Dual specific antibodies

Abstract

Provided is a bispecific antibody comprising a first antigen-binding portion and a second antigen-binding portion, wherein the paired domain of the first antigen-binding portion comprises the amino acid sequence of an engineered HLA-Iα3 and the amino acid sequence of an engineered β2M. Further provided are methods for expressing the bispecific antibody, polynucleotides encoding the bispecific antibody, vectors and host cells containing the polynucleotides, pharmaceutical compositions comprising the bispecific antibody, and fusion proteins and conjugates. The provided bispecific antibody prevents or reduces mismatches between heavy and light chains with different specificities.

Description

[Technical field]

[0001] The present invention relates to antibodies, and in particular to bispecific antibodies that ameliorate mismatches between the light and heavy chains of antibodies. [Background technology]

[0002] In the early 1960s, Nisonoff et al. first proposed the concept of bispecific antibodies (BsAb), an artificial antibody molecule with two different antigen-binding moieties (Nisonoff et al., Science 132, 1770-1771 (1960)). They successfully combined rabbit antigen-binding fragments (Fab) with different specificities by slight reoxidation, and expressed the agglutination of two different types of cells mediated by these two antigen-binding fragments (Fudenberg et al., AJ Exp. Med. 119, 151-166 (1964)). After a series of conceptual and technological innovations, and with groundbreaking advances in antibody engineering and antibody biology, the concept of bispecific antibodies is being commercialized by biotechnology and pharmaceutical companies to generate innovative and unique therapeutics.

[0003] In the research and development design of bispecific antibodies, it is necessary to ensure the accurate pairing of two pairs of light and heavy chains with different specificities based on clinical therapeutic purposes, and to maintain the independence of each binding domain of each monoclonal antibody and ensure that there is no steric interference when binding to different epitopes. In addition, it is also necessary that the antibody molecule is easy to express in mammalian cells and does not require complicated protein modification technology. Here, the bispecific antibody in the form of asymmetric molecular design retains the natural structure of natural antibodies as much as possible to maintain the relevant functional characteristics and advantageous quality properties. The chain mismatch problem is solved by breaking the symmetry of the antibody H2L2 element.

[0004] Due to its dual targeting advantage, bispecific antibodies have become one of the new drug research areas for tumor treatment. The construction and production technology of bispecific antibodies has also undergone a series of innovations, one of the aims of which is to solve the mismatch problem and improve production efficiency. Some bispecific antibodies need to co-express two different heavy chains and two different light chains, and the desired bispecific antibody must be selected from a variety of recombinant products, which has always been one of the challenges in the development of bispecific antibodies. To solve the chain mismatch, scientists have developed many strategies and technology platforms in recent years, allowing them to design and express bispecific antibodies with various structures. Currently, there are dozens of bispecific antibody structures based on different manufacturing platforms. The relatively mature ones are the Genentech / Roche knob-into-hole, Amgen BiTe, and Chugai / Roche ART-Ig platforms, among which Knob-into-hole is a classical technology for solving the heavy chain mismatch problem, and other technology platforms include DVD-Ig, TrioMab, DART, FIT-Ig, WuXiBody, etc.

[0005] Most asymmetric bispecific antibodies achieve the assembly of two different heavy chain heterodimers by introducing mutations in the Fc region (Ridgway et al., Protein Engineering, Design and Selection, 9(7), 617-621.) to forcefully correct and promote pairing between the heavy chains. The desired final product is then isolated by combining a differential Protein A binding design purification strategy with a sequential affinity chromatography and molecular weight difference chromatography purification strategy. Strategies to avoid mismatches between light and heavy chains usually involve expressing bispecific molecules separately as half molecules or parent antibodies, and then assembling the antibody half molecules, and heavy chain heterodimers can be achieved by designing mutations in the heavy chains. However, selective pairing of each light chain and heavy chain of an individual antibody remains difficult. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Science,(USA),1960,Vol.132,No.3441,p.1770-1771 [Non-Patent Document 2] Journal of Experimental Medicine,(US),1964,Vol.119(1),p.151-166 [Non-Patent Document 3] Protein Engineering, Design and Selection, (English), 1996, Vol. 9(7), p. 617-621. Summary of the Invention [Problem to be solved by the invention]

[0007] The present invention provides bispecific antibodies with improved structural designs that prevent or reduce mismatches between heavy and light chains with different specificities. [Means for solving the problem]

[0008] In one aspect, the present invention provides a method for producing a method for treating a cancer cell comprising: A bispecific antibody comprising a first antigen-binding portion and a second antigen-binding portion that specifically bind to two different antigens or different epitopes of the same antigen, wherein the first antigen-binding portion is a first polypeptide comprising, from N-terminus to C-terminus, a first heavy chain variable domain and a first paired domain operably linked to the first heavy chain variable domain; a second polypeptide comprising, from N-terminus to C-terminus, a first light chain variable domain and a second paired domain operably linked to the first light chain variable domain; Here, a bispecific antibody is provided, wherein one of the first and second paired domains comprises the amino acid sequence of an engineered HLA-Iα3, and the other paired domain comprises the amino acid sequence of an engineered β2 microglobulin.

[0009] In some embodiments, the first paired domain and the second paired domain can form a dimer, can form at least one non-natural inter-chain bond between the first paired domain and the second paired domain, and can stabilize the dimer.

[0010] In some embodiments, the amino acid sequence of the engineered HLA-Iα3 has at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence set forth in SEQ ID NO:1, and the amino acid sequence of the engineered β2 microglobulin has at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence set forth in SEQ ID NO:2.

[0011] In some specific embodiments, said engineered HLA-I α3 comprises the amino acid sequence of SEQ ID NO:3 and said engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO:4.

[0012] In some specific embodiments, said engineered HLA-I α3 comprises the amino acid sequence of SEQ ID NO:3 and said engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO:5.

[0013] In some embodiments, the second antigen binding portion comprises a second heavy chain variable domain and a second light chain variable domain.

[0014] In some embodiments, the second antigen-binding portion comprises a Fab. The Fab comprises the second heavy chain variable domain and a second light chain variable domain.

[0015] In some embodiments, a bispecific antibody comprises a first antigen-binding portion and a second antigen-binding portion that specifically bind to two different antigens or different epitopes of the same antigen, wherein the first antigen-binding portion: a first polypeptide comprising, from N-terminus to C-terminus, a first heavy chain variable domain and a first paired domain operably linked to the first heavy chain variable domain; a second polypeptide comprising, from N-terminus to C-terminus, a first light chain variable domain and a second paired domain operably linked to the first light chain variable domain; wherein one of the first and second paired domains comprises an amino acid sequence of an engineered HLA-Iα3 and the other paired domain comprises an amino acid sequence of an engineered β2 microglobulin; The amino acid sequence of the engineered HLA-Iα3, engineered β2 microglobulin has at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence shown in SEQ ID NO:1, SEQ ID NO:2, respectively, and the second antigen binding portion comprises a Fab. In some such embodiments, the amino acid at position F118 of the CL domain of the Fab is substituted with alanine according to EU numbering. In some embodiments, the engineered HLA-Iα3 comprises the amino acid sequence of SEQ ID NO:3 and the engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO:4 or SEQ ID NO:5.

[0016] In some embodiments, a bispecific antibody comprises a first antigen-binding portion and a second antigen-binding portion that specifically bind to two different antigens or different epitopes of the same antigen, wherein the first antigen-binding portion: a first polypeptide comprising, from N-terminus to C-terminus, a first heavy chain variable domain and a first paired domain operably linked to the first heavy chain variable domain; a second polypeptide comprising, from N-terminus to C-terminus, a first light chain variable domain and a second paired domain operably linked to the first light chain variable domain; wherein one of the first and second paired domains comprises an amino acid sequence of an engineered HLA-Iα3 and the other paired domain comprises an amino acid sequence of an engineered β2 microglobulin; The second antigen binding portion comprises a Fab, wherein the amino acid at position F118 of the CL domain of the Fab is substituted with alanine, according to EU numbering. In some such embodiments, the amino acid sequence of the engineered HLA-Iα3 and the engineered β2 microglobulin have at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity to the sequences set forth in SEQ ID NO:1 and SEQ ID NO:2, respectively. Further, in some embodiments, the engineered HLA-Iα3 comprises the amino acid sequence of SEQ ID NO:3 and the engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO:4 or SEQ ID NO:5.

[0017] In some embodiments of the above, the first paired domain comprises the amino acid sequence of an engineered HLA-Iα3 and the second paired domain comprises the amino acid sequence of an engineered β2 microglobulin.

[0018] In some embodiments of the above, the first paired domain comprises the amino acid sequence of an engineered β2 microglobulin and the second paired domain comprises the amino acid sequence of an engineered HLA-Iα3.

[0019] In some embodiments of the above, the bispecific antibody comprises an Fc.

[0020] In one aspect, the invention provides an isolated polynucleotide encoding a bispecific antibody according to the invention.

[0021] In one aspect, the present invention provides an isolated vector comprising a polynucleotide according to the present invention.

[0022] In one aspect, the invention provides a host cell comprising the isolated polynucleotide or isolated vector according to the invention.

[0023] In another aspect, the invention provides a method for expressing said bispecific antibody, comprising culturing said host cell under conditions whereby said bispecific antibody is expressed.

[0024] In another aspect, the invention provides a pharmaceutical composition comprising a bispecific antibody according to the invention and a pharma- ceutically acceptable carrier.

[0025] In another aspect, the invention provides a conjugate comprising a bispecific antibody according to the invention and a therapeutic agent linked or conjugated to said bispecific antibody.

[0026] In another aspect, the invention provides a fusion protein comprising a fusion moiety expressed as a fusion and the bispecific antibody described above.

[0027] In yet another aspect, the present invention provides a method for treating a disease in a subject in need thereof comprising administering to the subject a therapeutically effective amount of a bispecific antibody, pharmaceutical composition, conjugate or fusion protein according to the invention. Effect of the Invention

[0028] In the bispecific antibody of the present invention, the paired domain in the first antigen-binding portion can improve chain mismatching by utilizing an engineered HLA-Iα3 and an engineered β2 microglobulin, making it less likely that mismatching will occur between the first and second antigen-binding portions compared to when the first and second antigen-binding portions are both corresponding portions of a native Fab, thereby improving the yield and purity of the target antibody.

[0029] In particular, engineered bispecific antibodies maintain good affinity for the antigen and / or thermal stability. In some aspects, engineered bispecific antibodies reduce immunogenicity risks. [Brief description of the drawings]

[0030] [Figure 1] This is the identity sequence of the constant region of the α chain of the HLA-IB molecule. [Diagram 2] FIG. 1 is a schematic diagram of the structure of an anti-CD3 / CD38 bispecific antibody constructed based on MHC-I elements. [Diagram 3] FIG. 1 is a schematic diagram of the molecular structure of MHI383-ccff-IgG1. [Figure 4] SEC purification map of the MHI383-ccff-IgG1 molecule. [Diagram 5] 1 shows a mass spectrum of the complete molecular weight of MHI383-ccff-IgG1. [Figure 6] This is non-reducing SDS-PAGE, in which heavy and light chains containing MHCICα, MHCICβ, CH1, and CL domains are expressed in a paired combination. [Figure 7] Non-reducing SDS-PAGE of the expression products of HZ5G11VL-IgK or a mutant thereof with amino acid mutations and HZ14A9VH-Cα-IgG1. [Figure 8] FIG. 1 is a schematic diagram of the structure of the MHL147-3322-IgG1-wt bispecific antibody constructed on the basis of MHC-I elements. [Figure 9] FIG. 1 is a schematic diagram of the structure of the MHL147-3322-IgG1-F118A bispecific antibody constructed based on MHC-I elements. [Figure 10] Effect of different concentrations of bispecific antibodies MHL147-3322-IgG1-wt and MHL147-3322-IgG1-F118A on hemagglutination. [Figure 11]Non-reducing SDS-PAGE after introduction of F118A into the CL of antibodies HZ5G11 and HZ14A9. [Figure 12] 1 shows some exemplary structural forms of bispecific antibodies of the invention, in which (A) an MHC-modified antigen-binding moiety of different specificity (first antigen-binding moiety) and a Fab are linked to the N-terminus of an Fc polypeptide in a "1+1" configuration; (B) an MHC-modified antigen-binding moiety of different specificity (first antigen-binding moiety, the orientation of Cα and Cβ differs from that in (A)) and a Fab are linked to an Fc polypeptide in a "1+1" configuration; and (C) the first antigen-binding moiety and one Fab are linked to an Fc polypeptide, and the other Fab is linked to the N-terminus of the first antigen-binding moiety. (D) is a "2+1" form in which two Fabs are linked in series, unlike (C), (E) is a "2+1" form in which two Fabs are each linked to an Fc polypeptide and the first antigen-binding moiety is linked to the N-terminus of one Fab, (F) is a "2+1" form in which the first antigen-binding moiety is linked to the C-terminus of the Fc polypeptide, unlike (E), (G) is a "2+2" form in which the antigen-binding moiety is linked to the N-terminus of the Fc, and (H) is a "2+2" form in which all of the first antigen-binding moieties are linked to the C-terminus of the Fc. The Fc domain may also have amino acid substitutions such as KIH and the CL may have F118A, but in this example, no amino acid substitutions are shown. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0031] [term] The term "antibody" is used herein in the broadest sense to refer to a protein that contains an antigen-binding portion and encompasses natural and artificial antibodies of various structures, including, but not limited to, monoclonal antibodies, single-specific antibodies, multispecific antibodies (e.g., bispecific antibodies, trispecific antibodies, etc.), single-chain antibodies, etc.

[0032] The term "multispecific" means that an antibody can specifically bind to at least two different antigenic determinants. In general, a bispecific antibody contains two antigen-binding moieties, each specific for a different antigenic determinant. The different antigenic determinants can be expressed on the same or different cells. The different antigenic determinants can be different depending on the type of antigen (e.g., binding antigens PDL1 and CD47) or can be present on the same type of antigen. An antigenic determinant is a structure that has a specific chemical group of a certain composition and structure on the surface or other site of an antigenic substance molecule, and specifically binds to a corresponding antibody or sensitized lymphocyte. As an example of an antigenic determinant, the antigenic determinant CD47 is an antigenic substance that has multiple antigenic determinants with a defined structure or multiple antigenic determinants with an undefined structure. One particular bispecific antibody can, for example, bind to PDL1 and CD47.

[0033] The term "specific binding" means that the binding is selective for the antigen and can be distinguished from unwanted or non-specific interactions.

[0034] The term "valent" antibody refers to the number of antigen binding sites present in the antibody. A "bivalent" antibody means that there are two antigen binding sites in the antibody, and "trivalent" means that there are three antigen binding sites in the antibody. Naturally occurring human immunoglobulin molecules usually have two antigen binding sites, and Fabs usually have one antigen binding site. Single variable domains, ScFvs usually have a single antigen binding site.

[0035] The term "antigen-binding portion" refers to a polypeptide molecule that specifically binds to an antigenic determinant. Particular antigen-binding portions may be Fabs, ScFvs, single variable domains, etc. Antigenic determinants herein are synonymous with antigenic epitopes.

[0036] The terms "first," "second," or "third" as used herein in reference to antigen-binding moieties, antigens, heavy chain variable domains, light chain variable domains, Fc polypeptides, etc. are used to facilitate distinction when more than one of each type of moiety is present. Unless expressly indicated, the use of these terms is not intended to confer a particular order or orientation of the bispecific antibodies.

[0037] The term "operably linked" or "operably linked" refers to the juxtaposition of two or more biological sequences of interest, with or without spacers or linkers, in a relationship that allows them to act in a desired manner. When used in reference to a polypeptide, the term is intended to mean that the polypeptide sequences are linked such that the linked product has a desired biological function. For example, an antibody variable region can be operably linked to a constant region to form a stable product with antigen-binding activity. The term can also be used in reference to polynucleotides. For example, when a polynucleotide encoding a polypeptide is operably linked to a regulatory sequence (e.g., a promoter, enhancer, silencer sequence, etc.), the term is intended to mean that the polynucleotide sequences are linked in a manner that allows the polypeptide to be expressed in a manner regulated by the polynucleotide.

[0038] The major histocompatibility complex (MHC) is a collective term for a group of genes that code for the major histocompatibility antigens of animals. The mouse MHC is called the H-2 gene complex, while the human MHC is called the HLA (human leukocyte antigen) gene complex, and the coding products are called HLA molecules, HLA antigens, or human leukocyte antigens. The HLA gene complex is located on the short arm of human chromosome 6, 6p21.3, and contains more than 220 genes with different functions at the same time. Many of these genes code for proteins of the immune system. MHC class I molecules are expressed on the surface of almost all nucleated cells and are recognized by CD8+ T cells. MHC class II molecules are expressed on the surface of antigen-presenting cells and are recognized by CD4+ T cells. MHC is polymorphic, and most people mainly contain six kinds of classical MHC molecules, namely, three kinds of classical MHC-class I molecules (HLA-A, HLA-B, HLA-C) and three kinds of classical MHC-class II molecules (HLA-DR, HLA-DP, HLA-DQ). Human MHC-class I molecules (HLA-1) consist of a heavy chain (α chain) and β2 microglobulin (β2m or β chain), among which, the extracellular segment of the α chain has three domains (α1, α2, α3), the two domains α1 and α2 at the far end of the membrane constitute the antigen-binding groove, and the domain α3 is homologous to the constant region of human immunoglobulins.

[0039] The term "variable domain" or "variable region" refers to the domain of an antibody involved in the binding of the antibody to an antigen. For example, naturally occurring four-chain antibodies (e.g., from humans, mice, etc.) have a heavy chain variable region (VH) and a light chain variable region (VL), whereas only heavy chain antibodies from animals such as camelids or sharks have a single variable domain. Each variable domain of a naturally occurring antibody essentially consists of four "framework regions" and three "complementarity determining regions". The four framework regions are referred to as framework region 1 (or FR1), framework region 2 (or FR2), framework region 3 (or FR3), and framework region 4 (or FR4), respectively. The framework regions are spaced apart by three complementarity determining regions (or CDRs), which in the art and hereinafter are referred to as complementarity determining region 1 (or CDR1), complementarity determining region 2 (or CDR2), and complementarity determining region 3 (or CDR3), respectively. The general structure of a variable domain is therefore designated FR1-CDR1-FR2-CDR2-FR3-CDR3-FR4. The variable domain contains the antigen-binding site and thereby confers specificity to the antibody for the antigen.

[0040] "CDRs" (complementarity determining regions) are also called "hypervariable regions (HVRs)". Naturally occurring four-chain antibodies generally contain six CDRs, three in the heavy chain variable region (HCDR1, HCDR2, and HCDR3) and three in the light chain variable region (LCDR1, LCDR2, and LCDR3). Heavy chain antibodies alone or single variable domains usually have three CDRs (CDR1, CDR2, and CDR3).

[0041] Currently, there are several ways to divide and define CDRs. Among them, the Kabat definition divides CDRs based on sequence variability and is the most commonly used (Elvin A. Kabat, et al, Sequences of Proteins of Immunological Interest, 5th ed., Public Health Service, National Institutes of Health, Bethesda, Md. (1991)). On the other hand, the Chothia definition is based on the location of structural loops (Cyrus Chothia, et al, Canonical Structures for the Hypervariable Regions of Immunoglobulins, J. Mol. Biol. 196:901-917 (1987)). The AbM definition is a trade-off between the Kabat and Chothia definitions and is used in Oxford Molecular's AbM antibody modeling software. The basis for "contact" definition of CDRs is the analysis of available complex crystal structures. However, the boundaries of the CDRs of the same antibody variable region defined and obtained based on different methods may be different, i.e., the CDR sequences of the same antibody variable region defined and obtained based on different methods may be different. Thus, when it is related to defining an antibody using a specific CDR sequence defined in the present invention, the scope of said antibody also includes antibodies defined by CDR sequences converted to any other definition (e.g., Chothia, AbM definition, etc.).

[0042] The term "framework region" or "FR" refers to the amino acid residues of a variable domain other than the CDR residues as herein defined.

[0043] The term "Fab" refers to a protein consisting of the VH and CH1 domains of the heavy chain and the VL and CL domains of the light chain of an immunoglobulin. In the present specification, Fab means a native or modified Fab, which includes a Fab heavy chain (VH-CH1, NC-terminal direction) consisting of a heavy chain variable region and a constant region CH1, and a Fab light chain (VL-CL, NC-terminal direction) consisting of a light chain variable region and a constant region CL. The modified Fab may be, for example, a Fab in which an amino acid substitution has been introduced into the CH1 / CL domain or / and the VH / VL domain, and as a specific example, the modified Fab may be a Fab in which an amino acid substitution has been introduced into the CL domain.

[0044] The term "Fc structural domain", "Fc" or "Fc domain" is used herein to define a C-terminal region of an immunoglobulin heavy chain that comprises, at least in part, a constant region. This term includes native sequence Fc and variant Fc. The C-terminal lysine (Lys447) of Fc may or may not be present. Unless otherwise indicated, the numbering of amino acid residues in Fc is based on the EU numbering system, also known as the EU index, as described in (Kabat, EA et al., Sequences of Proteins of Immunological Interest, 5th ed., Public Health Service, National Institutes of Health, Bethesda, MD (1991), NIH Publication 91-3242). When the first paired domain and the second paired domain are described in terms of amino acid positions, the corresponding sequence is numbered sequentially starting from the first amino acid as position 1, for example, the R60C substitution in SEQ ID NO:1 refers to the substitution of R at position 60 in the sequence with C, with the first amino acid in SEQ ID NO:1 being numbered sequentially starting from position 1. As used herein, an "Fc polypeptide" of an Fc domain refers to one of the two polypeptides that form a dimeric Fc structural domain. For example, the Fc polypeptide of an IgG Fc domain includes the IgG CH2 and IgG CH3 constant regions.

[0045] The term "KD" as used herein refers to the equilibrium dissociation constant and is expressed as a molar concentration (M). The KD value of an antibody can be measured using methods known to those skilled in the art. One method for measuring antibody KD is to use surface plasmon resonance, for example, using a biosensor system such as the Biacore system. One method for measuring antibody KD is to use Bio-Layer Interferometry (BLI), for example, using the ForteBio system.

[0046] The term "treatment" refers to an attempt to modify the natural process of the disease being treated and can be a preventative or clinical intervention carried out during the clinical pathological process. The expected effects of treatment include, but are not limited to, prevention of disease onset or recurrence, alleviation of symptoms, reduction of any direct or indirect pathological consequences of the disease, prevention of metastasis, reduction of the rate of disease progression, improvement or palliation of the disease state, and eradication or improvement of prognosis.

[0047] The term "subject" includes any human or non-human animal. The term "non-human animal" includes all vertebrates, including mammals and non-mammals such as non-human primates, sheep, dogs, cats, horses, cows, chickens, amphibians, reptiles, etc. Preferably, a subject according to the present invention is a human. Unless otherwise indicated, the terms "patient" or "subject" may be used interchangeably.

[0048] The term "isolated" refers to a target compound (e.g. a VHH, a multispecific antibody, an antibody or a nucleic acid) that has been isolated from its natural environment.

[0049] "Percent identity" of an amino acid sequence refers to the percentage of amino acid residues in the sequence to be aligned that are the same as those in the specific amino acid sequence shown herein, after aligning the sequence to be aligned with the specific amino acid sequence shown herein and introducing caps, if necessary, to achieve the maximum percent of sequence identity, without considering any conservative substitutions as part of the sequence identity. Amino acid sequence alignment of identity can be performed by many methods in the art, such as BLAST, BLAST-2, ALIGN, or Megalign (DNASTAR) software. Those skilled in the art can determine appropriate parameters to use for sequence alignment, including any algorithms necessary to achieve maximum alignment over the entire length of the comparison sequences.

[0050] In the context of the present invention, amino acid substitutions are expressed as original amino acid-position-replaced amino acid, where the amino acid residue is expressed using a three-letter code or a one-letter code, including the codes Xaa and X. Thus, for example, "H435R" means the substitution of the amino acid H at position 435 with the amino acid R, and the substituted amino acid may comprise more than one, for example T366Y / W means the substitution of the amino acid T at position 366 with the amino acid Y or W.

[0051] The terms "comprise," "contain," or "including," and variations thereof, should be understood to mean "including but not limited to," and mean that other unspecified elements, components, and steps may be included, in addition to the listed elements, components, and steps.

[0052] As used herein, singular terms cover the plural and vice versa, unless the context clearly dictates otherwise.

[0053] All patents, patent applications, and other publications identified are hereby expressly incorporated herein by reference for purposes of illustration and disclosure. These publications are provided solely for their publication prior to the filing date of the application. Any indication as to the date or contents of these documents is based on the information available to the applicant and does not constitute any admission as to the correctness of the dates or contents of the documents. Moreover, any citation to these publications herein does not constitute an acknowledgement that the publications are part of the common general knowledge in the field in any country. Various aspects of the invention are described in further detail in the following sections.

[0054] Various aspects of the invention are described in further detail in the following sections.

[0055] [Bispecific antibody] The present invention provides a bispecific antibody capable of preventing or reducing mismatching between heavy and light chains of different specificities by modifying one of the constant regions of the antigen-binding portion, thereby improving the yield and / or purity of the target antibody. Specifically, the α3 domain and β2 microglobulin of a human MHC-I molecule are engineered to construct a first paired domain and a second paired domain to replace the CH1 and CL domains of the antibody, thereby improving the mismatching of the chains so that mismatching is less likely to occur between the first and second antigen-binding portions than when, for example, the first and second antigen-binding portions are all the corresponding portions of a natural Fab.

[0056] At present, the immunogenicity of biological drugs is mainly divided into endogenous and exogenous. Intrinsic immunogenicity comes from non-human amino acid sequences, non-human structural forms, and a series of chemical modifications (point mutations, domain fusions, glycoform modifications, etc.). Exogenous immunogenicity mainly comes from production technology, drug transport, patient genetic background, drug dosage, frequency and mode of administration, etc. With the gradual increase in the types and amounts of biological drugs and the expansion of their applications, immunogenicity and related issues have also attracted more attention from people. MHC (major histocompatibility complex) is the basis of recognition by the immune system and is involved in the adaptive immune process of the body. Designing bispecific antibodies based on the composition of MHC molecules has good compatibility with the immune environment of the human body. Compared to MHC-class II molecules, MHC-class I molecules are widely distributed, expressed on the surface of almost all nucleated cells, and recognized by CD8+ T cells. Since the number of human MHC-I molecules of the HLA-IB class is the largest among HLA-I molecules, the risk of immunogenicity can be reduced by constructing the structural elements of a bispecific antibody based on the constant regions of human MHC-I molecules of the HLA-IB class.

[0057] Furthermore, the constant regions of human MHC-I molecules of the HLA-IB class contain no N-linked and no O-linked glycosylation sites, making them more developable and less immunogenic.

[0058] Through experiments, we found that the engineered and constructed bispecific antibody still had good antigen-binding ability.

[0059] A bispecific antibody according to the invention comprises a first antigen-binding portion and a second antigen-binding portion that specifically bind to two different antigens or different epitopes of the same antigen, said first antigen-binding portion comprising: a first polypeptide comprising, from N-terminus to C-terminus, a first heavy chain variable domain and a first paired domain operably linked to the first heavy chain variable domain; a second polypeptide comprising, from N-terminus to C-terminus, a first light chain variable domain and a second paired domain operably linked to the first light chain variable domain; Here, one of the first paired domain and the second paired domain comprises the amino acid sequence of an engineered HLA-Iα3, and the other paired domain comprises the amino acid sequence of an engineered β2 microglobulin.

[0060] In some embodiments, the first paired domain and the second paired domain can form a dimer, can form at least one non-natural inter-chain bond between the first paired domain and the second paired domain, and can stabilize the dimer.

[0061] The first and second paired domains can form dimers with superior binding affinity to the antigen even when oriented at different positions in the first antigen-binding moiety (e.g., linked to the light chain variable region or the heavy chain variable region).

[0062] In one embodiment, the first paired domain comprises the amino acid sequence of an engineered HLA-Iα3 and the second paired domain comprises the amino acid sequence of an engineered β2 microglobulin.

[0063] In another embodiment, the first paired domain comprises the amino acid sequence of an engineered β2 microglobulin and the second paired domain comprises the amino acid sequence of an engineered HLA-Iα3.

[0064] The non-natural interchain bond can stabilize the dimer formed by the first paired domain and the second paired domain, and the amino acid residues forming the non-natural interchain bond are at the contact interface between the first paired domain and the second paired domain and can effectively form a bond. The term "contact interface" refers to a particular region of the polypeptide where the polypeptides interact / associate. The contact interface includes one or more amino acid residues and can interact with corresponding amino acid residues that are in contact or associated during the interaction. The amino acid residues at the contact interface may or may not be in a contiguous sequence. For example, if the interface is three-dimensional, the amino acid residues within the interface may be individually located at different positions on a linear sequence.

[0065] The number of said non-natural interchain bonds may be 1 to 3, for example 1, 2 or 3, and in order to be as similar as possible to a natural antibody, it is preferred that the number of said non-natural interchain bonds is 1. In one embodiment, said non-natural interchain bonds are disulfide bonds.

[0066] In some embodiments, the amino acid sequence of the engineered HLA-Iα3 has at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence set forth in SEQ ID NO:1, and the amino acid sequence of the engineered β2 microglobulin has at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence set forth in SEQ ID NO:2. In some embodiments, the amino acid sequence of the engineered HLA-I α3 and the engineered β2 microglobulin have at least 95%, 96%, 97%, 98%, 99%, or 100% identity to the sequences shown in SEQ ID NO:1 and SEQ ID NO:2, respectively.

[0067] The sequence shown in SEQ ID NO:1 is derived from a consensus sequence generated by aligning constant region sequences of multiple HLA-IB class molecules, and it can be used or modified to form paired domains with good versatility.

[0068] In some embodiments, the engineered HLA-Iα3 comprises an amino acid sequence having an amino acid substitution in SEQ ID NO:1, and the engineered β2 microglobulin comprises an amino acid sequence having an amino acid substitution in SEQ ID NO:2.

[0069] The amino acid substitutions in engineered HLA-Iα3, engineered β2 microglobulin include amino acid substitutions that form non-natural interchain bonds (e.g., disulfide bonds) or / and amino acid substitutions that improve dimeric or bispecific antibodies formed by paired domains.

[0070] In some embodiments, the amino acid substitutions in the engineered HLA-Iα3 and engineered β2 microglobulin both occur at both contact interfaces and comprise cysteine ​​residue substitutions that can form disulfide bonds with each other. In one specific embodiment, the cysteine ​​residues are substituted with R60C in SEQ ID NO:1 and Y26C in SEQ ID NO:2. In one specific embodiment, the cysteine ​​residues are substituted with A62C in SEQ ID NO:1 and R12C in SEQ ID NO:2. In one specific embodiment, the cysteine ​​residues are substituted with G63C in SEQ ID NO:1 and Y67C in SEQ ID NO:2. In another specific embodiment, the cysteine ​​residue substitutions may be two or three pairs selected from the above R60C / Y26C, A62C / R12C, G63C / Y67C.

[0071] In some embodiments, the amino acid substitutions in the engineered HLA-Iα3 and / or engineered β2 microglobulin include amino acid substitutions that improve the isoelectric point of the dimeric or bispecific antibody formed by the paired domains. These amino acid substitutions can occur in both paired domains or in any one paired domain. To facilitate the development of the production process and improve druggability, the isoelectric point of the dimeric or bispecific antibody formed by the paired domains is generally improved to 6.5-9.0. In some embodiments, the isoelectric point of the dimeric or bispecific antibody formed by the paired domains is improved to 6.5-8.5, 7.0-8.5, 7.0-9.0, 7.0-7.8, 7.0-8.0, 7.5-7.8, or 7.5-8.0. In some specific embodiments, the isoelectric point of the dimer or bispecific antibody formed by the paired domains is increased to 6.5, 6.7, 6.9, 7.1, 7.3, 7.5, 7.7, 7.8, 7.9, 8.0, 8.2, 8.3, 8.5, 8.7 or 9.0. The isoelectric point can be experimentally measured or theoretically calculated, for example using the online computational analysis tool Expasy-Compute pI / Mw tool (http: / / web.expasy.org / compute_pi / ).

[0072] In some embodiments, the amino acid substitutions that improve the isoelectric point include substitutions of positively charged amino acids at one or more of positions E3, D22, E48, D53, E90, E101 of the sequence of SEQ ID NO:1 in the engineered HLA-Iα3.

[0073] In some embodiments, the amino acid substitutions that improve the isoelectric point include substitutions with positively charged amino acids at one or more of positions E74, E47, E69, D34, E16, D53, E44, E50, and E36 of the engineered β2 microglobulin sequence of SEQ ID NO:2.

[0074] In some specific embodiments, the amino acid substitutions that improve the isoelectric point include the engineered HLA-Iα3 being substituted with positively charged amino acids at positions D22, E48 and D53 of the SEQ ID NO:1 sequence, and the engineered β2 microglobulin being substituted with a positively charged amino acid at position E69 of the SEQ ID NO:2 sequence.

[0075] In some embodiments, the positively charged amino acid is K or R.

[0076] In a more specific embodiment, the amino acid substitutions that improve the isoelectric point include the engineered HLA-Iα3 comprising the amino acid substitutions D22R, E48K and D53K in the sequence of SEQ ID NO:1, and the engineered β2 microglobulin comprising the amino acid substitution E69R in the sequence of SEQ ID NO:2.

[0077] In some more specific embodiments, the amino acid substitutions include the engineered HLA-Iα3 comprises amino acid substitutions A62C, D22R, E48K, and D53K in the sequence of SEQ ID NO:1, and the engineered β2 microglobulin comprises amino acid substitutions R12C, E69R in the sequence of SEQ ID NO:2. In a further aspect, the engineered β2 microglobulin comprises a "GP" added to the carboxyl terminus of the sequence of SEQ ID NO:2.

[0078] In a specific example, the engineered HLA-Iα3 comprises the amino acid sequence of SEQ ID NO:3, and the engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO:4.

[0079] In another example, the engineered HLA-Iα3 comprises the amino acid sequence of SEQ ID NO:3 and the engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO:5.

[0080] In some embodiments, the second antigen binding moiety comprises a second heavy chain variable domain and a second light chain variable domain. The second antigen binding moiety does not comprise a first paired domain and a second paired domain as described herein. In a specific embodiment, the second antigen binding moiety comprises a Fab. The first and second antigen binding moieties specifically bind to two different antigens or different epitopes of the same antigen, and in some embodiments, the first and second antigen binding moieties bind monovalently to their respective antigens.

[0081] In some embodiments, one or more of the amino acids at positions T / S114, T / F116, and F118 of the CL domain of the Fab are substituted with hydrophobic amino acids, which in some embodiments are selected from non-aromatic amino acids such as methionine, alanine, valine, isoleucine, or leucine, according to EU numbering. The mismatch between the light and heavy chains can be further reduced by introducing appropriate amino acid substitutions into the CL domain of the Fab of the second antigen-binding portion. As a specific example, the amino acid at position F118 of the CL domain of the Fab is substituted with alanine (A), according to EU numbering. In another embodiment, the CL domain of the Fab is substituted with alanine at position F118 of the EU numbering, and the CL domain also contains other substitutions to further reduce CL mismatch with paired domains, including, for example, engineered HLA-Iα3.

[0082] In some embodiments, the CL domain may be derived from a human kappa light chain CL domain or a human lambda light chain CL domain. As a specific example, the CL domain comprising F118A comprises the amino acid sequence set forth in SEQ ID NO:44. As a specific example, the amino acid sequence of the CL domain comprising F118A is set forth in SEQ ID NO:44.

[0083] In some embodiments, the CH1 domain of the Fab may be a CH1 domain of human IgG (IgG1-4), IgM, IgA, IgD or IgE, and in particular, the CH1 domain of the Fab is a CH1 domain of human IgG1. In particular, the CH1 domain comprises the amino acid sequence set forth in SEQ ID NO:34. In particular, the amino acid sequence of the CH1 domain is set forth in SEQ ID NO:34. In particular, the CH1 domain of the Fab is a CH1 domain of human IgG4.

[0084] The amino acid sequence represented by SEQ ID NO:34 is as follows: ASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPSNTKVDKKV The amino acid sequence represented by SEQ ID NO:35 is as follows: RTVAAPSVFIFPPSDEQLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC The amino acid sequence represented by SEQ ID NO:44 is as follows: RTVAAPSVFIAPPSDEQLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0085] In some specific embodiments, the amino acid at position F118 of the CL domain of said Fab is substituted with alanine, and the CH1 domain of said Fab does not comprise an amino acid substitution at position 128 (EU numbering), 141 (EU numbering), 185 (EU numbering), 11 (IMGT numbering), 28 (IMGT numbering), or 139 (Kabat numbering). In another embodiment, the CL domain of said Fab is substituted with alanine at position F118 according to EU numbering, and said CL domain further comprises another substitution (e.g., a substitution that further reduces CL mismatch with a paired domain comprising an engineered HLA-Iα3), and the CH1 domain of said Fab does not comprise an amino acid substitution at position 128 (EU numbering), 141 (EU numbering), 185 (EU numbering), 11 (IMGT numbering), 28 (IMGT numbering), or 139 (Kabat numbering).

[0086] In some specific embodiments, the amino acid at position F118 of the CL domain of said Fab is substituted with alanine, and the CH1 domain of said Fab does not contain amino acid substitutions at positions L128F (EU numbering), A141L (EU numbering), V185A / L (EU numbering), L11K / F / W (IMGT numbering), L28N / R (IMGT numbering), or A139W / V / I (Kabat numbering). In another embodiment, the CL domain of the Fab is substituted at position F118 according to EU numbering with alanine, the CL domain further comprises other substitutions (e.g., substitutions that further reduce CL mismatch with the paired domain comprising the engineered HLA-Iα3), and the CH1 domain of the Fab does not comprise amino acid substitutions at L128F (EU numbering), positions A141L (EU numbering), V185A / L (EU numbering), L11K / F / W (IMGT numbering), L28N / R (IMGT numbering), or A139W / V / I (Kabat numbering).

[0087] In some specific embodiments, the amino acid at position F118 of the CL domain of the Fab is substituted with alanine and the CH1 domain of the Fab uses a wild-type immunoglobulin CH1 domain, e.g., the CH1 domain of human IgG. In another embodiment, the CL domain of the Fab is substituted with alanine at position F118 according to EU numbering, the CL domain further comprises other substitutions (e.g., substitutions that further reduce CL mismatch with the paired domain comprising engineered HLA-Iα3), and the CL domain of the Fab uses a wild-type immunoglobulin CH1 domain, e.g., the CH1 domain of human IgG (IgG1, IgG2, IgG3, or IgG4).

[0088] In some specific embodiments, the CL domain of the Fab comprises an amino acid sequence having at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence set forth in SEQ ID NO: 44 and the CH1 domain of the Fab comprises an amino acid sequence having at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence set forth in SEQ ID NO: 34. Further, in this embodiment, the CL domain of the Fab comprises F118A. Furthermore, in this embodiment, the CH1 domain of said Fab is the CH1 domain of a wild type immunoglobulin or said CH1 domain of said Fab does not comprise an amino acid substitution at position 128 (EU numbering), 141 (EU numbering), 185 (EU numbering), position 11 (IMGT numbering), position 28 (IMGT numbering), or position 139 (Kabat numbering).

[0089] In some specific embodiments, the CL domain of said Fab comprises the amino acid sequence set forth in SEQ ID NO:44, and the CH1 domain of said Fab comprises the amino acid sequence set forth in SEQ ID NO:34.

[0090] In some specific embodiments, the amino acid sequence of the CL domain of said Fab is set forth in SEQ ID NO:44, and the amino acid sequence of the CH1 domain of said Fab is set forth in SEQ ID NO:34.

[0091] In some embodiments, one of the first and second antigen-binding moieties binds to a T cell-specific receptor molecule and / or a natural killer cell-specific receptor molecule, and the other binds to a tumor-associated antigen.

[0092] T cell specific receptor molecules allow T cells to bind and, in the presence of additional signals, become activated by and respond to epitopes / antigens presented by other cells called antigen presenting cells or APCs. T cell specific receptor molecules can be TCR, CD3, CD28, CD134 (also called OX40), 4-1BB, CD5, or CD95 (also called Fas receptor).

[0093] Examples of NK cell-specific receptor molecules include CD16, low affinity Fc receptor and NKG2D, and CD2.

[0094] The term "tumor associated antigen" refers to an antigen that is presented on the surface of tumor cells or may be presented on the surface of tumor cells and is present on or within tumor cells. In some embodiments, the tumor associated antigen may be presented only by tumor cells and not by normal cells, i.e., non-tumor cells. In some other embodiments, the tumor associated antigen may be expressed only on tumor cells or may display tumor-specific mutations compared to non-tumor cells. In some other embodiments, the tumor associated antigen may be found in tumor cells and non-tumor cells, but may be overexpressed on tumor cells compared to non-tumor cells, or may be bound to an antibody within tumor cells due to the less dense structure of tumor tissue compared to non-tumor tissue. In some embodiments, the tumor associated antigen is located on the vasculature of the tumor. Some exemplary tumor associated antigens are CD38, CD47, PD-L1, PD-1, etc.

[0095] In some embodiments, one of the first and second antigen-binding moieties specifically binds to CD3 and the other specifically binds to a tumor-associated antigen. For example, the first antigen-binding moiety specifically binds to CD3 and the second antigen-binding moiety specifically binds to a tumor-associated antigen, or the second antigen-binding moiety specifically binds to CD3 and the first antigen-binding moiety specifically binds to a tumor-associated antigen. In some embodiments, the first antigen-binding moiety specifically binds to CD3 and the second antigen-binding moiety specifically binds to CD38, or the first antigen-binding moiety specifically binds to CD38 and the second antigen-binding moiety specifically binds to CD3.

[0096] In some embodiments, the first antigen-binding moiety and the second antigen-binding moiety both specifically bind to a tumor-associated antigen, hi some specific examples, the first antigen-binding moiety specifically binds to PD-L1 and the second antigen-binding moiety specifically binds to CD47, or the first antigen-binding moiety specifically binds to CD47 and the second antigen-binding moiety specifically binds to PD-L1.

[0097] In some embodiments, the bispecific antibody further comprises an Fc domain consisting of two Fc polypeptides capable of stably associating, hi some embodiments, the bispecific antibody is bivalent.

[0098] In some embodiments, the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide and the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide. In some more specific embodiments, the first polypeptide of the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide and the second antigen-binding portion comprises a Fab and a Fab heavy chain operably linked at its C-terminus to the N-terminus of the other Fc polypeptide. In some more specific embodiments, the first polypeptide of the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide and the second antigen-binding portion comprises a Fab and a Fab light chain operably linked at its C-terminus to the N-terminus of the other Fc polypeptide.

[0099] In some embodiments, the bispecific antibody further comprises a third antigen-binding moiety. In some specific embodiments, the binding of the bispecific antibody to antigen is trivalent.

[0100] In some embodiments, the third antigen-binding moiety binds to the same antigen epitope as the second antigen-binding moiety, preferably the third antigen-binding moiety is identical to the second antigen-binding moiety, in other embodiments, the third antigen-binding moiety binds to the same antigen epitope as the first antigen-binding moiety, preferably the third antigen-binding moiety is identical to the first antigen-binding moiety.

[0101] In some embodiments, the third antigen-binding portion is operably linked to the N-terminus or C-terminus of the first antigen-binding portion.

[0102] In another embodiment, the third antigen-binding moiety is operably linked to the N-terminus or C-terminus of the second antigen-binding moiety.

[0103] In some embodiments, the bispecific antibody further comprises an Fc domain comprising a third antigen-binding portion and consisting of two Fc polypeptides capable of stably associating.

[0104] In some specific embodiments, the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of another Fc polypeptide, and the third antigen-binding portion is operably linked at its C-terminus to the N-terminus of the first antigen-binding portion or the N-terminus of the second antigen-binding portion.

[0105] In other specific embodiments, the third antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of the third antigen-binding portion, and the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide.

[0106] In other specific embodiments, the third antigen binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, the second antigen binding portion is operably linked at its C-terminus to the N-terminus of the third antigen binding portion, and the first antigen binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide.

[0107] The operative linkage of the third antigen-binding moiety to the first antigen-binding moiety may be via a peptide linker.

[0108] The operative linkage of the third antigen-binding moiety to the second antigen-binding moiety may be via a peptide linker. Any suitable peptide linker may be used, for example a charged and / or flexible linker polypeptide. In a specific embodiment, the peptide linker consists of 1 to 50 amino acids linked via peptide bonds, the amino acids being selected from the 20 naturally occurring amino acids, and in a more preferred embodiment, the 1 to 50 amino acids are selected from glycine, alanine, proline, serine, asparagine, glutamine and lysine. Thus, exemplary peptide linkers may be polyglycine (particularly (Gly)4, (Gly)5), poly(Gly-Ser), (Gly)3AsnGlySer(Gly)2, (Gly)3Cys(Gly)4, GlyProAsnGlyGly, or those disclosed in Table 4 of patent application WO2019195535, etc. In some embodiments, the peptide linker may be a peptide linker consisting of glycine and serine. In some embodiments, the peptide linker may comprise 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more than 20 amino acids.

[0109] When an antigen-binding moiety is operably linked to an Fc domain or Fc polypeptide, it is typically linked via a hinge region, which comprises, for example, amino acids from the hinge region of human IgG.

[0110] [Fc domain] The Fc domain of a bispecific antibody consists of a pair of polypeptide chains comprising the heavy chain domain of an immunoglobulin molecule. For example, the Fc domain of an immunoglobulin G (IgG) molecule is a dimer, each Fc polypeptide of which comprises the CH2 and CH3 IgG heavy chain constant regions. The two Fc polypeptides of an Fc domain can stably associate with each other. For example, the two Fc polypeptides are stably associated via one or more of a linker, a disulfide bond, a hydrogen bond, an electrostatic interaction, a salt bridge, and a hydrophobic-hydrophilic interaction. In one embodiment, the bispecific antibody of the present invention comprises one Fc domain.

[0111] In one embodiment, the Fc domain of the bispecific antibody is an IgG Fc domain. In one specific embodiment, the Fc domain is an IgG1 Fc domain. In another embodiment, the Fc domain is an IgG4 Fc domain. In one more specific embodiment, the Fc domain is an IgG4 Fc domain that comprises an amino acid substitution at position S228, in particular the amino acid substitution S228P, which reduces in vivo Fab arm exchange of IgG4 antibodies. In yet another specific embodiment, the Fc domain is a human Fc domain. In a more specific embodiment, the Fc domain is a human IgG1 Fc domain.

[0112] In some embodiments, the Fc domain comprises a modification, such as an amino acid substitution, which may be, for example, a modification that promotes heterodimerization, a modification that alters effector function, a modification that alters binding ability to protein A, etc.

[0113] In some embodiments, the Fc domain comprises a modification that promotes heterodimerization.

[0114] The bispecific antibodies of the invention comprise different antigen-binding moieties fused to one or the other of the two Fc polypeptides of the Fc domain, and thus the two Fc polypeptides are usually comprised in two different polypeptide chains. Recombinant co-expression of these polypeptides and subsequent dimerization generates several possible combinations of the two polypeptides. To improve the yield and purity of the multispecific antibodies in recombinant production, it is advantageous to introduce modifications in the Fc domain of this bispecific antibody that promote the desired polypeptide binding. Thus, in a specific embodiment, the Fc domain comprises amino acid substitutions that promote the association of the two Fc polypeptides of the Fc domain.

[0115] The most extensive protein-protein interaction site between two Fc polypeptides of a human IgG Fc domain is in the CH3 domain of the Fc domain, therefore, in one embodiment, the modification is in the CH3 domain of the Fc domain.

[0116] In a specific embodiment, the modification is a so-called knob-into-hole modification, which includes a "knob" modification in one of the two Fc polypeptides of the Fc domain and a "hole" modification in the other of the two Fc polypeptides of the Fc domain. In general, this method introduces a protrusion ("knob") at the interface of one Fc polypeptide and a corresponding recess ("hole") at the interface of the other Fc polypeptide, allowing the protrusion to be positioned within this recess, promoting heterodimer formation and preventing homodimer formation. The protrusion is constructed by replacing a small amino acid side chain from the interface of one Fc polypeptide with a larger side chain (such as tyrosine or tryptophan). A complementary recess having the same or similar size as the protrusion is formed in the interface of the other Fc polypeptide by replacing the large amino acid side chain with a smaller amino acid side chain (such as alanine or threonine).

[0117] Thus, in a specific embodiment, an amino acid residue in the CH3 domain of one of the Fc polypeptides of this bispecific antibody can be replaced with an amino acid residue having a larger side chain volume to form a protrusion in the CH3 domain of that Fc polypeptide that can be positioned in a recess in the CH3 domain of the other Fc polypeptide, and an amino acid residue in the CH3 domain of the other Fc polypeptide can be replaced with an amino acid residue having a smaller side chain volume to form a recess in the CH3 domain of that Fc polypeptide.

[0118] In some particular embodiments, one Fc polypeptide of the Fc domain comprises T366Y / W or / and S354C and the other Fc polypeptide comprises Y407T / V, Y349C, T366S or / and L368A. More particularly, one Fc polypeptide of the Fc domain comprises the amino acid substitutions T366Y / W and S354C and the other Fc polypeptide comprises the amino acid substitutions Y407T / V, Y349C, T366S and L368A. More particularly, the Fc may be a human IgG1 Fc.

[0119] In some embodiments, the Fc domain comprises a modification that reduces or eliminates binding between the CH3 region of one Fc polypeptide in the Fc domain and Protein A from Staphylococcus aureus. In some embodiments, the modification is an amino acid substitution. In some embodiments, the Fc domain comprises the amino acid substitutions H435R or / and Y436F that occur only in one Fc polypeptide and not in the other Fc polypeptide. In one specific embodiment, the Fc domain comprises the amino acid substitutions H435R or / and Y436F that occur only in one of the Fc polypeptides. In such a specific embodiment, the Fc is an IgG1 Fc, in particular a human IgG1 Fc.

[0120] [Composition] The present invention provides pharmaceutical compositions comprising a bispecific antibody, further comprising one or more pharma- ceutically acceptable carriers, including, for example, excipients, diluents, encapsulating agents, fillers, buffers, or other agents.

[0121] [Isolated Nucleic Acid] The present invention provides isolated polynucleotides encoding said bispecific antibodies. The nucleic acid sequences of the polypeptide chains of some bispecific antibodies are exemplarily listed in the sequence listing.

[0122] [vector] The present invention provides an isolated vector comprising said polynucleotide. In some embodiments, said vector is a cloning vector, in other embodiments, said vector is an expression vector, in particular the expression vector is pcDNA3.1. The expression vector may be any expression vector capable of expressing a multispecific antibody as described herein.

[0123] [Host cells] In some embodiments, the invention provides a host cell comprising a vector or polynucleotide of the invention, which is a host cell suitable for cloning or encoding a multispecific antibody. In some embodiments, the host cell is a prokaryotic cell. In other embodiments, the host cell is a eukaryotic cell. In some embodiments, the host cell is selected from a yeast cell, a mammalian cell, or other cells suitable for producing a multispecific antibody. The mammalian cell is, for example, a Chinese Hamster Ovary (CHO) cell, a CHO-S cell.

[0124] [Method for expressing bispecific antibodies] The invention provides a method for expressing a bispecific antibody of the invention, comprising culturing a host cell under conditions whereby said bispecific antibody is expressed.

[0125] To produce said bispecific antibody, a polynucleotide encoding said bispecific antibody is isolated and inserted into one or more vectors for further cloning and / or expression in a host cell. The polynucleotides can be obtained using various methods well known to those skilled in the art, such as gene splicing, chemical synthesis, etc.

[0126] [Conjugate] The invention provides a conjugate comprising a bispecific antibody according to the invention and a therapeutic agent that binds or is conjugated to the bispecific antibody.

[0127] [Fusion protein] The invention provides fusion proteins in which a fusion moiety is expressed fused to a bispecific antibody according to the invention. The construction and expression of the fusion protein molecule can be achieved by genetic engineering means.

[0128] [use] The invention provides a method for treating a disease in a subject in need thereof comprising administering to the subject a therapeutically effective amount of a bispecific antibody, pharmaceutical composition, conjugate or fusion protein according to the invention.

[0129] An example of the disease may be cancer, and some examples of cancer are selected from, but are not limited to, leukemia, lymphoma, myeloma, ovarian cancer, breast cancer, endometrial cancer, colon cancer, rectal cancer, kidney cancer, bladder cancer, urothelial cancer, lung cancer, bronchial cancer, bone cancer, prostate cancer, pancreatic cancer, gastric cancer, hepatocellular carcinoma, gallbladder cancer, bile duct cancer, esophageal cancer, renal cell carcinoma, thyroid cancer, head and neck cancer, testicular cancer, endocrine gland cancer, adrenal gland cancer, pituitary gland cancer, skin cancer, soft tissue cancer, blood vessel cancer, brain cancer, nerve cancer, eye cancer, meningeal cancer, oropharynx cancer, hypopharyngeal cancer, cervical cancer, uterine cancer, glioblastoma, medulloblastoma, astrocytoma, glioma, meningioma, gastrinoma, neuroblastoma, melanoma, myelodysplastic syndrome, or sarcoma.

[0130] The present invention also provides several specific embodiments below, but is not limited thereto.

[0131] Embodiment 1. A bispecific antibody comprising a first antigen-binding portion and a second antigen-binding portion that specifically bind to two different antigens or different epitopes of the same antigen, The first antigen-binding portion comprises a first polypeptide comprising, from N-terminus to C-terminus, a first heavy chain variable domain and a first paired domain operably linked to the first heavy chain variable domain; a second polypeptide comprising, from N-terminus to C-terminus, a first light chain variable domain and a second paired domain operably linked to the first light chain variable domain; Here, one of the first paired domain and the second paired domain comprises the amino acid sequence of an engineered modified HLA-Iα3, and the other paired domain comprises the amino acid sequence of an engineered modified β2 microglobulin, in a bispecific antibody. Embodiment 2. The bispecific antibody of embodiment 1, wherein the first paired domain and the second paired domain are capable of forming a dimer, and wherein at least one non-natural interchain bond is formed between the first paired domain and the second paired domain, and wherein the non-natural interchain bond is capable of stabilizing the dimer. Embodiment 3. The bispecific antibody of embodiment 1 or 2, wherein the first paired domain comprises the amino acid sequence of an engineered HLA-I α3 and the second paired domain comprises the amino acid sequence of an engineered β2 microglobulin. Embodiment 4. The bispecific antibody of embodiment 1 or 2, wherein the first paired domain comprises the amino acid sequence of an engineered β2 microglobulin and the second paired domain comprises the amino acid sequence of an engineered HLA-Iα3. Embodiment 5. The bispecific antibody of any one of Embodiments 1 to 4, wherein the non-native interchain bonds are 1 to 3, preferably 1. Embodiment 6. The bispecific antibody of any one of embodiments 1 to 5, wherein the amino acid residues that form the non-native interchain bond are at the contact interface between the first paired domain and the second paired domain. Embodiment 7. The bispecific antibody of any one of embodiments 1 to 6, wherein the non-native interchain bond is a disulfide bond. Embodiment 8. The bispecific antibody of any one of embodiments 1 to 7, wherein the amino acid sequence of the engineered HLA-I α3 has at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence shown in SEQ ID NO:1 and the amino acid sequence of the engineered β2 microglobulin has at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence shown in SEQ ID NO:2. Embodiment 9. The bispecific antibody of any one of embodiments 1 to 8, wherein the engineered HLA-I α3 comprises an amino acid sequence having an amino acid substitution in SEQ ID NO:1 and the engineered β2 microglobulin comprises an amino acid sequence having an amino acid substitution in SEQ ID NO:2. Embodiment 10. A bispecific antibody according to embodiment 9, wherein said amino acid substitutions in the engineered HLA-I α3 and in the engineered β2 microglobulin both occur at both contact interfaces and comprise cysteine ​​residue substitutions capable of forming disulfide bonds with each other. Embodiment 11. The bispecific antibody of embodiment 10, wherein the cysteine ​​residue substitutions are one to several pairs selected from the group consisting of: (1) R60C in SEQ ID NO:1 and Y26C in SEQ ID NO:2 (2) A62C in SEQ ID NO:1 and R12C in SEQ ID NO:2 (3) G63C in SEQ ID NO:1 and Y67C in SEQ ID NO:2 Embodiment 12. The bispecific antibody of any one of embodiments 9 to 11, wherein the amino acid substitutions in the engineered HLA-I α3 and / or engineered β2 microglobulin comprise amino acid substitutions that improve the isoelectric point of the dimeric or bispecific antibody formed by the paired domains. Embodiment 13. The bispecific antibody of embodiment 12, wherein the isoelectric point of the dimer or bispecific antibody formed by the paired domains is increased to between 6.5 and 9.0. Embodiment 14. The bispecific antibody according to any one of embodiments 12 to 13, wherein the amino acid substitutions that improve the isoelectric point comprise substitutions with positively charged amino acids at one or more of the following positions in the engineered HLA-Iα3: E3, D22, E48, D53, E90, E101 of SEQ ID NO:1. Embodiment 15. The bispecific antibody of any one of embodiments 12 to 14, wherein the amino acid substitutions that improve the isoelectric point include substitutions of positively charged amino acids at one or more of the following positions in the engineered β2 microglobulin: E74, E47, E69, D34, E16, D53, E44, E50, E36 of the sequence SEQ ID NO:2. Embodiment 16. The bispecific antibody of embodiment 15, wherein the amino acid substitutions that improve the isoelectric point comprise substitutions of positively charged amino acids at positions D22, E48 and D53 of sequence SEQ ID NO:1 in said engineered HLA-I α3 and substitutions of positively charged amino acids at position E69 of sequence SEQ ID NO:2 in said engineered β2 microglobulin. Embodiment 17. The bispecific antibody of any one of embodiments 14 to 16, wherein the positively charged amino acid is K or R. Embodiment 18. The bispecific antibody of any one of embodiments 12 to 13, wherein the amino acid substitutions that improve the isoelectric point comprise said engineered HLA-I α3 comprising the amino acid substitutions D22R, E48K and D53K in the sequence of SEQ ID NO:1, and said engineered β2 microglobulin comprising the amino acid substitution E69R in the sequence of SEQ ID NO:2. Embodiment 19. The bispecific antibody of any one of embodiments 1 to 18, wherein said engineered HLA-I α3 comprises the amino acid sequence of SEQ ID NO:3 and said engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO:4. Embodiment 20. The bispecific antibody of any one of embodiments 1 to 18, wherein said engineered HLA-I α3 comprises the amino acid sequence of SEQ ID NO:3 and said engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO:5. Embodiment 21. The bispecific antibody of any one of Embodiments 1 to 20, wherein the second antigen-binding portion comprises a second heavy chain variable domain and a second light chain variable domain. Embodiment 22. The bispecific antibody of any one of Embodiments 1 to 21, wherein the second antigen-binding portion comprises a Fab. Embodiment 23. The bispecific antibody of embodiment 22, wherein the amino acid at position F118 of the CL domain of said Fab has been substituted with alanine according to EU numbering. Embodiment 24. The bispecific antibody of embodiment 23, wherein the CL domain comprises the amino acid sequence set forth in SEQ ID NO:44. Embodiment 25. The bispecific antibody of any one of Embodiments 1 to 24, wherein one of the first and second antigen-binding portions binds to a T cell-specific receptor molecule and / or a natural killer cell-specific receptor molecule, and the other binds to a tumor-associated antigen. Embodiment 26. The bispecific antibody of any one of embodiments 1 to 24, wherein one of the first and second antigen-binding portions specifically binds to CD3, and the other specifically binds to a tumor-associated antigen. Embodiment 27. The bispecific antibody of any one of embodiments 1 to 24, wherein both the first antigen-binding portion and the second antigen-binding portion specifically bind to a tumor-associated antigen. Embodiment 28. The bispecific antibody of any one of embodiments 25 to 27, wherein the tumor-associated antigen is CD38, CD47, PD-L1 or PD-1. Embodiment 29. The bispecific antibody of embodiment 26, wherein the first antigen-binding portion specifically binds to CD3 and the second antigen-binding portion specifically binds to CD38, or the first antigen-binding portion specifically binds to CD38 and the second antigen-binding portion specifically binds to CD3. Embodiment 30. The bispecific antibody of embodiment 27, wherein the first antigen-binding portion specifically binds PD-L1 and the second antigen-binding portion specifically binds CD47, or the first antigen-binding portion specifically binds CD47 and the second antigen-binding portion specifically binds PD-L1. Embodiment 31 The bispecific antibody of any one of Embodiments 1 to 30, wherein the bispecific antibody is bivalent. Embodiment 32. The bispecific antibody of any one of embodiments 1 to 30, further comprising a third antigen-binding portion that binds to the same antigen epitope as the second antigen-binding portion, preferably wherein said third antigen-binding portion is identical to the second antigen-binding portion. Embodiment 33. The bispecific antibody of any one of embodiments 1 to 30, further comprising a third antigen-binding portion that binds to the same antigen epitope as the first antigen-binding portion, preferably wherein said third antigen-binding portion is identical to the first antigen-binding portion. Embodiment 34 The bispecific antibody of embodiment 32 or 33, wherein the third antigen-binding portion is operably linked to the N-terminus or C-terminus of the first antigen-binding portion. Embodiment 35 The bispecific antibody of embodiment 32 or 33, wherein the third antigen-binding portion is operably linked to the N-terminus or C-terminus of the second antigen-binding portion. Embodiment 36 The bispecific antibody of any one of Embodiments 32 to 35, wherein the bispecific antibody is trivalent. Embodiment 37. A bispecific antibody according to any one of embodiments 1 to 31, further comprising an Fc domain consisting of two Fc polypeptides capable of stably associating. Embodiment 38. The bispecific antibody of embodiment 37, wherein the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide and the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of another Fc polypeptide. Embodiment 39. The bispecific antibody of embodiment 38, wherein the first polypeptide of said first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, and said second antigen-binding portion comprises a Fab, the Fab heavy chain of which is operably linked at its C-terminus to the N-terminus of another Fc polypeptide. Embodiment 40. A bispecific antibody according to any one of embodiments 32 to 36, further comprising an Fc domain consisting of two Fc polypeptides capable of stably associating. Embodiment 41. The bispecific antibody of embodiment 40, wherein the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide, and the third antigen-binding portion is operably linked at its C-terminus to the N-terminus of the first antigen-binding portion or the N-terminus of the second antigen-binding portion. Embodiment 42. The bispecific antibody of embodiment 40, wherein the third antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of the third antigen-binding portion, and the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide. Embodiment 43. The bispecific antibody of embodiment 40, wherein the third antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of the third antigen-binding portion, and the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide. Embodiment 44. The bispecific antibody of embodiments 41 to 43, wherein the operative linkage of the third antigen-binding moiety to the first antigen-binding moiety is via a peptide linker, or the operative linkage of the third antigen-binding moiety to the second antigen-binding moiety is via a peptide linker. Embodiment 45. The bispecific antibody of any one of embodiments 37 to 44, wherein the Fc domain is an IgG Fc domain. Embodiment 46. The bispecific antibody according to embodiment 45, wherein said IgG Fc domain is a human IgG Fc domain, preferably a human IgG1 or human IgG4 Fc domain. Embodiment 47. A bispecific antibody according to any one of embodiments 37 to 46, wherein the two Fc polypeptides are stably associated via one or several of the following: a linker, a disulfide bond, a hydrogen bond, an electrostatic interaction, a salt bridge, a hydrophobic-hydrophilic interaction. Embodiment 48. The bispecific antibody of any one of embodiments 37 to 47, wherein the Fc domain comprises a modification that promotes association of the two Fc polypeptides of the Fc domain. Embodiment 49. A bispecific antibody according to any one of embodiments 37 to 48, wherein one Fc polypeptide of said Fc polypeptides comprises T366Y / W and S354C and the other Fc polypeptide comprises Y407T / V, Y349C, T366S and L368A according to EU numbering. Embodiment 50. The bispecific antibody of any one of embodiments 37 to 49, wherein the Fc domain comprises a modification that reduces or eliminates binding between the CH3 region of one Fc polypeptide within the Fc domain and Protein A. Embodiment 51. A bispecific antibody according to any one of embodiments 37 to 50, wherein the Fc domain comprises the amino acid substitutions H435R or / and Y436F, which occur only in one of the Fc polypeptides, according to EU numbering. Embodiment 52. An isolated polynucleotide encoding the bispecific antibody of any one of embodiments 1 to 51. Embodiment 53. An isolated vector comprising the polynucleotide of embodiment 52. Embodiment 54. A host cell comprising the isolated polynucleotide of embodiment 52 or the isolated vector of embodiment 51. Embodiment 55. A method for expressing a bispecific antibody according to any one of embodiments 1 to 51, comprising culturing a host cell according to embodiment 52 under conditions in which the bispecific antibody is expressed. Embodiment 56. A pharmaceutical composition comprising the bispecific antibody of any one of embodiments 1 to 51 and a pharma- ceutically acceptable carrier. Embodiment 57. A conjugate comprising the bispecific antibody of any one of Embodiments 1 to 51 and a therapeutic agent that binds or conjugates to the bispecific antibody. Embodiment 58. A fusion protein comprising a fusion moiety expressed as a fusion and a bispecific antibody according to any one of embodiments 1 to 51. Embodiment 59. A method for treating a disease in a subject in need thereof, comprising administering to the subject a therapeutically effective amount of a bispecific antibody described in any one of embodiments 1 to 51, a pharmaceutical composition described in embodiment 56, a conjugate described in embodiment 57, or a fusion protein described in embodiment 58. EXAMPLES

[0132] [Design of MHC-I element sequences] According to statistics (Table 1) from the IPD-IMGT / HLA (https: / / www.ebi.ac.uk / ipd / imgt / hla / stats.html) database (version 3.42), HLA-I has a higher number and a wider distribution of alleles than HLA-II, so designing structural elements of bispecific antibodies using HLA-I molecules reduces the risk of immunogenicity. According to the statistics of the IPD-IMGT / HLA database (version 3.42) (Table 2), HLA-I molecules have the highest number of HLA-IB alleles, so designing the structural elements of bispecific antibodies using HLA-IB molecules further reduces the immunogenicity risk. [Table 1] [Table 2]

[0133] 1.1 Determining the initial sequence of MHC-I elements A series of full-length sequences of natural HLA-IB molecule α chains (Table 3) were obtained from the Uniprot (https: / / www.uniprot.org / ) database. The full-length sequence of the HLA-I molecule β chain (also called β microglobulin) (Uniprot ID: P61769) was obtained from the Uniprot database. The cross-indexing system of the Uniprot database page further obtained information on protein crystal data containing the HLA-IB molecule α chain, and high-resolution crystal structures (Resolution<2Å) were selected to ensure the accuracy of the crystal data (Table 4). Sequence analysis was performed and the information on these crystal structures was combined to determine the relative start position of the constant region of the natural HLA-IB molecule α chain (here, defined as MHCICα, constant region alpha of MHC I, or abbreviated as Cα). The constant region of the HLA-IB molecule α chain has a certain degree of sequence diversity, so multiple sequence alignment was performed for the sequences in Table 3 using the online alignment tool ClustalW2 (https: / / www.ebi.ac.uk / Tools / phylogeny / simple_phylogeny / ), and the alignment results were analyzed and displayed using web Logo (http: / / weblogo.berkeley.edu / ). The consistent sequence of the constant region of the HLA-IB molecule α chain was obtained from the frequency of occurrence of amino acids at each position (Figure 1), and the consistent sequence of the constant region of the HLA-IB molecule α chain was used as the initial sequence for subsequent manipulation and modification. Since there is no sequence diversity in human β2 microglobulin, the sequence of the pairing region between native β2 microglobulin and MHCICα (herein defined as MHCICβ, constant region beta of MHC I, or abbreviated as Cβ) was determined by analyzing the sequence (Uniprot ID: P61769). The initial sequence of the MHC-I element is as follows: > MHCICαori (SEQ ID NO:1) GKETLQRADPPKTHVTHHPISDHEATLRCWALGFYPAEITLTWQRDGEDQTQDTELVETRPAGDRTFQKWAAVVVPSGEEQRYTCHVQHEGLPKPLTLRWEPS >MHCICβori (SEQ ID NO:2) IQRTPKIQVYSRHPAENGKSNFLNCYVSGFHPSDIEVDLLKNGERIEKVEHSDLSFSKDWSFYLLYYTEFTPTEKDEYACRVNHVTLSQPKIVKWDR [Table 3] [Table 4]

[0134] 1.2 Optimization of the initial sequence of MHC-I elements {1.2.1. Introduction of interchain disulfide bonds to enhance pairing stability of MHC-I elements} The above MHCICα ori and MHCICβ ori amino acid sequences were simultaneously loaded into MOE (Molecular Operating Environment) software, and heterodimer modeling was performed on them through the Homology modeling module of MOE, thereby obtaining the simulation structure of the MHC-I element. In the natural state, there are no disulfide bonds between the constant regions of MHC molecules. In order to increase the pairing stability of the MHC-I constant region and optimize its developability, cysteine ​​(Cys) can be introduced at the R60, A62, or G63 positions of the MHCICα ori at the contact interface between the MHCICα and MHCICβ domains, and cysteine ​​(Cys) can be introduced at the Y26, R12, or Y67 positions of the MHCICβ ori. Using the disulfide scan module of MOE, we simulated and calculated the delta stability after the introduction of disulfide bonds. When the A62C mutation was introduced into the MHCICα ori and the R12C mutation was introduced into the MHCICβ ori, the delta stability was the lowest at -1.62 kcal / mol, indicating that mutations at these two positions can form the most stable disulfide bond at the interface between Cα and Cβ (Table 5). [Table 5]

[0135] {1.2.2, Optimization of isoelectric point} The theoretical isoelectric point (pI) of the MHC-I element with disulfide bonds (A62C and R12C in Table 5) introduced into the initial sequence was 5.80, which was unfavorable for the development of subsequent processes. Therefore, the relevant charge modifications were made to the MHC-I element to improve its drug formability. The solvent accessible surface (SAS) and charged amino acid residues of the MHC-I element were analyzed through the Protein-properties module of the MOE software. In the MHC-I element, the amino acid residues with a solvent exposure degree of >40% and negatively charged / acidic are shown in Table 6. In addition, one or more of these acidic amino acid residues were replaced with basic amino acid residues (arginine (R) or lysine (K)) according to the structural environment, chemical environment, and application purpose in which the relevant residues are located, and the theoretical isoelectric point of the MHC-I element was improved to 6.5-9.0. Specifically, three mutations, D22R, E48K, and D53K, were introduced into MHCICα, and a single point mutation, E69R, was introduced into MHCICβ, thereby improving the theoretical isoelectric point of the MHC-I element to 7.8. [Table 6]

[0136] {1.2.3, other optimizations} Considering the large side chains of the four amino acids (-KWDR) at the end of the C region, this may lead to potential steric hindrance effects. The use of flexible G4S linking peptides reduces the conformational stability of the protein. Due to the structural characteristics of glycine (G) and proline (P), these two amino acids are mainly in irregular curled or bent positions, and glycine has no side chain group, which can well alleviate the steric hindrance effect, while proline has a five-membered ring structure in the side chain, which has a certain rigidity and can enhance the steric hindrance effect. Therefore, finally, adding GP to the carboxyl terminus of the Cβ region balances the potential steric hindrance effect and the conformational stability. The optimized sequence of the MHC-I element is as follows: > Cα(SEQ ID NO:3) GKETLQRADPPKTHVTHHPISRHEATLRCWALGFYPAEITLTWQRDGKDQTQKTELVETRPCGDRTFQKWAAVVVPSGEEQRYTCHVQHEGLPKPLTLRWEPS >Cβ (SEQ ID NO:4) IQRTPKIQVYSCHPAENGKSNFLNCYVSGFHPSDIEVDLLKNGERIEKVEHSDLSFSKDWSFYLLYYTRFTPTEKDEYACRVNHVTLSQPKIVKWDRGP EXAMPLES

[0137] Construction and characterization of MHC-I element-modified bispecific IgG4 antibodies To ensure that MHC-I elements can be used in bispecific antibodies, we selected the daratumumab antibody and the human targeting CD3 antibody SP34 to construct a bispecific antibody targeting CD3 and CD38, to verify the initial concept and to explore the effect of the relative orientation of Cα and βC on the MHC-I element-modified antigen-binding moiety. To avoid or mitigate mismatches between IgG heavy chains, all the bispecific antibodies subsequently constructed used a "knobs into holes" structure. The domains CH1 / CL of the IgG constant region of SP34 Fab were replaced with the corresponding Cα / Cβ and Cβ / Cα, respectively, to construct MHC-I modified antigen-binding moieties with different relative orientations, and the heavy chain of the SP34 arm used a "hole" structure, i.e., CH3-hole. The Fab of daratumumab was kept intact, and its heavy chain used a "knob" structure, i.e., CH3-knob. By combining the MHC-I modified CD3 targeting antigen binding moiety with daratumumab, two different bispecific antibodies were generated, designated MHI383-1122-IgG4 and MHI383-1133-IgG4 (see FIG. 2 for molecular structures), whose sequences are shown in Table 7. The MHC-I elements used were Cα of the sequence shown in SEQ ID NO:3 and Cβ of the sequence shown in SEQ ID NO:4. [Table 7] For each bispecific antibody, the DNA sequence encoding each polypeptide chain constituting the antibody was inserted into a vector to obtain an expression vector expressing the corresponding polypeptide chain. 25 μg of each expression vector encoding each multispecific antibody polypeptide chain was co-transfected into ExpiCHO-S cells (manufacturer: Shanghai Pharmaceutical Industry Research Institute, product number: 127200005), the transfection expression volume was 100 ml, and the cell density was 6E+06 cells / ml. After transfection, the cells were cultured at 32°C, 5% CO2, and 130 rpm for expression. On the 8th day, the cell supernatant was collected and the antibody protein was purified in one step using Protein A magnetic beads (manufacturer: GenScript, product number: L00695-20). The cell supernatant was added to 4 ml of magnetic beads, placed in a 50 ml centrifuge tube, and placed on a rotary destaining rocker (manufacturer: Shanghai Intelligent Technology, product number: ZHWY-304). The mixture was incubated at room temperature for 2 hours. The magnetic beads were adsorbed using a magnetic stand base, the supernatant was discarded, and the magnetic beads were thoroughly washed with buffer. The antibody protein was eluted using 0.1 mol / L glycine at pH 3.0 (manufacturer: Sinopharm Group Co., Ltd., product number: 62011516). The eluate was placed in an ultrafiltration tube (manufacturer: Millipore Corporation, product number: UFC501096) and concentrated. The buffer was replaced with 1x PBS (manufacturer: Shanghai Synthetic Technology Co., Ltd., product number: B040100-0005), and purified to obtain MHI383-1122-IgG4 and MHI383-1133-IgG4.

[0138] (2.1 Detection of binding affinity between bispecific antibodies and antigens by BLI) Using human CD38 protein (Beijing ACROBiosystems, product code: CD8-H5224) and human CD3E (Beijing ACROBiosystems, product code: CDE-H5223) as antigens, the affinity of MHI383-1122-IgG4 and MHI383-1133-IgG4 to human CD38 and CD3E proteins, respectively, was measured by biological interference membrane technology (BLI). The specific experimental method is as follows. Affinity assays were used to evaluate the affinity of MHI383-1122-IgG4 and MHI383-1133-IgG4 to human CD38 and CD3E proteins. The affinity assays were mainly performed by BLI technology. MHI383-1122-IgG4 or MHI383-1133-IgG4 was immobilized on a sensor, respectively, and the sensor was immersed in a constant concentration gradient of human CD38 and CD3E protein solutions, respectively. The binding and dissociation curves were collected in real time using Data Acquisition 11.0.0.64, and the data was analyzed using Data Analysis 11.0 to obtain the binding rate constant Ka, dissociation rate constant Kd, and equilibrium constant KD, and the equilibrium constant KD indicated the affinity. [Table 8] Experimental results showed that bispecific antibodies MHI383-1122-IgG4 and MHI383-1133-IgG4 with different Cα / Cβ orientations both showed good affinity to human CD38 and CD3E proteins, with the difference in affinity KD values ​​between them being <3-fold, mainly due to the loss of their dissociation rate constant Kd (Table 8). Therefore, MHC-I elements (Cα and Cβ) can be used to construct bispecific antibodies instead of CH1 and CL domains. EXAMPLES

[0139] Construction and characterization of MHC-I element-modified bispecific IgG1 antibodies To avoid or mitigate mismatches between IgG heavy chains, all the bispecific antibodies constructed used a "knobs into holes" structure. To investigate the effect of MHC-I elements (Cα of the sequence shown in SEQ ID NO:3 and Cβ of the sequence shown in SEQ ID NO:4) on bispecific antibodies of different IgG subtypes, the above MHI383-1122-IgG4 was converted to the IgG1 form to generate a new bispecific antibody named MHI383-ccff-IgG1 (the constructed structure is shown in Figure 3). Both CH2 domains of the two heavy chains of this antibody were double-mutated with LALA (L234A+L235A) to eliminate potential Fc effector effects. In this example, when the H435R+Y436F (EU numbering) double mutation is introduced into the CH3 domain of an MHC-I element-modified heavy chain, the heavy chain mismatch product containing the H435R+Y436F double mutation does not bind to protein A and is therefore removed in protein A affinity chromatography purification, improving the purity and purification efficiency of the desired bispecific antibody. MHI383-ccff-IgG1 has four polypeptide chains (Hc, Lc, Hf and Lf) with the following sequences: Hc: The amino acid sequence is shown in SEQ ID NO:18 and the nucleotide sequence is shown in SEQ ID NO:19. Lc: the amino acid sequence is shown in SEQ ID NO:20, and the nucleotide sequence is shown in SEQ ID NO:21. Hf: the amino acid sequence is shown in SEQ ID NO:22 and the nucleotide sequence is shown in SEQ ID NO:23. Lf: the amino acid sequence is shown in SEQ ID NO:24, and the nucleotide sequence is shown in SEQ ID NO:25. The eluted target protein peak was further purified by preparative size-exclusion chromatography using a SuperdexTM 200 Increase 10 / 300GL column and GE's AKTA system. The purification experiment was performed at a flow rate of 0.8 ml / min using an equilibration solution (50 mM Tris-HAc, 150 mM L-Arg-HCl, pH 7.5). Figure 4 shows the SEC purification map of MHI383-ccff-IgG1 molecule, and the main SEC peak (retention time RT = 7.174 min) was collected and used for subsequent mass spectrometry analysis. The main SEC peak was subjected to complete molecular weight identification by liquid chromatography-mass spectrometry.

[0140] A complete molecular weight analysis was performed using a Waters LC-MS system (Waters, Singapore, USA, UK). A MAbPac RP 4 μm 2.1 × 50 mm (Thermo, USA) column was used with A phase (0.1% formic acid in water) and B phase (0.1% formic acid in acetonitrile in water) as the mobile phase, and the detection wavelength was 280 nm. 1 μg of protein was injected into the liquid chromatography-mass spectrometer with a gradient of 5% B to 100% B within 5.5 min. The mass spectrometer used positive ion mode with a scan range of 200 to 4000 m / z. Data were collected with MassLynx 4.1 and processed with UNIFI 1.8.2.169. Mass spectrometry showed that the peak at 149934 Da (Figure 5) was consistent with the complete molecular weight of the desired correctly assembled bispecific antibody MHI383-ccff-IgG1.

[0141] Based on the sequence patterns, the amino acid sequences of the MHC-I elements were analyzed and no N-linked glycosylation site motifs (NXS / T, X = any amino acid other than proline) were observed. However, it was difficult to predict O-linked glycosylation sites from the sequences. Therefore, the N-sugars present in the main peak proteins of the above SEC were further hydrolytically cleaved using glycosidase PNGaseF. The proteins after N-sugar cleavage were characterized by peptidograms (protein sequencing) using tandem mass spectrometry to examine the presence or absence of possible O-glycosylation sites.

[0142] Peptide sequence coverage analysis was performed using a Waters LC-MS system (Waters, Singapore, USA, UK). An AdvanceBio Peptide Map 2.1 × 150 mm 2.7-Micron (Aglient, USA) column was used, with phase A (0.1% formic acid in water) and phase B (0.1% formic acid in acetonitrile in water) as the mobile phase, and the detection wavelength was 214 nm. Proteins treated with PNGaseF were denatured, reduced, and alkylated, and then enzymatically digested with trypsin overnight. An appropriate amount of the enzymatically digested protein was injected into a liquid chromatography-mass spectrometer for detection. The elution gradient was 0% B to 15% B within 13 min; 15% B to 40% B within 45 min. The mass spectrometer was in positive ion mode, with a scan range of 100 to 2000 m / z. Data were collected with MassLynx 4.1 and processed with UNIFI 1.8.2.169. The experimental results of peptide mapping experiments also revealed the absence of any O-linked glycosylation modifications in the MHI383-ccff-IgG1 molecule, thereby revealing the absence of any O-linked glycosylation sites in the MHC-I element. The MHC-I element-modified bispecific antibodies have lower glycoform heterogeneity than the TCR-modified based bispecific antibodies, which is more advantageous for downstream process development and quality control. EXAMPLES

[0143] [Design of mutation sites and verification of strand mismatch] The antibody HZ5G11 targets the extracellular segment of the human PD-L1 protein, and the antibody HZ14A9 targets the extracellular segment of the human CD47 protein. Both are monoclonal antibodies obtained by immunizing mice with the corresponding antigens, screening hybridomas, and then humanizing them. The human-targeting PD-L1 monoclonal IgG1K antibody HZ5G11 is composed of a heavy chain HZ5G11VH-CH1-IgG1 (heavy chain amino acid sequence shown in SEQ ID NO:26, nucleotide sequence shown in SEQ ID NO:27) and a light chain HZ5G11VL-IgK (light chain amino acid sequence shown in SEQ ID NO:28, nucleotide sequence shown in SEQ ID NO:29), and the human-targeting CD47 monoclonal IgG1K antibody HZ14A9 is composed of a heavy chain HZ14A9VH-CH1-IgG1 (heavy chain amino acid sequence shown in SEQ ID NO:45) and a light chain HZ14A9VL-IgK (light chain amino acid sequence shown in SEQ ID NO:46). Further, variants related to HZ14A9 and HZ5G11 based on MHC-I elements were constructed, namely, the domain CH1 of the IgG1 constant region of the HZ14A9 antibody heavy chain was replaced with MHCICα (SEQ ID NO:3) in the corresponding MHC-I element to generate the heavy chain HZ14A9VH-Cα-IgG1 (heavy chain amino acid sequence shown in SEQ ID NO:47), the domain CL of the IgK constant region of the HZ14A9 antibody light chain was replaced with MHCICβ (SEQ ID NO:4) in the corresponding MHC-I element to generate the light chain HZ14A9VL-Cβ (light chain amino acid sequence shown in SEQ ID NO:48), and the domain CH1 of the IgG1 constant region of the HZ5G11 antibody heavy chain was replaced with MHCICα in the corresponding MHC-I element to generate the heavy chain HZ5G11VH-Cα-IgG1 (heavy chain amino acid sequence shown in SEQ ID NO:49). The CL domain of the IgK constant region of the HZ5G11 antibody light chain was replaced with the corresponding MHC-I constant region element MHCICβ to generate light chain HZ5G11VL-Cβ (light chain amino acid sequence is shown in SEQ ID NO:32, nucleotide sequence is shown in SEQ ID NO:33).

[0144] In order to verify whether there is a possible interaction between the MHC-I elements MHCICα and MHCICβ and the antibody constant region CH1 and the light chain constant region CL, plasmids encoding different heavy and light chains were constructed, combined, paired, expressed, and purified as shown in Table 9. Specifically, DNA sequences were synthesized, and the DNA sequences encoding the heavy and light chains were inserted into vectors (e.g., the pcDNA3.1 vector disclosed in CN107001463A, the pCHO1.0 vector disclosed in CN109422811A, etc.) to obtain recombinant expression vectors expressing the corresponding heavy or light chains. The recombinant expression vectors expressing the heavy and light chains were co-transfected into ExpiCHO-S cells (manufacturer: Shanghai Institute of Pharmaceutical Industry, product number: 127200005) at a mass ratio of 1:1, and then transfected into ExpiCHO-S (manufacturer: Shanghai Institute of Pharmaceutical Industry, product number: 127200005) with an expression volume of 0.1-1 L and a cell density of 6E+06 cells / ml. After transfection, cells were TM The cells were cultured in a medium (manufacturer: Thermo, product number: A2910001) at 32°C, 5% CO2, and 130 rpm for expression. On the 10th day, the cell supernatant was collected and the target protein was purified by protein A affinity chromatography using a GE AKTA Pure 25L system and MabSelect SuRe LX packing material. After elution, the protein concentration obtained by elution was measured using a NanoDrop Lite (Thermo Fisher Scientific), and the eluted protein purity was detected by SEC-HPLC. The sample protein purity used in the test was 80% or higher. A total of 5 mg of protein was taken from each sample, subjected to non-reducing SDS-PAGE electrophoresis, and after Coomassie Brilliant Blue staining, the electrophoretic results were observed. [Table 9] As shown in Figure 6, the experimental results show that the heterozygous pairing product is still generated based on the exchange of light and heavy chains between HZ5G11 and HZ14A9 monoclonal antibodies, and the molecular weight of the heterozygous pairing product is between HZ5G11 and HZ14A9 monoclonal antibodies. After co-expression and purification of HZ5G11VH-Cα-IgG1 and HZ14A9VL-IgK, and HZ14A9VH-Cα-IgG1 and HZ5G11VL-IgK, various products are generated. Among them, the molecular weight of the heterozygous pairing product of the combination of HZ5G11VH-Cα-IgG1 and HZ14A9VL-IgK is between 130-180kDa, which is close to HZ5G11 and HZ14A9 monoclonal antibodies. However, the combination of HZ14A9VH-Cα-IgG1 and HZ5G11VL-IgK has a molecular weight of the heterozygous pairing product between 100 and 130 kDa, which is easy to distinguish from HZ5G11 and HZ14A9 monoclonals. The above results indicate that there is an interaction between the CL domain and MHCICα, which further promotes the production of light and heavy chain mismatch products. After co-expression and purification of HZ14A9VH-CH1-IgG1 and HZ5G11VL-Cβ, and HZ5G11VH-CH1-IgG1 and HZ14A9VL-Cβ, no obvious bands were found in the corresponding lanes of non-reducing SDS-PAGE, i.e., there was no obvious generation of heterozygous pairing products, indicating no significant interaction between the CH1 domain and MHCICβ.

[0145] In order to solve the problem of mismatch between the heavy chain containing MHCICα and the light chain containing the CL domain, structural analysis was performed on the contact interfaces of the domains MHCICα (SEQ ID NO:3) / MHCICβ (SEQ ID NO:4) and IgG1 constant region CH1 (SEQ ID NO:34) / IgK constant region CL (SEQ ID NO:35). Combined with sequence analysis, it is believed that the sites 114 (EU numbering), 116 (EU numbering), and 118 (EU numbering) of the CL domain may be involved in the exchange process between MHCICβ and the CL domain in the subsequent bispecific antibody assembly, generating exchange mismatch products. To verify the effect of single point mutations at S114 (EU number), F116 (EU number), and F118 (EU number) in the CL domain of the IgK constant region on the pairing process of MHCICα and the CL domain, non-aromatic hydrophobic amino acids (M: methionine, A: alanine, V: valine, I: isoleucine, or L: leucine) were introduced into the S114 (EU number), F116 (EU number), or F118 (EU number) sites of the CL domain of the light chain HZ5G11VL-IgK, respectively, and expressed in combination with the heavy chain HZ14A9VH-Cα-IgG1. Table 10 shows the light chain sequence information of some of the antibodies containing the constructed mutations, and Table 11 shows some of the combinations of different heavy chain / light chain pairing expression. [Table 10] [Table 11] In the development of bispecific antibodies, mismatched products with molecular weights >100 kDa may not be effectively separated from the target product or may be difficult to separate, which may affect the quality of the bispecific antibody product. The results show that the pairing combination of HZ5G11VL-IgK-F118A and HZ14A9VH-Cα-IgG1 can significantly reduce the generation of pairing products with molecular weights >100 kDa (Figure 7), and the single point mutation of F118A can reduce the interaction between MHCICα and CL domains, thereby further resolving or mitigating the exchange of MHCICβ and CL domains in the subsequent bispecific antibody assembly, thereby improving the quality of the bispecific antibody product. EXAMPLES

[0146] [Effect of F118A mutation in the CL domain on monoclonal antibody assembly] To verify the effect of single point mutation of F118A (EU number) in the CL domain of IgK constant region on monoclonal antibody aggregation, monoclonal antibodies (HZ14A9 and HZ5G11) targeting CD47 and PD-L1 were subjected to single point mutation of F118A in the light chain IgK constant region, respectively, and the mutated light chains were expressed by pairing with the original heavy chains, respectively, to obtain HZ14A9-F118A and HZ5G11-F118A. In addition, wild-type monoclonal antibodies (HZ14A9 and HZ5G11) targeting CD47 and PD-L1 were expressed as controls, respectively. The construction and expression method of the antibody is described in Example 4. As shown in Figure 11, the results of the non-reducing SDS-PAGE experiment showed that the bands of all the detection samples were around 150 kDa, indicating that the single point mutation of F118A (EU numbering) in the CL domain of the IgK constant region does not affect the assembly of the monoclonal antibody. EXAMPLES

[0147] Construction and performance characterization of anti-PDL1*CD47 bispecific antibodies (5.1 Construction and Representation) To verify the effect of the single point mutation of F118A (EU numbering) in the CL domain of the IgK constant region on the bispecific antibody, two bispecific antibodies targeting CD47 and PD-L1 were constructed, in which the antigen-binding moiety targeting PD-L1 is an MHC-I element-modified antigen-binding moiety, in which the CH1 / CL of antibody HZ5G11 is replaced by MHCICα / MHCICβ, and the antigen-binding moiety targeting CD47 is a Fab, constructed based on antibody HZ14A9, in which the CL domain of the constant region of the Fab light chain has no mutation at the F118 (EU numbering) site. squid, F118A single point mutation had The CH2 domains of the two heavy chains of all antibodies carried the LALA (L234A+L235A) double mutation to eliminate potential Fc effector effects, and the CH3 domains carried the KIH mutation to promote the formation of heterodimers. This resulted in the generation of two bispecific antibodies that simultaneously target CD47 and PD-L1, namely MHL147-3322-IgG1-wt (see molecular structure Figure 8) and MHL147-3322-IgG1-F118A (see molecular structure Figure 9). [Table 12-A1] [Table 12-A2]

[0148] The DNA sequence constituting each polypeptide chain was inserted into a vector (e.g., pcDNA3.1 vector disclosed in CN107001463A, pCHO1.0 vector disclosed in CN109422811A, etc.) to obtain an expression vector expressing the corresponding polypeptide chain. The expression vectors encoding various polypeptide chains were co-transfected into ExpiCHO-S cells (manufacturer: Shanghai Institute of Pharmaceutical Industry Research, product number: 127200005) at an equal mass ratio, and then transfected into ExpiCHO-S (manufacturer: Shanghai Institute of Pharmaceutical Industry Research, product number: 127200005) with an expression volume of 0.1-1 L and a cell density of 6E+06 cells / ml. After transfection, the cells were transformed into ExpiCHO TM The cells were cultured in a medium (manufacturer: Thermo, product number: A2910001) at 32°C, 5% CO2, and 130 rpm for expression. On the 10th day, the cell supernatant was collected and protein A affinity chromatography of the target protein was performed using a GE AKTA Pure 25L system and MabSelect SuRe LX packing material. After elution, the protein concentration obtained by elution was measured using a NanoDrop Lite (Thermo Fisher Scientific), and the purity of the eluted protein was detected by SEC-HPLC. [Table 12-B] As a result of purification by one-step protein A affinity chromatography, MHL147-3322-IgG1-F118A showed higher expression level and purity than MHL147-3322-IgG1-wt ( Table 12-B ), a single point mutation of F118A (EU numbering) in the CL domain could improve the assembly and pairing process of bispecific antibodies in cells, reduce the formation of mismatch products, and increase the expression yield and purity of bispecific antibodies.

[0149] (5.2 Thermal stability) The thermal stability of HZ14A9, HZ5G11, MHL147-3322-IgG1-wt and MHL147-3322-IgG1-F118A was characterized using Nano DSF (Nano Temper). The concentration of the sample to be measured was adjusted to about 1.0 mg / ml, and then a small amount of the sample was sucked into a quartz capillary tube and placed in a Nano DSF sample suspension. A program was run at a constant rate of 1.0 °C / min from 20.0 °C to 95.0 °C, and the optical signal changes of the protein sample were collected to analyze the thermal stability of the protein. [Table 13] Nano DSF experiments showed that HZ14A9, HZ5G11, MHL147-3322-IgG1-wt and MHL147-3322-IgG1-F118A all had good thermal stability, with the melting temperature Tm1 of all samples being higher than 65°C (Table 13). The single point mutation of F118A (EU number) in the CL domain did not significantly affect the thermal stability of the bispecific antibody.

[0150] (5.3 Affinity) The affinities of the HZ5G11 monoclonal antibody, the HZ14A9 monoclonal antibody, the MHL147-3322-IgG1-wt and the MHL147-3322-IgG1-F118A bispecific antibodies were characterized by Biacore. By using the capture method, HZ5G11 monoclonal antibody, HZ14A9 monoclonal antibody, MHL147-3322-IgG1-wt and MHL147-3322-IgG1-F118A bispecific antibody were respectively immobilized on a sensor chip, and a constant concentration gradient of human PDL1 (manufacturer: ACROBiosystems, product code: PD1-H52H3) or human CD47 (manufacturer: ACROBiosystems, product code: CD7-H5227) protein antigen solution was injected, and the binding and dissociation curves were collected in real time using Biacore 8K Control Software 3.0. Then, the data was analyzed using Biacore Insight Evaluation Software 3.0 to obtain the binding rate constant Ka, the dissociation rate constant Kd and the equilibrium constant KD. [Table 14] [Table 15] The experimental results showed that both the bispecific antibodies MHL147-3322-IgG1-wt and MHL147-3322-IgG1-F118A showed good affinity for human PDL1 and CD47 proteins (Tables 15 and 16).

[0151] (5.4 Hemagglutination experiment) The potential hemosafety of MHL147-3322-IgG1-wt and MHL147-3322-IgG1-F118A was characterized by human erythrocyte aggregation experiments. Red blood cells were thoroughly washed to remove plasma attached to the surface of the red blood cell membrane. Red blood cells were washed three times with isotonic diluent, the first two times centrifuged at 2000 rpm / min for 5 min, and the last time centrifuged at 2000 rpm / min for 10 min, and the supernatant was discarded. Red blood cells were made into a 2% cell suspension with PBS, and 100 μL was aspirated and placed into a 96-well U-plate, centrifuged at 1500 rpm / min for 5 min, and the supernatant was discarded. MHL147-3322-IgG1-wt, MHL147-3322-IgG1-F118A and isotype control hIgG1 samples were prepared into eight concentration gradients (0, 0.1, 0.2, 0.7, 2, 6, 17, 50 μg / ml), antibody diluent was added to a 96-well U-plate, the total reaction volume was 100 μL, and the samples were resuspended and mixed uniformly, then placed in a 37°C incubator for 2 h, and the results were observed after incubation for 2 h. As shown in Figure 10, under the condition of a bispecific antibody concentration of ≧2 μg / ml, MHL147-3322-IgG1-F118A had a weaker hemagglutination ability than MHL147-3322-IgG1-wt, suggesting that the F118A (EU number) mutation in the CL domain of the IgK constant region effectively reduced the generation of mismatch products and improved the uniformity of the bispecific antibody.

Claims

1. 1. A bispecific antibody comprising a first antigen-binding portion and a second antigen-binding portion that specifically bind to two different antigens or different epitopes of the same antigen, The first antigen-binding portion comprises a first polypeptide comprising, from N-terminus to C-terminus, a first heavy chain variable domain and a first paired domain operably linked to the first heavy chain variable domain; a second polypeptide comprising, from N-terminus to C-terminus, a first light chain variable domain and a second paired domain operably linked to the first light chain variable domain; Here, one of the first paired domain and the second paired domain comprises the amino acid sequence of an engineered HLA-Iα3, and the other paired domain comprises the amino acid sequence of an engineered β2 microglobulin.

2. 2. The bispecific antibody of claim 1, wherein the first paired domain and the second paired domain are capable of forming a dimer, and wherein at least one non-natural interchain bond is formed between the first paired domain and the second paired domain, and wherein the non-natural interchain bond is capable of stabilizing the dimer.

3. the first paired domain comprises the amino acid sequence of an engineered HLA-Iα3 and the second paired domain comprises the amino acid sequence of an engineered β2 microglobulin; or 2. The bispecific antibody of claim 1 , wherein the first paired domain comprises the amino acid sequence of an engineered β2 microglobulin and the second paired domain comprises the amino acid sequence of an engineered HLA-I α3.

4. 2. The bispecific antibody of claim 1 , wherein the amino acid sequence of the engineered HLA-I α3 has at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence shown in SEQ ID NO:1 and the amino acid sequence of the engineered β2 microglobulin has at least 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% identity to the sequence shown in SEQ ID NO:

2.

5. 2. The bispecific antibody of claim 1 , wherein said engineered HLA-I α3 comprises an amino acid sequence having an amino acid substitution in SEQ ID NO:1 and said engineered β2 microglobulin comprises an amino acid sequence having an amino acid substitution in SEQ ID NO:

2.

6. said amino acid substitutions in both engineered HLA-I α3 and engineered β2 microglobulin occur at both contact interfaces and include cysteine ​​residue substitutions capable of forming disulfide bonds with each other; The bispecific antibody of claim 5 , wherein the cysteine ​​residue substitutions are one to multiple pairs selected from the following group: (1) R60C in SEQ ID NO: 1 and Y26C in SEQ ID NO: 2 (2) A62C in SEQ ID NO: 1 and R12C in SEQ ID NO: 2 (3) G63C in SEQ ID NO: 1 and Y67C in SEQ ID NO: 2

7. the amino acid substitutions in the engineered HLA-I α3 or / and engineered β2 microglobulin include amino acid substitutions that improve the isoelectric point of the dimeric or bispecific antibody formed by the paired domains; wherein the amino acid substitutions that improve the isoelectric point include substitutions of positively charged amino acids at one or more of the following positions in the engineered HLA-Iα3: E3, D22, E48, D53, E90, E101 of the sequence of SEQ ID NO: 1; or / and 6. The bispecific antibody of claim 5, wherein the amino acid substitutions that improve the isoelectric point comprise substitutions with positively charged amino acids at one or more of the following positions in the engineered β2 microglobulin: E74, E47, E69, D34, E16, D53, E44, E50, E36 of the sequence of SEQ ID NO:

2.

8. The bispecific antibody of claim 7, wherein the amino acid substitutions which improve the isoelectric point comprise the engineered HLA-I α3 comprising the amino acid substitutions D22R, E48K and D53K in the sequence of SEQ ID NO:1, and the engineered β2 microglobulin comprising the amino acid substitution E69R in the sequence of SEQ ID NO:

2.

9. the engineered HLA-I α3 comprises the amino acid sequence of SEQ ID NO: 3 and the engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO: 4; or 2. The bispecific antibody of claim 1 , wherein said engineered HLA-I α3 comprises the amino acid sequence of SEQ ID NO: 3 and said engineered β2 microglobulin comprises the amino acid sequence of SEQ ID NO:

5.

10. The bispecific antibody of claim 1 , wherein the second antigen-binding portion comprises a Fab.

11. The bispecific antibody of claim 10, wherein the amino acid at position F118 of the CL domain of the Fab is replaced with alanine according to EU numbering.

12. The bispecific antibody of claim 1 , wherein the bispecific antibody is bivalent.

13. binds to the same antigenic epitope as the second antigen-binding moiety; or binds to the same antigen epitope as the first antigen-binding portion The bispecific antibody of claim 1 , further comprising a third antigen-binding portion.

14. The bispecific antibody of claim 13, wherein the bispecific antibody is trivalent.

15. The bispecific antibody of claim 1 , further comprising an Fc domain consisting of two Fc polypeptides capable of stably associating.

16. the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide and the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide; or 16. The bispecific antibody of claim 15, wherein a first polypeptide of the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide and the second antigen-binding portion comprises a Fab, the Fab heavy chain of which is operably linked at its C-terminus to the N-terminus of another Fc polypeptide.

17. The bispecific antibody of claim 13, further comprising an Fc domain consisting of two Fc polypeptides capable of stably associating.

18. the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide, and the third antigen-binding portion is operably linked at its C-terminus to the N-terminus of the first antigen-binding portion or the N-terminus of the second antigen-binding portion; or the third antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of the third antigen-binding portion, and the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide; or 18. The bispecific antibody of claim 17, wherein the third antigen-binding portion is operably linked at its C-terminus to the N-terminus of one Fc polypeptide, the second antigen-binding portion is operably linked at its C-terminus to the N-terminus of the third antigen-binding portion, and the first antigen-binding portion is operably linked at its C-terminus to the N-terminus of the other Fc polypeptide.

19. An isolated polynucleotide encoding a bispecific antibody according to any one of claims 1 to 18.

20. An isolated vector comprising the polynucleotide of claim 19.

21. A host cell comprising the isolated polynucleotide of claim 19, or the isolated vector of claim 20.

22. A method for expressing a bispecific antibody according to any one of claims 1 to 18, comprising culturing a host cell according to claim 21 under conditions in which the bispecific antibody is expressed.

23. A pharmaceutical composition comprising the bispecific antibody of claim 1 and a pharma- ceutically acceptable carrier.

24. A conjugate comprising the bispecific antibody of claim 1 and a therapeutic agent that binds or conjugates to the bispecific antibody.

25. A fusion protein comprising a fusion component expressed by fusion and the bispecific antibody described in claim 1.

26. Use of a bispecific antibody according to any one of claims 1 to 18, a pharmaceutical composition according to claim 23, a conjugate according to claim 24 or a fusion protein according to claim 25 in the manufacture of a medicament for treating a disease in a subject in need thereof, comprising: The disease comprises leukemia, lymphoma, myeloma, ovarian cancer, breast cancer, endometrial cancer, colon cancer, rectal cancer, kidney cancer, bladder cancer, urothelial cancer, lung cancer, bronchial cancer, bone cancer, prostate cancer, pancreatic cancer, gastric cancer, hepatocellular carcinoma, gallbladder cancer, bile duct cancer, esophageal cancer, renal cell carcinoma, thyroid cancer, head and neck cancer, testicular cancer, endocrine gland cancer, adrenal gland cancer, pituitary gland cancer, skin cancer, soft tissue cancer, blood vessel cancer, brain cancer, nerve cancer, eye cancer, meningeal cancer, oropharynx cancer, hypopharyngeal cancer, cervical cancer, uterine cancer, glioblastoma, medulloblastoma, astrocytoma, glioma, meningioma, gastrinoma, neuroblastoma, melanoma, myelodysplastic syndrome or sarcoma.

Images