Bispecific binding construct

Abstract

New formats of bispecific binding constructs and methods of making the same are described. In addition, uses in therapeutic indications are described.

Description

[0001] This application is a divisional application of the invention patent filed on June 5, 2020, with application number 202080055055.6 (PCT / US2020 / 036464) and entitled "Bispecific Binding Construct".

[0002] Cross-references to related applications This application claims priority to U.S. Provisional Application No. 62 / 858,509, filed June 7, 2019, and U.S. Provisional Application No. 62 / 858,630, filed June 7, 2019. Each of these applications is incorporated herein by reference for all purposes. Technical Field

[0003] This invention belongs to the field of protein engineering. Background Technology

[0004] In recent years, bispecific binding constructs have shown therapeutic promise. For example, bispecific T-cell adaptors (BiTE) have been used to... ® Bispecific constructs targeting both CD3 and CD19 in this form showed impressive efficacy at low doses. Bargou et al. (2008), Science 321: 974-978. This BiTE... ® The form comprises two scFvs linked by a flexible connector, one targeting CD3 and the other targeting the tumor antigen CD19. This unique design allows the bispecific construct to bring activated T cells close to target cells, resulting in cytolytic killing of the target cells. See, for example, WO 99 / 54440 A1 (US Patent No. 7,112,324 B1) and WO 2005 / 040220 (US Patent Application Publication No. 2013 / 0224205 A1). A later development was the bispecific construct that binds to a context-independent epitope at the N-terminus of the CD3ε chain (see WO 2008 / 119567; US Patent Application Publication No. 2016 / 0152707 A1).

[0005] In the biopharmaceutical industry, molecules are typically produced on a large scale to meet commercial demands for supplying a large number of patients, and several properties can be evaluated to mitigate the risk that the molecules are unsuitable for large-scale production and purification. Efficient expression of these complex recombinant peptides can be a persistent challenge. Furthermore, even once expressed, peptides are often not as stable as desired in pharmaceutical compositions. Therefore, there is a need in the art for bispecific therapeutic agents that possess favorable pharmacokinetic properties and therapeutic efficacy, as well as forms that provide efficient production and increased stability. Summary of the Invention

[0006] This article describes several novel forms of bispecific antibodies. In one embodiment, the present invention provides a bispecific binding construct comprising a polypeptide chain having an amino acid sequence of the formula VH1-L1-VH2-L2-VL1-L3-VL2, wherein VH1 and VH2 are variable regions of the immunoglobulin heavy chain, VL1 and VL2 are variable regions of the immunoglobulin light chain, and L1, L2, and L3 are linkers, wherein L1 has at least 10 amino acids, L2 has at least 15 amino acids, and L3 has at least 10 amino acids, and wherein the bispecific binding construct is capable of binding to immune effector cells and target cells.

[0007] In another embodiment, the present invention provides a bispecific binding construct comprising a polypeptide chain having an amino acid sequence of the formula VH1-L1-VH2-L2-VL1-L3-VL2, wherein VH1 and VH2 are variable regions of the immunoglobulin heavy chain, VL1 and VL2 are variable regions of the immunoglobulin light chain, and L1, L2 and L3 are linkers, wherein L1 has at least 10 amino acids, L2 has at least 10 amino acids, and L3 has at least 10 amino acids, and wherein the total amino acid content of L1, L2 and L3 is at least 35 amino acids, and wherein the bispecific binding construct is capable of binding to immune effector cells and target cells.

[0008] In another embodiment, the present invention provides a bispecific binding construct comprising a polypeptide chain having an amino acid sequence of the formula VH1-L1-VH2-L2-VL1-L3-VL2-Fc, wherein VH1 and VH2 are variable regions of the immunoglobulin heavy chain, VL1 and VL2 are variable regions of the immunoglobulin light chain, Fc comprises an antibody Fc region (e.g., scFc), and L1, L2, L3, and L4 are linkers, wherein L1 has at least 10 amino acids, L2 has at least 10 amino acids, and L3 has at least 10 amino acids, and wherein the total amino acid content of L1, L2, and L3 is at least 35 amino acids, and wherein the bispecific binding construct is capable of binding to immune effector cells and target cells.

[0009] In another embodiment, the present invention provides nucleic acids encoding the bispecific binding constructs described herein, and vectors comprising these nucleic acids. Furthermore, the present invention provides host cells comprising the vectors described herein.

[0010] In yet another embodiment, the present invention provides a method for manufacturing the bispecific binding construct described herein, the method comprising (1) culturing host cells under conditions expressing the bispecific binding construct, and (2) recovering the binding construct from a cell cluster or cell culture supernatant, wherein the host cells contain one or more nucleic acids encoding any one of the bispecific binding constructs described herein.

[0011] In other embodiments, the present invention provides a method for treating a cancer patient, the method comprising administering to the patient a therapeutically effective amount of the bispecific binding construct described herein.

[0012] In other embodiments, the present invention provides a method for treating a patient with an infectious disease, the method comprising administering to the patient a therapeutically effective amount of the bispecific binding construct described herein.

[0013] In other embodiments, the present invention provides a method for treating a patient with an autoimmune, inflammatory, or fibrotic condition, the method comprising administering to the patient a therapeutically effective amount of the bispecific binding construct described herein.

[0014] In another embodiment, the present invention provides a pharmaceutical composition comprising the bispecific binding construct described herein. Attached Figure Description

[0015] Figure 1 A representative diagram of an exemplary embodiment of the .HHLL form.

[0016] Figure 2 Representative diagrams of various molecular forms (including the standard BiTE form of HLHL and various embodiments of the HHLL form with different joint lengths).

[0017] Figure 3 Compared to the HHLL form of the connector length in a representative embodiment, the HLHL BiTE form and the representative diagram of the connector length are shown.

[0018] Figure 4A - Figure 4D . Figure 4A A representative molecular model of the neutral-oriented HHLL form was described. Figure 4B A representative molecular model of the HHLL form, rotated 90° around the Y-axis, is depicted. Figure 4C A representative molecular model of the HHLL form, rotated 180° around the Y-axis, is depicted. Figure 4D Representative molecular models in the HHLL form, rotated 180° around the Y-axis and 90° around the X-axis, are depicted. These models illustrate how the linker lengths relate to each other and how molecules can correctly form their structures.

[0019] Figure 5A and Figure 5B Representative examples of purification chromatograms and gel results for different wild-type (“WT”) (HLHL) constructs.

[0020] Figure 6A and Figure 6B Representative examples of purification chromatograms and gel electrophoresis results for different HHLL constructs.

[0021] Figure 7 Representative results of various HHLL construct expressions compared to WT.

[0022] Figure 8 Representative results on the chemical stability of various HHLL constructs compared to WT.

[0023] Figure 9A and Figure 9B Representative results of thermal stability of various HHLL constructs compared to WT.

[0024] Figure 10 Representative results from accelerated stability studies of various HHLL constructs at 40 °C compared to WT. The “332” construct was designed as a negative control based on molecular modeling.

[0025] Figure 11 Representative results of shearing studies on various HHLL constructs compared to WT.

[0026] Figure 12 Representative results of the combination study of various HHLL constructs compared to WT.

[0027] Figure 13 Representative results of target cell killing and potency of various HHLL constructs compared to WT.

[0028] Figure 14 Representative results of various HHLL constructs at weeks 0, 2, 4 and 8 in stability determination at -20°C compared to wild type (WT).

[0029] Figure 15 Representative results of various HHLL constructs in freeze-thaw assays compared to wild type (WT). Detailed Implementation

[0030] This article describes a novel form of a bispecific binding construct. This construct comprises a single polypeptide chain containing two immunoglobulin variable heavy chain (VH) regions, two immunoglobulin variable light chain (VL) regions, and optionally an Fc region (e.g., scFc) arranged in the following order: VH-connector-VH-connector-VL-connector-VL (“HHLL”) or VH-connector-VH-connector-VL-connector-VL-connector-scFc. Compared to, for example, the HLHL form, this bispecific binding construct, the HHLL form, provides enhanced stability and increased in vitro expression, while maintaining the intended function of binding to the desired target on immune effector cells and target cells. Therefore, this HHLL form provides a bispecific molecule that can be produced more efficiently and with greater stability, features sought in pharmaceutical compositions.

[0031] It should be understood that the foregoing general description and the following detailed description are exemplary and illustrative only, and do not limit the claimed invention. In this application, the singular is used to include the plural unless otherwise specified. In this application, the word “or” is used to mean “and / or” unless otherwise specified. Furthermore, the use of the term “including” and other forms such as “includes” and “included” is not limiting. Similarly, unless otherwise specified, terms such as “element” or “component” cover both elements and components comprising one unit and elements and components comprising more than one subunit. Moreover, the use of the term “portion” can include a moiety or the entire moiety.

[0032] Unless otherwise defined herein, the scientific and technical terms used in connection with this invention will have the meanings commonly understood by one of ordinary skill in the art. Furthermore, unless the context requires otherwise, singular terms will include plural and plural terms will include singular. Generally, the nomenclature and techniques used in connection with cell and tissue culture, molecular biology, immunology, microbiology, genetics, and protein and nucleic acid chemistry and hybridization described herein are those well-known and commonly used in the art. The methods and techniques of this invention are generally performed according to conventional methods well-known in the art and as described in various general and more specific references.

[0033] Polynucleotide and polypeptide sequences are represented using standard one- or three-letter abbreviations. Unless otherwise specified, the amino terminus of a polypeptide sequence is on the left and its carboxyl terminus is on the right; for single-stranded and double-stranded nucleic acid sequences, the 5' end of the upper strand is on the left and its 3' end is on the right. Specific portions of a polypeptide can be designated by the number of amino acid residues (e.g., amino acids 1 to 50) or by the actual residues at that site (e.g., asparagine to proline). Specific polypeptide or polynucleotide sequences can also be described by explaining their differences from a reference sequence.

[0034] definition The term "isolation" for a molecule (in the case of a molecule that is, for example, a polypeptide, a polynucleotide, a bispecific binding construct, or an antibody) means that, depending on its origin or derived source, it (1) is not associated with the naturally associated components that accompany it in its native state; (2) is substantially free of other molecules from the same species; (3) is expressed by cells of a different species; or (4) is not present in nature. Thus, a molecule (which is chemically synthesized or expressed in a cellular system different from the cell from which it is of natural origin) will be "isolated" from its naturally associated components. Purification techniques well known in the art can also be used to isolate a molecule to make it substantially free of its naturally associated components. Molecular purity or homogeneity can be determined in a variety of ways well known in the art. For example, the purity of a polypeptide sample can be determined using polyacrylamide gel electrophoresis and staining the gel with techniques well known in the art to visualize the polypeptide. For certain purposes, higher resolution can be provided using HPLC or other purification methods well known in the art.

[0035] The terms “polynucleotide,” “oligonucleotide,” and “nucleic acid” are used interchangeably throughout the text and include DNA molecules (e.g., cDNA or genomic DNA), RNA molecules (e.g., mRNA), DNA or RNA analogs produced using nucleotide analogs (e.g., peptide nucleic acids and non-naturally occurring nucleotide analogs), and hybrids thereof. Nucleic acid molecules can be single-stranded or double-stranded. In one embodiment, the nucleic acid molecule of the present invention comprises consecutive open reading frames encoding the binding construct or fragments, derivatives, mutant proteins, or variants thereof of the present invention.

[0036] A "vector" is a nucleic acid that can be used to introduce another nucleic acid, to which it is linked, into a cell. One type of vector is a "plasmid," which is a linear or circular double-stranded DNA molecule in which an additional nucleic acid segment can be linked. Another type of vector is a viral vector (e.g., replication-defective retroviruses, adenoviruses, and adeno-associated viruses), in which an additional DNA segment can be introduced into the viral genome. Some vectors (e.g., bacterial vectors containing bacterial origins of replication and free-living mammalian vectors) are capable of autonomous replication in the host cells in which they are introduced. Other vectors (e.g., non-free-living mammalian vectors) are integrated into the host cell's genome after introduction and thus replicate along with the host genome. An "expression vector" is a vector that can direct the expression of selected polynucleotides.

[0037] If a regulatory sequence affects the expression of a nucleotide sequence (e.g., the level, timing, or location of expression), the nucleotide sequence is "operably linked" to the regulatory sequence. A "regulatory sequence" is a nucleic acid that affects the expression (e.g., the level, timing, or location) of the nucleic acid to which it is operably linked. For example, a regulatory sequence may act directly on the regulated nucleic acid or through the action of one or more other molecules (e.g., a polypeptide that binds to the regulatory sequence and / or the nucleic acid). Examples of regulatory sequences include promoters, enhancers, and other expression control elements (e.g., polyadenylation signals).

[0038] "Host cell" is a cell that can be used to express nucleic acids (e.g., the nucleic acids of this invention). Host cells can be prokaryotes (e.g., *Escherichia coli*), or eukaryotes (e.g., single-celled eukaryotes (e.g., yeast or other fungi), plant cells (e.g., tobacco or tomato plant cells), animal cells (e.g., human cells, monkey cells, hamster cells, rat cells, mouse cells, or insect cells)) or hybridomas. Typically, a host cell is a cultured cell that can be transformed or transfected with a nucleic acid encoded by a polypeptide, which can then be expressed in the host cell. The phrase "recombinant host cell" can be used to refer to a host cell that has been transformed or transfected with the nucleic acid to be expressed. A host cell can also be a cell that contains a nucleic acid but does not express it at the desired level (unless a regulatory sequence is introduced into the host cell to operatively link the host cell to the nucleic acid). It should be understood that the term host cell refers not only to the specific test cell but also to the offspring or potential offspring of such cells. Because certain modifications may occur in subsequent generations due to, for example, mutations or environmental influences, such offspring may not be completely identical to the parent cell, but are still included within the scope of the terminology used herein.

[0039] A “single-chain variable fragment” (“scFv”) is a fusion protein in which the VL and VH regions are linked by a linker (e.g., a synthetic sequence of amino acid residues) to form a continuous protein chain, wherein the linker is long enough to allow the protein chain to fold back and form a monovalent antigen-binding site (see, for example, Bird et al., Science 242: 423-26 (1988) and Huston et al., 1988, Proc. Natl. Acad. Sci. USA 85: 5879-83 (1988)). For example, when in the context of other additional parts (e.g., the Fc region), scFv can be arranged as VH-linker-VL or VL-linker-VH.

[0040] The term "CDR" refers to the complementarity-determining region (also known as the "minimum recognition unit" or "hypervariant region") within the variable sequence of an antibody, and the bispecific binding construct of the present invention comprises heavy chain and / or light chain CDRs. CDRs allow the binding construct to bind specifically to a particular target antigen. There are three heavy chain variable region CDRs (CDRH1, CDRH2, and CDRH3) and three light chain variable region CDRs (CDRL1, CDRL2, and CDRL3). The CDRs in each of the two chains are typically aligned by framework regions to form a structure that specifically binds to a specific epitope or domain on the target protein. From the N-terminus to the C-terminus, the naturally occurring light and heavy chain variable regions typically follow the following order of components: FR1, CDR1, FR2, CDR2, FR3, CDR3, and FR4. The numbering system is obtained by assigning numbers to the amino acids occupying positions in each of these domains. This numbering system is defined in the following literature: Kabat Sequences of Proteins of Immunological Interest (1987 and 1991, NIH, Bethesda, Maryland); or Chothia and Lesk, 1987, J. Mol. Biol. 196: 901-917; Chothia et al., 1989, Nature 342: 878-883. The complementarity-determining region (CDR) and framework region (FR) of a given antibody can be identified using this system. Other numbering systems for amino acids in immunoglobulin chains include IMGT® (International ImMunoGeneTics Information System; Lefranc et al., Dev. Comp. Immunol. [Developmental and Comparative Immunology] 29: 185-203; 2005) and AHo (Honegger and Pluckthun, J. Mol. Biol. [Journal of Molecular Biology] 309 (3): 657-670; 2001). One or more CDRs can be covalently or non-covalently incorporated into the molecule to make it a binding construct.

[0041] The “binding domain” of the binding construct according to the invention may, for example, include the CDR group mentioned above. Preferably, those CDRs are contained within the framework of the antibody light chain variable region (VL) and antibody heavy chain variable region (VH) contained in the bispecific binding construct of the invention. Alternatively, in the terminology used herein, “L” and “H” are variable regions (e.g., “HHLL”).

[0042] The term "human antibody" includes antibodies having antibody regions that substantially correspond to variable and constant regions or domains of human immunoglobulin sequences known in the art, including, for example, those described by Kabat et al. (1991) (cited above). Human antibodies as referred to herein may include, for example, amino acid residues in CDR, and particularly CDR3, that are not encoded by human immunoglobulin sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or by somatic mutations in vivo). Human antibodies may have at least one, two, three, four, five, or more positions substituted by amino acid residues not encoded by human immunoglobulin sequences. The definition of human antibody as used herein also considers fully human antibodies, which consist only of human antibody sequences that are not artificially and / or genetically altered, such as those that can be obtained using techniques or systems known in the art (e.g., phage display technology or transgenic mouse technology, including but not limited to Xenomouse). ® Those derived from [the present invention]. In the context of this invention, variable regions derived from human antibodies can be used in the intended bispecific binding construct form.

[0043] Compared to non-human antibodies, when humanized antibodies are administered to human subjects, the sequence of the humanized antibody differs from that of antibodies derived from non-human species in that one or more amino acid substitutions, deletions, and / or additions are made so that the humanized antibody is less likely to induce an immune response and / or induces a less severe immune response. In one embodiment, certain amino acid mutations are made in the framework and constant domains of the heavy chain and / or light chain of a non-human antibody to produce a humanized antibody. In another embodiment, one or more constant domains from a human antibody are fused to one or more variable domains from a non-human species. In yet another embodiment, when a non-human antibody is administered to a human subject, one or more amino acid residues in one or more CDR sequences of the non-human antibody are altered to reduce the potential immunogenicity of the non-human antibody, wherein the altered amino acid residues are not critical for the immune-specific binding of the antibody or binding construct to its antigen, or the alteration to the amino acid sequence is a conserved alteration such that the binding of the humanized antibody to the antigen is not significantly inferior to the binding of the non-human antibody to the antigen. Examples of how to prepare humanized antibodies can be found in U.S. Patent Nos. 6,054,297, 5,886,152, and 5,877,293. In the context of this invention, the variable region derived from a humanized antibody can be used in the desired bispecific binding construct form.

[0044] The term "chimeric antibody" refers to an antibody containing one or more regions from one antibody and one or more regions from one or more other antibodies. In one embodiment, one or more of the CDRs are derived from a human antibody. In another embodiment, all CDRs are derived from a human antibody. In yet another embodiment, CDRs from more than one human antibody are mixed and matched in the chimeric antibody. For example, a chimeric antibody may comprise CDR1 from the light chain of a first human antibody, CDR2 and CDR3 from the light chain of a second human antibody, and CDRs from the heavy chain of a third antibody. Furthermore, the frame region may be derived from one of the same antibodies, from one or more different antibodies (such as human antibodies), or from a humanized antibody. In one instance of a chimeric antibody, a portion of the heavy chain and / or light chain is identical, homologous to, or derived from an antibody from a particular species or belonging to a particular antibody class or subclass, while the remaining portion of the chain is identical, homologous to, or derived from an antibody from another species or belonging to another antibody class or subclass. Fragments of such antibodies exhibiting the desired biological activity are also included. In the context of this invention, variable regions from chimeric antibodies can be used in the intended bispecific binding construct form.

[0045] This invention provides a bispecific binding construct comprising the form of HHLL. In the most general sense, a bispecific binding construct as described herein comprises several polypeptide chains with different amino acid sequences, which, when linked together, can bind to two different antigens. Optionally, the HHLL molecule further comprises a half-life extension portion. In some embodiments, the half-life extension portion is an Fc polypeptide chain. In other embodiments, the half-life extension portion is a single-chain Fc. In still other embodiments, the half-life extension portion is an iso-Fc. In still other embodiments, the half-life extension portion is human albumin.

[0046] connector Between the immunoglobulin variable regions are peptide linkers, which can be identical linkers or different linkers of varying lengths. Linkers play a crucial role in the structure of the bispecific binding construct, and the invention described herein provides not only suitable linker sequences but also suitable linker lengths at each position in the bispecific binding construct of the invention. If the linker is too short, it will allow the appropriate variable regions on a single polypeptide chain sufficient flexibility to interact to form antigen-binding sites. If the linker length is appropriate, it will allow one variable region to interact with another variable region on the same polypeptide chain to form an antigen-binding site. In some embodiments, the HHLL form comprises both disulfide bonds within the domain (within H1, L1) and between the domains (between H1 and L1). To achieve proper expression and conformation of the bispecific binding construct of the invention, in some embodiments, specific linkers are used between various immunoglobulin regions (see, for example, those described herein). Figure 1Exemplary linkers are provided in Table 1 of this document. In some embodiments, increasing the linker length may result in increased protein shearing (an undesirable property). Therefore, it is desirable to achieve an appropriate balance between linker lengths to allow for proper peptide structure and activity without causing increased shearing.

[0047] As used herein, a “linker” is a peptide that connects two polypeptides. In some embodiments, the linker can connect two immunoglobulin variable regions in the case of a bispecific binding construct. The linker length can be 2-30 amino acids. In some embodiments, the linker can be 2-25, 2-20, or 3-18 amino acids long. In some embodiments, the linker can be a peptide not exceeding 14, 13, 12, 11, 10, 9, 8, 7, 6, or 5 amino acids in length. In other embodiments, the linker can be 5-25, 5-15, 4-11, 10-20, or 20-30 amino acids long. In other embodiments, the linker can be about 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acids long. Exemplary linkers include, for example, the amino acid sequences GGGGS (SEQ ID NO: 1), GGGSGGGGS (SEQ ID NO: 2), GGGGSGGGGSGGGGS (SEQ ID NO: 3), GGGGSGGGGSGGGGSGGGGS (SEQ ID NO: 4), GGGGSGGGGSGGGGSGGGGSGGGGS (SEQ ID NO: 5), GGGGQ (SEQ ID NO: 6), GGGGQGGGGQ (SEQ ID NO: 7), GGGGQGGGGQGGGGQ (SEQ ID NO: 8), GGGGQGGGGQGGGGQGGGGQ (SEQ ID NO: 9), GGGGQGGGGQGGGGQGGGGQGGGGQ (SEQ ID NO: 10), GGGGSAAA (SEQ ID NO: 11), TVAAP (SEQ ID NO: 12), ASTKGP (SEQ ID NO: 13), and AAA (SEQ ID NO: 14), including repeats of the above amino acid sequences or subunits of the amino acid sequences (e.g., repeats of GGGGS (SEQ ID NO: 1) or GGGGQ (SEQ ID NO: 6)).

[0048] In some embodiments within the context of the HHLL molecule of the present invention, the linker sequence of linker 1 is at least 10 amino acids. In other embodiments, linker 1 is at least 15 amino acids. In other embodiments, linker 1 is at least 20 amino acids. In other embodiments, linker 1 is at least 25 amino acids. In other embodiments, linker 1 is at least 30 amino acids. In other embodiments, linker 1 is 10-30 amino acids. In other embodiments, linker 1 is 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acids. In still other embodiments, linker 1 is greater than 30 amino acids.

[0049] In some embodiments within the context of the HHLL molecule of the present invention, the linker sequence of linker 2 is at least 15 amino acids. In other embodiments, linker 2 is at least 20 amino acids. In other embodiments, linker 2 is at least 25 amino acids. In other embodiments, linker 2 is at least 30 amino acids. In other embodiments, linker 2 is 15-30 amino acids. In other embodiments, linker 2 is 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acids. In still other embodiments, linker 2 is greater than 30 amino acids.

[0050] In some embodiments within the context of the HHLL molecule of the present invention, the linker sequence of linker 3 is at least 15 amino acids. In other embodiments, linker 3 is at least 20 amino acids. In other embodiments, linker 3 is at least 25 amino acids. In other embodiments, linker 3 is at least 30 amino acids. In other embodiments, linker 3 is 15-30 amino acids. In other embodiments, linker 3 is 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acids. In still other embodiments, linker 3 is greater than 30 amino acids.

[0051] In some embodiments within the context of the HHLL molecule of the present invention, the linker sequence of linker 4 is at least 5 amino acids. In other embodiments, linker 4 is at least 10 amino acids. In other embodiments, linker 4 is at least 15 amino acids. In other embodiments, linker 4 is at least 20 amino acids. In other embodiments, linker 4 is at least 25 amino acids. In other embodiments, linker 4 is at least 30 amino acids. In other embodiments, linker 4 is 5-30 amino acids. In other embodiments, linker 4 is 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acids. In still other embodiments, linker 4 is greater than 30 amino acids.

[0052] Figures 4A-4D Molecular models of the HHLL construct in various orientations are depicted, demonstrating how a linker of a specific length is necessary for the HHLL construct to adopt an appropriate conformation and for both HL-binding domains to function. A, B, and C represent the distances between the C-α atoms of the terminal residues of one domain and the starting residues of the other domain. Using this information, skilled technicians can model the desired HHLL construct and adjust the linker length according to the needs of specific HL-binding domains, thereby enabling the HHLL construct to express and function as required.

[0053] In certain embodiments within the context of the HHLL molecule of the present invention, the linker sequence and position are listed in Table 1 below, wherein the linker position is related to... Figure 1 The corresponding ones listed in the table, and if the Fc region is also attached to the HHLL molecule, then adapter 4 may be used optionally.

[0054] Table 1 * Numerical subscripts indicate the number of repetitions, for example, (GGGGS) 2 =GGGGSGGGGS (SEQ ID NO: 2) Note that the 3-3-2 connector was intentionally designed to be of a non-optimal length as a negative control.

[0055] amino acid sequence of the binding region In the exemplary embodiments described herein, the bispecific binding constructs maintain desired binding to various desired targets because they assume appropriate conformations that allow this binding. The immunoglobulin variable region contains VH and VL domains that associate to form a variable domain for binding the desired target.

[0056] The variable domain can be obtained from any immunoglobulin with the desired characteristics, and methods for achieving this are further described herein. In one embodiment, VH1 and VL1 associate and bind to CD3ε, and VH2 and VL2 associate and bind to different targets. In another embodiment, VH2 and VL2 bind to CD3ε, and VH1 and VL1 bind to different targets.

[0057] In another embodiment, the light chain variable domain comprises an amino acid sequence having at least 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity with the sequence of the light chain variable domain described herein.

[0058] In another embodiment, the light chain variable domain comprises a sequence of amino acids encoded by a nucleotide sequence having at least 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity with the polynucleotide sequence described herein. In another embodiment, the light chain variable domain comprises a sequence of amino acids encoded by a polynucleotide that hybridizes under moderately stringent conditions with a complementary sequence of a polynucleotide encoding a light chain variable domain selected from the sequences described herein. In yet another embodiment, the light chain variable domain comprises a sequence of amino acids encoded by a polynucleotide that hybridizes under stringent conditions with a complementary sequence of a polynucleotide encoding a light chain variable domain selected from the group consisting of the sequences described herein.

[0059] In another embodiment, the heavy chain variable domain comprises an amino acid sequence having at least 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity with a sequence of the heavy chain variable domain selected from the sequences described herein. In another embodiment, the heavy chain variable domain comprises a sequence of amino acids encoded by a nucleotide sequence having at least 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity with a nucleotide sequence encoding a heavy chain variable domain selected from the sequences described herein. In another embodiment, the heavy chain variable domain comprises a sequence of amino acids encoded by a polynucleotide that hybridizes under moderately stringent conditions to a complementary sequence of a polynucleotide encoding a heavy chain variable domain selected from the sequences described herein. In another embodiment, the heavy chain variable domain comprises a sequence of amino acids encoded by a polynucleotide that hybridizes under stringent conditions to a complementary sequence of a polynucleotide encoding a sequence selected from those described herein.

[0060] replace It should be understood that the bispecific binding constructs of the present invention may have at least one amino acid substitution, provided that the binding construct retains the same or better desired binding specificity (e.g., binding to CD3). Therefore, modifications to the structure of the binding construct are within the scope of the present invention. In one embodiment, the binding construct comprises sequences that each independently differ from the CDR sequences described herein by 5, 4, 3, 2, 1, or 0 single amino acid additions, substitutions, and / or deletions. As used herein, a CDR sequence differing from the CDR sequences described herein by no more than, for example, four single amino acid additions, substitutions, and / or deletions refers to a sequence having 4, 3, 2, 1, or 0 single amino acid additions, substitutions, and / or deletions compared to the sequences described herein. These may include amino acid substitutions, which may be conserved or non-conserved and do not impair the desired binding ability of the binding construct. Conserved amino acid substitutions may encompass non-naturally present amino acid residues typically incorporated through chemical peptide synthesis rather than through synthesis in biological systems. These include peptide mimics and other reverse or inverse forms of the amino acid moiety. Conservative amino acid substitution may also involve replacing native amino acid residues with standard residues, such that the polarity or charge of the amino acid residue at that position has little or no effect.

[0061] Non-conservative substitution may involve replacing a member of one class of amino acids or amino acid mimics with a member of another class that has different physical properties (e.g., size, polarity, hydrophobicity, charge). In some embodiments, such substituted residues may be introduced into a human antibody region homologous to a non-human antibody or a non-homologous region of the molecule, thus enabling the generation of the binding constructs of the present invention.

[0062] Furthermore, those skilled in the art can generate test variants containing a single amino acid substitution at each desired amino acid residue. These variants can then be screened using activity assays known to those skilled in the art. Such variants can be used to gather information about suitable variants. For example, variants with such changes can be avoided if a change to a particular amino acid residue is found to result in compromised, undesirably reduced, or unsuitable activity. In other words, based on information gathered from such routine experiments, those skilled in the art can readily identify further substitutions of amino acids, either alone or in combination with other mutations, that should be avoided.

[0063] Those skilled in the art will be able to use well-known techniques to determine suitable variants of the binding constructs as described herein. In some embodiments, those skilled in the art can identify suitable regions of the molecule that have been altered without destroying activity by targeting regions believed to be unimportant to activity. In some embodiments, conserved residues and portions of the molecule among the similar peptides described above can be identified. In some embodiments, conserved amino acid substitutions may be made even in regions that may be important to biological activity or structure, without destroying biological activity or adversely affecting the peptide structure.

[0064] Furthermore, those skilled in the art can review structure-function studies that identify activity- or structurally important residues in similar peptides. Based on this comparison, the importance of amino acid residues in proteins can be predicted, as these residues correspond to activity- or structurally important amino acid residues in similar proteins. Those skilled in the art can then select chemically similar amino acid substitutions for these predicted important amino acid residues.

[0065] In some embodiments, those skilled in the art can identify modifiable residues that produce the desired enhancing properties. For example, amino acid substitutions (conserved or non-conserved) may result in enhanced binding affinity to a desired target.

[0066] Those skilled in the art can also analyze the three-dimensional structure and amino acid sequence associated with the structure of similar peptides. Given this information, those skilled in the art can predict the alignment of amino acid residues of an antibody based on its three-dimensional structure. In some embodiments, those skilled in the art may choose not to drastically alter the amino acid residues expected to be on the protein surface, as such residues may be involved in important interactions with other molecules. Numerous scientific publications have dedicated themselves to the prediction of secondary structures. See Moult J., Curr. Op. in Biotech. [Current Perspectives on Biotechnology], 7 (4):422-427 (1996); Chou et al., Biochemistry, 13 (2): 222-245 (1974); Chou et al., Biochemistry, 113 (2): 211-222 (1974); Chou et al., Adv. Enzymol. Relat. Areas Mol. Biol. [Advances in Enzymology and Molecular Biology], 47: 45-148 (1978); Chou et al., Ann. Rev. Biochem. [Annals of Biochemistry], 47: 251-276; and Chou et al., Biophys. J. [Journal of Biophysics], 26: 367-384 (1979). Furthermore, computer programs are currently available to assist in the prediction of secondary structures. One method for predicting secondary structures is based on homology modeling. For example, two polypeptides or proteins with greater than 30% sequence identity or greater than 40% similarity often have similar structural topologies. The growth of protein structure databases (PDBs) provides enhanced predictability of secondary structures, including the structure of polypeptides or the potential number of folds in the structure of proteins. See Holm et al., Nucl. Acid. Res. [Nucleic Acid Research], 27 (1): 244-247 (1999).Other methods for predicting secondary structures include threading (Jones, D., Curr. Opin. Struct. Biol. [New Insights in Structural Biology], 7 (3): 377-87 (1997); Sippl et al., Structure, 4 (1): 15-19 (1996)), profile analysis (Bowie et al., Science, 253: 164-170 (1991); Gribskov et al., Meth. Enzym. [Enzymological Methods], 183: 146-159 (1990); Gribskov et al., Proc. Nat. Acad. Sci. [Proceedings of the National Academy of Sciences], 84 (13):4355-4358 (1987)), and evolutionary linkage (see Holm, Ibid. (1999), and Brenner, Ibid. (1997).

[0067] In some embodiments, variants of the binding construct include glycosylation variants in which the number and / or type of glycosylation sites have been altered compared to the amino acid sequence of the parent polypeptide. In some embodiments, the variants contain more or fewer N-linked glycosylation sites than the native protein. Alternatively, substitution to eliminate this sequence will remove the existing N-linked carbohydrate chain. Rearrangements of the N-linked carbohydrate chain are also provided, in which one or more N-linked glycosylation sites (typically naturally occurring) are eliminated and one or more new N-linked sites are created. Additional variants include cysteine ​​variants in which one or more cysteine ​​residues are missing or substituted for another amino acid (e.g., serine) compared to the parent amino acid sequence. Cysteine ​​variants can be useful when antibodies or bispecific constructs must refold into a biologically active conformation, such as after the separation of insoluble inclusion bodies. Cysteine ​​variants generally have fewer cysteine ​​residues than the native protein, and typically an even number, to minimize interactions caused by unpaired cysteines.

[0068] The desired amino acid substitutions (whether conserved or non-conserved) can be determined by those skilled in the art when such substitutions are required. In some embodiments, amino acid substitutions can be used to identify important residues of the binding construct against a target, or to increase or decrease the affinity of the binding construct for the target described herein.

[0069] According to some embodiments, the desired amino acid substitutions are those that: (1) reduce sensitivity to protein hydrolysis; (2) reduce sensitivity to oxidation; (3) alter the binding affinity for forming protein complexes; (4) alter the binding affinity and / or (5) impart or alter other physiological or functional properties to such polypeptides. According to some embodiments, single or multiple amino acid substitutions (in some embodiments, conserved amino acid substitutions) may be made in naturally occurring sequences (in some embodiments, in polypeptide portions outside one or more domains forming intermolecular contacts). In some embodiments, conserved amino acid substitutions typically do not substantially alter the structural characteristics of the parental sequence (e.g., the substituted amino acid should not tend to cause helical breaks present in the parental sequence, or disrupt other types of secondary structures characteristic of the parental sequence). Examples of recognized examples of peptide secondary and tertiary structures are described in Proteins, Structures and Molecular Principles (edited by Creighton, WH Freeman and Company, New York (1984)); Introduction to Protein Structure (edited by C. Branden and J. Tooze, Garland Publishing, New York, NY (1991)); and Thornton et al., Nature 354: 105 (1991), which are incorporated herein by reference.

[0070] Extended half-life and Fc region In some embodiments, it is desirable to extend the in vivo half-life of the bispecific binding construct of the present invention. This can be achieved by including a half-life extension portion as part of the bispecific binding construct. Non-limiting examples of the half-life extension portion include Fc peptides, albumin, albumin fragments, portions that bind to albumin or neonatal Fc receptors (FcRn), fibronectin derivatives engineered to bind albumin or fragments thereof, peptides, single-domain protein fragments, or other peptides that can increase serum half-life. In alternative embodiments, the half-life extension portion can be a non-peptide molecule, such as polyethylene glycol (PEG).

[0071] As used herein, the term "Fc polypeptide" includes polypeptides in both native and mutant protein forms derived from the Fc region of an antibody. It also includes truncated forms of such polypeptides containing a hinge region that promotes dimerization. In addition to the other properties described herein, polypeptides containing the Fc moiety offer the advantage of purification via affinity chromatography, for example, using protein A or protein G columns.

[0072] In some embodiments, the half-life extension portion is the Fc region of the antibody. The Fc region may be located at the N-terminus of the HHLL bispecific binding construct, or it may be located at the C-terminus of the HHLL bispecific binding construct. A linker may be present between the HHLL bispecific binding construct and the Fc region, but is not required. As explained above, the Fc polypeptide chain may contain all or part of the hinge region, followed by CH2 and CH3 regions. The Fc polypeptide chain may be from mammals (e.g., humans, mice, rats, rabbits, dromedary camels, or new or old world monkeys), birds, or sharks. Additionally, as explained above, the Fc polypeptide chain may include a limited number of modifications. For example, the Fc polypeptide chain may include one or more heterodimerization modifications, one or more modifications that inhibit or enhance binding to FcγR, or one or more modifications that increase binding to FcRn.

[0073] In a particular embodiment, the Fc used for half-life extension is a single-chain Fc (“scFc”).

[0074] In some embodiments, the amino acid sequence of the Fc polypeptide can be a mammalian (e.g., human) amino acid sequence. Isotypes of the Fc polypeptide can be IgG (such as IgG1, IgG2, IgG3, or IgG4), IgA, IgD, IgE, or IgM. Table 2 below shows the amino acid sequence alignments of human IgG1, IgG2, IgG3, and IgG4 Fc polypeptide chains.

[0075] Sequences of human IgG1, IgG2, IgG3, and IgG4 Fc peptides that can be used are provided in SEQ ID NO: 36-39. Variants of these sequences containing one or more heterodimerization alterations, one or more Fc alterations that prolong the half-life, one or more alterations that enhance ADCC, and / or one or more alterations that inhibit Fcγ receptor (FcγR) binding are also considered, as are other close variants containing no more than 10 single amino acid deletions, insertions, or substitutions per 100 amino acids.

[0076] The numbering shown in Table 2 is based on the EU numbering system, which is based on the sequential numbering of the constant regions of the IgG1 antibody. Edelman et al. (1969), Proc. Natl. Acad. Sci. [Proceedings of the National Academy of Sciences] 63: 78-85. Therefore, it cannot accommodate the additional length of the IgG3 hinge pore. Nevertheless, it is still used here to specify the position in the Fc region, as it is still commonly used in the art to refer to the position in the Fc region. The hinge regions of the IgG1, IgG2, and IgG4 Fc peptides extend from about position 216 to about position 230. It is clear from the comparison that the IgG2 and IgG4 hinge regions are each three amino acids shorter than the IgG1 hinge. The IgG3 hinge is longer, extending upstream by an additional 47 amino acids. The CH2 region extends from about position 231 to position 340, and the CH3 region extends from about position 341 to position 447.

[0077] The naturally occurring amino acid sequence of the Fc polypeptide can vary slightly. Such variations can include single amino acid insertions, deletions, or substitutions of no more than 10 out of every 100 amino acids in the sequence of the naturally occurring Fc polypeptide chain. If substitutions are present, these substitutions can be conserved amino acid substitutions as defined above. The amino acid sequences of the Fc polypeptide on the first and second polypeptide chains can differ. In some embodiments, they can include “heterodimerization alterations” that promote heterodimer formation, such as charge-pair substitutions as defined above. Furthermore, the Fc polypeptide moiety of PABP may also contain alterations that inhibit or enhance FcγR binding. Such mutations are described above and in Xu et al. (2000), Cell Immunol. [Cell Immunology] 200 (1): 16-26 (the relevant portions of which are incorporated herein by reference). The Fc peptide moiety may also include "Fc alterations that prolong half-life" as described above, including those described in the following literature: for example, U.S. Patents 7,037,784, 7,670,600, and 7,371,827, U.S. Patent Application Publication 2010 / 0234575, and International Application PCT / US 2012 / 070146, all of which are incorporated herein by reference in their relevant portions. Additionally, the Fc peptide may include "alterations that enhance ADCC" as defined above.

[0078] Another suitable Fc polypeptide, described in PCT application WO 93 / 10151 (incorporated hereby by reference), is a single-chain polypeptide extending the N-terminal hinge region of the Fc region of a human IgG1 antibody to the native C-terminus. Another useful Fc polypeptide is an Fc mutant protein, described in U.S. Patent 5,457,035 and Baum et al., 1994, EMBO J. [Journal of the European Society for Molecular Biology] 13: 3992-4001. This mutant protein has the same amino acid sequence as the native Fc sequence presented in WO 93 / 10151, except that amino acid 19 has been changed from Leu to Ala, amino acid 20 from Leu to Glu, and amino acid 22 from Gly to Ala. The mutant protein exhibits reduced affinity for the Fc receptor.

[0079] By introducing one or more mutations into the Fc, the effector function of an antibody can be increased or decreased. Embodiments of the present invention include IL-2 mutant Fc fusion proteins with an engineered Fc to increase effector function (US 7,317,091 and Strohl, Curr. Opin. Biotech. [Biotechnology Review Now], 20: 685-691, 2009; both are incorporated herein by reference in their entirety). For certain therapeutic indications, it may be desirable to increase effector function. For other therapeutic indications, it may be desirable to decrease effector function.

[0080] Exemplary IgG1 Fc molecules with enhanced effector functionality include those with the following substitutions: Another approach to increasing the effector function of proteins containing IgG Fc is by reducing Fc fucosylation. Removing the core fucosylate from the biantennary complex oligosaccharide attached to the Fc increases ADCC effector function without altering antigen-binding or CDC effector function. Several methods are known to reduce or eliminate fucosylation of Fc-containing molecules (e.g., antibodies). These include recombinant expression in certain mammalian cell lines, including FUT8 knockout cell lines, the variant CHO cell line Lec13, the rat hybridoma cell line YB2 / 0, cell lines containing small interfering RNA specifically targeting the FUT8 gene, and cell lines co-expressing α-1,4- N - Cell lines containing acetylglucosamine transferase III and Golgi α-mannosidase II. Alternatively, Fc-containing molecules can be expressed in non-mammalian cells (e.g., plant cells, yeast, or prokaryotic cells, such as Escherichia coli).

[0081] In some embodiments of the invention, the bispecific binding construct comprises an Fc engineered to reduce effector function. Exemplary Fc molecules with reduced effector function include those having the following substitutions: Human IgG1 is known to have a glycosylation site at N297 (EU numbering system), and glycosylation contributes to the effector function of IgG1 antibodies. An exemplary IgG1 sequence is provided in SEQ ID NO: 36. N297 can be mutated to form glycosylated-free antibodies. For example, the mutation can be performed by replacing N297 with an amino acid with similar physiological and chemical properties to asparagine (such as glutamine (N297Q)) or with alanine (N297A) (which mimics asparagine without a polar group).

[0082] In some embodiments, mutating amino acid N297 of human IgG1 to glycine, i.e., N297G, provides significantly superior purification efficiency and biophysical properties compared to other amino acid substitutions at this residue. See, for example, U.S. Patent Nos. 9,546,203 and 10,093,711. In a particular embodiment, the bispecific binding construct of the present invention comprises human IgG1 Fc with the N297G substitution.

[0083] The bispecific binding construct of the present invention comprising human IgG1 Fc with the N297G mutation may further comprise additional insertions, deletions, and substitutions. In some embodiments, human IgG1 Fc comprises an N297G substitution and has at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% identity with the amino acid sequence listed in SEQ ID NO: 36. In a particularly preferred embodiment, a C-terminal lysine residue is substituted or deleted.

[0084] In some cases, molecules containing glycosylated IgG1 Fc may be less stable than those containing glycosylated IgG1 Fc. Therefore, the Fc region can be further engineered to increase the stability of the glycosylated molecule. In some embodiments, one or more amino acids are substituted with cysteine ​​to form disulfide bonds in a dimer state. In specific embodiments, cysteine ​​may be used to replace residues V259, A287, R292, V302, L306, V323, or I332 of the amino acid sequence listed in SEQ ID NO: 36. In other embodiments, specific residue pairs are substituted such that they preferentially form disulfide bonds with each other, thus limiting or preventing disulfide bond disorder. In specific embodiments, pairs include, but are not limited to, A287C and L306C, V259C and L306C, R292C and V302C, and V323C and I332C.

[0085] As discussed above in the linker section, in some embodiments, the bispecific binding construct of the present invention includes a linker between the Fc and HHLL bispecific binding constructs (specifically, linking Fc to VL2). In some embodiments, one or more copies of the peptide consist of GGGGS (SEQ ID NO: 1), GGNGT (SEQ ID NO: 15), or YGNGT (SEQ ID NO: 16) between the Fc and HHLL peptides. In some embodiments, the peptide region between the Fc region and the HHLL peptide comprises a single copy of GGGGS (SEQ ID NO: 1), GGNGT (SEQ ID NO: 15), or YGNGT (SEQ ID NO: 16). In some embodiments, the linker GGNGT (SEQ ID NO: 15) or YGNGT (SEQ ID NO: 16) is glycosylated when expressed in appropriate cells, and this glycosylation helps stabilize the protein in solution and / or during in vivo administration. Therefore, in some embodiments, the bispecific binding construct of the present invention includes a glycosylated linker between the Fc region and the HHLL peptide.

[0086] Nucleic acid encoding bispecific binding construct In another embodiment, the present invention provides isolated nucleic acid molecules encoding the bispecific binding constructs of the present invention. Additionally, vectors containing these nucleic acids, cells containing these nucleic acids, and methods for preparing the binding constructs of the present invention are provided. These nucleic acids contain, for example, polynucleotides encoding all or part of the bispecific binding construct (e.g., fragments, derivatives, mutant proteins, or variants thereof), polynucleotides sufficient to serve as hybridization probes, PCR primers or sequencing primers for identifying, analyzing, mutating, or amplifying polynucleotides encoding polypeptides, antisense nucleic acids for inhibiting polynucleotide expression, and the aforementioned complementary sequences. Where appropriate, these nucleic acids can be of any length for the desired use or function and may contain one or more additional sequences, such as regulatory sequences, and / or as part of a longer nucleic acid, such as a vector. The nucleic acids can be single-stranded or double-stranded and may contain RNA and / or DNA nucleotides and their artificial variants (e.g., peptide nucleic acids).

[0087] Nucleic acids encoding polypeptides (e.g., heavy or light chains, variable domains only, or full-length) can be isolated from B cells of mice immunized with antigens. Nucleic acids can be isolated using routine procedures such as polymerase chain reaction (PCR).

[0088] Nucleic acid sequences encoding the variable regions of the heavy and light chains are included herein. Those skilled in the art will understand that, due to the degeneracy of the genetic code, each polypeptide sequence disclosed herein is encoded by a large number of other nucleic acid sequences. This invention provides each degenerate nucleotide sequence encoding each binding construct of this invention.

[0089] This invention further provides nucleic acids that can hybridize with other nucleic acids under specific hybridization conditions. Methods for hybridizing nucleic acids are well known in the art. See, for example, Current Protocols in Molecular Biology, John Wiley & Sons, New York (1989), 6.3.1–6.3.6. As defined herein, for example, moderately stringent hybridization conditions use a prewash solution containing 5X sodium chloride / sodium citrate (SSC), 0.5% SDS, and 1.0 mM EDTA (pH 8.0), a hybridization buffer containing approximately 50% formamide, 6X SSC, and a hybridization temperature of 55°C (or other similar hybridization solutions, such as a hybridization solution containing approximately 50% formamide and a hybridization temperature of 42°C), and wash conditions at 60°C in 0.5X SSC and 0.1% SDS. Strict hybridization conditions involve hybridization at 45°C in 6X SSC, followed by washing once or multiple times at 68°C in 0.1X SSC and 0.2% SDS. Additionally, those skilled in the art can manipulate these hybridization and / or washing conditions to increase or decrease the stringency of hybridization, such that nucleic acids containing nucleotide sequences having at least 65%, 70%, 75%, 80%, 85%, 90%, 95%, 98%, or 99% identity with each other typically maintain hybridization with one another. The fundamental parameters influencing the selection of hybridization conditions and guidelines for designing suitable conditions are set forth in, for example, the following: Sambrook, Fritsch, and Maniatis (1989, *Molecular Cloning: A Laboratory Manual*, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, Chapters 9 and 11; and *Current Protocols in Molecular Biology*, 1995, edited by Ausubel et al., John Wiley & Sons, Inc., Sections 2.10 and 6.3-6.4), and can be readily determined by those skilled in the art based, for example, the length and / or base composition of DNA. Changes can be introduced into nucleic acids by mutation, thereby causing a change in the amino acid sequence of the polypeptide (e.g., a binding construct) it encodes. Any technique known in the art can be used to introduce mutations. In one embodiment, a site-directed mutagenesis protocol is used to alter one or more specific amino acid residues. In another embodiment, a random mutagenesis protocol is used to alter one or more randomly selected residues. Regardless of the modifications made, mutant peptides can be expressed and screened for the desired properties.

[0090] Mutations can be introduced into nucleic acids without significantly altering the biological activity of the encoded polypeptide. For example, nucleotide substitutions can be performed, thereby substituting amino acid residues at non-essential amino acid residues. In one embodiment, the nucleotide sequence of the binding construct provided herein, or a desired fragment, variant, or derivative thereof, is mutated to encode an amino acid sequence comprising one or more deleted or substituted amino acid residues, which, for the light chain or heavy chain of the binding construct of the present invention, are shown herein as residues at two or more sequence differences. In another embodiment, mutagenesis inserts an amino acid adjacent to one or more amino acid residues in the light chain or heavy chain of the binding construct of the present invention, as two or more sequence differences. Alternatively, one or more mutations can be introduced into the nucleic acid to selectively alter the biological activity of the encoded polypeptide.

[0091] In another embodiment, the present invention provides a vector comprising a nucleic acid encoding a polypeptide or a portion thereof of the present invention. Examples of vectors include, but are not limited to, plasmids, viral vectors, non-free mammalian vectors, and expression vectors (e.g., recombinant expression vectors).

[0092] The recombinant expression vector of the present invention may contain the nucleic acid of the present invention in a form suitable for expression in host cells. The recombinant expression vector includes one or more regulatory sequences selected based on the host cell to be used for expression, operatively linked to the nucleic acid sequence to be expressed. The regulatory sequences include those that direct constitutive expression of the nucleotide sequence in many types of host cells (e.g., SV40 early gene enhancer, Laureth sarcoma virus promoter, and cytomegalovirus promoter), those that direct expression of the nucleotide sequence only in certain host cells (e.g., tissue-specific regulatory sequences, see Voss et al., 1986, Trends Biochem. Sci. 11: 287; Maniatis et al., 1987, Science 236: 1237, which are incorporated herein by reference in their entirety), and those that direct inducible expression of the nucleotide sequence in response to a specific treatment or condition (e.g., metallothionein promoters in mammalian cells and tetracycline-reactive and / or streptomycin-reactive promoters in prokaryotic and eukaryotic systems (see ibid.)). Those skilled in the art will understand that the design of the expression vector can depend on factors such as the choice of host cells to be transformed and the desired expression level of the protein. The expression vectors of the present invention can be introduced into host cells, thereby producing proteins or peptides encoded by nucleic acids as described herein, including fusion proteins or peptides.

[0093] In another embodiment, the present invention provides a host cell into which the recombinant expression vector of the present invention has been incorporated. The host cell can be any prokaryotic or eukaryotic cell. Prokaryotic host cells include Gram-negative or Gram-positive organisms, such as *Escherichia coli* or bacilli. Higher eukaryotic cells include insect cells, yeast cells, and established cell lines of mammalian origin. Examples of suitable mammalian host cell lines include Chinese hamster ovary (CHO) cells or derivatives thereof (such as Veggie CHO) and related cell lines grown in serum-free medium (see Rasmussen et al., 1998, *Cytotechnology* 28: 31) or the DHFR-deficient CHO strain DXB-11 (see Urlaub et al., 1980, *Proc. Natl. Acad. Sci. USA* 77: 4216-20). Other CHO cell lines include CHO-K1 (ATCC#CCL-61), EM9 (ATCC#CRL-1861), and UV20 (ATCC#CRL-1862). Other host cell lines include the COS-7 lineage of monkey kidney cells (ATCC CRL 1651) (see Gluzman et al., 1981, Cell [Cell] 23: 175), L cells, C127 cells, 3T3 cells (ATCCCCCL 163), AM-1 / D cells (described in US Patent No. 6,210,924), HeLa cells, the BHK (ATCC CRL 10) cell line, the CV1 / EBNA cell line derived from the African green monkey kidney cell line CV1 (ATCC CCL 70) (see McMahan et al., 1991, EMBO J. [Journal of the European Society for Molecular Biology] 10: 2821), and human embryonic kidney cells (such as 293, 293 EBNA, or MSR). 293), human epidermal A431 cells, human Colo205 cells, other transformed primate cell lines, normal diploid cells, in vitro cultured cell lines derived from primary tissues and primary explants, HL-60, U937, HaK, or Jurkat cells. Appropriate cloning and expression vectors for use with bacterial, fungal, yeast, and mammalian cell hosts are described in: Pouwels et al. (Cloning Vectors: A Laboratory Manual, Elsevier, New York, 1985).

[0094] Vector DNA can be introduced into prokaryotic or eukaryotic cells via conventional transformation or transfection techniques. For stable transfection of mammalian cells, it is known that only a small fraction of cells will integrate the foreign DNA into their genome, depending on the expression vector and transfection technique used. To identify and select these integrators, genes encoding selectable markers (e.g., antibiotic resistance) are typically introduced into the host cell along with the target gene. Other selectable markers include those conferring resistance to drugs such as G418, hygromycin, and methotrexate. Among other methods, cells stably transfected with the introduced nucleic acid can be identified by drug selection (e.g., cells with incorporated selectable marker genes will survive, while other cells will die).

[0095] Transformed cells can be cultured under conditions that promote peptide expression, and the peptides can be recovered using standard protein purification procedures. Peptides considered for use in this paper include substantially homogeneous recombinant mammalian peptides that are largely free of contaminating endogenous substances.

[0096] Cells containing nucleic acids encoding the bispecific binding construct of this invention also include hybridomas. The generation and culture of hybridomas are discussed herein.

[0097] In some embodiments, a vector comprising nucleic acid molecules as described herein is provided. In some embodiments, the invention comprises a host cell containing nucleic acid molecules as described herein.

[0098] In some embodiments, a nucleic acid molecule encoding a bispecific binding construct as described herein is provided.

[0099] In some embodiments, a pharmaceutical composition comprising at least one bispecific binding construct described herein is provided.

[0100] Production methods The binding constructs of the present invention can be generated by any method known in the art for synthesizing proteins (e.g., antibodies), particularly by chemical synthesis or preferably by recombinant expression techniques.

[0101] Recombinant expression of the bispecific binding construct requires the construction of an expression vector containing a polynucleotide encoding the bispecific binding construct. Once the polynucleotide encoding the bispecific binding construct is obtained, a vector for generating the bispecific binding construct can be produced using recombinant DNA technology. Expression vectors containing the coding sequence of the bispecific binding construct and appropriate transcription and translation control signals are constructed. These methods include, for example, in vitro recombinant DNA technology, synthetic techniques, and in vivo gene recombination.

[0102] The expression vector is transferred to host cells using conventional techniques, and the transfected cells are then cultured using conventional techniques to produce the bispecific binding construct of the present invention.

[0103] The bispecific binding constructs of the present invention can be expressed using a variety of host expression vector systems. Such host expression systems represent both the medium (through which the target coding sequence is generated and subsequently purified) and the cell that can express the molecules of the present invention in situ upon transformation or transfection with appropriate nucleotide coding sequences. Bacterial cells (such as Escherichia coli) and eukaryotic cells are commonly used to express recombinant binding molecules, especially for expressing whole recombinant binding molecules. For example, mammalian cells (such as Chinese hamster ovary cells (CHO)) combined with vectors (such as promoter elements of major metaphase-early genes from human cytomegalovirus) are efficient antibody expression systems (Foecking et al., Gene 45: 101 (1986); Cockett et al., Bio / Technology 8: 2 (1990)).

[0104] Additionally, host cell lines can be selectively regulated to modulate the expression of the inserted sequence or to modify and process the gene product in a desired specific manner. Such modifications (e.g., glycosylation) and processing (e.g., cleavage) of protein products can be important for protein function. Different host cells have characteristic and specific mechanisms for post-translational processing and modification of proteins and gene products. Appropriate cell lines or host systems can be selected to ensure proper modification and processing of the expressed exogenous protein. For this purpose, eukaryotic host cells with appropriate cellular machinery for the processing of primary transcripts, glycosylation, and phosphorylation of gene products can be used. Such mammalian host cells include, but are not limited to, CHO, COS, 293, 3T3, or myeloma cells.

[0105] Stable expression is preferred for long-term, high-yield production of recombinant proteins. For example, cell lines that stably express binding molecules can be engineered. Instead of using expression vectors containing viral origins of replication, host cells can be transformed with DNA controlled by appropriate expression control elements (e.g., promoters, enhancers, sequences, transcription terminators, polyadenylation sites, etc.) and selectable markers. After introducing exogenous DNA, engineered cells can be allowed to grow in enriched media for 1–2 days and then switched to selective media. Selective markers in the recombinant plasmid confer resistance to selection and allow cells to stably integrate the plasmid into their chromosomes and grow to form foci, which can then be cloned and expanded into cell lines. This method can be advantageously used to engineer cell lines expressing binding molecules. Such engineered cell lines can be particularly used for screening and evaluating compounds that interact directly or indirectly with binding molecules.

[0106] Many selection systems can be used, including but not limited to the herpes simplex virus thymidine kinase (Wigler et al., Cell 11: 223 (1977)), hypoxanthine-guanine phosphoribosyltransferase (Szybalska and Szybalski, Proc. Natl. Acad. Sci. USA 48: 202 (1992)), and adenine phosphoribosyltransferase (Lowy et al., Cell 22: 817 (1980)) genes, which can be used in tk-, hgprt-, or aprt- cells, respectively. Additionally, antimetabolite resistance can be used as a basis for selecting the following genes: dhfr, which confers resistance to methotrexate (Wigler et al., Proc. Natl. Acad. Sci. USA [Proceedings of the National Academy of Sciences] 77: 357 (1980); O'Hare et al., Proc. Natl. Acad. Sci. USA [Proceedings of the National Academy of Sciences] 78: 1527 (1981)); gpt, which confers resistance to mycophenolic acid (Mulligan and Berg, Proc. Natl. Acad. Sci. USA [Proceedings of the National Academy of Sciences] 78: 2072 (1981)); neo, which confers resistance to aminoglycoside G-418 (Wu and Wu, Biotherapy 3: 87-95 (1991)); and hygro, which confers resistance to hygromycin (Santerre et al., Gene 30: 147 (1984)).Methods of recombinant DNA techniques known in the art can be routinely applied to the selection of desired recombinant clones, and such methods are described, for example, Ausubel et al. (eds.), Current Protocols in Molecular Biology, John Wiley & Sons, New York (1993); Kriegler, Gene Transfer and Expression, A Laboratory Manual, Stockton Press, New York (1990); and Chapters 12 and 13, Dracopoli et al. (eds.), Current Protocols in Human Genetics, John Wiley & Sons, New York (1994); Colberre-Garapin et al., J. Mol. Biol. 150: 1 (1981), which are incorporated herein by reference in their entirety.

[0107] The expression level of binding molecules can be increased by vector amplification (for a review, see Bebbington and Henschel, “The use of vectors based on gene amplification for the expression of cloned genes in mammalian cells” (DNA Cloning, Vol. 3. Academic Press, New York, 1987). When the marker in the expression-binding vector system is amplifiable, the increased level of inhibitors present in the host cell culture will increase the copy number of the marker gene. Since the amplified region is associated with the gene, protein production will also increase (Crouse et al., Mol. Cell. Biol. 3: 257(1983)).

[0108] Host cells can be co-transfected with multiple expression vectors of the present invention. The vectors can contain the same selectable markers, which allows for equivalent expression of the expressed peptides. Alternatively, a single vector can be used that encodes and is capable of expressing, for example, the peptides of the present invention. The coding sequence can comprise cDNA or genomic DNA.

[0109] Once the binding molecule of the present invention has been generated through animal, chemical synthesis, or recombinant expression, it can be purified using any method known in the art for purifying immunoglobulin molecules, such as chromatography (e.g., ion exchange, affinity chromatography (particularly by the affinity of a specific antigen for protein A), and size exclusion chromatography), centrifugation, differential solubility, or any other standard technique for purifying proteins. Furthermore, the binding construct of the present invention or fragments thereof may be fused to heterologous polypeptide sequences described herein or other known in the art to facilitate purification. Purification techniques can vary depending on whether the Fc region (e.g., scFC) is attached to the bispecific binding construct of the present invention.

[0110] In some embodiments, the present invention covers binding constructs of recombinant fusions or chemical conjugates (including covalent and non-covalent conjugates) to peptides. The fusion or conjugate binding constructs of the present invention can be used for convenient purification. See, for example, Harbor et al., ibid., and PCT disclosure WO 93 / 21232; EP 439,095; Naramura et al., Immunol. Lett. [Immunology Letters] 39: 91-99 (1994); U.S. Patent No. 5,474,981; Gillies et al., Proc. Natl. Acad. Sci. [Proceedings of the National Academy of Sciences] 89: 1428-1432 (1992); Fell et al., J. Immunol. [Journal of Immunology] 146: 2446-2452 (1991).

[0111] Furthermore, the binding constructs or fragments thereof of the present invention can be fused with marker sequences such as peptides to facilitate purification. In a preferred embodiment, the marker amino acid sequence is a hexahistine peptide (SEQ ID NO: 58), such as the tag provided in the pQE vector (QIAGEN, Inc., 9259 Eton Avenue, Chatsworth, CA 91311), many of which are commercially available. As described in Gentz ​​et al., Proc. Natl. Acad. Sci. USA [Proceedings of the National Academy of Sciences] 86: 821-824 (1989), hexahistine (SEQ ID NO: 58) provides convenient purification of the fusion protein. Other peptide tags that can be used for purification include, but are not limited to, “HA” tags and “flag” tags corresponding to epitopes derived from influenza hemagglutinin proteins (Wilson et al., Cell [Cell] 37: 767 (1984)).

[0112] Generation of bispecific binding constructs In a general sense, the bispecific binding construct of the present invention is constructed by selecting VH and VL regions from a desired antibody and linking them together using a peptide linker as described herein to form an HHLL bispecific binding construct, optionally with an Fc region attached. More specifically, nucleic acids encoding VH, VL, and the linker, and optionally Fc, are combined to produce an HHLL nucleic acid construct encoding the bispecific binding construct of the present invention.

[0113] Antibody production In some embodiments, a monospecific antibody with binding specificity to a desired target is first generated before generating the bispecific binding construct of the present invention.

[0114] Antibodies that can be used to generate the bispecific binding constructs of the present invention can be prepared by techniques well known to those skilled in the art. For example, they can be prepared by immunizing animals (such as mice, rats, or rabbits) and then by immortalizing spleen cells collected from the animals after the completion of the immunization protocol. Spleen cells can be immortalized using any technique known in the art, such as by fusing spleen cells with myeloma cells to generate hybridomas. See, for example, Antibodies; Harlow and Lane, Cold Spring Harbor Laboratory Press, 1st edition, e.g., from 1988, or 2nd edition, e.g., from 2014).

[0115] In one embodiment, the humanized monoclonal antibody comprises a variable domain (or all or part of its antigen-binding site) of a mouse antibody and a constant domain derived from a human antibody. Alternatively, the humanized antibody fragment may comprise an antigen-binding site of a mouse monoclonal antibody and a variable domain fragment derived from a human antibody (lacking the antigen-binding site). Procedures for producing engineered monoclonal antibodies include those described in Riechmann et al., 1988, Nature 332: 323, Liu et al., 1987, Proc. Nat. Acad. Sci. USA 84: 3439, Larrick et al., 1989, Bio / Technology 7: 934, and Winter et al., 1993, TIPS 14: 139. In one embodiment, the chimeric antibody is a CDR-grafted antibody. Technical discussions for humanized antibodies are found, for example, in U.S. Patent Nos. 5,869,619; 5,225,539; 5,821,337; 5,859,205; 6,881,557; Padlan et al., 1995, FASEB J. [Journal of the Federation of American Societies for Experimental Biology] 9: 133-39; Tamura et al., 2000, J. Immunol. [Journal of Immunology] 164: 1432-41; Zhang, W. et al., Molecular Immunology. [Molecular Immunology] 42 (12): 1445-1451, 2005; Hwang W. et al., Methods. [Methods] 36 (1): 35-42, 2005; Dall'Acqua WF et al., Methods [Methods] 36 (1): 43-60, 2005; and Clark, M., Immunology Today. 21 (8): 397-402, 2000.

[0116] The binding molecules of the present invention may also contain regions of fully human monoclonal antibodies. Fully human monoclonal antibodies can be generated by any number of techniques familiar to those skilled in the art. Such methods include, but are not limited to, Epstein-Barr virus (EBV) transformation of human peripheral blood cells (e.g., containing B lymphocytes), in vitro immunization with human B cells, fusion of spleen cells from immunized transgenic mice carrying inserted human immunoglobulin genes, isolation from a phage library of the human immunoglobulin V region, or other procedures known in the art and based on the disclosure herein.

[0117] Procedures for generating human monoclonal antibodies in non-human animals have been developed. For example, mice inactivated by various means have been prepared using one or more endogenous immunoglobulin genes. Human immunoglobulin genes have been introduced into the mice to replace the inactivated mouse genes. Through this technique, elements of human heavy and light chain loci are introduced into mouse strains derived from embryonic stem cell lines containing targeted disruption of endogenous heavy and light chain loci (see also Bruggemann et al., Curr. Opin. Biotechnol. [Current Views on Biotechnology] 8: 455-58 (1997)). For example, human immunoglobulin transgenes can be microgene constructs or trans loci on yeast artificial chromosomes that undergo B cell-specific DNA rearrangements and hypermutations in mouse lymphoid tissues.

[0118] The antibodies produced in the animals incorporate human immunoglobulin polypeptide chains encoded by human genetic material introduced into the animals. In one embodiment, a non-human animal (such as a transgenic mouse) is immunized with a suitable immunogen.

[0119] Examples of techniques for producing and using transgenic animals to generate human or partially human antibodies are described in: U.S. Patents 5,814,318, 5,569,825, and 5,545,806; Davis et al., Production of human antibodies from transgenic mice, Lo; Antibody Engineering: Methods and Protocols, Humana Press, NJ: 191-200 (2003); Kellermann et al., 2002, Curr Opin Biotechnol. 13: 593-97; Russell et al., 2000, Infect Immun. 68: 1820-26; Gallo et al., 2000, Eur J Immun. 30: 534-40; Davis et al., 1999, Cancer Metastasis Rev. [Cancer and Metastasis Review] 18: 421-25, Green, 1999; J Immunol Methods. [Journal of Immunological Methods] 231: 11-23, Jakobovits, 1998; Advanced Drug Delivery Reviews 31: 33-42, Green et al., 1998; J Exp Med. [Journal of Experimental Medicine] 188: 483-95, Jakobovits A, 1998; Exp. Opin. Invest. Drugs. [Expert Reviews on Drugs] 7: 607-14, Tsuda et al., 1997; Genomics. [Genomics] 42: 413-21, Mendez et al., 1997; Nat Genet. [Nature Genetics] 15: 146-56, Jakobovits, 1994; Curr Biol. [Current Biology] 4: 761-63, Arbones et al., 1994, Immunity. [Immunity] 1: 247-60, Green et al., 1994, Nat Genet. [Nature Genetics] 7: 13-21, Jakobovits et al., 1993, Nature. [Nature] 362: 255-58, Jakobovits et al., 1993, Proc Natl Acad Sci USA.[Proceedings of the National Academy of Sciences of the United States of America] 90: 2551-55. Chen, J., M. Trounstine, FW Alt, F. Young, C. Kurahara, J.Loring, D. Huszar. "Immunoglobulin gene rearrangement in B-cell deficient mice generated by targeted deletion of the JH locus."[Immunoglobulin gene rearrangement in B-cell-deficient mice generated by targeting the deletion of the JH locus] International Immunology 5 (1993): 647-656, Choi et al., 1993; Nature Genetics 4: 117-23, Fishwild et al., 1996; Nature Biotechnology 14: 845-51, Harding et al., 1995; Annals of the New York Academy of Sciences, Lonberg et al., 1994; Nature 368: 856-59, Lonberg, 1994; Transgenic Approaches to Human Monoclonal Antibodies in Handbook of Experimental Pharmacology 113: 49-101, Lonberg et al., 1995; Internal Review of Immunology [Internal Review of Immunology] 13: 65-93, Neuberger, 1996, Nature Biotechnology; [Nature Biotechnology] 14: 826, Taylor et al., 1992, Nucleic Acids Research; [Nucleic Acids Research] 20: 6287-95, Taylor et al., 1994, International Immunology; [International Immunology] 6: 579-91, Tomizuka et al., 1997, Nature Genetics; [Nature Genetics] 16: 133-43, Tomizuka et al., 2000, Proceedings of the National Academy of Sciences USA; [Proceedings of the National Academy of Sciences USA] 97: 722-27, Tuaillon et al., 1993, Proceedings of the National Academy of Sciences USA; [Proceedings of the National Academy of Sciences USA] 90: 3720-24, and Tuaillon et al., 1994, Journal of Immunology. [Journal of Immunology] 152: 2912-20.Lonberg et al., Nature 368: 856, 1994; Taylor et al., Int. Immun. 6: 579, 1994; US Patent No. 5,877,397; Bruggemann et al., 1997 Curr. Opin. Biotechnol. 8: 455-58; Jakobovits et al., 1995 Ann. NY Acad. Sci. 764: 525-35. Furthermore, embodiments relating to XenoMouse® (Abgenix, now Amgen, Inc.) are described in, for example, US 05 / 0118643 and WO 05 / 694879, WO 98 / 24838, WO 00 / 76310, and US Patent 7,064,244. .

[0120] For example, lymphocytes from immunized transgenic mice are fused with myeloma cells to generate hybridomas. The myeloma cells used in the fusion procedure for generating hybridomas are preferably non-antibody-generated, exhibiting high fusion efficiency and enzyme deficiencies (preventing them from growing in certain selective media that only support the growth of the desired fused cells (hybridomas)). Examples of cell lines suitable for such fusions include Sp-20, P3-X63 / Ag8, P3-X63-Ag8.653, NS1 / 1.Ag4 1, Sp210-Ag14, FO, NSO / U, MPC-11, MPC11-X45-GTG 1.7, and S194 / 5XX0 Bul; examples of cell lines used for rat fusions include R210.RCY3, Y3-Ag 1.2.3, IR983F, and 4B210. Other cell lines that can be used for cell fusion are U-266, GM1500-GRG2, LICR-LON-HMy2, and UC729-6.

[0121] Lymphoma cells (e.g., spleen) and myeloma cells can be combined with a membrane fusion promoter (e.g., polyethylene glycol or a nonionic detergent) for several minutes, and then plated at low density on a selective medium that supports hybridoma cell growth but not the growth of unfused myeloma cells. One selective medium is HAT (hypoxanthine, aminopterin, thymidine). After a sufficient period of time (typically about one to two weeks), cell colonies are observed. Single colonies are isolated, and the binding activity of the cell-produced antibodies to the desired target can be tested using any of the various immunoassays known in the art and described herein. The hybridomas are cloned (e.g., by limiting dilution clones or by soft agar plaque isolation), and positive clones that produce molecules specific to the desired target are selected and cultured. The binding molecules from the hybridoma culture can be isolated from the supernatant of the hybridoma culture. Thus, the present invention provides hybridomas containing polynucleotides encoding the binding constructs of the present invention on the chromosomes of cells. These hybridomas can be cultured according to the methods described herein and known in the art.

[0122] Another method for generating human antibodies that can be used to generate the bispecific binding molecules of the present invention includes immortalizing human peripheral blood cells via EBV transformation. See, for example, U.S. Patent No. 4,464,456. Such immortalized B cell lines (or lymphoblastoid cell lines) that generate monoclonal antibodies that specifically bind to the desired target can be identified by immunoassay methods (e.g., ELISA) as provided herein and then isolated by standard cloning techniques. The stability of the antibody-generating lymphoblastoid cell lines can be improved according to methods known in the art by fusing the transformed cell lines with mouse myeloma to generate mouse-human hybrid cell lines (see, for example, Glasky et al., Hybridoma 8: 377-89 (1989)). Yet another method for generating human monoclonal antibodies is in vitro immunization, which involves inducing human spleen B cells with an antigen and then fusing the induced B cells with a heterologous hybrid fusion partner. See, for example, Boerner et al., 1991 J. Immunol. 147: 86-95.

[0123] In some embodiments, B cells that produce the desired antibody are selected, and the light and heavy chain variable regions are cloned from the B cells according to techniques known in the art (WO 92 / 02551; U.S. Patent 5,627,052; Babcook et al., Proc. Natl. Acad. Sci. USA [Proceedings of the National Academy of Sciences] 93: 7843-48 (1996)) and the molecular biology techniques described herein. B cells from immunized animals can be isolated from spleen, lymph node, or peripheral blood samples by selecting cells that produce the desired antibody. B cells can also be isolated from humans (e.g., from peripheral blood samples). Methods for detecting individual B cells that produce antibodies with the desired specificity (e.g., by plaque formation, fluorescence-activated cell sorting, in vitro stimulation followed by detection of specific antibodies, etc.) are well known in the art. Methods for selecting B cells that produce specific antibodies include, for example, preparing a single-cell suspension of B cells in soft agar containing the antigen. The binding of specific antibodies produced by B cells to antigens results in the formation of complexes, which can be visible as immunoprecipitates. According to the methods known in the art and described herein, after selecting B cells that produce the desired antibodies, specific antibody genes can be cloned by isolating and amplifying DNA or mRNA, and these specific antibody genes are used to generate the bispecific binding constructs of the present invention.

[0124] Another method for obtaining antibodies that can be used to generate the bispecific binding constructs of the present invention is via phage display. See, for example, Winter et al., 1994 Annu. Rev. Immunol. [Annual Review of Immunology] 12:433-55; Burton et al., 1994 Adv. Immunol. [Advances in Immunology] 57: 191-280. Human or mouse immunoglobulin variable region gene combinatorial libraries can be generated in phage vectors, which can be screened to select Ig fragments (Fab, Fv, sFv, or multimers thereof) that specifically bind to TGF-β binding proteins or their variants or fragments. See, for example, U.S. Patent No. 5,223,409; Huse et al., 1989 Science 246: 1275-81; Sastry et al., Proc. Natl. Acad. Sci. USA 86: 5728-32 (1989); Alting-Mees et al., Strategies in Molecular Biology 3: 1-9 (1990); Kang et al., 1991 Proc. Natl. Acad. Sci. USA 88: 4363-66; Hoogenboom et al., 1992 J. Molec. Biol. 227: 381-388; Schlebusch et al., 1997 Hybridoma 16: 47-52 and the references cited therein. For example, a library containing multiple polynucleotide sequences encoding an Ig variable region fragment can be co-framed with a sequence encoding a phage coat protein into the genome of a filamentous phage (e.g., M13 or a variant thereof). The fusion protein can be a fusion of the coat protein with a light chain variable region domain and / or with a heavy chain variable region domain. According to some embodiments, immunoglobulin Fab fragments may also be displayed on the phage particle (see, for example, U.S. Patent No. 5,698,426).

[0125] Heavy and light chain immunoglobulin cDNA expression libraries can also be prepared in λ phage, for example, using λimmunoZap™(H) and λImmunoZap™(L) vectors (Stratagene, La Jolla, CA). In short, mRNA is isolated from a population of B cells and used to generate heavy and light chain immunoglobulin cDNA expression libraries in λImmunoZap(H) and λImmunoZap(L) vectors. These vectors can be screened individually or co-expressed to form Fab fragments or antibodies (see Huse et al., ibid.; also see Sastry et al., ibid.). Positive plaques can then be converted into non-lytic plasmids that allow for high-level expression of monoclonal antibody fragments from *E. coli*.

[0126] In one embodiment, in a hybridoma, variable regions of the gene expressing the target monoclonal antibody are amplified using nucleotide primers. These primers can be synthesized by those skilled in the art or purchased from commercially available sources. (See, for example, Stratagene (La Jolla, California), which sells primers for mouse and human variable regions, including, in particular, primers for the VHa, VHb, VHc, VHd, CH1, VL, and CL regions.) These primers can be used to amplify heavy or light chain variable regions, which can then be inserted, respectively, into vectors (such as ImmunoZAPTMH or ImmunoZAPTML (Stratagene)). These vectors can then be introduced into E. coli, yeast, or mammalian-based systems for expression. Using these methods, large quantities of single-chain proteins containing fusions of the VH and VL domains can be produced (see Bird et al., Science 242: 423-426, 1988).

[0127] In some embodiments, the binding constructs of the present invention are obtained from transgenic animals (e.g., mice) that produce “heavy chain only” antibodies or “HCAbs”. HCAbs are similar to naturally occurring camel and llama single-chain VHH antibodies. See, for example, U.S. Patent Nos. 8,507,748 and 8,502,014, and U.S. Patent Application Publications Nos. US 2009 / 0285805 A1, US 2009 / 0169548 A1, US 2009 / 0307787 A1, US 2011 / 0314563 A1, US 2012 / 0151610 A1, WO 2008 / 122886 A2, and WO 2009 / 013620 A2.

[0128] After obtaining cells that produce antibodies according to the invention using any of the above-described immunization and other techniques, specific antibody genes can be cloned according to the standard procedures described herein by isolating and amplifying DNA or mRNA therefrom, and then used to generate the bispecific constructs of the invention. The resulting antibodies can be sequenced and CDRs identified, and the DNA encoding these CDRs can be manipulated as previously described to generate other bispecific constructs according to the invention.

[0129] Molecular evolution of the complementarity-determining region (CDR) at the center of the antibody binding site has also been used to isolate antibodies with increased affinity, such as those described by Schier et al., 1996, J. Mol. Biol. [Journal of Molecular Biology] 263: 551. Therefore, such techniques can be used to prepare the binding constructs of the present invention.

[0130] While human, partially human, or humanized antibodies will be suitable for many applications, particularly those of the present invention, other types of binding constructs will be suitable for certain applications. These non-human antibodies can, for example, be derived from any antibody-producing animal, such as mice, rats, rabbits, goats, donkeys, or non-human primates (e.g., monkeys such as cynomolgus monkeys or rhesus monkeys, or apes such as chimpanzees)). Antibodies from a particular species can be prepared, for example, by immunizing an animal of that species with a desired immunogen or using an artificial system for generating antibodies of that species (e.g., a bacterial or phage display-based system for generating antibodies of a particular species), or by converting an antibody from one species into an antibody from another species (by, for example, replacing the constant region of the antibody with a constant region from another species, or by replacing one or more amino acid residues of the antibody to make it more closely resemble the sequence of an antibody from another species). In one embodiment, the antibody is a chimeric antibody comprising amino acid sequences derived from antibodies from two or more different species. The desired binding region sequence can then be used to generate the bispecific binding constructs of the present invention.

[0131] In cases where it is desired to improve the affinity of binding constructs according to the invention containing one or more of the aforementioned CDRs, this can be achieved through a number of affinity maturation schemes, including maintaining CDRs (Yang et al., J. Mol. Biol. [Journal of Molecular Biology], 254, 392-403, 1995), strand shuffling (Marks et al., Bio / Technology [Biology / Technology], 10, 779-783, 1992), using E. coli mutant strains (Low et al., J. Mol. Biol. [Journal of Molecular Biology], 250, 350-368, 1996), DNA shuffling (Patten et al., Curr. Opin. Biotechnol. [State of Biotechnology], 8, 724-733, 1997), and phage display (Thompson et al., J. Mol. Biol. [Journal of Molecular Biology], 256, 7-88, (1996) and other PCR techniques (Crameri et al., Nature, 391, 288-291, 1998). All these affinity maturation methods are discussed by Vaughan et al. (Nature Biotechnology, 16, 535-539, 1998).

[0132] In some embodiments, to generate the HHLL bispecific binding construct of the present invention, it may first be desirable to generate a more typical single-chain antibody that can be formed by linking heavy and light chain variable domain (Fv region) fragments via amino acid bridges (short peptide linkers) to form a single polypeptide chain. Such single-chain Fv (scFv) has been prepared by fusing DNA encoding a peptide linker between DNA encoding two variable domain polypeptides (VL and VH). The resulting polypeptide can fold itself to form an antigen-binding monomer, or it can form a multimer (e.g., a dimer, trimer, or tetramer) depending on the length of the flexible linker between the two variable domains (Kortt et al., 1997, Prot. Eng. [Protein Engineering] 10: 423; Kortt et al., 2001, Biomol. Eng. [Biomolecular Engineering] 18: 95-108). Techniques for developing single-chain antibodies include those described below: U.S. Patent No. 4,946,778; Bird, 1988, Science 242: 423; Huston et al., 1988, Proc. Natl. Acad. Sci. USA 85: 5879; Ward et al., 1989, Nature 334: 544; de Graaf et al., 2002, Methods Mol Biol. 178: 379-87. These single-chain antibodies are distinct from and different from the bispecific binding constructs of this invention.

[0133] Antigen-binding fragments derived from antibodies can also be obtained, for example, by proteolytic hydrolysis of the antibody (e.g., digestion of the whole antibody with pepsin or papain according to conventional methods). For instance, antibody fragments can be generated by enzymatic cleavage of the antibody with pepsin to provide a 5S fragment called F(ab')2. This fragment can be further cleaved with a thiol reducing agent to generate a 3.5S Fab' monovalent fragment. Optionally, the cleavage reaction can be performed using a blocking group of the thiol group generated by disulfide bond cleavage. Alternatively, enzymatic cleavage with papain can directly generate two monovalent Fab fragments and one Fc fragment. These methods are described in, for example, the following literature: Goldenborg, U.S. Patent No. 4,331,647; Nisonoff et al., Arch. Biochem. Biophys. 89: 230, 1960; Porter, Biochem. J. 73: 119, 1959; Edelman et al., in Methods in Enzymology 1: 422 (Academic Press 1967); and Andrews, SM and Titus, JA, in Current Protocols in Immunology (Coligan JE et al., eds.), John Wiley & Sons, New York (2003), pp. 2.8.1–2.8.10 and 2.10A.1–2.10A.5. Other methods of antibody cleavage can also be used, such as separating the heavy chain to form a monovalent light-heavy chain fragment (Fd), further cleaving the fragment or other enzymatic, chemical or genetic techniques, as long as the fragment can bind to an antigen that can be recognized by the intact antibody.

[0134] In some embodiments, the bispecific binding construct includes one or more complementarity-determining regions (CDRs) of the antibody. The CDR can be obtained by constructing a polynucleotide encoding the target CDR. For example, such polynucleotides can be prepared by using polymerase chain reaction to synthesize variable regions using mRNA from antibody-producing cells as templates (see, for example, Larrick et al., Methods: A Companion to Methods in Enzymology 2: 106, 1991; Courtenay-Luck, “Genetic Manipulation of Monoclonal Antibodies,” in Monoclonal Antibodies: Production, Engineering and Clinical Application, Ritter et al. (ed.), p. 166 (Cambridge University Press, 1995); and Ward et al., “Genetic Manipulation and Expression of Antibodies,” in Monoclonal Antibodies: Principles and Applications, Birch et al. (ed.), p. 137 (Wiley-Liss, Inc., 1995). The antibody fragment may further include at least one variable region domain of the antibody described herein. Thus, for example, as described herein, the V region domain may be a monomer and may be a VH or VL domain capable of independently binding to a desired target (e.g., human CD3) with an affinity of at least 10⁻⁷ M or less.

[0135] The variable region can be any naturally occurring variable domain or its engineered version. An engineered version refers to a variable region generated using recombinant DNA engineering techniques. Such engineered versions include, for example, those generated from the variable region of a specific antibody through insertion, deletion, or alteration of the amino acid sequence of or thereof of a specific antibody. Those skilled in the art can use any known method to identify amino acid residues suitable for engineering. Further examples include engineered variable regions containing at least one CDR from a first antibody and optionally one or more framework amino acids, as well as the remainder of the variable region domain from a second antibody. Engineered versions of antibody variable domains can be generated using any number of techniques familiar to those skilled in the art.

[0136] The variable region can be covalently attached at its C-terminal amino acid to at least one other antibody domain or fragment thereof. Thus, for example, the VH domain present in the variable region can be linked to the immunoglobulin CH1 domain. Similarly, the VL domain can be linked to the CK domain. In this way, for example, the antibody can be a Fab fragment, wherein the antigen-binding domain contains the associated VH and VL domains (covalently linked at their C-termini to the CH1 and CK domains, respectively). The CH1 domain can be extended with additional amino acids, for example, to provide a hinge region or a portion of a hinge region domain as found in the Fab' fragment, or to provide additional domains, such as the antibody CH2 and CH3 domains.

[0137] Combination specificity If an antibody or bispecific binding construct binds to an antigen with a tight-binding affinity determined by an equilibrium dissociation constant of 10⁻⁷ M or less (KD as defined below, or the corresponding KD), then the antibody or bispecific binding construct “specifically binds” the antigen.

[0138] Affinity can be determined using a variety of techniques known in the art, such as, but not limited to, equilibrium methods (e.g., enzyme-linked immunosorbent assay (ELISA); KinExA, Rathanaswami et al. Analytical Biochemistry, Vol. 373: 52-60, 2008; or radioimmunoassay (RIA)), or by surface plasmon resonance assays or other kinetic-based assays (e.g., BIACORE® assays or Octet® assays (forteBIO)), and other methods such as indirect binding assays, competitive binding assays, fluorescence resonance energy transfer (FRET), gel electrophoresis, and chromatography (e.g., gel filtration). These and other methods can label one or more components to be examined and / or use multiple detection methods, including but not limited to chromogenic labeling, fluorescent labeling, luminescent labeling, or isotopic labeling. A detailed description of affinity and kinetics can be found in Paul, WE, ed., Fundamental Immunology, 4th ed., Lippincott-Raven, Philadelphia (1999), which focuses on antibody-immunogen interactions. An example of competitive binding assays is radioimmunoassay, which involves incubating a labeled antigen with a target antibody in the presence of an increased amount of unlabeled antigen and detecting the antibody binding to the labeled antigen. The affinity of the target antibody for a specific antigen and the binding off-rate can be determined from the data using scatchard plot analysis. Competition with a secondary antibody can also be determined using radioimmunoassay. In this case, the antigen is incubated with a target antibody conjugated to a labeled compound in the presence of an increased amount of unlabeled secondary antibody. This type of assay can be readily adapted for use with the bispecific binding constructs of the present invention.

[0139] Further embodiments of the invention provide bispecific binding constructs that bind to desired targets with an equilibrium dissociation constant or KD (koff / kon) of less than 10⁻⁷ M, or less than 10⁻⁸ M, or less than 10⁻⁹ M, or less than 10⁻¹⁰ M, or less than 10⁻¹¹ M, or less than 10⁻¹² M, or less than 10⁻¹³ M, or less than 5 x 10⁻¹³ M (lower values ​​indicate tighter binding affinity). Still other embodiments of the invention are bispecific binding constructs that bind to desired targets with an equilibrium dissociation constant or KD (koff / kon) of less than about 10⁻⁷ M, or less than about 10⁻⁸ M, or less than about 10⁻⁹ M, or less than about 10⁻¹⁰ M, or less than about 10⁻¹¹ M, or less than about 10⁻¹² M, or less than about 10⁻¹³ M, or less than about 5 x 10⁻¹³ M.

[0140] In yet another embodiment, the bispecific binding construct that binds to the desired target has an equilibrium dissociation constant or KD (koff / kon) between about 10⁻⁷ M and about 10⁻⁸ M, between about 10⁻⁸ M and about 10⁻⁹ M, between about 10⁻⁹ M and about 10⁻¹⁰ M, between about 10⁻¹⁰ M and about 10⁻¹¹ M, between about 10⁻¹¹ M and about 10⁻¹² M, and between about 10⁻¹² M and about 10⁻¹³ M. In yet another embodiment, the bispecific construct of the present invention has an equilibrium dissociation constant or KD (koff / kon) between 10⁻⁷ M and 10⁻⁸ M, between 10⁻⁸ M and 10⁻⁹ M, between 10⁻⁹ M and 10⁻¹⁰ M, between 10⁻¹⁰ M and 10⁻¹¹ M, between 10⁻¹¹ M and 10⁻¹² M, and between 10⁻¹² M and 10⁻¹³ M.

[0141] Molecular stability Various aspects of molecular stability may be desired, particularly in the context of biopharmaceutical therapeutic molecules. For example, stability at different temperatures (“thermal stability”) may be desired. In some embodiments, this may cover stability within physiological temperature ranges (e.g., at or about 37°C, or from 32°C to 42°C). In other embodiments, this may cover stability within higher temperature ranges (e.g., 42°C to 60°C). In still other embodiments, this may cover stability within colder temperature ranges (e.g., 20°C to 32°C). In yet another embodiment, this may cover stability when frozen (e.g., 0°C or lower).

[0142] Determining the thermal stability of protein molecules is known in the art. For example, the fully automated UNcle platform (Unchained Labs) is used, and further described in the examples, which allows for the simultaneous acquisition of intrinsic protein fluorescence and static light scattering (SLS) data during thermal temperature variations. Furthermore, the thermal stability and aggregation determinations described in the examples herein (such as differential scanning fluorescence (DSF) and static light scattering (SLS)) can also be used to measure thermal melting (Tm) and thermal aggregation (Tagg), respectively.

[0143] Alternatively, and as illustrated in the examples herein, accelerated stress studies can be performed on the molecules. In short, this involves incubating protein molecules at a specific temperature (e.g., 40°C) and then measuring aggregation at different time points using size exclusion chromatography (SEC), where lower aggregation indicates better protein stability.

[0144] Alternatively, the thermal stability parameter can be determined based on the molecular aggregation temperature as follows: A molecular solution with a concentration of 250 µg / ml is transferred to a disposable cuvette and placed in a dynamic light scattering (DLS) device. The sample is heated from 40°C to 70°C at a constant heating rate of 0.5°C / min, and the measured radius is obtained. The molecular aggregation temperature is calculated using the increase in radius indicating the melting of the protein and aggregates.

[0145] Alternatively, temperature melting profiles can be determined using differential scanning calorimetry (DSC) to ascertain the intrinsic biophysical protein stability of the binding construct. These experiments were performed using a VP-DSC device from MicroCal LLC (Northampton, MA, USA). Energy uptake of samples containing the binding construct was recorded from 20°C to 90°C compared to samples containing only the preparation buffer. For example, the binding construct was adjusted to a final concentration of 250 μg / ml in SEC running buffer. To record the corresponding melting profiles, the sample temperature was gradually increased. At each temperature T, the energy uptake of the sample and the preparation buffer reference was recorded. The difference between the energy uptake Cp (kcal / mol / °C) of the sample and the reference was plotted against the corresponding temperature. The melting temperature was defined as the temperature at which the first maximum energy uptake occurred.

[0146] In another embodiment, the bispecific binding construct according to the invention is stable at approximately physiological pH (i.e., approximately pH 7.4). In other embodiments, the bispecific binding construct is stable at lower pH (e.g., as low as pH 6.0). In other embodiments, the bispecific binding construct is stable at higher pH (e.g., up to pH 9.0). In one embodiment, the bispecific binding construct is stable at pH 6.0 to 9.0. In another embodiment, the bispecific binding construct is stable at pH 6.0 to 8.0. In yet another embodiment, the bispecific binding construct is stable at pH 7.0 to 9.0.

[0147] In some embodiments, the greater the tolerance of the bispecific binding construct to non-physiological pH (e.g., pH 6.0), the higher the recovery rate of molecules eluted from the ion exchange column relative to the total amount of loaded protein. In one embodiment, the molecule recovery rate from the ion (e.g., cation) exchange column is ≥ 30%. In another embodiment, the molecule recovery rate from the ion (e.g., cation) exchange column is ≥ 40%. In another embodiment, the molecule recovery rate from the ion (e.g., cation) exchange column is ≥ 50%. In another embodiment, the molecule recovery rate from the ion (e.g., cation) exchange column is ≥ 60%. In another embodiment, the molecule recovery rate from the ion (e.g., cation) exchange column is ≥ 70%. In another embodiment, the molecule recovery rate from the ion (e.g., cation) exchange column is ≥ 80%. In another embodiment, the molecule recovery rate from the ion (e.g., cation) exchange column is ≥ 90%. In another embodiment, the molecule recovery rate from the ion (e.g., cation) exchange column is ≥ 95%. In another embodiment, the molecule recovery rate from the ion (e.g., cation) exchange column is ≥ 99%.

[0148] In some embodiments, it may be desirable to determine the chemical stability of the molecule. As further described in the examples herein, the determination of the chemical stability of the bispecific binding construct can be performed via isothermal chemical denaturation (“ICD”) by monitoring the intrinsic protein fluorescence. ICD produces C1 / 2 and ΔG, which can serve as good measures of protein stability. C1 / 2 is the amount of chemical denaturant required to denature 50% of the protein and is used to derive ΔG (or unfolding energy).

[0149] For biopharmaceuticals, protein chain cleavage is another critical product quality attribute that requires careful monitoring and reporting. Typically, the expectation of longer and / or fewer structured linkers leads to increased cleavage (as a function of incubation time and temperature). Cleavage is a critical issue for bispecific binding constructs because cleavage of linkers connecting to the target or T-cell binding domain has ultimately adverse effects on pharmacodynamics and efficacy. Cleavage at other sites, including scFc, can affect pharmacodynamic / pharmacokinetic properties. Increased cleavage is an attribute to be avoided in pharmaceutical products. Therefore, in some embodiments, protein cleavage can be determined as described in the examples herein.

[0150] Immune effector cells and effector cell proteins The bispecific binding construct can bind to a molecule expressed on the surface of immune effector cells (referred to herein as "effector cell protein") and another molecule expressed on the surface of target cells (referred to herein as "target cell protein"). Immune effector cells can be T cells, NK cells, macrophages, or neutrophils. In some embodiments, the effector cell protein is a protein included in the T cell receptor (TCR)-CD3 complex. The TCR-CD3 complex is a heterodimer comprising TCRα and TCRβ or TCRγ and TCRδ, and various CD3 chains among CD3 zeta (CD3ζ), CD3 epsilon (CD3ε), CD3 gamma (CD3γ), and CD3 delta (CD3δ) chains.

[0151] The CD3 receptor complex is a protein complex composed of four chains. In mammals, the complex contains a CD3γ (gamma) chain, a CD3δ (delta) chain, and two CD3ε (epsilon) chains. These chains associate with the T cell receptor (TCR) and the so-called zeta chain to form the T cell receptor CD3 complex and generate an activation signal in T lymphocytes. The CD3γ (gamma), CD3δ (delta), and CD3ε (epsilon) chains are highly correlated cell surface proteins of the immunoglobulin superfamily containing a single extracellular immunoglobulin domain. The intracellular tail of the CD3 molecule contains a single conserved motif essential for TCR signaling, called the immunoreceptor tyrosine-based activation motif, or ITAM for short. The CD3ε molecule is a polypeptide encoded in humans by the CD3E gene located on chromosome 11. The most preferred epitopes of CD3ε are contained within amino acid residues 1-27 of the extracellular domain of the human CD3ε. It is envisioned that the bispecific binding construct according to the invention typically and advantageously exhibits less nonspecific T cell activation, which is undesirable in specific immunotherapy. This implies a reduced risk of side effects.

[0152] In some embodiments, the effector cell protein may be a human CD3 epsilon (CD3ε) chain (the mature amino acid sequence of which is disclosed in SEQ ID NO: 40), which may be part of a multimeric protein. Alternatively, the effector cell protein may be a human or / or cynomolgus monkey TCRα, TCRβ, TCRδ, TCRγ, CD3 beta (CD3β) chain, CD3 gamma (CD3γ) chain, CD3 delta (CD3δ) chain, or CD3 zeta (CD3ζ) chain.

[0153] Furthermore, in some embodiments, the bispecific binding construct can also bind to the CD3ε chain from non-human species such as mice, rats, rabbits, New World monkeys, and / or Old World monkey species. Such species include, but are not limited to, the following mammalian species: house mouse; black mouse (Rattus rattus); brown mouse; cynomolgus monkey (Macacafascicularis); Arabian baboon (Hamadryas baboon / Papiohamadryas); Guinea baboon (Papio papio); East African baboon (Olive baboon / Papioanubis); steppe baboon (Yellowbaboon / Papio cynocephalus); pig-tailed baboon (Chacma baboon / Papioursinus); common marmoset (Callithrix jacchus); woolly-crowned tamarin (Saguinus Oedipus); and squirrel monkey (Saimiri sciureus). The mature amino acid sequence of the cynomolgus monkey CD3ε chain is provided in SEQ ID NO: 41. The availability of therapeutic molecules with similar activities in humans and species commonly used in preclinical trials, such as mice and monkeys, can streamline, accelerate, and ultimately deliver improved drug development outcomes. These advantages may be crucial in the long and costly process of bringing drugs to market.

[0154] In some embodiments, the bispecific binding construct may bind to an epitope (SEQ ID NO: 43) within the first 27 amino acids of the CD3ε chain, which may be a human CD3ε chain or a CD3ε chain from a different species (particularly one of the mammalian species listed above). This epitope may contain the amino acid sequence Gln-Asp-Gly-Asn-Glu (SEQ ID NO: 59). The advantages of molecules binding such epitopes are described in detail in U.S. Patent Application Publication 2010 / 0183615 A1 (the relevant portion of which is incorporated herein by reference). The epitope bound by the antibody or bispecific binding construct can be determined by an alanine scan, as described, for example, in U.S. Patent Application Publication 2010 / 0183615 A1, the relevant portion of which is incorporated herein by reference. In other embodiments, the bispecific binding construct may bind to an epitope (SEQ ID NO: 42) within the extracellular domain of CD3ε.

[0155] In embodiments where T cells are immune effector cells, the bispecific binding construct can bind effector cell proteins including, but not limited to, CD3ε, CD3γ, CD3δ, CD3ζ chains, TCRα, TCRβ, TCRγ, and TCRδ. In embodiments where NK cells or cytotoxic T cells are immune effector cells, NKG2D, CD352, NKp46, or CD16a can be, for example, effector cell proteins. In embodiments where CD8+ T cells are immune effector cells, 4-1BB or NKG2D can be, for example, effector cell proteins. Alternatively, in other embodiments, the bispecific binding construct can bind to other effector cell proteins expressed on T cells, NK cells, macrophages, or neutrophils.

[0156] Target cells and target cell proteins expressed on target cells As explained above, the bispecific binding construct can bind to both effector cell proteins and target cell proteins. Target cell proteins can be expressed, for example, on the surface of cancer cells, cells infected with pathogens, or cells mediating diseases such as inflammation, autoimmune diseases, and / or fibrotic conditions. In some embodiments, target cell proteins can be highly expressed on target cells, although high levels of expression are not necessarily required.

[0157] When the target cells are cancer cells, the bispecific binding constructs described herein can bind to cancer cell antigens as described above. Cancer cell antigens can be human proteins or proteins from other species. For example, the bispecific binding constructs can bind to target cell proteins from mouse, rat, rabbit, New World monkey, and / or Old World monkey species. These species include, but are not limited to, the following: house mouse; black mouse (Rattus rattus); brown mouse; cynomolgus monkey (Macaca fascicularis); Arabian baboon (Hamadryas baboon / Papiohamadryas); Guinea baboon (Papio papio); East African baboon (Olive baboon / Papioanubis); grassland baboon (Yellow baboon / Papio cynocephalus); pig-tailed baboon (Chacmababoon / Papioursinus), common marmoset, woolly tamarin, and squirrel monkey.

[0158] In some instances, the target cell protein may be a protein selectively expressed on the infected cell. For example, in the case of HBV or HCV infection, the target cell protein may be an HBV or HCV envelope protein expressed on the surface of the infected cell. In other embodiments, the target cell protein may be gp120 encoded by HIV on the infected cell.

[0159] In other respects, target cells can be cells that mediate autoimmune or inflammatory diseases. For example, human eosinophils in asthma can be target cells, in which case an EGF-like module containing mucin-like hormone receptor (EMR1) can be, for example, a target cell protein. Alternatively, excess human B cells in patients with systemic lupus erythematosus can be target cells, in which case CD19 or CD20 can be, for example, a target cell protein. In other autoimmune disorders, excess human Th2 T cells can be target cells, in which case CCR4 can be, for example, a target cell protein. Similarly, target cells can be fibrotic cells that mediate diseases such as atherosclerosis, chronic obstructive pulmonary disease (COPD), cirrhosis, scleroderma, kidney transplant fibrosis, kidney transplant nephropathy, or pulmonary fibrosis, including idiopathic pulmonary fibrosis and / or idiopathic pulmonary hypertension. For such fibrotic disorders, fibroblast activation protein α (FAPα) can be, for example, a target cell protein.

[0160] Treatment methods and compositions Bispecific binding constructs can be used to treat a variety of conditions, including, for example, various forms of cancer, infections, autoimmune or inflammatory conditions, and / or fibrotic conditions.

[0161] Therefore, in the embodiments, this document provides bispecific binding constructs for use in the prevention, treatment or improvement of disease.

[0162] Another embodiment provides the use of the binding construct of the present invention (or the binding construct produced according to the method of the present invention) in the manufacture of a medicament for the prevention, treatment or relief of disease.

[0163] This document provides pharmaceutical compositions comprising a bispecific binding construct. These pharmaceutical compositions comprise a therapeutically effective amount of the bispecific binding construct and one or more additional components, such as physiologically acceptable carriers, excipients, or diluents. In some embodiments, these additional components may include buffers, carbohydrates, polyols, amino acids, chelating agents, stabilizers, and / or preservatives.

[0164] In some embodiments, bispecific binding constructs can be used to treat proliferative disorders, including cancers, involving uncontrolled and / or inappropriate cell proliferation, sometimes accompanied by destruction of adjacent tissues and the growth of new blood vessels, which can allow cancer cells to invade new areas, i.e., metastasis. Conditions treatable with bispecific binding constructs include non-malignant conditions involving inappropriate cell growth, including colorectal polyps, cerebral ischemia, giant cystic disease, polycystic kidney disease, benign prostatic hyperplasia, and endometriosis. Bispecific binding constructs can also be used to treat hematologic malignancies or solid tumor malignancies. More specifically, proliferative disorders treatable with bispecific binding constructs include, for example, cancers such as mesothelioma, squamous cell carcinoma, myeloma, osteosarcoma, glioblastoma, glioma, carcinomas, adenocarcinoma, melanoma, sarcoma, acute and chronic leukemia, lymphoma, and meningioma, Hodgkin's disease, and Sézary syndrome. Multiple myeloma, and lung cancer, non-small cell lung cancer, small cell lung cancer, laryngeal cancer, breast cancer, head and neck cancer, bladder cancer, ovarian cancer, skin cancer, prostate cancer, cervical cancer, vaginal cancer, gastric cancer, renal cell carcinoma, kidney cancer, pancreatic cancer, colorectal cancer, endometrial cancer and esophageal cancer, hepatobiliary cancer, bone cancer, skin cancer and blood cancer, as well as nasal cavity cancer and paranasal sinus cancer, nasopharyngeal cancer, oral cavity cancer, oropharyngeal cancer, laryngeal cancer, lower laryngeal cancer, salivary gland cancer, mediastinal cancer, gastric cancer, small intestine cancer, colon cancer, rectal cancer and anal region cancer, ureteral cancer, urethral cancer, penile cancer, testicular cancer, vulvar cancer, endocrine system cancer, central nervous system cancer and plasma cell cancer.

[0165] Texts providing guidance for cancer therapy include *Cancer, Principles and Practice of Oncology*, 4th edition, edited by DeVita et al., JB Lippincott Co., Philadelphia, PA (1993). Appropriate treatment methods should be selected based on the specific cancer type and other factors recognized in the relevant field, such as the patient's general condition. In treating cancer patients, bispecific binding constructs can be added to treatment regimens using other antitumor agents.

[0166] In some embodiments, the bispecific binding construct can be administered concurrently with, before, or after a variety of drugs and treatments widely used in cancer treatment (e.g., chemotherapeutic agents, non-chemotherapeutic agents, antitumor agents, and / or radiation). For example, chemotherapy and / or radiation can be administered before, during, and / or after any of the treatments described herein. Examples of chemotherapeutic agents discussed above include, but are not limited to, cisplatin, paclitaxel, etoposide, mitoxantrone (Novantrone®), actinomycin D, cycloheximide, camptothecin (or its water-soluble derivatives), methotrexate, mitomycin (e.g., mitomycin C), dacarbazine (DTIC), antitumor antibiotics (such as doxorubicin (doxorubicin) and daunomycin), and all of the chemotherapeutic agents mentioned above.

[0167] Bispecific binding constructs can also be used to treat infectious diseases such as chronic hepatitis B virus (HBV) infection, hepatitis C virus (HCV) infection, human immunodeficiency virus (HIV) infection, Epstein-Barr virus (EBV) infection, or cytomegalovirus (CMV) infection.

[0168] Bispecific binding constructs can be further used in other types of diseases where the depletion of certain cell types is beneficial. For example, the depletion of human eosinophils in asthma, the depletion of excess human B cells in systemic lupus erythematosus, the depletion of excess human Th2 T cells in autoimmune diseases, or the depletion of pathogen-infected cells in infectious diseases may be beneficial. In fibrotic diseases, the depletion of cells that form fibrotic tissue may be useful.

[0169] A therapeutically effective dose of the bispecific binding construct can be administered. The amount of bispecific binding construct constituting a therapeutic dose can vary depending on the indication for treatment, the patient's weight, and the calculated patient skin surface area. The administration of the bispecific binding construct can be adjusted to achieve the desired effect. In many cases, repeated administration may be necessary.

[0170] Bispecific binding constructs or pharmaceutical compositions containing such molecule can be administered by any feasible method. Protein therapeutics are typically administered via parenteral routes (e.g., by injection) because, without certain specific formulations or conditions, oral administration would result in the hydrolysis of the protein in the acidic environment of the stomach. Subcutaneous, intramuscular, intravenous, intra-arterial, intralesional, or peritoneal bolus injections are possible routes of administration. Bispecific binding constructs can also be administered by infusion (e.g., intravenous or subcutaneous infusion). Topical administration is also possible, particularly for diseases involving the skin. Alternatively, bispecific binding constructs can be administered by contact with mucous membranes, such as intranasal, sublingual, vaginal, or rectal administration, or as an inhaler. Alternatively, certain suitable pharmaceutical compositions containing bispecific binding constructs can be administered orally.

[0171] The term "treatment" encompasses the reduction of at least one symptom or other manifestation of a disorder, or the reduction of disease severity, etc. The bispecific binding construct according to the invention does not need to achieve a complete cure or eradication of every symptom or manifestation of the disease to be considered a viable therapeutic agent. As recognized in the relevant art, a medicine used as a therapeutic agent can reduce the severity of a given disease state, but does not need to eliminate every manifestation of the disease to be considered a useful therapeutic agent. It is sufficient to merely reduce the effects of the disease (e.g., by reducing the frequency or severity of its symptoms, or by increasing the effectiveness of another treatment, or by producing another beneficial effect) or reduce the likelihood of the disease occurring or worsening in a subject. One embodiment of the invention relates to a method comprising administering to a patient a certain amount of the bispecific binding construct of the invention for a period of time, the amount and duration of which are sufficient to cause a sustained improvement compared to a baseline indicative of the severity of a particular disorder.

[0172] The term "prevention" encompasses the prevention of at least one symptom or other manifestation of the disorder. A treatment combining prophylactic administration of the bispecific binding construct according to the invention does not need to be completely effective in preventing the onset of the condition to be considered a viable preventative agent. Simply reducing the likelihood of the disease occurring or worsening in the subject is sufficient.

[0173] As understood in the relevant art, a pharmaceutical composition comprising a bispecific binding construct is administered to a subject in a manner appropriate to the indication and the composition. The pharmaceutical composition may be administered by any suitable technique, including but not limited to parenteral, topical, or inhalation. If administered by injection, the pharmaceutical composition may be administered, for example, via intra-articular, intravenous, intramuscular, intralesional, intraperitoneal, or subcutaneous routes, by bolus injection or continuous infusion. Inhalation delivery includes, for example, nasal or oral inhalation, use of a nebulizer, inhalation of the binding construct in aerosol form, etc. Other alternatives include oral formulations, including pills, syrups, or tablets.

[0174] The bispecific binding construct can be administered as a composition comprising one or more additional components, such as physiologically acceptable carriers, excipients, or diluents. Optionally, the composition further comprises one or more physiologically active agents. In various specific embodiments, in addition to one or more bispecific binding constructs, the composition also comprises one, two, three, four, five, or six physiologically active agents.

[0175] Kits are provided for use by licensed physicians, comprising one or more bispecific binding constructs, and labeling or other instructions for use in treating any of the conditions discussed herein. In one embodiment, the kit comprises a sterile formulation of one or more bispecific binding constructs, which may be in the form of the compositions disclosed herein and may be in one or more vials.

[0176] The dosage and frequency of administration may vary depending on factors such as the route of administration, the specific bispecific binding construct used, the nature and severity of the disease to be treated, whether the condition is acute or chronic, and the size and general condition of the subject.

[0177] The invention has been described above in general terms, and the following examples are provided for illustration rather than limitation.

[0178] Example Example 1 Generation and expression of bispecific HHLL binding constructs To evaluate whether the proposed HHLL form produces improved stability, we designed, expressed, and purified a set of BiTEs targeting Flt3, Msln, and Dll3, respectively. These included: standard or wild-type (WT), HLE-BiTE (HLE = extended half-life), and three different linker length configurations: tight, intermediate, or loose, representing the proposed HHLL BiTE architecture.

[0179] We utilize the repeating connections of the GlyGlyGlyGlySer (G4S) (SEQ ID NO: 1) connector to link various chains containing HLE-BiTE. In WT HLE-BiTE, the heavy and light chains (H1 and L1) of the target domain are linked by three G4S repeats (SEQ ID NO: 3). The anti-CD3 domain also contains heavy and light chain pairs (H2 and L2) linked by three G4S repeats (SEQ ID NO: 3). To connect the anti-target scFv domain and the anti-CD3 scFv domain, a single G4S connector (SEQ ID NO: 1) is required. This can also be described alternatively by the following nomenclature: H1-(G4S)3-L1-(G4S)1-H2-(G4S)3-L2.

[0180] In the proposed “tight” HHLL configuration, two G4S repeats (SEQ ID NO: 2) link H1 to H2, followed by three G4S repeats (SEQ ID NO: 3) linking this segment to the L1 chain. Then, three more G4S repeats (SEQ ID NO: 3) are needed to link H1H2L1 to L2, allowing for the correct folding as determined by molecular modeling. This “tight” configuration can also be described as: H1-(G4S)2-H2-(G4S)3-L1(G4S)3-L2, or simply 233.

[0181] The “medium” configuration connects all the various chains (H1, H2, L1 and L2) through a series of quadruple G4S connectors (SEQ ID NO: 4) and can be described as follows: H1-(G4S)4-H2-(G4S)4-L1(G4S)4-L2, or simply 444.

[0182] The “loose” configuration utilizes a series of five-repeating G4S connectors (SEQ ID NO: 5) and can be described as follows: H1-(G4S)5-H2-(G4S)-L1(G4S)5-L2, or simply 444.

[0183] plasmid: The expression plasmid carrying the target BiTE gene with an N-terminal signal peptide was cloned into the pTT5 vector.

[0184] Expression and purification: All BiTE proteins were generated using transiently transfected HEK293-6E cells. In short, plasmid DNA encoding a BiTE target sequence with an N-terminal signal secretion peptide was introduced into cells using PEI MAX transfection reagent (at approximately 99.9% viability and a cell density of 1.5e6). Cells were maintained at 37°C, 5% CO2, and 150 RPM for 6 days to allow for protein overproduction. Cells were then harvested by centrifugation (4000 RPM for 30 min), and the resulting cell culture supernatant was filtered and stored for purification.

[0185] BiTE expressing a single-chain Fc region (“scFc”) was purified using protein A affinity chromatography (GE Healthcare, HiTrap mAb SelectSuRe). Protein A resin was equilibrated in binding buffer (25 mM Tris, 100 mM NaCl, pH 7.4) and the protein was eluted with 100 mM sodium acetate (pH 3.6). To rapidly remove BiTE from the elution buffer, it was desalted in 10 mM potassium phosphate, 75 mM lysine, and 4% trehalose (pH 8.0) before separation by size exclusion chromatography (GE Healthcare, HiLoad Superdex 200). scFc-free BiTE was purified in a similar manner using protein L resin (GE Healthcare, HiTrap Protein L) instead of protein A. Purity was verified by SDS-PAGE. The protein was then prepared at a concentration of 1 mg / ml in 10 mM glutamic acid, 9% sucrose, and 0.01% polysorbate 80 (pH 4.2). The protein was stored at -80°C before use.

[0186] Example 2 Analytical HPLC size exclusion chromatography To quantify protein aggregation and stability, temperature studies over several time periods were performed using analytical size exclusion chromatography (SEC). The same SEC method was employed to simplify analysis regardless of the study temperature. Approximately 2 µg of sample was injected onto an SEC column (Waters ACQUITY UPLCProtein BEH SEC 200 Å) equilibrated in 100 mM sodium phosphate and 250 mM NaCl (pH 6.8). Absorbance data were collected at 280 nm and the Chromeleon integration peak was used. The data were plotted and analyzed in a GraphPad Prism.

[0187] Example 3 Reduction capillary electrophoresis To examine the formation of low molecular weight (LMW) species in our temperature stability assay, reductive capillary electrophoresis-sodium dodecyl sulfate (rCE-SDS) analysis was performed. Prior to analysis, the sample was denatured and reduced by SDS and heat, as well as in the presence of β-mercaptoethanol. Approximately 10 µg of sample was then loaded onto the CE cassette, and proteins and fragments were monitored by UV detection. Purity was determined using Chromeleon quantification of peak area percentage.

[0188] Example 4 Chemical stability Isothermal chemical denaturation The determination of BiTE chemical stability can be performed via isothermal chemical denaturation (“ICD”) by monitoring intrinsic protein fluorescence. ICD generates C1 / 2 and ΔG, which can serve as good measures of protein stability. C1 / 2 is the amount of chemical denaturant required to denature 50% of the protein and is used to derive ΔG (or unfolding energy). To monitor protein unfolding as a function of the chemical denaturant, we measured intrinsic protein fluorescence. To investigate the chemical stability profile of the HHLL form, we used three different targets, Flt3, Msln, and Dll3. This process was fully automated using a HUNK instrument (Unchained Labs). Thirty-two independent denaturation data points were generated, ranging from 0 to 5.52 M guanidine HCl (GuHCl), and the resulting 350 / 330 nm fluorescence intensity ratio was plotted and fitted to determine the denaturation fraction and derive C1 / 2 and ΔG values. Forty-two data points were collected to define the Flt3 data as having the highest certainty. The figure shows the normalized data fit. Data were fitted to either a two-state or three-state model. Figure 8 The results of this experiment are described.

[0189] In the case of Flt3, we observed the highest degree of improvement in chemical stability. Here, as can be seen from the approximately 0.8 M shift in C1 / 2 and the approximately 1–6 kcal / mol improvement in ΔG, Flt3-444 and 555, or the “moderate” and “loose” variants, showed a significant improvement over Flt3-WT. In contrast, no dramatic shift in C1 / 2 was observed in the cases of Msln and Dll3. However, as can be seen from the lower ΔG values, Msln-555, or the loose conformation, was slightly less stable than Msln-WT. Based on these data, we conclude that, as observed in the Flt3-444 and Flt-555 variants, the chemical stability of the HHLL form is either comparable to that of the standard HLHL form or, ideally, can outperform the WT protein in a target domain-dependent manner.

[0190] Table 3 below summarizes some of the results: Table 3 Construct C 1 / 2 (1) ∆G（1） M（1） C 1 / 2 (2) ∆G（2） M（2） Flt3-WT 2.81 2.85 ± 0.07 1.01 ± 0.03 Flt3-233 3.62 4.46 ± 0.64 1.72 ± 0.40 Flt3-444 3.60 5.98 ± 0.70 1.66 ± 0.20 Flt3-555 3.33 3.99 ± 0.44 1.2 ± 0.20 Msln-WT 1.66 7.79 ± 1.09 4.71 ± 0.15 2.58 2.58 ± 0.17 1.59 ± 0.44 Msln 233 1.5 11.15 ± 2.5 7.41 ± 0.25 2.66 4.96 ± 0.20 1.86 ± 0.95 Msln 355 1.99 3.36 ± 0.31 1.69 ± 0.38 2.73 6.31 ± 0.68 2.25 ± 0.15 Msln-444 1.6 9.24 ± 1.6 5.79 ± 0.19 2.58 4.57 ± 0.20 1.77 ± 0.64 Msln-444-No scFc 1.48 9.57 ± 1.93 6.46 ± 0.40 2.74 8.95 ± 0.44 3.27 ± 0.72 Msln-555 1.93 3.53 ± 0.32 1.83 ± 0.43 2.68 5.16 ± 0.71 1.92 ± 0.16 Example 5 thermal stability Differential scanning fluorometer Tm and static light scattering Tagg measurement To determine the different Tm and Tagg values ​​for our BiTE protein, we used the fully automated UNcle platform (Unchained Labs), which allows for the simultaneous acquisition of intrinsic protein fluorescence and static light scattering (SLS) data during thermal temperature changes. In short, during data acquisition, a 1 mg / ml protein sample underwent thermal temperature changes from 20°C to 90°C. Tm and Tagg values ​​were derived from the averages of three replicates using UNcle analysis software.

[0191] Further characterization of the HHLL form using thermal stability and aggregation assays also revealed enhanced stability parameters. The results of these different experiments are described in... Figure 9A , Figure 9B and Figure 10 We screened a group of BiTE molecules using differential scanning fluorescence (DSF) and static light scattering (SLS) to measure thermal melting (Tm) and thermal aggregation (Tagg), respectively. Interestingly, again in the case of Flt3, we observed the highest degree of improvement in the Tm curve. As observed from the shifts in the DSF data and derivative curves, all HHLL Flt3 variants (Flt3-233, Flt3-444, and Flt3-555) showed an improvement in Tm of approximately 3 °C. Similar to Flt3-444 and Flt3-555, the thermochromic static light scattering data for Flt3 did not show this improvement, although they had comparable Tagg values ​​compared to Flt3-WT. Notably, the compact configuration of Flt3-233 exhibited a significantly reduced Tagg compared to Flt3-444, Flt3-555, and Flt3-WT, which may indicate that this variant may be prone to aggregation.

[0192] In the case of the Msln HHLL variants, we did not observe any significant changes in the DSF or derivative curves, and therefore Tm remained unchanged. However, interestingly, we did observe significant differences in the Tagg properties of the Msln HHLL variants. Here, we observed improvements in Tagg values ​​of approximately 2-3 °C for Msln-233, Msln-444, and Msln-555. This indicates that the Msln HHLL variants do indeed possess improved aggregation properties compared to the standard Msln BiTE. We observed a similar trend in Dll3 HHLL and the standard BiTE. The data suggest that the HHLL form can indeed outperform the standard form.

[0193] Accelerated stress research The performance of the HHLL form under accelerated stress conditions was evaluated. A set of BiTEs was incubated at 40 °C, and aggregation was measured by analytical size exclusion chromatography (SEC) at time points T0, week 2 (2W), and week 4 (4W). We quantified the aggregation level by integrating the peaks of the high molecular weight (HMW) (major) and low molecular weight (LMW). Interestingly, unlike previous data (where the Flt3 HHLL variant showed the highest level of improvement in protein stability (i.e., ICD and Tm), we did not observe a significant improvement in its aggregation properties. However, this is indeed consistent with the Tagg data for Flt3 obtained by SLS, where we observed no difference in Tagg. It should also be noted that the SLS data suggest that the Flt-233, or “tight” conformation, may have a higher tendency to aggregate, as indicated by the reduced Tagg values, and we observed this in experimental SEC measurements. Improved aggregation properties were observed by SEC in the Msln and Dll3 HHLL variants with improved Tagg properties. Msln / Dll3-444 and Msln-555 variants showed the greatest improvement via BiTE; among them, the 444 variant exhibited half the aggregation level compared to WT over a 4-week time period. Interestingly, the Msln-555 variant appeared to have fully reversible aggregation, and the HMW species decreased to undetectable levels after 4 weeks. The Dll3-332 HHLL variant with a non-ideal linker length was included as a negative control. As observed by molecular modeling, this protein contains a short final linker that may hinder the folding of the final light chain. This protein aggregated readily over time, more readily than the WT or HHLL variants used to validate our molecular modeling approach. These data collectively suggest that HHLL forms with optimized linker lengths can improve aggregation or at least be equivalent to WT.

[0194] Protein shearing In addition to investigating accelerated stress-induced aggregation, protein cleavage was examined. Cleavage is a critical issue for BiTEs because cleavage of the linker connecting the target or T-cell junctional domain has ultimately adverse effects on pharmacological efficacy and efficacy. Cleavage at other sites, including scFc, can also affect pharmacodynamic / pharmacokinetic properties. For these reasons, reducing capillary electrophoresis (rCE) was used to assess protein cleavage at T0, 2 weeks, and 4 weeks for a set of BiTEs. The results of this experiment are depicted on... Figure 11For clarity, we present exemplary rCE traces of the Flt3-444 variant at 40°C at T0 and 4W, with LMW and main peak regions marked. Flt3 BiTE exhibits the most variable level of shear, however, Flt-444 outperforms WT HLE-BiTE, with a 10%–15% reduction in shear events. It should be noted that the “tight” Flt3-233 protein behaves similarly to WT, while the “loose” Flt3-555 protein, with the longest linker, exhibits the highest degree of shear at 2W, and these sheared segments are further degraded at 4W. Surprisingly, for Msln and Dll3 “medium and loose,” or 444 and 555 proteins, we observed a 5%–10% reduction in shear in the stressed samples at 4W 40°C. A linker length-dependent effect appears to exist on protein shear. The data suggest that the 444 HHLL form generally appears to be well-tolerated throughout the HLE-BiTE group.

[0195] Stability in frozen state (-20℃) Triple 100 µL aliquots of each protein sample (1 mg / mL) were frozen at -20 °C. The protein samples were thawed at t0, week 2, week 4, and week 8 (at room temperature). Immediately after thawing, the samples were run on size exclusion chromatography (HPLC) to quantify the percentage of the main peak. HHLL constructs with different linker lengths (e.g., 444 and 555 constructs) provided freeze-state stability comparable to that of the WT bispecific constructs.

[0196] Example 6 Freeze-thaw stability determination Triple 100 μL samples of each protein (1 mg / mL) were frozen in 1.5 mL microcentrifuge tubes. Proteins were thawed by removing them from a -80°C freezer and allowing them to reach room temperature, then returning them to a -80°C freezer for refreezing. 1X, 5X, and 10X freeze-thaw cycles were performed on the samples. Results are plotted on... Figure 15 In the study, HHLL constructs with different connector lengths (e.g., 444 and 555 constructs) provided stability comparable to the WT bispecific constructs. Although the HMW of HHLL was higher at 1X freeze-thaw cycles, the HMW decreased uniformly throughout the freeze-thaw cycle, indicating that the HHLL form is compatible with freeze-thaw cycles.

[0197] Example 7 Cell viability assay Target cell viability was determined by quantification of constitutively expressed firefly luciferase using the Steady-Glo luciferase assay system (Promega). Briefly, in this assay, HuT-78 cells (a CD3-expressing human skin T-cell lymphocyte line) were co-incubated with OVCAR-8-Luc cells (a mesothelin-expressing human ovarian cancer cell line) and engineered to constitutively express luciferase as a marker of cell number and viability. Msln BiTE protein was triple-diluted at concentrations ranging from 0.01 ng / mL to 5 ng / mL into 96-well, full-area, flat-bottomed, tissue culture-treated, sterile, white polystyrene plates (Costar, #3917). Cells were added to the plates at a ratio of 10:1 (HutT-78 T cells to OVCAR-8-Luc Msln-expressing cells) and incubated for at least 24 hours prior to the addition of Steady-Glo reagent. Following Promega's protocol, reconstituted Steady-Glo reagent (25 µL per well) was added, and the assay plate was incubated at room temperature for 30 min. Quantification of luminescence was performed using an EnVision multi-label reader (PerkinElmer) with an ultrasensitive luminescence detector. Data were normalized relative to 100% using Msln-WT as a reference / control.

[0198] The activity of Flt3 and Msln HHLL BiTE groups in T cell-mediated cell killing was determined using this cell-based cytotoxicity assay. Human T cells incubated with either Flt3 or BiTE WT and HHLL molecules and human ovarian cancer cell lines expressing both Flt3 and Msln markers, as well as a luciferase reporter gene, were used in this assay. For simplicity, potency data were normalized relative to their respective WT HLE-BiTE values. Flt3-HHLL BiTEs retained similar potency to WT. Msln-HHLL BiTEs showed an overall potency reduction of approximately 75% compared to WT-Msln HLE-BiTEs. The results of this experiment are depicted in… Figure 13 middle.

[0199] Every reference cited in this article is incorporated in full by way of citation for all purposes.

[0200] This invention is not limited in scope to the specific embodiments described herein, which are intended as individual illustrations of various embodiments of the invention and functionally equivalent methods and components. In fact, various modifications to the invention will become apparent to those skilled in the art from the foregoing description and drawings, in addition to those shown and described herein. Such modifications are intended to be included within the scope of the claims.

[0201] sequence Exemplary connector sequence Note that the following sequence contains the N-terminal signal peptide that was removed during expression. The signal peptide sequence is: MDMRVPAQLLGLLLLWLRGARC (SEQ ID NO: 52) Target 1: dll3_332 (SEQ ID NO: 17) Target 2: dll3_444 (SEQ ID NO: 18) Target 3: dll3_555 (SEQ ID NO: 19) Target 4: dll3_WT_complete_HLHL_scFC from Flt3 (SEQ ID NO: 20) Target 5: FLT3_AT_Kozak (WT) (SEQ ID NO: 21) Target 6: FLT3_233 (SEQ ID NO: 22) Target 7: FLT3_444 (SEQ ID NO: 23) Target 8: FLT3_555 (SEQ ID NO: 24) Target 9: MSLN_444_without scFC (SEQ ID NO: 25) Target 10: MSLN_H1L1H2L2_None_scFC (WT) (SEQ ID NO: 26) Target 11: MSLN_ _H1L2L1H2_ No SCFC (SEQ ID NO: 27) Target 12: MSLN_L1H2H1L2_without scFC (SEQ ID NO: 28) Target 13: MSLN L1L2H1H2_without scFC (SEQ ID NO: 29) Target 14: MSLN_L2H1H2L1_without scFC (SEQ ID NO: 30) Target 15: MSLN_233 (SEQ ID NO: 31) Target 16: MSLN 355 (SEQ ID NO: 32) Target 17: MSLN_444 (SEQ ID NO: 33) Target 18: MSLN_555 (SEQ ID NO: 34) Target 19: MSLN_with scFC(WT) (SEQ ID NO: 35) Fc area (36-39) The amino acid sequence of mature human CD3ε SEQ ID NO: 40 The amino acid sequence of mature CD3ε in cynomolgus monkeys (SEQ ID NO: 41). The amino acid sequence of the extracellular domain of human CD3ε (SEQ ID NO: 42). Human CD3ε SEQ ID NO: 43 amino acids 1-27 qdgneemgg itqtpykvsi sgttvilt DLL3 (SEQ ID NO: 44) DLL3 Light (SEQ ID NO: 45) FLT3 heavy (SEQ ID NO: 46) FLT3 light (SEQ ID NO: 47) MSLN heavy (SEQ ID NO: 48) MSLN (SEQ ID NO: 49) CD3 heavy (SEQ ID NO: 50) CD3 Light (SEQ ID NO: 51) .

Claims

1. A bispecific binding construct comprising a polypeptide chain having an amino acid sequence of the formula VH1-L1-VH2-L2-VL1-L3-VL2, wherein VH1 and VH2 are variable regions of the immunoglobulin heavy chain, VL1 and VL2 are variable regions of the immunoglobulin light chain, and L1, L2, and L3 are linkers, wherein L1, L2, and L3 are 25-30 amino acids, and L3 are 25-30 amino acids, wherein the linker is SEQ ID NO: The bispecific binding construct binds to a human T-cell receptor (TCR)-CD3 complex on a T cell and to a target cell, wherein the target cell is a cancer cell, wherein the target is Dll3, and wherein the bispecific binding construct includes a half-life extension portion following VL2, the half-life extension portion including an additional linker (L4) and a single-chain immunoglobulin Fc region (scFc) from a human IgG1, IgG2, or IgG4 antibody.

2. The bispecific binding construct of claim 1, wherein the scFc polypeptide chain comprises one or more alterations that inhibit Fcγ receptor (FcγR) binding and / or one or more alterations that prolong the half-life.

3. The bispecific binding construct of claim 1, wherein VH1, VH2, VL1, and VL2 all have different sequences.

4. The bispecific binding construct of claim 1, wherein the VH2 sequence comprises SEQ ID NO: 50 and the VL2 sequence comprises SEQ ID NO: 51, and the VH1 sequence comprises SEQ ID NO: 44 and the VL1 sequence comprises SEQ ID NO:

45.

5. The bispecific binding construct of claim 1, wherein L1, L2, and L3 have different lengths.

6. The bispecific binding construct of claim 1, wherein L1, L2, and L3 are of equal length.

7. The bispecific binding construct of claim 1, wherein L1 and L2 are of the same length.

8. The bispecific binding construct of claim 1, wherein L1 and L3 are of the same length.

9. The bispecific binding construct of claim 1, wherein L2 and L3 are of the same length.

10. The bispecific binding construct of claim 1, wherein the bispecific binding construct exhibits enhanced stability compared to a bispecific binding construct having the formula VH1-L1-VL1-L2-VH2-L3-VL2.

11. The bispecific antibody construct of claim 1, wherein the bispecific antibody exhibits increased in vitro expression compared to a bispecific binding construct having the formula VH1-L1-VL1-L2-VH2-L3-VL2.

12. The bispecific binding construct of claim 1, wherein the effector cellular protein is the CD3ε chain.

13. A nucleic acid encoding a bispecific binding construct as described in claims 1-12.

14. A vector comprising the nucleic acid as described in claim 13.

15. A host cell comprising the vector as described in claim 14.

16. A method for manufacturing a bispecific binding construct as claimed in claim 1, the method comprising (1) culturing host cells under conditions expressing the bispecific binding construct, and (2) recovering the bispecific binding construct from a cell cluster or cell culture supernatant, wherein the host cells contain one or more nucleic acids encoding the bispecific binding construct as claimed in any one of claims 1-7.

17. Use of the bispecific binding construct as described in any one of claims 1-7 in the manufacture of a medicament for treating cancer.

18. A pharmaceutical composition comprising the bispecific binding construct as described in any one of claims 1-12.

Images