 58 shows the amino acid sequence of T2C-0001. Figure 59 shows the amino acid sequence of T2C-0003. Figure 60 shows the amino acid sequence of T2C-0005. Figure 61 shows the amino acid sequence of T2C-0006. Figure 62 shows the binding of the anti-TROP2-CD3 bispecific molecule, T2C-0001, T2C-0003, T2C-0005, T2C-0006 to TROP2-positive cell lines. Figure A shows the binding to the pharyngeal squamous cell carcinoma cell line FaDu. Figure B shows the binding to the pancreatic cancer cell line HPAF-II. Figure 63 shows the binding of the anti-TROP2-CD3 bispecific molecule, T2C-0001, T2C-0003, T2C-0005, T2C-0006 to human CD3 (PBMC). Figure 64 shows the binding of the anti-TROP2-CD3 bispecific molecule, T2C-0001, T2C-0003, T2C-0005, T2C-0006 to cynomolgus monkey CD3 (PBMC). Figure 65 A shows the expression of TROP2 in the pharyngeal squamous cell carcinoma cell line FaDu. B shows the expression of TROP2 in the pancreatic cancer cell line HPAF-II. C shows the absence of TROP2 expression in the lung cancer cell line Calu-6. Figure 66 A shows the cytotoxic activity of the anti-TROP2-CD3 bispecific molecule, T2C-0001, T2C-0003, T2C-0005, T2C-0006 against the pharyngeal squamous cell carcinoma cell line FaDu in the presence of human PBMC. B shows the cytotoxic activity of the anti-TROP2-CD3 bispecific molecule, T2C-0001, T2C-0003, T2C-0005, T2C-0006 against the pancreatic cancer cell line HPAF-II in the presence of human PBMC. C shows that the anti-TROP2-CD3 bispecific molecule, T2C-0001, T2C-0003, T2C-0005, T2C-0006 does not show cytotoxic activity against the human lung cancer cell line Calu-6 in the presence of human PBMC. Figure 67 shows the nucleotide sequence encoding C3E-7078. Figure 68 shows the amino acid sequence of C3E-7078. Figure 69 shows the nucleotide sequence encoding C3E-7079.Figure 70 is a figure showing the amino acid sequence of C3E-7079. Figure 71 is a figure showing the nucleotide sequence encoding C3E-7085. Figure 72 is a figure showing the amino acid sequence of C3E-7085. Figure 73 is a figure showing the nucleotide sequence encoding C3E-7086. Figure 74 is a figure showing the amino acid sequence of C3E-7086. Figure 75 is a figure showing the nucleotide sequence encoding C3E-7087. Figure 76 is a figure showing the amino acid sequence of C3E-7087. Figure 77 is a figure showing the nucleotide sequence encoding C3E-7088. Figure 78 is a figure showing the amino acid sequence of C3E-7088. Figure 79 is a figure showing the nucleotide sequence encoding C3E-7089. Figure 80 is a figure showing the amino acid sequence of C3E-7089. Figure 81 is a figure showing the nucleotide sequence encoding C3E-7090. Figure 82 is a figure showing the amino acid sequence of C3E-7090. Figure 83 is a figure showing the nucleotide sequence encoding C3E-7091. Figure 84 is a figure showing the amino acid sequence of C3E-7091. Figure 85 is a figure showing the nucleotide sequence encoding C3E-7092. Figure 86 is a figure showing the amino acid sequence of C3E-7092. Figure 87 is a figure showing the nucleotide sequence encoding C3E-7093. Figure 88 is a figure showing the amino acid sequence of C3E-7093. Figure 89 is a figure showing the nucleotide sequence encoding C3E-7094. Figure 90 is a figure showing the amino acid sequence of C3E-7094. Figure 91 is a figure showing the nucleotide sequence encoding C3E-7095. Figure 92 is a figure showing the amino acid sequence of C3E-7095. Figure 93 is a figure showing the nucleotide sequence encoding primer HN53R Fw. Figure 94 is a figure showing the nucleotide sequence encoding primer HN53R Rv. Figure 95 is a figure showing the nucleotide sequence encoding primer HN53S Fw. Figure 96 is a figure showing the nucleotide sequence encoding primer HN53S Rv. Figure 97 is a figure showing the ORF nucleotide sequence encoding AXC-0001. Figure 98 is a figure showing the amino acid sequence of AXC-0001. Figure 99 is a figure showing the ORF nucleotide sequence encoding AXC-0002. Figure 100 is a figure showing the amino acid sequence of AXC-0002. Figure 101 is a figure showing the ORF nucleotide sequence encoding MGC-0001. Figure 102 is a figure showing the amino acid sequence of MGC-0001. Figure 103 is a figure showing the ORF nucleotide sequence encoding MGC-0002. Figure 104 is a figure showing the amino acid sequence of MGC-0002. Figure 105 is a list of primers for the production of anti-CD3 antibody CDR variants. Figure 106-1 is a figure showing the binding of various anti-CD3 antibody CDR variants to human and cynomolgus monkey CD3 (PBMC).Figure 106-2 is a graph showing the binding of various anti-CD3 antibody CDR variants to human and cynomolgus monkey CD3 (PBMC). Figure 107 is a graph showing the expression of Axl in human lung cancer cell line A549 (A), human pancreatic cancer cell line PANC-1 (B), human pancreatic cancer cell line MIA PaCa-2 (C), human myeloma cell line U266B1 (D), mantle cell lymphoma cell line Jeko-1 (E). Figures 108 A, B, C are graphs showing that anti-Axl-CD3 bispecific molecules, AXC-0001, AXC-0002 have cytotoxic activity against Axl-expressing cell lines in the presence of human PBMC. Figures 108 D, E are graphs showing that anti-Axl-CD3 bispecific molecules, AXC-0001, AXC-0002 do not exhibit cytotoxic activity against Axl non-expressing cell lines in the presence of human PBMC. Figure 109 is a graph showing the binding of MAG-032scFv in MAGEC1 lymphoma (A) or human lymphoblastoid fusion cell line T2 cells added with DMSO (B). Figure 110 A is a graph showing that anti-HLA-A2 / MAGEC1-CD3 bispecific molecules, MGC-0001, MGC-0002 have cytotoxic activity against T2 cells added with MAGEC1 peptide in the presence of human PBMC. Figure 110 B is a graph showing that anti-HLA-A2 / MAGEC1-CD3 bispecific molecules, MGC-0001, MGC-0002 do not show cytotoxic activity against T2 added with DMSO in the presence of human PBMC. Figure 111 is a graph showing the amino acid sequence of MAGEC1 peptide. Figure 112 is a graph showing the amino acid sequence of the CDRH2 region of the CDR variant. The first X. aa and the second X aa are each an arbitrary natural amino acid residue. Figure 113 is a graph showing the amino acid sequence of the CDRL2 region of the CDR variant. X aa is an arbitrary natural amino acid residue. Figure 114 is a graph showing the heavy chain amino acid sequence of the CDR variant of C3E-7034. The first X aa and the second X aa are each an arbitrary natural amino acid residue. Figure 115 is a graph showing the light chain amino acid sequence of the CDR variant of C3E-7034. The X of CDRL2 aais an arbitrary natural amino acid residue. Figure 116 is a diagram showing the light chain amino acid sequence of the CDR variant of C3E-7035. X of CDRL2 aa is an arbitrary natural amino acid residue. Figure 117 is a diagram showing the light chain amino acid sequence of the CDR variant of C3E-7036. X of CDRL2 aa is an arbitrary natural amino acid residue. None.

Claims

1. An antibody or an antigen-binding fragment of the antibody, comprising: a heavy chain variable region comprising amino acid residues from positions 2 to 119 of sequence identification number 22, and a light chain variable region comprising amino acid residues from positions 137 to 243 of sequence identification number 22; or a heavy chain variable region comprising amino acid residues from positions 2 to 119 of sequence identification number 25, and a light chain variable region comprising amino acid residues from positions 135 to 241 of sequence identification number 25; wherein the antibody or the antigen-binding fragment of the antibody binds to human CD3 and cynomolgus monkey CD3.

2. The antibody of claim 1 or the antigen-binding fragment of the antibody, wherein the heavy chain variable region binds to the light chain variable region in this order, or the light chain variable region binds to the heavy chain variable region in this order, starting from the amino acid terminus, and optionally: i) having a linker between the two variable regions; ii) having a glycine residue at the amino terminus of the variable region on the amino terminus side; and iii) having a linker, a FLAG tag, and / or an HIS tag at the carboxyl terminus of the variable region on the carboxyl terminus side.

3. The antibody or antigen-binding fragment of the antibody as requested in claim 1, wherein the antibody is IgG.

4. The antibody or antigen-binding fragment of the antibody as requested in any of items 1 to 3, wherein the antigen-binding fragment of the antibody is selected from the group consisting of Fab, F(ab)', Fv, and scFv.

5. The antibody or antigen-binding fragment of the antibody as claimed in claim 4, wherein the antibody or antigen-binding fragment of the antibody is an scFv.

6. The antibody or antigen-binding fragment of the antibody as claimed in claim 5, wherein the antibody or antigen-binding fragment comprises: an amino acid sequence comprising amino acid residues at positions 2 to 243 or 269 of sequence identification number 22, or an amino acid sequence comprising amino acid residues at positions 2 to 241 or 267 of sequence identification number 25.

7. A polynucleotide comprising a nucleotide sequence encoding an antibody or an antigen-binding fragment of such antibody as claimed in any one of claims 1 to 6.

8. The polynucleotide of claim 7, wherein the polynucleotide comprises a nucleotide sequence encoding the following amino acid sequences: an amino acid sequence comprising amino acid residues at positions 2 to 243 or 269 of sequence identification number 22, or an amino acid sequence comprising amino acid residues at positions 2 to 241 or 267 of sequence identification number 25.

9. A vector comprising a polynucleotide as claimed in claim 7 or 8.

10. A cell comprising a polynucleotide as claimed in claim 7 or 8 or a carrier as claimed in claim 9.

11. Use of a cell as claimed in claim 10 for producing an antibody or an antigen-binding fragment of such antibody as claimed in any one of claims 1 to 6.

12. A method for manufacturing an antibody or an antigen-binding fragment of the antibody that binds to human CD3 and cynomolgus monkey CD3, the method comprising the steps of: culturing cells as claimed in claim 10; and recovering the antibody or the antigen-binding fragment of the antibody that binds to human CD3 from the culture.

13. A pharmaceutical composition comprising an antibody or an antigen-binding fragment of such antibody as claimed in any one of claims 1 to 6 as an active ingredient.

14. A molecule having antigen-binding activity, comprising an antibody as claimed in any one of claims 1 to 6 or an antigen-binding fragment of such antibody.

15. The molecule of claim 14, wherein the molecule is multispecific.

16. The molecule of claim 14, in addition to the antibody or antigen-binding fragment of the antibody as described in any one of claims 1 to 6, further comprises one or two other antibodies or antigen-binding fragments of the other antibodies.

17. The molecule of claim 16, wherein the antigen-binding fragment of the additional antibody is Fab, F(ab)', Fv, scFv, or sdAb.

18. The molecule of claim 17, wherein the molecule contains Fc.

19. The molecule of any one of claims 16 to 18, wherein the additional antibody is a humanized antibody or a human antibody containing the constant region of human immunoglobulin.

20. The molecule of any one of claims 16 to 18, wherein the additional antibody or the antigen-binding fragment of the additional antibody is bound to the antibody or the antigen-binding fragment of the antibody of any one of claims 1 to 6 via a linker or not via a linker.

21. The molecule of claim 20, wherein the carboxyl terminus of the amino acid sequence of the additional antibody or the antigen-binding fragment of the additional antibody is bound to a linker, and the carboxyl terminus of the amino acid sequence of the linker is further bound to the antibody or the antigen-binding fragment of the antibody of any one of claims 1 to 6.

22. A molecule of any one of claims 16 to 18, comprising an antibody of any one of claims 1 to 6 or an antigen-binding fragment of such antibody, wherein the heavy chain variable region binds to the light chain variable region in such order, or the light chain variable region binds to the heavy chain variable region in such order, starting from the amino terminus, and optionally: i) having a linker between the two variable regions; ii) having a glycine residue at the amino terminus of the variable region on the amino terminus side; and iii) having a linker, a FLAG tag, and / or an HIS tag at the carboxyl terminus of the variable region on the carboxyl terminus side.

23. The molecule of any one of claims 16 to 18 comprises: (i) an amino acid sequence comprising amino acid residues at positions 2 to 243 or 269 of sequence identification number 22, or (ii) an amino acid sequence comprising amino acid residues at positions 2 to 241 or 267 of sequence identification number 25.

24. The molecule of any one of claims 16 to 18, wherein the additional antibody is an anticancer target antibody.

25. A molecule as claimed in any of claims 16 to 18, wherein the molecule is a polypeptide.

26. A polynucleotide comprising a nucleotide sequence encoding an amino acid sequence of a molecule as claimed in claim 25.

27. A vector comprising the polynucleotide as claimed in claim 26.

28. A cell comprising a polynucleotide as claimed in claim 26 or a carrier as claimed in claim 27.

29. Use of a cell as claimed in claim 28 for producing molecules as claimed in claim 25.

30. A method for producing molecules that bind to human CD3 and cynomolgus monkey CD3, the method comprising the steps of: culturing cells as claimed in claim 28; and recovering the molecules that bind to human CD3 from the culture.

31. A pharmaceutical composition comprising a molecule as an active ingredient of any one of claims 16 to 25.

32. A pharmaceutical composition of claim 31, used to treat cancer, wherein the cancer treatment includes inducing cytotoxicity in the target tumor cells by redirection of T cells to the target tumor cells; wherein the molecule is multispecific; and wherein the molecule further comprises an antitumor antigen antibody.

33. Use of a molecule as claimed in any one of claims 16 to 25 in the manufacture of a medicament for treating cancer, wherein the cancer treatment comprises inducing cytotoxicity in the target tumor cells by redirecting T cells to the target tumor cells; wherein the molecule is multispecific; and wherein the molecule further comprises an antitumor antigen antibody.

Images