Tetravalent bispecific antibodies targeting PD-1 and VEGF, their production method and use

Abstract

To provide a structurally symmetrical tetravalent bispecific antibody based on a common light chain and a construction method therefor, wherein the inventive bispecific antibody has similar or even better biological activity and physical and chemical properties than monoclonal antibodies, and can be applied for the treatment of various inflammatory diseases, cancer, and other diseases.SOLUTION: The present invention relates to a tetravalent bispecific antibody against PD-1 and VEGF, wherein the tetravalent bispecific antibody comprises two polypeptide chains and four common light chains, wherein the amino acid sequence of the polypeptide chain is shown as a specific amino acid sequence, and the amino acid sequence of the common light chain is shown as another specific amino acid sequence.SELECTED DRAWING: None

Description

[Technical field]

[0001] The present invention relates to the field of antibody engineering, and more specifically to a class of tetravalent bispecific antibodies, their preparation methods and uses. [Background technology]

[0002] Bispecific antibodies refer to antibody molecules that can specifically bind to two antigens or two binding epitopes simultaneously. Bispecific antibodies are classified into symmetric and asymmetric molecules based on their structural symmetry, and into bivalent, trivalent, tetravalent and multivalent molecules based on the number of binding sites.

[0003] A single chain antibody (hereinafter also referred to as "ScFv") is composed of a heavy chain variable region (hereinafter also referred to as "VH region") and a light chain variable region (hereinafter also referred to as "VL region"), of which the VH region and the VL region are linked together via a flexible peptide linker. In ScFv, the structural domains are arranged in the order of VH region-linker-VL region or VL region-linker-VH region, and various molecules have already been reported as linkers connecting the VH region and the VL region, such as a short alanine linker, a linker rich in glycine and serine, a linker exhibiting a helical structure, and a linker molecule derived from an immunoglobulin or a non-immunoglobulin can be used (Ahmad ZA, Yeap SK, Ali AM, et al., scFv antibody: principles and clinical application [J]. Clinical and developmental immunology, 2012: 980-250). ScFv refers to the smallest binding site of an antibody, and bispecific antibodies can be made by linking two ScFvs together using a single linker. Bispecific antibodies are bivalent, have one binding site per antigen, and generally have a molecular weight of about 50-60 kDa. When two ScFv fragments are linked in tandem, each ScFv fragment folds independently to form its own antigen-binding site. Tandem scFvs exhibit bispecificity and are currently already widely used in cancer immunotherapy, where T cells are used to retarget tumor cells and tumor-associated cells in the tumor microenvironment. This antibody morphology forms the molecular basis of bispecific T-cell engagers (BiTEs), and the first BiTE molecule to receive marketing approval was blinatumomab (Huehls AM, Coupet TA, Sentman C L. Bispecific T-cell engagers for cancer immunotherapy[J]. Immunology and cell biology, 2015, 93(3):290-296).

[0004] A diabody (Db) is a bivalent molecule formed by two chains, each of which contains one VH and one VL structural domain derived from the same or different antibodies. In a diabody, the two variable structural domains are linked via a short linker molecule, which usually consists of five amino acid residues, such as GGGGS. Since the linker molecule is obviously shorter than the length required for intrachain organization, the two ScFv chains dimerize along the head-to-tail direction to form a diabody molecule with a molecular weight equivalent to that of a tandem ScFv and tightly organized (Wu C.Diabodies:molecular engineering and therapeutic applications[J].Drug News Perspect,2009,22(8):453). When expressing the two different chains of a diabody molecule in a cell, there is a risk of mismatching of the variable structural domains, and since the two chains are not linked by non-covalent bonds, the diabody molecule still has stability problems (Kipriyanov SM, Moldenhauer G, Braunagel M, et al., Effect of domain order on the activity of bacterially produced bispecific single-chain Fv antibodies[J]. Journal of molecular biology, 2003, 330(1):99-111).

[0005] By fusing the diabody with human Fc, a molecule more similar to immunoglobulin can be formed, and the resulting antibody molecule is called a "Di-Diabody" (Lu D, Zhang H, Koo H, et al., A fully human recombinant IgG-like bispecific antibody to both the epidermal growth factor receptor and the insulin-like growth factor receptor for enhanced antitumor activity[J]. Journal of Biological Chemistry, 2005, 280(20):19665-19672). In addition, by fusing two tandemly linked ScFvs with human Fc, it is possible to form a molecule that is even more similar to immunoglobulin, and the resulting antibody molecule is called "TaFv-Fc" (Chen Z, Xie W, Acheampong DO, et al., A human IgG-like bispecific antibody co-targeting epidermal growth factor receptor and the vascular endothelial growth factor receptor 2 for enhanced antitumor activity[J]. Cancer biology & therapy, 2016, 17(2): 139-150).

[0006] Other than the above-mentioned Di-Diabody and TaFv-Fc, other types of tetravalent bispecific antibodies exhibiting structural symmetry include, but are not limited to, the following:

[0007] By attaching ScFv to the N-terminus or C-terminus of the heavy or light chain of natural IgG via a peptide linker, it is possible to form an IgG-ScFv type bispecific antibody (Coloma MJ, Morrison S L., Design and production of novel tetravalent bispecific antibodies[J]. Nature biotechnology, 1997, 15(2):159).

[0008] A DVD-Ig type bispecific antibody can be formed by attaching the VL and VH regions of another antibody via a peptide linker to the N-terminus of the light and heavy chains of a natural antibody (Wu C, Ying H, Grinnell C, et al., Simultaneous targeting of multiple disease mediators by a dual-variable-domain immunoglobulin[J]. Nature biotechnology, 2007, 25(11):1290).

[0009] CrossMAb is a technology based on the crossing of functional domains in the Fab arm of an antibody, and is a technology platform established by Roche. This technology solves the problem of accurate assembly of light and heavy chains derived from the same species, further increasing the success rate of constructing antibodies. Currently, several bispecific antibodies based on CrossMAb have already been reported (Klein C, Schaefer W, Regula J T., The use of CrossMAb technology for the generation of bi-and multispecific antibodies[C] / / MAbs. Taylor & Francis, 2016, 8(6):1010-1020).

[0010] A FIT-IgG type bispecific antibody can be formed by attaching the light chain of another antibody via a peptide linker to the N-terminus of the heavy chain of a natural antibody and associating VH+CH1 as an independent short chain with the light chain linked to the N-terminus (US10266608B2).

[0011] IgG-TCR is a molecule that has been developed by replacing the CH1 region of the heavy chain and the CL region of the light chain with the constant regions of the α and β chains of T cell receptors (TCR), respectively, to improve the association efficiency of the related heavy and light chains (Wu X, Sereno AJ, Huang F, et al., Protein design of IgG / TCR chimeras for the co-expression of Fab-like moieties within bispecific antibodies[C] / / MAbs. Taylor & Francis, 2015, 7(2):364-376). WuXiBody has been further improved based on IgG-TCR, and has further improved the association efficiency of the related heavy and light chains and the stability of the bispecific antibody molecule by providing an artificial disulfide bond between the α chain constant region and the β chain constant region (WO2019 / 057122A1).

[0012] Recently, a new method for producing tetravalent bispecific antibodies has been reported, in which the interface between the heavy and light chains of a Fab is artificially modified based on knowledge of the crystal structure, and an orthogonal Fab interface is experimentally selected, so that the heavy and light chains of the orthogonal Fab can specifically associate with each other and do not erroneously associate with the wild-type heavy and light chains (Lewis SM, Wu X, Pustilnik A, et al., Generation of bispecific IgG antibodies by structure-based design of an orthogonal Fab interface[J]. Nature biotechnology, 2014, 32(2):191; Wu X, Sereno AJ, Huang F, et al., Fab-based bispecific antibody formats with robust biophysical properties and biological activity[C] / / MAbs. Taylor & Francis, 2015, 7(3):470-482).

[0013] Furthermore, bispecific antibodies that are not structurally symmetric further include, but are not limited to, some of the structures shown below.

[0014] The knob-in-hole structure (hereinafter referred to as "KIH") mainly plays a role in promoting the heterodimerization of two different heavy chains of bispecific antibodies. The characteristic of the knob-in-hole structure is that a protruding "knob" is formed when a mutation occurs in the CH3 region of one heavy chain that constitutes the bispecific antibody, and a groove-like "knob" is formed when a mutation occurs in the CH3 region of the other heavy chain. By using the knob-in-hole structure, the heavy chains of two different types of antibodies can be accurately associated with each other (Merchant AM, Zhu Z, Yuan JQ, et al., An efficient route to human bispecific IgG[J]. Nature biotechnology, 1998, 16(7):677). DuetMab has used KIH technology to achieve heterodimerization of two heterologous heavy chains, and further improved the association efficiency of related heavy and light chains by introducing an artificial disulfide bond to replace the natural disulfide bond between CH1 and CL (Mazor Y, Oganesyan V, Yang C, et al., Improving target cell specificity using a novel monovalent bispecific IgG design[C] / / MAbs. Taylor & Francis, 2015, 7(2):377-389).

[0015] Another method for promoting heterodimerization of heterologous heavy chains is to use electrostatic steering mutations. This method is established based on charged residues that exhibit electrostatic interactions, and selectively changes the charge polarity of the CH3 interface, so that Fc structural domains with matching charge states form heterodimers through electrostatic interactions and suppress homodimerization through electrostatic repulsion (US10011858B2; Gunasekaran K, Pentony M, Shen M, et al., Enhancing antibody Fc heterodimer formation through electrostatic steering effects applications to bispecific molecules and monovalent IgG[J]. Journal of Biological Chemistry, 2010, 285(25):19637-19646).

[0016] Common light chains are also a commonly used technique for constructing bispecific antibodies. Common light chains can reduce the generation of by-products caused by mismatching of heavy and light chains and increase the yield of bispecific antibodies (Brinkmann U, Kontermann R E., The making of bispecific antibodies[C] / / MAbs. Taylor & Francis, 2017, 9(2):182-212). Common light chains are selected through experimental screening, typically using hybridoma or phage display technology. Typically, high-throughput screening of antibody libraries for antibody A and antibody B is required, and there are certain requirements for the capacity of the two antibody libraries (Sampei Z, Igawa T, Soeda T, et al., Identification and multidimensional optimization of an asymmetric bispecific IgG antibody mimicking the function of factor VIII cofactor activity[J]. PloS one, 2013, 8(2): e57479; US9657102B2).

[0017] Bispecific antibodies are becoming a kind of novel therapeutic antibodies, which can be applied to the treatment of various inflammatory diseases, cancer and other diseases. Recently, many novel configurations of bispecific antibodies have been reported, but obtaining the correct association of molecules is also the main technical difficulty in producing bispecific antibodies. All of the above-mentioned bispecific antibodies have the problem of mismatching, and the incorrect association will result in the formation of one or several by-products or aggregates, which will affect the yield, purity and physicochemical stability of the desired bispecific antibodies, and will also affect the safety and efficacy of the bispecific antibodies in the body. DISCLOSURE OF THEINVENTION [Problem to be solved by the invention]

[0018] The present invention has been made in view of the above-mentioned conventional problems, and provides a tetravalent bispecific antibody based on the structural symmetry of a common light chain and a method for producing the same. Usually, mismatching occurs often between the heavy chain and light chain of a bispecific antibody, but there are also cases where the heavy chains erroneously associate with each other, and the higher the efficiency of accurate association, the lower the possibility of mismatching. Therefore, in the present invention, the problem of mismatching between the heavy chain and the light chain is solved by using a common light chain, and one long heavy chain is constructed by linking the heavy chain variable region VH-B and CH1 of an antibody, attaching the heavy chain variable region VH-A of another antibody via a peptide linker, and further attaching the heavy chain constant region CH1-CH2-CH3. When the genes of this long heavy chain and the common light chain are expressed in the same cell, each long heavy chain associates with two common light chains, and at this time, the two light chains associated with the long heavy chain are the same, so there is no risk of mismatching occurring between the long heavy chain and the common light chain, and since the long heavy chains form homodimers rather than heterodimers, there is no risk of the long heavy chains erroneously associating with each other. The tetravalent bispecific antibody of the present invention does not require Fc modification, can be easily produced, and has biological activity and physicochemical properties comparable to or superior to those of monoclonal antibodies.

[0019] Furthermore, as a result of extensive research by the present inventors into the above-mentioned tetravalent bispecific antibodies, they have surprisingly found that the human PD-1 antibody mAb1-25-Hu (hereinafter also abbreviated as "609") binds to the PD-1 molecule mainly via its heavy chain, and does not depend much on its light chain. Based on this finding, a bispecific antibody capable of binding to PD-1 and other targets can be prepared by combining the heavy chain variable region and / or heavy chain of 609 with the heavy chain variable region and / or heavy chain and light chain variable region and / or light chain of an antibody that recognizes other targets (i.e., a common light chain), and this strategy is not limited to the tetravalent bispecific antibody of the present invention, but can also be applied to the preparation of other bispecific antibodies containing a common light chain known in the art.

[0020] Thus, in a first aspect of the invention there is provided a tetravalent bispecific antibody.

[0021] In a second aspect of the invention, there is provided an isolated nucleotide sequence encoding the tetravalent bispecific antibody.

[0022] In a third aspect of the present invention, there is provided an expression vector comprising the nucleotide sequence described above.

[0023] In a fourth aspect of the invention, there is provided a host cell comprising the expression vector.

[0024] In a fifth aspect of the present invention, there is provided a method for producing the tetravalent bispecific antibody.

[0025] In a sixth aspect of the invention, there is provided a pharmaceutical composition comprising the tetravalent bispecific antibody.

[0026] In a seventh aspect of the invention, there is provided the use of said tetravalent bispecific antibody or said pharmaceutical composition in the manufacture of a medicament for the treatment of cancer, inflammatory diseases, autoimmune diseases and other conditions.

[0027] In an eighth aspect of the invention, a method for producing a tetravalent bispecific antibody is provided.

[0028] A ninth aspect of the present invention provides a bispecific antibody, the heavy chain variable region of which has an amino acid sequence set forth in SEQ ID NO:83.

[0029] A tenth aspect of the present invention provides the use of a heavy chain variable region whose amino acid sequence is set forth in SEQ ID NO:83 in the production of a bispecific antibody. [Means for solving the problem]

[0030] In order to achieve the above object, the present invention comprises the following technical solutions.

[0031] In a first aspect of the present invention there is provided a tetravalent bispecific antibody, the tetravalent bispecific antibody comprising: Two polypeptide chains comprising VH-B-CH1-peptide linker-VH-A-CH1-CH2-CH3 from the N-terminus to the C-terminus, and four common light chains comprising VL-CL from the N-terminus to the C-terminus, wherein the VH-A is a heavy chain variable region of a first antibody, the VH-B is a heavy chain variable region of a second antibody, the CH1 is a first structural domain of a heavy chain constant region, the CH2 is a second structural domain of a heavy chain constant region, and the CH3 is a third structural domain of a heavy chain constant region; the first antibody specifically binds to a first antigen, and the second antibody specifically binds to a second antigen; The VL is a light chain variable region, the CL is a light chain constant region, the VH-A-CH1 of the polypeptide chain and the VH-B-CH1 of the polypeptide chain each associate with the VL-CL of the common light chain, the VH-A and the VL form a first antigen-binding site, and the VH-B and the VL form a second antigen-binding site.

[0032] The structure of the bispecific antibody according to the present invention is as shown in FIG.

[0033] According to the present invention, the common light chain is obtained by screening using the following method: By exchanging the heavy chain and the light chain of the first antibody and the light chain of the second antibody, respectively, a hybrid antibody formed by a combination of the heavy chain of the first antibody and the light chain of the second antibody, and a combination of the heavy chain of the second antibody and the light chain of the first antibody is obtained, and at this time, a) if the hybrid antibody formed by combining the heavy chain of the first antibody and the light chain of the second antibody can specifically bind to the first antigen, the light chain of the second antibody is used as a common light chain; b) if the hybrid antibody formed by combining the heavy chain of the second antibody with the light chain of the first antibody can specifically bind to the second antigen, the light chain of the first antibody is used as the common light chain; c) if the hybrid antibody formed by combining the heavy chain of the first antibody with the light chain of the second antibody cannot specifically bind to the first antigen and the hybrid antibody formed by combining the heavy chain of the second antibody with the light chain of the first antibody cannot specifically bind to the second antigen, introducing a back mutation into the light chain of the first antibody so that the hybrid antibody formed by combining the heavy chain of the second antibody with the mutated light chain of the first antibody can specifically bind to the second antigen, and making the mutated light chain of the first antibody the common light chain; or d) if the hybrid antibody formed by combining the heavy chain of the first antibody with the light chain of the second antibody cannot specifically bind to the first antigen and the hybrid antibody formed by combining the heavy chain of the second antibody with the light chain of the first antibody cannot specifically bind to the second antigen, then a back mutation is introduced into the light chain of the second antibody so that the hybrid antibody formed by combining the heavy chain of the first antibody with the mutated light chain of the second antibody can specifically bind to the first antigen, and the mutated light chain of the second antibody is made the common light chain.

[0034] According to a preferred embodiment of the present invention, when the hybrid antibody can specifically bind to the antigen in the above a) and b), the affinity of the hybrid antibody binding to the antigen can be further increased by introducing back mutations into the light chain in the same manner as described above. Therefore, the common light chain can be obtained by further screening using the following method.

[0035] By exchanging the heavy chain and the light chain of the first antibody and the light chain of the second antibody, respectively, a hybrid antibody formed by a combination of the heavy chain of the first antibody and the light chain of the second antibody, and a combination of the heavy chain of the second antibody and the light chain of the first antibody is obtained, and at this time, a') when a hybrid antibody formed by combining the heavy chain of the first antibody and the light chain of the second antibody can specifically bind to the first antigen, introducing a back mutation into the light chain of the second antibody so that the hybrid antibody combining the heavy chain of the first antibody and the mutated light chain of the second antibody has a higher affinity for the first antigen than the hybrid antibody combining the heavy chain of the first antibody and the light chain of the second antibody, and using the mutated light chain of the second antibody as a common light chain; b') If a hybrid antibody formed by combining the heavy chain of the second antibody with the light chain of the first antibody can specifically bind to a second antigen, then by introducing a back mutation into the light chain of the first antibody, the hybrid antibody combining the heavy chain of the second antibody with the mutated light chain of the first antibody has a higher affinity for the second antigen than the hybrid antibody combining the heavy chain of the second antibody with the light chain of the first antibody, and the mutated light chain of the first antibody serves as the common light chain.

[0036] According to the present invention, the peptide linker is an artificial linker. Preferably, the artificial linker is any one selected from the group consisting of G, GS, SG, GGS, GSG, SGG, GGG, GGGS, SGGG, GGGGS, GGGGSGS, GGGGSGGS, GGGGSGGGGS, GGGGSGGGGSGGGGS (SEQ ID NO: 150), AKTTPKLEEGEFSEAR, AKTTPKLEEGEFSEARV, AKTTPKLGG, SAKTTPKLGG, SAKTTP, RADAAP, RADAAPTVS, RADAAAAAGGPGS, SAKTTPKLEEGEFSEARV, ADAAP, ADAATVSIFPP, TVAAP, TVAAPSVFIFPP, QPKAAP, QPKAAPSVTLFPP, AKTTPP, AKTTPPSVTPLAP, AKTTAPSVYPLAP, ASTKGP, ASTKGPSVFPLAP, GENKVEYAPALMALS, GPAKELTPLKEAKVS and GHEAAAVMQVQYPAS. More preferably, the artificial linker is GGGGSGGGSGGGGS (SEQ ID NO: 150).

[0037] According to the present invention, the CH1 structural domain and CH1-CH2-CH3 near the N-terminus in the polypeptide chain may be derived from heavy chain constant regions of the same or different subtypes, for example, derived from a heavy chain constant region selected from the group consisting of an IgG1 heavy chain constant region, an IgG2 heavy chain constant region, an IgG3 heavy chain constant region, and an IgG4 heavy chain constant region, preferably derived from an IgG1 or IgG4 heavy chain constant region, and preferably containing S228P (EU numbering) as a site mutation in the IgG4.

[0038] According to the invention, said CL is the constant region of a κ light chain or the constant region of a λ light chain.

[0039] In a preferred embodiment of the present invention, the first antigen and the second antigen are any one selected from the group consisting of VEGF / PD-1, PD-1 / VEGF, TGF-β / PD-1, PD-1 / TGF-β, HER2 / CD47, CD47 / HER2, HER2 / CD137, CD137 / HER2, PD-1 / CD137, CD137 / PD-1, PD-1 / CD40, CD40 / PD-1, PD-1 / EGFR, EGFR / PD-1, PD-1 / HER2, HER2 / PD-1, PD-1 / CTLA-4, CTLA-4 / PD-1, PD-1 / LAG-3, and LAG-3 / PD-1.

[0040] In a preferred embodiment of the present invention, the VH-A or the VH-B has a sequence represented by SEQ ID NO:83.

[0041] In a preferred embodiment of the present invention, the sequences of the VH-A, the VH-B and the VL are SEQ ID NO:1, 11, 15, SEQ ID NO:11, 1, 15, SEQ ID NO:20, 11, 15, SEQ ID NO:11, 20, 15, SEQ ID NO:29, 41, 42, SEQ ID NO:41, 29, 42, SEQ ID NO:31, 41, 42, SEQ ID NO:41, 31, 42, SEQ ID NO:53, 61, 54, SEQ ID NO:61, 53, 54, SEQ ID NO:53, 77, 54, SEQ ID NO:77, 53 , 54, SEQ ID NO:83, 91, 92, SEQ ID NO:91, 83, 92, SEQ ID NO:83, 105, 106, SEQ ID NO:105, 83, 106, SEQ ID NO:83, 113, 114, SEQ ID NO:113, 83, 114, SEQ ID NO:83, 117, 118, SEQ ID NO:117, 83, 118, SEQ ID NO:83, 119, 120, SEQ ID NO:119, 83, 120, SEQ ID NO:83, 121, 122, and SEQ ID NO:121, 83, 122.

[0042] In a preferred embodiment of the invention, the sequences of the polypeptide chains and the common light chain are selected from the group consisting of SEQ ID NO:16, 13, SEQ ID NO:18, 13, SEQ ID NO:21, 13, SEQ ID NO:45, 43, SEQ ID NO:47, 43, SEQ ID NO:49, 43, SEQ ID NO:51, 43, SEQ ID NO:65, 63, SEQ ID NO:67, 63, SEQ ID NO:79, 63, SEQ ID NO:81, 63, SEQ ID NO:95, 93, SEQ ID NO:97, 93, SEQ ID NO:109, 107, SEQ ID NO:111, 107, SEQ ID NO:131, 123, SEQ ID NO:133, 125, SEQ ID NO:135, 127, SEQ ID NO:137, 127, SEQ ID NO:139, 127, SEQ ID NO:141, 127, SEQ ID NO:143, 129, and SEQ ID NO:145, 129.

[0043] In a second aspect of the invention, there is provided an isolated nucleotide sequence encoding the tetravalent bispecific antibody.

[0044] In a third aspect of the present invention, there is provided an expression vector comprising the nucleotide sequence described above.

[0045] In a preferred embodiment of the invention, said nucleotides encode said polypeptide chain and said common light chain, and said nucleotides have a sequence selected from the group consisting of SEQ ID NO:17, 14, SEQ ID NO:19, 14, SEQ ID NO:22, 14, SEQ ID NO:46, 44, SEQ ID NO:48, 44, SEQ ID NO:50, 44, SEQ ID NO:52, 44, SEQ ID NO:66, 64, SEQ ID NO:68, 64, SEQ ID NO:80, 64, SEQ ID NO:82, 64, SEQ ID NO:96, 94, SEQ ID NO:98, 94, SEQ ID NO:110, 108, SEQ ID NO:112, 108, SEQ ID NO:132, 124, SEQ ID NO:134, 126, SEQ ID NO:136, 128, SEQ ID NO:138, 128, SEQ ID NO:140, 128, SEQ ID NO:142, 128, SEQ ID NO:144, 130, and SEQ ID NO:146, 130.

[0046] In a fourth aspect of the invention, there is provided a host cell comprising the expression vector.

[0047] In a fifth aspect of the present invention, there is provided a method for producing the tetravalent bispecific antibody, the method comprising: a step a) expressing the tetravalent bispecific antibody by culturing the host cell under expression conditions; and a step b) isolating and purifying the tetravalent bispecific antibody expressed in step a.

[0048] In a sixth aspect of the invention, there is provided a pharmaceutical composition comprising the above-mentioned tetravalent bispecific antibody and a pharma- ceutically acceptable carrier.

[0049] In a seventh aspect of the present invention, there is provided a use of the tetravalent bispecific antibody or the pharmaceutical composition in the manufacture of a medicament for treating cancer and inflammatory diseases, autoimmune diseases and other diseases. Furthermore, the present invention provides a method for treating cancer and inflammatory diseases, autoimmune diseases and other diseases by administering the tetravalent bispecific antibody or the pharmaceutical composition to a subject in need of treatment. The cancers include, but are not limited to, melanoma (e.g., metastatic malignant melanoma), kidney cancer (e.g., clear cell renal cell carcinoma), prostate cancer (e.g., hormone refractory prostate cancer), pancreatic cancer, breast cancer, colon cancer, lung cancer (e.g., non-small cell lung cancer), esophageal cancer, head and neck squamous cell carcinoma, liver cancer, ovarian cancer, cervical cancer, thyroid cancer, glioblastoma, glioma, leukemia, lymphoma and other neoplastic malignant diseases. The inflammatory disease, autoimmune disease and other diseases include, but are not limited to, ophthalmology, fibrosis, asthma, rheumatoid arthritis, systemic lupus erythematosus, multiple sclerosis, psoriasis, atopic dermatitis, etc. There is no limitation on the subject, but the subject is generally a human.

[0050] In an eighth aspect of the present invention, there is provided a method for producing a tetravalent bispecific antibody, the method comprising the steps of: Step I links a heavy chain variable region VH-B of a second antibody that specifically binds to a second antigen to a first structural domain CH1 of a heavy chain constant region; Step II, in which the conjugate obtained in step I is linked via a peptide linker to a heavy chain variable region VH-A of a first antibody that specifically binds to a first antigen; Step III, in which the conjugate obtained in step II is linked to the heavy chain constant region CH1-CH2-CH3 to form a polypeptide chain; and The method includes step IV of introducing the polypeptide chains and the common light chain VL-CL obtained in step III into expression vectors, and combining and expressing the obtained expression vectors to obtain the desired tetravalent bispecific antibody.

[0051] Among them, the CH1 is the first structural domain of the heavy chain constant region, the CH2 is the second structural domain of the heavy chain constant region, the CH3 is the third structural domain of the heavy chain constant region, the VL is a light chain variable region, the CL is a light chain constant region, the VH-A-CH1 and the VH-B-CH1 associate with the VL-CL, respectively, the VH-A forms a first antigen-binding site with the VL, and the VH-B forms a second antigen-binding site with the VL.

[0052] In a ninth aspect of the present invention, there is provided a bispecific antibody, the bispecific antibody comprising at least two different heavy chain variable regions and at least two common light chains, the common light chains comprising the same light chain variable region, the heavy chain variable regions and the light chain variable regions forming an antigen-binding site, and the heavy chain variable region comprising the amino acid sequence represented by SEQ ID NO:83.

[0053] A tenth aspect of the present invention provides a use of a heavy chain variable region comprising the amino acid sequence represented by SEQ ID NO: 83 in producing a bispecific antibody, the bispecific antibody comprising at least two different heavy chain variable regions and at least two common light chains, the common light chains comprising the same light chain variable region, and the heavy chain variable regions and the light chain variable regions forming an antigen-binding site. Effect of the Invention

[0054] The present invention provides a tetravalent bispecific antibody based on a common light chain and having a symmetric structure, and a method for producing the same. The bispecific antibody according to the present invention has biological activity and physicochemical properties comparable to or superior to those of monoclonal antibodies, and can be applied to the treatment of various inflammatory diseases, cancers, and other diseases. [Brief description of the drawings]

[0055] [Figure 1] FIG. 1 is a schematic diagram of the structure of a bispecific antibody of the present invention, in which VH-A represents the heavy chain variable region of a first antibody, VH-B represents the heavy chain variable region of a second antibody, VL represents the light chain variable region of a common light chain, CH1, CH2, and CH3 are three structural domains in the heavy chain constant region, and CL is the light chain constant region of the common light chain. The disulfide bonds between the heavy chains and the disulfide bonds between the heavy and light chains are each represented by a line segment, the artificial linker between the CH1 region and the VH-A region adjacent to the N-terminus of the polypeptide chain is represented by a line segment, and the line segment between the CH1 region and the CH2 region adjacent to the C-terminus of the polypeptide chain represents the original linker and hinge region of the antibody (when the heavy chain is a human IgG4 subtype, the hinge region contains S228P as a site mutation). [Diagram 2] 2A and 2B show the ELISA results for 601, 20-Hu, and their hybrid antibodies. [Diagram 3] Figures 3A and 3B show the ELISA results for 20-Fab-601-IgG4-V94L and 601-Fab-20-IgG4-V94L. [Figure 4] Figures 4A and 4B are HPLC-SEC spectra of 20-Fab-601-IgG4-V94L. [Diagram 5] Figures 5A and 5B are HPLC-IEC spectra of 20-Fab-601-IgG4-V94L. [Figure 6] Figures 6A to 6D are CE-SDS spectra of 20-Fab-601-IgG4-V94L. [Figure 7]Figures 7A and 7B are DSC spectra of 20-Fab-601-IgG4-V94L. [Figure 8] FIG. 1 shows the results of measuring the neutralizing ability of 20-Fab-0313-IgG4-V94L against VEGF bioactivity. [Figure 9] Figures 9A and 9B show the enhancing effect of 20-Fab-0313-IgG4-V94L on MLR activity. [Figure 10] FIG. 1 shows the binding properties of 20-Fab-0313-IgG4-V94L when binding to both PD-1 and VEGF165 simultaneously. [Figure 11] FIG. 1 shows the pharmacokinetic properties of 20-Fab-0313-IgG4-V94L. [Figure 12] FIG. 1 shows the in vivo antitumor activity of a bispecific antibody targeting PD-1 and VEGF in mice. [Figure 13] 13A and 13B show the ELISA results for 1D11-Hu, 14-Hu, and their hybrid antibodies. [Figure 14] FIG. 1 shows the ELISA results of mAb127, 14-Hu and their hybrid antibodies. [Figure 15] 15A and 15B show the ELISA results for 14-Fab-1D11-IgG4, 1D11-Fab-14-IgG4, 14-Fab-127-IgG4, and 127-Fab-14-IgG4. [Figure 16] 16A and 16B show the enhancing effect of 14-Fab-127-IgG4 on MLR activity. [Figure 17] FIG. 1 shows the binding properties of 14-Fab-127-IgG4 when binding to both TGF-β1 and PD-1 simultaneously. [Figure 18] FIG. 1 shows the in vivo antitumor activity of a bispecific antibody targeting PD-1 and TGF-beta in mice. [Figure 19]19A and 19B show the ELISA results for 19H6-Hu, Anti-CD47B-Hu, and their hybrid antibodies. [Figure 20] 20A and 20B show the ELISA results of CD47B-Fab-19H6-IgG1 and 19H6-Fab-CD47B-IgG1. [Figure 21] 21A and 21B show the ELISA results for 19H6-Hu, 94-Hu, and their hybrid antibodies. [Figure 22] 22A and 22B show the ELISA results for 94-Fab-19H6-IgG1-LALA and 19H6-Fab-94-IgG1-LALA. [Figure 23] 23A and 23B show the ELISA results for 609, Anti-CD137-Hu, and their hybrid antibodies. [Figure 24] 24A and 24B show the ELISA results of 609-Fab-137-IgG4 and 137-Fab-609-IgG4. [Diagram 25] 25A and 25B show the ELISA results for 609, Anti-CD40-Hu, and their hybrid antibodies. [Figure 26] Figures 26A and 26B show the ELISA results of 609-Fab-40-IgG4 and 40-Fab-609-IgG4. [Figure 27] FIG. 1 shows ELISA results for 609-HC+Cetuximab-LC, 609-HC+Bevacizumab-LC, 609-HC+Trastuzumab-LC, 609-HC+Pertuzumab-LC, 609-HC+Ipilimumab-LC and 609-HC+5E7-Hu-LC. [Figure 28]Figure 1 shows the inhibitory effects of 609-HC+Cetuximab-LC, 609-HC+Bevacizumab-LC, 609-HC+Trastuzumab-LC, 609-HC+Pertuzumab-LC, 609-HC+Ipilimumab-LC and 609-HC+5E7-Hu-LC on the interaction between PD-1 and PD-L1. [Figure 29] FIG. 1 shows the enhancing effect of 609-HC+Cetuximab-LC, 609-HC+Bevacizumab-LC, 609-HC+Trastuzumab-LC, 609-HC+Pertuzumab-LC, 609-HC+Ipilimumab-LC and 609-HC+5E7-Hu-LC on mixed lymphocyte reaction. [Diagram 30] FIG. 1 shows the results of measuring the binding characteristics of 609, 609-HC+Cetuximab-LC, 609-HC+Bevacizumab-LC, 609-HC+Pertuzumab-LC, 609-HC+Ipilimumab-LC and 609-HC+5E7-Hu-LC to cell surface PD-1 using flow cytometry. [Diagram 31] FIG. 1 shows the results of alanine scanning of the light chain variable region of 609. [Diagram 32] Figures 32A and 32B show the ELISA results of 609-Fab-Cetuximab-IgG4. [Diagram 33] Figures 33A and 33B show the ELISA results of 609-Fab-Pertuzumab-IgG4. [Diagram 34] Figures 34A and 34B show the ELISA results for 609-Fab-Ipilimumab-IgG1, Ipilimumab-Fab-609-IgG1, 609-Fab-Ipilimumab-IgG4, and Ipilimumab-Fab-609-IgG4. [Diagram 35] FIG. 1 shows the results when measuring the functional activity of a bispecific antibody targeting PD-1 and CTLA-4. [Diagram 36]FIG. 1 shows the results when measuring the ADCC activity of a bispecific antibody targeting PD-1 and CTLA-4. [Figure 37] FIG. 1 shows the pharmacokinetic properties of a bispecific antibody targeting PD-1 and CTLA-4 in rats. [Figure 38] FIG. 1 shows the in vivo antitumor activity of a bispecific antibody targeting PD-1 and CTLA-4 in mice. [Figure 39] Figures 39A and 39B show the ELISA results of 609-Fab-5E7-IgG4 and 5E7-Fab-609-IgG4. [Diagram 40] FIG. 1 shows the binding characteristics of a bispecific antibody targeting PD-1 and LAG-3 when it simultaneously binds to two different antigens. [Diagram 41] FIG. 1 shows the results when measuring the functional activity of a bispecific antibody targeting PD-1 and LAG-3. [Diagram 42] FIG. 1 shows the in vivo antitumor activity of a bispecific antibody targeting PD-1 and LAG-3 in mice. [Diagram 43] 43A to 43E show the results of verifying the specificity of the hybrid antibody. [Diagram 44] Figures 44A to 44H are HPLC-SEC spectra of 609-Fab-Cetuximab-IgG4, 609-Fab-Pertuzumab-IgG4, 609-Fab-Ipilimumab-IgG1, Ipilimumab-Fab-609-IgG1, 609-Fab-Ipilimumab-IgG4, Ipilimumab-Fab-609-IgG4, 609-Fab-5E7-IgG4, and 5E7-Fab-609-IgG4. [Diagram 45] Figures 45A and 45B are HPLC-IEC spectra of 609-Fab-Ipilimumab-IgG1 and 609-Fab-5E7-IgG4. [Diagram 46]Figures 46A to 46D show NR-CE-SDS and R-CE-SDS spectra of 609-Fab-Ipilimumab-IgG1 and 609-Fab-5E7-IgG4. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0056] In the present invention, "antibody (Ab)" and "immunoglobulin G (IgG)" are heterotetrameric glycoproteins of about 150,000 daltons with the same structural characteristics, which are composed of two identical light chains (L chains) and two identical heavy chains (H chains). Each light chain is linked to a heavy chain via one covalent disulfide bond, and the number of disulfide bonds between heavy chain isotypes in different immunoglobulins varies. The heavy and light chains each have intrachain disulfide bonds at regular intervals, and each heavy chain has a variable region (VH region) at one end, followed by a constant region consisting of three structural domains CH1, CH2 and CH3. Each light chain has a variable region (VL region) at one end, and a constant region consisting of one CL structural domain at the other end, where the constant region of the light chain associates with the first structural domain CH1 in the constant region of the heavy chain, and the variable region of the light chain associates with the variable region of the heavy chain. The constant region is not directly involved in the binding of an antibody to an antigen, but exhibits other biological functions, such as antibody-dependent cell-mediated cytotoxicity (hereinafter also abbreviated as "ADCC"). The constant region of the heavy chain includes subtypes of IgG1, IgG2, IgG3, and IgG4, and the constant region of the light chain includes subtypes of κ (Kappa) or λ (Lambda) constant region. The heavy and light chains of an antibody are linked together via a covalent disulfide bond between the CH1 structural domain of the heavy chain and the CL structural domain of the light chain, and the two heavy chains are linked together via an interchain disulfide bond formed between the hinge regions.

[0057] In the present invention, the term "bispecific antibody" refers to an antibody molecule that can simultaneously recognize and specifically bind to two antigens (targets) or two epitopes.

[0058] In the present invention, the term "monoclonal antibody" refers to an antibody obtained from a nearly homogeneous group of one type, that is, the single antibody contained in the group is the same, except for the possibility of natural mutation in a very small portion. A monoclonal antibody recognizes and binds to one antigen site with high specificity. Unlike ordinary polyclonal antibody preparations (these polyclonal antibody preparations are usually mixtures containing multiple antibodies that can recognize different determinants), a monoclonal antibody recognizes and binds to a single determinant on an antigen. In addition to the above-mentioned specificity, a monoclonal antibody has the advantage that it is obtained by culturing a hybridoma and is not contaminated by other immunoglobulins. In addition, in the present invention, "monoclonal" refers to the characteristics of an antibody, and should not be interpreted as being obtained from a nearly homogeneous antibody group and produced by any special method.

[0059] In the present invention, "Fab" and "Fc" refer to two completely identical Fab fragments and one Fc fragment obtained by digesting an antibody with papain. The Fab fragment is composed of the VH region and CH1 structural domain of the antibody heavy chain, and the VL region and CL structural domain of the light chain. The Fc fragment is also called fragment crystallizable (Fc), and is composed of the CH2 structural domain and CH3 structural domain of the antibody. The Fc fragment does not exhibit binding activity to an antigen, and is the site where the antibody interacts with its effector molecule or cell.

[0060] In the present invention, "variable" refers to a structure in which there is a sequence difference in a part of the antibody variable region, and which determines the binding property and specificity of the antibody to a specific antigen. In addition, the sequence that determines the degree of variability is not uniformly distributed throughout the variable region of the antibody, but is concentrated in the complementarity-determining region (CDR) in the variable regions of the light chain and heavy chain, or in three fragments in the hypervariable region. The relatively highly conserved part of the variable region is the framework region (FR), and each of the variable regions of natural heavy chains and light chains contains four FR regions. These FR regions generally form a β-fold structure and are connected to each other by three CDRs that form a connecting ring, but in some cases they can also form a partial β-fold structure. The CDR regions in each chain are closely adjacent to each other via the FR regions, and together with the CDR regions of the other chain, form the antigen-binding site of the antibody [Kabat et al., NIH Publ. No. 91-3242, Vol. I, pp. 647-669 (1991)].

[0061] In the present invention, the term "mouse-derived antibody" refers to an antibody derived from a rat or mouse, with mouse-derived antibodies being particularly preferred.

[0062] In the present invention, the term "humanized antibody" refers to an antibody whose CDR region is derived from an antibody of an animal other than human (e.g., mouse) and whose framework region and constant region are derived from a human antibody. From the viewpoint of maintaining binding affinity, it is also possible to modify residues in the framework region.

[0063] In the present invention, "specific binding" and "binding" refer to a non-random binding reaction between two molecules, for example, a specific reaction between an antibody and the antigen to which it binds. Antibodies are typically found in concentrations of 1×10 -7 M, e.g., about 1×10 -8 Less than M, more preferably less than 1×10 -9 Less than M, more preferably less than 1×10 -10 Less than M, 1×10 -11The antibody binds to the antigen with an equilibrium constant of less than M or less. In the present invention, "KD" is the equilibrium dissociation constant when a specific antibody interacts with an antigen, and represents the binding affinity between the antibody and the antigen. The smaller the equilibrium dissociation constant, the stronger the binding between the antibody and the antigen, i.e., the higher the binding affinity between the antibody and the antigen. The binding affinity between the antibody and the antigen can be measured, for example, using a Biacore device by using Surface Plasmon Resonance (SPR), or the relative affinity when the antibody and the antigen bind can be measured by using an ELISA method.

[0064] In the present invention, "(antibody) valency" is a term that refers to the number of antigen-binding sites present in an antibody molecule. The bispecific antibody of the present invention preferably has four antigen-binding sites and is tetravalent. In the present invention, the antigen-binding sites include a heavy chain variable region (VH region) and a light chain variable region (VL region).

[0065] In the present invention, the term "binding epitope" refers to a peptide determinant to which an antibody specifically binds, and is a region within an antigen capable of specifically binding to an antibody.

[0066] In the present invention, the term "peptide linker" refers to a peptide consisting of an amino acid sequence. The peptide linker referred to in the present invention may be a natural peptide linker or an artificial peptide linker, and is preferably an artificial peptide linker. Examples of peptide linkers of the present invention include G, GS, SG, GGS, GSG, SGG, GGG, GGGS, SGGG, GGGGS, GGGGSGS, GGGGSGGS, GGGGSGGGGS, GGGSGGGGSGGGGGS (sequence number 150), AKTTPKLEEGEFSEAR, AKTTPKLEEGEFSEARV, AKTTPKLGG, SAKTTPKLGG, SAKTTP, RADAAP, RADAAPTVS, RADAAAAAGGPGS, SAKTTPKLEEGEFSEARV, ADAAP, ADAAPTVSIFPP, TVAAP, TVAAPSVFIFPP, QPKAAP, QPKAAPSVTLFPP, AKTTPP, AKTTPPSVTPLAP, AKTTAPSVYPLAP, ASTKGP, ASTKGPSVFPLAP, GENKVEYAPALMALS, GPAKELTPLKEAKVS, or GHEAAAVMQVQYPAS. The peptide linker referred to in the present invention may be a peptide linker that is biodegradable, an enzyme (e.g., MMP) sensitive linker, a linker having a reductively degradable disulfide bond, etc., and as such a peptide linker, reference may be made to "Fusion Protein Technologies for Biopharmaceuticals: Applications and Challenges" edited by Stefan R. Schmidt, or any degradable linker commonly used in the art may be used. These degradable linkers can be applied to the dissociation of the Fab at the head of the antibody molecule in the body, and can enhance the penetration and distribution in body tissues and the binding to the target, while simultaneously reducing potential side effects, and can improve the in vivo function and half-life of the two different Fab regions. Most preferably, the artificial linker of the present invention is GGGGSGGGSGGGGS (SEQ ID NO: 150).

[0067] In the present invention, a "common light chain" refers to a light chain that contains the same light chain variable region and light chain constant region, and can form a binding site that specifically binds to a first antigen by associating with a heavy chain of a first antibody that binds to a first antigen, and can form a binding site that specifically binds to a second antigen by associating with a heavy chain of a second antibody that binds to a second antigen. Furthermore, the light chain variable region of the common light chain and the heavy chain variable region of the first antibody form a first antigen-binding site, and the light chain variable region of the common light chain and the heavy chain variable region of the second antibody form a second antigen-binding site.

[0068] In the present invention, there are no particular limitations on the expression vector, and for example, pTT5, pSECtag, pCGS3, pCDNA expression vectors, as well as vectors used in other mammalian expression systems, can be used. These expression vectors comprise a fusion DNA sequence linked to appropriate transcriptional and translational regulatory sequences.

[0069] In the present invention, host cells that can be used include cells that can express the above-mentioned expression vector. For example, the fusion protein of the present invention can be expressed using eukaryotic cells of a mammalian or insect host cell culture system. CHO (Chinese Hamster Ovary), HEK293, COS, BHK, and modified cells derived therefrom can also be used in the present invention.

[0070] In the present invention, a "drug composition" refers to a drug formulation comprising a tetravalent bispecific antibody of the present invention and a pharma- ceutically acceptable carrier, which is capable of exerting a stable therapeutic effect while at the same time maintaining the structural integrity of the amino acid core sequence of the tetravalent bispecific antibody of the present invention, and further protecting the polyfunctional groups of the protein from degradation (e.g., aggregation, deamination, and oxidation).

[0071] The protein expression and purification methods are described below in the Examples, but the present invention is not limited thereto. Specifically, a foreign gene was introduced into a pcDNA3.4 expression vector (Thermo Fisher Scientific), and the resulting expression vectors were combined and introduced into HEK293F cells (Thermo Fisher Scientific) using the polyethyleneimine (PEI) method to express the antibody, which was then purified using a Protein-A column.

[0072] The methods for measuring physicochemical properties in the examples will be described below.

[0073] The ELISA detection in the examples was carried out as follows. That is, the recombinant proteins were used to coat the ELISA plates, and the ELISA plates were blocked with PBST (phosphate buffer containing 0.05% Tween-20) containing 1% bovine serum albumin. The antibodies to be measured were diluted stepwise to form a concentration gradient, and then transferred to the ELISA plate coated with the recombinant proteins described above. The antibodies were then incubated at room temperature for 30 minutes, and the plate was then washed. Horseradish peroxidase (HRP)-labeled goat anti-human antibodies (Fc or Fab specific, Sigma) were appropriately diluted and added to the plate, incubated at room temperature for 30 minutes, and the plate was then washed. 100 μL of a coloring solution using 3,3',5,5'-tetramethylbenzidine (hereinafter also referred to as "TMB") as a substrate was added to each well, incubated at room temperature for 1 to 5 minutes, and then 2M H2SO4 was added as a stop solution. 2 SO 4 The reaction was stopped by adding 50 μL of 450 The values ​​were measured, the measurement data was processed with GraphPad Prism6 to create an analysis graph, and the EC 50 was calculated.

[0074] The mixed lymphocyte reaction (hereinafter also referred to as "MLR") of the embodiment was carried out as follows. That is, peripheral blood mononuclear cells (hereinafter also referred to as "PBMC") were separated from human blood using Histopaque (Sigma), and mononuclear cells of the PBMC were separated using an adhesion method. Then, the mononuclear cells were induced to differentiate into dendritic cells using IL-4 (25 ng / mL) plus GM-CSF (25 ng / mL), and after 7 days, the induced cells were digested to recover the dendritic cells. PBMC were also separated from the blood of another donor in the same manner as above, and CD4 positive T cells were separated from the PBMC using a MACS magnetic column and CD4 microbeads (Miltenyi Biotec). The induced dendritic cells (10 4 cells / well) and CD4+ T cells (10 5 The antibodies (100 μL / well) were mixed uniformly at an appropriate ratio and seeded in each well of a 96-well plate at 150 μL each. After several hours, 50 μL of serially diluted antibodies were added to the 96-well plate and incubated in a 37°C incubator for 3 days. In the above experiment, AIM-V medium (Thermo Fisher Scientific) was used for cell culture, and the secretion levels of IL-2 and IFN-γ were measured according to the standard procedure described in the manual. IL-2 and IFN-γ were detected by standard sandwich ELISA, and the antibody reagents used in the sandwich ELISA were from BD Biosciences. OD was measured using a SpectraMax 190 plate reader. 450 The values ​​were read out, and the obtained data was processed with GraphPad Prism6 to create an analysis graph, and EC 50 was calculated.

[0075] HPLC-SEC analysis was performed as follows. Antibodies are high molecular weight proteins with highly complex secondary and tertiary structures, and therefore undergo post-translational modifications, aggregation, degradation, and other changes, and are heterogeneous in terms of chemical and physical properties. Therefore, when bispecific antibodies are analyzed using separation techniques, denatured bodies, aggregates, and cleavage fragments are often found, the presence of which can cause great damage to the safety and efficacy of antibodies, and they are also prone to the formation of aggregates, cleavage fragments, and poorly organized molecules during antibody production and storage. In the present invention, high performance liquid chromatography-size exclusion chromatography (HPLC-SEC) was used to detect the amount of impurities in samples based on the principle that the retention time of the peak of aggregates, which have a larger molecular weight than the monomer, is short, while the molecular weight of degraded fragments (cleavage fragments) and poorly organized molecules is small, and the retention time of the associated peaks is long. The HPLC-SEC analysis was performed using a Dionex Ultimate 3000. The mobile phase was a 20 mM mother liquor of sodium dihydrogen phosphate adjusted to pH 6.8±0.1 with 20 mM disodium hydrogen phosphate. The sample injection amount was 20 g. The column used was a TSK G3000SWXL, column specification 7.8×300 mm 5 μm. The experiment was performed under the following conditions: flow rate 0.5 mL / 1 min, elution time 30 min, column temperature 25°C, sample chamber temperature 10°C, detection wavelength 214 nm.

[0076] HPLC-IEC analysis was performed as follows. The surface charge of an antibody is directly or indirectly changed by post-translational modifications such as N-glycosylation, modification of C-terminal lysine residues, cyclization of N-terminal glutamine or glutamic acid, deamidation of asparagine, isomerization of aspartic acid, and oxidation of amino acid residues, resulting in heterogeneity in the charge. Charge variants are also easily generated during production and storage of antibodies, so charge variants can be separated and analyzed based on their charge. Such charge variants can be analyzed using ion exchange chromatography (CEX) and anion exchange chromatography (AEX), and are analyzed by measuring the retention times of acidic and basic species relative to the main peak during analysis by ion exchange chromatography. Specifically, acidic species are eluted before the main peak of CEX or after the main peak of AEX, and basic species are eluted after the main peak of CEX or before the main peak of AEX, and the peaks corresponding to acidic species and basic species are called acidic peak and basic peak, respectively. In the present invention, the charging characteristics (i.e., charge uniformity) of the samples were analyzed using high performance liquid chromatography-ion exchange chromatography (HPLC-IEC). The HPLC-IEC analyzer used was Dionex Ultimate 3000, and the column was Thermo Propac TM The experiment was performed using a WCX-10, 20 mM PB (pH 6.3) as mobile phase A, and 20 mM PB + 200 mM NaCl (pH 6.3) as mobile phase B. The ratio of the two mobile phases was adjusted over time according to a pre-set program, and the conditions were a flow rate of 1.0 mL / min, a column temperature of 30°C, a sample chamber temperature of 10°C, a sample injection amount of 20 g, and a detection wavelength of 214 nm.

[0077] In the examples, the amount of cleaved fragments and the amount of incompletely organized molecules contained in a sample were measured using a capillary electrophoresis-SDS (hereinafter also referred to as "CE-SDS") method. Capillary electrophoresis is divided into non-reduced and reduced types, and samples subjected to non-reduced capillary electrophoresis do not require the use of DTT as a reducing agent to break intramolecular disulfide bonds, while samples subjected to reduced capillary electrophoresis are subjected to breaking intramolecular disulfide bonds using DTT. Hereinafter, non-reduced and reduced CE-SDS may be abbreviated as NR-CE-SDS and R-CE-SDS, respectively. The capillary electrophoresis apparatus used was the ProteomeLab CE-SDS manufactured by Beckman Coulter. TM The capillary electrophoresis apparatus was equipped with a 214 nm UV detector, the capillary model was a Bare Fused-Silica Capillary, with a specification of 30.7 cm × 50 μm and an effective length of 20.5 cm, and all related reagents were from Beckman Coulter. Capillary electrophoresis was performed under the conditions of capillary and sample chamber temperatures of 20 ± 2°C and separation voltage of 15 kV.

[0078] Differential scanning calorimetry (hereinafter also referred to as "DSC") is a method for evaluating the thermal stability of biological samples by measuring the associated changes in heat of biomolecules while changing the temperature according to a controllable heating or cooling program. The principle is that proteins absorb heat during the heating process, causing the folded structure to unfold, and at this time, a temperature difference occurs in the sample chamber, and the energy that compensates for such a temperature difference can be measured and recorded by a detection device. These heat changes appear as peaks in the detection spectrum, and the peak temperature value when the folded structure of the protein unfolds is, so to speak, the melting temperature of the protein (hereinafter also referred to as "Tm"). Tm is an important index of the thermal stability of a protein, and the higher the Tm, the higher the stability of the protein.

[0079] The sequence information according to the present invention is summarized in Table 1 below. (Table 1) Sequence number Sequence name 1. Amino acid sequence of the heavy chain variable region of Bevacizumab (601) 2. Amino acid sequence of the light chain variable region of Bevacizumab (601) 3. Amino acid sequence of the heavy chain variable region of mouse antibody No. 20 4. Amino acid sequence of the light chain variable region of mouse antibody No. 20 5. Amino acid sequence of the heavy chain complementarity determining region H-CDR1 of mouse antibody No. 20 6. Amino acid sequence of the heavy chain complementarity determining region H-CDR2 of mouse antibody No. 20 7. Amino acid sequence of the heavy chain complementarity determining region H-CDR3 of mouse antibody No. 20 8. Amino acid sequence of the light chain complementarity determining region (L-CDR1) of mouse antibody No. 20 9. Amino acid sequence of the light chain complementarity determining region (L-CDR2) of mouse antibody No. 20 10 Amino acid sequence of the light chain complementarity determining region (L-CDR3) of mouse antibody No. 20 11 Amino acid sequence of the heavy chain variable region of 20-Hu 12 Amino acid sequence of the light chain variable region of 20-Hu 13 Amino acid sequence of 601-LC-V94L Nucleotide sequence of 14601-LC-V94L Amino acid sequence of the light chain variable region of 15 601-LC-V94L 16 Amino acid sequence of 20-Fab-601-IgG4 17. Nucleotide sequence of 20-Fab-601-IgG4 18 Amino acid sequence of 601-Fab-20-IgG4 Nucleotide sequence of 19601-Fab-20-IgG4 20 Amino acid sequence of the heavy chain variable region of Y0313-1 21 Amino acid sequence of 20-Fab-0313-IgG4 22 Nucleotide sequence of 20-Fab-0313-IgG4 23 Amino acid sequence of the heavy chain complementarity determining region H-CDR1 of mouse-derived antibody 1D11 24 Amino acid sequence of the heavy chain complementarity determining region H-CDR2 of mouse-derived antibody 1D11 25 Amino acid sequence of the heavy chain complementarity determining region H-CDR3 of mouse-derived 1D11 antibody 26 Amino acid sequence of the light chain complementarity determining region L-CDR1 of mouse antibody 1D11 27 Amino acid sequence of the light chain complementarity determining region (L-CDR2) of mouse antibody 1D11 28 Amino acid sequence of the light chain complementarity determining region (L-CDR3) of mouse-derived 1D11 antibody 29 Amino acid sequence of the heavy chain variable region of 1D11-Hu 30 Amino acid sequence of the light chain variable region of 1D11-Hu 31 Amino acid sequence of the heavy chain variable region of mAb127 32 Amino acid sequence of the light chain variable region of mAb127 33 Amino acid sequence of the heavy chain variable region of mouse antibody No. 14 34 Amino acid sequence of the light chain variable region of mouse antibody No. 14 35 Amino acid sequence of heavy chain complementarity determining region H-CDR1 of mouse antibody No. 14 36 Amino acid sequence of the heavy chain complementarity determining region H-CDR2 of mouse antibody No. 14 37 Amino acid sequence of the heavy chain complementarity determining region H-CDR3 of mouse antibody No. 14 38 Amino acid sequence of the light chain complementarity determining region (L-CDR1) of mouse antibody No. 14 39 Amino acid sequence of the light chain complementarity determining region (L-CDR2) of mouse antibody No. 14 40 Amino acid sequence of the light chain complementarity determining region (L-CDR3) of mouse antibody No. 14 41 Amino acid sequence of the heavy chain variable region of 14-Hu 42 Amino acid sequence of the light chain variable region of 14-Hu 43 Amino acid sequence of 14-Hu-LC 44 Nucleotide sequence of 14-Hu-LC 45 Amino acid sequence of 14-Fab-1D11-IgG4 46 Nucleotide sequence of 14-Fab-1D11-IgG4 47 Amino acid sequence of 1D11-Fab-14-IgG4 48 Nucleotide sequence of 1D11-Fab-14-IgG4 49 Amino acid sequence of 14-Fab-127-IgG4 Nucleotide sequence of 5014-Fab-127-IgG4 51 Amino acid sequence of 127-Fab-14-IgG4 52 Nucleotide sequence of 127-Fab-14-IgG4 53 Amino acid sequence of the heavy chain variable region of 19H6-Hu 54 Amino acid sequence of the light chain variable region of 19H6-Hu Amino acid sequence of the heavy chain complementarity determining region H-CDR1 of 55 Anti-CD47B 56 Amino acid sequence of the heavy chain complementarity determining region H-CDR2 of Anti-CD47B Amino acid sequence of the heavy chain complementarity determining region H-CDR3 of 57 Anti-CD47B Amino acid sequence of the light chain complementarity determining region L-CDR1 of 58 Anti-CD47B 59 Amino acid sequence of the light chain complementarity determining region L-CDR2 of Anti-CD47B 60 Amino acid sequence of the light chain complementarity determining region (L-CDR3) of Anti-CD47B 61 Amino acid sequence of the heavy chain variable region of Anti-CD47B-Hu Amino acid sequence of the light chain variable region of 62 Anti-CD47B-Hu 63 Amino acid sequence of 19H6-Hu-LC 64 Nucleotide sequence of 19H6-Hu-LC 65 Amino acid sequence of CD47B-Fab-19H6-IgG1 Nucleotide sequence of 66CD47B-Fab-19H6-IgG1 67 Amino acid sequence of 19H6-Fab-CD47B-IgG1 68 Nucleotide sequence of 19H6-Fab-CD47B-IgG1 69 Amino acid sequence of the heavy chain variable region of mouse antibody No. 94 70 Amino acid sequence of the light chain variable region of mouse antibody No. 94 71 Amino acid sequence of the heavy chain complementarity determining region H-CDR1 of mouse antibody No. 94 72 Amino acid sequence of the heavy chain complementarity determining region H-CDR2 of mouse antibody No. 94 73 Amino acid sequence of the heavy chain complementarity determining region H-CDR3 of mouse antibody No. 94 74 Amino acid sequence of the light chain complementarity determining region L-CDR1 of mouse antibody No. 94 75 Amino acid sequence of the light chain complementarity determining region (L-CDR2) of mouse antibody No. 94 76 Amino acid sequence of the light chain complementarity determining region (L-CDR3) of mouse antibody No. 94 77 Amino acid sequence of the heavy chain variable region of 94-Hu 78 Amino acid sequence of the light chain variable region of 94-Hu 79 Amino acid sequence of 94-Fab-19H6-IgG1-LALA Nucleotide sequence of 8094-Fab-19H6-IgG1-LALA

[0080] 81 Amino acid sequence of 19H6-Fab-94-IgG1-LALA 82 Nucleotide sequence of 19H6-Fab-94-IgG1-LALA 83 Amino acid sequence of the heavy chain variable region of mAb1-25-Hu(609) 84 Amino acid sequence of the light chain variable region of mAb1-25-Hu(609) Amino acid sequence of the heavy chain complementarity determining region H-CDR1 of 85 Anti-CD137 Amino acid sequence of the heavy chain complementarity determining region H-CDR2 of 86 Anti-CD137 Amino acid sequence of the heavy chain complementarity determining region H-CDR3 of 87 Anti-CD137 Amino acid sequence of the light chain complementarity determining region L-CDR1 of 88 Anti-CD137 Amino acid sequence of the light chain complementarity determining region L-CDR2 of 89 Anti-CD137 90 Amino acid sequence of the light chain complementarity determining region (L-CDR3) of Anti-CD137 91 Amino acid sequence of the heavy chain variable region of Anti-CD137-Hu Amino acid sequence of the light chain variable region of 92 Anti-CD137-Hu 93 Amino acid sequence of Anti-CD137-Hu-LC Nucleotide sequence of 94 Anti-CD137-Hu-LC 95 Amino acid sequence of 609-Fab-137-IgG4 Nucleotide sequence of 96609-Fab-137-IgG4 97 Amino acid sequence of 137-Fab-609-IgG4 98 Nucleotide sequence of 137-Fab-609-IgG4 99 Amino acid sequence of the heavy chain complementarity determining region H-CDR1 of Anti-CD40 100 Amino acid sequence of the heavy chain complementarity determining region H-CDR2 of Anti-CD40 Amino acid sequence of the heavy chain complementarity determining region H-CDR3 of 101 Anti-CD40 102 Amino acid sequence of the light chain complementarity determining region L-CDR1 of Anti-CD40 Amino acid sequence of the light chain complementarity determining region L-CDR2 of 103 Anti-CD40 Amino acid sequence of the light chain complementarity determining region (L-CDR3) of 104 Anti-CD40 Amino acid sequence of the heavy chain variable region of 105 Anti-CD40-Hu Amino acid sequence of the light chain variable region of 106 Anti-CD40-Hu Amino acid sequence of 107 Anti-CD40-Hu-LC Nucleotide sequence of 108 Anti-CD40-Hu-LC Amino acid sequence of 109 609-Fab-40-IgG4 Nucleotide sequence of 110609-Fab-40-IgG4 111 Amino acid sequence of 40-Fab-609-IgG4 112 Nucleotide sequence of 40-Fab-609-IgG4 113 Amino acid sequence of the heavy chain variable region of cetuximab 114 Amino acid sequence of the light chain variable region of cetuximab 115 Amino acid sequence of the heavy chain variable region of Trastuzumab 116 Amino acid sequence of the light chain variable region of Trastuzumab 117 Amino acid sequence of the heavy chain variable region of Pertuzumab 118 Amino acid sequence of the light chain variable region of Pertuzumab 119 Amino acid sequence of the heavy chain variable region of 10D1 (Ipilimumab) 120 Amino acid sequence of the light chain variable region of 10D1 (Ipilimumab) 121 Amino acid sequence of the heavy chain variable region of 5E7-Hu 122 Amino acid sequence of the light chain variable region of 5E7-Hu 123 Amino acid sequence of Cetuximab-LC 124 Nucleotide sequence of Cetuximab-LC 125 Amino acid sequence of Pertuzumab-LC 126 Nucleotide sequence of Pertuzumab-LC 127 Amino acid sequence of Ipilimumab-LC 128 Nucleotide sequence of Ipilimumab-LC 129 Amino acid sequence of 5E7-Hu-LC 130 Nucleotide sequence of 5E7-Hu-LC Amino acid sequence of 131 609-Fab-Cetuximab-IgG4 Nucleotide sequence of 132 609-Fab-Cetuximab-IgG4 Amino acid sequence of 133 609-Fab-Pertuzumab-IgG4 Nucleotide sequence of 134 609-Fab-Pertuzumab-IgG4 Amino acid sequence of 135 609-Fab-Ipilimumab-IgG1 Nucleotide sequence of 136 609-Fab-Ipilimumab-IgG1 137 Amino acid sequence of Ipilimumab-Fab-609-IgG1 138 Nucleotide sequence of Ipilimumab-Fab-609-IgG1 Amino acid sequence of 139 609-Fab-Ipilimumab-IgG4 Nucleotide sequence of 140609-Fab-Ipilimumab-IgG4 141 Amino acid sequence of Ipilimumab-Fab-609-IgG4 142 Nucleotide sequence of Ipilimumab-Fab-609-IgG4 Amino acid sequence of 143 609-Fab-5E7-IgG4 Nucleotide sequence of 144 609-Fab-5E7-IgG4 145 Amino acid sequence of 5E7-Fab-609-IgG4 146 Nucleotide sequence of 5E7-Fab-609-IgG4 147 Amino acid sequence of IgG1 heavy chain constant region 148 Amino acid sequence of the heavy chain constant region of IgG4(S228P) 149 Amino acid sequence of the κ light chain constant region 150 Linker (GGGGSGGGGSGGGGS) 151 Fourth framework region of heavy chain (WGQGTLVTVSS) 152 Fourth framework region of light chain (FGQGTKVEIK) 153 Fourth framework region of light chain (FGGGTKVELK)

[0081] The present invention will be described in detail below with reference to examples and experimental examples, but the present invention is not limited to these examples and experimental examples. Detailed explanations of well-known techniques, such as methods for constructing vectors and plasmids, methods for introducing genes encoding proteins into vectors and plasmids, and methods for introducing plasmids into host cells, may be omitted below. These techniques are well known to those skilled in the art, and are described in detail in publications such as Sambrook, J., Fritsch, EF and Maniais, T. (1989), Molecular Cloning: A Laboratory Manual, 2nd edition, Cold Spring Harbor Laboratory Press.

[0082] Example 1: Generation of bispecific antibodies targeting PD-1 and VEGF 1.1) Array The heavy chain variable region and light chain variable region sequences (SEQ ID NO: 1 and SEQ ID NO: 2) of the bevacizumab (hereinafter also referred to as "601") antibody were those described in the literature (Magdelaine-Beuzelin C, Kaas Q, Wehbi V, et al., Structure-function relationships of the variable domains of monoclonal antibodies approved for cancer treatment[J]. Critical reviews in oncology / hematology, 2007, 64(3):210-225). We commissioned Shanghai Synthetic Biotech Co., Ltd. to synthesize DNA encoding the above variable regions, and linked the heavy chain variable region (601-VH region) and light chain variable region (601-VL region) of 601 to the heavy chain constant region of human IgG1 (sequence number 147) and the light chain constant region of human Kappa (sequence number 149), respectively, to construct the full-length heavy chain gene and full-length light chain gene of the 601 antibody, which we named 601-HC and 601-LC, respectively.

[0083] According to Examples 1 to 5 of the specification of WO2018 / 137576A1, ultimately, based on the screening results, mouse-derived human PD-1 monoclonal antibody No. 20 was selected as the lead antibody, and the nucleotide sequences of its heavy chain variable region and light chain variable region were obtained and converted into amino acid sequences (SEQ ID NO: 3 and SEQ ID NO: 4).

[0084] The amino acid sequences of the heavy and light chain variable regions of antibody No. 20 were analyzed, and the antigen complementarity determining regions and framework regions of the heavy and light chains of antibody No. 20 were determined according to the KABAT rules. The amino acid sequences of the heavy chain CDRs of antibody No. 20 were as follows: H-CDR1 was NYDMS (SEQ ID NO: 5), H-CDR2 was TISGGGGYTYYSDSVKG (SEQ ID NO: 6), and H-CDR3 was PYGHYGFEY (SEQ ID NO: 7), and the amino acid sequences of the light chain CDRs were as follows: L-CDR1 was SASQGISNFLS (SEQ ID NO: 8), L-CDR2 was YTSSLHS (SEQ ID NO: 9), and L-CDR3 was QQYSNLPWT (SEQ ID NO: 10).

[0085] The heavy chain variable region of mouse-derived antibody No. 20 was compared for homology with human IgG germline sequences at https: / / www.ncbi.nlm.nih.gov / igblast / , and IGHV3-21*01 was selected as a template for heavy chain CDR grafting. The heavy chain CDR of mouse-derived antibody No. 20 was grafted into the frame region of IGHV3-21*01, and WGQGTLVTVSS (sequence number 151) was added after H-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted heavy chain variable region. Similarly, the light chain variable region of mouse-derived antibody No. 20 was compared with the human IgG germline sequence for homology, IGKV1-39*01 was selected as a light chain CDR grafting template, the light chain CDR of mouse-derived antibody No. 20 was grafted into the frame region of IGKV1-39*01, and FGQGTKVEIK (SEQ ID NO: 152) was added after L-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted light chain variable region. Furthermore, based on the CDR-grafted variable region, back mutations (back mutations mean changing some amino acids in the human-derived framework region to amino acids at the same positions in the mouse-derived framework region, and the back mutation sites are usually very important for maintaining the antibody structure and / or affinity) were introduced into amino acid sites in some framework regions, and when introducing the back mutations, KABAT editing was performed on the amino acid sequence, and the site positions were changed to KABAT numbering.

[0086] Preferably, the CDR-grafted heavy chain variable region is subjected to KABAT editing, changing T at position 28 of the KABAT numbering to V derived from mouse, G at position 44 to R, and R at position 94 to S, and in the CDR-grafted light chain variable region, A at position 43 to T, P at position 44 to V, and F at position 71 to Y.

[0087] The heavy chain variable region and light chain variable region containing the above-mentioned reversion mutation site were made into humanized heavy chain variable region and light chain variable region (SEQ ID NO: 11 and SEQ ID NO: 12), respectively, and DNA encoding the above-mentioned humanized heavy chain variable region and light chain variable region was synthesized by Shanghai Synchro Biotech Co., Ltd. The obtained humanized heavy chain variable region was linked to a human IgG4 (containing S228P as a site mutation in the hinge region) constant region (SEQ ID NO: 148) to obtain a full-length humanized heavy chain gene, which was named 20-Hu-HC, and the humanized light chain variable region was linked to a human Kappa chain constant region (SEQ ID NO: 149) to obtain a full-length humanized light chain gene, which was named 20-Hu-LC.

[0088] The heavy and light chain genes of the above antibodies were introduced into pcDNA3.4 expression vectors, and the resulting heavy and light chain expression vectors were introduced into HEK293F cells to express the antibodies. HEK293F cells were cultured in Free Style293 expression medium for 5 days, and the supernatant was collected and the antibodies were purified using a Protein-A column. The antibody that combines 20-Hu-HC and 20-Hu-LC was named 20-Hu.

[0089] 1.2) Selection of common light chain A comparative amino acid sequence analysis was performed using BLAST on the light chain variable regions of 20-Hu and 601. It was found that 89% (Identities) of the amino acids were completely identical between the two, and 94% (Positives) of the amino acids had similar physicochemical properties.

[0090] The gene sequences of 601-HC and 601-LC were respectively introduced into pcDNA3.4 expression vector, and the expression vectors of 20-Hu-HC, 20-Hu-LC, 601-HC and 601-LC were combined as follows: 20-Hu-HC+20-Hu-LC, 601-HC+601-LC, 20-Hu-HC+601-LC and 601-HC+20-Hu-LC, and the antibodies were expressed and purified, and the obtained antibodies were named 20-Hu, 601, 20-Hu-HC+601-LC and 601-HC+20-Hu-LC, respectively.

[0091] The measurement by ELISA method was carried out as follows: 6xHis tagged human PD-1 extracellular domain (the extracellular domain of PD-1 is disclosed in WO2018 / 137576A1) was prepared by ourselves, and the recombinant protein was designated as PD1-His. 6xHis tagged human VEGF165 (the sequence is from NCBI, the accession number is AAM03108) was also prepared by ourselves, and the recombinant protein was designated as VEGF165-His. PD1-His and VEGF165-His were respectively used to coat the ELISA plate, and the coating concentration was 20ng / well and 10ng / well, respectively.

[0092] As shown in Figure 2A, 20-Hu and 20-Hu-HC+601-LC effectively bound to PD1-His and inhibited EC 50 The EC values ​​of 601 and 601-HC+20-Hu-LC were 0.2062 nM and 0.9747 nM, respectively. On the other hand, 601 and 601-HC+20-Hu-LC could not bind to PD1-His and had low EC 50 As shown in Figure 2B, 601 and 601-HC+20-Hu-LC effectively bound to VEGF165-His with EC50 of 0.4681 nM and 8.217 nM, respectively, while 20-Hu and 20-Hu-HC+601-LC were unable to bind to VEGF165-His.

[0093] The relative affinity of 20-Hu-HC+601-LC for PD1-His was significantly lower than that of 20-Hu, and the relative affinity of 601-HC+20-Hu-LC for VEGF165-His was also significantly lower than that of 601. In order to increase the relative affinity for PD1-His, a search experiment was conducted to introduce back mutations into 20-Hu-HC+601-LC. As a result, 12 amino acid residues were found to be different between the 601 light chain variable region and the 20-Hu light chain variable region, and among them, the amino acid residues at positions 28, 32, 34, 46, 50, 71, 93, and 94 in the KABAT numbering were confirmed to be extremely important for maintaining affinity to the antibody. In addition, site mutations were introduced into the above positions of 601-LC, and these amino acid residues were changed to the amino acid residues at the corresponding positions of 20-Hu-LC, respectively, and 601-LCs having these site mutations were designated as 601-LC-D28G, 601-LC-Y32F, 601-LC-N34S, 601-LC-V46L, 601-LC-F50Y, 601-LC-F71Y, 601-LC-T93N and 601-LC-V94L, respectively.

[0094] The gene sequences of the light chains were introduced into pcDNA3.4 expression vectors, and the expression vectors of 20-Hu-HC and 601-LC carrying the above site mutations were combined to express and purify the antibodies. The resulting antibodies were named 20-Hu-HC+601-LC-D28G, 20-Hu-HC+601-LC-Y32F, 20-Hu-HC+601-LC-N34S, 20-Hu-HC+601-LC-V46L, 20-Hu-HC+601-LC-F50Y, 20-Hu-HC+601-LC-F71Y, 20-Hu-HC+601-LC-T93N, and 20-Hu-HC+601-LC-V94L. The relative affinities of the antibodies to bind to PD-1 were measured using the ELISA method described above, with 20-Hu-HC+601-LC used as a control. The ELISA results confirmed that the EC50 values ​​of the mutants and 20-Hu-HC+601-LC were 0.4849 nM, 0.4561 nM, 0.1751 nM, 0.5333 nM, 0.5255 nM, 1.0345 nM, 0.4859 nM, 0.3079 nM, and 0.6251 nM, respectively. Compared to 20-Hu-HC+601-LC, 20-Hu-HC+601-LC-N34S and 20-Hu-HC+601-LC-V94L showed higher relative affinity for PD-1, while the relative affinity of the other mutant antibodies was almost the same as that of 20-Hu-HC+601-LC or was even lower than that of 20-Hu-HC+601-LC.

[0095] The expression vector of 601-HC and the expression vector of 601-LC carrying the above site mutation were respectively combined to express and purify the antibodies, and the obtained antibodies were named 601-HC+601-LC-D28G, 601-HC+601-LC-Y32F, 601-HC+601-LC-N34S, 601-HC+601-LC-V46L, 601-HC+601-LC-F50Y, 601-HC+601-LC-F71Y, 601-HC+601-LC-T93N, and 601-HC+601-LC-V94L. The above-mentioned ELISA method was used to measure the relative affinity of the above antibodies to bind to VEGF165, and 601 was used as the control. The ELISA measurement results showed that the EC of the above mutants and 601 were significantly higher than that of the control.50 were confirmed to be 0.1328nM, 0.1254nM, 0.2081nM, 0.3256nM, 0.1400nM, 0.1481nM, 0.1259nM, 0.1243nM and 0.1291nM, respectively. Furthermore, it was confirmed that 601-HC+601-LC-N34S and 601-HC+601-LC-V46L had significantly reduced relative affinity for binding to VEGF165 compared to 601, while the other mutant antibodies had almost the same level of relative affinity for binding to VEGF165 as 601.

[0096] In summary, it was confirmed that the V94L mutant does not reduce the relative affinity of 601 to bind to VEGF165, and at the same time, it can increase the relative affinity of 20-Hu-HC+601-LC to bind to PD-1. Based on this, we decided to prepare a bispecific antibody using 601-LC-V94L (SEQ ID NO: 13 and SEQ ID NO: 14) as a common light chain.

[0097] 1.3) Generation of bispecific antibodies The heavy chain variable region of 20-Hu and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 601 was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain gene containing two heavy chain variable regions and two CH1 structural domains, and this was named 20-Fab-601-IgG4 (SEQ ID NO: 16 and SEQ ID NO: 17). Similarly, the heavy chain variable region of 601 and the CH1 structural domain of human IgG4 were linked, and then the heavy chain variable region of 20-Hu was attached via an artificial linker GGGGSGGGGSGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain gene containing two heavy chain variable regions and two CH1 structural domains, and this was named 601-Fab-20-IgG4 (SEQ ID NO: 18 and SEQ ID NO: 19).

[0098] The above sequences were each introduced into a pcDNA3.4 expression vector, and the 20-Fab-601-IgG4 expression vector and the 601-Fab-20-IgG4 expression vector were each combined with the 601-LC-V94L expression vector to express and purify the antibodies. The resulting antibodies were named 20-Fab-601-IgG4-V94L and 601-Fab-20-IgG4-V94L, respectively.

[0099] 1.4) Measurement of relative affinity by ELISA method As shown in Figure 3A, 20-Hu-HC+601-LC-V94L, 20-Fab-601-IgG4-V94L and 601-Fab-20-IgG4-V94L effectively bound to PD1-His and inhibited EC 50 The EC50s of 601-HC+601-LC-V94L, 20-Fab-601-IgG4-V94L, and 601-Fab-20-IgG4-V94L were 0.3314 nM, 0.4768 nM, and 1.772 nM, respectively. As shown in Figure 3B, 601-HC+601-LC-V94L, 20-Fab-601-IgG4-V94L, and 601-Fab-20-IgG4-V94L effectively bound to VEGF165-His with EC50s of 0.01872 nM, 0.05859 nM, and 0.03886 nM. In addition, 20-Fab-601-IgG4-V94L and 601-Fab-20-IgG4-V94L were able to bind to both PD-1 and VEGF, demonstrating that they are bispecific antibodies.

[0100] 1.5) Measurement of affinity by Biacore method The affinity of the above antibodies to PD-1 or VEGF was measured using a Biacore 8K instrument manufactured by GE Healthcare. Specifically, in the Biacore 8K instrument, each antibody was captured on a chip immobilized with Protein-A / G, and the recombinant protein PD1-His or VEGF165-His was then injected to obtain binding-dissociation graphs, and a 6M guanidine hydrochloride solution was used for chip regeneration. The data obtained was analyzed using analysis software dedicated to the Biacore 8K instrument, and the results are summarized in Table 2 below.

[0101] [Table 1]

[0102] As shown in Table 2, 20-Hu-HC+601-LC-V94L and 20-Fab-601-IgG4-V94L showed particularly close values ​​for the equilibrium dissociation constant (KD) for PD-1, with KDs of 1.59E-09 and 8.85E-10, respectively. In addition, 601-HC+601-LC-V94L and 20-Fab-601-IgG4-V94L showed almost the same binding constant (Kon) and dissociation constant (Koff) for VEGF165-His, and the equilibrium dissociation constants (KD) were also at almost the same level, with KDs of 7.46E-12 and 9.26E-12, respectively. Here, the equilibrium dissociation constant (KD) and affinity were inversely proportional to each other.

[0103] 1.6) Evaluation of physicochemical properties 1.6.1) HPLC-SEC FIG. 4A shows the HPLC-SEC spectrum of monoclonal antibody 601-HC+601-LC-V94L. In the spectrum, peaks 1 and 2 are particularly prominent, with the occupancy rates of each peak being 0.7% and 99.3%, respectively. Peak 1 has a shorter retention time than peak 2, the main peak, and is thought to be due to aggregates. No peaks representing decomposition fragments (cleavage fragments) or incompletely organized molecules were observed in the spectrum. FIG. 4B shows the HPLC-SEC spectrum of 20-Fab-601-IgG4-V94L. Peaks 1 and 2 are particularly prominent in the spectrum, with the occupancy rates of each peak being 0.7% and 99.3%, respectively. Peak 1 has a shorter retention time than peak 2, the main peak, and is thought to be due to aggregates. No peaks representing decomposition fragments (cleavage fragments) or incompletely organized molecules were observed in the spectrum.

[0104] 1.6.2) HPLC-IEC Figures 5A and 5B show the HPLC-IEC spectra of 601-HC+601-LC-V94L and 20-Fab-601-IgG4-V94L. The occupancy rates of the major peaks were 79.31% and 80.64%, respectively, confirming that 20-Fab-601-IgG4-V94L was equivalent to 601-HC+601-LC-V94L in terms of charging properties.

[0105] 1.6.3) CE-SDS Figures 6A and 6B are the NR-CE-SDS and R-CE-SDS spectra of 601-HC+601-LC-V94L, respectively. In the NR-CE-SDS spectrum, the occupancy rate of the main peak, peak 9, was 97.90%. In the R-CE-SDS spectrum, the occupancy rates of the two main peaks, peak 6 (peak corresponding to the light chain) and peak 12 (peak corresponding to the heavy chain), were 30.92% and 65.27%, respectively, and the area ratio of the two peaks was 1:2.1. Figures 6C and 6D are the NR-CE-SDS and R-CE-SDS spectra of 20-Fab-601-IgG4-V94L, respectively. In the NR-CE-SDS spectrum, the occupancy rate of the main peak, peak 13, was 96.74%. In the R-CE-SDS spectrum, the occupancy rates of the two major peaks, peak 3 (peak corresponding to the light chain) and peak 12 (peak corresponding to the heavy chain), were 38.42% and 59.74%, respectively, and the area ratio of the two peaks was 2:3.1. 601-HC+601-LC-V94L and 20-Fab-601-IgG4-V94L showed particularly close values ​​for the occupancy rate of the major peaks in the NR-CE-SDS spectrum, and the peak area ratio of the light chain to the heavy chain showed the expected value in the R-CE-SDS spectrum.

[0106] 1.6.4) DSC Figures 7A and 7B show the DSC spectra of 601-HC+601-LC-V94L and 20-Fab-601-IgG4-V94L, respectively. The TmOnset and Tm of 601-HC+601-LC-V94L were 66.46°C and 75.37°C, respectively, and the TmOnset and Tm of 20-Fab-601-IgG4-V94L were 65.92°C and 74.28°C, respectively. These results demonstrated that 20-Fab-601-IgG4-V94L has thermal stability very similar to that of 601-HC+601-LC-V94L.

[0107] 1.7) Generation of improved bispecific antibodies 1.7.1) Generation of bispecific antibodies In U.S. Patent Application Publication US20020032315A1, the inventors used phage display technology to improve the heavy chain variable region and light chain variable region of bevacizumab, thereby obtaining an amino acid sequence of the heavy chain variable region, Y0313-1 (sequence number 114 described in US20020032315A1, the amino acid sequence represented by sequence number 20 in the present invention), which has superior affinity and neutralizing activity.

[0108] Here, the 601 (bevacizumab) heavy chain variable region in 20-Fab-601-IgG4-V94L was replaced with Y0313-1. Specifically, the heavy chain variable region of 20-Hu was linked to the CH1 structural domain of human IgG4, and the heavy chain variable region of Y0313-1 was attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain gene containing two heavy chain variable regions and two CH1 structural domains, which was named 20-Fab-0313-IgG4 (SEQ ID NO: 21 and SEQ ID NO: 22).

[0109] The above sequence was introduced into a pcDNA3.4 expression vector, and the expression vector for 20-Fab-0313-IgG4 was combined with the expression vector for 601-LC-V94L to express and purify the antibody. The resulting antibody was named 20-Fab-0313-IgG4-V94L.

[0110] 1.7.2) Evaluation of neutralizing ability against VEGF bioactivity Human umbilical vein endothelial cells (HUVEC) were purchased from AllCells Biotech, and HUVEC cells were subcultured using complete medium for umbilical vein endothelial cells (product number and specification H-004 / 500mL, manufactured by AllCells Biotech). When HUVEC cells were grown in the logarithmic growth phase, they were digested with pancreatin to release the cells, centrifuged, collected, and resuspended in complete medium. Then, the cells were seeded in a 96-well cell culture plate at a density of 8000 cells / well, and after 24 hours, the complete medium in the 96-well plate was replaced with basal medium (product number and specification H-004B / 500mL, manufactured by AllCells Biotech), and the amount of basal medium was 150 μL per well. A concentration gradient was formed by stepwise diluting the antibody using a base medium containing 400 ng / mL recombinant VEGF165 (product number VE5-H4210, Acrobiosystems). The mixture of VEGF165 and the antibody was added to a 96-well plate at 50 μL per well and incubated at 37°C, 5% CO 2 The plates were incubated in an incubator for 3 days. After 3 days, 20 μL of CCK-8 (Dojindo) solution was added to each well, and the plates were incubated in an incubator for another 4 hours. OD was measured using a plate reader. 450 The values ​​are read out, the obtained data is processed in GraphPad Prism6 to create an analysis graph, and the IC 50 was calculated.

[0111] As shown in Figure 8, 601, 20-Fab-601-IgG4-V94L, 601-Fab-20-IgG4-V94L and 20-Fab-0313-IgG4-V94L could effectively inhibit the proliferation of HUVEC cells induced by VEGF165, and the IC 50 and were 4.422 nM, 9.039 nM, 3.84 nM, and 1.632 nM, respectively. These results confirmed that 20-Fab-0313-IgG4-V94L showed the strongest neutralizing ability against VEGF165 bioactivity.

[0112] 1.7.3) Evaluation of MLR enhancing activity As shown in Figures 9A and 9B, 20-Humanized (i.e., 20-Hu), 20-Hu-HC+601-LC-V94L, and 20-Fab-0313-IgG4-V94L can all effectively stimulate IL-2 and IFN-γ secretion from the MLR, and the EC 50 The EC values ​​for stimulating IFN-γ secretion from MLR were 0.2571 nM, 0.3703 nM, and 0.7554 nM, respectively. 50 were 0.1426 nM, 0.247 nM and 1.036 nM, respectively.

[0113] 1.7.4) Evaluation of the binding ability of bispecific antibodies against PD-1 and VEGF A microplate was coated with VEGF165 (product number VE5-H4210, manufactured by Acrobiosystems), and the antibody to be measured was stepwise diluted with PBST containing 1% bovine serum albumin to form a concentration gradient, which was then added to the microplate and incubated at room temperature for about 30 minutes, after which the plate was washed three times with PBST. Biotinylated human PD-1 extracellular domain (product number 10377-H08H-B, manufactured by Sino Biological) was diluted to 200 ng / mL and added to the microplate, incubated at room temperature for about 30 minutes, and then the plate was washed three times with PBST. HRP-labeled streptavidin (product number 554066, manufactured by BD Biosciences) was diluted 1000-fold and added to the microplate, incubated at room temperature for about 30 minutes, and then the plate was washed three times with PBST. TMB staining solution (100 μL / well) was added and incubated at room temperature for 1 to 5 minutes, after which the color reaction was stopped by adding stop solution (50 μL / well). The OD was measured using a plate reader. 450 The values ​​were read out, and the obtained data was processed with GraphPad Prism6 to create an analysis graph, and EC 50 was calculated.

[0114] As shown in Figure 10, 20-Fab-0313-IgG4-V94L can effectively bind to human PD-1 in addition to VEGF165, and inhibits EC 50 was 0.3293 nM. On the other hand, 601 and 20-Hu-HC+601-LC-V94L were unable to bind to PD-1 and VEGF165 simultaneously.

[0115] 1.7.5) Pharmacokinetic evaluation of a bispecific antibody targeting PD-1 and VEGF in rats The pharmacokinetics of bispecific antibodies was studied using SD rats (Sprague-Dawley rats, Zhejiang Weitong Lihua Laboratory Animal Company). Four rats were assigned to each group, with an average weight of approximately 200 g. The rats were administered 1 mg of antibody via tail vein injection (IV). Blood was collected from the orbit at the designated time points after administration, and serum was collected by centrifugation after waiting for the blood to clot naturally.

[0116] The concentration of the target antibody in the serum was measured as follows. That is, an ELISA plate was coated with two antigens corresponding to the bispecific antibody (VEGF165 was manufactured by Acrobiosystems, and the product number was VE5-H4210. The 6×His-tagged human PD-1 extracellular domain was that described in 1.2 above), and the ELISA plate was blocked with PBST containing 1% bovine serum albumin. Appropriately diluted rat serum was added to the ELISA plate coated with the above two antigens, incubated at room temperature for 1 hour, and then the plate was washed. Then, an HRP-labeled goat anti-human antibody (Fc-specific, manufactured by Sigma) was added, and the antibody was adsorbed to eliminate interspecies cross-reactivity and did not recognize the rat antibody. After incubation at room temperature for 30 minutes, the plate was washed, and 100 μL of a coloring solution using TMB substrate was added to each well, incubated at room temperature for 1 to 5 minutes, and then 2 M H was added as a stop solution. 2 SO 4 Stop the reaction by adding 50 μL of 450 The OD value was measured using a calibration curve. 450 The values ​​were converted to serum antibody concentrations, and the resulting data was processed with GraphPad Prism6 to create analytical graphs. The half-life of the antibody drug in the rat body was calculated using Phoenix software.

[0117] As shown in Figure 11, 20-Fab-0313-IgG4-V94L had a half-life of 16.9 days when PD-1 was used as the antigen, and a half-life of 17.3 days when VEGF165 was used as the antigen. These results confirmed that 20-Fab-0313-IgG4-V94L has good pharmacokinetic properties.

[0118] 1.7.6) Antitumor effect of a bispecific antibody targeting PD-1 and VEGF in mice Human peripheral blood mononuclear cells (PBMCs) were used to reconstitute a humanized immune system in NSG mice, which were then used to generate a subcutaneous tumor model of the human lung cancer line NCI-H460. The mouse tumor model combines T cells expressing human PD-1 and human tumor cells expressing human VEGF, allowing the evaluation of the in vivo antitumor activity of bispecific antibodies targeting PD-1 and VEGF. Specifically, the experiment was performed as follows. Human non-small cell lung cancer NCI-H460 cell line (ATCC® HTB-177) cultured in vitro was used to generate a subcutaneous tumor model of the human lung cancer line NCI-H460. TM ) and the cell suspension was diluted to a concentration of 1 × 10 8 After adjusting the concentration to 1 × 10 cells / mL, it was mixed with Matrigel (BD Biosciences, product number 356234) at a volume ratio of 1:1. In addition, purchased PBMC cells (Allcells Biotech, product number PB005-C) were revived and resuspended in PBS to prepare a PBMC suspension with a concentration of 1 × 10 7 The tumor cell suspension and PBMC suspension were mixed at a volume ratio of 1:1, and 200 μL of the mixture was inoculated subcutaneously into the right back of M-NSG mice (obtained from Shanghai Southern Model Biological Research Center) under aseptic conditions. Immediately after inoculation, the inoculated mice were randomly divided into several groups according to body weight, with 10 mice per group. The mice were administered saline to the control group, 10 mg / kg of Avastin (Roche), a VEGF positive control antibody, 10 mg / kg of Opdivo (Bristol-Myers Squibb), a PD-1 positive control antibody, and 16 mg / kg of 20-Fab-0313-IgG4-V94L to the 20-Fab-0313-IgG4-V94L group. Since bispecific antibodies and monoclonal antibodies have different molecular weights, in this experiment the antibodies were converted to amounts of substance and each antibody was administered at the same amount of substance. Then, according to the above-mentioned administration schedule, the antibodies were administered twice a week for a total of 10 times, and the tumor volume was measured twice a week. Finally, the tumor growth status of each group was organized in chronological order, and a growth graph was created as shown in Figure 12.

[0119] The tumor inhibition rate was calculated according to [tumor inhibition rate = (average tumor volume in the control group - average tumor volume in the experimental group) / average tumor volume in the control group × 100%]. At the end of the experiment on day 31, the tumor inhibition rates of the Avastin group, Opdivo group, and 20-Fab-0313-IgG4-V94L group were confirmed to be 84.5%, 35.8%, and 96.6%, respectively, and a superior tumor inhibition effect was confirmed in the 20-Fab-0313-IgG4-V94L group compared to the Avastin group and Opdivo group.

[0120] Example 2: Generation of bispecific antibodies targeting PD-1 and TGF-Beta 2.1) Array US Patent Application Publication US5571714A discloses a series of TGF-β (Transforming growth factor beta) monoclonal antibodies, among which mouse-derived monoclonal antibody 1D11.16 (hereinafter also abbreviated as "1D11") can effectively bind to both TGF-β1 and TGF-β2. A hybridoma corresponding to 1D11 has already been deposited under the accession number ATCC (registered trademark) HB-9849. TM The mouse 1D11 was deposited at the American Type Culture Collection and preserved therein, and the sequence of mouse 1D11 was that described in U.S. Patent Application Publication US20180244763A1.

[0121] The amino acid sequences of the heavy and light chain variable regions of antibody 1D11 were analyzed, and the antigen complementarity determining regions and framework regions of the heavy and light chains of antibody 1D11 were determined according to the KABAT rules. The amino acid sequences of the heavy chain CDRs of antibody 1D11 were: H-CDR1 was TYWMN (SEQ ID NO: 23), H-CDR2 was QIFPASGSTNYNEMFEG (SEQ ID NO: 24), and H-CDR3 was GDGNYALDAMDY (SEQ ID NO: 25), and the amino acid sequences of the light chain CDRs were: L-CDR1 was RASESVDSYGNSFMH (SEQ ID NO: 26), L-CDR2 was LASNLES (SEQ ID NO: 27), and L-CDR3 was QQNNEDPLT (SEQ ID NO: 28).

[0122] The heavy chain variable region of mouse-derived antibody 1D11 was compared for homology with human IgG germline sequences at https: / / www.ncbi.nlm.nih.gov / igblast / , and IGHV1-46*01 was selected as a template for heavy chain CDR grafting. The heavy chain CDR of mouse-derived antibody 1D11 was grafted into the frame region of IGHV1-46*01, and WGQGTLVTVSS (sequence number 151) was added after H-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted heavy chain variable region. Similarly, the light chain variable region of mouse-derived antibody 1D11 was compared with the human IgG germline sequence for homology, IGKV7-3*01 was selected as a template for light chain CDR grafting, and the light chain CDR of mouse-derived antibody 1D11 was grafted into the frame region of IGKV7-3*01, and FGGGTKVELK (SEQ ID NO: 153) was added after L-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted light chain variable region. Furthermore, based on the CDR-grafted variable region, back mutations were introduced into amino acid sites in some of the framework regions. When introducing back mutations, KABAT editing was performed on the amino acid sequence, and the site positions were changed to KABAT numbering.

[0123] Preferably, the CDR-grafted heavy chain variable region is subjected to KABAT editing to change T at position 28 to mouse-derived I, T at position 30 to I, M at position 48 to I, V at position 67 to A, M at position 69 to L, R at position 71 to V, and V at position 78 to A, and in the CDR-grafted light chain variable region, N at position 81 is changed to D.

[0124] The DNA encoding the above humanized heavy chain variable region and light chain variable region (SEQ ID NO: 29 and SEQ ID NO: 30) was synthesized by Shanghai Synthetic Biotech Co., Ltd. The obtained humanized heavy chain variable region was linked to the heavy chain constant region (SEQ ID NO: 148) of human IgG4 (containing S228P as a site mutation in the hinge region) to obtain a full-length humanized heavy chain gene named 1D11-Hu-HC, and the humanized light chain variable region was linked to the constant region of human Kappa light chain (SEQ ID NO: 149) to obtain a full-length humanized light chain gene named 1D11-Hu-LC.

[0125] The heavy and light chain genes of the above antibodies were introduced into pcDNA3.4 expression vectors, respectively, and the antibodies were expressed and purified. The antibody obtained by combining 1D11-Hu-HC and 1D11-Hu-LC was designated as 1D11-Hu.

[0126] U.S. Patent Application Publication US20100136021A1 discloses a series of TGF-β antibodies, among which monoclonal antibody mAb12.7 (hereinafter also abbreviated as "mAb127") has the ability to effectively bind to and neutralize TGF-β1, TGF-β2, and TGF-β3. Therefore, the following experiments were performed using the heavy chain variable region and light chain variable region sequences (sequence numbers 31 and 32) of mAb127 described in US20100136021A1.

[0127] We commissioned Shanghai Synthetic Biotech Co., Ltd. to synthesize DNA encoding the above variable regions, and linked the resulting heavy chain variable region to the heavy chain constant region (sequence number 148) of human IgG4 (containing S228P as a site mutation in the hinge region) to obtain a full-length heavy chain gene that was named mAb127-HC. We also linked the light chain variable region to the human Kappa light chain constant region (sequence number 149) to obtain a full-length light chain gene that was named mAb127-LC.

[0128] According to Examples 1 to 5 of the specification of WO2018 / 137576A1, mouse-derived monoclonal antibody No. 14 was selected as the lead antibody, and amino acid sequences (sequence numbers 33 and 34) were derived based on the nucleotide sequences of the heavy chain variable region and light chain variable region.

[0129] The amino acid sequences of the heavy and light chain variable regions of antibody No. 14 were analyzed, and the antigen complementarity determining regions and framework regions of the heavy and light chains of antibody No. 14 were determined according to the KABAT rules. The amino acid sequences of the heavy chain CDRs of antibody No. 14 were as follows: H-CDR1 was GYTMN (SEQ ID NO: 35), H-CDR2 was LINPYNGDTSYNQKFKG (SEQ ID NO: 36), and H-CDR3 was WRYTMDY (SEQ ID NO: 37), and the amino acid sequences of the light chain CDRs were as follows: L-CDR1 was RASESVDNYGNSFMN (SEQ ID NO: 38), L-CDR2 was FASNLES (SEQ ID NO: 39), and L-CDR3 was QQNNEAPPT (SEQ ID NO: 40).

[0130] The heavy chain variable region of mouse-derived antibody No. 14 was compared for homology with human IgG germline sequences at https: / / www.ncbi.nlm.nih.gov / igblast / , and IGHV1-46*01 was selected as a template for heavy chain CDR grafting. The heavy chain CDR of mouse-derived antibody No. 14 was grafted into the frame region of IGHV1-46*01, and WGQGTLVTVSS (sequence number 151) was added after H-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted heavy chain variable region. Similarly, the light chain variable region of mouse-derived antibody No. 14 was compared with the human IgG germline sequence for homology, and IGKV7-3*01 was selected as a light chain CDR grafting template. The light chain CDR of mouse-derived antibody No. 14 was grafted into the frame region of IGKV7-3*01, and FGQGTKVEIK (SEQ ID NO: 152) was added after L-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted light chain variable region. Based on the CDR-grafted variable region, back mutations were introduced into amino acid sites in some of the framework regions. When introducing back mutations, KABAT editing was performed on the amino acid sequence, and the site positions were changed to KABAT numbering.

[0131] Preferably, the CDR-grafted heavy chain variable region is subjected to KABAT editing to change T at position 28 (KABAT numbering) to mouse-derived S, M at position 48 to I, V at position 67 to A, M at position 69 to V, R at position 71 to V, T at position 73 to K, and V at position 78 to A; and the CDR-grafted light chain variable region is also subjected to KABAT editing to change L at position 46 (KABAT numbering) to P, G at position 68 to R, and N at position 81 to D.

[0132] The heavy and light chain variable regions containing the above mutation sites were designated as humanized heavy chain variable region and light chain variable region (SEQ ID NO: 41 and SEQ ID NO: 42), respectively, and DNA encoding the above humanized heavy chain variable region and light chain variable region was synthesized by Shanghai Synchro Biotech Co., Ltd. The obtained humanized heavy chain variable region was linked to a human IgG4 (containing S228P as a site mutation in the hinge region) heavy chain constant region (SEQ ID NO: 148) to obtain a full-length humanized heavy chain gene, which was named 14-Hu-HC, and the humanized light chain variable region was linked to a human Kappa light chain constant region (SEQ ID NO: 149) to obtain a full-length humanized light chain gene, which was named 14-Hu-LC.

[0133] The heavy and light chain genes of the above antibodies were each introduced into a pcDNA3.4 expression vector, and the antibodies were expressed and purified. The antibody obtained by combining 14-Hu-HC and 14-Hu-LC was designated as 14-Hu.

[0134] 2.2) Selection of common light chain A comparative amino acid sequence analysis was performed using BLAST on the light chain variable regions of 14-Hu and 1D11-Hu. As a result, it was confirmed that 92% (Identities) of the amino acids were completely identical between the two, and 94% (Positives) of the amino acids were similar in physicochemical properties.

[0135] The expression vectors of 14-Hu-HC, 14-Hu-LC, 1D11-Hu-HC and 1D11-Hu-LC were combined as follows: 14-Hu-HC+14-Hu-LC, 1D11-Hu-HC+1D11-Hu-LC, 14-Hu-HC+1D11-Hu-LC and 1D11-Hu-HC+14-Hu-LC, and the antibodies were expressed and purified. The resulting antibodies were named 14-Hu, 1D11-Hu, 14-Hu-HC+1D11-Hu-LC and 1D11-Hu-HC+14-Hu-LC, respectively.

[0136] Detection by ELISA was performed as follows: A 6xHis-tagged human PD-1 extracellular domain (PD-1 extracellular domain is described in WO2018 / 137576A1) was prepared by ourselves, the resulting recombinant protein was labeled as PD1-His, and the ELISA plate (20 ng / well) was coated with the PD1-His. The ELISA plate was coated with TGF-β1 (manufactured by Sino Biological Co., Ltd.) at a coating concentration of 5 ng / well.

[0137] As shown in Figure 13A, 14-Hu effectively bound to PD1-His and inhibited EC 50 The binding activity of 1D11-Hu was 0.3924 nM, whereas 1D11-Hu, 14-Hu-HC+1D11-Hu-LC, and 1D11-Hu-HC+14-Hu-LC were unable to bind to PD1-His. As shown in FIG. 13B, 1D11-Hu effectively bound to TGF-β1 and inhibited EC 50 was 0.06624 nM, and 1D11-Hu-HC+14-Hu-LC could also bind to TGF-β1, and the EC 50 was 0.5255 nM, and 14-Hu and 14-Hu-HC+1D11-Hu-LC were unable to bind to TGF-β1. In light of these results, we decided to select 14-Hu-LC (SEQ ID NO: 43 and SEQ ID NO: 44) as the common light chain to generate a bispecific antibody.

[0138] A comparative amino acid sequence analysis using BLAST for the 14-Hu light chain variable region and the mAb127 light chain variable region confirmed that 75% (Identities) of the amino acids were completely identical between the two, and 83% (Positives) of the amino acids had similar physicochemical properties.

[0139] The expression vectors for 14-Hu-LC, mAb127-HC and mAb127-LC were combined as follows: mAb127-HC+mAb127-LC and mAb127-HC+14-Hu-LC, and the antibodies were expressed and purified. The resulting antibodies were named mAb127 and mAb127-HC+14-Hu-LC, respectively.

[0140] Using the above-mentioned ELISA method, the relative affinities of 1D11-Hu, 1D11-Hu-HC+14-Hu-LC, mAb127, and mAb127-HC+14-Hu-LC for TGF-β1 were measured, and the obtained measurement data was processed with GraphPad Prism6 to create an analytical graph, and the EC 50 was calculated.

[0141] As shown in FIG. 14, 1D11-Hu, mAb127, and mAb127-HC+14-Hu-LC all effectively bound to TGF-β1 and inhibited EC 50 The EC values ​​of mAb127 and mAb127-HC+14-Hu-LC were 0.1338 nM, 0.04136 nM, and 0.07105 nM, respectively. In addition, the EC values ​​of mAb127 and mAb127-HC+14-Hu-LC were significantly higher than those of 1D11-Hu. 50 It was confirmed that the binding signal was stronger and the affinity of both antibodies to TGF-β1 was higher. Therefore, 14-Hu-LC (SEQ ID NO: 43 and SEQ ID NO: 44) was selected as the common light chain to prepare a bispecific antibody.

[0142] 2.3) Generation of bispecific antibodies The heavy chain variable region of 14-Hu and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 1D11 was then attached via the artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain gene containing two heavy chain variable regions and two CH1 structural domains, and this was named 14-Fab-1D11-IgG4 (SEQ ID NO: 45 and SEQ ID NO: 46). Similarly, the heavy chain variable region of 1D11 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 14-Hu was then attached via the artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain gene containing two heavy chain variable regions and two CH1 structural domains, and this was designated 1D11-Fab-14-IgG4 (SEQ ID NO: 47 and SEQ ID NO: 48).

[0143] In the same manner as above, the heavy chain variable region of 14-Hu and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of mAb127 was then attached via the artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to prepare a long heavy chain gene containing two heavy chain variable regions and two CH1 structural domains, and this was named 14-Fab-127-IgG4 (SEQ ID NO: 49 and SEQ ID NO: 50). In addition, the heavy chain variable region of mAb127 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 14-Hu was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain gene containing two heavy chain variable regions and two CH1 structural domains, and this was named 127-Fab-14-IgG4 (SEQ ID NO: 51 and SEQ ID NO: 52).

[0144] The above sequences were each introduced into pcDNA3.4, and the expression vectors for 14-Fab-1D11-IgG4, 1D11-Fab-14-IgG4, 14-Fab-127-IgG4, and 127-Fab-14-IgG4 were each combined with an expression vector for 14-Hu-LC. The antibodies were expressed and purified, and the resulting antibodies were named 14-Fab-1D11-IgG4, 1D11-Fab-14-IgG4, 14-Fab-127-IgG4, and 127-Fab-14-IgG4, respectively.

[0145] 2.4) Measurement of relative affinity by ELISA method As shown in Figure 15A, 14-Hu, 14-Fab-1D11-IgG4, 1D11-Fab-14-IgG4, 14-Fab-127-IgG4, and 127-Fab-14-IgG4 all effectively bound to PD1-His and inhibited EC 50 were 0.4321 nM, 0.4367 nM, 1.996 nM, 0.3873 nM, and 3.955 nM, respectively. Among them, 1D11-Fab-14-IgG4 and 127-Fab-14-IgG4 had lower relative affinity for PD1-His than 14-Fab-1D11-IgG4 and 14-Fab-127-IgG4, which is considered to be due to the effect of steric hindrance. As shown in Figure 15B, 1D11-Hu-HC+14-Hu-LC, mAb127-HC+14-Hu-LC, 14-Fab-1D11-IgG4, 1D11-Fab-14-IgG4, 14-Fab-127-IgG4, and 127-Fab-14-IgG4 were all able to bind to TGF-β1 and inhibit EC 50 were 1.267 nM, 0.0803 nM, 0.6985 nM, 0.3628 nM, 0.1525 nM, and 0.1083 nM, respectively. Of these, it was confirmed that mAb127-HC+14-Hu-LC, 14-Fab-127-IgG4, and 127-Fab-14-IgG4 had higher relative affinity for TGF-β1 than 1D11-Hu-HC+14-Hu-LC, 14-Fab-1D11-IgG4, and 1D11-Fab-14-IgG4.

[0146] 2.5) Evaluation of MLR enhancing activity As shown in Figures 16A and 16B, both 14-Hu and 14-Fab-127-IgG4 effectively stimulated IL-2 and IFN-γ secretion from the MLR, and the EC 50 The EC values ​​for stimulating IFN-γ secretion from MLR were 0.1008 nM and 0.3185 nM, respectively. 50 were 0.04716 nM and 0.5871 nM, respectively. It was also demonstrated that, compared with 14-Hu, 14-Fab-127-IgG4 was able to further enhance IL-2 and IFN-γ secretion from the MLR at a high concentration.

[0147] 2.6) Evaluation of the binding ability of bispecific antibodies against PD-1 and TGF-Beta A microplate was coated with TGF-β1 (product number 10804-HNAC, manufactured by Sino Biological Co., Ltd.), and the antibody for measurement was serially diluted with PBST containing 1% bovine serum albumin to form a concentration gradient, and then added to the microplate and incubated at room temperature for about 30 minutes.Then, evaluation was performed in the same manner as in 1.7.4 above.

[0148] As shown in Figure 17, 14-Fab-127-IgG4 can effectively bind to human PD-1 even after binding to TGF-β1, and EC 50 was 0.2784 nM. On the other hand, 14-Hu and mAb127-HC+14-Hu-LC were unable to bind to PD-1 and TGF-β1 simultaneously.

[0149] 2.7) Antitumor effect of a bispecific antibody targeting PD-1 and TGF-beta in mice Human PD-1 transgenic mice (breed background: C57BL / 6) and MC38 mouse colon cancer cell line were purchased from Shanghai Southern Model Biological Research Center, and the mouse homologous portion was replaced in the transgenic mice with the extracellular domain of the human-derived PD-1 gene. The bispecific antibody of the present invention can recognize the PD-1 molecule of the transgenic mouse and also bind to mouse endogenous TGF-beta. Specifically, the MC38 cell line was cultured in vitro, using DMEM containing 10% serum (both serum and medium are manufactured by Gibco) as the medium. The cultured MC38 cells were then transferred to human PD-1 transgenic mice at a concentration of 2 × 10 per mouse. 6 The tumor cells were inoculated by subcutaneous injection. The volume of the inoculated tumor cells was approximately 100 mm 3 When the mice grew to 1000 mg / kg, they were randomly divided into groups, with 8 mice in each group. The mice in each group were administered saline to the control group, Keytruda (Merck) 2mg / kg or 10mg / kg, a PD-1 positive control antibody, to the Keytruda group (two administration groups), and 14-Fab-127-IgG4 3.2mg / kg or 16mg / kg to the 14-Fab-127-IgG4 group (two administration groups). Since bispecific antibodies and monoclonal antibodies have different molecular weights, in this experiment the antibodies were converted to substance amounts, and each antibody was administered at the same substance amount. Then, according to the administration schedule described above, the tumors were administered twice a week for a total of six times, and the tumor volume was measured twice a week. Finally, the tumor growth status of each group was organized according to time series, and a growth graph was created as shown in Figure 18.

[0150] The tumor inhibition rate was calculated according to [tumor inhibition rate = (average tumor volume of control group - average tumor volume of experimental group) / average tumor volume of control group x 100%]. At the end of the experiment on day 25, the tumor inhibition rates of the Keytruda group (two treatment groups) and the 14-Fab-127-IgG4 group (two treatment groups) were 79.2% (2 mg / kg), 77.7% (10 mg / kg), 85.1% (3.2 mg / kg), and 100% (16 mg / kg), respectively. Compared to the Keytruda group, the 14-Fab-127-IgG4 group showed a superior tumor inhibition effect, and it was confirmed that when 14-Fab-127-IgG4 was administered at a high dose, the tumors in all mice were completely regressed.

[0151] Example 3: Generation of bispecific antibodies targeting HER-2 and CD47 3.1) Array 19H6-Hu is a humanized monoclonal antibody targeting human HER2, and the sequences of its heavy and light chain variable regions are described in WO2020 / 025013A1. In the present invention, the humanized heavy and light chain variable regions of 19H6-Hu are designated as 19H6-Hu-VH and 19H6-Hu-VL (SEQ ID NO: 53 and SEQ ID NO: 54), respectively.

[0152] Shanghai Synthetic Biotechnology Co., Ltd. was commissioned to synthesize DNA encoding the humanized heavy chain variable region and light chain variable region of the above 19H6-Hu. The synthesized humanized heavy chain variable region was linked to the heavy chain constant region of human IgG1 (SEQ ID NO: 147) to obtain a full-length humanized heavy chain gene, which was named 19H6-Hu-HC, and the humanized light chain variable region was linked to the constant region of human Kappa chain (SEQ ID NO: 149) to obtain a full-length humanized light chain gene, which was named 19H6-Hu-LC. The genes of 19H6-Hu-HC and 19H6-Hu-LC were introduced into pcDNA3.4 expression vectors, respectively, and the antibodies were expressed and purified, and the obtained antibodies were named 19H6-Hu.

[0153] MABL-2 (hereinafter referred to as "Anti-CD47B") is a mouse-derived monoclonal antibody that targets human CD47, and the amino acid sequences of its heavy chain variable region and light chain variable region are SEQ ID NO: 12 and SEQ ID NO: 10 described in the specification of US20030108546A.

[0154] The amino acid sequences of the heavy and light chain variable regions of Anti-CD47B were analyzed, and the antigen complementarity determining regions and framework regions of the heavy and light chains of Anti-CD47B were determined according to the KABAT rules. The amino acid sequences of the heavy chain CDRs of Anti-CD47B were H-CDR1, NHVIH (SEQ ID NO: 55), H-CDR2, YIYPYNDGTKYNEKFKD (SEQ ID NO: 56), and H-CDR3, GGYYTYDD (SEQ ID NO: 57), and the amino acid sequences of the light chain CDRs of Anti-CD47B were L-CDR1, RSSQSLVHSNGKTYLH (SEQ ID NO: 58), L-CDR2, KVSNRFS (SEQ ID NO: 59), and L-CDR3, SQSTHVPYT (SEQ ID NO: 60).

[0155] The heavy chain variable region of mouse-derived Anti-CD47B was compared for homology with human IgG germline sequences at https: / / www.ncbi.nlm.nih.gov / igblast / . IGHV1-46*01 was selected as a template for grafting of heavy chain CDRs. The heavy chain CDR of mouse-derived Anti-CD47B was grafted into the frame region of IGHV1-46*01, and WGQGTLVTVSS (sequence number 151) was added after H-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted heavy chain variable region. Similarly, the light chain variable region of mouse-derived Anti-CD47B was also compared with the human IgG germline sequence for homology, IGKV2-30*01 was selected as a light chain CDR grafting template, the light chain CDR of mouse-derived Anti-CD47B was grafted into the frame region of IGKV2-30*01, and FGQGTKVEIK (SEQ ID NO: 152) was added after L-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted light chain variable region. Based on the CDR-grafted variable region, back mutations were introduced into amino acid sites in some of the framework regions. When introducing back mutations, KABAT editing was performed on the amino acid sequence, and the site positions were changed to KABAT numbering.

[0156] Preferably, the CDR-grafted heavy chain variable region is subjected to KABAT editing, changing T at position 30 of the KABAT numbering to A, M at position 69 to L, R at position 71 to S and T at position 73 to K, and the CDR-grafted light chain variable region is also subjected to KABAT editing, changing F at position 36 of the KABAT numbering to Y and R at position 46 to L.

[0157] The heavy chain variable region and light chain variable region carrying the above site mutations were designated as humanized heavy chain variable region and light chain variable region, respectively, and designated as Anti-CD47B-Hu-VH and Anti-CD47B-Hu-VL (sequence numbers 61 and 62).

[0158] The DNA encoding the humanized heavy chain variable region and light chain variable region was synthesized by Shanghai Synthetic Biotechnology Co., Ltd. The synthesized humanized heavy chain variable region was linked to the heavy chain constant region of human IgG1 (SEQ ID NO: 147) to obtain a full-length humanized heavy chain gene, which was named Anti-CD47B-Hu-HC, and the humanized light chain variable region was linked to the constant region of human Kappa chain (SEQ ID NO: 149) to obtain a full-length humanized light chain gene, which was named Anti-CD47B-Hu-LC. The genes of Anti-CD47B-Hu-HC and Anti-CD47B-Hu-LC were introduced into pcDNA3.4 expression vectors, respectively, and the antibodies were expressed and purified. The obtained antibodies were named Anti-CD47B-Hu.

[0159] 3.2) Selection of common light chain The 19H6-Hu light chain variable region was subjected to amino acid sequence analysis with the Anti-CD47B-Hu light chain variable region using BLAST. It was confirmed that 96% of the amino acids (Identities) were completely identical between the two sequences, and 99% (Positives) of the amino acids had similar physicochemical properties.

[0160] The heavy and light chain genes of 19H6-Hu and Anti-CD47B-Hu were combined as 19H6-Hu-HC+Anti-CD47B-Hu-LC, and Anti-CD47B-Hu-HC+19H6-Hu-LC, and the antibodies were expressed and purified. The resulting antibodies were named 19H6-Hu-HC+Anti-CD47B-Hu-LC and Anti-CD47B-Hu-HC+19H6-Hu-LC, respectively.

[0161] The measurement by ELISA method was carried out as follows: the His-tagged human Her-2 extracellular domain was manufactured by ACROBiosystems, with the product number HE2-H5225, and the His-tagged human CD47 extracellular domain was manufactured by Sino Biological, with the product number 12283-H08H, and these two recombinant proteins were designated as HER2-ECD-His and CD47-ECD-His, respectively. The ELISA plate was coated with HER2-ECD-His and CD47-ECD-His, respectively, at a coating concentration of 10 ng / well.

[0162] As shown in Figure 19A, both 19H6-Hu and 19H6-Hu-HC+Anti-CD47B-Hu-LC effectively bound to HER2-ECD-His and inhibited EC 50 The binding activator activity was 0.07701 nM and 0.1388 nM, respectively. On the other hand, Anti-CD47B-Hu and Anti-CD47B-Hu-HC+19H6-Hu-LC were unable to bind to HER2-ECD-His. As shown in FIG. 19B, both Anti-CD47B-Hu and Anti-CD47B-Hu-HC+19H6-Hu-LC effectively bound to CD47-ECD-His and inhibited EC 50 were 0.04276 nM and 0.0541 nM, respectively, and 19H6-Hu and 19H6-Hu-HC+Anti-CD47B-Hu-LC were unable to bind to CD47-ECD-His. In light of these results, we decided to select 19H6-Hu-LC (SEQ ID NO: 63 and SEQ ID NO: 64) as the common light chain to generate a bispecific antibody.

[0163] 3.3) Generation of bispecific antibodies The heavy chain variable region of Anti-CD47B-Hu was linked to the CH1 structural domain of human IgG4, and the heavy chain variable region of 19H6-Hu was then attached via an artificial linker, GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG1 (CH1+CH2+CH3) was attached to prepare a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named CD47B-Fab-19H6-IgG1 (sequence numbers 65 and 66). Similarly, the heavy chain variable region of 19H6-Hu was linked to the CH1 structural domain of human IgG4, and the heavy chain variable region of Anti-CD47B-Hu was then attached via the artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG1 (CH1+CH2+CH3) was attached to prepare a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 19H6-Fab-CD47B-IgG1 (sequence numbers 67 and 68).

[0164] The above sequences were each introduced into a pcDNA3.4 expression vector. Furthermore, the expression vectors for CD47B-Fab-19H6-IgG1 and 19H6-Fab-CD47B-IgG1 were each combined with an expression vector for 19H6-Hu-LC to express and purify the antibodies. The resulting antibodies were named CD47B-Fab-19H6-IgG1 and 19H6-Fab-CD47B-IgG1, respectively (for convenience, the antibodies are named by their heavy chains).

[0165] 3.4) Measurement of relative affinity by ELISA method As shown in Figure 20A, 19H6-Hu, CD47B-Fab-19H6-IgG1, and 19H6-Fab-CD47B-IgG1 all effectively bound to HER2-ECD-His and inhibited EC 50 The binding efficiencies of the antibodies were 0.1262 nM, 0.1057 nM, and 0.1543 nM, respectively, and as shown in FIG. 20B, Anti-CD47B-Hu-HC+19H6-Hu-LC, CD47B-Fab-19H6-IgG1, and 19H6-Fab-CD47B-IgG1 all effectively bound to CD47-ECD-His and inhibited EC 50were 0.06166 nM, 0.07817 nM, and 0.1945 nM, respectively. These results confirmed that CD47B-Fab-19H6-IgG1 and 19H6-Fab-CD47B-IgG1 can effectively bind to both HER-2 and CD47, and are bispecific antibodies.

[0166] Example 4: Generation of bispecific antibodies targeting HER-2 and CD137 4.1) Arrays As described in 3.1 above, 19H6-Hu is a humanized monoclonal antibody targeting human HER2. Antibody No. 94 is a mouse-derived human CD137 antibody created by the present inventors, and can be produced according to Examples 1 to 5 of WO2018 / 137576A1, but is slightly different in that mice are immunized with recombinant human CD137 protein and hybridoma cells are screened using recombinant human CD137 protein using ELISA, and the amino acid sequences of the heavy chain variable region and the light chain variable region are represented by SEQ ID NO: 69 and SEQ ID NO: 70, respectively.

[0167] The heavy and light chain variable regions of antibody No. 94 were subjected to amino acid sequence analysis, and the antigen complementarity determining regions and framework regions of the heavy and light chains of antibody No. 94 were determined according to the KABAT rules. The amino acid sequences of the heavy chain CDRs of antibody No. 94 were as follows: H-CDR1 was SYDIS (SEQ ID NO: 71), H-CDR2 was VIWTGGGTNYNSAFMS (SEQ ID NO: 72), and H-CDR3 was MDY (SEQ ID NO: 73), and the amino acid sequences of the light chain CDRs of antibody No. 94 were as follows: L-CDR1 was RSSQSLLHSNGNTYLH (SEQ ID NO: 74), L-CDR2 was KVSNRFS (SEQ ID NO: 75), and L-CDR3 was SQSTHVPWT (SEQ ID NO: 76).

[0168] The heavy chain variable region of mouse-derived antibody No. 94 was compared for homology with human IgG germline sequences at https: / / www.ncbi.nlm.nih.gov / igblast / , and IGHV4-59*01 was selected as a template for heavy chain CDR grafting. The heavy chain CDR of mouse-derived antibody No. 94 was grafted into the frame region of IGHV4-59*01, and WGQGTLVTVSS (sequence number 151) was added after H-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted heavy chain variable region. Similarly, the light chain variable region of mouse-derived antibody No. 94 was also compared with the human IgG germline sequence for homology, and IGKV2-30*01 was selected as a light chain CDR grafting template. The light chain CDR of mouse-derived antibody No. 94 was grafted into the frame region of IGKV2-30*01, and FGQGTKVEIK (SEQ ID NO: 152) was added after L-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted light chain variable region. Based on the CDR-grafted variable region, back mutations were introduced into amino acid sites in some of the framework regions. When introducing back mutations, KABAT editing was performed on the amino acid sequence, and the site positions were changed to KABAT numbering.

[0169] Preferably, the CDR-grafted heavy chain variable region is subjected to KABAT editing to change G at position 27 to F, I at position 29 to L, I at position 48 to L, R at position 71 to S, V at position 71 to K, T at position 73 to N, N at position 76 to S, F at position 78 to V and A at position 93 to V in the KABAT numbering; and the CDR-grafted light chain variable region is also subjected to KABAT editing to change F at position 36 to Y, R at position 46 to L, I at position 48 to F and Y at position 87 to F in the KABAT numbering.

[0170] The heavy chain variable region and light chain variable region carrying the above site mutations were designated as humanized heavy chain variable region and light chain variable region, respectively, and designated as 94-Hu-VH and 94-Hu-VL (sequence numbers 77 and 78).

[0171] The DNA encoding the humanized heavy chain variable region and light chain variable region was synthesized by Shanghai Synthetic Biotechnology Co., Ltd. The synthesized humanized heavy chain variable region was linked to the heavy chain constant region of human IgG1 (SEQ ID NO: 147) to obtain a full-length humanized heavy chain gene, which was named 94-Hu-HC, and the humanized light chain variable region was linked to the constant region of human Kappa chain (SEQ ID NO: 149) to obtain a full-length humanized light chain gene, which was named 94-Hu-LC. The 94-Hu-HC and 94-Hu-LC genes were introduced into pcDNA3.4 expression vectors, respectively, and the antibodies were expressed and purified, and the obtained antibodies were named 94-Hu.

[0172] 4.2) Selection of common light chain Amino acid sequence analysis of the 19H6-Hu light chain variable region and the 94-Hu light chain variable region using BLAST confirmed that 97% (Identities) of the amino acids were completely identical between the two, and 99% (Positives) of the amino acids were similar in physicochemical properties.

[0173] The heavy and light chain genes of 19H6-Hu and 94-Hu were combined as 19H6-Hu-HC+94-Hu-LC and 94-Hu-HC+19H6-Hu-LC, and the antibodies were expressed and purified. The resulting antibodies were named 19H6-Hu-HC+94-Hu-LC and 94-Hu-HC+19H6-Hu-LC, respectively.

[0174] Measurement by ELISA was performed as follows. That is, the His-tagged human Her-2 extracellular domain was manufactured by ACROBiosystems, with the product number HE2-H5225, and the Fc-tagged human CD137 extracellular domain was manufactured by Sino Biological, with the product number 10041-H02H. In the present invention, these recombinant proteins were designated as HER2-ECD-His and CD137-ECD-Fc, respectively. HER2-ECD-His and CD137-ECD-Fc were used to coat an ELISA plate, and the coating concentration was 10 ng / well for each.

[0175] As shown in Figure 21A, both 19H6-Hu and 19H6-Hu-HC+94-Hu-LC effectively bound to HER2-ECD-His and inhibited EC 50 The IgG binding sites were 0.1222 nM and 0.1391 nM, respectively, and 94-Hu and 94-Hu-HC+19H6-Hu-LC were unable to bind to HER2-ECD-His. As shown in FIG. 21B, both 94-Hu and 94-Hu-HC+19H6-Hu-LC effectively bound to CD137-ECD-Fc and EC 50 were 0.3913 nM and 0.634 nM, respectively, and 19H6-Hu and 19H6-Hu-HC+94-Hu-LC were unable to bind to CD137-ECD-Fc. In view of these results, 19H6-Hu-LC (SEQ ID NO: 63 and SEQ ID NO: 64) was selected as the common light chain to produce a bispecific antibody. In addition, a bispecific antibody can also be produced using 94-Hu-LC as the common light chain, and the present invention is not particularly limited thereto.

[0176] 4.3) Generation of bispecific antibodies A long heavy chain containing two heavy chain variable regions and two CH1 structural domains was prepared by linking the heavy chain variable region of 94-Hu and the CH1 structural domain of human IgG4, and then attaching the heavy chain variable region of 19H6-Hu via an artificial linker GGGGSGGGSGGGGGS, and finally attaching the heavy chain constant region of human IgG1 (CH1+CH2+CH3), and named 94-Fab-19H6-IgG1. Similarly, a long heavy chain containing two heavy chain variable regions and two CH1 structural domains was prepared by linking the heavy chain variable region of 19H6-Hu and the CH1 structural domain of human IgG4, and then attaching the heavy chain variable region of 94-Hu via an artificial linker GGGGSGGGSGGGGGS, and finally attaching the heavy chain constant region of human IgG1 (CH1+CH2+CH3), and named 19H6-Fab-94-IgG1.

[0177] In the present invention, in order to suppress the interaction between the Fc fragment of the above-mentioned dual antibody molecule and FcγRs (Fc gamma receptors) and to reduce potential toxicity in the body, both leucine at position 234 and leucine at position 235 of the Fc fragment were changed to alanine, and such a mutation was designated as LALA (Wang X, Mathieu M, Brezski R J., IgG Fc engineering to modulate antibody effector functions [J]. Protein & cell, 2018, 9 (1): 63-73; Ho SK, Xu Z, Thakur A, et al., Epitope and Fc-mediated Crosslinking, but not High Affinity, Are Critical for Antitumor Activity of CD137 Agonist Antibody with Reduced Liver Toxicity [J]. Molecular Cancer Therapeutics, 2020, pp. 1040-1051). The double antibody molecules into which the LALA mutation was introduced were named 94-Fab-19H6-IgG1-LALA (sequence numbers 79 and 80) and 19H6-Fab-94-IgG1-LALA (sequence numbers 81 and 82), respectively.

[0178] The above sequences were each introduced into a pcDNA3.4 expression vector. Furthermore, the expression vectors for 94-Fab-19H6-IgG1-LALA and 19H6-Fab-94-IgG1-LALA were each combined with an expression vector for 19H6-Hu-LC to express and purify the antibodies. The resulting antibodies were named 94-Fab-19H6-IgG1-LALA and 19H6-Fab-94-IgG1-LALA, respectively (for convenience, the antibodies are named by the heavy chain names herein).

[0179] 4.4) Measurement of relative affinity by ELISA method As shown in Figure 22A, 19H6-Hu, 94-Fab-19H6-IgG1-LALA, and 19H6-Fab-94-IgG1-LALA all effectively bound to HER2-ECD-His and inhibited EC 50 The binding efficiencies of 94-Hu-HC+19H6-Hu-LC and 94-Fab-19H6-IgG1-LALA were 0.1933 nM, 0.1579 nM, and 0.1201 nM, respectively, as shown in FIG. 22B. 50 The binding of 19H6-Fab-94-IgG1-LALA to CD137-ECD-Fc was weak, and the binding of 19H6-Fab-94-IgG1-LALA to CD137-ECD-Fc was weak. 50 The antibody titer of 94-Fab-19H6-IgG1-LALA was 27.56 nM. These results demonstrated that 94-Fab-19H6-IgG1-LALA and 19H6-Fab-94-IgG1-LALA were capable of binding to both HER-2 and CD137, i.e., they were bispecific antibodies.

[0180] Example 5: Generation of bispecific antibodies targeting PD-1 and CD137 5.1) Arrays mAb1-25-Hu (hereinafter also abbreviated as "609") is a humanized monoclonal antibody targeting human PD-1, and the sequences of its heavy chain variable region and light chain variable region are disclosed in WO2018 / 137576A1. The full-length heavy chain gene and full-length light chain gene of the humanized mAb1-25-Hu monoclonal antibody (609) were obtained by linking the humanized heavy chain variable region and light chain variable region (sequence numbers 83 and 84) with the heavy chain constant region (sequence number 148) and the Kappa light chain constant region (sequence number 149), respectively, of human IgG4 (S228P).

[0181] 4B4-1-1 (hereinafter also referred to as "Anti-CD137") is a mouse-derived monoclonal antibody that targets human CD137, and the amino acid sequences of its heavy chain variable region and light chain variable region are SEQ ID NO: 10 and SEQ ID NO: 11 as described in the specification of US5928893.

[0182] The amino acid sequences of the heavy and light chain variable regions of anti-CD137 were analyzed, and the antigen complementarity determining regions and framework regions of the heavy and light chains of anti-CD137 were determined according to the KABAT rules. The amino acid sequences of the heavy chain CDRs of anti-CD137 were H-CDR1 SYWMH (SEQ ID NO: 85), H-CDR2 EINPGNGHTNYNEKFKS (SEQ ID NO: 86), and H-CDR3 SFTTARGFAY (SEQ ID NO: 87), and the amino acid sequences of the light chain CDRs were L-CDR1 RASQTISDYLH (SEQ ID NO: 88), L-CDR2 YASQSIS (SEQ ID NO: 89), and L-CDR3 QDGHSFPPT (SEQ ID NO: 90).

[0183] The heavy chain variable region of mouse-derived Anti-CD137 was compared for homology with human IgG germline sequences at https: / / www.ncbi.nlm.nih.gov / igblast / , and IGHV1-46*01 was selected as a template for heavy chain CDR grafting. The heavy chain CDR of mouse-derived Anti-CD137 was grafted into the frame region of IGHV1-46*01, and WGQGTLVTVSS (sequence number 151) was added after H-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted heavy chain variable region. Similarly, the light chain variable region of mouse-derived Anti-CD137 was also compared with the human IgG germline sequence for homology, IGKV6-21*02 was selected as a light chain CDR grafting template, the light chain CDR of mouse-derived Anti-CD137 was grafted into the frame region of IGKV6-21*02, and FGQGTKVEIK (SEQ ID NO: 152) was added after L-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted light chain variable region. Based on the CDR-grafted variable region, back mutations were introduced into amino acid sites in some framework regions. When introducing back mutations, KABAT editing was performed on the amino acid sequence, and the site position was changed to KABAT numbering.

[0184] Preferably, the CDR-grafted heavy chain variable region is subjected to KABAT editing to change T at position 30 of the KABAT numbering to S, M at position 69 to L, R at position 71 to V and T at position 73 to K, and the CDR-grafted light chain variable region is also subjected to KABAT editing to change L at position 4 of the KABAT numbering to M, V at position 58 to I and T at position 69 to S.

[0185] The heavy chain variable region and light chain variable region harboring the above site mutations were designated as humanized heavy chain variable region and light chain variable region, respectively, and named Anti-CD137-Hu-VH and Anti-CD137-Hu-VL (sequence numbers 91 and 92).

[0186] Shanghai Synthetic Biotech Co., Ltd. was commissioned to synthesize DNA encoding the above humanized heavy chain variable region and light chain variable region. The synthesized humanized heavy chain variable region was linked to the heavy chain constant region of human IgG4 (S228P) (sequence number 148) to obtain a full-length humanized heavy chain gene, which was named Anti-CD137-Hu-HC. The humanized light chain variable region was linked to the constant region of human Kappa chain (sequence number 149) to obtain a full-length humanized light chain gene, which was named Anti-CD137-Hu-LC. The Anti-CD137-Hu-HC and Anti-CD137-Hu-LC genes were introduced into the pcDNA3.4 expression vector, respectively, and the antibody was expressed and purified. The obtained antibody was named Anti-CD137-Hu.

[0187] 5.2) Selection of common light chain When the 609 light chain variable region was subjected to amino acid sequence analysis with the Anti-CD137-Hu light chain variable region using BLAST, it was confirmed that 73% (Identities) of the amino acids were completely identical between the two, and 88% (Positives) of the amino acids had similar physicochemical properties.

[0188] The heavy and light chain genes of 609 and Anti-CD137-Hu were combined as 609-HC+Anti-CD137-Hu-LC and Anti-CD137-Hu-HC+609-LC, and the antibodies were expressed and purified. The resulting antibodies were named 609-HC+Anti-CD137-Hu-LC and Anti-CD137-Hu-HC+609-LC, respectively.

[0189] The ELISA assay was performed as follows: His-tagged human PD-1 extracellular domain was manufactured by Sino Biological Co., Ltd., with product number 10377-H08H, and Fc-tagged human CD137 extracellular domain was manufactured by Sino Biological Co., Ltd., with product number 10041-H02H. In the present invention, the above two recombinant proteins are designated as PD1-ECD-His and CD137-ECD-Fc, respectively, and ELISA plates were coated with PD1-ECD-His and CD137-ECD-Fc, respectively, at a coating concentration of 10 ng / well.

[0190] As shown in Figure 23A, both 609 and 609-HC+Anti-CD137-Hu-LC effectively bound to PD1-ECD-His and inhibited EC 50 The binding efficiencies of Anti-CD137-Hu and Anti-CD137-Hu-HC+609-LC were 0.1358 nM and 0.2067 nM, respectively, and Anti-CD137-Hu and Anti-CD137-Hu-HC+609-LC were unable to bind to PD1-ECD-His. As shown in Figure 23B, Anti-CD137-Hu effectively bound to CD137-ECD-Fc and EC 50 The binding affinity of 609, 609-HC+Anti-CD137-Hu-LC, and Anti-CD137-Hu-HC+609-LC to CD137-ECD-Fc was 0.461 nM, and 609, 609-HC+Anti-CD137-Hu-LC, and Anti-CD137-Hu-HC+609-LC were unable to bind to CD137-ECD-Fc. Therefore, we decided to prepare a bispecific antibody using Anti-CD137-Hu-LC (SEQ ID NO: 93 and SEQ ID NO: 94) as the common light chain.

[0191] 5.3) Generation of bispecific antibodies The heavy chain variable region of 609 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of Anti-CD137-Hu was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 609-Fab-137-IgG4 (SEQ ID NO: 95 and SEQ ID NO: 96). Similarly, the heavy chain variable region of Anti-CD137-Hu and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 609 was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 137-Fab-609-IgG4 (SEQ ID NO: 97 and SEQ ID NO: 98).

[0192] The above sequences were each introduced into a pcDNA3.4 expression vector. Furthermore, the expression vectors for 609-Fab-137-IgG4 and 137-Fab-609-IgG4 were each combined with an expression vector for Anti-CD137-Hu-LC to express and purify the antibodies. The resulting antibodies were named 609-Fab-137-IgG4 and 137-Fab-609-IgG4, respectively (for convenience, the antibodies are named by the heavy chain name here).

[0193] 5.4) Measurement of relative affinity by ELISA As shown in Figure 24A, 609-HC+Anti-CD137-Hu-LC, 609-Fab-137-IgG4, and 137-Fab-609-IgG4 all effectively bound to PD1-ECD-His and inhibited EC 50 The binding efficiencies of Anti-CD137-Hu, 609-Fab-137-IgG4, and 137-Fab-609-IgG4 were 0.2067 nM, 0.2293 nM, and 1.415 nM, respectively. As shown in FIG. 24B, Anti-CD137-Hu, 609-Fab-137-IgG4, and 137-Fab-609-IgG4 all effectively bound to CD137-ECD-Fc and inhibited EC 50were 0.461 nM, 0.3572 nM, and 0.2424 nM, respectively. This confirmed that 609-Fab-137-IgG4 and 137-Fab-609-IgG4 were able to bind to CD137 in addition to PD-1, and were therefore bispecific antibodies.

[0194] Example 6: Generation of bispecific antibodies targeting PD-1 and CD40 6.1) Arrays 609 is a humanized monoclonal antibody targeting human PD-1, and the details can be seen in 5.1 above. MAb2.220 (hereinafter also referred to as "Anti-CD40") is a mouse-derived monoclonal antibody targeting human CD40, and the amino acid sequences of its heavy chain variable region and light chain variable region are SEQ ID NO: 2 and SEQ ID NO: 1 described in the specification of US6312693.

[0195] The amino acid sequences of the heavy and light chain variable regions of the anti-CD40 antibody were analyzed, and the antigen complementarity determining regions and framework regions of the heavy and light chains of anti-CD40 were determined according to the KABAT rules. The amino acid sequences of the heavy chain CDRs of anti-CD40 were H-CDR1 TTGMQ (SEQ ID NO: 99), H-CDR2 WINTHSGVPKYVEDFKG (SEQ ID NO: 100), and H-CDR3 SGNGNYDLAYFAY (SEQ ID NO: 101), and the amino acid sequences of the light chain CDRs were L-CDR1 RASQSISDYLH (SEQ ID NO: 102), L-CDR2 YASHSIS (SEQ ID NO: 103), and L-CDR3 QHGHSFPWT (SEQ ID NO: 104).

[0196] The heavy chain variable region of mouse-derived Anti-CD40 was compared for homology with human IgG germline sequences at https: / / www.ncbi.nlm.nih.gov / igblast / . IGHV7-4-1*02 was selected as a template for grafting of heavy chain CDRs, and the heavy chain CDR of mouse-derived Anti-CD40 was grafted into the frame region of IGHV7-4-1*02. WGQGTLVTVSS (sequence number 151) was added after H-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted heavy chain variable region. Similarly, the light chain variable region of mouse-derived Anti-CD40 was also compared with the human IgG germline sequence for homology, IGKV3-11*01 was selected as a light chain CDR grafting template, the light chain CDR of mouse-derived Anti-CD40 was grafted into the frame region of IGKV3-11*01, and FGQGTKVEIK (SEQ ID NO: 152) was added after L-CDR3 to form the fourth framework region, thereby obtaining the sequence of the CDR-grafted light chain variable region. Based on the CDR-grafted variable region, back mutations were introduced into amino acid sites in some framework regions, and when introducing mutations, KABAT editing was performed on the amino acid sequence, and the position of the site was changed to KABAT numbering.

[0197] Preferably, the variable region of the CDR-grafted heavy chain is subjected to KABAT editing to change V at position 2 of the KABAT numbering to I, T at position 28 to A, Q at position 39 to E, M at position 48 to I, S at position 76 to N and A at position 93 to V, and the variable region of the CDR-grafted light chain is also subjected to KABAT editing to change A at position 43 to S, Y at position 49 to K and T at position 69 to S, in the KABAT numbering.

[0198] The heavy chain variable region and light chain variable region harboring the above site mutations were designated as humanized heavy chain variable region and light chain variable region, respectively, and named Anti-CD40-Hu-VH and Anti-CD40-Hu-VL (sequence numbers 105 and 106).

[0199] The DNA encoding the humanized heavy chain variable region and light chain variable region was synthesized by Shanghai Synchro Biotech Co., Ltd. The synthesized humanized heavy chain variable region was linked to the heavy chain constant region of human IgG4 (S228P) (SEQ ID NO: 148) to obtain a full-length humanized heavy chain gene, which was named Anti-CD40-Hu-HC. The humanized light chain variable region was linked to the constant region of human Kappa chain (SEQ ID NO: 149) to obtain a full-length humanized light chain gene, which was named Anti-CD40-Hu-LC. The Anti-CD40-Hu-HC gene and the Anti-CD40-Hu-LC gene were introduced into the pcDNA3.4 expression vector, respectively, and the antibody was expressed and purified. The obtained antibody was named Anti-CD40-Hu.

[0200] 6.2) Selection of common light chain The 609 light chain variable region was subjected to amino acid sequence analysis with the Anti-CD40-Hu light chain variable region using BLAST. It was confirmed that 90% of the amino acids (Identities) were completely identical between the two sequences, and 96% (Positives) were similar in physicochemical properties.

[0201] The heavy and light chain genes of 609 and Anti-CD40-Hu were combined as follows: 609-HC+Anti-CD40-Hu-LC and Anti-CD40-Hu-HC+609-LC. The antibodies were expressed and purified. The resulting antibodies were named 609-HC+Anti-CD40-Hu-LC and Anti-CD40-Hu-HC+609-LC, respectively.

[0202] The ELISA assay was performed as follows: His-tagged human PD-1 extracellular domain was manufactured by Sino Biological Co., Ltd., with product number 10377-H08H, and His-tagged human CD40 extracellular domain was manufactured by Sino Biological Co., Ltd., with product number 10774-H08H. In the present invention, these two recombinant proteins are designated as PD1-ECD-His and CD40-ECD-His, respectively, and ELISA plates were coated with PD1-ECD-His and CD40-ECD-His, respectively, at a coating concentration of 10 ng / well.

[0203] As shown in Figure 25A, both 609 and 609-HC+Anti-CD40-Hu-LC effectively bound to PD1-ECD-His and inhibited EC 50 The binding efficiencies of Anti-CD40-Hu and Anti-CD40-Hu-HC+609-LC were 0.1263 nM and 0.1387 nM, respectively, and Anti-CD40-Hu and Anti-CD40-Hu-HC+609-LC were unable to bind to PD1-ECD-His. As shown in FIG. 25B, Anti-CD40-Hu effectively bound to CD40-ECD-His and inhibited EC 50 The binding affinity of the antibody was 0.1104 nM, and none of 609, 609-HC+Anti-CD40-Hu-LC, and Anti-CD40-Hu-HC+609-LC could bind to CD40-ECD-His. Therefore, we decided to prepare a bispecific antibody using Anti-CD40-Hu-LC (SEQ ID NO: 107 and SEQ ID NO: 108) as the common light chain.

[0204] 6.3) Generation of bispecific antibodies The heavy chain variable region of 609 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of Anti-CD40-Hu was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 609-Fab-40-IgG4 (SEQ ID NO: 109 and SEQ ID NO: 110). Similarly, the heavy chain variable region of Anti-CD40-Hu and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 609 was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 40-Fab-609-IgG4 (SEQ ID NO: 111 and SEQ ID NO: 112).

[0205] The above sequences were each introduced into a pcDNA3.4 expression vector. Furthermore, the expression vectors for 609-Fab-40-IgG4 and 40-Fab-609-IgG4 were each combined with an expression vector for Anti-CD40-Hu-LC to express and purify the antibodies. The resulting antibodies were named 609-Fab-40-IgG4 and 40-Fab-609-IgG4, respectively (for convenience, the antibodies are named by the heavy chain name here).

[0206] 6.4) Relative affinity measurement by ELISA As shown in Figure 26A, 609-HC+Anti-CD40-Hu-LC, 609-Fab-40-IgG4, and 40-Fab-609-IgG4 all effectively bound to PD1-ECD-His and inhibited EC 50 The binding efficiencies of Anti-CD40-Hu, 609-Fab-40-IgG4, and 40-Fab-609-IgG4 were 0.1387 nM, 0.1723 nM, and 1.017 nM, respectively, and as shown in FIG. 26B, Anti-CD40-Hu, 609-Fab-40-IgG4, and 40-Fab-609-IgG4 all effectively bound to CD40-ECD-His and inhibited EC 50were 0.1104 nM, 0.1047 nM, and 0.09556 nM, respectively. This confirmed that 609-Fab-40-IgG4 and 40-Fab-609-IgG4 were capable of binding to CD40 in addition to PD-1, and were therefore bispecific antibodies.

[0207] Example 7: Generation of bispecific antibodies targeting PD-1 and other protein molecules 7.1) Arrays 609 is a humanized monoclonal antibody targeting human PD-1, and for details, see 5.1 above. On the other hand, the amino acid sequences of the heavy chain variable region and light chain variable region of antibody preparations such as cetuximab, bevacizumab, trastuzumab, and pertuzumab (SEQ ID NOs: 1-2 and 113-118) were already reported in the literature (Magdelaine-Beuzelin C, Kaas Q, Wehbi V, et al., Structure-function relationships of the variable domains of monoclonal antibodies approved for cancer treatment[J]. Critical reviews in oncology / hematology, 2007, 64(3):210-225). In addition, 10D1 (hereinafter also referred to as "Ipilimumab") is a monoclonal antibody that targets human CTLA-4, and the amino acid sequences of its heavy chain variable region and light chain variable region are SEQ ID NO: 17 and SEQ ID NO: 7 (in the present invention, SEQ ID NO: 119 and SEQ ID NO: 120) described in the specification of US20020086014A1.

[0208] The heavy chain variable region and the light chain variable region were synthesized by Shanghai Synchro Biotech Co., Ltd. The sequences encoding the heavy chain variable region and the light chain variable region were linked to the heavy chain constant region of human IgG1 (SEQ ID NO: 147) and the human Kappa light chain constant region (SEQ ID NO: 149), respectively, to prepare the full-length heavy chain gene and the full-length light chain gene of the antibody. The heavy chain genes of the antibody were named Cetuximab-HC, Bevacizumab-HC, Trastuzumab-HC, Pertuzumab-HC, and Ipilimumab-HC, respectively, and the light chain genes of the antibody were named Cetuximab-LC, Bevacizumab-LC, Trastuzumab-LC, Pertuzumab-LC, and Ipilimumab-LC, respectively. The above heavy chain genes and light chain genes were each introduced into a pcDNA3.4 expression vector, and the heavy chain and light chain genes with corresponding names were combined to express and purify antibodies. The obtained antibodies were named Cetuximab-IgG1, Bevacizumab-IgG1, Trastuzumab-IgG1, Pertuzumab-IgG1, and Ipilimumab-IgG1, respectively.

[0209] 5E7-Hu is a humanized antibody targeting human LAG-3, and the sequences of its heavy and light chain variable regions were SEQ ID NO: 26 and SEQ ID NO: 28 (SEQ ID NO: 121 and SEQ ID NO: 122 in the present invention) described in the specification of PCT / CN2020 / 076023. Shanghai Synthetic Biotech Co., Ltd. was commissioned to synthesize DNA encoding the above humanized heavy and light chain variable regions. The synthesized humanized heavy chain variable region was linked to the human IgG4 (S228P) heavy chain constant region (SEQ ID NO: 148) to obtain a full-length humanized heavy chain gene, which was named 5E7-Hu-HC. In addition, the humanized light chain variable region was linked to the human Kappa chain constant region (SEQ ID NO: 149) to obtain a full-length humanized light chain gene, which was named 5E7-Hu-LC. The 5E7-Hu-HC gene and the 5E7-Hu-LC gene were each introduced into a pcDNA3.4 expression vector, and the antibody was expressed and purified. The resulting antibody was newly named 5E7-Hu.

[0210] 7.2) Selection of common light chain 7.2.1) Measurement of hybrid antibody-antigen binding affinity The heavy chain gene of 609 was combined with the light chain genes of the above antibodies, and the antibodies were expressed and purified. The resulting antibodies were named 609-HC+Cetuximab-LC, 609-HC+Bevacizumab-LC, 609-HC+Trastuzumab-LC, 609-HC+Pertuzumab-LC, 609-HC+Ipilimumab-LC and 609-HC+5E7-Hu-LC, respectively.

[0211] The ELISA assay was performed as follows: His-tagged human PD-1 extracellular domain was manufactured by Sino Biological Co., Ltd., with the product number 10377-H08H, and the recombinant protein was designated as PD1-ECD-His in the present invention. PD1-ECD-His was used to coat the ELISA plate, and the coating concentration was 10 ng / well. As shown in Figure 27, Opdivo (Bristol-Myers Squibb), 609, 609-HC+Cetuximab-LC, 609-HC+Bevacizumab-LC, 609-HC+Pertuzumab-LC, 609-HC+Ipilimumab-LC and 609-HC+5E7-Hu-LC all effectively bound to PD1-ECD-His and inhibited EC 50 The affinity of 609-HC+Trastuzumab-LC to PD1-ECD-His was 0.2887 nM, 0.1100 nM, 0.2424 nM, 0.1530 nM, 0.2244 nM, 0.1547 nM, and 0.1709 nM, respectively. On the other hand, 609-HC+Trastuzumab-LC had a weaker relative affinity to PD1-ECD-His and a higher EC 50was 0.7219 nM. This shows that related bispecific antibodies can be prepared using Cetuximab-LC (SEQ ID NO: 123 and SEQ ID NO: 124), Bevacizumab-LC, Pertuzumab-LC (SEQ ID NO: 125 and SEQ ID NO: 126), Ipilimumab-LC (SEQ ID NO: 127 and SEQ ID NO: 128) and 5E7-Hu-LC (SEQ ID NO: 129 and SEQ ID NO: 130) as a common light chain. In this case, a human IgG4 antibody that does not bind to PD-1 was used as the allogeneic control antibody.

[0212] 7.2.2) Evaluation of the inhibitory effect of hybrid antibodies on the interaction between PD-1 and PD-L1 The inhibitory effect of antibodies on the interaction between PD-1 and PD-L1 was evaluated as follows. Human Fc-tagged human PD-1 and human PD-L1 extracellular domains were prepared by ourselves with reference to the specification of WO2018 / 137576A1, and these two recombinant proteins were designated as PD1-ECD-hFc and PD-L1-ECD-hFc, respectively. Biotinylation reagent, biotin n-hydroxysuccinimide ester (Sigma, product number / specification H1759-100MG), was prepared in anhydrous DMSO to a mother solution of 100 mM, and PD-L1-ECD-hFc was converted to a substance amount concentration based on the molecular weight and solution concentration. An appropriate amount of PD-L1-ECD-hFc fusion protein solution was taken, and the substance amount was calculated, and then the solution was mixed uniformly with biotin n-hydroxysuccinimide ester at a ratio of 1:20, and allowed to stand at room temperature for 1 hour for labeling. Dialysis was performed and the protein concentration was measured using UV spectroscopy. Human PD-1-ECD-hFc was then diluted with coating buffer to a concentration of 2 μg / mL, pipetted into a 96-well ELISA plate at 100 μL per well, and incubated at room temperature for 4 hours. After rinsing once with PBST, 200 μL of PBST containing 1% BSA was added to each well and incubated at room temperature for 2 hours for blocking. The plate was tapped to remove the blocking solution and temporarily stored at 4°C for further processing. In the 96-well plate, biotinylated PD-L1-ECD-hFc was diluted with PBST containing 1% BSA (bovine serum albumin) to a concentration of 500 ng / mL, and the diluted biotinylated fusion protein was used to stepwise dilute the PD-1 antibody to form a concentration gradient. The mixture of the diluted antibody and biotinylated fusion protein was added to the ELISA plate coated with human PD-1-ECD-hFc and incubated at room temperature for 1 hour. The plate was washed three times with PBST, and HRP-labeled streptavidin (BD Biosciences) diluted 1:1000 in PBST containing 1% BSA was added and incubated at room temperature for 45 minutes. The plate was then washed three times with PBST.100 μL of the staining solution using TMB as a substrate was added to each well, and after incubation at room temperature for 1 to 5 minutes, 50 μL of the stop solution was added to stop the color reaction. The OD was measured using a plate reader. 450 The values ​​are read out, the measurement data is processed in GraphPad Prism6 to create an analysis graph, and the IC 50 was calculated.

[0213] As shown in Figure 28, Opdivo, 609, 609-HC+Cetuximab-LC, 609-HC+Bevacizumab-LC, 609-HC+Trastuzumab-LC, 609-HC+Pertuzumab-LC, 609-HC+Ipilimumab-LC and 609-HC+5E7-Hu-LC could all effectively inhibit the interaction between PD-1 and PD-L1, and showed a significant increase in IC 50 were 0.05729 nM, 0.1309 nM, 0.1199 nM, 0.1191 nM, 0.1162 nM, 0.09876 nM, 0.1052 nM, and 0.1312 nM, respectively. At this time, a human IgG4 antibody that does not bind to PD-1 was used as the allogeneic control antibody.

[0214] 7.2.3) Evaluation of the enhancing effect of hybrid antibodies on mixed lymphocyte reaction Furthermore, the enhancing effect of the above antibodies on the mixed lymphocyte reaction was measured. As shown in Figure 29, 609, 609-HC+Cetuximab-LC, 609-HC+Bevacizumab-LC, 609-HC+Pertuzumab-LC, 609-HC+Ipilimumab-LC and 609-HC+5E7-Hu-LC were all able to enhance the mixed lymphocyte reaction and promote IL-2 secretion, and EC 50 were 0.08623 nM, 0.2510 nM, 0.1211 nM, 0.5171 nM, 0.2040 nM, and 0.09101 nM, respectively. On the other hand, 609-HC + Trastuzumab-LC showed no stimulatory effect on mixed lymphocyte reaction and IL-2 secretion. In this experiment, a human IgG4 antibody that does not bind to PD-1 was used as the allogeneic control antibody.

[0215] 7.2.4) Evaluation of the binding activity of hybrid antibodies to cell surface PD-1 The binding ability of the hybrid antibody to cell surface PD-1 was evaluated using flow cytometry. The TF-1 cell line expressing PD-1 was established as follows. Specifically, the full-length gene sequence of human PD-1 was the sequence registered as Q15116 in UniProt, and the gene sequence was introduced into a lentivirus expression vector, pLVX-Puro (manufactured by Clontech). Using Lipofectamine 3000 (manufactured by Thermo Fisher Scientific, product number L3000001), pLVX-Puro containing a lentivirus packaging vector and a target gene was introduced into HEK293FT cells (manufactured by Thermo Fisher Scientific, product number R70007) and incubated for 48 hours. The cell culture solution was centrifuged to collect the supernatant, which was then filtered through a 0.45 μm filter membrane to remove cell debris. The supernatant containing the above viral particles was used to transfect TF-1 cells (manufactured by ATCC, product number CRL-2003). TM ) and treated the cells with puromycin 48 hours later. Cell lines that stably expressed the target gene were selected by screening. The TF-1 cell line that stably expressed PD-1 was named TF1-PD1.

[0216] The binding of the antibody to the cells was measured using flow cytometry as follows. Specifically, TF1-PD1 cells were seeded at 200,000 cells per well in a 96-well round-bottom plate, centrifuged to discard the supernatant, and serially diluted antibodies were added and incubated at room temperature for about 30 minutes. The cells were washed twice with PBS, centrifuged to discard the supernatant, and appropriately diluted Anti-Human IgG-FITC antibodies (Fc-specific, Sigma, product number F9512) were added to each well and incubated at room temperature for about 30 minutes, after which the cells were washed twice with PBS. The supernatant was discarded, and the cells were fixed by adding Fixation Buffer I (BD Biosciences) and incubating at room temperature for 5 minutes, after which the cells were washed twice with PBS, and finally the cells were resuspended in 200 μL of PBS. The fluorescence intensity of the FITC channel was measured using a flow cytometer, and the measurement data was processed using the software installed in the flow cytometer and then converted into Excel data. The Excel data was then processed using GraphPad Prism6 to create an analysis graph and calculate the EC 50 was calculated.

[0217] As shown in Figure 30, 609, 609-HC+Cetuximab-LC, 609-HC+Bevacizumab-LC, 609-HC+Pertuzumab-LC, 609-HC+Ipilimumab-LC and 609-HC+5E7-Hu-LC all effectively bound to PD-1 on the cell surface and inhibited EC 50 The binding efficiencies of 609-HC+Trastuzumab-LC to TF1-PD1 were 0.3761 nM, 0.577 nM, 0.5193 nM, 0.4302 nM, 0.4773 nM, and 0.3864 nM, respectively. Among them, 609-HC+Trastuzumab-LC showed significantly reduced binding to TF1-PD1 compared to the other hybrid antibodies. In this experiment, a human IgG4 antibody that does not bind to PD-1 was used as the allogeneic control antibody.

[0218] 7.2.5) Effect of 609 light chain variable region CDRs on 609 binding to PD-1 The above experimental results demonstrate that the hybrid antibody, which combines the heavy chain of 609 with another antibody light chain, can also effectively bind to the PD-1 molecule, inhibit the interaction between PD-1 and PD-L1, and enhance the binding to PD-1 on the cell surface and in the mixed lymphocyte reaction. In this experiment, the effect of the 609 light chain variable region CDR on the binding of 609 to PD-1 was evaluated using alanine scanning technology. Specifically, the amino acid residues in the 609 light chain CDR were each converted to alanine residues, while the alanine residues inherent in the CDR were left unchanged. The 609 heavy chain was combined with each of the light chain mutants, and the antibodies were expressed and purified according to the method described above, and the relative affinity of the antibodies to PD-1 was measured using the ELISA method described above. The measurement results are summarized in Table 3 below. In Table 3, R24A in 609-HC+609-LC-R24A means that the arginine residue at position 24 was converted to an alanine residue, and in this experiment, the positions of the amino acid residues are also indicated by the KABAT numbering.

[0219] [Table 2]

[0220] As shown in Figure 31 and Table 4, the results of the alanine scanning experiment showed that converting each of the amino acid residues in the light chain CDR to alanine residues had no significant effect on the binding of the antibody to PD-1, demonstrating that 609 mainly uses the heavy chain to bind to the PD-1 molecule, and the binding of 609 to PD-1 is not significantly dependent on the light chain of 609.

[0221] 7.3) Generation of bispecific antibodies targeting PD-1 and EGFR The heavy chain variable region of 609 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of Cetuximab-IgG1 was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached, to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 609-Fab-Cetuximab-IgG4 (SEQ ID NO: 131 and SEQ ID NO: 132).

[0222] Each of the above sequences was introduced into a pcDNA3.4 expression vector, and the 609-Fab-Cetuximab-IgG4 expression vector and the Cetuximab-LC expression vector were combined to express and purify the antibody. The resulting antibody was named 609-Fab-Cetuximab-IgG4 (for convenience, the antibody is named by the heavy chain name here).

[0223] The ELISA assay was performed as follows: His-tagged human PD-1 extracellular domain was manufactured by Sino Biological Co., Ltd., product number 10377-H08H), and His-tagged human EGFR extracellular domain was also manufactured by Sino Biological Co., Ltd., product number 10001-H08H. In the present invention, these two recombinant proteins are designated as PD1-ECD-His and EGFR-ECD-His, and PD1-ECD-His and EGFR-ECD-His were used to coat ELISA plates at coating concentrations of 10 ng / well and 20 ng / well, respectively.

[0224] As shown in Figure 32A, both 609-HC+Cetuximab-LC and 609-Fab-Cetuximab-IgG4 effectively bound to PD1-ECD-His and inhibited EC 50 The binding efficiencies of Cetuximab-IgG1 and 609-Fab-Cetuximab-IgG4 were 0.7172 nM and 0.2616 nM, respectively. As shown in FIG. 32B, both Cetuximab-IgG1 and 609-Fab-Cetuximab-IgG4 effectively bound to EGFR-ECD-His and inhibited EC 50and 0.07609 nM and 0.09327 nM, respectively. These results demonstrated that 609-Fab-Cetuximab-IgG4 can bind to EGFR in addition to PD-1, and is a bispecific antibody.

[0225] 7.4) Generation of bispecific antibodies targeting PD-1 and HER-2 The heavy chain variable region of 609 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of Pertuzumab-IgG1 was attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached, to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 609-Fab-Pertuzumab-IgG4 (sequence numbers 133 and 134).

[0226] Each of the above sequences was introduced into a pcDNA3.4 expression vector, and the expression vector for 609-Fab-Pertuzumab-IgG4 and the expression vector for Pertuzumab-LC were combined to express and purify the antibody, which was named 609-Fab-Pertuzumab-IgG4 (for convenience, the antibody is named by the heavy chain name here).

[0227] The ELISA assay was performed as follows: His-tagged human PD-1 extracellular domain was manufactured by Sino Biological Co., Ltd., with product number 10377-H08H, and His-tagged human Her-2 extracellular domain was manufactured by ACROBiosystems Co., Ltd., with product number HE2-H5225. In the present invention, these two recombinant proteins are designated as PD1-ECD-His and HER2-ECD-His, and PD1-ECD-His and HER2-ECD-His were used to coat the ELISA plate, respectively, at a coating concentration of 10 ng / well.

[0228] As shown in Figure 33A, both 609-HC+Pertuzumab-LC and 609-Fab-Pertuzumab-IgG4 effectively bound to PD1-ECD-His and inhibited EC 50 The IgGl and 609-Fab-Pertuzumab-IgG4 binding domains were 0.1422 nM and 0.1196 nM, respectively. As shown in FIG. 33B, both Pertuzumab-IgG1 and 609-Fab-Pertuzumab-IgG4 effectively bound to HER2-ECD-His and inhibited EC 50 were 0.5352 nM and 2.616 nM, respectively. These results demonstrated that 609-Fab-Pertuzumab-IgG4 can bind to HER-2 in addition to PD-1, and is a bispecific antibody.

[0229] 7.5) Creation of bispecific antibodies targeting PD-1 and CTLA-4 7.5.1) Generation of bispecific antibodies The heavy chain variable region of 609 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of Ipilimumab-IgG1 was then attached via the artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG1 (CH1+CH2+CH3) was attached to create a long heavy chain gene containing two heavy chain variable regions and two CH1 structural domains, and this gene was named 609-Fab-Ipilimumab-IgG1 (SEQ ID NO: 135 and SEQ ID NO: 136). Similarly, the heavy chain variable region of Ipilimumab-IgG1 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 609 was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG1 (CH1+CH2+CH3) was attached to create a long heavy chain gene containing two heavy chain variable regions and two CH1 structural domains, and this gene was named Ipilimumab-Fab-609-IgG1 (sequence numbers 137 and 138).

[0230] The heavy chain variable region of 609 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of Ipilimumab-IgG1 was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached, to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 609-Fab-Ipilimumab-IgG4 (SEQ ID NO: 139 and SEQ ID NO: 140). Similarly, the heavy chain variable region of Ipilimumab-IgG1 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 609 was attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named Ipilimumab-Fab-609-IgG4 (SEQ ID NO: 141 and SEQ ID NO: 142).

[0231] The above sequences were each introduced into a pcDNA3.4 expression vector. Furthermore, the expression vectors for 609-Fab-Ipilimumab-IgG1, Ipilimumab-Fab-609-IgG1, 609-Fab-Ipilimumab-IgG4 and Ipilimumab-Fab-609-IgG4 were each combined with an expression vector for Ipilimumab-LC to express and purify the antibodies. The resulting antibodies were named 609-Fab-Ipilimumab-IgG1, Ipilimumab-Fab-609-IgG1, 609-Fab-Ipilimumab-IgG4 and Ipilimumab-Fab-609-IgG4 (for convenience, the antibodies are named by the heavy chain name here), respectively.

[0232] Measurement by ELISA was performed as follows: His-tagged human PD-1 extracellular domain was manufactured by Sino Biological Co., Ltd., product number 10377-H08H), and human Fc-tagged human CTLA-4 extracellular domain was manufactured by Sino Biological Co., Ltd., product number 11159-H31H5. In the present invention, these two recombinant proteins are designated as PD1-ECD-His and CTLA4-ECD-Fc, and PD1-ECD-His and CTLA4-ECD-Fc were used to coat the ELISA plate, respectively, at a coating concentration of 10 ng / well.

[0233] As shown in Figure 34A, 609-HC+Ipilimumab-LC, 609-Fab-Ipilimumab-IgG1, Ipilimumab-Fab-609-IgG1, 609-Fab-Ipilimumab-IgG4, and Ipilimumab-Fab-609-IgG4 all effectively bound to PD1-ECD-His and inhibited EC 50 The IgGl, 609-Fab-Ipilimumab-IgGl, Ipilimumab-Fab-609-IgGl, 609-Fab-Ipilimumab-IgG4 and Ipilimumab-Fab-609-IgG4 all effectively bound to CTLA4-ECD-Fc and inhibited the EC 50 were 0.8354 nM, 2.123 nM, 0.3376 nM, 2.626 nM, and 0.392 nM, respectively. These results demonstrated that 609-Fab-Ipilimumab-IgG1, Ipilimumab-Fab-609-IgG1, 609-Fab-Ipilimumab-IgG4, and Ipilimumab-Fab-609-IgG4 are capable of binding to CTLA-4 in addition to PD-1, and are therefore bispecific antibodies.

[0234] 7.5.2) Bioactivity Evaluation of Bispecific Antibodies Targeting PD-1 and CTLA-4 RPMI1640 complete medium was prepared by adding 10% fetal bovine serum, 1% MEM non-essential amino acid solution, 1% sodium pyruvate, 1 HEPES, 0.1% 2-mercaptoethanol, 1% penicillin-streptomycin, and 1% GlutaMAX supplement to RPMI1640, all of which were from Thermo Fisher Scientific. Freshly isolated PBMC cells (Allcells Biotech, product number PB005-C) were washed and resuspended in the RPMI1640 complete medium, and a certain amount of staphylococcal enterotoxin B (SEB) was added as a superantigen. The amino acid sequence of the superantigen is registered at https: / / www.uniprot.org / uniprot / P01552 and was prepared by the inventor himself using E. coli. PBMC cell suspension was seeded in a 96-well round-bottom cell culture plate at 150 μL per well at 200,000 cells, and 50 μL of serially diluted antibodies were added and incubated in a 37°C incubator for 4 days. An appropriate amount of supernatant was taken from the 96-well plate, and the amount of IL-2 in the supernatant was measured using a sandwich ELISA method according to the standard operating manual to detect the amount of IL-2 secretion. The paired antibodies used in this experiment were manufactured by BD Biosciences, and were measured using a SpectraMax 190 plate reader at OD . 450 The measured values ​​were then processed using GraphPad Prism6 to create an analytical graph, and the EC 50 was calculated.

[0235] [Table 3]

[0236] As shown in Figure 35 and Table 4 above, 609, 609-HC+Ipilimumab-LC and Ipilimumab-IgG1 were significantly increased in EC 50and activity values ​​(Top value) were particularly high, and it was confirmed that these three antibodies have equivalent bioactivity. In addition, the bioactivity of bispecific antibodies targeting PD-1 and CTLA-4 is, in descending order, 609-Fab-Ipilimumab-IgG1>Ipilimumab-Fab-609-IgG1>609-Fab-Ipilimumab-IgG4>Ipilimumab-Fab-609-IgG4, and it was confirmed that 609-Fab-Ipilimumab-IgG1 and Ipilimumab-Fab-609-IgG1 clearly exceed the bioactivity of monoclonal antibodies 609 and Ipilimumab-IgG1, and hybrid antibody 609-HC+Ipilimumab-LC, and that the bioactivity is at the same level as when 609-HC+Ipilimumab-LC and Ipilimumab-IgG1 are used in combination. In addition, in Table 5 above, "609-HC+Ipilimumab-LC / Ipilimumab-IgG1" refers to the case where the two antibodies were used in combination at a substance ratio of 1:1.

[0237] 7.5.3) Evaluation of ADCC activity of bispecific antibodies targeting PD-1 and CTLA-4 Antibody-dependent cell-mediated cytotoxicity (ADCC) is a universal function of human IgG antibodies, and ADCC activity is closely related to the antibody subtype. Bispecific antibodies of IgG1 subtype targeting PD-1 and CTLA-4 may exhibit potential cytotoxicity against cells expressing PD-1. Therefore, the ADCC activity was measured using freshly isolated PBMC cells (Allcells Biotech, Catalog No. PB005-C) as effector cells and TF-1 cells expressing PD-1 as target cells.

[0238] As shown in FIG. 36, the IgG4 allogeneic control antibody (i.e., a monoclonal antibody that does not bind to PD-1 and CTLA-4) did not show ADCC activity, and Ipilimumab-IgG1 also did not show a clear ADCC effect. This is thought to be because TF1-PD1 cells do not express CTLA-4. 609 has a heavy chain constant region of IgG4 subtype, and has weak ADCC activity, with target cell lysis remaining at about 10%. On the other hand, 609-IgG1, in which the heavy chain constant region of 609 is replaced with the IgG1 subtype, showed slightly stronger ADCC activity, with target cell lysis reaching a maximum of about 50%. This indicates that there is a difference in ADCC activity depending on whether the heavy chain constant region is the IgG4 subtype or the IgG1 subtype, and it was confirmed that IgG1 shows stronger ADCC activity than IgG4. Furthermore, 609-Fab-Ipilimumab-IgG1 exhibited ADCC activity similar to that of 609, and Ipilimumab-Fab-609-IgG1 exhibited ADCC activity similar to that of 609-IgG1. The ADCC activity of 609-Fab-Ipilimumab-IgG1 was significantly reduced compared to that of Ipilimumab-Fab-609-IgG1, and the difference in ADCC activity between the two is considered to be due to the configuration of the bispecific antibody.

[0239] 7.5.4) Pharmacokinetic evaluation of bispecific antibodies targeting PD-1 and CTLA-4 in rats The pharmacokinetic properties of the bispecific antibodies targeting PD-1 and CTLA-4 in vivo in rats were evaluated as described above in 1.7.5, except that the two antigens PD1-ECD-His and CTLA4-ECD-Fc corresponding to the bispecific antibodies targeting PD-1 and CTLA-4 prepared in 7.5.1 above were used as antigens to coat the ELISA plate.

[0240] As shown in Figure 37, the half-life of 609-Fab-Ipilimumab-IgG1 was 15.2 days when PD-1 was used as the antigen, and 14.6 days when CTLA-4 was used as the antigen. These results confirmed that 609-Fab-Ipilimumab-IgG1 has good pharmacokinetic properties.

[0241] 7.5.5) Antitumor effects of bispecific antibodies targeting PD-1 and CTLA-4 in mice Human PD-1 / CTLA-4 double transgenic mice (breed background: C57BL / 6) were purchased from Beijing Biocytogen Co., Ltd., and MC38 mouse colon cancer cell line was purchased from Guangzhou JennioBiotech Co., Ltd. In the PD-1 / CTLA-4 double transgenic mice, the extracellular domains of human-derived PD-1 and CTLA-4 genes were substituted for the mouse homologous regions, so that the bispecific antibody of the present invention can recognize and bind to the PD-1 and CTLA-4 of the transgenic mice. The experiment was performed as follows. MC38 was cultured in vitro using DMEM medium containing 10% serum, and the serum and medium were both manufactured by Gibco. The MC38 cells grown by culture were introduced into human PD-1 transgenic mice at 2 × 10 per mouse. 6 The tumor cells were inoculated by subcutaneous injection. The volume of the inoculated tumor cells was approximately 100 mm 3When the mice grew to 100 mm, they were randomly divided into groups, with 8 mice in each group. The mice were administered saline to the control group, 10 mg / kg of 609, a PD-1 antibody, to the 609 group, 10 mg / kg of Yervoy (Bristol-Myers Squibb), a CTLA-4 positive control antibody, to the Yervoy group, 10 mg / kg of 609 and 10 mg / kg of Yervoy to the 609+Yervoy group, and 16 mg / kg of 609-Fab-Ipilimumab-IgG1 to the 609-Fab-Ipilimumab-IgG1 group. Since bispecific antibodies and monoclonal antibodies have different molecular weights, in this experiment the antibodies were converted to the amount of substance, and each antibody was administered at the same amount of substance. The above-mentioned administration schedule was followed, and the tumor volume was measured twice a week for a total of four times. Finally, the tumor growth status of each group was arranged in chronological order to create a growth graph as shown in FIG.

[0242] The tumor inhibition rate was calculated according to [tumor inhibition rate = (average tumor volume of the control group - average tumor volume of the experimental group) / average tumor volume of the control group x 100%], and the tumor inhibition rates of the 609 group, the Yervoy group, the 609 + Yervoy group and the 609-Fab-Ipilimumab-IgG1 group at the end of the experiment on day 14 were 48.6%, 79.1%, 85.9% and 92.2%, respectively. It was also confirmed that 609-Fab-Ipilimumab-IgG1 inhibited tumor growth more effectively than the single administration of 609 and Yervoy, and showed the same inhibitory effect as the combined administration of 609 and Yervoy.

[0243] 7.6) Generation of bispecific antibodies targeting PD-1 and LAG-3 7.6.1) Generation of bispecific antibodies The heavy chain variable region of 609 and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 5E7-Hu was then attached via an artificial linker, GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 609-Fab-5E7-IgG4 (SEQ ID NO: 143 and SEQ ID NO: 144). Similarly, the heavy chain variable region of 5E7-Hu and the CH1 structural domain of human IgG4 were linked, and the heavy chain variable region of 609 was then attached via an artificial linker GGGGSGGGSGGGGGS, and finally the heavy chain constant region of human IgG4 (CH1+CH2+CH3, containing S228P as a site mutation in the hinge region) was attached to create a long heavy chain containing two heavy chain variable regions and two CH1 structural domains, and this was named 5E7-Fab-609-IgG4 (SEQ ID NO: 145 and SEQ ID NO: 146).

[0244] The above sequences were each introduced into a pcDNA3.4 expression vector, and the expression vectors for 609-Fab-5E7-IgG4 and 5E7-Fab-609-IgG4 were each combined with an expression vector for 5E7-Hu-LC. The antibodies were expressed and purified using the methods described above, and the resulting antibodies were named 609-Fab-5E7-IgG4 and 5E7-Fab-609-IgG4, respectively (for convenience, the antibodies are named by their heavy chains).

[0245] The ELISA assay was performed as follows: His-tagged human PD-1 extracellular domain was manufactured by Sino Biological Co., Ltd., with product number 10377-H08H, and His-tagged human LAG-3 extracellular domain was also manufactured by Sino Biological Co., Ltd., with product number 16498-H08H. In the present invention, these two recombinant proteins are designated as PD1-ECD-His and LAG3-ECD-His, and PD1-ECD-His and LAG3-ECD-His were used to coat the ELISA plate, respectively, at a coating concentration of 10 ng / well.

[0246] As shown in Figure 39A, 609-HC+5E7-Hu-LC, 609-Fab-5E7-IgG4, and 5E7-Fab-609-IgG4 all effectively bound to PD1-ECD-His and inhibited EC 50 The EC binding sites of 5E7-Hu, 609-Fab-5E7-IgG4, and 5E7-Fab-609-IgG4 were 0.1523 nM, 0.161 nM, and 0.8138 nM, respectively. As shown in FIG. 39B, 5E7-Hu, 609-Fab-5E7-IgG4, and 5E7-Fab-609-IgG4 all effectively bound to LAG3-ECD-His and exhibited EC 50 were 0.1472 nM, 0.2082 nM, and 0.1529 nM, respectively. These results demonstrated that 609-Fab-5E7-IgG4 and 5E7-Fab-609-IgG4 can bind to LAG-3 in addition to PD-1, and that both are bispecific antibodies.

[0247] 7.6.2) Evaluation of the binding ability of bispecific antibodies to PD-1 and LAG-3 A microplate was coated with LAG3-ECD-His, and the antibody to be measured was serially diluted with PBST containing 1% bovine serum albumin to form a concentration gradient, which was then added to the coated microplate and incubated at room temperature for about 30 minutes.Then, an experiment was performed in the same manner as in 1.7.4 above.

[0248] As shown in Figure 40, 609-Fab-5E7-IgG4 effectively binds human PD-1 even after binding to LAG-3, and inhibits EC 50 On the other hand, neither 5E7-Hu nor 609-HC+5E7-Hu-LC was able to bind to PD-1 and LAG-3 simultaneously.

[0249] 7.6.3) Bioactivity Evaluation of Bispecific Antibodies Targeting PD-1 and LAG-3 In this study, the biological activity of bispecific antibodies targeting PD-1 and LAG-3 was evaluated according to the methods described in 7.5.2 above.

[0250] [Table 4]

[0251] As shown in FIG. 41 and Table 5, the stimulatory effect of 5E7-Hu on IL-2 secretion was particularly weak. On the other hand, 609 and 609-HC+5E7-Hu-LC suppressed the EC 50 and the actual measured value (Top value) were at the same level, and it was confirmed that both have similar biological activities. At a concentration of over 1 nM, 609-Fab-5E7-IgG4 stimulated IL-2 secretion more effectively than monoclonal antibodies 5E7-Hu and 609, and hybrid antibody 609-HC+5E7-Hu-LC, and showed a stimulating effect similar to that of the combined use of 5E7-Hu and 609-HC+5E7-Hu-LC. In addition, in Table 5, "5E7-Hu / 609-HC+5E7-Hu-LC" means the combined use of 5E7-Hu and 609-HC+5E7-Hu-LC at a substance amount ratio of 1:1.

[0252] 7.6.4) Antitumor effect of bispecific antibodies targeting PD-1 and LAG-3 in mice Human PD-1 / LAG-3 double transgenic mice (breed background: C57BL / 6) were purchased from Beijing Biocytogen Co., Ltd., and MC38 mouse colon cancer cell line was purchased from Guangzhou JennioBiotech Co., Ltd. In the PD-1 / LAG-3 double transgenic mice, the mouse homologous regions were replaced with the extracellular domains of human-derived PD-1 and LAG-3 genes, so that the bispecific antibody of the present invention can recognize and bind to PD-1 and LAG-3 of the transgenic mice. The experiment was performed as follows. MC38 was cultured in vitro using DMEM medium containing 10% serum, and the serum and medium were both manufactured by Gibco. The MC38 cells grown by culture were introduced into human PD-1 transgenic mice at 2 × 10 per mouse. 6 The tumor cells were inoculated by subcutaneous injection. The volume of the inoculated tumor cells was approximately 100 mm 3When the mice grew to 100 mm, they were randomly divided into groups, with 8 mice in each group. Each group was administered saline to the control group, 20 mg / kg of PD-1 antibody 609 to the 609 group, 20 mg / kg of LAG-3 antibody 5E7-Hu to the 5E7-Hu group, 20 mg / kg of 609 and 20 mg / kg of 5E7-Hu to the 609+5E7-Hu group, and 32 mg / kg of 609-Fab-5E7-IgG4 to the 609-Fab-5E7-IgG4 group. Since bispecific antibodies and monoclonal antibodies have different molecular weights, the antibodies were converted to substance amounts in this experiment, and each antibody was administered at the same substance amount. Then, according to the above-mentioned administration schedule, the administration was performed twice a week for a total of four times, and the tumor volume was measured twice a week. Finally, the tumor growth status of each group was arranged in chronological order to create a growth graph as shown in FIG.

[0253] The tumor inhibition rate was calculated according to [tumor inhibition rate = (average tumor volume of the control group - average tumor volume of the experimental group) / average tumor volume of the control group x 100%]. At the end of the experiment on the 14th day, the tumor inhibition rates of the 609 group, the 5E7-Hu group, the 609 + 5E7-Hu group, and the 609-Fab-5E7-IgG4 group were 70.8%, 13.1%, 71.5%, and 82.8%, respectively. In addition, 609-Fab-5E7-IgG4 inhibited tumor growth more effectively than 609 or 5E7-Hu alone, and the combination of 609 and 5E7-Hu did not show a superior inhibitory effect compared to the administration of 609 alone. From these results, it was confirmed that 609-Fab-5E7-IgG4 exerts a synergistic effect as a bispecific antibody targeting PD-1 and LAG-3.

[0254] 7.7) Evaluation of hybrid antibody specificity Measurements by ELISA were performed as follows. The above-mentioned antigens EGFR-ECD-His, VEGF165-His, HER2-ECD-His, LAG3-ECD-His and CTLA4-ECD-Fc were used to coat an ELISA plate under the above-mentioned experimental conditions, respectively, to confirm whether the hybrid antibodies combining the heavy chain of 609 and the light chain of other antibodies could recognize these targets. In this experiment, antibodies such as Cetuximab-IgG1, 601, Trastuzumab-IgG1, 5E7-Hu and Ipilimumab-IgG1 were derived as described above and used as positive control antibodies that bind to each antigen. Among them, Trastuzumab-IgG1 has a variable region as described in 7.1 above, a constant region similar to Ipilimumab-IgG1, and was prepared in the same manner as the other antibodies.

[0255] As shown in Figures 43A to 43E, hybrid antibodies combining the heavy chain of 609 with the light chain of other antibodies were unable to recognize other targets, confirming that these hybrid antibodies have good specificity.

[0256] 7.8) HPLC-SEC Figure 44A is the HPLC-SEC spectrum of 609-Fab-Cetuximab-IgG4, with the occupancy rate of the main peak being 99.13%. Figure 44B is the HPLC-SEC spectrum of 609-Fab-Pertuzumab-IgG4, with the occupancy rate of the main peak being 99.2%. Figure 44C is the HPLC-SEC spectrum of 609-Fab-Ipilimumab-IgG1, with the occupancy rate of the main peak being 99.3%. Figure 44D is the HPLC-SEC spectrum of Ipilimumab-Fab-609-IgG1, with the occupancy rate of the main peak being 99. 2%, Figure 44E is the HPLC-SEC spectrum of 609-Fab-Ipilimumab-IgG4, with a main peak occupancy of 99.3%, Figure 44F is the HPLC-SEC spectrum of Ipilimumab-Fab-609-IgG4, with a main peak occupancy of 99.1%, Figure 44G is the HPLC-SEC spectrum of 609-Fab-5E7-IgG4, with a main peak occupancy of 99.2%, and Figure 44H is the HPLC-SEC spectrum of 5E7-Fab-609-IgG4, with a main peak occupancy of 99.0%.

[0257] 7.9) HPLC-IEC Figure 45A shows the HPLC-IEC spectrum of 609-Fab-Ipilimumab-IgG1, with the occupancy rate of the main peak being 83.52%. This result confirmed that 609-Fab-Ipilimumab-IgG1 has good charge heterogeneity.

[0258] Figure 45B shows the HPLC-IEC spectrum of 609-Fab-5E7-IgG4, with the occupancy rate of the main peak being 85.43%. This result confirmed that 609-Fab-5E7-IgG4 has good charge heterogeneity.

[0259] 7.10) CE-SDS Figures 46A and 46B are the NR-CE-SDS and R-CE-SDS spectra of 609-Fab-Ipilimumab-IgG1, respectively. In the NR-CE-SDS spectrum, the occupancy rate of the main peak, peak 13, was 97.02%, and in the R-CE-SDS spectrum, the occupancy rates of the two main peaks, peak 2 (peak corresponding to the light chain) and peak 12 (peak corresponding to the long heavy chain), were 39.14% and 59.13%, respectively, with the area ratio of both peaks being 2:3.0. In addition, the peak area ratio of the light chain and the long heavy chain of 609-Fab-Ipilimumab-IgG1 in the R-CE-SDS spectrum was as expected.

[0260] Figures 46C and 46D show the NR-CE-SDS and R-CE-SDS spectra of 609-Fab-5E7-IgG4, respectively, in which the occupancy rate of the main peak, peak 11, was 94.73% in the NR-CE-SDS spectrum, and the occupancy rates of the two main peaks, peak 7 (peak corresponding to the light chain) and peak 16 (peak corresponding to the long heavy chain), were 38.32% and 59.58%, respectively, in the R-CE-SDS spectrum, with the area ratio of the two peaks being 2:3.1. Furthermore, the peak area ratio of the light chain and long heavy chain of 609-Fab-5E7-IgG4 in the R-CE-SDS spectrum was as expected.

Claims

1. A tetravalent bispecific antibody targeting PD-1 and VEGF, The tetravalent bispecific antibody comprises two polypeptide chains and four common light chains, A tetravalent bispecific antibody, characterized in that the amino acid sequences of the polypeptide chains are represented by SEQ ID NO:21 and the amino acid sequence of the common light chain is represented by SEQ ID NO:

13.

2. 1. An isolated nucleotide comprising: A nucleotide encoding the tetravalent bispecific antibody of claim 1, which targets PD-1 and VEGF and comprises two polypeptide chains and four common light chains.

3. The nucleotide sequence encoding the polypeptide chain is represented in SEQ ID NO:22, The nucleotide sequence encoding the common light chain is represented by SEQ ID NO:

14.

4. An expression vector, comprising the nucleotide sequence according to claim 2 or 3.

5. A host cell, characterized in that it comprises a nucleotide according to claim 2 or claim 3 or an expression vector according to claim 4.

6. A pharmaceutical composition comprising the tetravalent bispecific antibody targeting PD-1 and VEGF described in claim 1.

7. Use of the tetravalent bispecific antibody targeting PD-1 and VEGF described in claim 1 or the pharmaceutical composition described in claim 6 in the manufacture of a drug for treating cancer, inflammatory diseases and autoimmune diseases.

8. The use according to claim 7, wherein the tetravalent bispecific antibody targeting PD-1 and VEGF or the pharmaceutical composition is used for the treatment of cancer.

9. A method for producing a tetravalent bispecific antibody targeting PD-1 and VEGF according to claim 1, comprising: A step a of constructing the expression vector according to claim 4 and introducing it into a host cell; Step b expressing the tetravalent bispecific antibody targeting PD-1 and VEGF by culturing the host cell obtained in step a under expression conditions; step c of isolating and purifying the tetravalent bispecific antibody targeting PD-1 and VEGF expressed in step b; A manufacturing method comprising the steps of:

Images