Antibodies binding to HLA-a02 / MAGE-a4 complex and uses thereof

Abstract

The present application provides a bispecific antibody comprising a first antigen-binding fragment that binds to an HLA-A02 / MAGE-A4 complex and / or an HLA-A02 / MAGE-A8 complex and a second antigen-binding fragment that binds to an activated T cell antigen, a single-domain antibody that binds to an HLA-A02 / MAGE-A4 complex and / or an HLA-A02 / MAGE-A8 complex, a pharmaceutical composition comprising the bispecific antibody or the single-domain antibody, and uses thereof.

Description

Multiple antibodies that bind to the HLA-A02 / MAGE-A4 complex and their uses Cross-reference to related applications This application claims priority to Chinese Patent Application No. 202311125718.9, filed on September 1, 2023, the entire content of which is incorporated herein by reference in its entirety. Field of the invention The present invention generally relates to the fields of genetic engineering and biopharmaceuticals; specifically, the present application relates to bispecific antibodies that bind to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex, single-domain antibodies that bind to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex, and their uses. Background of the invention Melanoma-associated antigen A (MAGE-A), belonging to the type I MAGE family, is an important member of the cancer-testis associated antigens [1] . The MAGE-A gene family is located in the region of X chromosome q28 and has 12 members (MAGE-A1 to MAGE-A12) [2] . All MAGE-A members have the same MAGE homology domain (MHD), which is a highly conserved domain composed of approximately 170 amino acids and has a high degree of sequence homology among each other [3] . Melanoma-associated antigen 4 protein (MAGE-A4) is a member of the MAGE-A gene family. MAGE-A4 is localized in the cytoplasm; it has been found that MAGE-A4 is also detected in the nucleus. The function of MAGE-A4 is not yet clear. MAGE-A4 may be involved in cell cycle progression, transcriptional control, cell survival or apoptosis. In malignant tumors, overexpressed MAGE-A4 may be associated with tumor growth and metastasis and poor prognosis of patients. The expression pattern of MAGE-A4 protein is consistent with that of MAGE-A family proteins, with restricted expression in local testicular tissue, partial expression in ovarian and placental tissues, and abnormally high expression in cancers [4,5] . Clinically histologically identified, MAGE-A4 positivity has been detected in many cancers such as urothelial carcinoma, bladder cancer, lung cancer, ovarian cancer, esophageal squamous cell carcinoma, oral squamous cell carcinoma, etc [6-11] . As an intracellular protein, MAGE-A4 can be degraded in the proteasome and, after processing, presented on the cell surface through the major histocompatibility complex I (MHC I) as a T cell antigen and then recognized by the T cell receptor (TCR) to induce an immune response, such as the polypeptide MAGE-A4 derived from MAGE-A4 230-239(GVYDGREHTV) is presented on the cell surface in complex with HLA-A02, triggering T cell recognition and specifically killing tumor cells in a major histocompatibility complex (MHC)-dependent manner.

[0012] Due to the limited expression of MAGE-A4, it may become a potential molecular target for tumor diagnosis and immunotherapy. [13,14] . Bispecific antibodies (BsAbs) that simultaneously bind to CD3 on the surface of T cells and antigens on the surface of target cells (such as tumor cells) establish an interaction between the target cells and T cells, causing the activation of cytotoxic T cells and lysing the target cells in a major histocompatibility complex (MHC)-independent manner. Such bispecific antibodies show promising prospects in the treatment of various tumors. T cell receptor-like antibodies (TCR-like antibodies), also known as TCR mimicking antibodies (TCRm), can effectively mimic TCRs and recognize MHC complexes of tumor-specific antigen peptides on the cell surface. TCRm that binds to intracellular proteins such as MAGE-A4 can serve as the part that binds the target antigen in CD3 bispecific antibodies, and such bispecific antibodies may be more attractive due to the restricted expression of the target antigen. In 2020, there were 19.29 million newly diagnosed cancer cases globally. Among them, 4.57 million new cancer cases occurred in China, accounting for 23.7% of the global total. In 2020, the global cancer death toll reached 9.96 million.

[0015] , including 1.8 million cases of lung cancer (18%), 0.92 million cases of colorectal cancer (9.2%), 0.83 million cases of liver cancer (8.3%), 0.77 million cases of gastric cancer (7.7%), 0.68 million cases of breast cancer (6.8%), 0.54 million cases of esophageal cancer (5.4%), 0.47 million cases of pancreatic cancer (4.7%), 0.38 million cases of prostate cancer (3.8%), 0.34 million cases of cervical cancer (3.4%), and 0.31 million cases of leukemia (3.1%). These ten types of cancers accounted for 70.4% of the total cancer deaths.

[0015] . Most of these cancers do not have good treatment methods, so there is an urgent need for new treatment strategies and drugs. As an intracellular antigen, MAGE-A4 is expressed in a variety of solid tumors and may become a new target for the treatment of solid tumors. Based on clinical needs, exploring and developing bispecific antibodies targeting MAGE-A4-MHC has important clinical significance. Summary of the Invention In a first aspect, the present application provides a bispecific antibody comprising a first antigen-binding fragment that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex, and a second antigen-binding fragment that binds to an activating T cell antigen. In some embodiments of the first aspect, the bispecific antibody is capable of mediating HLA-A02 + / MAGE-A4+ Tumor cells activate T cells, and / or the bispecific antibody is capable of mediating the killing of HLA-A02 + / MAGE-A4 + by PBMCs In some embodiments of the first aspect, the first antigen-binding fragment comprises an HCDR1 as set forth in SEQ ID NO:32, an HCDR2 as set forth in SEQ ID NO:33, and an HCDR3 as set forth in SEQ ID NO:34; wherein the amino acid sequences of the HCDRs are defined according to Kabat. In some embodiments of the first aspect, the first antigen-binding fragment is in the form of a single-domain antibody. In some embodiments of the first aspect, the first antigen-binding fragment comprises a monovalent or multivalent single-domain antibody that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex. In some embodiments of the first aspect, the second antigen-binding fragment comprises an HCDR1 as set forth in SEQ ID NO:37, an HCDR2 as set forth in SEQ ID NO:38, an HCDR3 as set forth in SEQ ID NO:39, an LCDR1 as set forth in SEQ ID NO:40, an LCDR2 as set forth in SEQ ID NO:41, and an LCDR3 as set forth in SEQ ID NO:42; wherein the amino acid sequences of the HCDRs are defined according to Kabat. In some embodiments of the first aspect, the second antigen-binding fragment is a single-chain antibody (scFv) or a Fab fragment. In some embodiments of the first aspect, the first antigen-binding fragment comprises a bivalent single-domain antibody that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex; optionally, the first antigen-binding fragment comprises the heavy-chain variable regions of two monovalent single-domain antibodies that bind to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex. In some embodiments of the first aspect, the first antigen-binding fragment comprises an amino acid sequence as set forth in SEQ ID NO:35, 36, 71 or 72; and / or the second antigen-binding fragment comprises a heavy-chain variable region having an amino acid sequence as set forth in SEQ ID NO:43 and a light-chain variable region having an amino acid sequence as set forth in SEQ ID NO:44. In some embodiments of the first aspect, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO: 35; and the second antigen-binding fragment comprises a heavy chain variable region with the amino acid sequence shown in SEQ ID NO: 43 and a light chain variable region with the amino acid sequence shown in SEQ ID NO: 44; The first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO: 36; and the second antigen-binding fragment comprises a heavy chain variable region with the amino acid sequence shown in SEQ ID NO: 43 and a light chain variable region with the amino acid sequence shown in SEQ ID NO: 44; The first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO: 71; and the second antigen-binding fragment comprises a heavy chain variable region with the amino acid sequence shown in SEQ ID NO: 43 and a light chain variable region with the amino acid sequence shown in SEQ ID NO: 44; or The first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO: 72; and the second antigen-binding fragment comprises a heavy chain variable region with the amino acid sequence shown in SEQ ID NO: 43 and a light chain variable region with the amino acid sequence shown in SEQ ID NO: 44. In some embodiments of the first aspect, the first antigen-binding fragment and the second antigen-binding fragment are linked by an Fc fragment of an antibody heavy chain constant region, and the Fc fragment of the antibody heavy chain constant region includes a first Fc fragment and a second Fc fragment, wherein the amino acids at positions 354 and 366 of the first Fc fragment are C and W respectively, or the amino acids at positions 349, 366, 368 and 407 of the first Fc fragment are C, S, A and V respectively; and the amino acids at positions 354 and 366 of the second Fc fragment are C and W respectively, or the amino acids at positions 349, 366, 368 and 407 of the second Fc fragment are C, S, A and V respectively; wherein the amino acid positions of the antibody constant region are determined according to EU numbering. In some embodiments of the first aspect, the first antigen-binding fragment and the second antigen-binding fragment are linked by an Fc fragment of an antibody heavy chain constant region, and the Fc fragment of the antibody heavy chain constant region includes a first Fc fragment and a second Fc fragment, wherein the amino acids at positions 234, 235 and 331 of the first Fc fragment and the second Fc fragment are F, E and S respectively; wherein the amino acid positions of the antibody constant region are determined according to EU numbering. In some embodiments of the first aspect, the first antigen-binding fragment and the second antigen-binding fragment are linked by an Fc fragment of an antibody heavy-chain constant region, and the Fc fragment of the antibody heavy-chain constant region includes a first Fc fragment and a second Fc fragment, wherein one of the first Fc fragment and the second Fc fragment is linked to the first antigen-binding fragment, and the other of the first Fc fragment and the second Fc fragment is linked to the second antigen-binding fragment. In some embodiments of the first aspect, the bispecific antibody comprises a first arm that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex and a second arm that binds to an activating T cell antigen, wherein the first arm comprises the amino acid sequence shown in SEQ ID NO:45 or 74; and the second arm comprises the amino acid sequence shown in SEQ ID NO:46. In a second aspect, the present application provides a single-domain antibody that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex, which comprises HCDR1 as shown in SEQ ID NO:32, HCDR2 as shown in SEQ ID NO:33, and HCDR3 as shown in SEQ ID NO:34; wherein, the amino acid sequences of the HCDRs are defined according to Kabat. In some embodiments of the second aspect, the single-domain antibody comprises the amino acid sequence shown in SEQ ID NO:35, 36, 71 or 72. In a third aspect, the present application provides a nucleic acid molecule that encodes the bispecific antibody described in the first aspect or the single-domain antibody described in the second aspect. In a fourth aspect, the present application provides a pharmaceutical composition that comprises the bispecific antibody described in the first aspect or the single-domain antibody described in the second aspect and a pharmaceutically acceptable excipient, diluent or carrier. In a fifth aspect, the present application provides the use of the bispecific antibody described in the first aspect, the single-domain antibody described in the second aspect or the pharmaceutical composition described in the fourth aspect in the preparation of a drug for preventing or treating HLA-A02 / MAGE-A4 positive tumors and / or HLA-A02 / MAGE-A8 positive tumors. Brief Description of the Drawings Figure 1 shows the results of ELISA detection of the binding of the TCRm antibody N18D10-Fc against the HLA-A02 / MAGE-A4 complex to the recombinant protein MHC-P1. Figure 2 shows the result of the binding of the TCRm antibody N18D10-Fc against the HLA-A02 / MAGE-A4 complex to T2 cells pulsed with the antigenic peptide MAGE-A4 230-239 and Figure 3 shows the result of the binding of the TCRm antibody N18D10 against the HLA-A02 / MAGE-A4 complex to T2 cells pulsed with the alanine-scanning mutant peptide of MAGE-A4 230- 239 (GVYDGREHTV). Figure 4 shows the result of the detection by ELISA of the binding of the TCRm antibody N18D10-h16-Fc to the recombinant protein MHC-P1 Figure 5 shows the result of the detection by ELISA of the simultaneous binding of the TCRm×CD3 bispecific antibody to MHC-P1 and CD3 Figure 6 shows the result of the activation of Jurkat-Dual cells by the TCRm×CD3 bispecific antibody-mediated antigenic peptide-pulsed T2 cells Figure 7 shows the result of the activation of Jurkat-Dual cells by the TCRm×CD3 bispecific antibody-mediated HLA-A02 + / MAGE-A4 + tumor cells Figure 8 shows the result of the killing of antigenic peptide-pulsed T2 cells by the TCRm×CD3 bispecific antibody-mediated PBMC cells Figure 9 shows the result of the killing of A375 cells by the TCRm×CD3 bispecific antibody-mediated PBMC cells Figure 10 shows the result of the killing of various antigen-positive tumor cells by the TCRm×CD3 bispecific antibody-mediated PBMC cells Figure 11 shows the result of the activation of Jurkat-Dual cells by the TCRm×CD3 bispecific antibody-mediated T2 cells pulsed with the alanine-scanning peptide of MAGE-A4 230-239 (GVYDGREHTV). Figure 12 shows the result of the activation of Jurkat-Dual cells by the TCRm×CD3 bispecific antibody-mediated T2 cells pulsed with a similar peptide Figure 13 shows the result that the TCRm×CD3 bispecific antibody does not mediate the killing of antigen-negative tumor cells by PBMC Figure 14 shows the result of the inhibition of the growth of A375 tumor cells in hPBMC-immunoreconstituted mice by the TCRm×CD3 bispecific antibody Sequence Listing SEQ ID NO:1 shows the antigenic peptide MAGE-A4 derived from Homo sapiens230-239 The amino acid sequence. SEQ ID NO:2 shows the amino acid sequence of MAGE-A4-S1 (MAGE-A4 230-239 similar peptide segment 1) derived from Homo sapiens. SEQ ID NO:3 shows the amino acid sequence of MAGE-A4-S2 derived from Homo sapiens. SEQ ID NO:4 shows the amino acid sequence of MAGE-A4-S3 derived from Homo sapiens. SEQ ID NO:5 shows the amino acid sequence of MAGE-A4-S4 derived from Homo sapiens. SEQ ID NO:6 shows the amino acid sequence of MAGE-A4-S5 derived from Homo sapiens. SEQ ID NO:7 shows the amino acid sequence of MAGE-A4-S6 derived from Homo sapiens. SEQ ID NO:8 shows the amino acid sequence of MAGE-A4-S7 derived from Homo sapiens. SEQ ID NO:9 shows the amino acid sequence of MAGE-A4-S8 derived from Homo sapiens. SEQ ID NO:10 shows the amino acid sequence of MAGE-A4-S9 derived from Homo sapiens. SEQ ID NO:11 shows the amino acid sequence of MAGE-A4-S10 derived from Homo sapiens. SEQ ID NO:12 shows the amino acid sequence of MAGE-A4-S11 derived from Homo sapiens. SEQ ID NO:13 shows the amino acid sequence of MAGE-A4-S12 derived from Homo sapiens. SEQ ID NO:14 shows the amino acid sequence of MAGE-A4-S13 derived from Homo sapiens. SEQ ID NO:15 shows the amino acid sequence of MAGE-A4-S14 derived from Homo sapiens. SEQ ID NO:16 shows the amino acid sequence of MAGE-A4-S15 derived from Homo sapiens. SEQ ID NO:17 shows the amino acid sequence of MAGE-A4-S16 derived from Homo sapiens. SEQ ID NO:18 shows the amino acid sequence of MAGE-A4-S17 derived from Homo sapiens. SEQ ID NO:19 shows the amino acid sequence of MAGE-A4-S18 derived from Homo sapiens. SEQ ID NO:20 shows the amino acid sequence of MAGE-A4-S19 derived from Homo sapiens. SEQ ID NO:21 shows the amino acid sequence of MAGE-A4-S20 derived from Homo sapiens. SEQ ID NO:22 shows the amino acid sequence of MAGE-A4-S21 derived from Homo sapiens. SEQ ID NO:23 shows the amino acid sequence of MAGE-A4-S22 derived from Homo sapiens. SEQ ID NO:24 shows the amino acid sequence of MAGE-A4-S23 derived from Homo sapiens. SEQ ID NO:25 shows the amino acid sequence of MAGE-A4-S24 derived from Homo sapiens. SEQ ID NO:26 shows the amino acid sequence of MAGE-A4-S25 derived from Homo sapiens. SEQ ID NO:27 shows the amino acid sequence of MAGE-A4-S26 derived from Homo sapiens. SEQ ID NO:28 shows the amino acid sequence of MAGE-A4-S27 derived from Homo sapiens. SEQ ID NO:29 shows the amino acid sequence of MAGE-A4-S28 derived from Homo sapiens. SEQ ID NO:30 shows the amino acid sequence of MAGE-A4-S29 derived from Homo sapiens. SEQ ID NO:31 shows the amino acid sequence of MAGE-A4-S30 derived from Homo sapiens. SEQ ID NO:32 - 34 show the amino acid sequences of HCDR1, HCDR2, and HCDR3 of the heavy chain variable regions of camel (Camelus) single domain antibodies N18D10 and N18D10-h16, respectively. SEQ ID NO:35 shows the amino acid sequence of the heavy chain variable region of camel (Camelus) single domain antibody N18D10. SEQ ID NO:36 shows the amino acid sequence of the heavy chain variable region of humanized single domain antibody N18D10-h16. SEQ ID NO:37 - 39 show the amino acid sequences of HCDR1, HCDR2, and HCDR3 of the heavy chain variable region of anti-human CD3 antibody IMCR, respectively. SEQ ID NO:40 - 42 show the amino acid sequences of LCDR1, LCDR2, and LCDR3 of the light chain variable region of anti-human CD3 antibody IMCR, respectively. SEQ ID NO:43 shows the amino acid sequence of the heavy chain variable region of anti-human CD3 antibody IMCR. SEQ ID NO:44 shows the amino acid sequence of the light chain variable region of anti-human CD3 antibody IMCR. SEQ ID NO:45 shows the amino acid sequence of the arm 2N18D10-h16-G1m3-FcHIn1 in bispecific antibody 2N18D10-h16+IMCR. SEQ ID NO:46 shows the amino acid sequence of the arm IMCR-anti-CD3-scFv-G1m3-FcKn1 in bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR. SEQ ID NO:47 shows the amino acid sequence of the HLA-A02 complex (MHC-P1) of GVYDGREHTV. SEQ ID NO:48 shows the amino acid sequence of Homo sapiens MAGE-A4. SEQ ID NO:49 shows the amino acid sequence of the linker. SEQ ID NO:50 shows the nucleotide sequence of primer PCal-CH2R. SEQ ID NO:51 shows the amino acid sequence of the extracellular region of human (Homo sapiens) CD3E (hCD3E). SEQ ID NO:52 shows the amino acid sequence of the extracellular region of human (homo sapiens) CD3D (hCD3D). SEQ ID NO:53 shows the amino acid sequence of the His tag. SEQ ID NO:54 shows the amino acid sequence of the Fc region (hFc) of human (homo sapiens) IgG1 antibody. SEQ ID NO:55 shows the amino acid sequence of the Fc region (mFc) of mouse (mus musculus) IgG2a antibody. SEQ ID NO:56 shows the amino acid sequence of the heavy chain constant region of human (homo sapiens) IgG1 subtype. SEQ ID NO:57 shows the amino acid sequence of the mutant IgG1H of the heavy chain constant region of human IgG1 subtype antibody. SEQ ID NO:58 shows the amino acid sequence of the mutant IgG1K of the heavy chain constant region of human IgG1 subtype antibody. SEQ ID NO:59 shows the amino acid sequence of the mutant IgG1m3-H of the heavy chain constant region of human IgG1 subtype antibody. SEQ ID NO:60 shows the amino acid sequence of the mutant IgG1m3-K of the heavy chain constant region of human IgG1 subtype antibody. SEQ ID NO:61 shows the amino acid sequence of the mutant IgG1m3-H1n1 of the heavy chain constant region of human IgG1 subtype antibody. SEQ ID NO:62 shows the amino acid sequence of the mutant IgG1m3-Kn1 of the heavy chain constant region of human IgG1 subtype antibody. SEQ ID NO:63 shows the amino acid sequence of the constant region of human (homo sapiens) lambda subtype light chain. SEQ ID NO:64 shows the amino acid sequence of the constant region of human (homo sapiens) kappa subtype light chain SEQ ID NO:65 shows the amino acid sequence of the variable region of the heavy chain (VH) of the germline gene antibody DP47. SEQ ID NO:66 shows the amino acid sequence of the variable region of the light chain (VK) of the germline gene antibody DP47. SEQ ID NO:67 shows the amino acid sequence of the variable region of the alpha chain (Vα) of the control TCR that binds to HLA-A02 / MAGE-4. SEQ ID NO:68 shows the amino acid sequence of the variable region of the beta chain (Vβ) of the control TCR that binds to HLA-A02 / MAGE-4. SEQ ID NO:69 shows the amino acid sequence of the first binding arm TCRA-Des-G1m3-FcKn1 of the control IMC-C103C bifunctional molecule. SEQ ID NO:70 shows the amino acid sequence of TCRB-Des-G1m3-FcH1n1 in the control TCR. SEQ ID NO:71 shows the amino acid sequence of the tandem heavy chain variable regions of two camel single domain antibodies N18D10. SEQ ID NO:72 shows the amino acid sequence of the tandem heavy chain variable regions of two camel single domain antibodies N18D10-h16. SEQ ID NO:73 shows the amino acid sequence of the single-chain antibody of the anti-human (homo sapiens) CD3 antibody IMCR. SEQ ID NO:74 shows the amino acid sequence of 2N18D10-G1m3-FcHIn1 in the bispecific antibody 2N18D10+IMCR. SEQ ID NO:75 shows the amino acid sequence of the second binding arm IMCR-anti-CD3-scFv-TCRB-Des-G1m3-FcH1n1 of the control IMC-C103C bifunctional molecule. SEQ ID NO:76 shows the amino acid sequence of the heavy chain variable region (057G03VH) of the control antibody 057D03 that binds to the HLA-A02 / MAGE-A4 complex. SEQ ID NO:77 shows the amino acid sequence of the light chain variable region (057G03VK) of the control antibody 057D03 that binds to the HLA-A02 / MAGE-A4 complex. SEQ ID NO:78 shows the amino acid sequence of the first component 057G03-Fd-EE-IMCR-anti-CD3VK-G1m3-Kn1 of the control anti-HLA-A02 / MAGE-A4 complex and CD3 bispecific antibody 057G03. SEQ ID NO:79 shows the amino acid sequence of the second component 057G03-Fd-EE-G1m3-FcH1n1 of the control anti-HLA-A02 / MAGE-A4 complex and CD3 bispecific antibody 057G03. SEQ ID NO:80 shows the amino acid sequence of the third component 057G03VK+CK-RK of the control anti-HLA-A02 / MAGE-A4 complex and CD3 bispecific antibody 057G03. SEQ ID NO:81 shows the amino acid sequence of the fourth component, IMCR-anti-CD3VH-CK, of the control anti-HLA-A02 / MAGE-A4 complex and CD3 bispecific antibody 057G03. SEQ ID NO:82 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position G1. SEQ ID NO:83 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position V2. SEQ ID NO:84 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position Y3. SEQ ID NO:85 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position D4. SEQ ID NO:86 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position G5. SEQ ID NO:87 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position R6. SEQ ID NO:88 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position E7. SEQ ID NO:89 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position H8. SEQ ID NO:90 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position T9. SEQ ID NO:91 shows MAGE-A4 230-239 The amino acid sequence of the alanine-scanning mutant peptide at position V10. SEQ ID NO:92 shows the amino acid sequence of the human IgG1 subtype antibody Fc segment mutant IgG1m3-FcH1n1. SEQ ID NO:93 shows the amino acid sequence of the human IgG1 subtype antibody Fc segment mutant IgG1m3-FcKn1. SEQ ID NO:94 shows the amino acid sequence of human (homo sapiens) MAGE-A8. DETAILED DESCRIPTION OF THE INVENTION The inventors of the present application prepared a bispecific antibody (such as a TCRm bispecific antibody) by genetic engineering means, which contains a first antigen-binding fragment that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex and a second antigen-binding fragment that binds to an activating T cell antigen (such as the CD3 molecule). By binding to MAGE-A4 presented on the surface of target cells (tumor cells) through the first antigen-binding fragment that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex 230-239 (GVYDGREHTV) complex with HLA-A02 and / or MAGE-A8 232-241 (GLYDGREHSV) complex with HLA-A02, while the second antigen-binding fragment binds to an activating T cell antigen (such as the CD3 molecule), thereby establishing an interaction between the target cell and the T cell, causing activation of cytotoxic T cells, and lysing the target cell (such as a tumor cell) in a major histocompatibility complex (MHC)-dependent manner, so as to achieve the purpose of treating diseases (such as tumors). In multiple aspects of the present application, there are provided new bispecific antibodies containing a first antigen-binding fragment that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex and a second antigen-binding fragment that binds to an activating T cell antigen (such as the CD3 molecule), single-domain antibodies that bind to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex, nucleic acid molecules encoding the bispecific antibody or the single-domain antibody, vectors containing the nucleic acid molecules, host cells containing the nucleic acid molecules or vectors, methods for preparing and purifying the bispecific antibody or the single-domain antibody, and medical and biological applications of the bispecific antibody or the single-domain antibody. According to the sequences of the bispecific antibodies or single-domain antibodies provided by the present application, bispecific antibodies or single-domain antibodies that bind to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex can be constructed and used as drugs clinically for preventing or treating tumors. Unless otherwise specified, the implementation of the present application uses conventional molecular biology, microbiology, cell biology, biochemistry, and immunology techniques in the art. Unless otherwise specified, the terms used in the present application have the meanings commonly understood by those skilled in the art. Definitions As used herein, the term "HLA-A02 / MAGE-A4 complex" refers to a complex of an HLA-A02 molecule and MAGE-A4. In a specific embodiment of the present application, the "HLA-A02 / MAGE-A4 complex" refers to a complex of an HLA-A02 molecule and the MAGE-A4 of positions 230-239 shown in SEQ ID NO:48 (MAGE-A4 230-239 ) of the HLA-A02 / MAGE-A4 complex 230-239 complex. As used herein, the term "HLA-A02 / MAGE-A8 complex" refers to a complex of an HLA-A02 molecule and MAGE-A8. In a specific embodiment of the present application, the "HLA-A02 / MAGE-A8 complex" refers to a complex of an HLA-A02 molecule and the MAGE-A84 of positions 232-241 shown in SEQ ID NO:94 (MAGE-A8 232-241 ) of the HLA-A02 / MAGE-A8 complex 232-241 complex. As used herein, the term "activating T cell antigen" refers to an antigenic determinant expressed on the surface of T lymphocytes, particularly cytotoxic T lymphocytes, which can induce T cell activation upon interaction with an antigen-binding molecule. Specifically, the interaction between the antigen-binding molecule and the activating T cell antigen can induce T cell activation by triggering the signal transduction cascade of the T cell receptor complex. In certain aspects, the activating T cell antigen is CD3. As used herein, the term "antibody" refers to an immunoglobulin molecule capable of specifically binding to a target via at least one antigen recognition site located in the variable region of the immunoglobulin molecule. Targets include, but are not limited to, carbohydrates, polynucleotides, lipids, polypeptides, etc. The "antibody" used herein includes not only intact (i.e., full-length) antibodies, but also their binding fragments (e.g., Fab, Fab’, F(ab’)2, Fv), their variants, fusion proteins containing antibody portions, humanized antibodies, chimeric antibodies, bispecific antibodies, linear antibodies, single-chain antibodies, single-domain antibodies, multispecific antibodies (e.g., bispecific antibodies), and any other modified configurations of immunoglobulin molecules containing the antigen recognition site with the required specificity, including glycosylation variants of antibodies, amino acid sequence variants of antibodies, and covalently modified antibodies. Typically, a complete or full-length antibody comprises two heavy chains and two light chains. Each heavy chain contains a heavy chain variable region (VH) and the first, second and third constant regions (CH1, CH2 and CH3). Each light chain contains a light chain variable region (VL) and a constant region (CL). A full-length antibody can be any type of antibody, such as IgD, IgE, IgG, IgA or IgM (or subclasses thereof), but the antibody need not belong to any particular class. Immunoglobulins can be assigned to different classes based on the antibody amino acid sequence of the constant region of the heavy chain. Typically, there are five major classes of immunoglobulins: IgA, IgD, IgE, IgG and IgM, and several of these classes can be further divided into subclasses (isotypes), such as IgG1, IgG2, IgG3, IgG4, IgA1 and IgA2. The heavy chain constant regions corresponding to different classes of immunoglobulins are called α, δ, ε, γ, and μ, respectively. The subunit structure and three-dimensional structure of different classes of immunoglobulins are well known. As used herein, the term "bispecific antibody" refers to an antibody that has the ability to bind to two antigen epitopes at the same time. The two antigen epitopes can be on different antigens or on the same antigen. Bispecific antibodies can have a variety of structural configurations. For example, a bispecific antibody can be composed of two Fc fragments and two binding parts fused thereto respectively (similar to natural antibodies, the difference is that the two arms bind to different antigen targets or epitopes), and the antigen binding part can be a single domain antibody, a single chain antibody (scFv) or a Fab fragment. When targeting the epitopes of two given antigens, the two different binding parts of the bispecific antibody each bind to the N-terminus of an Fc fragment, and the antigen binding part configuration of the two arms can have four combinations: single domain antibody + Fab fragment, single domain antibody + scFv, scFv + single domain antibody and Fab fragment + single domain antibody. The Fc fragment can contain mutations that can ensure heavy chain heteropolymerization, and the KIH technology (knob-in-hole, KIH) is a strategy to solve heavy chain heteropolymerization. Generally, KIH technology refers to the formation of a structure that is conducive to the pairing of heterologous half antibodies by modifying the amino acid sequence of the CH3 region, which can form a bispecific antibody while maintaining the structure of a normal antibody as much as possible. For guidance on KIH technology, see, for example, "An efficient route to human bispecific IgG", A. Margaret Merchant et al., Nature Biotechnology, Volume 16, 1998, which is incorporated herein by reference in its entirety. As used herein, the terms "binding portion" or "binding fragment" are used interchangeably and refer to a part or region of a complete antibody molecule responsible for binding an antigen. The antigen-binding domain can include the heavy chain variable region (VH), the light chain variable region (VL), or both. Each of VH and VL typically contains three complementarity determining regions CDR1, CDR2, and CDR3. It is well known to those skilled in the art that the complementarity determining regions (CDRs, usually CDR1, CDR2, and CDR3) are the regions in the variable region that have the greatest impact on the affinity and specificity of an antibody. There are two common ways to define the CDR sequences of VH or VL, namely the Kabat definition and the Chothia definition. (See, e.g., Kabat, "Sequences of Proteins of Immunological Interest", National Institutes of Health, Bethesda, Md. (1991); A1-Lazikani et al., J. Mol. Biol. 273:927-948 (1997); and Martin et al., Proc. Natl. Acad. Sci. USA 86:9268-9272 (1989)). For a given variable region sequence of an antibody, the CDR region sequences in the VH and VL sequences can be determined according to the Kabat definition or the Chothia definition. In the embodiments of the present application, the Kabat definition is used to define the CDR sequences. For a given variable region sequence of an antibody, the CDR region sequences in the variable region sequence can be analyzed in a variety of ways. For example, they can be determined using the online software Abysis (http: / / www.abysis.org / ). Examples of antigen-binding fragments for a general antibody include, but are not limited to: (1) Fab fragments, which can be monovalent fragments having a VL-CL chain and a VH-CH1 chain; (2) F(ab’)2 fragments, which can be bivalent fragments having two Fab’ fragments connected by a disulfide bridge in the hinge region (i.e., a dimer of Fab’); (3) Fv fragments having the VL and VH domains of a single arm of an antibody; (4) single-chain Fv (scFv), which can be a single polypeptide chain composed of a VH domain and a VL domain via a peptide linker; (5) (scFv)2, which can contain two VH domains and two VL domains connected by a peptide linker, and the two VL domains are combined with the two VH domains via a disulfide bridge; and (6) single-domain antibody forms. In the construction of bispecific antibodies, the "binding portion" includes, but is not limited to, single-domain antibody forms, Fab fragment forms, and / or single-chain antibody (scFv) forms. As used herein, the term "single chain fragment variable (scFv)" refers to an antibody having a single chain structure generally constructed by genetic engineering techniques, and comprising a polypeptide chain having a heavy chain variable region (VH) and a light chain variable region (VL). A flexible linker is usually designed between the heavy chain variable region and the light chain variable region so that the heavy chain variable region and the light chain variable region can fold into a correct conformation capable of binding an antigen. As used herein, the terms "Fab (fragment antigen binding) fragment", "Fab portion" or similar terms refer to an antibody fragment capable of binding an antigen produced by treating a whole antibody with papain, and include a whole light chain (VL-CL), a heavy chain variable region and a CH1 fragment (VH-CH1). As used herein, the term "single domain antibody" refers to a heavy chain single variable domain antibody that is naturally lacking a light chain. Such an antibody comprises a heavy chain variable region (VHH) and conventional CH2 and CH3 regions (e.g., one or two sets (heavy chain variable region (VHH) and conventional CH2 and CH3 regions)). The VHH structure cloned and expressed alone has a structural stability comparable to that of the original heavy chain antibody and a binding activity to an antigen, and is the smallest unit known to be capable of binding a target antigen. A single domain antibody is also referred to as a Nanobody (Nb). As used herein, the terms "Fc fragment", "Fc domain" and "Fc portion" are used interchangeably and refer to a part of the constant region of an antibody heavy chain, including a hinge region, a CH2 fragment and a CH3 fragment of the heavy chain constant region, and are determined with reference to the EU numbering of human IgG1 antibody. As used herein, the term "specifically bind" refers to a non-random binding reaction between two molecules, such as the binding of an antibody to an antigen epitope. As used herein, the term "tumor" refers to a neoplasm or solid lesion formed by abnormal cell growth. A tumor can be benign, pre-malignant or malignant. As used herein, the term "malignant tumor" refers to or describes a physiological condition in a mammal, typically characterized by unregulated cell growth. Exemplary malignant tumors include: carcinoma, solid tumor, melanoma, sarcoma, hematologic tumor, germ cell tumor, and embryonal cell tumor. More specific examples of malignant tumors include: esophageal cancer, urothelial cancer, oral cancer, sarcoma, head and neck cancer, bladder cancer, melanoma, ovarian cancer, colorectal cancer, breast cancer, renal cell cancer, lung cancer including small cell lung cancer, non-small cell lung cancer, lung adenocarcinoma, and lung squamous cell carcinoma, cervical cancer, hepatic carcinoma, gastric cancer including gastrointestinal cancer, prostate cancer, pancreatic cancer, peritoneal cancer, hepatocellular carcinoma, glioblastoma, liver cancer, urinary tract cancer, hepatoma, endometrial or uterine cancer, salivary gland cancer, squamous cell carcinoma (e.g., squamous epithelial cell carcinoma), vulvar cancer, thyroid cancer, anal cancer, penile cancer, multiple myeloma, and B-cell lymphoma, brain cancer, and related metastases. As used herein, the term "hematologic tumor" refers to a condition caused by uncontrolled growth and proliferation of abnormal cells, which in most cases originate in the bone marrow, the site where blood cells are produced. Exemplary hematologic tumors include various types of leukemia, multiple myeloma, and malignant lymphoma. More specific examples of hematologic tumors include: acute lymphoblastic leukemia (ALL), chronic lymphocytic leukemia (CLL), acute myeloid leukemia (AML), chronic myelogenous leukemia (CML), hairy cell leukemia (HCL), T-cell prolymphocytic leukemia, large granular lymphocyte leukemia, juvenile myelomonocytic leukemia, B-cell prolymphocytic leukemia, Burkitt leukemia, and adult T-cell leukemia, non-Hodgkin lymphoma, B-cell lymphoma, small lymphocytic lymphoma, lymphoplasmacytic lymphoma, Waldenström macroglobulinemia ( macroglobulinemia), splenic marginal zone lymphoma, plasmacytoma, extranodal marginal zone B-cell lymphoma, MALT lymphoma, nodal marginal zone B-cell lymphoma (NMZL), follicular lymphoma, mantle cell lymphoma, diffuse large B-cell lymphoma, mediastinal (thymic) large B-cell lymphoma, intravascular large B-cell lymphoma, primary effusion lymphoma, Burkitt lymphoma, B-cell chronic lymphocytic lymphoma, classical Hodgkin lymphoma, nodular lymphocyte-predominant Hodgkin lymphoma, adult T-cell lymphoma, extranodal nasal type NK / T-cell lymphoma, enteropathy-type T-cell lymphoma, hepatosplenic T-cell lymphoma, blastic NK-cell lymphoma, mycosis fungoides, Sézary syndrome, primary cutaneous CD30-positive T-cell lymphoproliferative disease, primary cutaneous anaplastic large cell lymphoma, lymphomatoid papulosis, angioimmunoblastic T-cell lymphoma, peripheral T-cell lymphoma, unspecified, and anaplastic large cell lymphoma. As used herein, the term "solid tumor" refers to a palpable mass that can be detected by clinical examinations such as radiography, CT scan, B-ultrasound, or palpation. Clinically diagnosed solid tumors are divided into two types: malignant and benign. Malignant solid tumors include: esophageal cancer (such as esophageal squamous cell carcinoma), urothelial carcinoma, oral cancer (such as oral squamous cell carcinoma), sarcoma (such as synovial sarcoma), head and neck cancer, liver cancer, lung cancer, bladder cancer, melanoma (metastatic melanoma), ovarian cancer, colorectal cancer, breast cancer, childhood Hodgkin lymphoma: lymphocyte-predominant type, nodular sclerosis type, mixed cell type, lymphocyte depletion type; childhood non-Hodgkin lymphoma: prelymphoblastic lymphoma, small non-cleaved cell lymphoma (Burkitt / non-Burkitt lymphoma), diffuse large B-cell lymphoma, anaplastic large cell lymphoma, etc.; childhood kidney tumors: nephroblastoma (Wilms tumor), renal clear cell carcinoma, renal rhabdoid tumor, renal clear cell sarcoma, primitive neuroectodermal tumor of the kidney, etc.; childhood neuroblastoma: neuroblastoma, ganglioneuroblastoma, ganglioneuroma; childhood extracranial germ cell tumors: mature teratoma, immature teratoma, endodermal sinus tumor (yolk sac tumor), seminoma, dysgerminoma, choriocarcinoma, embryonal carcinoma, etc.; osteosarcoma and chondrosarcoma; childhood rhabdomyosarcoma: embryonal type, alveolar type, pleomorphic type, etc.; soft tissue sarcoma: fibrosarcoma, malignant fibrous histiocytoma, liposarcoma, leiomyosarcoma, angiosarcoma, lymphangiosarcoma, malignant schwannoma, alveolar soft part sarcoma, epithelioid sarcoma, clear cell sarcoma, malignant melanoma, synovial sarcoma, desmoplastic small round cell tumor, etc.; Ewing family sarcoma: Ewing sarcoma, primitive neuroectodermal tumor; childhood liver tumors: hepatoblastoma (embryonal type, fetal type, undifferentiated type), hepatocellular carcinoma; retinoblastoma; other tumors: posterior fossa medulloblastoma, nasopharyngeal carcinoma, papillary thyroid carcinoma, thymoma, pulmonary blastoma, pancreaticoblastoma, islet cell tumor, ileocecal carcinoid, mesothelioma, etc. Benign solid tumors include: lymphangioma, hemangioma, thyroglossal duct cyst, etc. In a first aspect, the present application provides a bispecific antibody that comprises a first antigen-binding fragment that binds to an HLA-A02 / MAGE-A4 complex and / or an HLA-A02 / MAGE-A8 complex and a second antigen-binding fragment that binds to an activating T cell antigen. In some embodiments of the first aspect, the bispecific antibody is capable of mediating the activation of T cells by HLA-A02 + / MAGE-A4 + tumor cells, and / or the bispecific antibody is capable of mediating the killing of HLA-A02 + / MAGE-A4 + tumor cells by PBMCs. In some embodiments of the first aspect, the bispecific antibody is capable of mediating HLA-A02 + / MAGE-A4 + Tumor cells activate Jurkat-Dual cells. In some embodiments of the first aspect, the binding epitope of the first antigen-binding fragment to the HLA-A02 / MAGE-A4 complex comprises positions 230-239 of MAGE-A4 as shown in reference to SEQ ID NO: 48. In some embodiments, the amino acid sequence of positions 230-239 of MAGE-A4 as shown in reference to SEQ ID NO: 48 is GVYDGREHTV (SEQ ID NO: 1). In some embodiments of the first aspect, the binding epitope of the first antigen-binding fragment to the HLA-A02 / MAGE-A4 complex comprises at least one of residues 232, 234, 235, 237, and 239 of MAGE-A4 as shown in reference to SEQ ID NO: 48. In some specific embodiments, the binding epitope of the first antigen-binding fragment to the HLA-A02 / MAGE-A4 complex comprises residues 232, 234, 235, 237, and 239 of MAGE-A4 as shown in reference to SEQ ID NO: 48. In some embodiments, residues 232, 234, 235, 237, and 239 of MAGE-A4 as shown in reference to SEQ ID NO: 48 correspond to residues 3, 5, 6, 8, and 10 of the MAGE-A4 230-239 (GVYDGREHTV) polypeptide. In some embodiments of the first aspect, the binding epitope of the first antigen-binding fragment to the HLA-A02 / MAGE-A8 complex comprises positions 232-241 of MAGE-A8 as shown in reference to SEQ ID NO: 94. In some embodiments, the amino acid sequence of positions 232-241 of MAGE-A8 as shown in reference to SEQ ID NO: 94 is GLYDGREHSV (SEQ ID NO: 2). In some embodiments of the first aspect, the activating T cell antigen is the CD3 molecule. In some embodiments of the first aspect, the first antigen-binding fragment comprises HCDR1 as shown in SEQ ID NO: 32, HCDR2 as shown in SEQ ID NO: 33, and HCDR3 as shown in SEQ ID NO: 34; wherein the amino acid sequences of the HCDRs are defined according to Kabat. In some embodiments of the first aspect, the first antigen-binding fragment is in the form of a single-domain antibody. In some embodiments of the first aspect, the first antigen-binding fragment comprises a monovalent or multivalent (e.g., 1, 2, 3, 4, 5, or 6-valent) single-domain antibody that binds to the HLA-A02 / MAGE-A4 complex. In some specific embodiments of the first aspect, the first antigen-binding fragment comprises a bivalent single-domain antibody that binds to the HLA-A02 / MAGE-A4 complex. In some embodiments, the first antigen-binding fragment comprises the heavy chain variable regions of two monovalent single-domain antibodies that bind to the HLA-A02 / MAGE-A4 complex. In some embodiments, the first antigen-binding fragment comprises the heavy chain variable regions of two monovalent single-domain antibodies that bind to the HLA-A02 / MAGE-A4 complex and are linked by a linker. In some embodiments, the linker is a linker comprising a GS flexible peptide, such as (GGGGS) n , where n is an integer ≥ 1. In some specific embodiments, the linker is GGGGSGGGGSGGGGS (SEQ ID NO:49). In some embodiments of the first aspect, the first antigen-binding fragment comprises a monovalent or multivalent (e.g., 1, 2, 3, 4, 5, or 6-valent) single-domain antibody that binds to the HLA-A02 / MAGE-A8 complex. In some specific embodiments of the first aspect, the first antigen-binding fragment comprises a bivalent single-domain antibody that binds to the HLA-A02 / MAGE-A8 complex. In some embodiments, the first antigen-binding fragment comprises the heavy chain variable regions of two monovalent single-domain antibodies that bind to the HLA-A02 / MAGE-A8 complex. In some embodiments, the first antigen-binding fragment comprises the heavy chain variable regions of two monovalent single-domain antibodies that bind to the HLA-A02 / MAGE-A8 complex and are linked by a linker. In some embodiments, the linker is a linker comprising a GS flexible peptide, such as (GGGGS) n , where n is an integer ≥ 1. In some specific embodiments, the linker is GGGGSGGGGSGGGGS (SEQ ID NO:49). In some embodiments of the first aspect, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:35. In some embodiments of the first aspect, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:36. In some embodiments of the first aspect, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:71. In some embodiments of the first aspect, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:72. In some embodiments of the first aspect, the amino acid sequence shown in SEQ ID NO:71 comprises two amino acid sequences shown in SEQ ID NO:35. In some embodiments of the first aspect, the amino acid sequence shown in SEQ ID NO:72 comprises two amino acid sequences shown in SEQ ID NO:36. In some embodiments of the first aspect, the second antigen-binding fragment comprises a heavy chain variable region having the amino acid sequence shown in SEQ ID NO:43 and a light chain variable region having the amino acid sequence shown in SEQ ID NO:44. In some embodiments of the first aspect, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:35; and the second antigen-binding fragment comprises a heavy chain variable region having the amino acid sequence shown in SEQ ID NO:43 and a light chain variable region having the amino acid sequence shown in SEQ ID NO:44. In some embodiments of the first aspect, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:36; and the second antigen-binding fragment comprises a heavy chain variable region having the amino acid sequence shown in SEQ ID NO:43 and a light chain variable region having the amino acid sequence shown in SEQ ID NO:44. In some embodiments of the first aspect, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:71; and the second antigen-binding fragment comprises a heavy chain variable region having the amino acid sequence shown in SEQ ID NO:43 and a light chain variable region having the amino acid sequence shown in SEQ ID NO:44. In some embodiments of the first aspect, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:72; and the second antigen-binding fragment comprises a heavy chain variable region having the amino acid sequence shown in SEQ ID NO:43 and a light chain variable region having the amino acid sequence shown in SEQ ID NO:44. In some embodiments of the first aspect, the amino acid sequence of the first antigen-binding fragment differs from the amino acid sequence shown in SEQ ID NO:35, 36, 71 or 72 by about 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid substitutions, deletions and / or additions. In some embodiments of the first aspect, the amino acid sequence of the first antigen-binding fragment has at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or higher homology with the amino acid sequence shown in SEQ ID NO: 35, 36, 71 or 72. In some embodiments of the first aspect, the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 35, 36, 71 or 72 may also be truncated by about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids, while still maintaining a similar function of the first antigen-binding fragment. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids may also be added to the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 35, 36, 71 or 72, and the resulting amino acid sequence still maintains a similar function of the first antigen-binding fragment. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids may be added or deleted from a region other than the C-terminal or N-terminal of the amino acid sequence shown in SEQ ID NO: 35, 36, 71 or 72, as long as the altered amino acid sequence substantially maintains a similar function of the first antigen-binding fragment. In some embodiments of the first aspect, the amino acid sequence of the heavy chain variable region of the second antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO: 43. In some embodiments of the first aspect, the amino acid sequence of the heavy chain variable region of the second antigen-binding fragment differs from the amino acid sequence shown in SEQ ID NO: 43 by about 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid substitutions, deletions, and / or additions. In some embodiments of the first aspect, the amino acid sequence of the heavy chain variable region of the second antigen-binding fragment has at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or higher homology with the amino acid sequence shown in SEQ ID NO: 43. In some embodiments of the first aspect, the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 43 can also be truncated by about 1, 2, 3, 4, 5, 6,7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids while still maintaining a function similar to that of the heavy chain variable region of the second antigen-binding fragment. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids can also be added to the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 43, and the resulting amino acid sequence still maintains a function similar to that of the heavy chain variable region of the second antigen-binding fragment. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids can be added or deleted from a region other than the C-terminal or N-terminal of the amino acid sequence shown in SEQ ID NO: 43, as long as the altered amino acid sequence substantially maintains a function similar to that of the heavy chain variable region of the second antigen-binding fragment. In some embodiments of the first aspect, the amino acid sequence of the light chain variable region of the second antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO: 44. In some embodiments of the first aspect, the amino acid sequence of the light chain variable region of the second antigen-binding fragment differs from the amino acid sequence shown in SEQ ID NO: 44 by about 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid substitutions, deletions, and / or additions. In some embodiments of the first aspect, the amino acid sequence of the light chain variable region of the second antigen-binding fragment has at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or higher homology with the amino acid sequence shown in SEQ ID NO: 44. In some embodiments of the first aspect, the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 44 can also be truncated by about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids while still maintaining a function similar to that of the light chain variable region of the second antigen-binding fragment. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids can also be added to the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 44, and the resulting amino acid sequence still maintains a similar function to the light chain variable region of the second antigen-binding fragment. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids can also be added or deleted from a region other than the C-terminal or N-terminal of the amino acid sequence shown in SEQ ID NO: 44, as long as the altered amino acid sequence substantially maintains a similar function to the light chain variable region of the second antigen-binding fragment. In some embodiments of the first aspect, the first antigen-binding fragment and the second antigen-binding fragment are linked by an Fc fragment of an antibody heavy chain constant region, and the Fc fragment of the antibody heavy chain constant region includes a first Fc fragment and a second Fc fragment, wherein the amino acids at positions 354 and 366 of the first Fc fragment are C and W respectively, or the amino acids at positions 349, 366, 368 and 407 of the first Fc fragment are C, S, A and V respectively; and the amino acids at positions 354 and 366 of the second Fc fragment are C and W respectively, or the amino acids at positions 349, 366, 368 and 407 of the second Fc fragment are C, S, A and V respectively; wherein the amino acid positions of the antibody constant region are determined according to EU numbering. In some specific embodiments of the first aspect, the first antigen-binding fragment and the second antigen-binding fragment are linked by an Fc fragment of an antibody heavy chain constant region, and the Fc fragment of the antibody heavy chain constant region includes a first Fc fragment and a second Fc fragment, wherein the amino acids at positions 354 and 366 of the first Fc fragment are C and W respectively, and the amino acids at positions 349, 366, 368 and 407 of the second Fc fragment are C, S, A and V respectively; wherein the amino acid positions of the antibody constant region are determined according to EU numbering. In some embodiments of the first aspect, the first antigen-binding fragment and the second antigen-binding fragment are linked by an Fc fragment of an antibody heavy chain constant region, and the Fc fragment of the antibody heavy chain constant region includes a first Fc fragment and a second Fc fragment, wherein the amino acids at positions 234, 235 and 331 of the first Fc fragment and the second Fc fragment are F, E and S respectively; wherein the amino acid positions of the antibody constant region are determined according to EU numbering. In some embodiments of the first aspect, the first antigen-binding fragment and the second antigen-binding fragment are linked by an Fc fragment of an antibody heavy chain constant region, and the Fc fragment of the antibody heavy chain constant region includes a first Fc fragment and a second Fc fragment, wherein one of the first Fc fragment and the second Fc fragment is linked to the first antigen-binding fragment, and the other of the first Fc fragment and the second Fc fragment is linked to the second antigen-binding fragment. In some embodiments of the first aspect, the first antigen-binding fragment (e.g., at the C-terminus) is linked to the N-terminus of the first Fc fragment (e.g., having an amino acid sequence such as SEQ ID NO: 92). In some embodiments of the first aspect, the second antigen-binding fragment (e.g., at the C-terminus) is linked to the N-terminus of the second Fc fragment (e.g., having an amino acid sequence such as SEQ ID NO: 93). In some embodiments of the first aspect, the Fc fragment of the antibody heavy chain constant region is an Fc fragment of the IgG1 subtype. In some embodiments, the first Fc fragment is an Fc fragment of the IgG1 subtype; and / or the second Fc fragment is an Fc fragment of the IgG1 subtype. In some embodiments of the first aspect, the Fc fragment of the antibody heavy chain constant region is an Fc fragment of the IgG1m3 subtype. In some embodiments, the first Fc fragment is an Fc fragment of the IgG1m3 subtype; and / or the second Fc fragment is an Fc fragment of the IgG1m3 subtype. In some embodiments of the first aspect, the bispecific antibody comprises a first arm that binds to the HLA-A02 / MAGE-A4 complex and a second arm that binds to an activating T cell antigen, wherein the first arm comprises the amino acid sequence shown in SEQ ID NO: 45 or 74; and the second arm comprises the amino acid sequence shown in SEQ ID NO: 46. In some embodiments of the first aspect, the bispecific antibody comprises a first arm that binds to the HLA-A02 / MAGE-A8 complex and a second arm that binds to an activating T cell antigen, wherein the first arm comprises the amino acid sequence shown in SEQ ID NO: 45 or 74; and the second arm comprises the amino acid sequence shown in SEQ ID NO: 46. In some embodiments of the first aspect, the amino acid sequence of the first arm comprises the amino acid sequence shown in SEQ ID NO: 45 or 74. In some embodiments of the first aspect, the amino acid sequence of the first arm differs from the amino acid sequence shown in SEQ ID NO: 45 or 74 by about 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid substitutions, deletions, and / or additions. In some embodiments of the first aspect, the amino acid sequence of the first arm has at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or higher homology with the amino acid sequence shown in SEQ ID NO: 45 or 74. In some embodiments of the first aspect, the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 45 or 74 can also be truncated by about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids while still maintaining a similar function of the first arm. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids can also be added to the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 45 or 74, and the resulting amino acid sequence still maintains a similar function of the first arm. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids can be added or deleted from a region other than the C-terminal or N-terminal of the amino acid sequence shown in SEQ ID NO: 45 or 74, as long as the altered amino acid sequence substantially maintains a similar function of the first arm. In some embodiments of the first aspect, the amino acid sequence of the second arm comprises the amino acid sequence shown in SEQ ID NO: 46. In some embodiments of the first aspect, the amino acid sequence of the second arm differs from the amino acid sequence shown in SEQ ID NO: 46 by about 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid substitutions, deletions, and / or additions. In some embodiments of the first aspect, the amino acid sequence of the second arm has at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or higher homology with the amino acid sequence shown in SEQ ID NO: 46. In some embodiments of the first aspect, the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 46 may also be truncated by about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids, while still maintaining the function of the second arm similar to that described above. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids may also be added to the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 46, and the resulting amino acid sequence still maintains the function of the second arm similar to that described above. In some embodiments of the first aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids may be added or deleted in a region other than the C-terminal or N-terminal of the amino acid sequence shown in SEQ ID NO: 46, as long as the altered amino acid sequence substantially maintains the function of the second arm similar to that described above. In a second aspect, the present application provides a single-domain antibody that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex, which comprises HCDR1 as shown in SEQ ID NO: 32, HCDR2 as shown in SEQ ID NO: 33, and HCDR3 as shown in SEQ ID NO: 34; wherein, the amino acid sequence of the HCDR is defined according to Kabat. In some embodiments of the second aspect, the single-domain antibody comprises the amino acid sequence shown in SEQ ID NO: 35. In some embodiments of the second aspect, the single-domain antibody comprises the amino acid sequence shown in SEQ ID NO: 36. In some embodiments of the second aspect, the single-domain antibody comprises the amino acid sequence shown in SEQ ID NO: 71. In some embodiments of the second aspect, the single-domain antibody comprises the amino acid sequence shown in SEQ ID NO: 72. In some embodiments of the second aspect, the amino acid sequence of the single-domain antibody differs from the amino acid sequence shown in SEQ ID NO: 35, 36, 71 or 72 by about 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid substitutions, deletions, and / or additions. In some embodiments of the second aspect, the amino acid sequence of the single-domain antibody has at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or higher homology with SEQ ID NO: 35, 36, 71 or 72. In some embodiments of the second aspect, the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 35, 36, 71 or 72 may also be truncated by about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids, while still maintaining the function of the similar single-domain antibody. In some embodiments of the second aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids may also be added to the C-terminal or N-terminal region of the amino acid sequence shown in SEQ ID NO: 35, 36, 71 or 72, and the resulting amino acid sequence still maintains the function of the similar single-domain antibody. In some embodiments of the second aspect, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25 or more amino acids may also be added or deleted from a region other than the C-terminal or N-terminal of the amino acid sequence shown in SEQ ID NO: 35, 36, 71 or 72, as long as the modified amino acid sequence substantially maintains the function of the similar single-domain antibody. In a third aspect, the present application provides a nucleic acid molecule encoding the bispecific antibody described in the first aspect or the single-domain antibody described in the second aspect. In some embodiments of the third aspect, the nucleic acid molecule may include a DNA molecule and an RNA molecule. The nucleic acid molecule may be single-stranded or double-stranded and may be cDNA. In some embodiments of the third aspect, the nucleic acid molecule is operably linked to a regulatory amino acid sequence that can be recognized by a host cell transformed with the vector. In a fourth aspect, the present application provides a pharmaceutical composition comprising the bispecific antibody described in the first aspect or the single-domain antibody described in the second aspect, and a pharmaceutically acceptable excipient, diluent or carrier. In some embodiments of the fourth aspect, the pharmaceutical composition is used for preventing or treating HLA-A02 / MAGE-A4 positive tumors. In some embodiments of the fourth aspect, the HLA-A02 / MAGE-A4 positive tumors are selected from: esophageal cancer (such as esophageal squamous cell carcinoma), urothelial carcinoma, oral cancer (such as oral squamous cell carcinoma), sarcoma (such as synovial sarcoma), head and neck cancer, lung cancer, liver cancer, bladder cancer, melanoma (such as metastatic melanoma), ovarian cancer, colorectal cancer, and breast cancer. In some embodiments of the fourth aspect, the pharmaceutical composition is used for preventing or treating HLA-A02 / MAGE-A8 positive tumors. In some embodiments of the fourth aspect, the HLA-A02 / MAGE-A8 positive tumors are selected from: esophageal cancer (such as esophageal squamous cell carcinoma), urothelial carcinoma, oral cancer (such as oral squamous cell carcinoma), sarcoma (such as synovial sarcoma), head and neck cancer, lung cancer, liver cancer, bladder cancer, melanoma (such as metastatic melanoma), ovarian cancer, colorectal cancer, and breast cancer. In some embodiments of the fourth aspect, the pharmaceutical composition may further comprise one or more of the following substances: lubricants such as talc, magnesium stearate, and mineral oil; wetting agents; emulsifiers; suspending agents; preservatives such as benzoic acid, sorbic acid, and calcium propionate; sweeteners and / or flavoring agents, etc. In some embodiments of the fourth aspect, the pharmaceutical composition in the present application can be formulated into forms such as tablets, pills, powders, lozenges, elixirs, suspensions, emulsions, solutions, syrups, suppositories, or capsules. In some embodiments of the fourth aspect, the pharmaceutical composition of the present application can be delivered by any physiologically acceptable administration method, including but not limited to: oral administration, parenteral administration, nasal administration, rectal administration, intraperitoneal administration, intravascular injection, subcutaneous administration, transdermal administration, inhalation administration, etc. In some embodiments of the fourth aspect, the pharmaceutical composition for therapeutic use can be formulated for storage in the form of a lyophilized preparation or an aqueous solution by mixing reagents with the desired purity with pharmaceutically acceptable carriers, excipients, etc. as appropriate. In a fifth aspect, the present application provides the use of the bispecific antibody described in the first aspect, the single-domain antibody described in the second aspect, or the pharmaceutical composition described in the fourth aspect in the preparation of a drug for preventing or treating HLA-A02 / MAGE-A4 positive tumors and / or HLA-A02 / MAGE-A8 positive tumors. In some embodiments of the fifth aspect, the HLA-A02 / MAGE-A4 positive tumors are selected from: esophageal cancer (such as esophageal squamous cell carcinoma), urothelial carcinoma, oral cancer (such as oral squamous cell carcinoma), sarcoma (such as synovial sarcoma), head and neck cancer, lung cancer, liver cancer, bladder cancer, melanoma (such as metastatic melanoma), ovarian cancer, colorectal cancer, and breast cancer. In some embodiments of the fifth aspect, the HLA-A02 / MAGE-A8 positive tumors are selected from: esophageal cancer (such as esophageal squamous cell carcinoma), urothelial carcinoma, oral cancer (such as oral squamous cell carcinoma), sarcoma (such as synovial sarcoma), head and neck cancer, lung cancer, liver cancer, bladder cancer, melanoma (such as metastatic melanoma), ovarian cancer, colorectal cancer, and breast cancer. In a sixth aspect, the present application provides a method for preventing or treating HLA-A02 / MAGE-A4 positive tumors and / or HLA-A02 / MAGE-A8 positive tumors, comprising administering to an individual in need thereof the bispecific antibody described in the first aspect, the single-domain antibody described in the second aspect, or the pharmaceutical composition described in the fourth aspect. In some embodiments of the sixth aspect, the HLA-A02 / MAGE-A4 positive tumors are selected from: esophageal cancer (such as esophageal squamous cell carcinoma), urothelial carcinoma, oral cancer (such as oral squamous cell carcinoma), sarcoma (such as synovial sarcoma), head and neck cancer, lung cancer, liver cancer, bladder cancer, melanoma (such as metastatic melanoma), ovarian cancer, colorectal cancer, and breast cancer. In some embodiments of the sixth aspect, the HLA-A02 / MAGE-A8 positive tumors are selected from: esophageal cancer (such as esophageal squamous cell carcinoma), urothelial carcinoma, oral cancer (such as oral squamous cell carcinoma), sarcoma (such as synovial sarcoma), head and neck cancer, lung cancer, liver cancer, bladder cancer, melanoma (such as metastatic melanoma), ovarian cancer, colorectal cancer, and breast cancer. The present application also provides a vector comprising a nucleic acid molecule encoding the bispecific antibody described in the first aspect or the single-domain antibody described in the second aspect, and a host cell comprising the nucleic acid molecule or the vector. In other aspects, the present application also provides a method for producing the bispecific antibody described in the first aspect or the single-domain antibody described in the second aspect. In some embodiments, the method for producing the bispecific antibody described in the first aspect or the single-domain antibody described in the second aspect comprises culturing a host cell to facilitate the expression of the nucleic acid molecule. In some embodiments, the method for producing the bispecific antibody described in the first aspect or the single-domain antibody described in the second aspect further comprises recovering the bispecific antibody or the single-domain antibody from the host cell culture medium. It should be understood that the above detailed description is only to make those skilled in the art understand the content of the present application more clearly, and is not intended to limit in any way. Those skilled in the art can make various changes and modifications to the described embodiments. The following examples are for illustrative purposes only and are not intended to limit the scope of the present application. Example Example 1: Preparation of Recombinant Protein Refer to the literature

[0016] Prepare the antigen peptide MAGE-A4 230-239 (GVYDGREHTV) disulfide trap single chain trimer (dtSCT) structure MHC class I complex, namely HLA-A02 complex of GVYDGREHTV (MHC-P1, SEQ ID NO: 47); and extracellular region of human CD3E (hCD3E, SEQ ID NO: 51), extracellular region of human CD3D (hCD3D, SEQ ID NO: 52). These recombinant proteins have a large number of post-translational modifications (such as glycosylation or disulfide bonds, etc.), so using a mammalian cell expression system will be more conducive to maintaining the structure and function of the recombinant protein. Add a His tag (His, SEQ ID NO: 53) or the Fc segment of human antibody IgG1 (Fc, SEQ ID NO: 54) or the Fc segment of murine antibody IgG2a (mFc, SEQ ID NO: 55) to the C-terminus of the recombinant protein, which will be more conducive to the purification of the recombinant protein and the identification of monoclonal antibody function. When preparing recombinant antibodies, the heavy chain constant region of the antibody can be human IgG1 subtype (SEQ ID NO: 56) or various mutants of the selected human IgG1 subtype, such as: IgG1H (SEQ ID NO: 57), IgG1K (SEQ ID NO: 58), IgG1m3-H (SEQ ID NO: 59), IgG1m3-K (SEQ ID NO: 60), IgG1m3-H1n1 (SEQ ID NO: 61) or IgG1m3-Kn1 (SEQ ID NO: 62), and the light chain constant region can be human λ subtype (SEQ ID NO: 63) or human κ subtype (SEQ ID NO: 64). Recombinant protein amino acid sequences were reorganized for various purposes according to the Uniprot database, and genes of the above recombinant proteins (including His-tag, mFc or Fc-encoding genes) were designed and synthesized. The synthesized recombinant protein genes were cloned into appropriate eukaryotic expression vectors (such as pcDNA3.1 from Invitrogen) using conventional molecular biology techniques, and then the prepared recombinant protein expression plasmids were transfected into HEK293 cells (such as HEK293F cells from invitrogen) using liposomes (such as 293fectin from Invitrogen) or other cationic transfection reagents (such as PEI), and cultured for 3 - 4 days under serum-free suspension culture conditions. Then the culture supernatant was harvested by centrifugation or other means. The recombinant protein with His-tag fusion expression was purified in one step from the recombinant protein in the culture supernatant using a metal chelate affinity chromatography column (such as HisTrap FF from GE). The recombinant protein with mFc / Fc fusion expression was purified in one step using a Protein A / G affinity chromatography column (such as Mabselect SURE from GE). Then the storage buffer of the recombinant protein was replaced with PBS (pH 7.0) or other appropriate buffers using a desalting column (such as Hitrap desaulting from GE). If necessary, the antibody sample can be filtered and sterilized, and then aliquoted and stored at -20°C. Example 2: Screening antibodies that bind to the HAL-A02 / MAGE-A4 complex from a camel immune repertoire 2.1 Construction of a camel immune repertoire One healthy adult Bactrian camel was selected, and blood was collected before immunization to retain background serum. For the first immunization, 1 mg of MHC-P1-His fusion protein was emulsified with Freund's complete adjuvant and injected subcutaneously at multiple points; after a two-week interval, a booster immunization was given. 1 mg of MHC-P1-His fusion protein was emulsified with Freund's incomplete adjuvant and injected subcutaneously at multiple points. A total of 5 booster immunizations were carried out, and blood was collected before each immunization for antibody titer analysis; for the seventh immunization, no adjuvant was added, and 1 mg of MHC-P1-His fusion protein was used as the antigen and injected subcutaneously at multiple points for a boost immunization. 200 ml of peripheral blood was collected 3 days later for lymphocyte separation. Lymphocytes were isolated from 200 mL of camel peripheral blood using a camel peripheral blood lymphocyte isolation kit (Solarbio, CAT#P5750); total RNA of lymphocytes was extracted using a cell total RNA extraction kit (Tiangen Biochemical Technology (Beijing) Co., Ltd., CAT#DP430); the extracted total RNA was used as a template, and the first-strand cDNA synthesis kit (Thermo scientific, CAT#K1621) was used to synthesize the camel single-domain antibody heavy chain variable region (VHH), and the reverse transcription primer used a gene-specific primer, and the primer pairing region was located in the CH2 domain of the antibody heavy chain constant region, and the specific sequence was PCal-CH2R: TCCTTCCCCGTCAGCCAGTCCT (SEQ ID NO: 50). The synthesized cDNA was immediately stored at -70°C for future use; then the cDNA obtained by reverse transcription was used as a template, and references were used.

[0017] Primers were synthesized and the camel single domain antibody VHH gene was isolated by nested PCR amplification. Finally, the amplified VHH gene was cloned into the vector pADSCFV-S (see Chinese Patent Application No. 201510097117.0).

[0018] ) to construct a VHH library. The library capacity of this antibody library reached 1.5E+08, and the accuracy was 75%. 2.2 Screening of camel immune library The recombinant MHC-P1-His prepared in Example 1 was used as an antigen and a solid phase screening strategy was used (the experimental scheme refers to Phage Display: General Experimental Guide / Clackson, T., Lowman, HB, ed.; Ma Lan et al., Chemical Industry Press, May 2008).

[0019] ) The phage library displaying camel single-domain antibodies constructed in Example 2.1 was screened, and three rounds of screening were performed by binding, elution, neutralization, infection, and amplification, and finally the TCR mimic (TCRm) antibody N18D10 that specifically binds to MHC-P1-His was obtained. Using conventional molecular biological methods, the nucleotide sequence encoding the N18D10 heavy chain variable region (SEQ ID NO: 35) was cloned into a eukaryotic expression vector (such as Invitrogen's pcDNA3.1, etc.) fused with a nucleotide sequence encoding the Fc segment of human antibody IgG1 to express the N18D10-Fc recombinant protein. At the same time, DP47 (germline gene antibody, refer to US patent application US20160200833A1) was prepared.

[0020] , the amino acid sequence of DP47VH is shown in SEQ ID NO: 65; the amino acid sequence of DP47VK is shown in SEQ ID NO: 66) monoclonal antibody DP47-IgG1 was used as a negative control. Reference patent application WO 2017175006A1

[0021] , the variable region genes of Vα (TCRA, SEQ ID NO: 67) and Vβ (TCRB, SEQ ID NO: 68) for synthesizing IMC-C103C were respectively cloned into a eukaryotic expression vector fused with the nucleotide sequence of the Fc fragment containing the Knob mutation and Cα, and a eukaryotic expression vector fused with the nucleotide sequence of the Fc fragment containing the Hole mutation and Cβ. The two chains of the TCR molecule were TCRA-Des-G1m3-FcKn1 (SEQ ID NO: 69) and TCRB-Des-G1m3-FcH1n1 (SEQ ID NO: 70) respectively, and the TCR control molecule binding to MHC-P1 was co-expressed and named P1-TCR. Example 3: Identification of TCRm antibodies binding to the HLA-A02 / MAGE-A4 complex 3.1 Affinity analysis of TCRm antibodies against the HLA-A02 / MAGE-A4 complex The affinity of the TCRm antibody N18D10 against HLA-A02 / MAGE-A4 was determined by surface plasmon resonance technology using Biacore T200. Related reagents and consumables such as the amine coupling kit (BR-1000-50), human antibody capture kit (BR-1008-39), S series CM5 chips (14100530), and 10×HBS-EP (BR100669) at pH 7.4 were all purchased from GE healthcare. According to the instructions in the kit, the surface of the carboxylated CM5 chip was activated with 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS). The anti-human IgG (Fc) antibody (capture antibody) was diluted to 25 μg / mL with 10 mM sodium acetate at pH 5.0 and then injected at a flow rate of 10 μL / min to achieve a coupling amount of approximately 10,000 response units (RU). After injecting the capture antibody, 1 M ethanolamine was injected to block the unreacted groups. For kinetic measurements, the TCRm antibody N18D10-Fc was diluted to 1 μg / mL and injected at 10 μL / min to ensure that about 60 RU of the antibody was captured by the anti-human Fc antibody. Then a series of concentration gradients of MHC-P1-His (such as 1.23 nM, 3.74 nM, 11.1 nM, 33.3 nM, 100 nM) were set, and injected from low concentration to high concentration at 30 μL / min at 25 °C. The binding time was 90 s and the dissociation time was 1200 s. A 3 M MgCl2 solution was injected at 10 μL / min for 30 s to regenerate the chip surface. Using Biacore T200 evaluation software version 3.2.1, the association rate (K a ) and dissociation rate (K d ) were calculated by fitting the association and dissociation sensorgrams with a 1:1 binding model. The dissociation equilibrium constant (K D ) was calculated as the ratio K d / K a . The fitting results are shown in Table 1. Table 1. Affinity of the TCRm antibody N18D10-Fc against the HLA-A02 / MAGE-A4 complex binding to MHC-P1 3.2 Binding of the TCRm antibody against the HLA-A02 / MAGE-A4 complex to the recombinant protein MHC-P1 Coat the 96-well ELISA plate with the TCRm antibody N18D10-Fc against the HLA-A02 / MAGE-A4 complex at a concentration of 5 μg / mL, 100 μL per well, and incubate overnight at 4°C. Block the plate with the blocking solution PBS-0.1% Tween 20-3% skim milk at 37°C for 1 hour. Gradient dilute the recombinant protein MHC-P1-His with PBS, starting from a concentration of 30 μg / mL, with a 3-fold gradient dilution, for a total of 10 concentration gradients. Add 100 μL per well to the blocked 96-well ELISA plate and incubate at 37°C for 1 hour. Wash the ELISA plate with PBS-0.1% Tween 20, add the HRP-labeled anti-His tag mouse monoclonal antibody (Beijing ComWin Biotech Co., Ltd., CW0285M), and incubate at 37°C for 1 hour. Wash the ELISA plate with PBS-0.1% Tween 20, add the OPD substrate chromogenic solution, and terminate the color development with 1 M H2SO4 after 5-10 minutes. Measure the optical density value at 492 nm / 630 nm using a microplate reader. The results in Figure 1 show that N18D10-Fc specifically binds to MHC-P1. 3.3 Binding of the TCRm antibody against the HLA-A02 / MAGE-A4 complex to antigen peptide-pulsed T2 cells Collect T2 cells (human lymphoma cells, HLA-A02 + , purchased from Shanghai Hongshun Biotech Co., Ltd.) in the logarithmic growth phase, resuspend them in the growth medium to 1×10 6 cells / mL after centrifugation, and plate 1 mL per well in a 24-well plate. Chemically synthesize the peptide MAGE-A4 230-239 (GVYDGREHTV) (GenScript Biotech Corporation), and add it to the 24-well plate at a final concentration of 50 μg / mL for overnight pulsing. After 16 hours, centrifuge, and resuspend the cells in PBS buffer containing 1% BSA to 2×10 6 cells / mL. Plate 100 μL per well in a 96-well V-bottom plate, and remove the supernatant after centrifugation. Prepare samples of the antibody N18D10-Fc, the positive control sample P1-TCR, and the negative control DP47-IgG1 with PBS at a starting final concentration of 200 nM and 4-fold dilutions. Add 100 μL per well to the wells containing cells and incubate at 4°C for 1 hour. Then wash 3 times with 200 μL PBS, incubate with goat anti-human IgG-FITC (Beijing Zhongshan Jinqiao Biotechnology Co., Ltd., ZF-0308), 100 μL per well, and incubate in the dark at 4°C for 30 minutes. Then wash 3 times with 200 μL PBS, resuspend in 100 μL PBS, and detect the FITC channel using a flow cytometer (ACEA, Novocyte). The results in Figure 2 show that N18D10-Fc specifically binds to the antigen peptide MAGE-A4. 230-239The T2 cells of the pulse, namely the HLA-A02 / MAGE-A4 complex on the surface of N18D10-Fc binding cells. 3.4 Key amino acids of the TCRm antibody against the HLA-A02 / MAGE-A4 complex binding to MAGE-A4 Chemically synthesize MAGE-A4 230-239 (GVYDGREHTV) alanine-scanning mutant peptides. Refer to Example 3.3 to detect the binding of N18D10 to MAGE-A4 230-239 (GVYDGREHTV) alanine-scanning mutant peptide-pulsed T2 cells. Collect T2 cells in the logarithmic growth phase, centrifuge and resuspend them in growth medium to 1×10 6 cells / mL, plate 1 mL per well in a 24-well plate. MAGE-A4 230-239 (GVYDGREHTV) and alanine-scanning mutant peptides are added to the 24-well plate at a final concentration of 50 μg / mL and pulsed overnight. After 16 hours, centrifuge, and after centrifugation, resuspend in PBS buffer containing 1% BSA to 2×10 6 cells / mL, plate 100 μL per well in a 96-well V-bottom plate, and remove the supernatant after centrifugation. Prepare the antibody N18D10-Fc in PBS to a final concentration of 10 nM, add 100 μL per well to the cell plate, and incubate at 4°C for 1 hour. Then wash 3 times with 200 μL PBS, incubate with goat anti-human IgG-FITC (Beijing Zhongshan Jinqiao Biotechnology Co., Ltd., ZF-0308), 100 μL per well, and incubate in the dark at 4°C for 30 minutes. Then wash 3 times with 200 μL PBS, resuspend with 100 μL PBS, and detect the FITC channel with a flow cytometer (ACEA, Novocyte). Calculate the value of the binding of N18D10 to the alanine-scanning peptide relative to the original peptide. The results in Table 2 and Figure 3 show that when the 3rd, 5th, 6th, 8th, or 10th position of MAGE-A4 230-239 (GVYDGREHTV) is substituted with alanine, the binding ability of the TCRm antibody N18D10 decreases by more than 90%, that is, the 3rd, 5th, 6th, 8th, and 10th positions of MAGE-A4 230-239 (GVYDGREHTV) are the key amino acids for N18D10 binding. Table 2. Binding of the TCRm antibody N18D10 to T2 cells pulsed with alanine-scanning mutant peptides of MAGE-A4 230-239 (GVYDGREHTV) Example 4: Humanization and identification of the TCRm antibody against the HLA-A02 / MAGE-A4 complex 4.1 Humanization of the TCRm antibody against the HLA-A02 / MAGE-A4 complex The TCRm antibody N18D10 was humanized to reduce its immunogenicity. The humanization protocol adopted the classical framework transplantation strategy.

[0022] . The amino acid sequence of N18D10 was compared with the human antibody germline gene sequences in the IMGT database. Suitable germline gene sequences were selected to provide framework regions 1 to 3 (FR1+FR2+FR3) of the antibody, and a suitable J region gene sequence was selected to provide framework region 4 (FR4). This template can be selected based on various factors, such as: the relative total length of the antibody, the size of the CDRs, the amino acid residues at the junctions between the antibody framework regions (FRs) and hypervariable regions (CDRs), the overall sequence homology, etc. The selected template can be a mixture of multiple sequences or can be a consensus template, aiming to maintain the appropriate conformation of the parental complementarity-determining regions (CDRs) as much as possible. At the same time, considering the characteristics of the humanized antibody such as solubility, stability, and expression yield, back mutations were performed on 4 hotspot amino acids 37F / 44E / 45R / 47F in its FR2, and finally the humanized molecule N18D10-h16 was obtained. Using conventional molecular biology techniques, the nucleotide sequence encoding the variable region of N18D10-h16 (SEQ ID NO:36) was cloned into a eukaryotic expression vector (such as pcDNA3.1 from Invitrogen) fused with the nucleotide sequence encoding the Fc segment of human antibody IgG1 to express the N18D10-h16-Fc recombinant protein. 4.2 Affinity analysis of the humanized molecule of the TCRm antibody against the HLA-A02 / MAGE-A4 complex Referring to Example 3.1, the affinity of the recombinant TCRm antibody N18D10-h16-Fc against the HLA-A02 / MAGE-A4 complex was analyzed using a Biacore T200, and the results are shown in Table 3. Table 3. Affinity of the TCRm antibody N18D10-h16-Fc against the HLA-A02 / MAGE-A4 complex for binding to the MHC-P1 complex 4.3 Binding of the TCRm antibody N18D10-h16-Fc against the HLA-A02 / MAGE-A4 complex to MHC-P1 Referring to Example 3.2, the binding activity of the antibody N18D10-h16-Fc against the HLA-A02 / MAGE-A4 complex to the recombinant protein MHC-P1-His was measured. The results in Figure 4 showed that N18D10-h16-Fc specifically bound to the recombinant protein MHC-P1-His, that is, N18D10-h16-Fc bound to the HLA-A02 / MAGE-A4 complex. Example 5: Preparation and identification of the TCRm×CD3 bispecific antibody 5.1 Preparation of TCRm×CD3 Bispecific Antibody The nucleotide sequences encoding the heavy chain variable regions of the bivalent 2N18D10 or bivalent 2N18D10-h16 single-domain antibodies against the HLA-A02 / MAGE-A4 complex and the nucleotide sequence encoding the single-chain antibody against CD3 (IMCR-anti-CD3-scFv, see anti-CD3 v9 in US Patent Application No. US 5821337A

[0023] ) were cloned into appropriate eukaryotic expression vectors to co-express the bispecific antibody against the HLA-A02 / MAGE-A4 complex and CD3. That is, the nucleotide sequences encoding 2N18D10 (SEQ ID NO:71) or 2N18D10-h16 (SEQ ID NO:72) were cloned into the eukaryotic expression vector fused with the nucleotide sequence of the Fc fragment IgG1m3-FcH1n1 (SEQ ID NO:92) encoding the Hole mutation, and the nucleotide sequence encoding IMCR-anti-CD3-ScFv (SEQ ID NO:73) was cloned into the eukaryotic expression vector fused with the nucleotide sequence of the Fc fragment IgG1m3-FcKn1 (SEQ ID NO:93) encoding the Knob mutation. Meanwhile, the DP47 bispecific antibody with the same structure was prepared as a negative control. The eukaryotic expression vectors expressing 2N18D10-G1m3-FcHIn1 (SEQ ID NO:74) or 2N18D10-h16-G1m3-FcH1n1 (SEQ ID NO:45) were co-transfected into HEK293F cells with the eukaryotic expression vector expressing IMCR-anti-CD3-ScFv-G1m3-FcKn1 (SEQ ID NO:46) using liposomes and cultured for 3 - 5 days under serum-free suspension culture conditions, and then the culture supernatant was harvested by centrifugation or other means. The bispecific antibody in the culture supernatant was purified using a Protein A affinity chromatography column (such as Mabselect SURE from GE Healthcare), and then the storage buffer of the recombinant protein was replaced with PBS (pH 7.0) or other appropriate buffers using a desalting column (such as Hitrap desaulting from GE Healthcare). The desalted protein solution was purified by size exclusion chromatography (SEC) using Superdex200 (GE) to obtain the target proteins, named 2N18D10+IMCR and 2N18D10-h16+IMCR respectively. If necessary, the antibody samples can be filtered and sterilized, and then aliquoted and stored at -20°C for later use. Referring to the patent application WO 2017175006A1, a single-chain antibody of CD3, IMCR-anti-CD3-scFv, was fused to the N-terminus of the TCRB-Des-G1m3-FcH1n1 chain to prepare a bifunctional molecule IMC-C103C based on TCR and CD3. The two chains constituting the TCR and CD3 bifunctional molecule are TCRA-Des-G1m3-FcKn1 (SEQ ID NO:69) and IMCR-anti-CD3-scFv-TCRB-Des-G1m3-FcH1n1 (SEQ ID NO:75), respectively. Referring to the patent application WO 2021122875A1

[0024] , the variable region sequences of 057D03 were synthesized (the amino acid sequence of 057G03VH is shown in SEQ ID NO:76; the amino acid sequence of 057G03VK is shown in SEQ ID NO:77), and a bispecific antibody 057G03 against the HLA-A02 / MAGE-A4 complex and CD3 based on the structure in the patent was prepared. The four chains constituting the bispecific antibody are 057G03-Fd-EE-IMCR-anti-CD3VK-G1m3-Kn1 (SEQ ID NO:78), 057G03-Fd-EE-G1m3-FcH1n1 (SEQ ID NO:79), 057G03VK+CK-RK (SEQ ID NO:80), and IMCR-anti-CD3VH-CK (SEQ ID NO:81). The expression and purification of the above IMC-C103C and 057G03 bispecific antibodies were referred to the expression and purification of the above 2N18D10+IMCR and 2N18D10-h16+IMCR. Then, the samples were filtered and sterilized, and then aliquoted and stored at -20 °C for later use. 5.2 Affinity analysis of TCRm×CD3 bispecific antibodies Referring to Example 3.1, the affinity analysis of the recombinant TCRm×CD3 bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR was performed using Biacore T200, and the results are shown in Tables 4 and 5. Table 4. Affinity of TCRm×CD3 bispecific antibodies for MHC-P1 Table 5. Affinity of TCRm×CD3 bispecific antibodies for CD3D-CD3E-mFc 5.3 TCRm×CD3 bispecific antibodies simultaneously recognize the antigens MHC-P1 and CD3 Detect the simultaneous binding of TCRm×CD3 bispecific antibody to the bidirectional antigens CD3 and MHC-P1 using the conventional ELISA method. Coat a 96-well ELISA plate with CD3D-CD3E-mFc antigen (3 μg / mL, 100 μL per well) and incubate overnight at 4°C. Block with blocking solution PBS-0.1% Tween 20-3% skim milk at 37°C for 1 hour. Gradient dilute the TCRm×CD3 bispecific antibody with PBS, starting from a concentration of 10 μg / mL, with a 3-fold dilution for a total of 10 concentration gradients, and add 100 μL per well to the blocked 96-well ELISA plate. Incubate at 37°C for 1 hour. Wash the ELISA plate with PBS-0.1% Tween 20, add MHC-P1-His antigen (10 μg / mL, 100 μL per well), and incubate at 37°C for 1 hour. Wash the ELISA plate with PBS-0.1% Tween 20, then add HRP-labeled anti-His tag mouse monoclonal antibody (Beijing ComWin Biotech Co., Ltd., cw0285M) and incubate at 37°C for 1 hour. Wash the ELISA plate with PBS-0.1% Tween 20, add OPD substrate chromogenic solution, terminate the color development with 1 M H2SO4 after 5-10 minutes, and measure the optical density value at 492 nm / 630 nm using an ELISA reader. The results of ELISA analysis are shown in Figure 5. The TCRm×CD3 bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR can simultaneously recognize the bidirectional antigens CD3 and MHC-P1. 5.4 Activation of Jurkat-Dual cells mediated by TCRm×CD3 bispecific antibody 5.4.1 Activation of Jurkat-Dual cells by TCRm×CD3 bispecific antibody-mediated antigen peptide-pulsed T2 cells Collect T2 cells in the logarithmic growth phase, centrifuge and resuspend with growth medium to 1×10 6 cells / mL, and plate 1 mL per well in a 24-well plate. Peptide MAGE-A4 230-239 (GVYDGREHTV) is added to the 24-well plate at a final concentration of 50 μg / mL and pulsed overnight. After 16 hours, centrifuge and resuspend with RPMI1640 medium to 4×10 5 cells / mL, and plate 50 μL per well in a cell plate. Collect Jurkat-Dual cells (purchased from Invivogen) in the logarithmic growth phase, centrifuge and resuspend with RPMI1640 medium to 4×10 5cells / mL, add 50 μL per well to the cell plate seeded with T2 cells above to obtain a final E:T ratio of 1:1. Then add bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR (initial concentration is 7.5 nM, 4-fold serial dilution, 8 concentration points, 100 μL per well). Negative control sample 2DP47+IMCR, positive control samples 057G03 and IMC-C103C, the usage concentrations are the same as those of bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR. Incubate for 24 h, take the supernatant, and refer to the QUANTI-Luc TM Instruction Manual (QUANTI-Luc TM , Invivogen, rep-qlc2) to detect and analyze the specific activation of antigen peptide-pulsed T2 cells by TCRm×CD3 bispecific antibodies against Jurkat-Dual cells. The results showed that bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR could specifically mediate the activation of Jurkat-Dual cells by antigen peptide-pulsed T2 cells. Compared with 057G03 and IMC-C103C, the activation of Jurkat-Dual cells by bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR was more obvious (Figure 6), and the EC 50 values are shown in Table 6. Table 6. EC 50 values of TCRm×CD3 bispecific antibody-mediated activation of Jurkat-Dual cells by exogenous pulsed antigen peptide segments 5.4.2 TCRm×CD3 bispecific antibody-mediated HLA-A02 + / MAGE-A4 + tumor cells activate Jurkat-Dual cells Collect A375 cells (human malignant melanoma cells, HLA-A02 + / MAGE-A4 + , purchased from Nanjing Kebai Biotechnology Co., Ltd.) in the logarithmic growth phase. After digestion and centrifugation, resuspend with RPMI1640 medium to 4×10 5 cells / mL, and seed 50 μL per well in the cell plate. Collect Jurkat-Dual cells in the logarithmic growth phase, and after centrifugation, resuspend with RPMI1640 medium to 4×10 5cells / mL, 50 μL per well was added to the cell plate to obtain a final E:T ratio of 1:1. Then bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR (starting concentration of 30 nM, 4-fold serial dilution, 10 concentration points, 100 μL per well) were added. Negative control sample 2DP47+IMCR, positive control samples 057G03 and IMC-C103C were used at the same concentrations as bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR. After incubation for 48 hours, the supernatant was taken and the specific activation of tumor cells on Jurkat-Dual cells mediated under different conditions was detected and analyzed with reference to the QUANTI-Luc TM Instruction Manual (QUANTI-Luc TM , Invivogen, rep-qlc2). The results showed that TCRm×CD3 bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR could mediate HLA-A02 + / MAGE-A4 + tumor cells A375 to activate Jurkat-Dual cells. Compared with 057G03 and IMC-C103C, TCRm×CD3 bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR had stronger activation activity on Jurkat-Dual cells (Figure 7), and the EC 50 values are shown in Table 7. Table 7 TCRm×CD3 bispecific antibodies mediate HLA-A02 + / MAGE-A4 + tumor cells to activate Jurkat-Dual cells 5.5 TCRm×CD3 bispecific antibodies mediate specific killing of target cells by PBMC 5.5.1 Isolation of human peripheral blood mononuclear cells (PBMC) Blood (50 mL each) was collected from normal volunteers, and all volunteers had signed informed consent forms. The volunteer inclusion criteria were: aged over 18 years old; no HIV and HBV infections; normal blood routine test; not pregnant or lactating women. PBMC were isolated from the whole blood of volunteers by Ficoll density gradient centrifugation and cultured in RPMI1640 medium. 5.5.2 TCRm×CD3 bispecific antibodies mediate the killing of antigen peptide-pulsed T2 cells by PBMC Collect T2 cells in the logarithmic growth phase, resuspend them in growth medium at 1×10 6 cells / mL after centrifugation, and plate 1 mL per well in a 24-well plate. Peptide MAGE-A4 230-239 (GVYDGREHTV) was added to the 24-well plate at a final concentration of 50 μg / mL and pulsed overnight. After 16 hours, centrifuge and resuspend in RPMI1640 medium at 4×10 5 cells / mL, and plate 50 μL per well in a 96-well cell plate. Then add bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR (starting concentration of 30 nM, 4-fold serial dilution, 10 concentration points, 100 μL per well). Negative control sample 2DP47+IMCR, positive control samples 057G03 and IMC-C103C, and the use concentrations are the same as those of bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR. PBMC (effector cells) were resuspended at 4×10 6 cells / mL, added to the above cell plates, 50 μL per well, and the final effector-to-target ratio was 10:1. At the same time, set up separate target cell controls (antigen peptide-pulsed T2 cells), separate effector cell controls (PBMC), and separate medium blank controls, and fill the volume of each well to 200 μL with medium. After incubation for 24 hours, take the supernatant and refer to the instruction manual of non-radioactive cytotoxicity detection reagent ( Non-Radioactive Cytotoxicity Assay, promega, G1780) to detect and analyze the killing rate of PBMC cells mediated by TCRm×CD3 bispecific antibodies against antigen peptide-pulsed T2 cells. The results showed that TCRm×CD3 bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR could specifically mediate the killing of antigen peptide-pulsed T2 cells by PBMC cells; compared with 057G03 and IMC-C103C, the activities of 2N18D10+IMCR and 2N18D10-h16+IMCR bispecific antibodies in mediating the killing of antigen peptide-pulsed T2 cells by PBMC cells were higher (Table 8 and Figure 8). Table 8. TCRm×CD3 bispecific antibodies mediate the killing of antigen peptide-pulsed T2 cells by PBMC cells 5.5.3 TCRm×CD3 bispecific antibodies mediate PBMC killing of HLA-A02 + / MAGE-A4 + tumor cells Collect different antigen double-positive (HLA-A02+ / MAGE-A4 + ) The cells A375 cells (human malignant melanoma cells), Hs695T (human malignant melanoma cells), U2-OS (human osteosarcoma cells), and the negative control cells NIC-H520 (human lung squamous carcinoma cells, HLA-A02 - / MAGE-A4 + ) and HepG2 (human hepatocarcinoma cells, HLA-A02 + / MAGE-A4 - ). After digestion and centrifugation, the cells were resuspended in RPMI 1640 medium to 4×10 5 cells / mL, and 50 μL per well was plated in a cell plate. Then, the bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR (starting concentration 30 nM, 4-fold serial dilution, 10 concentration points, 100 μL per well) were added. The negative control sample 2DP47+IMCR, the positive control samples 057G03 and IMC-C103C were used at the same concentrations as the bispecific antibodies 2N18D10+IMCR and 2N18D10-h16+IMCR. PBMC (effector cells) were resuspended to 4×10 6 cells / mL, added to the above cell plate, 50 μL per well, and the final effector-to-target ratio was 10:1. At the same time, separate target cell controls, separate effector cell controls (PBMC), and separate medium blank controls were set up, and the volume of each well was made up to 200 μL with medium. After incubation for 24 - 48 hours, the supernatant was taken, and the killing rate of PBMC cells mediated by the TCRm×CD3 bispecific antibody against tumor cells was detected and analyzed with reference to the Non-Radioactive Cytotoxicity Assay reagent instruction manual ( Non-Radioactive Cytotoxicity Assay, promega, G1780). The results showed that 2N18D10-h16+IMCR mediated specific killing of tumor cells by PBMC (Figure 10), and the killing activity of 2N18D10+IMCR and 2N18D10-h16+IMCR against A375 cells was stronger than that of 057G03 and IMC-C103C (Figure 9, EC 50 values are shown in Table 9). Table 9. EC 50 values of PBMC cell killing of A375 cells mediated by the TCRm×CD3 bispecific antibody Example 6: Key amino acid analysis of TCRm×CD3 bispecific antibody recognition of MAGE-A4 230-239 (GVYDGREHTV) Chemical synthesis of MAGE-A4 230-239 (GVYDGREHTV) alanine-scanning peptide. Collect T2 cells in the logarithmic growth phase. After centrifugation, resuspend with growth medium to 1×10 6 cells / mL, and plate 1 mL per well in a 24-well plate. MAGE-A4 230- 239 (GVYDGREHTV) and alanine-scanning mutant peptides were added to the 24-well plate at a final concentration of 50 μg / mL and pulsed overnight. After 16 hours, centrifuge and resuspend with RPMI 1640 medium to 4×10 5 cells / mL, and add 50 μL per well to a 96-well cell plate. Collect Jurkat-Dual cells in the logarithmic growth phase. After centrifugation, resuspend with RPMI 1640 medium to 4×10 5 cells / mL, and add 50 μL per well to the previous cell plate to obtain a final E:T ratio of 1:1. Then add 2N18D10-h16+IMCR bispecific antibody (starting concentration of 30 nM, 4-fold serial dilution, 8 concentration points, 100 μL per well). Incubate for 24 h, take the supernatant, and refer to the QUANTI-Luc TM instruction manual (QUANTI-Luc TM , Invivogen, rep-qlc2) to detect and analyze the specific activation of Jurkat-Dual cells by T2 cells pulsed with different peptides. The results showed that 2N18D10-h16+IMCR could not mediate the activation of Jurkat-Dual cells by T2 cells pulsed with alanine-scanning mutant peptides at positions 3, 5, 6, 8, or 10 (Figure 11), that is, the amino acids at positions 3, 5, 6, 8, and 10 of MAGE-A4 230-239 (GVYDGREHTV) are the key amino acids for the binding of 2N18D10-h16+IMCR. Example 7: Analysis of the interaction between TCRm×CD3 bispecific antibody and similar peptides Chemically synthesize the similar peptides of MAGE-A4 230-239 (GVYDGREHTV) in Table 10. Table 10. Similar peptide information of MAGE-A4 230-239 (GVYDGREHTV) Collect T2 cells in the logarithmic growth phase. After centrifugation, resuspend with growth medium to 1×10 6 cells / mL, and plate 1 mL per well in a 24-well plate. MAGE-A4 230-239(GVYDGREHTV) and similar peptide segments were added to a 24-well plate at a final concentration of 50 μg / mL and pulsed overnight. After 16 hours, centrifugation was performed, and after centrifugation, the cells were resuspended in RPMI 1640 medium to 4×10 5 cells / mL, and 50 μL per well was added to a 96-well cell plate. Jurkat-Dual cells in the logarithmic growth phase were collected, centrifuged, and resuspended in RPMI 1640 medium to 4×10 5 cells / mL, and 50 μL per well was added to the above cell plate to obtain a final E:T ratio of 1:1. Then, the bispecific antibody 2N18D10-h16 + IMCR (starting initial concentration of 30 nM, 4-fold serial dilution, 8 concentration points, 100 μL per well) was added. Incubation was carried out for 24 h, the supernatant was taken, and the activation of Jurkat-Dual cells by T2 cells pulsed with different similar peptide segments was detected and analyzed with reference to the QUANTI-Luc TM instruction manual (QUANTI-Luc TM , Invivogen, rep-qlc2). The results in Figure 12 showed that 2N18D10-h16 + IMCR could not only specifically mediate the activation of Jurkat-Dual cells by T2 cells pulsed with MAGE-A4 230-239 (i.e., MAGE-A4-M1), but also mediate the activation of Jurkat-Dual cells by cells pulsed with the homologous sequence MAGE-A4-S1 (the antigen peptide of MAGE-A8 presented by HLA-A02); it did not mediate the activation of Jurkat-Dual cells by other T2 cells loaded with similar peptide segments. MAGE-A8 also belongs to the MAGE family and has the same expression pattern as MAGE-A4. It is restrictedly expressed in testicular and placental tissues in healthy individuals and specifically expressed in some tumor tissues in cancer patients. Example 8: Analysis of the interaction between TCRm×CD3 bispecific antibody and different antigen-negative tumor cells PBMC was sorted with reference to Section 5.5.1. Different tumor cells were collected, and the antigen information of the cells is shown in Table 11. With reference to Section 5.5.3, a killing experiment of 2N18D10-h16 + IMCR bispecific antibody-mediated PBMC against different antigen-negative tumor cells was carried out. The results in Figure 13 showed that 2N18D10-h16 + IMCR only mediated the cytotoxicity of PBMC against HLA-A02 + / MAGE-A4 + cell A375 and did not mediate the killing of antigen-negative tumor cells by PBMC. Table 11. Antigen expression information of different tumor cells Note: "+" represents positive, and "-" represents negative. Example 9: Tumor Suppression Effect of TCRm×CD3 Bispecific Antibody in hPBMC Immunoreconstituted Mice PBMC was sorted according to Section 5.5.1. Forty 5-6-week-old female NOG mice (purchased from Beijing Vitalstar Biotechnology Co., Ltd.) were selected, and each mouse was intraperitoneally inoculated with 8×10 6 personal PBMC. On the seventh day after inoculation with human PBMC, 1×10 7 A375 cells were subcutaneously inoculated on the right side of NPG mice. When the average tumor volume reached 70 mm 3 , the mice were randomly grouped according to tumor size. The grouping day was defined as day 0, and drug administration was carried out. The experiment was divided into a 1 mg / kg group and a 0.1 mg / kg group of the bispecific antibody 2N18D10-h16+IMCR, a negative control group of 2DP47+IMCR, and a positive control group of 057G03, with a total of 4 groups, 6 mice in each group. Administration was by tail vein injection, once a week for three times, and continuous observation was carried out. The therapeutic effect was evaluated according to the tumor growth inhibition rate TGI. During the experiment, the mental state of the animals was generally good. The test drug, the bispecific antibody 2N18D10-h16+IMCR, at a dose of 0.1 mg / kg (Group 3), had a significant effect on inhibiting tumor growth, with a relative tumor growth inhibition rate TGI (%) of 99.09; and the tumors of 3 animals completely regressed, showing a highly significant difference compared with the negative control group (Group 1) (p<0.001). 2N18D10-h16+IMCR at a dose of 1 mg / kg (Group 4) had a more significant effect on inhibiting tumor growth, with a relative tumor growth inhibition rate TGI (%) of 104.2%; and the tumors of 3 animals completely regressed, showing a highly significant difference compared with the negative control group (Group 1) (p<0.001). The bispecific antibody 2N18D10-h16+IMCR effectively inhibited tumor growth and was better than 057G03 (Group 2) (Figure 14). No animals died in each treatment group, and no obvious drug toxicity was shown. All patents, patent application publications, and non-patent literatures mentioned and / or listed in this application are hereby incorporated by reference in their entirety. The above describes the exemplary embodiments of the inventions of this application. However, without departing from the essence and scope of this application, those skilled in the art can modify or improve the exemplary embodiments described in this application, and the resulting variant or equivalent embodiments also fall within the scope of this application. Sequence Information References [1] Plaen, E.D., Arden, K., Traversarietal, C., et al. (1994) Structure, chromosomal localization, and expression of 12 genes of the MAGE family. Immunogenetics 40(5), 360–369. [2] Rogner, U.C., Wilke, K., Steck, E., et al. (1994) The melanoma antigen gene (MAGE) family is clustered in the chromosomal band Xq28. Genomics 29(3), 725–731. [3] Hao, Y.H., Doyle, J.M., Ramanathan, S., et al. (2013) Regulation of WASH-dependent actin polymerization and protein trafficking by ubiquitination. Cell 152, 1051-1064. [4]. Chomez, P., De, B.O., Bertrand, M., et al. (2001). An overview of the mage gene family with the identification of all human members of the family. Cancer Research 61(14), 5544-5551. [5] Simpson, A., Caballero, O.L., Jungbluth, A., et al. (2005) Cancer / testis antigens, gametogenesis and cancer. Nature Reviews Cancer 5(8), 615. [6]Sharma, P., Shen, Y., Wen, S., et al. (2006) Cancer-testis antigens: expression and correlation with survival in human urothelial carcinoma. Clinical Cancer Research an Official Journal of the American Association for Cancer Research 12(18), 5442-5447. [7]Bergeron, A., Valérie, P., Hélène, L., et al. (2010) High frequency of mage-a4 and mage-a9 expression in high-risk bladder cancer. International Journal of Cancer 125(6), 1365-1371. [8]Peikert, T., Specks, U., Farver, C., et al., (2006) Melanoma antigen A4 is expressed in non-small cell lung cancers and promotes apoptosis. Cancer Research 66(9), 4693–4700. [9]Yakirevich, E., Sabo, E., Lavie, O., et al. (2003) Expression of the MAGE-A4 and NY-ESO-1 cancer-testis antigens in serous ovarian neoplasms. Clinical Cancer Research an Official Journal of the American Association for Cancer Research 9(17), 6453-6460.

[0010] Quillien, V., Raoul, J. L., Heresbach, D., et al. (1997) Expression of MAGE genes in esophageal squamous-cell carcinoma. Anticancer Research 17(1a), 387–391.

[0011] Ries, J., Schultze-Mosgau, S., Neukam, F., et al. (2005) Investigation of the expression of melanoma antigen-encoding genes (MAGE-A1 to-A6) in oral squamous cell carcinomas to determine potential targets for gene-based cancer immunotherapy. International Journal of Oncology 26(3), 817–824.

[0012] M. T. Duffour, P. Chaux, C. Lurquin, G., et al. (1999) A MAGE-A4 peptide presented by HLA-A02 is recognized by cytolytic T lymphocytes. European Journal of Immunology 29(10), 3329–3337.

[0013] .Sanderson, J. P., Crowley, D. J., Wiedermann, G. E., et al. (2019) Preclinical evaluation of an affinity-enhanced MAGE-A4-specific T-cell receptor for adop-tive T-cell therapy. Oncoimmunology 9, 1682381.

[0014] Ottaviani, S., Colau, D., Bruggen, P. V. D., et al. (2006) A new mage-4 antigenic peptide recognized by cytolytict lymphocytes on HLA-A024 carcinoma cells. Cancer Immunology Immunotherapy 55(7), 867-872.

[0015] Sung, H., Ferlay, J., Siegel, R. L., et al. (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA: a cancer journal for clinicians 71(3), 209-249.

[0016] Kotsiou, E., Brzostek, J., Lenart, I., et al. (2010) Dimerization of soluble disulfide trap single-chain major histocompatibility complex class I molecules dependent on peptide binding affinity. Antioxid Redox Signal 15(3): 635-644.

[0017] E., Poignavent, V., Vincke, C., et al. (2018) Ritzenthaler C, Muyldermans S, Monsion B. Construction of High-Quality Camel Immune Antibody Libraries. Methods Mol Biol 1701: 169-187.

[0018] CN 201510097117.0

[0019] Phage Display: A General Experimental Guide / Edited by Clarkson, T. (USA) and Lowman, H. B. (USA); Translated by Ma Lan, et al. Chemical Industry Press, May 2008.

[0020] US 20160200833A1

[0021] WO 2017175006A1

[0022] Tan, P., Mitchell, D. A., Buss, T. N., et al. (2022) "Superhumanized" antibodies: reduction of immunogenic potential by complementarity-determining region grafting with human germline sequences: application to an anti-CD28. J Immunol 169(2):1119 - 1125.

[0023] US 5821337A

[0024] WO 2021122875A1

Claims

1. A bispecific antibody comprising a first antigen-binding fragment that binds to an HLA-A02 / MAGE-A4 complex and / or an HLA-A02 / MAGE-A8 complex and a second antigen-binding fragment that binds to an activating T cell antigen; Preferably, the binding epitope of the first antigen-binding fragment to the HLA-A02 / MAGE-A4 complex includes residues 230-239 of MAGE-A4 as shown in SEQ ID NO:48; more preferably, the binding epitope of the first antigen-binding fragment to the HLA-A02 / MAGE-A4 complex includes at least one of residues 232, 234, 235, 237 and 239 of MAGE-A4 as shown in SEQ ID NO:48; and / or The binding epitope of the first antigen-binding fragment to the HLA-A02 / MAGE-A8 complex includes positions 232-241 of MAGE-A8 shown in SEQ ID NO: 94; and / or Preferably, the activating T cell antigen is a CD3 molecule.

2. The bispecific antibody according to claim 1, wherein The first antigen-binding fragment comprises HCDR1 as shown in SEQ ID NO:32, HCDR2 as shown in SEQ ID NO:33 and HCDR3 as shown in SEQ ID NO:34; preferably, the first antigen-binding fragment is in the form of a single-domain antibody; more preferably, the first antigen-binding fragment comprises a single-domain antibody that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex in a monovalent or multivalent manner; and / or The second antigen-binding fragment comprises HCDR1 as shown in SEQ ID NO:37, HCDR2 as shown in SEQ ID NO:38, HCDR3 as shown in SEQ ID NO:39, LCDR1 as shown in SEQ ID NO:40, LCDR2 as shown in SEQ ID NO:41, and LCDR3 as shown in SEQ ID NO:42; preferably, the second antigen-binding fragment is a single-chain antibody (scFv) or a Fab fragment; more preferably, the second antigen-binding fragment is a scFv; in, The amino acid sequences of the HCDRs are according to the Kabat definition.

3. The bispecific antibody of claim 1 or 2, wherein the first antigen-binding fragment comprises a bivalent single domain antibody that binds to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex; Optionally, the first antigen-binding fragment comprises two heavy chain variable regions of monovalent single-domain antibodies that bind to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex; preferably, the first antigen-binding fragment comprises two heavy chain variable regions of monovalent single-domain antibodies that bind to the HLA-A02 / MAGE-A4 complex and / or the HLA-A02 / MAGE-A8 complex connected by a linker; more preferably, the linker is (GGGGS) n , wherein n is an integer ≥1, for example GGGGSGGGGSGGGGS (SEQ ID NO: 49).

4. The bispecific antibody according to any one of claims 1 to 3, wherein The first antigen-binding fragment comprises an amino acid sequence as shown in SEQ ID NO: 35, 36, 71 or 72; and / or The second antigen-binding fragment comprises a heavy chain variable region with an amino acid sequence as shown in SEQ ID NO:43 and a light chain variable region with an amino acid sequence as shown in SEQ ID NO:44; Preferably, the first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:35; and the second antigen-binding fragment comprises the heavy chain variable region with the amino acid sequence shown in SEQ ID NO:43 and the light chain variable region with the amino acid sequence shown in SEQ ID NO:44; The first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO: 36; and the second antigen-binding fragment comprises the heavy chain variable region with the amino acid sequence shown in SEQ ID NO: 43 and the light chain variable region with the amino acid sequence shown in SEQ ID NO: 44; or The first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO: 71; and the second antigen-binding fragment comprises the heavy chain variable region with the amino acid sequence shown in SEQ ID NO: 43 and the light chain variable region with the amino acid sequence shown in SEQ ID NO: 44; or The first antigen-binding fragment comprises the amino acid sequence shown in SEQ ID NO:72; and the second antigen-binding fragment comprises the heavy chain variable region shown in SEQ ID NO:43 and the light chain variable region shown in SEQ ID NO:

44.

5. The bispecific antibody according to any one of claims 1 to 4, wherein The first antigen-binding fragment and the second antigen-binding fragment are connected through an antibody heavy chain constant region Fc fragment, and the antibody heavy chain constant region Fc fragment includes a first Fc fragment and a second Fc fragment; preferably, the antibody heavy chain constant region Fc fragment is an Fc fragment of the IgG1 subtype; more preferably, the antibody heavy chain constant region Fc fragment is an Fc fragment of the IgG1m3 subtype. Fc fragment; wherein The amino acids at positions 354 and 366 of the first Fc fragment are C and W, respectively, or the amino acids at positions 349, 366, 368 and 407 of the first Fc fragment are C, S, A and V, respectively; and the amino acids at positions 354 and 366 of the second Fc fragment are C and W, respectively, or the amino acids at positions 349, 366, 368 and 407 of the second Fc fragment are C, S, A and V, respectively; preferably, the amino acids at positions 354 and 366 of the first Fc fragment are C and W, respectively, and the amino acids at positions 349, 366, 368 and 407 of the second Fc fragment are C, S, A and V, respectively; and / or The amino acids at positions 234, 235 and 331 of the first Fc fragment and the second Fc fragment are F, E and S, respectively; and / or One of the first Fc fragment and the second Fc fragment is linked to the first antigen-binding fragment, and the other of the first Fc fragment and the second Fc fragment is linked to the second antigen-binding fragment; in, The amino acid positions of antibody constant regions are determined according to EU numbering.

6. The bispecific antibody according to any one of claims 1 to 5, comprising a first arm that binds to an HLA-A02 / MAGE-A4 complex and / or an HLA-A02 / MAGE-A8 complex and a second arm that binds to an activating T cell antigen, wherein The first arm comprises the amino acid sequence shown in SEQ ID NO: 45 or 74; and The second arm comprises the amino acid sequence shown in SEQ ID NO:

46.

7. A single domain antibody that binds to an HLA-A02 / MAGE-A4 complex and / or an HLA-A02 / MAGE-A8 complex, comprising HCDR1 as shown in SEQ ID NO:32, HCDR2 as shown in SEQ ID NO:33, and HCDR3 as shown in SEQ ID NO:34; Preferably, the single-domain antibody comprises an amino acid sequence as shown in SEQ ID NO: 35, 36, 71 or 72; in, The amino acid sequences of the HCDRs are according to the Kabat definition.

8. A nucleic acid molecule encoding the bispecific antibody of any one of claims 1 to 6 or the single domain antibody of claim 7.

9. A pharmaceutical composition comprising the bispecific antibody according to any one of claims 1 to 6 or the single domain antibody according to claim 7 and a pharmaceutically acceptable excipient, diluent or carrier.

10. The pharmaceutical composition according to claim 9, which is used for preventing or treating HLA-A02 / MAGE-A4 positive tumors and / or HLA-A02 / MAGE-A8 positive tumors; preferably, the HLA-A02 / MAGE-A4 positive tumors and / or HLA-A02 / MAGE-A8 positive tumors are selected from: esophageal cancer, urothelial carcinoma, oral cancer, sarcoma, head and neck cancer, lung cancer, liver cancer, bladder cancer, melanoma, ovarian cancer, colorectal cancer and breast cancer.

11. Use of the bispecific antibody of any one of claims 1 to 6, the single domain antibody of claim 7, or the pharmaceutical composition of claim 9 or 10 in the preparation of a medicament for preventing or treating HLA-A02 / MAGE-A4-positive tumors and / or HLA-A02 / MAGE-A8-positive tumors; preferably, the HLA-A02 / MAGE-A4-positive tumors and / or HLA-A02 / MAGE-A8-positive tumors are selected from the group consisting of esophageal cancer, urothelial cancer, oral cancer, sarcoma, head and neck cancer, lung cancer, liver cancer, bladder cancer, melanoma, ovarian cancer, colorectal cancer and breast cancer.

12. A method for preventing or treating HLA-A02 / MAGE-A4-positive tumors and / or HLA-A02 / MAGE-A8-positive tumors, comprising administering to an individual in need thereof the bispecific antibody of any one of claims 1 to 6, the single domain antibody of claim 7, or the pharmaceutical composition of claim 9 or 10; preferably, the HLA-A02 / MAGE-A4-positive tumor and / or HLA-A02 / MAGE-A8-positive tumor is selected from the group consisting of esophageal cancer, urothelial carcinoma, oral cancer, sarcoma, head and neck cancer, lung cancer, liver cancer, bladder cancer, melanoma, ovarian cancer, colorectal cancer and breast cancer.

Images