Bispecific / multispecific antibodies and uses thereof

Abstract

The invention provides a novel four-chain protein. The protein is a bispecific or multispecific protein. The invention also provides isolated polynucleotides for coding heavy chains 1 and 2 and light chains 1 or 2 of the protein, a group of isolated polynucleotides, vectors for coding the heavy chains 1 and 2 and the light chains 1 or 2 respectively, and host cells containing the isolated polynucleotides / the group of isolated polynucleotides / vectors. The invention also provides a preparation method of the novel quadruplex protein, a pharmaceutical composition containing the protein, a method for preparing the novel protein with improved thermal stability, application of the protein / separated polynucleotide / carrier / host cell / pharmaceutical composition in preparation of medicines for preventing and treating diseases, and a method for preventing and treating diseases.

Description

Technical Field

[0001] The present invention relates to a novel protein comprising at least two antigen-binding domains. Specifically, the present invention relates to a novel four-chain protein, which is derived from an IgG-like or IgG-based antibody and is a bispecific or multispecific protein. The present invention also relates to an isolated polynucleotide encoding the heavy chain 1, light chain 1, heavy chain 2 or light chain 2 of the novel protein; a set of isolated polynucleotides comprising a polynucleotide encoding the heavy chain 1 of the novel protein, a polynucleotide encoding the light chain 1 of the novel protein, a polynucleotide encoding the heavy chain 2 of the novel protein, and a polynucleotide encoding the light chain 2 of the novel protein; a vector comprising the isolated polynucleotide; a host cell containing the isolated polynucleotide, the set of isolated polynucleotides or the vector; a method for preparing a novel protein with regulated thermal stability; a method for preparing the novel protein; a pharmaceutical composition comprising the novel protein, or the isolated polynucleotide, or the set of isolated polynucleotides, or the vector or the host cell, and a pharmaceutically acceptable carrier. The present invention also relates to the use of the novel protein, the isolated polynucleotide, the set of isolated polynucleotides, the vector, the host cell or the pharmaceutical composition in the preparation of a drug for preventing or treating a disease, or in the preparation of a disease diagnostic kit. The present invention also relates to a method for preventing or treating a disease in a subject in need thereof, comprising administering to the subject a therapeutically effective amount of the novel protein, isolated polynucleotide, set of isolated polynucleotides, vector, host cell or pharmaceutical composition. Background Art

[0002] The statements in this section are intended to provide background information related to the present invention, and the content thereof does not necessarily constitute prior art.

[0003] Bispecific / multispecific antibodies are rapidly emerging as an important class of therapeutic agents. Bispecific antibodies (BsAbs) can bind two different antigens or two different epitopes of the same antigen, thereby enhancing the targeting and therapeutic potential of the antibody. Multispecific antibodies (MsAbs) can bind more than two different antigens or multiple epitopes of the same antigen, further enhancing the targeting and therapeutic potential of the antibody. Bispecific antibodies (BsAbs) / multispecific antibodies (MsAbs) can be constructed in a variety of different conformations, such as CrossMab, dual-variable-domain immunoglobulin (DVD-Ig), single-chain variable fragment-Fc dimer (scFv-Fc dimer), dual-affinity retargeting antibody (DART), etc.

[0004] To promote the heterodimerization of two different heavy chains and inhibit the formation of heavy chain homodimers, a variety of techniques have been developed, such as the knobs-into-holes (KIH, Ridgway et al., 1996), CH3 electrostatic steering technology (Gunasekaran et al., 2010), SEED technology (Davis et al., 2010), DuoBody technology (Labrijn et al., 2013), Azymetric platform (Escobar-Cabrera et al., 2017), and XmAb bispecific platform (Moore et al., 2019).

[0005] However, ensuring the selective pairing of antibody light and heavy chains remains challenging. Currently, there are some solutions to light chain mispairing, such as culturing half-antibodies in two different transformed Escherichia coli cell lines and then assembling them to prepare IgG-like bispecific antibodies (Spiess et al., 2013), but bispecific antibodies expressed by a single cell line are preferred. In addition, a strategy of combining a common light chain with two different heavy chains can be adopted, but this scheme requires screening for suitable common light chains by other techniques, including transgenic animals with a single light chain, or extensive modification of the complementarity-determining regions (CDRs) (Kitazawa and Shima, 2020). Other solutions to light chain mispairing include the CH1 / CL exchange domain used in Roche's CrossMab platform (Regula et al., 2018), and engineered CH1 / CL interfaces with additional mutations in the variable domain (Dillon et al., 2017; Lewis et al., 2014; Liu et al., 2015; Zhao et al., 2021). Most of the above techniques require optimizing the chain ratio to achieve optimal LC / HC assembly.

[0006] Therefore, there is an urgent need to develop bispecific / multispecific antibody structures with selective light and heavy chain pairing functions to improve their expression, stability, and purity. Summary of the Invention

[0007] The present invention provides a novel protein comprising two pairs of polypeptides, wherein each pair of polypeptides consists of one heavy chain and one light chain. The heavy chain comprises a heavy chain variable domain (VH) and a first heavy chain constant domain (CH1), and the light chain comprises a light chain variable domain (VL) and a light chain constant domain (CL). The VH domain and the VL domain of the first pair of polypeptides can form a first Fv fragment, and the first Fv fragment can specifically bind to a first antigen. The VH domain and the VL domain of the second pair of polypeptides can form a second Fv fragment, and the second Fv fragment can specifically bind to a second antigen. Among them, at least one pair of dimerization domains is introduced into the first pair of polypeptides, and each pair of dimerization domains can form a dimer alone or form a dimer with other proteins together. The dimerization domain is a homo / hetero-dimerization domain.

[0008] In some embodiments, the first antigen and the second antigen are different antigens. In some embodiments, the first antigen and the second antigen are different epitopes of the same antigen.

[0009] In some embodiments, each pair of polypeptides consists of a heavy chain and a light chain, wherein the heavy chain comprises a VH domain, a first heavy chain constant (CH1) domain, a second heavy chain constant (CH2) domain, and a third heavy chain constant (CH3) domain, forming a long chain of VH-CH1-CH2-CH3; the light chain comprises a VL domain and a CL domain, forming a short chain of VL-CL.

[0010] In some embodiments, the novel protein further comprises one or more modifications that can improve the correct assembly rate and thermal stability of the protein. In some embodiments, the modification is located in the CH3 domain, such as the knob-into-hole (KIH, Ridgway et al., 1996), CH3 electrostatic steering technology (Gunasekaran et al., 2010), SEED technology (Davis et al., 2010), XmAb bispecific platform (Moore et al., 2019), Azymetric platform (Escobar-Cabrera et al., 2017), and DuoBody technology (Labrijn et al., 2013).

[0011] In some embodiments, the novel protein is derived from an IgG-type antibody and is a bispecific or multispecific protein. The present invention provides a novel protein with four chains, which are respectively named heavy chain 1 (H chain 1), light chain 1 (L chain 1), heavy chain 2 (H chain 2), and light chain 2 (L chain 2). In some embodiments, the novel protein comprises four polypeptides, which are heavy chain 1, light chain 1, heavy chain 2, and light chain 2 respectively.

[0012] Both heavy chain 1 and heavy chain 2 contain a VH domain, a CH1 domain, a CH2 domain, and a CH3 domain, respectively forming a long chain of VH-CH1-CH2-CH3; both light chain 1 and light chain 2 contain a VL domain and a CL domain, respectively forming a short chain of VL-CL.

[0013] Heavy chain 1 and light chain 1 are derived from antibody A, and a single heavy chain 1 pairs with a single light chain 1 to form half-antibody A; heavy chain 2 and light chain 2 are derived from antibody B, and a single heavy chain 2 pairs with a single light chain 2 to form half-antibody B.

[0014] Among them, the CH1 domain of heavy chain 1 pairs with the CL domain of light chain 1, named the CH1 / CL combination in half-antibody A; the CH1 domain of heavy chain 2 pairs with the CL domain of light chain 2, named the CH1 / CL combination in half-antibody B.

[0015] In some embodiments, the CH1 / CL combination in half-antibody A is replaced by a dimerization domain, and the dimerization domain includes a homodimer and a heterodimer. Each pair of homodimeric or heterodimeric domains can form a dimer alone or together with other proteins.

[0016] In some embodiments, at least one pair of dimerization domains is introduced to replace the CH1 / CL combination in half-antibody A and / or half-antibody B.

[0017] In the present invention, a "homodimerization domain" refers to a domain that can form a homodimer alone or together with other proteins. The homodimerization domain is derived from humans or other mammals for replacing the CH1 / CL combination in half-antibody A or half-antibody B.

[0018] In some embodiments, the homodimerization domain is derived from humans or other mammals, and specifically can be selected from the IgG CH3 / CH3 domain, the IgM CH2 / CH2 domain, or the IgE CH2 / CH2 domain. In some embodiments, each homodimerization domain carries different mutations to optimize the dimerization effect.

[0019] In some embodiments, the CH1 / CL combination in half-antibody A is replaced by a homodimerization domain or a modified homodimerization domain, while the CH1 / CL combination in half-antibody B remains unchanged.

[0020] In other embodiments, the CH1 / CL combinations in both half-antibody A and half-antibody B are replaced by a homodimerization domain or a modified homodimerization domain.

[0021] Among them, "the CH1 / CL combination in half-antibody A is replaced by CH3 IgG / CH3 IgG"Replacement" means replacing the CH1 domain in half-antibody A with the IgG1 wild-type CH3 (WT CH3) domain and simultaneously replacing the CL domain with the IgG1 wild-type CH3 domain, that is, replacing the CH1 domain of heavy chain 1 with the IgG1 wild-type CH3 domain and simultaneously replacing the CL domain of light chain 1 with the IgG1 wild-type CH3 domain.

[0022] Among them, "the CH1 / CL combination in half-antibody A is replaced by CH2 IgM / CH2 IgM "Replacement" means replacing the CH1 domain in half-antibody A with the IgM wild-type CH2 (WT CH2) domain and the CL domain with the IgM wild-type CH2 domain, that is, replacing the CH1 domain of heavy chain 1 with the IgM wild-type CH2 domain and simultaneously replacing the CL domain of light chain 1 with the IgM wild-type CH2 domain.

[0023] Among them, "the CH1 / CL combination in half-antibody A is replaced by a knob IgGCH3 / hole IgGCH3 "Replacement" means replacing the CH1 domain in half-antibody A with the knob-shaped IgG1 CH3 domain and the CL domain with the hole-shaped IgG1 CH3 domain, that is, replacing the CH1 domain of heavy chain 1 with the knob-shaped IgG1 CH3 domain and simultaneously replacing the CL domain of light chain 1 with the hole-shaped IgG1 CH3 domain.

[0024] Among them, "the CH1 / CL combination in half-antibody A is replaced by a hole IgGCH3 / knob IgGCH3 "Replacement" means replacing the CH1 domain in half-antibody A with the hole-shaped IgG1 CH3 domain and the CL domain with the knob-shaped IgG1 CH3 domain, that is, replacing the CH1 domain of heavy chain 1 with the hole-shaped IgG1 CH3 domain and simultaneously replacing the CL domain of light chain 1 with the knob-shaped IgG1 CH3 domain.

[0025] Among them, "the CH1 / CL combination in half-antibody A is replaced by a knob IgM CH2 / hole IgM CH2 "Replacement" means replacing the CH1 domain in half-antibody A with the knob-shaped IgM CH2 domain and the CL domain with the hole-shaped IgM CH2 domain, that is, replacing the CH1 domain of heavy chain 1 with the knob-shaped IgM CH2 domain and simultaneously replacing the CL domain of light chain 1 with the hole-shaped IgM CH2 domain.

[0026] Among them, "the CH1 / CL combination in half-antibody A is replaced by a holeIgM CH2 / Knob IgM CH2 "Replacement" means replacing the CH1 domain in half-antibody A with the IgM CH2 domain of the socket type and the CL domain with the IgM CH2 domain of the knob type, that is, replacing the CH1 domain of heavy chain 1 with the IgM CH2 domain of the socket type, and at the same time replacing the CL domain of light chain 1 with the IgM CH2 domain of the knob type.

[0027] The "heterodimerization domain" described in the present invention refers to a domain that can form a heterodimer alone or form a heterodimer with other proteins, and is used to replace the CH1 / CL combination in half-antibody A. The heterodimerization domain is derived from humans or other mammals.

[0028] In some embodiments, at least one pair of heterodimerization domains is derived from humans or other mammals, and the heterodimerization domain does not include the TCRα / β or TCRβ / α constant domains. Preferably, the heterodimerization domain does not include the TCRα / β, TCRβ / α constant domains or the HLA / β2m, β2m / HLA domains.

[0029] In some embodiments, the CH1 / CL combination in half-antibody A is replaced by other heterodimerization domains, while half-antibody B remains unchanged.

[0030] In some embodiments, the CH1 / CL combinations in both half-antibody A and half-antibody B are replaced by other heterodimerization domains.

[0031] In some embodiments, the heterodimerization domain can be selected from one of the IgG-like domain combinations, and the IgG-like domain combinations include HLA / β2m, β2m / HLA, CD1s / β2m and β2m / CD1s, such as CD1a / β2m, β2m / CD1a, CD1b / β2m and β2m / CD1b, etc. Alternatively, the heterodimerization domain is selected from one of the non-IgG-like domain combinations, and the non-IgG-like domain combinations include GABA1 / GABA2, GABA2 / GABA1, SIRPα / CD47 and CD47 / SIRPα, etc.

[0032] Among them, "the CH1 / CL combination in half-antibody A is replaced by HLA / β2m" means replacing the CH1 domain of half-antibody A with the α3 domain of human leukocyte antigen (HLA) and the CL domain with human β2-microglobulin (β2m), that is, replacing the CH1 domain of heavy chain 1 with the α3 domain of human leukocyte antigen (HLA), and at the same time replacing the CL domain of light chain 1 with human β2-microglobulin (β2m).

[0033] Among them, "the CH1 / CL combination in semi - antibody A is replaced by β2m / HLA" means that the CH1 domain of semi - antibody A is replaced by β2m, and the CL domain is replaced by the α3 domain of HLA, that is, the CH1 domain of heavy chain 1 is replaced by β2m, and at the same time, the CL domain of light chain 1 is replaced by the α3 domain of HLA.

[0034] Among them, "the CH1 / CL combination in semi - antibody A is replaced by CD1s / β2m" means that the CH1 domain of semi - antibody A is replaced by the α3 domain of CD1 molecule (CD1s), and the CL domain is replaced by β2m, that is, the CH1 domain of heavy chain 1 is replaced by the α3 domain of CD1 molecule (CD1s), and at the same time, the CL domain of light chain 1 is replaced by β2m.

[0035] Among them, "the CH1 / CL combination in semi - antibody A is replaced by β2m / CD1s" means that the CH1 domain of semi - antibody A is replaced by β2m, and the CL domain is replaced by the α3 domain of CD1 molecule, that is, the CH1 domain of heavy chain 1 is replaced by β2m, and the CL domain of light chain 1 is replaced by the α3 domain of CD1 molecule.

[0036] Among them, "the CH1 / CL combination in semi - antibody A is replaced by GABA1 / GABA2" means that the CH1 domain of semi - antibody A is replaced by the α1 subunit of γ - aminobutyric acid receptor (GABA1), and the CL domain is replaced by the subunit 2 of γ - aminobutyric acid type B receptor (GABA2), that is, the CH1 domain of heavy chain 1 is replaced by the α1 subunit of γ - aminobutyric acid receptor (GABA1), and the CL domain of light chain 1 is replaced by the subunit 2 of γ - aminobutyric acid type B receptor (GABA2).

[0037] Among them, "the CH1 / CL combination in semi - antibody A is replaced by GABA2 / GABA1" means that the CH1 domain of semi - antibody A is replaced by GABA2, and the CL domain is replaced by GABA1, that is, the CH1 domain of heavy chain 1 is replaced by GABA2, and the CL domain of light chain 1 is replaced by GABA1.

[0038] Among them, "the CH1 / CL combination in semi - antibody A is replaced by SIRPα / CD47" means that the CH1 domain of semi - antibody A is replaced by signal regulatory protein α (SIRPα), and the CL domain is replaced by cluster of differentiation 47 (CD47), that is, the CH1 domain of heavy chain 1 is replaced by signal regulatory protein α (SIRPα), and the CL domain of light chain 1 is replaced by cluster of differentiation 47 (CD47).

[0039] Among them, "the CH1 / CL combination in semi-antibody A is replaced by CD47 / SIRPα" means that the CH1 domain of semi-antibody A is replaced by CD47 and the CL domain is replaced by SIRPα, that is, the CH1 domain of heavy chain 1 is replaced by CD47 and the CL domain of light chain 1 is replaced by SIRPα.

[0040] In the present invention, "HLA" refers to the α3 domain of the major histocompatibility complex class I protein (MHC-I) encoded by the human leukocyte antigen (HLA) locus. The present invention takes the α3 domain of HLA subtype A*02:01 (amino acids 182-276+LSS in heavy chain 1 or amino acids 185-276+LSS in light chain 1) as an example, and its sequence is shown in SEQ ID NO:48.

[0041] In the present invention, "β2m" refers to human β2-microglobulin (amino acids: 1-99), which can bind to the α3 domain of MHC-I or MHC-I-like proteins to form a non-covalently bound heterodimer, and its protein sequence is shown in SEQ ID NO:49.

[0042] In the present invention, "human CD1s" or "CD1s" refers to human CD1 molecules that can present lipids and glycolipids on the cell surface for T cell recognition. In some embodiments, "CD1s" includes CD1a, CD1b, CD1c, and CD1d. In some embodiments, "CD1s" specifically refers to CD1a or CD1b.

[0043] Among them, "CD1a" refers to the α3 domain of human CD1a protein (amino acids: 184-278, counted from the 18th amino acid, derived from PDB 7KPI, and its sequence is shown in SEQ ID NO:50), which can form a heterodimer with other protein domains through covalent or non-covalent bonds.

[0044] "CD1b" refers to the α3 domain of human CD1b protein (amino acids: 184-278, counted from the 18th amino acid, derived from PDB 6D64, and its sequence is shown in SEQ ID NO:51), which can form a heterodimer with other protein domains through covalent or non-covalent bonds.

[0045] In the present invention, "GABA1" refers to the α1 subunit of the γ-aminobutyric acid receptor (amino acids: 878-919, and its sequence is shown in SEQ ID NO:52), which can form a heterodimer with other protein domains through covalent or non-covalent bonds.

[0046] In the present invention, "GABA2" refers to the R797H natural variant of the gamma-aminobutyric acid type B receptor subunit 2 (amino acids: 779-819, the sequence of which is shown in SEQ ID NO: 53), which can form a heterodimer with other protein domains through covalent or non-covalent bonds.

[0047] In the present invention, the term "domain" refers to a part of a molecule or structure that has common physical, chemical, or structural characteristics (such as similar hydrophobic or polar characteristics). Typical domains include protein-binding domains, DNA-binding domains, ATP-binding domains, or similar folded structures with globular or helical characteristics. The domain can be identified based on its homology to conserved structural or functional motifs.

[0048] In the present invention, the term "protein" is defined as a biopolymer containing units derived from amino acids linked by peptide bonds, which may contain one or more chains.

[0049] In the present invention, the terms "pharmaceutically acceptable" or "physiologically tolerable" and their grammatical variants are used interchangeably, and refer to compositions, carriers, diluents, and reagents that can be administered in or on the human body without producing adverse physiological reactions (such as nausea, dizziness, stomach discomfort, etc.) that impede therapeutic use.

[0050] In the present invention, the terms "cancer", "tumor" or "neoplasia" are used synonymously and refer to any one of several diseases characterized by the uncontrolled proliferation of abnormal cells, the local spread of the affected cells or their spread to other parts of the body (metastasis) via the bloodstream and lymphatic system, and any one of several characteristic structural and / or molecular features. "Cancerous tumor" or "malignant cell" means a cell having specific structural properties, including reduced differentiation and increased ability to invade and metastasize. Cancers treatable by the antibodies of the present invention include solid tumors and hematological cancers. For example, cancers treatable by the antibodies of the present invention include breast cancer, lung cancer, brain cancer, bone cancer, liver cancer, kidney cancer, colon cancer, head and neck cancer, ovarian cancer, hematopoietic cancers (such as leukemia), and prostate cancer, etc. Other examples of cancers treatable by multivalent and multispecific antibodies include, but are not limited to, carcinoma, lymphoma, blastoma, sarcoma, and leukemia, etc. The most specific examples of these cancers include squamous cell carcinoma, small cell lung cancer, non-small cell lung cancer, lung adenocarcinoma, lung squamous cell carcinoma, peritoneal carcinoma, hepatocellular carcinoma, gastrointestinal cancer, pancreatic cancer, glioblastoma, cervical cancer, ovarian cancer, liver cancer, bladder cancer, hepatoma, breast cancer, colon cancer, colorectal cancer, endometrial cancer or uterine cancer, salivary gland cancer, kidney cancer, liver cancer, prostate cancer, vulvar cancer, hepatic carcinoma, thyroid cancer, and various head and neck cancers. Other types of cancers treatable by multivalent and multispecific antibodies are also included within the scope of the present invention or are known in the art.

[0051] In the present invention, an "effective amount" of an antibody refers to the dose required to achieve a particular therapeutic purpose and to cause an observable change in one or more biological activity levels of the target cells bound by the antibody, which change may be manifested as an increase in target activity or as a decrease in target activity. The "effective amount" can be determined empirically for the stated therapeutic purpose by conventional experimental methods.

[0052] In the present invention, the term "therapeutically effective amount" refers to the dose of an antibody, other multivalent multispecific drugs of the present invention, or other drugs that can effectively "treat" a disease or disorder in a patient or mammal. In the case of cancer, the therapeutically effective amount of a drug can be manifested as: reducing angiogenesis and neovascularization; reducing the number of cancer cells; shrinking the tumor volume; inhibiting (i.e., slowing down or preventing to a certain extent) the infiltration of cancer cells into surrounding organs; inhibiting (i.e., slowing down or preventing to some extent) tumor metastasis; inhibiting tumor growth or tumor incidence to a certain extent; stimulating an immune response against cancer cells; and / or partially relieving one or more symptoms associated with cancer (see the definition of "treatment" in the present invention). "Therapeutically effective amount" can also refer to the effective amount of the necessary dose and treatment cycle to achieve the desired therapeutic effect. The therapeutically effective amount of the compositions of the present invention may vary depending on factors such as the pathological condition, the age, sex, weight of the individual, and the ability of the composition to elicit the expected response in the individual. The therapeutically effective amount also needs to ensure that the toxic or harmful effects of the therapeutic composition are offset by its therapeutic beneficial effects.

[0053] Antibodies that can be used to construct multivalent multispecific antibodies include, but are not limited to, monoclonal antibodies, multispecific antibodies, human antibodies, humanized antibodies, primatized antibodies, and chimeric antibodies. The immunoglobulin or antibody molecules of the present invention can be of any type (such as IgG, IgE, IgM, IgD, IgA, and IgY), any class (such as IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2), or any subclass of immunoglobulin molecules.

[0054] In certain embodiments, antibody A and antibody B can be the same or different.

[0055] In certain embodiments, antibody A or antibody B can be selected from any of the following categories: cytotoxic antibodies, cell proliferation inhibitors, cell activation and interaction regulators, human immune system regulators, and antigen-neutralizing antibodies.

[0056] In certain embodiments, antibody A or antibody B is optionally selected from the group consisting of: anti-HER2 antibody, anti-CCR8 antibody, anti-FAP antibody, anti-OX-40 antibody, anti-41BB antibody, anti-Angiopoietin-2 antibody, anti-IL-4Rα antibody, anti-BCMA antibody, anti-Blys antibody, anti-BTNO2 antibody, anti-C5 antibody, anti-CD122 antibody, anti-CD13 antibody, anti-CD133 antibody, anti-CD137 antibody, anti-CD138 antibody, anti-CD16a antibody, anti-CD19 antibody, anti-CD20 antibody, anti-CD22 antibody, anti-CD27 antibody, anti-CD28 antibody, anti-CD3 antibody, anti-CD30 antibody, anti-CD33 antibody, anti-CD38 antibody, anti-CD40 antibody, anti-CD47 antibody, anti-CD8 antibody, anti-CEA antibody, anti-CGPR / CGRPR antibody, anti-CSPGs antibody, anti-CTLA4 antibody, anti-CTLA-4 domain antibody, anti-DLL-4 antibody, anti-EGFR antibody, anti-EpCAM antibody, anti-Factor IXa antibody, anti-Factor X antibody, anti-GITR antibody, anti-GP130 antibody, anti-Her3 antibody, anti-HSG antibody, anti-ICOS antibody, anti-IGF1 antibody, anti-IGF1 / 2 antibody, anti-IGF-1R antibody, anti-IGF2 antibody, anti-IGFR antibody, anti-IL-1 antibody, anti-IL-12 antibody, anti-IL-12p40 antibody, anti-IL-13 antibody, anti-IL-17A antibody, anti-IL-1β antibody, anti-IL-23 antibody, anti-IL-5 antibody, anti-IL-6 antibody, anti-IL-6R antibody, anti-Lag-3 antibody, anti-LAG3 antibody, anti-MAG antibody, anti-Met antibody, anti-NgR antibody, anti-NogoA antibody, anti-OMGp antibody, anti-OX40 antibody, anti-PD-1 antibody, anti-PDGFR antibody, anti-PDL-1 antibody, anti-PSMA antibody, anti-RGMA antibody, anti-RGMB antibody, anti-SARS-CoV-2 antibody, anti-Te38 antibody, anti-TIM-3 antibody, anti-TNF antibody, anti-TNFα antibody, anti-TROP-2 antibody, anti-TWEAK antibody, anti-VEGF antibody, and anti-VEGFR antibody.

[0057] In certain embodiments, antibody A is an anti-FAP antibody and antibody B is an anti-HER2 antibody. In certain embodiments, antibody A is the anti-HER2 antibody trastuzumab, which is approved for the treatment of human epidermal growth factor receptor 2 positive (HER2+) breast cancer. The trade name of this antibody is Herceptin. The sequences of the VH domain and VL domain of this antibody are shown as SEQ ID NO:31 and SEQ ID NO:5, respectively. Antibody B is an anti-FAP antibody that targets fibroblast activation protein.

[0058] In some embodiments, the half antibody B remains unmodified. In some embodiments, the half antibody B is modified.

[0059] In some embodiments, a tag is fused to the half antibody B. Preferably, the tag is fused to the C-terminus of the light chain 2 of the half antibody B.

[0060] The tag is used for the purification or labeling of the target protein, including SUMO tag, HIS tag, Flag tag, HA tag, MYC tag, SBP tag, CBD tag, GST tag, MBP tag, pMAL tag, IMPACT tag, Protein A, and GFP, etc.

[0061] In some embodiments, the SUMO tag is fused to the C-terminus of the light chain 2 of the half antibody B and separated by a thrombin cleavage sequence.

[0062] In some embodiments, the half antibody A and the half antibody B can be assembled in any form.

[0063] In some embodiments, the half antibody A and the half antibody B are assembled in the form of a knob-into-hole (KIH) structure.

[0064] In some embodiments, the half antibody A and the half antibody B are assembled by combining the knob-into-hole structure with cysteine mutations.

[0065] In the present invention, "Knobs-into-holes" refers to that the CH3 domain of one heavy chain carries knob mutations, and the CH3 domain of the other heavy chain carries hole mutations. Specifically, the "knob-into-hole" structure involves the interface modification between the CH3 domains of two antibody heavy chains: i) In the CH3 domain of one heavy chain, by replacing an amino acid residue with an amino acid having a larger side chain, a protrusion ("knob") is formed at the interface of the CH3 domain of the first heavy chain; ii) In the CH3 domain of the other heavy chain, the amino acid residue is replaced with an amino acid residue having a smaller side chain volume, so as to form a groove ("hole") at the interface of the CH3 domain of the second heavy chain, so that the "knob" of the CH3 domain of the first heavy chain can be inserted into the "hole" of the CH3 domain of the second heavy chain.

[0066] In some embodiments, the half antibody A and the half antibody B are assembled through the knob-into-hole structure of the CH3 domain, wherein heavy chain 1 has a "knob"-type mutant CH3 domain and is paired and assembled with heavy chain 2 having a "hole"-type mutant CH3 domain. In some embodiments, heavy chain 1 has a "hole"-type mutant CH3 domain and is paired and assembled with heavy chain 2 having a "knob"-type mutant CH3 domain.

[0067] In certain embodiments, the present invention also provides an isolated polynucleotide encoding heavy chain 1, light chain 1, heavy chain 2, or light chain 2 of the novel protein.

[0068] In certain embodiments, the present invention also provides a group of isolated polynucleotides, including isolated polynucleotides encoding heavy chain 1, light chain 1, heavy chain 2, and light chain 2 of the novel protein.

[0069] In certain embodiments, the present invention also provides an isolated vector containing the isolated polynucleotide.

[0070] In certain embodiments, the present invention also provides a host cell containing the isolated polynucleotide, a combination of a group of isolated polynucleotides, or the isolated vector.

[0071] In certain embodiments, the present invention also provides a pharmaceutical composition containing the novel protein, the isolated polynucleotide, a group of isolated polynucleotides, the isolated vector, or the host cell and a pharmaceutically acceptable carrier.

[0072] In certain embodiments, the present invention also provides the use of the novel protein, the isolated polynucleotide, the group of isolated polynucleotides, the isolated vector, the host cell, or the pharmaceutical composition in the preparation of a drug for preventing or treating a disease, or in the preparation of a disease diagnostic kit.

[0073] In certain embodiments, the present invention also provides a method for preventing or treating a disease in a subject in need thereof, including administering to the subject a therapeutically effective amount of the novel protein, the isolated polynucleotide, a group of isolated polynucleotides, the isolated vector, the host cell, or the pharmaceutical composition.

[0074] The present invention also provides a method for preparing the novel protein, including: co-introducing a heavy chain 1 expression vector, a light chain 1 expression vector, a heavy chain 2 expression vector, and a light chain 2 expression vector into an expression host cell, or introducing a combination of these expression vectors into different host cells and expressing them under suitable conditions.

[0075] In certain embodiments, the method includes introducing four expression vectors into the same expression host, and the host expresses four protein chains under suitable conditions. The heavy chain 1 expression vector contains a heavy chain 1 expression construct, the light chain 1 expression vector contains a light chain 1 expression construct, the heavy chain 2 expression vector contains a heavy chain 2 expression construct, and the light chain 2 expression vector contains a light chain 2 expression construct. The expression construct can be a plasmid or other expression form.

[0076] In certain embodiments, the correct pairing of the novel protein is insensitive to the molar ratio of heavy chain 1 expression vector: light chain 1 expression vector: heavy chain 2 expression vector: light chain 2 expression vector. In certain embodiments, there is no limitation on the molar ratio of heavy chain 1 expression vector: light chain 1 expression vector: heavy chain 2 expression vector: light chain 2 expression vector. The molar ratio of heavy chain 1 expression vector: light chain 1 expression vector: heavy chain 2 expression vector: light chain 2 expression vector can be 1:1:1:1, 2:2:1:1, 1:1:2:2, 1:3:1:1 or 1:1:1:3.

[0077] In certain embodiments, the host cell is a eukaryotic cell. In certain embodiments, the host cell is a mammalian cell.

[0078] In certain embodiments, at least one pair of homodimerization domains is selected from IgG1 CH3 / CH3, IgM CH2 / CH2, and IgE CH2 / CH2.

[0079] In certain embodiments, at least one pair of heterodimerization domains is selected from: a pair of HLA and β2m, a pair of β2m and CD1b, a pair of CD1a and β2m, a pair of GABA1 and GABA2, or a pair of SIRPα and CD47.

[0080] In certain embodiments, the heterodimerization domain is selected from at least one of the following combinations: HLA / β2m, β2m / HLA, CD1b / β2m, β2m / CD1b, β2m / CD1a, and CD1a / β2m.

[0081] In certain embodiments, the heterodimerization domain is selected from β2m / CD1b, or CD1b / β2m, or β2m / CD1a, or CD1a / β2m. Preferably, the CH1 domain of heavy chain 1 is replaced with the α3 domain of CD1a, and the CL domain of light chain 1 is replaced with β2m.

[0082] In certain embodiments, β2m / CD1a, or CD1a / β2m, or β2m / CD1b or CD1b / β2m contains a mutation. In certain embodiments, in β2m / CD1a, or CD1a / β2m, or β2m / CD1b or CD1b / β2m, one or more disulfide bonds are formed by introducing cysteine mutations.

[0083] In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1a D240C , and the CL domain of light chain 1 is replaced with β2m R12C . In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1a S238C , and the CL domain of light chain 1 is replaced with β2m R12C。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1a D234C and the CL domain of light chain 1 is replaced with β2m Q8C 。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1a G194C and the CL domain of light chain 1 is replaced with β2m M99C 。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1a W190C and the CL domain of light chain 1 is replaced with β2m P14C 。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1a G194C and the CL domain of light chain 1 is replaced with β2m 100C 。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1a G194C and the CL domain of light chain 1 is replaced with β2m Q8K / 100C 。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1a G194C-G233S and the CL domain of light chain 1 is replaced with β2m Q8K / 100C 。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1a G194C / G233S / D234E and the CL domain of light chain 1 is replaced with β2m Q8K / 100C 。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1b G194C and the CL domain of light chain 1 is replaced with β2m Q8K / 100C 。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1b G194C / G233S and the CL domain of light chain 1 is replaced with β2m Q8K / 100C 。In certain embodiments, the CH1 domain of heavy chain 1 is replaced with CD1b G194C / G233S / D234E and the CL domain of light chain 1 is replaced with β2m Q8K / 100C 。

[0084] In certain embodiments, the novel protein is selected from protein ABC060, ABC061, ABC570, ABC571, ABC572, ABC132, ABC133, ABC074, ABC075, ABC131, ABC171, ABC172, ABC173, ABC174, ABC215, ABC588, ABC591, ABC270, ABC271, ABC272, ABC373, ABC374, ABC405, ABC478, ABC603, ABC672, AB C673, ABC513, ABC604, ABC605, ABC674, ABC675, ABC736, ABC737, ABC738, ABC739, ABC723, ABC724, ABC752 or ABC753.

[0085] Among them, ABC060 is composed of four chains shown in SEQ ID NOs: 13, 11, 2, and 4; ABC061 is composed of four chains shown in SEQ ID NOs: 14, 12, 2, and 4; ABC570 is composed of four chains shown in SEQ ID NOs: 82, 83, 3, and 4; ABC571 is composed of four chains shown in SEQ ID NOs: 84, 83, 3, and 4; ABC572 is composed of four chains shown in SEQ ID NOs: 82, 85, 3, and 4; ABC132 is composed of four chains shown in SEQ ID NOs: 14, 26, 2, and 4; ABC133 is composed of four chains shown in SEQ ID NOs: 27, 11, 2, and 4; ABC074 is composed of four chains shown in SEQ ID NOs: 22, 21, 2, and 4; ABC075 is composed of four chains shown in SEQ ID NOs: 24, 23, 2, and 4; ABC131 is composed of four chains shown in SEQ ID NOs: 19, 20, 2, and 4; ABC171 is composed of four chains shown in SEQ ID NOs: 14, 25, 3, and 1; ABC172 is composed of four chains shown in SEQ ID NOs: 14, 26, 3, and 1; ABC173 is composed of four chains shown in SEQ ID NOs: 27, 11, 3, and 1; ABC174 is composed of four chains shown in SEQ ID NOs: 28, 11, 3, and 1; ABC734 is composed of four chains shown in SEQ ID NOs: 13, 11, 3, and 1; ABC735 is composed of four chains shown in SEQ ID NOs: 14, 12, 3, and 1; ABC215 is composed of four chains shown in SEQ ID NOs: 38, 41, 45, and 44; ABC588 is composed of four chains shown in SEQ ID NOs: 27, 11, 47, and 46; ABC591 is composed of four chains shown in SEQ ID NOs: 36, 37, 47, and 46; ABC270 is composed of four chains shown in SEQ ID NOs: 66, 65, 3, and 1; ABC271 is composed of four chains shown in SEQ ID NOs: 67, 65, 3, and 1; ABC272 is composed of four chains shown in SEQ ID NOs: 68, 69, 3, and 1; ABC373 is composed of four chains shown in SEQ ID NOs: 73, 71, 3, and 1; ABC374 is composed of four chains shown in SEQ ID NOs: 72, 70, 3, and 1; ABC405 is composed of four chains shown in SEQ ID NOs: 73, 74, 3, and 1; ABC478 is composed of four chains shown in SEQ ID NOs: 76, 74, 3, and 1; ABC603 is composed of four chains shown in SEQ ID NOs: 76, 75, 3, and 1.ABC672 is composed of four chains shown by SEQ ID NOs: 77, 75, 3 and 1. ABC673 is composed of four chains shown by SEQ ID NOs: 78, 75, 3 and 1. ABC513 is composed of four chains shown by SEQ ID NOs: 79, 74, 3 and 1. ABC604 is composed of four chains shown by SEQ ID NOs: 80, 74, 3 and 1. ABC605 is composed of four chains shown by SEQ ID NOs: 80, 75, 3 and 1. ABC674 is composed of four chains shown by SEQ ID NOs: 81, 75, 3 and 1. ABC675 is composed of four chains shown by SEQ ID NOs: 59, 75, 3 and 1. ABC736 is composed of four chains shown by SEQ ID NOs: 86, 75, 2 and 1. ABC737 is composed of four chains shown by SEQ ID NOs: 87, 75, 2 and 1. ABC738 is composed of four chains shown by SEQ ID NOs: 88, 75, 2 and 1. ABC739 is composed of four chains shown by SEQ ID NOs: 89, 75, 2 and 1. ABC723 is composed of four chains shown by SEQ ID NOs: 76, 75, 47 and 46. ABC724 is composed of four chains shown by SEQ ID NOs: 59, 75, 47 and 46. ABC752 is composed of four chains shown by SEQ ID NOs: 106, 105, 47 and 46. ABC753 is composed of four chains shown by SEQ ID NOs: 107, 105, 47 and 46.,

[0086] Preferably, the novel protein is selected from the group consisting of proteins ABC570, ABC571, ABC572, ABC132, ABC133, ABC074, ABC075, ABC131, ABC171, ABC172, ABC173, ABC174, ABC215, ABC588, ABC591, ABC270, ABC271, ABC272, ABC373, ABC374, ABC405, ABC478, ABC603, ABC672, ABC673, ABC513, ABC604, ABC605, ABC674, ABC675, ABC736, ABC737, ABC738, ABC739, ABC723, ABC724, ABC752 and ABC753.

[0087] In certain embodiments, portions of the half-antibody A and / or half-antibody B are replaced by at least one pair of homologous / heterologous dimerization domains, and each pair of homologous / heterologous dimerization domains can individually form a homologous / heterologous dimer, or bind to other proteins to form a homologous / heterologous dimer. Preferably, CH1 / CL in half-antibody A or CH1 / CL in half-antibody B is replaced by at least one pair of homologous / heterologous dimerization domains. In certain embodiments, CH2 of heavy chain 1 and CH2 of heavy chain 2 are replaced by at least one pair of homologous / heterologous dimerization domains. In certain embodiments, CH3 of heavy chain 1 and CH3 of heavy chain 2 are replaced by at least one pair of homologous / heterologous dimerization domains.

[0088] In certain embodiments, the homologous / heterologous dimerization domains can be introduced by means other than replacement. In certain embodiments, the homologous / heterologous dimerization domains can be introduced by insertion.

[0089] In the present invention, the term "insertion" refers to the addition of one or more amino acid residues or domains between two existing amino acids.

[0090] In certain embodiments, the homologous / heterologous dimerization domains can be inserted into an antibody, its fragment, or variant. For example, the homologous / heterologous dimerization domains can be directly inserted between the CH1 / CL domain and the CH2 / CH2 domain. For another example, the homologous / heterologous dimerization domains can be directly inserted between the CH1 / CL domain and the VH / VL domain. For another example, the homologous / heterologous dimerization domains can be directly inserted after the CH3 / CH3 domain. In certain embodiments, at least one pair of homologous / heterologous dimerization domains are inserted into the novel protein, and each pair of homologous / heterologous dimerization domains can individually form a homodimer or a heterodimer, or bind to other proteins to form a homodimer or a heterodimer.

[0091] In certain embodiments, at least one pair of homologous / heterologous dimerization domains can be fused with the novel protein, and each pair of homologous / heterologous dimerization domains can individually form a homodimer or a heterodimer, or bind to other proteins to form a homodimer or a heterodimer.

[0092] In the present invention, the term "fusion" refers to the linking of a protein with another biomaterial (such as a protein, a nucleic acid molecule, or any other biomolecule or part thereof). A fusion protein can be defined as a protein composed of at least two linked domains.

[0093] In certain embodiments, at least one pair of homodimerization / heterodimerization domains is fused to a novel protein, and each pair of homodimerization / heterodimerization domains can individually form a homodimer or a heterodimer, or bind to other proteins to form a homodimer or a heterodimer. In certain embodiments, at least one pair of homodimerization / heterodimerization domains is fused to any two chains of the novel protein. In certain embodiments, at least one pair of homodimerization / heterodimerization domains is fused to light chain 1 and / or heavy chain 1 of the novel protein. In certain embodiments, at least one pair of homodimerization / heterodimerization domains is fused to heavy chain 1 and / or heavy chain 2 of the novel protein. In certain embodiments, at least one pair of homodimerization / heterodimerization domains is fused to light chain 2 and / or heavy chain 2 of the novel protein. In certain embodiments, two pairs of homodimerization / heterodimerization domains are respectively fused to light chain 1 and heavy chain 1 in half antibody A and light chain 2 and heavy chain 2 in half antibody B.

[0094] In certain embodiments, CH1 / CL in half antibody A and / or CH1 / CL in half antibody B are replaced by at least one pair of homodimerization / heterodimerization domains, and each pair of homodimerization / heterodimerization domains can individually form a homodimer or a heterodimer, or jointly form a homodimer or a heterodimer with other proteins.

[0095] In certain embodiments, a protein can be mutated by fusing another homodimerization / heterodimerization domain. For example, an IgG antibody can be fused with another homodimerization / heterodimerization domain. For another example, the N-terminus of an IgG antibody can be fused with another homodimerization / heterodimerization domain. In certain embodiments, an IgG antibody can be fused with another homodimerization / heterodimerization domain at its C-terminus or within its domain. For example, a homodimerization / heterodimerization domain can be fused at the VH / VL or CH1 / CL domain of an antibody or a half antibody. For another example, a homodimerization / heterodimerization domain can be fused at the CH1 / CL domain of a half antibody through a covalent bond (such as a disulfide bond).

[0096] In the present invention, the term "thermal stability" relates to the ability of a protein (such as an antibody) to resist the action of heat and maintain its properties (such as strength, toughness or elasticity) at a given temperature. During the denaturation process caused by heating, the entire protein or protein domain undergoes one or more unfoldings, resulting in one or more transitions in the melting curve.

[0097] In certain embodiments, "Tm" can be any one, any two or all of Tm1, Tm2, Tm3.

[0098] In the present invention, the term "Tm1" refers to the midpoint of the first melting / unfolding transition.

[0099] In the present invention, the term "Tm2" refers to the midpoint of the second melting / unfolding transition.

[0100] In the present invention, the term "Tm3" refers to the midpoint of the third melting / unfolding transition.

[0101] In certain embodiments, the present invention provides a method for preparing a novel protein with regulated thermal stability, particularly with substantially the same or improved thermal stability, particularly with substantially the same or improved Tm compared to the original protein. The original protein has four polypeptide chains, namely heavy chain 1, light chain 1, heavy chain 2, and light chain 2, wherein heavy chain 1 and heavy chain 2 contain VH domain, CH1 domain, CH2 domain, and CH3 domain, and respectively form VH-CH1-CH2-CH3 long chains, and light chain 1 and light chain 2 contain VL domain and CL domain, and respectively form VL-CL short chains. Among them, heavy chain 1 and light chain 1 are derived from antibody A, and a single heavy chain 1 pairs with a single light chain 1 to form half-antibody A; heavy chain 2 and light chain 2 are derived from antibody B, and a single heavy chain 2 pairs with a single light chain 2 to form half-antibody B. Among them, the CH1 domain of heavy chain 1 pairs with the CL domain of light chain 1, named the CH1 / CL combination of half-antibody A, and the CH1 domain of heavy chain 2 pairs with the CL domain of light chain 2, named the CH1 / CL combination of half-antibody B. The method is characterized in that the CH1 / CL combination in half-antibody A is replaced by a homodimerization domain or other heterodimerization domains.

[0102] In certain embodiments, half-antibody B is not altered. In certain embodiments, half-antibody B is altered.

[0103] In certain embodiments, compared to the original protein, Tm is substantially the same or increased by at least 0.1 °C, 0.2 °C, 0.3 °C, 0.4 °C, 0.5 °C, 0.6 °C, 0.7 °C, 0.8 °C, 0.9 °C, 1 °C, 1.5 °C, 2 °C, 2.5 °C, 3 °C, 3.5 °C, 4 °C, 4.5 °C, 5 °C, 5.5 °C, 6 °C, 6.5 °C, 7 °C, 7.5 °C, 8 °C, 8.5 °C, 9 °C, 9.5 °C, 10 °C, 11 °C, 12 °C, 13 °C, 14 °C, 15 °C, 16 °C, 17 °C, 18 °C, 19 °C or 20 °C, preferably increased by at least 0.3 °C, 0.5 °C, 1.5 °C, 2 °C or 3 °C. In certain embodiments, compared to the original protein, Tm is increased or decreased by at least 0.3 °C, 0.5 °C, 1.5 °C or 3 °C. In certain embodiments, compared to the original protein, Tm is increased or decreased by more than 10 °C.

[0104] In certain embodiments, the thermal stability of the mutant antibody is improved when the mutant antibody maintains substantially the same Tm value as the original antibody. In certain embodiments, the thermal stability of the mutant antibody is improved when the mutant antibody has an increased Tm value compared to the original antibody.

[0105] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the heterodimerization domain is derived from a human or other mammal, and the heterodimerization domain is not the TCRα / β or TCRβ / α constant domain. In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the heterodimerization domain is derived from a human or other mammal, and the heterodimerization domain is not the TCRα / β or TCRβ / α constant domain or the HLA / β2m or β2m / HLA domain.

[0106] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the heterodimerization domain is selected from the combination of HLA and β2m, the combination of β2m and CD1b, the combination of CD1a and β2m, the combination of GABA1 and GABA2, and the combination of SIRPα and CD47.

[0107] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the heterodimerization domain is selected from at least one combination of HLA / β2m, β2m / HLA, CD1b / β2m, β2m / CD1b, β2m / CD1a, and CD1a / β2m.

[0108] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the heterodimerization domain is β2m / CD1b or CD1b / β2m or β2m / CD1a or CD1a / β2m. Preferably, the CH1 domain of heavy chain 1 is replaced with the α3 domain of CD1a, and the CL domain of light chain 1 is replaced with β2m.

[0109] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the homodimerization domain is derived from a human or other mammal, and the homodimerization domain is selected from the IgG CH3 / CH3 domain, the IgM CH2 / CH2 domain, and the IgE CH2 / CH2 domain.

[0110] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, further comprising altering the amino acids of the novel protein by mutagenesis.

[0111] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein amino acids are altered by introducing hydrogen bonds or disulfide bonds into the antibody, preferably into the CH1 / CL domain of the half-antibody A.

[0112] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein β2m / CD1a or CD1a / β2m, or β2m / CD1b or CD1b / β2m contains mutations.

[0113] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein β2m / CD1a or CD1a / β2m, or β2m / CD1b or CD1b / β2m contains mutations that can form one or more disulfide bonds and at the same time achieve better domain stacking.

[0114] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the CH1 domain of heavy chain 1 is replaced with CD1a D240C , and the CL domain of light chain 1 is replaced with β2m R12C . In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the CH1 domain of heavy chain 1 is replaced with CD1a S238C , and the CL domain of light chain 1 is replaced with β2m R12C . In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the CH1 domain of heavy chain 1 is replaced with CD1a D234C , and the CL domain of light chain 1 is replaced with β2m Q8C . In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the CH1 domain of heavy chain 1 is replaced with CD1a G194C , and the CL domain of light chain 1 is replaced with β2m M99C . In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the CH1 domain of heavy chain 1 is replaced with CD1a W190C , and the CL domain of light chain 1 is replaced with β2m P14C . In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability, wherein the CH1 domain of heavy chain 1 is replaced with CD1a G194C , and the CL domain of light chain 1 is replaced with β2m 100C .

[0115] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability and / or yield, wherein the CH1 domain of heavy chain 1 is replaced with CD1a G194C , and the CL domain of light chain 1 is replaced with β2m 100C / Q8K ; or the CH1 domain of heavy chain 1 is replaced with CD1a G194C / G233S , and the CL domain of light chain 1 is replaced with β2m 100C / Q8K ; or the CH1 domain of heavy chain 1 is replaced with CD1a G194C / G233S / D234E , and the CL domain of light chain 1 is replaced with β2m 100C / Q8K .

[0116] In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability and / or yield, wherein the CH1 domain of heavy chain 1 is replaced with CD1b G194C , and the CL domain of light chain 1 is replaced with β2m 100C . In certain embodiments, the present invention provides a method for preparing a novel protein with improved thermal stability and / or yield, wherein the CH1 domain of heavy chain 1 is replaced with CD1b G194C , and the CL domain of light chain 1 is replaced with β2m 100C / Q8K ; or the CH1 domain of heavy chain 1 is replaced with CD1b G194C / G233S , and the CL domain of light chain 1 is replaced with β2m 100C / Q8K ; or the CH1 domain of heavy chain 1 is replaced with CD1b G194C / G233S / D234E , and the CL domain of light chain 1 is replaced with β2m 100C / Q8K .

[0117] Thermal stability can be determined by any technique known to those skilled in the art. In certain embodiments, the melting temperature of the novel protein with improved thermal stability differs from that of the corresponding parental protein by about 5 °C or less. In certain embodiments, the melting temperature of the novel protein with improved thermal stability differs from that of the corresponding parental protein by about 4 °C or less. In certain embodiments, the melting temperature of the novel protein with improved thermal stability differs from that of the corresponding parental protein by about 3 °C or less. In certain embodiments, the melting temperature of the novel protein with improved thermal stability differs from that of the corresponding parental protein by about 2 °C or less. In certain embodiments, the melting temperature of the novel protein with improved thermal stability differs from that of the corresponding parental protein by about 1 °C or less.

[0118] In certain embodiments, the melting temperature of the novel protein with improved thermal stability is at least about 5 °C higher than that of the corresponding parental protein. In certain embodiments, the melting temperature of the novel protein with improved thermal stability is at least about 4 °C higher than that of the corresponding parental protein. In certain embodiments, the melting temperature of the novel protein with improved thermal stability is at least about 3 °C higher than that of the corresponding parental protein. In certain embodiments, the melting temperature of the novel protein with improved thermal stability is at least about 2 °C higher than that of the corresponding parental protein. In certain embodiments, the melting temperature of the novel protein with improved thermal stability is at least 1 °C higher than that of the corresponding parental protein.

[0119] In certain embodiments, the yield of the novel protein is at least 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50, 60, 70, 80, 90, or 100 times that of the corresponding parental protein. BRIEF DESCRIPTION OF THE DRAWINGS

[0120] The following is a brief description of the drawings, which are only used to illustrate the exemplary embodiments of the present invention and not to limit the scope of the claims.

[0121] Figure 1 Shows the screening results of the homologous / heterologous dimerization domains or proteins used to replace the CH1 / CL domain in half antibody A.

[0122] Figure 1 Panel A shows non-reducing and reducing SDS-PAGE gel analyses of protein A-purified anti-HER2 X FAP heterologous IgG variants, where the CH1 / CL of half antibody A is replaced with a non-IgG-like heterologous dimer domain. The heterologous dimer domain replacements include Fos / Jun (ABC064), GABA1 / GABA2 (ABC074), and GABA2 / GABA1 (ABC076).

[0123] Figure 1 Panel B shows non-reducing and reducing SDS-PAGE gel analyses of protein A-purified anti-HER2 X FAP heterologous IgG variants, where the CH1 / CL of half antibody A is replaced with an IgG-like domain. The heterologous dimer domain replacements include IL2Rβ / γ (ABC058, ABC059) and CD8a / CD8β (ABC134).

[0124] Figure 1Panel C shows non-reducing and reducing SDS-PAGE gel analyses of protein A-purified anti-HER2 X FAP heterologous IgG variants in which the CH1 / CL of half antibody A was replaced by an IgG-like homodimeric domain. The homodimeric domain replacement includes CH3 IgG / CH3 IgG (ABC055), knob IgGCH3 / hole IgGCH3 (ABC056), knob IgGCH3 / hole IgGCH3 (ABC057), CH2 Ig M / CH3 IgM (ABC570), hole IgMCH2 / knob IgMCH2 (ABC571), knob IgMCH2 / hole IgMCH2 (ABC572). Wherein knob IgGCH3 / hole IgGCH3 indicates that the CH1 domain of heavy chain 1 (HC1) was replaced by an IgG1 CH3 domain with a knob mutation (S354C, T336W), and the CL domain of light chain 1 (LC1) was replaced by an IgG1 CH3 domain with a hole mutation (Y349C, T336S, L368A, Y407V). Wherein knob IgMCH2 / hole IgMCH2 indicates that the CH1 domain of heavy chain (HC1) was replaced by an IgM CH2 domain with a knob mutation (R245C, C337S, I258Y, T302Y, residue numbering based on Uniprot P0DOX6), and the CL domain of light chain 1 (LC1) was replaced by an IgM CH2 domain with a hole mutation. All anti-HER2 XFAP heterologous IgG variants (IgG domains) were analyzed by 4-20% SDS-PAGE gel under non-reducing and reducing treatment conditions after protein A purification.

[0125] Figure 1 Panel D shows non-reducing and reducing SDS-PAGE gel analyses of protein A-purified anti-HER2 X FAP heterologous IgG variants in which the CH1 / CL in half antibody A was replaced by an IgG-like heterodimeric domain. The heterodimeric domains include HLA / β2m (ABC060), β2m / HLA (ABC061), SIPRα / CD47 (ABC131), β2m / CD1b (ABC132) and CD1a / β2m (ABC133). In addition, deglycosylated anti-HER2 X FAP heterologous IgG variants after PNGaseF treatment were also analyzed.

[0126] Figure 2 SDS-PAGE gel analysis of anti-HER2 X FAP heterologous IgG with different light chain 1 / light chain 2 (LC1 / LC2) chain ratios is shown. Four combinations (β2m / CD1a, CD1a / β2m, β2m / CD1b, and CD1b / β2m) were tested. Four plasmids with different ratios (Table 3) were co-transfected into a host cell line. The protein A-purified products were analyzed by 4-20% SDS-PAGE gels under non-reducing and reducing conditions. The results showed that the orthogonal Fab pairing was insensitive to the LC1 / LC2 ratio.

[0127] Figure 3 A shows the SDS-PAGE gel analysis of ABC171-174, ABC734, and ABC735 under non-reducing and reducing conditions. The chain ratio used in all designs was 1:1:1:1.

[0128] Figure 3 B shows the SEC-HPLC chromatograms of bispecific antibodies (BsAbs) and monoclonal antibodies (mAbs) after protein A purification, with the percentage of the target product marked. Among them, Herceptin (ABC141), anti-FAP X HER2-CrossMab bispecific antibody (ABC203), anti-FAP X HER2-Wuxibody bispecific antibody (ABC204), anti-FAP X HER2-HLA / β2m (ABC734), and anti-FAP X HER2-HLA / β2m (ABC735) were used as controls.

[0129] Figure 3 C shows the sequence alignment of the α3 domains (amino acids: 179-278) of common HLA-A and HLA-B subtypes.

[0130] Figure 4 The thermal stabilities of variants of anti-HER2 / FAP bispecific antibodies and their parental monoclonal IgGs are shown. Figure 4 A: Normalized melting curves recorded using a Bio-Rad CFX Real-Time PCR System. Figure 4 B: Plot of the derivative of the melting curve (dF / dT), where the peak value represents the derivative Tm value. Figure 4 C summarizes the melting temperature (Tm) measured by differential scanning fluorimetry.

[0131] Figure 5Shows the LC-MS analysis of the intact ABC171-ABC174 bispecific antibody using a quadrupole time-of-flight (Q-TOF) liquid chromatography-mass spectrometry in non-reducing mode. Since light chain 1 (LC1) was not covalently bound to the remaining three chains, two peaks were detected in non-reducing mode. Mass spectrometry analysis of the IgG bispecific antibody showed correct chain assembly. Figure 5 A: Deconvoluted mass spectrum of ABC171. Among them, the theoretical molecular weight (MW) of LC1 is 23340.26 Da, the observed mass is 23340.00 Da, and the theoretical MW of HC1 / HC2 / LC2 is 122622.17 Da, the observed mass is 122622.25 Da. Figure 5 B: Deconvoluted mass spectrum of ABC173. Among them, the theoretical molecular weight (MW) of LC1 is 23662.51 Da, the observed mass is 23662.75 Da, and the theoretical MW of HC1 / HC2 / LC2 is 121985.58 Da, the observed mass is 121986.5 Da. Figure 5 C: Deconvoluted mass spectrum of ABC172. Among them, the theoretical molecular weight (MW) of LC1 is 24722.80 Da, the observed mass is 24721.4 Da. The theoretical MW of HC1 / HC2 / LC2 is 122622.17 Da, the observed mass is 122620.4 Da.

[0132] Figure 6 Shows the binding kinetics analysis of the anti-HER2 X anti-FAP bispecific antibody to the human HER2 antigen based on biolayer interferometry (BLI). Among them, the processed results of the kinetic data sets of ABC141 (Herceptin, positive control (A)), ABC171 (B), ABC172 (C), ABC173 (D) and ABC174 (E) binding to the human HER2 antigen are shown. The smooth line represents the global fit of the data to the 1:1 interaction model.

[0133] Figure 7 Shows the simultaneous binding of the anti-HER2×FAP Abio bispecific antibody based on BLI to human HER2 and human FAP antigens. The top panel shows the analyte loading order. First, biotinylated human HER2 antigen was immobilized on the SA sensor tip, then the bispecific antibody was loaded, and then immersed in the second antigen human FAP solution. Taking ABC171, ABC172 and ABC173 as examples. ABC141 (Herceptin), ABC089 (isotype control antibody) and assay buffer were used as controls.

[0134] Figure 8 Shows that the engineered Fab in the ABC173 format has a similar conformation to the wild-type (WT) Fab. Figure 8A: Schematic diagram of the Herceptin Fab domain (PDB ID: 6MH2). Figure 8 B: Engineered Fab structure of ABC173 predicted by AlphaFold2. Figure 8 C: Structural alignment of WT Fab and engineered Fab in the ABC173 format.

[0135] Figure 9 Showing that the engineered orthogonal Fab design can be widely applied to different antibody combinations.

[0136] Figure 9 A: SDS-PAGE gel analysis of anti-Ab1 / Ab2 antibodies in the ABC173, CrossMab, and Wuxibody formats under non-reducing and reducing conditions. The Ab1 / Ab2 combination mCD20 / mCD3 indicates that the variable domain sequence of half-antibody A is from an anti-mouse CD20 antibody, and the variable domain of half-antibody B is from an anti-mouse CD3 antibody, where the CH1 / CL of half-antibody A is replaced by CD1a / β2m (ABC215, labeled 173), or CL / CH1 (ABC216, labeled CrossMab), or the TCRβ / α constant region (ABC217, labeled Wuxi). In the HER2 / hCD3 combination, the variable domain of half-antibody A is from Herceptin, and the variable domain of half-antibody B is from the anti-CD3 part in IMCgp100-CD3, where the CH1 / CL of half-antibody A is replaced by CD1a / β2m (ABC588, labeled 173), or CL / CH1 (ABC589, labeled CrossMab), or the TCRβ / α constant region (ABC590, labeled Wuxi). In the hCCR8 / hCD3 combination, the variable domain of half-antibody A is from the anti-hCCR8 antibody ABC138 (the light chain sequence of the anti-hCCR8 antibody is shown in SEQ ID NO: 37, and the heavy chain sequence is shown in SEQ ID NO: 36), and the variable domain of half-antibody B is from the anti-CD3 part in IMCgp100-CD3, where the CH1 / CL of half-antibody A is replaced by CD1a / β2m (ABC591, labeled 173), or CL / CH1 (ABC592, labeled CrossMab), or the TCRβ / α constant region (ABC593, labeled Wuxi).

[0137] Figure 9 B: SEC-HPLC analysis of ABC215, ABC588, and ABC591 after protein A purification.

[0138] Figure 9C: Deconvoluted mass spectrum of ABC588. The theoretical molecular weight (MW) of LC1 is 23662.1 Da, and the observed mass is 23662.5 Da. The theoretical MW of HC1 / HC2 / LC2 is 123134.8 Da, and the observed mass is 123133.4 Da.

[0139] Figure 10 Show that the thermal stability is improved by introducing a disulfide bond between the CD1a α3 domain and β2m. Figure 10 A shows the interaction between the CD1a α3 domain and β2m (PDB ID: 1XZO). Cysteine mutations were introduced based on the illustrated structure (①②③④) to enhance the interaction between the CD1a α3 domain and β2m. Figure 10 B: Non-reducing SDS-PAGE analysis shows six groups of cysteine mutant CD1a D240C / β2m R12C (ABC270), CD1a S238C / β2m R12C (ABC271), CD1a D234C / β2m Q8C (ABC272), CD1a G194C / β2m M99C (ABC373), CD1a W190C / β2m P14C (ABC374) and CD1a G194C / β2m 100C (ABC405) produced a product of approximately 150 kDa. Position numbering was based on the crystal structure of CD1a / β2m (PDB ID 1XZO).

[0140] Figure 10 C: In terms of expression yield and thermal stability, ABC405 is the most promising among the six tested variants.

[0141] Figure 11 : Optimization of the purity and stability of bispecific variants. Figure 11 A shows non-reducing and reducing SDS-PAGE gel analysis of bispecific antibodies designed based on the ABC173 format. Figure 11 B shows non-reducing and reducing SDS-PAGE gel analysis of bispecific antibodies designed based on the ABC174 format. Figure 11 C and 11D show SEC-HPLC analysis of these variants after protein A purification. Figure 11E and 11F show the deconvoluted mass spectra of mutants ABC603 and ABC605. The theoretical molecular weight (MW) of ABC603 is 145274.8 Da, and the observed mass is 145274.5 Da. The theoretical MW of ABC605 is 145226.1 Da, and the observed mass is 145225.0 Da.

[0142] Figure 12 : Thermal stability analysis of bispecific variants. Figure 12 A: Original melting curve of the bispecific antibody based on the ABC173 format recorded by the Bio-Rad CFX96 Real-Time PCR System. Figure 12 B: Derivative (dF / dT) plot of the melting curve processed by the Bio-Rad CFX96 Real-Time PCR System. Figure 12 C: Original melting curve of the bispecific antibody based on the ABC174 format recorded by the Bio-Rad CFX96 Real-Time PCR System. Figure 12 D: Derivative (dF / dT) plot of the melting curve processed by the Bio-Rad CFX96 Real-Time PCR System. Figure 12 E: Original melting curve of the bispecific antibody without N297A recorded by the Bio-Rad CFX96 Real-Time PCR System. Figure 12 F: Derivative (dF / dT) plot of the melting curve processed by the Bio-Rad CFX96 Real-Time PCR System. Among them, ABC001 (anti-FAP parental antibody) and ABC141 (Herceptin, anti-HER2 parental antibody) are used as controls.

[0143] Figure 12 G: Original melting curves of Herceptin Fab variant (CH / CL replaced by CD1a / β2m) and Fab WT recorded by the Bio-Rad CFX96 Real-Time PCR System. Figure 12 H: Derivative (dF / dT) plot of the melting curve processed by the Bio-Rad CFX96 Real-Time PCR System.

[0144] Figure 13 : Cell-based binding affinity experiments show dual binding of bispecific antibodies in the ABC603, ABC605, ABC673, or ABC675 format. Figure 13 A: Binding curves of ABC723_HER2 / hCD3_603, ABC724_HER2 / hCD3_605, and parental antibody ABC141_Herceptin to the HER2 antigen (SKBR3 cells). Figure 13 B: Binding curves of ABC723_HER2 / hCD3_603, ABC724_HER2 / hCD3_605, and parental antibody ABC728_anti-hCD3 to the hCD3 antigen (Jurkat cells). Figure 13C: Binding curves of ABC752_hCCR8 / hCD3_673, ABC753_hCCR8 / hCD3_675 and parental antibody ABC138_anti-hCCR8 with hCCR8 antigen (hCCR8-transfected 293T cells). Figure 13 D: Binding curves of ABC752_hCCR8 / hCD3_673, ABC753_hCCR8 / hCD3_675 and parental antibody ABC728_anti-hCD3 with hCD3 antigen (Jurkat cells). Detailed implementation manners

[0145] The content of the present invention will be described in more detail below. The following description is not intended to list all possible implementation manners of the present invention or all features that the present invention may include. For example, the features shown in a certain embodiment may be incorporated into other embodiments, and the features shown in a specific embodiment may also be deleted from that embodiment. In addition, based on the content of the present invention, those skilled in the art can make a large number of modifications and supplements to the embodiments of the present invention without departing from the present invention. Therefore, the following description is only to clarify some specific embodiments of the present invention, rather than an exhaustive list of all possible permutations, combinations and their variants.

[0146] Unless otherwise defined, the technical terms and scientific terms used in the present invention have the conventional meanings understood by those of ordinary skill in the art. Although any methods and materials similar or equivalent to those of the present invention can be used to test the content of the present invention, the preferred materials and methods are still as described in the present invention. When describing the claims of the content of the present invention, the following terms will be used.

[0147] In the present invention, the term "antibody" refers to any antibody-like molecule having an antigen-binding region, including antibody fragments containing an antigen-binding domain, such as Fab, F(ab')2, single-domain antibodies (DABs), TandAbs dimers, Fv, scFv (single-chain Fv), dsFv, ds-scFv, Fd, linear antibodies, minibodies, diabodies, bispecific antibody fragments, bispecific antibodies, trispecific antibodies (scFv-Fab fusions, bispecific or trispecific); sc-bispecific antibodies; κ(λ) antibodies (scFv-CL fusions); DVD-Ig (dual variable domain antibodies, bispecific format); SIP (small immunoproteins, a type of minibody); SMIP (small modular immunopharmaceuticals); scFv-Fc dimers; DART (ds-stabilized bispecific antibodies, bispecific affinity retargeting antibodies); and all antibody mimics containing one or more CDRs, etc. Techniques for preparing and using various antibody-based constructs and fragments are well known in the art (see Kabat et al., 1991, which is hereby incorporated by reference into the present invention).

[0148] As an example, we employed a Fab fragment that binds to FAP (fibroblast activation protein) (the CH3 domain contains a knob mutation) and a Fab fragment that binds to HER2 (human epidermal growth factor receptor 2) (the CH3 domain contains a hole mutation and the CH2 domain contains an N297A mutation). To facilitate heterologous Fc assembly, we introduced a knob-into-hole structure or other CH3 / CH3 mutations that contribute to dimerization in the heavy chains (HCs).

[0149] In the present invention, the term "bispecific antibody" refers to an antibody that is specific for two target molecules, i.e., an antibody that can specifically bind to at least two different epitopes or two different targets (usually non-overlapping epitopes).

[0150] The term "multispecific antibody" refers to an antibody that is specific for more than one target molecule, i.e., an antibody that can specifically bind to at least two, three or more different epitopes or targets.

[0151] The term "IgG-type antibody" refers to an antibody belonging to the IgG immunoglobulin, which contains two light chains and two heavy chains. The light chain contains VL and CL domains, and the heavy chain contains VH, CH1, CH2 and CH3 domains. At the same time, it can also be defined as a four-chain protein containing a first chain, a second chain, a third chain and a fourth chain, where the first chain and the fourth chain are VL-CL chains, and the second chain and the third chain are VH-CH1-CH2-CH3 chains.

[0152] The term "human homologous / heterologous dimerization domain" refers to a domain from a human protein that can form a homologous / heterologous dimer as a part of a complete protein or a full-length protein.

[0153] The term "homologous / heterologous dimerization domain" refers to a homologous dimerization domain and / or a heterologous dimerization domain.

[0154] The term "homologous / heterologous dimer" refers to a homologous dimer and / or a heterologous dimer.

[0155] In the present invention, the term "knob-into-hole" refers to both a strategy for modifying the heavy chain homologous dimer of an antibody to achieve heterologous dimerization and a domain in which HC1 and HC2 carry "knob" and "hole" amino acid mutations respectively (or vice versa).

[0156] The term "trastuzumab" refers to the approved anti-human HER2 protein monoclonal antibody (Herceptin).

[0157] The term "SUMO tag" refers to the SUMO protein fused to the C-terminus of LC2.

[0158] The term "β2-microglobulin (β2m)" refers to the full-length human β2m protein (amino acids: 1-99).

[0159] The term "extracellular α3 domain of MHC class I molecule" refers to the α3 domain (amino acids: 180-278) of the human MHC class I protein.

[0160] The term "Wuxibody" refers to the bispecific antibody format described in Patent WO2019057122 A1, in which the CH1 / CL domain of half-antibody A is replaced by the TCRβ / α constant domain.

[0161] The term "anti-HER2 X anti-FAP bispecific antibody" refers to a bispecific antibody that can bind to the antigens HER2 and FAP, wherein the VH / VL of half-antibody A targets the antigen HER2 and the VH / VL of half-antibody B targets the antigen FAP.

[0162] The term "half-antibody" refers to a dimer in which a light chain binds to its cognate heavy chain.

[0163] The term "CrossMab" refers to a bispecific antibody format invented by Roche (US20170129962 A1), in which the CH1 / CL domains of half-antibody A are mutually exchanged.

[0164] The terms "Abio antibody", "Abio platform" and "Abio antibody platform" are used interchangeably in the present invention and refer to antibodies (including but not limited to bispecific and multispecific antibodies) in any format (whether IgG-type antibody or not), in which the CH1 and CL of the half-antibody are replaced by CD1s and β2m respectively (or vice versa). CD1s includes but is not limited to CD1a and CD1b, and CD1s and β2m can be wild-type or mutant.

[0165] In the present invention, the term "BsAbs" refers to bispecific immunoglobulin G formats that contain one binding part for each antigen. Exemplary BsAb formats include, but are not limited to: CrossMab, DAF (two-in-one), DAF (four-in-one), DutaMab, DT-IgG, shared light chain of the mortar and pestle structure, mortar and pestle structure assembly, charge pairing, DuoBody, SEEDbody, triomab, LUZ-Y, Fcab, κλ antibody, orthogonal Fab, etc. Approved BsAbs include: Amivantamab (GenMab, Janssen Biotech) against EGFR / against cMET, Teclistamab (GenMab and Jansen Biotech) against BCMA / against CD3, Emicizumab (Roche) against coagulation factor X / against coagulation factor IX, Fabricimab (Roche) against VEGFA / against ANG2, Mosunetuzumab (Roche) against CD20 / against CD3. In certain embodiments, the BsAb contains heavy chains engineered to enable heterodimerization, for example, by the knobs-into-holes (KIH) (Ridgway et al., 1996), CH3 electrostatic steering (Gunasekaran et al., 2010), SEED technology (Davis et al., 2010), XmAb bispecific platform (Moore et al., 2019), Azymetric platform (Escobar-Cabrera et al., 2017), DuoBody (Labrijn et al., 2013), and other heterologous Fc technologies.

[0166] The terms "Abio BsAb", "Abio BsIgG", "Abio bispecific antibody", "Abio bispecific antibody", "Abio BsAb format", and "Abio bispecific format" are used interchangeably in the present invention to refer to a bispecific immunoglobulin G in which the CH1 and CL of the half-antibody are replaced by CD1s and β2m, respectively, or vice versa. CD1s and β2m can be wild-type or mutant. In one embodiment, CD1s represents CD1a or CD1b. In certain embodiments, a tag, such as a SUMO tag or an Avi tag, is fused to the C-terminus of one light chain of the Abio BsAb.

[0167] The term "ABC173 format" or "173 format" refers to a bispecific immunoglobulin G in which the CH1 and CL of one half-antibody are replaced by CD1a and β2m, respectively.

[0168] The term "ABC174 format" or "174 format" refers to a bispecific immunoglobulin G in which the CH1 and CL of one of the half-antibodies are replaced by CD1b and β2m, respectively.

[0169] The terms "mutated", "mutation" and "mutant" are used interchangeably in the present invention. Generally, the mutated amino acid or mutation refers to the replacement, insertion, deletion or combination thereof of one amino acid by one or more amino acids. Preferably, the mutated amino acid or mutation refers to the replacement of a single amino acid by another different single amino acid.

[0170] Examples

[0171] The amino acid sequences of SEQ ID NOs: 1-47, 57-89 and 101-111 are shown in the following table.

[0172] Example 1: Design and engineering of antibodies with different human homologous / heterologous dimerization domains

[0173] 1. Materials, methods and strategies

[0174] 1) Antibodies

[0175] Antibody A is the anti-HER2 antibody Trastuzumab developed by Roche. Antibody A provides light chain 1 and heavy chain 1.

[0176] Antibody B is the anti-FAP antibody clone 3F2 developed by Roche, and its sequence is derived from patent WO 2012 / 020006. Antibody B provides light chain 2 and heavy chain 2.

[0177] 2) Bispecific antibodies (BsAbs)

[0178] All bispecific antibodies used in this experiment contain four chains: heavy chain 1, light chain 1, heavy chain 2, and light chain 2. Heavy chain 1 and light chain 1 can form half-antibody A, and heavy chain 2 and light chain 2 can form half-antibody B.

[0179] To solve the problem of heavy chain mispairing, we introduced a pestle (T336W)-mortar (T336S, L368A, Y407V) mutation into the CH3 domain of the bispecific antibody and added a stabilizing disulfide bond (S354C of heavy chain 1 and Y349C of heavy chain 2) to the CH3. Specifically, the heavy chain 1 of trastuzumab was cloned into the mortar chain containing the N297A mutation (SEQ ID NO:6), while the heavy chain 2 of the anti-FAP antibody was cloned into the pestle chain (SEQ ID NO:2). To better evaluate the assembly effect of heavy chain and light chain, a SUMO tag (SEQ ID NO:4) was fused to the C-terminus of the anti-FAP light chain 2.

[0180] The combinations of heavy and light chains of the bispecific antibody are summarized in Table 1. The sequences of the homologous / heterologous dimerization domains used to replace CH1 or CL and the tag sequences are summarized in Table 2.

[0181] Table 1

[0182] Table 2 Important sequences and tag sequences used to replace CH1 or CL

[0183] 3) Method: Expression and purification

[0184] The target DNA sequences of the heavy and light chains were inserted between the NheI and Xmal sites of the mammalian expression vector pCI (Promega #E1731). The bispecific antibody was produced by co-transfecting 4 plasmids (HC1, LC1, HC2, LC2) into Expi293F cells (ThermoFisher #A14527). Expi293F cells were cultured in suspension. The supernatant was collected by centrifugation six days after transfection and filtered through a 0.22 μm filter membrane. The antibody was purified from the filtered supernatant using protein A magnetic beads. The antibody concentration was measured at a wavelength of 280 nm using a Nanodrop (NanoDrop2000, Thermo scientific).

[0185] 4) Evaluation of heavy and light chain pairing: The corresponding LC / HC pairing was analyzed by SDS-PAGE

[0186] Method: SDS-PAGE Analysis of Bispecific Antibodies

[0187] After purification by Protein A, the assembly of bispecific antibodies was analyzed by SDS-PAGE. After mixing the samples with loading buffer, they were heated at 95 °C for 5 minutes with / without DTT and electrophoresed through a 4-20% gel in MOPS electrophoresis buffer (Bio-Rad). The gels were stained with Coomassie blue (Thermo Scientific) and then washed and decolorized with water. Equal amounts of protein (6 μg) were loaded for each sample. To better distinguish the two heavy chains, 20 μg of bispecific antibodies (ABC056, ABC057, ABC060, ABC061, ABC131, ABC132, and ABC133) were treated with 2 μl of PNGase F 500 U / μl (NEB#P0704L) at 37 °C for 1 hour for deglycosylation.

[0188] Results:

[0189] Bispecific antibodies can be produced by co-expressing two different heavy chains (HCs) and two different light chains (LCs) in a single host cell line by replacing the CH1 / CL of half-antibody A with other human homologous / heterologous dimerization domains or proteins (including non-IgG-like and IgG-like heterodimers). Non-IgG-like heterodimers used in the design include Fos / Jun, GABA1 / GABA2, and GABA2 / GABA1. IgG-like domains used for bispecific assembly include: interleukin receptor pairs (IL2Rβ / γ), CD3 subunit pairs, CD8a / CD8β, CH3 / CH3 of IgG1 (knob-into-hole structure), CH2 / CH2 of IgM (wild type and mutants), HLA-α3 / β2m, β2m / HLA-α3, SIPRα / CD47, β2m / CD1s-α3, and CD1s-α3 / β2m. All engineered bispecific antibodies were expressed at a molar ratio of HC1:LC1:HC2:LC2 of 1:1:1:1 during primary screening. After purification of the heterologous IgG1 mixture containing bispecific antibodies by Protein A magnetic beads, the corresponding LC / HC pairing (i.e., the binding of the heavy chain to its original paired light chain) was evaluated by SDS-PAGE gels.

[0190] The parental anti-HER2 / anti-FAP bispecific antibody without modification of the CH1 / CL domain of half-antibody A showed three bands at approximately 160 kDa ( Figure 1A, 1C, ABC055). The upper and lower bands represent a light chain mispairing product, and the middle band corresponds to the correct bispecific antibody configuration and / or light chain exchange product. Therefore, less than 30% of the correctly assembled bispecific antibody was observed in the IgG1 mixture.

[0191] Since there are no interchain disulfide bonds in the engineered CH1 / CL domain of the half-antibody A, the engineered light chain 1 (LC1) should be separated from the BsAb on a non-reducing SDS-PAGE gel. Therefore, the successfully assembled bispecific antibody should show two bands under non-reducing conditions: one corresponding to the HC1 / HC2 / LC2-SUMO configuration and the other corresponding to the unlinked LC1. In addition, four bands are expected to be detected under reducing conditions, although the two heavy chains may be difficult to distinguish due to their close molecular weights. Based on these criteria, the Fos / Jun (ABC064) group substitution failed because the engineered LC1 was not detected under non-reducing conditions ( Figure 1 A), indicating that the engineered LC1 did not assemble well with HC1 / HC2 / LC2-SUMO. The GABA1 / GABA2 group substitution (ABC074 and ABC075) showed the correct bands ( Figure 1 A). For the IgG-like domain replacement ( Figure 1 B), replacing the CH1 / CL with the IL2Rβ / γ (ABC058) and CD8a / CD8β (ABC134) domains may result in suboptimal assembly because LC1 was not observed under non-reducing conditions ( Figure 1 B).

[0192] As Figure 1 shown in Figure 1 C, the CH3 / CH3 knob-into-hole domain with interchain disulfide bonds between the CH3 / CH3 domains (ABC056 and ABC057) results in better bispecific antibody assembly compared to the wild type (

[0193] As Figure 1As shown in D, when the CH1 / CL domains of HC1 and LC1 were replaced with HLA-α3 / β2m or β2m / HLA-α3 (ABC060 and ABC061), SIRPα / CD47 (ABC131), β2m / CD1b (ABC132), CD1a / β2m (ABC133), good HC / LC pairing was observed because two expected bands appeared under non-reducing conditions and all four chains were detected under reducing conditions after PNGase F treatment. ABC133 (CD1a / β2m) is the most promising design with high purity of the bispecific product and minimal half-antibody content. Table 3 summarizes the IgG-like heterodimer substitutions of CH1 / CL showing correct HC / LC pairing.

[0194] Table 3. IgG-like homo / heterodimer replacements of CH1 / CL showing correct HC / LC pairing

[0195] Example 2: Orthogonal Fab pairing without optimizing the light chain ratio (correct pairing is insensitive to the light chain ratio)

[0196] Method: Expression and SDS-PAGE analysis

[0197] The CD1s / β2m design with a SUMO tag introduced into light chain 2 was further evaluated with an LC1 / LC2 plasmid ratio ranging from 3:1 to 1:3, while the HC1 / HC2 plasmid ratio was kept at 1:1 and an N297A mutation was introduced on both heavy chains. Four plasmids (HC1, LC1, HC2, LC2) were co-transfected into Expi293F cells (ThermoFisher #A14527) at different light chain plasmid ratios to prepare bispecific antibodies. Expi293F cells were cultured in suspension. The supernatant was collected by centrifugation six days after transfection and filtered through a 0.22 μm filter membrane. The IgG1 mixed product containing the bispecific antibody was purified by protein A magnetic beads and analyzed by SDS-PAGE gels under non-reducing and reducing conditions.

[0198] Materials: The antibodies used in the test and the HC1:LC1:HC2:LC2 ratios are shown in Table 4

[0199] Table 4

[0200] Results:

[0201] The CH1 / CL orthogonal Fab format of Genentech requires optimization of the chain ratio to ensure correct HC / LC pairing, as the ratio of correctly assembled bispecific antibodies is 63%-84% when the LC1 / LC2 ratio varies (Dillon et al., 2017).

[0202] As described above, engineered LC1 separates from the assembly of the other three chains on a non-reducing SDS-PAGE gel. As Figure 2 shown, regardless of whether β2m replaces CH1 or CL, the CD1s / β2m design shows two main bands (the band at approximately 135 kDa corresponds to HC1 / HC2 / LC2-SUMO, and the band at approximately 25 kDa corresponds to LC1). When the LC1 / LC2 ratio varies from 3:1 to 1:3, no other bands appear near the 135 kDa band. These results indicate that in our design, the light chain of the parental antibody can accurately assemble with its original paired heavy chain.

[0203] Example 3: Characterization of the Abio BsAb format by SEC-HPLC

[0204] Method:

[0205] The anti-HER2 / FAP bispecific antibody was produced by co-transfecting 4 plasmids (HC1, LC1, HC2, LC2) at different ratios. LC2 does not contain any SUMO tags, and both heavy chains contain the N297A mutation. The required plasmid combinations were transfected into Expi293F cells, and the supernatant was purified by protein A magnetic beads. The eluate was analyzed by SDS-PAGE gel under reducing and non-reducing conditions.

[0206] SEC-HPLC was used to evaluate the purity and monomer content of the bispecific antibody. Analysis was performed using an Agilent liquid chromatograph equipped with an AdvanceBio SEC300A 2.7 μm 4.6X 300 mm column (Agilent#PL1580-5301). The flow rate was 0.35 ml / min, the mobile phase was PBS (pH = 7.4), 2.5 μl of the protein A-purified bispecific antibody was injected, and isocratic elution was performed for 15 minutes. Elution was monitored by ultraviolet absorption at 280 nm. The percentage of each component was calculated based on the relative peak area detected by the software.

[0207] Table 5. Evaluation of the assembly of CD1s / β2m or β2m / CD1s bispecific antibody formats without SUMO tags

[0208] Results:

[0209] We evaluated the assembly of CD1s / β2m or β2m / CD1s bispecific antibody formats without the SUMO tag (hereinafter, the bispecific antibody in which CH1 / CL of the half antibody is replaced by CD1s / β2m or β2m / CD1s is referred to as the Abio bispecific antibody). All variants showed comparable expression levels to the parental antibody. As Figure 3 shown in A, in all four designs, two bands under non-reducing conditions and four bands under reducing conditions were observed as expected. The upper band is the bispecific antibody with engineered light chain deletion (about 130 kDa), and the lower band is engineered light chain 1 (about 23 kDa). These four CD1s / β2m bispecific antibodies were further characterized by SEC-HPLC ( Figure 3 B). First, the main monomer components of ABC171, ABC172, ABC173, and ABC174 were eluted, and the elution times were similar to those of the parental mAb. Second, the Abio bispecific antibody exhibited properties similar to those of the mAb, i.e., only a small amount of aggregation. In terms of the percentage of monomeric bispecific antibody, the quality of the Abio bispecific antibody (about 95% monomer) was comparable to that of the anti-FAP / HER2-CrossMab bispecific antibody (ABC203, 55%) and the anti-FAP / HER2-Wuxibody (ABC204, 92%). Although the HLA / β2m (ABC734) or β2m / HLA (ABC735) formats showed correct assembly and similar yields to the Abio bispecific antibody ( Figure 3 A and 3B), due to the polymorphism of HLA allotypes, they have a higher risk of immunogenicity ( Figure 3 C).

[0210] Example 4: Thermal stability of Abio bispecific antibody

[0211] Method: Differential scanning fluorimetry

[0212] The protein was purified by protein A and size exclusion chromatography (Cytiva #28-9909-44, Superdex 200, 10 / 300GL). Differential scanning fluorimetry was performed using a QuantStudio 6Flex (Thermo Fisher) and Protein ThermalShift TM dye (Thermo #4461146). A PBS solution (containing 12.5 μl of protein) with a concentration of 0.3 - 0.5 mg / ml was mixed with 2.5 μl of 8X dye solution (diluted 1:1000 in PBS buffer) and 5 μl of reaction buffer. The melting curve was measured in the range of 25 °C to 99 °C with a temperature scanning rate of 0.05 °C / s, and three replicate experiments were conducted. The melting temperature (Tm) represents the peak in the d(fluorescence) / dT plot applied to the experimental curve. The Tm values of the anti-HER2 / FAP BsAb variants and the corresponding monoclonal IgG (ABC001 FAP IgG and ABC002 FAP IgG N297A) as well as the bispecific antibody of the Wuxibody version (ABC204) are shown in Table 6. The T m 1 value of the Abio bispecific antibody is comparable to that of the bispecific antibody generated using a non-Abio platform.

[0213] Table 6: Melting temperatures of bispecific antibodies and monoclonal antibodies.

[0214] In the present invention, T m is defined as the temperature at which half of the protein is in the unfolded state and is an important protein property. It is known that deglycosylation of CH2 reduces T m 1. In fact, the T m 1 of anti-FAP hIgG1 (ABC002) containing the N297A mutation is 60.4 °C, lower than that of anti-FAP hIgG1 wild type (ABC001) at 71.33 °C and anti-HER2 hIgG wild type (ABC141) at 70.53 °C.

[0215] First, the anti-HER2 / FAP bispecific antibody variants (ABC171, ABC172, ABC173, ABC174) containing the N297A mutation have a T m 1 similar to that of the parental FAP IgG N297A mAb (ABC002), although their T m1. The peak is higher. The Tm of these bispecific antibodies is comparable to that of other human mAbs or humanized mAbs, ranging from 57°C to 82°C (Garber and Demarest, 2007). Secondly, our design has comparable T m 1 to the anti-HER2 / FAP-Wuxibody format (ABC204) and the anti-HER2 / FAP-HLA / β2m format (ABC734 and ABC735). m 1. Thirdly, the T

[0216] Example 5: Mass spectrometry analysis of IgG bispecific antibodies shows correct chain assembly

[0217] Method:

[0218] The chain assembly of bispecific antibodies was evaluated by mass spectrometry. Proteins were purified by Protein A and size exclusion chromatography (Cytiva #28-9909-44, Superdex 200, 10 / 300GL). Subsequently, desalting was performed using a C4 column (Waters, ACQUITY UPLC BEH300 C4 1.7um 2.1*50mm) on an ACQUITY UPLC system at a flow rate of 0.3 ml / min with a 10-minute gradient elution using mobile phase A (0.1% formic acid) and mobile phase B (0.1% formic acid and 100% acetonitrile) at 80°C. The desalted protein (about 1 mg / ml, 1 - 3 μL) was loaded into the autosampler and analyzed on a quadrupole time-of-flight (Q-TOF) liquid chromatography-mass spectrometry instrument (Xevo G2-XS QTof, Waters) with a spray voltage of 3000 - 3500 V. The data acquisition range was 500 - 4000 m / z. The raw electrospray mass spectra were combined, and then the multi-charged molecular ions were deconvoluted into molecular weight spectra by UNIFI software (1.8.2, Waters, MaxEnt1 processing). The molecular weights of the light and heavy chain components were determined using the ExPASy-ProtParam tool.

[0219] Results:

[0220] Since there are no inter-chain disulfide bonds in the engineered Fab, two mass spectrometry peaks were detected for ABC171: one peak matched the engineered anti-HER2 LC1 (the theoretical molecular weight of ABC171 is 233340.26 Da, and the measured molecular weight is 23340.00 Da), and the other peak matched the molecular weight of the remaining molecule (the theoretical molecular weight is 122622.17 Da, and the measured molecular weight is 122622.25 Da, Figure 5A). Bispecific antibodies in other Abio bispecific antibody formats also showed the correct molecular weight ( Figure 5 B-C).

[0221] Therefore, LC-MS data for all molecules indicated that the Abio bispecific format ensured orthogonal light chain pairing, and no mismatched components were detected.

[0222] Example 6: Dual binding of Abio bispecific antibodies is comparable to parental antibodies

[0223] Method: Octet assay for binding affinity

[0224] The binding kinetics of Abio bispecific antibodies to antigens were determined by Biolayer Interferometry (BLI) on an Octet R8 (Sartorius) at 30 °C and 1000 rpm. The sensors were incubated in the detection buffer (PBS 7.4 + 0.02% Tween 20 + 0.1% BSA) for at least 10 minutes before loading onto the instrument.

[0225] Human biotinylated HER2-Fc (Sino biological #10004-H02H-B) was diluted to 10 ng / ml in the detection buffer (PBS 7.4 + 0.02% Tween 20 + 0.1% BSA) and captured by an Octet streptavidin (SA) sensor tip (Sartorius ForteBio #18-5019). After the baseline was stable for 2 minutes, the sensor was immersed in the antibody solution for 2 minutes. Binding and dissociation experiments were performed at serial dilution concentrations (100 nM, 50 nM, 25 nM, 12.5 nM, 6.25 nM, 3.125 nM, 0 nM) of the bispecific antibodies and mAbs. The binding rate constant (ka), dissociation rate constant (kd), and equilibrium dissociation constant (KD) were determined by fitting all data using a 1:1 binding model with the instrument software.

[0226] The dual-antigen binding experiment of Abio bispecific antibodies was also performed on the Octet system. Human biotinylated HER2-Fc was immobilized on the Octet streptavidin (SA) sensor tip using the above method. After washing, the loaded biosensor was incubated with 100 nM bispecific antibody for 2 minutes, and then the tip was immersed in the second antigen (human FAP, Novoprotein #C14G) for 3 minutes. Monoclonal anti-HER2 (Herceptin, ABC141), anti-lysozyme.IgG1.S239D.I332E (ABC089), and the detection buffer were used as controls.

[0227] Results:

[0228] Determination of the binding affinity of bispecific antibodies to human HER2 by Biolayer Interferometry (BLI) Figure 6 ). The binding affinities of four formats of anti-HER2 / FAP bispecific antibodies to human HER2 were 1.05 ± 0.08 nM, 0.78 ± 0.07 nM, 0.43 ± 0.11 nM, and 0.61 ± 0.08 nM, respectively (Table 7), which were close to the apparent K D (0.23 ± 0.08 nM) of the parental Herceptin with two Fabs. These results indicate that the bispecific design of the present invention does not reduce the antigen-binding ability of the antibody.

[0229] In addition, we verified the dual-antigen binding ability of the Abio bispecific format. After the first antigen HER2 was loaded onto the tip, binding occurred for all bispecific antibodies and Herceptin mAb (ABC141). Subsequently, for the tips loaded with bispecific antibodies, binding to the second antigen FAP was observed, while this was not the case for Herceptin Figure 7 ). Therefore, the Abio bispecific antibody is capable of binding both HER2 and FAP simultaneously with an affinity comparable to that of the parental antibody.

[0230] Table 7: Summary of K D values Antibody <![CDATA[K D (nM)]]> ABC141 0.23±0.08 ABC171 1.05±0.08 ABC172 0.78±0.07 ABC173 0.43±0.11 ABC174 0.61±0.08

[0231] Example 7: The engineered Fab conformation in ABC173 is predicted to be similar to that of wild-type Fab

[0232] Methods:

[0233] The structure of the engineered Fab of ABC173 was predicted by AlphaFold-multimer and compared with the crystal structure of Herceptin Fab (PDB: 6MH2). The root mean square deviation (RMSD) was calculated based on backbone atoms, and all images were prepared using PyMOL software.

[0234] Results:

[0235] The structure of the engineered Fab of ABC173 was compared with that of the original Fab (PDB: 6MH2). As shown in Figure 8 A and 8B, the engineered Fab of ABC173 has a similar conformation to that of wild-type Fab. The alignment of the backbone atoms of these two structures shows an RMSD of ( Figure 8 C).

[0236] Example 8: The Abio bispecific antibody design can be widely applied to different antibody combinations

[0237] Methods:

[0238] Four plasmids (HC1, LC1, HC2, LC2) were co-transfected into Expi293F cells (ThermoFisher #A14527) to prepare bispecific antibodies. Both heavy chains contained the N297A mutation. Expi293F cells were cultured in suspension. Six days after transfection, the supernatant was collected by centrifugation and filtered through a 0.22 μm filter membrane. The antibody was purified from the filtered supernatant using protein A magnetic beads. After protein A purification, the assembled bispecific antibody was analyzed by SDS-PAGE to evaluate its purity and monomer content. SEC-HPLC analysis was performed using an Agilent liquid chromatography system equipped with an AdvanceBio SEC 300A 2.7 μm 4.6X 300 mm column (Agilent Technologies #PL1580-5301), with a flow rate of 0.35 ml / min and a mobile phase of PBS (pH = 7.4), and the elution process was monitored by ultraviolet absorption at 280 nm.

[0239] The ABC173 format, which is optimal in terms of chain assembly and purity, was applied to evaluate the affinity of three additional antibody combinations to verify the universality of this format. The first pair of antibody combinations was anti-mCD20 / mCD3 (ABC215), the second pair was anti-HER2 / hCD3 (ABC588), and the third pair was anti-hCCR8 / hCD3 (ABC591). In the mCD20 / CD3 bispecific antibody, the CH1 / CL of mCD20 was replaced by CD1a / β2m. Bispecific antibodies generated in the CrossMab and Wuxibody formats were also tested. All bispecific antibodies were prepared by co-expressing four plasmids in Expi293F cells at a molar ratio of 1:1:1:1 or 2:2:1:1 (Table 8).

[0240] Table 8: Information on bispecific antibodies used in Example 8

[0241] Results:

[0242] All tested antibody combinations (ABC215, ABC588, and ABC591) designed in the ABC173 format showed the expected bands Figure 9 A) under non-reducing conditions, with a high percentage of monomer content Figure 9 B), and the correct molecular weight was confirmed by LC-MS analysis. As Figure 9As shown in C, the intact molecular weight of ABC588 determined by mass spectrometry is consistent with the theoretical molecular weight, and no mismatched components are detected. The bispecific antibody in the ABC173 format seems to have a higher purity than the corresponding bispecific antibodies in the CrossMab or Wuxibody formats ( Figure 9 A).

[0243] Example 9: Improving the Thermal Stability of Abio Bispecific Antibodies by Further Engineering

[0244] Method: Mutations based on the ABC173 and ABC174 formats were introduced by standard site-directed mutagenesis techniques. The expression and purification methods of the variants were the same as those described in Example 3, and the chain ratio (LC1:HC1:LC2:HC2) was set to 1:1:1:1 or 2:2:1:1. After purification by protein A magnetic beads, the eluted protein was analyzed by non-reducing and reducing SDS-PAGE to examine the formation of the four-chain bispecific antibody. The purity and monomer content of the mutants were evaluated by SEC-HPLC, using an Agilent liquid chromatography system equipped with an AdvanceBio SEC 300A 2.7μm 4.6×300mm column (Agilent Technologies #PL1580-5301), with a flow rate of 0.35 ml / min and a mobile phase of PBS (pH = 7.4), and the elution process was monitored by ultraviolet absorption at 280 nm. Differential scanning fluorimetry analysis was performed according to the method described in Example 4.

[0245] Results:

[0246] To improve the biophysical properties of the Abio bispecific antibody, mutagenesis was performed on the interface between the CD1a α3 domain and β2m, and six pairs of cysteine mutations were tested ( Figure 10 A). The numbering is based on the CD1a / β2m crystal structure (PDB ID 1XZO). CD1a D240C / β2m R12C (ABC270), CD1a S238C / β2m R12C (ABC271), CD1a D234C / β2m Q8C (ABC272), CD1a G194C / β2m M99C (ABC373), CD1a W190C / β2m P14C (ABC374), and CD1a G194C / β2m 100C (ABC405) variant. The detailed chain information is summarized in Table 9.

[0247] Table 9

[0248] *β2m 100C Indicates the addition of an extra cysteine at the end of the β2m sequence.

[0249] As Figure 10 shown in B, full assembled four-chain bispecific antibodies with a molecular weight of 150 kDa were observed for all variants with cysteine mutations introduced at the CD1a / β2m interface. Although disulfide bonds were formed in all designed molecules, there were differences in the efficiency and homogeneity of disulfide bond formation.

[0250] Table 10

[0251] The thermal stability (T m 1 and T m 2) and yields of the antibodies are summarized in Table 10. Compared with the original molecule ABC173, all six pairs of cysteine mutations in CD1s / β2m showed an increase in T m 1, while T m 2 remained unchanged. These results indicate that introducing cysteine mutations in the CD1s / β2m heterodimerization domain can improve thermal stability. The CD1a G194C / β2m 100C mutation combination (ABC405) increased T m 1 by 2.6 °C.

[0252] Secondly, mutagenesis may affect the yield and solubility of engineered bispecific antibodies. For example, the expression level of ABC270 was significantly reduced ( Figure 10 B and 10C, ABC270), while the yield of ABC405 was higher than that of the parental ABC173. Therefore, in terms of expression yield and thermal stability, the bispecific format of ABC405 is the best design.

[0253] In previous bispecific molecules, the C-terminus of the CD1a α3 domain contained amino acids (GSLVPRG) that did not belong to the CD1a α3 domain. To reduce immunogenicity, the LVPRG amino acids were removed, and only GS was retained as the linker between the CD1a α3 domain and the CH2 domain. The resulting molecule ABC478 derived from ABC405 had a similar thermal stability to ABC405.

[0254] Based on the ABC478 format, we conducted a second round of optimization to improve the yield and stability of bispecific antibodies. By introducing mutations to form hydrogen bond and / or salt bridge interactions, the interaction between the CD1sα3 domain and β2m was improved (ABC603, ABC672, ABC673). As Figure 11 shown in A, all variants showed the expected 150 kDa band, and T m 1 was similar to ABC478 (Table 11). HPLC-SEC analysis indicated that these variants had little aggregation ( Figure 11 C).

[0255] The same strategy was used to engineer the ABC174 format. As Figure 11 shown in B and 11D, non-reducing SDS-PAGE showed that all variants derived from ABC174 formed the expected 150 kDa tetrabody.

[0256] Table 11

[0257] As shown in Table 11, all variants of ABC173 and ABC174 showed good stability. In particular, the T m 1 of ABC173-derived molecules was increased by 2.2 °C to 2.6 °C compared to ABC173, while the T m 1 of ABC174-derived molecules was increased by 3.7 - 4.3 °C compared to ABC174. Among them, ABC405, ABC478, ABC603, ABC672, and ABC673 had the best biophysical properties in terms of expression level and thermal stability.

[0258] Subsequently, bispecific antibodies with wild-type CH2 (ABC736, ABC737, ABC738, and ABC739) were constructed, and the thermal stability of these constructs was measured. The T m 1 of the bispecific antibodies was approximately 71 °C, and the T m 1 of the parental anti-FAP and anti-HER2 hIgGWT was also approximately 71 °C ( Figure 4 ). Therefore, the engineered bispecific antibodies had the same stability as natural antibodies.

[0259] To determine which melting peak corresponded to the unfolding of which domain, Fab variants of ABC603, ABC673, and ABC675 (ABC747, ABC748, and ABC749) were constructed and compared with Herceptin wild-type Fab (ABC750). The T mis approximately 71 °C, similar to the T of the CH2 domain m of Herceptin wild-type Fab, while the T m of Herceptin wild-type Fab is 85.5 °C ( Figure 12 G and 12H). Therefore, the first melting peak of the bispecific antibody corresponds to the unfolding peak common to the CH2 domain and the engineered Herceptin Fab.

[0260] Compared with MHC / β2m in Patent WO2022166728, the Abio bispecific antibody format of the present invention has the following advantages: 1) Due to the lack of polymorphism in CD1s, the bispecific antibody of the Abio platform of the present application has lower immunogenicity; 2) The T m 1 of the Abio bispecific antibody molecule using engineered CD1a / β2m is approximately 71 °C, comparable to the stability of mAb hIgG1, while the T m 1 of the reported MHL147-33322-IgG1-F118A is 67 °C.

[0261] Example 10: The Abio bispecific antibody has cell-based binding affinity comparable to that of the parental antibody

[0262] Method: Cell-based binding experiments were performed to determine the binding affinity of the bispecific antibody

[0263] The bispecific antibody and the parental antibody were expressed and purified according to the method described in Example 3, and the chain ratio (LC1:HC1:LC2:HC2) was 2:2:1:1. All antibodies were first purified by Protein A and then by SEC.

[0264] Antibody binding to the antigen was detected by flow cytometry (FACS): Cells expressing the antigen were dispensed at 300,000 per well into a 96-well U-bottom plate and incubated on ice for 15 minutes with 50 μl of Fc receptor blocking solution (Human TruStain FcX TM , BioLegend #422302, diluted 1:40). Subsequently, serially diluted antibodies or bispecific antibodies (highest final concentration 1000 nM, 3-fold serial dilution with PBS buffer) were added, incubated on ice for 30 minutes, washed twice with FACS buffer (PBS buffer containing 0.5 mg / ml BSA and 0.02% Tween), incubated on ice for 30 minutes with FITC anti-human IgG antibody (goat anti-human (IgG Fc) secondary antibody FITC, Abcam #ab97224, 1:1000), washed twice with FACS buffer, stained with 7-AAD (live cell staining solution, Invitrogen #00-6993-50), and finally resuspended in 150 μl of FACS buffer for flow cytometry analysis. Aurora). The data was analyzed using GraphPad Prism. The median fluorescence intensity was plotted against the antibody concentration, and the EC50 was calculated by fitting the curve with the log(agonist) vs. response-Variable slope (four-parameter) function.

[0265] Materials:

[0266] Cell lines: SK-BR-3 cell line (SKBR3, a cell line with high expression of HER2 antigen, ATCC#HTB-30), Jurkat clone E6-1 cell line (expressing hCD3 antigen, ATCC#TIB-152), 293T cell line transfected with hCCR8.

[0267] Table 12 *ABC722 is a negative control antibody.

[0268] Results:

[0269] Another two pairs of antibody combinations were generated in formats with improved thermal stability (such as ABC603, ABC605, ABC673, and ABC675) to evaluate their universality. Mutations based on the ABC603 and ABC605 formats were introduced into ABC588 to generate ABC723 (ABC723_HER2 / hCD3_603) and ABC724 (ABC724_HER2 / hCD3_605) respectively. Similarly, mutations based on the ABC673 and ABC675 formats were introduced into ABC591 to generate ABC752 (ABC752_hCCR8 / hCD3_673) and ABC753 (ABC753_hCCR8 / hCD3_675) respectively.

[0270] The antigen-binding ability of the generated bispecific antibodies was evaluated by flow cytometry (FACS). The EC50 values of the bispecific antibodies are summarized in Table 13. All bispecific antibodies showed comparable binding affinities with their corresponding parental antibodies ( Figure 13 ), indicating the broad applicability of the method of replacing CH1 / CL with engineered CD1s / β2m domains.

[0271] Table 13 EC50 values of bispecific antibodies against their corresponding antigens.

[0272] Davis, J.H., Aperlo, C., Li, Y., Kurosawa, E., Lan, Y., Lo, K.M., and Huston, J.S. (2010). SEED bodies: fusion proteins based on strand-exchange engineered domain (SEED) CH3 heterodimers in an Fc analogue platform for asymmetric binders or immunofusions and bispecific antibodies. Protein Eng Des Sel 23, 195 - 202. 10.1093 / protein / gzp094.

[0273] Dillon, M., Yin, Y., Zhou, J., McCarty, L., Ellerman, D., Slaga, D., Junttila, T.T., Han, G., Sandoval, W., Ovacik, M.A., et al. (2017). Efficient production of bispecific IgG of different isotypes and species of origin in single mammalian cells. MAbs 9, 213 - 230. 10.1080 / 19420862.2016.1267089.

[0274] Escobar-Cabrera, E., Lario, P., Baardsnes, J., Schrag, J., Durocher, Y., and Dixit, S. (2017). Asymmetric Fc Engineering for Bispecific Antibodies with Reduced Effector Function. Antibodies (Basel) 6. 10.3390 / antib6020007.

[0275] Gunasekaran, K., Pentony, M., Shen, M., Garrett, L., Forte, C., Woodward, A., Ng, S. B., Born, T., Retter, M., Manchulenko, K., et al. (2010). Enhancing antibody Fc heterodimer formation through electrostatic steering effects: applications to bispecific molecules and monovalent IgG. J Biol Chem 285, 19637 - 19646. 10.1074 / jbc.M110.117382.

[0276] Kitazawa, T., and Shima, M. (2020). Emicizumab, a humanized bispecific antibody to coagulation factors IXa and X with a factor VIIIa - cofactor activity. Int J Hematol 111, 20 - 30. 10.1007 / s12185 - 018 - 2545 - 9.

[0277] Labrijn, A. F., Meesters, J. I., de Goeij, B. E., van den Bremer, E. T., Neijssen, J., van Kampen, M. D., Strumane, K., Verploegen, S., Kundu, A., Gramer, M. J., et al. (2013). Efficient generation of stable bispecific IgG1 by controlled Fab - arm exchange. Proc Natl Acad Sci U S A 110, 5145 - 5150. 10.1073 / pnas.1220145110.

[0278] Lewis, S.M., Wu, X., Pustilnik, A., Sereno, A., Huang, F., Rick, H.L., Guntas, G., Leaver-Fay, A., Smith, E.M., Ho, C., et al. (2014). Generation of bispecific IgG antibodies by structure-based design of an orthogonal Fab interface. Nat Biotechnol 32, 191 - 198. 10.1038 / nbt.2797.

[0279] Liu, Z., Leng, E.C., Gunasekaran, K., Pentony, M., Shen, M., Howard, M., Stoops, J., Manchulenko, K., Razinkov, V., Liu, H., et al. (2015). A novel antibody engineering strategy for making monovalent bispecific heterodimeric IgG antibodies by electrostatic steering mechanism. J Biol Chem 290, 7535 - 7562. 10.1074 / jbc.M114.620260.

[0280] Moore, G.L., Bernett, M.J., Rashid, R., Pong, E.W., Nguyen, D.T., Jacinto, J., Eivazi, A., Nisthal, A., Diaz, J.E., Chu, S.Y., et al. (2019). A robust heterodimeric Fc platform engineered for efficient development of bispecific antibodies of multiple formats. Methods 154, 38 - 50. 10.1016 / j.ymeth.2018.10.006.

[0281] Regula, J.T., Imhof-Jung, S., Molhoj, M., Benz, J., Ehler, A., Bujotzek, A., Schaefer, W., and Klein, C. (2018). Variable heavy-variable light domain and Fab-arm CrossMabs with charged residue exchanges to enforce correct light chain assembly. Protein Eng Des Sel 31, 289-299. 10.1093 / protein / gzy021.

[0282] Ridgway, J.B., Presta, L.G., and Carter, P. (1996). 'Knobs-into-holes' engineering of antibody CH3 domains for heavy chain heterodimerization. Protein Eng 9, 617-621. 10.1093 / protein / 9.7.617.

[0283] Spiess, C., Merchant, M., Huang, A., Zheng, Z., Yang, N.Y., Peng, J., Ellerman, D., Shatz, W., Reilly, D., Yansura, D.G., and Scheer, J.M. (2013). Bispecific antibodies with natural architecture produced by co-culture of bacteria expressing two distinct half-antibodies. Nat Biotechnol 31, 753-758. 10.1038 / nbt.2621.

[0284] Zhao, C., Zhang, W., Gong, G., Xie, L., Wang, M.W., and Hu, Y. (2021). A new approach to produce IgG4-like bispecific antibodies. Sci Rep 11, 18630. 10.1038 / s41598-021-97393-2.

Claims

1. A novel protein comprising two pairs of polypeptides, wherein each pair of polypeptides consists of a heavy chain and a light chain, the heavy chain comprises a VH domain and a CH1 domain, and the light chain comprises a VL domain and a CL domain; and the VH domain and the VL domain of the first pair of polypeptides form a first Fv fragment that specifically binds to a first antigen, and the VH domain and the VL domain of the second pair of polypeptides form a second Fv fragment that specifically binds to a second antigen, and at least one pair of dimerization domains is introduced into the first pair of polypeptides, and each pair of dimerization domains forms a dimer alone or together with other proteins, wherein the dimerization domain is a homodimerization domain or a heterodimerization domain.

2. The novel protein according to claim 1, wherein at least one pair of the heterodimerization domains is derived from humans or other mammals, and the heterodimerization domain is not a TCRα / β or TCRβ / α constant domain, preferably, the heterodimerization domain is not a TCRα / β or TCRβ / α constant domain or an HLA / β2m or β2m / HLA domain.

3. The novel protein according to claim 1, wherein at least one pair of the heterodimerization domains is selected from: a pair of HLA and β2m, a pair of β2m and CD1b, a pair of CD1a and β2m, a pair of GABA1 and GABA2, or a pair of SIRPα and CD47.

4. The novel protein according to claim 1, wherein the heterodimerization domain is selected from at least one of the combinations of HLA / β2m, β2m / HLA, CD1b / β2m, β2m / CD1b, β2m / CD1a, and CD1a / β2m.

5. The novel protein according to claim 1, wherein the heterodimerization domain is selected from β2m / CD1b, or CD1b / β2m, or β2m / CD1a, or CD1a / β2m, preferably, the CH1 domain of heavy chain 1 is replaced by the α3 domain of CD1a, and the CL domain of light chain 1 is replaced by β2m.

6. The novel protein according to claim 5, wherein β2m / CD1a, CD1a / β2m, β2m / CD1b, or CD1b / β2m contains a mutation.

7. The novel protein according to claim 6, wherein β2m / CD1a, CD1a / β2m, β2m / CD1b, or CD1b / β2m contains one or more disulfide bonds by introducing a cysteine mutation.

8. The novel protein according to claim 1, wherein at least one pair of the homodimerization domains is derived from humans or other mammals, and the homodimerization domain is selected from IgG CH3 / CH3 domain, IgM CH2 / CH2 domain, and IgE CH2 / CH2 domain.

9. The novel protein according to claim 1, wherein the novel protein comprises four polypeptide chains, namely heavy chain 1, light chain 1, heavy chain 2, and light chain 2, Heavy chain 1 and heavy chain 2 contain a VH domain, a CH1 domain, a CH2 domain, and a CH3 domain and respectively form a VH-CH1-CH2-CH3 long chain; light chain 1 and light chain 2 contain a VL domain and a CL domain and respectively form a VL-CL short chain. Heavy chain 1 and light chain 1 are derived from antibody A, and a single heavy chain 1 pairs with a single light chain 1 to form a half-antibody A; heavy chain 2 and light chain 2 are derived from antibody B, and a single heavy chain 2 pairs with a single light chain 2 to form a half-antibody B; and wherein the CH1 domain of heavy chain 1 pairs with the CL domain of light chain 1, named the CH1 / CL combination in half-antibody A, and the CH1 domain of heavy chain 2 pairs with the CL domain of light chain 2, named the CH1 / CL combination in half-antibody B.

10. The novel protein according to any one of claims 1-9 further comprises one or more modifications that enhance the correct assembly and thermal stability of the novel protein.

11. The novel protein according to claim 10, wherein the modification comprises a mortise and tenon structure.

12. The novel protein according to claim 9, wherein part of half-antibody A and / or part of half-antibody B are replaced by at least one pair of homologous / heterologous dimerization domains, and each pair of homologous / heterologous dimerization domains alone or together with other proteins forms a homodimer or a heterodimer. Preferably, the CH1 / CL in half-antibody A and / or the CH1 / CL in half-antibody B are replaced by at least one pair of homologous / heterologous dimerization domains.

13. The novel protein according to claim 12, wherein the CH2 of heavy chain 1 and the CH2 of heavy chain 2 are replaced by at least one pair of homologous / heterologous dimerization domains.

14. The novel protein according to claim 12, wherein the CH3 of heavy chain 1 and the CH3 of heavy chain 2 are replaced by at least one pair of homologous / heterologous dimerization domains.

15. The novel protein according to claim 9, wherein at least one pair of the homologous / heterologous dimerization domains are inserted into the novel protein, and each pair of the homologous / heterologous dimerization domains alone or together with other proteins forms a homodimer or a heterodimer.

16. The novel protein according to claim 15, wherein at least one pair of the homologous / heterologous dimerization domains are directly inserted between the CH1 / CL domain and the CH2 / CH2 domain.

17. The novel protein according to claim 15, wherein at least one pair of the homologous / heterologous dimerization domains are directly inserted between the CH1 / CL domain and the VH / VL domain.

18. The novel protein according to claim 15, wherein at least one pair of the homologous / heterologous dimerization domains are directly inserted after the CH3 / CH3 domain.

19. The novel protein according to claim 18, wherein at least one pair of the homologous / heterologous dimerization domains are fused with the novel protein, and each pair of homologous / heterologous dimerization domains alone or together with other proteins forms a homodimer or a heterodimer.

20. The novel protein according to claim 19, wherein at least one pair of said homologous / heterologous dimerization domains is fused to any two chains of the novel protein.

21. The novel protein according to claim 19, wherein at least one pair of said homologous / heterologous dimerization domains is fused to light chain 1 and / or heavy chain 1.

22. The novel protein according to claim 19, wherein at least one pair of said homologous / heterologous dimerization domains is fused to heavy chain 1 and / or heavy chain 2.

23. The novel protein according to claim 19, wherein at least one pair of said homologous / heterologous dimerization domains is fused to light chain 2 and / or heavy chain 2.

24. The novel protein according to claim 19, wherein two pairs of homologous / heterologous dimerization domains are respectively fused to light chain 1 and heavy chain 1 in said half - antibody A, and light chain 2 and heavy chain 2 in said half - antibody B.

25. The novel protein according to claim 9, wherein CH1 / CL in said half - antibody A and / or CH1 / CL in said half - antibody B is replaced by at least one pair of homologous / heterologous dimerization domains, and each pair of homologous / heterologous dimerization domains forms a homodimer or a heterodimer alone or in combination with other proteins.

26. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1a D240C and the CL domain of light chain 1 is replaced by β2m R12C ​ 27. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1a S238C and the CL domain of light chain 1 is replaced by β2m R12C ​ 28. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1a D234C and the CL domain of light chain 1 is replaced by β2m Q8C ​ 29. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C and the CL domain of light chain 1 is replaced by β2m M99C ​ 30. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1a W190C and the CL domain of light chain 1 is replaced by β2m P14C ​ 31. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C and the CL domain of light chain 1 is replaced by β2m 100C ​ 32. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C and the CL domain of light chain 1 is replaced by β2m 100C / Q8K ​ 33. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C-G233S and the CL domain of light chain 1 is replaced by β2m 100C / Q8K ​ 34. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C-G233S / D234E and the CL domain of light chain 1 is replaced by β2m 100C / Q8K ​ 35. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1b G194C and the CL domain of light chain 1 is replaced by β2m 100C / Q8K and replaced 36. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1b G194C-G233S and the CL domain of light chain 1 is replaced by β2m 100C / Q8K ​ 37. The novel protein according to claim 9, wherein the CH1 domain of heavy chain 1 is replaced by CD1b G194C / G233S / D234E and the CL domain of light chain 1 is replaced by β2m 100C / Q8K ​ 38. The novel protein according to claim 1, which is selected from protein ABC060, ABC061, ABC570, ABC571, ABC572, ABC132, ABC133, ABC074, ABC075, ABC131, ABC171, ABC172, ABC173, ABC174, ABC215, ABC588, ABC591, ABC270, ABC271, ABC272, ABC373, ABC374, ABC405, ABC478, ABC603, ABC672, AB C673, ABC513, ABC604, ABC605, ABC674, ABC675, ABC736, ABC737, ABC738, ABC739, ABC723, ABC724, ABC752 and ABC753, Among them, ABC060 includes four chains shown in SEQ ID NOs: 13, 11, 2, and 4; ABC061 includes four chains shown in SEQ ID NOs: 14, 12, 2, and 4; ABC570 includes four chains shown in SEQ ID NOs: 82, 83, 3, and 4; ABC571 includes four chains shown in SEQ ID NOs: 84, 83, 3, and 4; ABC572 includes four chains shown in SEQ ID NOs: 82, 85, 3, and 4; ABC132 includes four chains shown in SEQ ID NOs: 14, 26, 2, and 4; ABC133 includes four chains shown in SEQ ID NOs: 27, 11, 2, and 4; ABC074 includes four chains shown in SEQ ID NOs: 22, 21, 2, and 4; ABC075 includes four chains shown in SEQ ID NOs: 24, 23, 2, and 4; ABC131 includes four chains shown in SEQ ID NOs: 19, 20, 2, and 4; ABC171 includes four chains shown in SEQ ID NOs: 14, 25, 3, and 1; ABC172 includes four chains shown in SEQ ID NOs: 14, 26, 3, and 1; ABC173 includes four chains shown in SEQ ID NOs: 27, 11, 3, and 1; ABC174 includes four chains shown in SEQ ID NOs: 28, 11, 3, and 1; ABC734 includes four chains shown in SEQ ID NOs: 13, 11, 3, and 1; ABC735 includes four chains shown in SEQ ID NOs: 14, 12, 3, and 1; ABC215 includes four chains shown in SEQ ID NOs: 38, 41, 45, and 44; ABC588 includes four chains shown in SEQ ID NOs: 27, 11, 47, and 46; ABC591 includes four chains shown in SEQ ID NOs: 36, 37, 47, and 46; ABC270 includes four chains shown in SEQ ID NOs: 66, 65, 3, and 1; ABC271 includes four chains shown in SEQ ID NOs: 67, 65, 3, and 1; ABC272 includes four chains shown in SEQ ID NOs: 68, 69, 3, and 1; ABC373 includes four chains shown in SEQ ID NOs: 73, 71, 3, and 1; ABC374 includes four chains shown in SEQ ID NOs: 72, 70, 3, and 1; ABC405 includes four chains shown in SEQ ID NOs: 73, 74, 3, and 1; ABC478 includes four chains shown in SEQ ID NOs: 76, 74, 3, and 1; ABC603 includes four chains shown in SEQ ID NOs: 76, 75, 3, and 1; ABC672 includes four chains shown in SEQ ID NOs: 77, 75, 3, and 1.ABC673 comprises four chains shown in SEQ ID NOs: 78, 75, 3, and 1. ABC513 comprises four chains shown in SEQ ID NOs: 79, 74, 3, and 1. ABC604 comprises four chains shown in SEQ ID NOs: 80, 74, 3, and 1. ABC605 comprises four chains shown in SEQ ID NOs: 80, 75, 3, and 1. ABC674 comprises four chains shown in SEQ ID NOs: 81, 75, 3, and 1. ABC675 comprises four chains shown in SEQ ID NOs: 59, 75, 3, and 1. ABC736 comprises four chains shown in SEQ ID NOs: 86, 75, 2, and 1. ABC737 comprises four chains shown in SEQ ID NOs: 87, 75, 2, and 1. ABC738 comprises four chains shown in SEQ ID NOs: 88, 75, 2, and 1. ABC739 comprises four chains shown in SEQ ID NOs: 89, 75, 2, and 1. ABC723 comprises four chains shown in SEQ ID NOs: 76, 75, 47, and 46. ABC724 comprises four chains shown in SEQ ID NOs: 59, 75, 47, and 46. ABC752 comprises four chains shown in SEQ ID NOs: 106, 105, 47, and 46. ABC753 comprises four chains shown in SEQ ID NOs: 107, 105, 47, and 46., 39. The novel protein according to claim 9, wherein antibody A and antibody B are the same or different.

40. The novel protein according to claim 9, wherein antibody A or antibody B is selected from any one of cytotoxic antibodies, cell proliferation inhibitors, cell activation and interaction regulators, human immune system regulators, and antigen neutralizers.

41. The novel protein according to claim 9, wherein antibody A or antibody B is selected from the group consisting of: anti-HER2 antibody, anti-CCR8 antibody, anti-FAP antibody, anti-OX-40 antibody, anti-41BB antibody, anti-angiopoietin-2 antibody, anti-IL-4Rα antibody, anti-BCMA antibody, anti-Blys antibody, anti-BTNO2 antibody, anti-C5 antibody, anti-CD122 antibody, anti-CD13 antibody, anti-CD133 antibody, anti-CD137 antibody, anti-CD138 antibody, anti-CD16a antibody, anti-CD19 antibody, anti-CD20 antibody, anti-CD22 antibody, anti-CD27 antibody, anti-CD28 antibody, anti-CD3 antibody, anti-CD30 antibody, anti-CD33 antibody, anti-CD38 antibody, anti-CD40 antibody, anti-CD47 antibody, anti-CD8 antibody, anti-CEA antibody, anti-CGPR / CGRPR antibody, anti-CSPGs antibody, anti-CTLA4 antibody, anti-CTLA-4 domain antibody, anti-DLL-4 antibody, anti-EGFR antibody, anti-EpCAM antibody, anti-factor IXa antibody, anti-factor X antibody, anti-GITR antibody, anti-GP130 antibody, anti-Her3 antibody, anti-HSG antibody, anti-ICOS antibody, anti-IGF1 antibody, anti-IGF1 / 2 antibody, anti-IGF-1R antibody, anti-IGF2 antibody, anti-IGFR antibody, anti-IL-1 antibody, anti-IL-12 antibody, anti-IL-12p40 antibody, anti-IL-13 antibody, anti-IL-17A antibody, anti-IL-1β antibody, anti-IL-23 antibody, anti-IL-5 antibody, anti-IL-6 antibody, anti-IL-6R antibody, anti-Lag-3 antibody, anti-LAG3 antibody, anti-MAG antibody, anti-Met antibody, anti-NgR antibody, anti-NogoA antibody, anti-OMGp antibody, anti-OX40 antibody, anti-PD-1 antibody, anti-PDGFR antibody, anti-PDL-1 antibody, anti-PSMA antibody, anti-RGMA antibody, anti-RGMB antibody, anti-SARS-CoV-2 antibody, anti-Te38 antibody, anti-TIM-3 antibody, anti-TNF antibody, anti-TNFα antibody, anti-TROP-2 antibody, anti-TWEAK antibody, anti-VEGF antibody and anti-VEGFR antibody.

42. The novel protein according to claim 9, wherein antibody B is an anti-FAP antibody and antibody A is an anti-HER2 antibody.

43. The novel protein according to claim 9, wherein the half-antibody B is unchanged or altered.

44. The novel protein according to claim 9, wherein a tag is fused to the half-antibody B.

45. The novel protein according to claim 44, wherein the tag is fused to the C-terminus of the light chain 2 of the half-antibody B.

46. The novel protein according to claim 44, wherein the tag is used for purifying or labeling the novel protein.

47. The novel protein according to claim 44, wherein the tag is selected from the group consisting of: SUMO tag, HIS tag, Flag tag, HA tag, MYC tag, SBP tag, CBD tag, GST tag, MBP tag, pMAL tag, IMPACT tag, Protein A, and GFP.

48. The novel protein according to claim 44, wherein the SUMO tag is fused to the C-terminus of the light chain 2 of half antibody B and separated by a thrombin cleavage sequence.

49. The novel protein according to claim 9, wherein half antibody A is assembled with half antibody B in any form.

50. The novel protein according to claim 9, wherein half antibody A is assembled with half antibody B in the form of a knob-into-hole structure.

51. The novel protein according to claim 9, wherein half antibody A is assembled with half antibody B through CH3 with a knob-into-hole mutation in heavy chain 1 or heavy chain 2.

52. An isolated polynucleotide encoding heavy chain 1, light chain 1, heavy chain 2, or light chain 2 of the novel protein according to any one of claims 1-51.

53. A set of isolated polynucleotides comprising polynucleotides encoding heavy chain 1, light chain 1, heavy chain 2, and light chain 2 of the novel protein according to any one of claims 1-51, respectively.

54. An isolated vector comprising the isolated polynucleotide according to claim 52.

55. A host cell comprising the isolated polynucleotide according to claim 52, the set of isolated polynucleotides according to claim 53, or the isolated vector according to claim 54.

56. A pharmaceutical composition comprising the novel protein according to any one of claims 1-51, the isolated polynucleotide according to claim 52, the set of isolated polynucleotides according to claim 53, the isolated vector according to claim 54, or the host cell according to claim 55, and a pharmaceutically acceptable carrier.

57. Use of the novel protein according to any one of claims 1-51, the isolated polynucleotide according to claim 52, the set of isolated polynucleotides according to claim 53, the isolated vector according to claim 54, the host cell according to claim 55, or the pharmaceutical composition according to claim 56 in the preparation of a drug for preventing or treating a disease, or in the preparation of a kit for diagnosing a disease.

58. A method for preventing or treating a disease in a subject in need thereof, comprising administering to the subject a therapeutically effective amount of the novel protein according to any one of claims 1-51, the isolated polynucleotide according to claim 52, the set of isolated polynucleotides according to claim 53, the isolated vector according to claim 54, the host cell according to claim 55, or the pharmaceutical composition according to claim 56.

59. A method for preparing the novel protein according to any one of claims 1-51, comprising co-introducing the expression vector of the heavy chain 1, the expression vector of the light chain 1, the expression vector of the heavy chain 2 and the expression vector of the light chain 2 into an expression host cell, or separately introducing combinations of these expression vectors into different expression host cells, and expressing under appropriate conditions.

60. The method according to claim 59, wherein the correct pairing of the novel protein is insensitive to the molar ratio of the heavy chain 1 expression vector: light chain 1 expression vector: heavy chain 2 expression vector: light chain 2 expression vector; preferably, the molar ratio of the heavy chain 1 expression vector: light chain 1 expression vector: heavy chain 2 expression vector: light chain 2 expression vector is not limited; more preferably, the molar ratio of the heavy chain 1 expression vector: light chain 1 expression vector: heavy chain 2 expression vector: light chain 2 expression vector is 1:1:1:1 or 2:2:1:1 or 1:1:2:2 or 1:3:1:1, or 1:1:1:

3.

61. The method according to claim 59 or 60, wherein the host cell is a eukaryotic cell.

62. The method according to claim 59 or 60, wherein the host cell is a mammalian cell.

63. A method for preparing a novel protein having substantially the same or improved thermal stability compared to the original protein, in particular, the novel protein has substantially the same or improved Tm compared to the original protein, wherein the original protein comprises four polypeptide chains, namely heavy chain 1, light chain 1, heavy chain 2 and light chain 2, wherein heavy chain 1 and heavy chain 2 comprise a VH domain, a CH1 domain, a CH2 domain and a CH3 domain, and respectively form a VH-CH1-CH2-CH3 long chain; light chain 1 and light chain 2 comprise a VL domain and a CL domain, and respectively form a VL-CL short chain; wherein heavy chain 1 and light chain 1 are derived from antibody A, and a single heavy chain 1 pairs with a single light chain 1 to form a half-antibody A; heavy chain 2 and light chain 2 are derived from antibody B, and a single heavy chain 2 pairs with a single light chain 2 to form a half-antibody B; wherein the CH1 domain of heavy chain 1 pairs with the CL domain of light chain 1, named the CH1 / CL combination of half-antibody A, and the CH1 domain of heavy chain 2 pairs with the CL domain of light chain 2, named the CH1 / CL combination of half-antibody B; the method is characterized in that the CH1 / CL combination of half-antibody A is replaced by a homologous dimerization domain or other heterologous dimerization domains.

64. The method according to claim 63, wherein half-antibody B is unchanged or altered.

65. The method according to claim 63, wherein the heterologous dimerization domain is derived from humans or other mammals, and the heterologous dimerization domain is not the TCRα / β or TCRβ / α constant domain, preferably, the heterologous dimerization domain is not the TCRα / β or TCRβ / α constant domain or the HLA / β2m or β2m / HLA domain.

66. The method according to claim 63, wherein the heterodimerization domain is selected from: a pair of HLA and β2m, a pair of β2m and CD1b, a pair of CD1a and β2m, a pair of GABA1 and GABA2, or a pair of SIRPα and CD47.

67. The method according to claim 63, wherein the heterodimerization domain is selected from at least one of the combinations of HLA / β2m, β2m / HLA, CD1b / β2m, β2m / CD1b, β2m / CD1a and CD1a / β2m.

68. The method according to claim 63, wherein the heterodimerization domain is selected from β2m / CD1b, or CD1b / β2m, or β2m / CD1a, or CD1a / β2m. Preferably, the CH1 domain of heavy chain 1 is replaced by the α3 domain of CD1a, and the CL domain of light chain 1 is replaced by β2m.

69. The method according to claim 63, further comprising changing the amino acids of the novel protein by mutation.

70. The method according to claim 63, wherein β2m / CD1a or CD1a / β2m, or β2m / CD1b or CD1b / β2m contains a mutation.

71. The method according to claim 63, wherein β2m / CD1a or CD1a / β2m, or β2m / CD1b or CD1b / β2m contains one or more disulfide bonds by introducing a cysteine mutation.

72. The method according to claim 63, wherein the homodimerization domain is derived from a human or other mammal, and the homodimerization domain is selected from the IgG CH3 / CH3 domain, the IgM CH2 / CH2 domain, and the IgE CH2 / CH2 domain.

73. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1a D240C and the CL domain of light chain 1 is replaced by β2m R12C ​ 74. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1a S238C and the CL domain of light chain 1 is replaced by β2m R12C ​ 75. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1a D234C and the CL domain of light chain 1 is replaced by β2m Q8C ​ 76. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C and the CL domain of light chain 1 is replaced by β2m M99C .

77. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1a W190C and the CL domain of light chain 1 is replaced by β2m P14C ​ 78. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C and the CL domain of light chain 1 is replaced by β2m 100C and replaced.

79. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C and the CL domain of light chain 1 is replaced by β2m 100C / Q8K and replaced.

80. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C / G233S and the CL domain of light chain 1 is replaced by β2m 100C / Q8K ​ 81. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1a G194C / G233S / D234E and the CL domain of light chain 1 is replaced by β2m 100C / Q8K .

82. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1b G194C and the CL domain of light chain 1 is replaced by β2m 100C / Q8K ​ 83. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1b G194C / G233S and the CL domain of light chain 1 is replaced by β2m 100C / Q8K and replaced.

84. The method according to claim 63, wherein the CH1 domain of heavy chain 1 is replaced by CD1b G194C / G233S / D234E and the CL domain of light chain 1 is replaced by β2m 100C / Q8K ​ 85. The method according to claim 63, characterized in that, The Tm of the novel protein is substantially the same as that of the original protein, or increased by at least 0.1 °C, 0.2 °C, 0.3 °C, 0.4 °C, 0.5 °C, 0.6 °C, 0.7 °C, 0.8 °C, 0.9 °C, 1 °C, 1.5 °C, 2 °C, 2.5 °C, 3 °C, 3.5 °C, 4 °C, 4.5 °C, 5 °C, 5.5 °C, 6 °C, 6.5 °C, 7 °C, 7.5 °C, 8 °C, 8.5 °C, 9 °C, 9.5 °C, 10 °C, 11 °C, 12 °C, 13 °C, 14 °C, 15 °C, 16 °C, 17 °C, 18 °C, 19 °C or 20 °C. Preferably, it is increased by at least 0.3 °C, 0.5 °C, 1.5 °C, 2 °C or 3 °C.

Images